{"id": "93d6fdd11d", "schema_version": "1.0", "status": "approved", "created_at": "2026-10-28T19:30:53Z", "approved_at": "2026-10-28T19:40:19Z", "paper": {"title": "Risk of out-of-hospital cardiac arrest among sepsis survivors in Taiwan: analysis of a nationwide population-based cohort", "authors": "Hsu WT, Sherrod CF, Tehrani B, Papaila A, Porta L, Hsu TC, Sheng WH, Lee CC.", "year": "2021", "journal": "BMJ open", "doi": "10.1136/bmjopen-2021-051502", "pmcid": "PMC8442105", "pmid": "34521676", "license": "cc by-nc"}, "study": {"study_design": "cohort", "population": "Patients with sepsis discharged from hospitals in Taiwan between 2000 and 2013, followed for up to 1 year post-discharge.", "exposure": {"name": "Sepsis", "definition": "Sepsis identified using ICD-9-CM codes for sepsis combined with acute organ dysfunction according to Martin criteria and Sepsis-3 definition."}, "outcome": {"name": "Out-of-hospital cardiac arrest (OHCA)", "definition": "OHCA identified by ICD-9-CM codes 427.5, 798, or procedure code 99.60 within 365 days of discharge."}, "reported_effect": {"measure": "HR", "value": "1.66", "ci": "1.27-2.16", "adjusted": true}}, "dag": {"nodes": [{"id": "sepsis", "label": "Sepsis", "role": "exposure", "adjusted": false, "snippet": "sepsis patients", "ai_suggested": false}, {"id": "ohca", "label": "Out-of-hospital cardiac arrest", "role": "outcome", "adjusted": false, "snippet": "OHCA events within 30 days", "ai_suggested": false}, {"id": "age", "label": "Age", "role": "confounder", "adjusted": true, "snippet": "matched according to admission date, 5-year age group", "ai_suggested": false}, {"id": "sex", "label": "Sex", "role": "confounder", "adjusted": true, "snippet": "matched according to ... sex", "ai_suggested": false}, {"id": "charlson", "label": "Charlson Comorbidity Score", "role": "confounder", "adjusted": true, "snippet": "matched according to Charlson Comorbidity Score quartile", "ai_suggested": false}, {"id": "cardiovascular_disease", "label": "Cardiovascular disease", "role": "confounder", "adjusted": true, "snippet": "presence of cardiovascular disease", "ai_suggested": false}, {"id": "diabetes", "label": "Diabetes", "role": "confounder", "adjusted": true, "snippet": "presence of diabetes", "ai_suggested": false}, {"id": "infection_source", "label": "Infection source", "role": "confounder", "adjusted": true, "snippet": "infection source", "ai_suggested": false}, {"id": "septic_shock", "label": "Septic shock", "role": "confounder", "adjusted": true, "snippet": "patients with septic shock", "ai_suggested": false}, {"id": "renal_dysfunction", "label": "Renal dysfunction", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "prior_arrhythmia_heart_failure_severity", "label": "Prior arrhythmia / heart failure severity", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "socioeconomic_status_income_urbanization", "label": "Socioeconomic status (income/urbanization)", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}], "edges": [{"source": "sepsis", "target": "ohca", "ai_suggested": false}, {"source": "age", "target": "sepsis", "ai_suggested": false}, {"source": "age", "target": "ohca", "ai_suggested": false}, {"source": "sex", "target": "sepsis", "ai_suggested": false}, {"source": "sex", "target": "ohca", "ai_suggested": false}, {"source": "charlson", "target": "sepsis", "ai_suggested": false}, {"source": "charlson", "target": "ohca", "ai_suggested": false}, {"source": "cardiovascular_disease", "target": "sepsis", "ai_suggested": false}, {"source": "cardiovascular_disease", "target": "ohca", "ai_suggested": false}, {"source": "diabetes", "target": "sepsis", "ai_suggested": false}, {"source": "diabetes", "target": "ohca", "ai_suggested": false}, {"source": "infection_source", "target": "sepsis", "ai_suggested": false}, {"source": "infection_source", "target": "ohca", "ai_suggested": false}, {"source": "septic_shock", "target": "sepsis", "ai_suggested": false}, {"source": "septic_shock", "target": "ohca", "ai_suggested": false}, {"source": "renal_dysfunction", "target": "sepsis", "ai_suggested": true}, {"source": "renal_dysfunction", "target": "ohca", "ai_suggested": true}, {"source": "prior_arrhythmia_heart_failure_severity", "target": "sepsis", "ai_suggested": true}, {"source": "prior_arrhythmia_heart_failure_severity", "target": "ohca", "ai_suggested": true}, {"source": "socioeconomic_status_income_urbanization", "target": "sepsis", "ai_suggested": true}, {"source": "socioeconomic_status_income_urbanization", "target": "ohca", "ai_suggested": true}]}, "dagitty": "dag {\nsepsis [exposure]\nohca [outcome]\nage [adjusted]\nsex [adjusted]\ncharlson [adjusted]\ncardiovascular_disease [adjusted]\ndiabetes [adjusted]\ninfection_source [adjusted]\nseptic_shock [adjusted]\nrenal_dysfunction\nprior_arrhythmia_heart_failure_severity\nsocioeconomic_status_income_urbanization\nsepsis -> ohca\nage -> sepsis\nage -> ohca\nsex -> sepsis\nsex -> ohca\ncharlson -> sepsis\ncharlson -> ohca\ncardiovascular_disease -> sepsis\ncardiovascular_disease -> ohca\ndiabetes -> sepsis\ndiabetes -> ohca\ninfection_source -> sepsis\ninfection_source -> ohca\nseptic_shock -> sepsis\nseptic_shock -> ohca\nrenal_dysfunction -> sepsis\nrenal_dysfunction -> ohca\nprior_arrhythmia_heart_failure_severity -> sepsis\nprior_arrhythmia_heart_failure_severity -> ohca\nsocioeconomic_status_income_urbanization -> sepsis\nsocioeconomic_status_income_urbanization -> ohca\n}", "open_backdoors": ["Renal dysfunction", "Prior arrhythmia / heart failure severity", "Socioeconomic status (income/urbanization)"], "critique": {"missing_confounders": [{"label": "Renal dysfunction", "rationale": "Chronic kidney disease increases susceptibility to sepsis and independently predisposes to arrhythmic cardiac arrest via electrolyte shifts and uremic cardiomyopathy.", "edges_to_add": [{"source": "Renal dysfunction", "target": "sepsis", "ai_suggested": false}, {"source": "Renal dysfunction", "target": "ohca", "ai_suggested": false}], "evidence": [{"title": "Renal dysfunction is common following resuscitation from out-of-hospital cardiac arrest", "url": "https://www.sciencedirect.com/science/article/abs/pii/S0300957213002219"}, {"title": "Acute kidney injury after out-of-hospital cardiac arrest | Critical Care | Springer Nature Link", "url": "https://link.springer.com/article/10.1186/s13054-024-04936-w"}, {"title": "Seeking a Treatable Cause of Out-of-Hospital Cardiac Arrest during and after Resuscitation", "url": "https://www.mdpi.com/2077-0383/13/19/5804"}], "status": "accepted"}, {"label": "QT\u2011prolonging medication use", "rationale": "Patients with sepsis often receive antibiotics (e.g., macrolides, fluoroquinolones) or antipsychotics that prolong QT interval, raising OHCA risk; prescribing patterns also correlate with sepsis severity.", "edges_to_add": [{"source": "QT\u2011prolonging medication use", "target": "sepsis", "ai_suggested": false}, {"source": "QT\u2011prolonging medication use", "target": "ohca", "ai_suggested": false}], "evidence": [{"title": "Out-of-hospital cardiac arrest and differential risk of cardiac and non-cardiac QT-prolonging drugs in 37 000 cases - University of Southern Denmark", "url": "https://portal.findresearcher.sdu.dk/en/publications/out-of-hospital-cardiac-arrest-and-differential-risk-of-cardiac-a"}, {"title": "Cardiac Arrest in a Woman with UTI: A Case of QT Prolongation | PSNet", "url": "https://psnet.ahrq.gov/web-mm/cardiac-arrest-woman-uti-case-qt-prolongation"}, {"title": "Understanding New-Onset QT Prolongation in Sepsis: Characteristics, Outcomes and Predictors | Physicians Weekly", "url": "https://www.physiciansweekly.com/post/understanding-new-onset-qt-prolongation-in-sepsis-characteristics-outcomes-and-predictors"}], "status": "rejected"}, {"label": "Prior arrhythmia / heart failure severity", "rationale": "A history of ventricular arrhythmias or severe heart failure raises both the likelihood of developing sepsis (due to reduced reserve) and the risk of OHCA after discharge.", "edges_to_add": [{"source": "Prior arrhythmia / heart failure severity", "target": "sepsis", "ai_suggested": false}, {"source": "Prior arrhythmia / heart failure severity", "target": "ohca", "ai_suggested": false}], "evidence": [{"title": "Severe sepsis and cardiac arrhythmias", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC5787725"}, {"title": "Risk factors and outcome of extracorporeal cardiopulmonary resuscitation patients with out-of-hospital cardiac arrest and in-hospital cardiac arrest: a nationwide prospective and observational study o", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC12335052"}, {"title": "OHCA LHC - What is Out of Hospital Cardiac Arrest", "url": "https://ohca.med.umich.edu/about-us/what-is-out-of-hospital-cardiac-arrest"}], "status": "accepted"}, {"label": "Socioeconomic status (income/urbanization)", "rationale": "Lower income or rural residence is linked to higher sepsis incidence (limited access to care) and higher OHCA risk (delayed emergency response, fewer resuscitative resources).", "edges_to_add": [{"source": "Socioeconomic status", "target": "sepsis", "ai_suggested": false}, {"source": "Socioeconomic status", "target": "ohca", "ai_suggested": false}], "evidence": [{"title": "Socioeconomic Status and Out-Of-Hospital Cardiac Arrest: A ...", "url": "https://www.heartlungcirc.org/article/S1443-9506(25)01172-2/abstract"}, {"title": "Racial and Socioeconomic Disparities in Out-Of-Hospital ...Mediators of the Association Between Socioeconomic Status and ...Association Between Income and Risk of Out-of-Hospital ...", "url": "https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8635948"}, {"title": "Relationship between socioeconomic status and incidence of ...", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC7577091"}], "status": "accepted"}, {"label": "Sepsis severity biomarkers (e.g., lactate, SOFA score)", "rationale": "Higher lactate or SOFA scores indicate more severe sepsis, which both defines the exposure more strongly and directly increases myocardial injury and arrhythmogenic potential leading to OHCA.", "edges_to_add": [{"source": "Sepsis severity biomarkers", "target": "sepsis", "ai_suggested": false}, {"source": "Sepsis severity biomarkers", "target": "ohca", "ai_suggested": false}], "evidence": [{"title": "Sepsis and Septic Shock - Critical Care Medicine - Merck Manual Professional Edition", "url": "https://www.merckmanuals.com/professional/critical-care-medicine/sepsis-and-septic-shock/sepsis-and-septic-shock"}, {"title": "The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3)", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC4968574"}, {"title": "To Determine Blood Lactate Levels in Patients with Sepsis Admitted to a Respiratory Intensive Care Unit and to Correlate with their Hospital Outcomes", "url": "https://clinmedjournals.org/articles/ijccem/international-journal-of-critical-care-and-emergency-medicine-ijccem-4-045.php?jid=ijccem"}], "status": "rejected"}], "bias_risks": [{"type": "Immortal time bias", "explanation": "The cohort includes only patients who survived sepsis hospitalization to discharge; those who died in-hospital are excluded, potentially inflating the observed post\u2011discharge OHCA risk among survivors.", "severity": "medium"}, {"type": "Measurement error", "explanation": "Sepsis and OHCA are identified solely from ICD\u20119\u2011CM codes; misclassification (e.g., coding errors, lack of clinical confirmation) can bias the hazard ratio toward or away from the null.", "severity": "medium"}, {"type": "Residual confounding", "explanation": "Important severity markers (lactate, SOFA, ICU admission) and medication exposures were not adjusted for, leaving open the possibility that sicker sepsis patients drive the OHCA association.", "severity": "high"}, {"type": "Selection bias (collider stratification)", "explanation": "Propensity\u2011score matching on variables that may be consequences of sepsis (e.g., infection source, septic shock) can induce collider bias, distorting the sepsis\u2013OHCA relationship.", "severity": "medium"}, {"type": "Loss to follow\u2011up", "explanation": "Although the NHIRD is near\u2011complete, emigration or death outside the health system could lead to censoring that differs by sepsis status, potentially biasing estimates.", "severity": "low"}], "overall_assessment": "The study provides evidence of an association between sepsis hospitalization and increased risk of out\u2011of\u2011hospital cardiac arrest up to one year later, but the causal interpretation is weakened by potential immortal time bias, measurement error, and residual confounding from unmeasured sepsis severity and medication exposures. While adjustment for demographics, comorbidities, and infection characteristics reduces confounding, important prognostic factors remain uncontrolled. Consequently, the observed hazard ratios should be viewed as suggestive rather than definitive proof of a causal effect.", "label": "AI-suggested"}, "models": {"extraction": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B", "critique": "nvidia/nemotron-3-super-120b-a12b", "evidence_search": "tavily", "provider": "Nebius Token Factory"}, "review": {"by": "human", "model": null}, "dataset_license": "CC-BY-4.0"}
{"id": "2efbc1e830", "schema_version": "1.0", "status": "approved", "created_at": "2026-10-28T20:10:59Z", "approved_at": "2026-10-28T21:24:16Z", "paper": {"title": "Association of Physical Activity With Risk of Mortality Among Breast Cancer Survivors", "authors": "Chen LH, Irwin MR, Olmstead R, Haque R.", "year": "2022", "journal": "JAMA network open", "doi": "10.1001/jamanetworkopen.2022.42660", "pmcid": "PMC9672970", "pmid": "36394875", "license": "cc by"}, "study": {"study_design": "cohort", "population": "Postmenopausal breast cancer survivors diagnosed 2 or more years prior, enrolled in a California health plan between 1996 and 2012, followed until death or study end.", "sample_size": "315", "follow_up": "median 7.8 years [IQR, 7.3-8.3 years]", "exposure": {"name": "Physical activity level", "definition": "Categorized as active, moderately active, or insufficiently active using the Godin-Shephard Leisure-Time Physical Activity Questionnaire at baseline"}, "outcome": {"name": "All-cause mortality", "definition": "Death from any cause ascertained via health plan and state/national death databases"}, "reported_effect": {"measure": "HR", "value": "0.42", "ci": "0.21-0.85", "adjusted": true}}, "dag": {"nodes": [{"id": "exposure", "label": "Physical activity level", "role": "exposure", "adjusted": false, "snippet": "active, moderately active, or insufficiently active at baseline", "ai_suggested": false}, {"id": "outcome", "label": "All-cause mortality", "role": "outcome", "adjusted": false, "snippet": "death from any cause ascertained via health plan and state/national death databases", "ai_suggested": false}, {"id": "age_baseline", "label": "Age at baseline", "role": "confounder", "adjusted": true, "snippet": "adjusted for age at baseline", "ai_suggested": false}, {"id": "cancer_stage", "label": "Breast cancer stage", "role": "confounder", "adjusted": true, "snippet": "adjusted for breast cancer stage", "ai_suggested": false}, {"id": "fatigue", "label": "Fatigue severity", "role": "confounder", "adjusted": true, "snippet": "adjusted for fatigue", "ai_suggested": false}, {"id": "charlson_index", "label": "Charlson Comorbidity Index", "role": "confounder", "adjusted": true, "snippet": "adjusted for Charlson Comorbidity Index", "ai_suggested": false}, {"id": "years_since_diagnosis", "label": "Years since cancer diagnosis", "role": "confounder", "adjusted": true, "snippet": "adjusted for years since cancer diagnosis", "ai_suggested": false}, {"id": "race_ethnicity", "label": "Race and ethnicity", "role": "confounder", "adjusted": true, "snippet": "adjusted for self-reported race and ethnicity", "ai_suggested": false}, {"id": "insomnia", "label": "Lifetime history of insomnia", "role": "confounder", "adjusted": true, "snippet": "adjusted for lifetime history of insomnia", "ai_suggested": false}, {"id": "depression", "label": "Lifetime history of depression", "role": "confounder", "adjusted": true, "snippet": "adjusted for lifetime history of depression", "ai_suggested": false}, {"id": "adjuvant_treatments", "label": "Adjuvant cancer treatments", "role": "confounder", "adjusted": true, "snippet": "adjusted for adjuvant cancer treatments (endocrine, chemotherapy and radiation)", "ai_suggested": false}, {"id": "socioeconomic_status", "label": "Socioeconomic status", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "smoking_status", "label": "Smoking status", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}], "edges": [{"source": "exposure", "target": "outcome", "ai_suggested": false}, {"source": "age_baseline", "target": "exposure", "ai_suggested": false}, {"source": "age_baseline", "target": "outcome", "ai_suggested": false}, {"source": "cancer_stage", "target": "exposure", "ai_suggested": false}, {"source": "cancer_stage", "target": "outcome", "ai_suggested": false}, {"source": "fatigue", "target": "exposure", "ai_suggested": false}, {"source": "fatigue", "target": "outcome", "ai_suggested": false}, {"source": "charlson_index", "target": "exposure", "ai_suggested": false}, {"source": "charlson_index", "target": "outcome", "ai_suggested": false}, {"source": "years_since_diagnosis", "target": "exposure", "ai_suggested": false}, {"source": "years_since_diagnosis", "target": "outcome", "ai_suggested": false}, {"source": "race_ethnicity", "target": "exposure", "ai_suggested": false}, {"source": "race_ethnicity", "target": "outcome", "ai_suggested": false}, {"source": "insomnia", "target": "exposure", "ai_suggested": false}, {"source": "insomnia", "target": "outcome", "ai_suggested": false}, {"source": "depression", "target": "exposure", "ai_suggested": false}, {"source": "depression", "target": "outcome", "ai_suggested": false}, {"source": "adjuvant_treatments", "target": "exposure", "ai_suggested": false}, {"source": "adjuvant_treatments", "target": "outcome", "ai_suggested": false}, {"source": "socioeconomic_status", "target": "exposure", "ai_suggested": true}, {"source": "socioeconomic_status", "target": "outcome", "ai_suggested": true}, {"source": "smoking_status", "target": "exposure", "ai_suggested": true}, {"source": "smoking_status", "target": "outcome", "ai_suggested": true}]}, "dagitty": "dag {\nexposure [exposure]\noutcome [outcome]\nage_baseline [adjusted]\ncancer_stage [adjusted]\nfatigue [adjusted]\ncharlson_index [adjusted]\nyears_since_diagnosis [adjusted]\nrace_ethnicity [adjusted]\ninsomnia [adjusted]\ndepression [adjusted]\nadjuvant_treatments [adjusted]\nsocioeconomic_status\nsmoking_status\nexposure -> outcome\nage_baseline -> exposure\nage_baseline -> outcome\ncancer_stage -> exposure\ncancer_stage -> outcome\nfatigue -> exposure\nfatigue -> outcome\ncharlson_index -> exposure\ncharlson_index -> outcome\nyears_since_diagnosis -> exposure\nyears_since_diagnosis -> outcome\nrace_ethnicity -> exposure\nrace_ethnicity -> outcome\ninsomnia -> exposure\ninsomnia -> outcome\ndepression -> exposure\ndepression -> outcome\nadjuvant_treatments -> exposure\nadjuvant_treatments -> outcome\nsocioeconomic_status -> exposure\nsocioeconomic_status -> outcome\nsmoking_status -> exposure\nsmoking_status -> outcome\n}", "open_backdoors": ["Socioeconomic status", "Smoking status"], "critique": {"missing_confounders": [{"label": "Socioeconomic status", "rationale": "Lower income or education limits opportunities for physical activity and is associated with higher mortality through multiple pathways.", "edges_to_add": [{"source": "Socioeconomic status", "target": "exposure", "ai_suggested": false}, {"source": "Socioeconomic status", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Changes in physical activity, associated with lower mortality", "url": "https://sciencemediacentre.es/en/small-and-feasible-changes-physical-activity-are-associated-lower-mortality-risk"}, {"title": "Socioeconomic status may influence the association ...", "url": "https://www.2minutemedicine.com/socioeconomic-status-may-influence-the-association-between-physical-activity-and-health-outcomes"}, {"title": "Association of physical activity and socio-economic status on ...", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC11196573"}], "status": "accepted", "review_note": "Accept: Socioeconomic status influences both opportunities for physical activity and mortality risk, and is not accounted for by the adjusted variables."}, {"label": "Body mass index", "rationale": "Higher BMI reduces likelihood of being active and independently increases risk of death from cardiovascular and other causes.", "edges_to_add": [{"source": "Body mass index", "target": "exposure", "ai_suggested": false}, {"source": "Body mass index", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Physical activity is associated with lower mortality in adults with obesity: a systematic review with meta-analysis", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC11245862"}, {"title": "Even low levels of physical activity are associated with lower all-cause mortality: A cohort study of 594,000 US adults", "url": "https://www.sciencedirect.com/science/article/pii/S1047279726001481"}, {"title": "Percentage of Deaths Associated With Inadequate Physical Activity in the United States", "url": "https://www.cdc.gov/pcd/issues/2018/17_0354.htm"}], "status": "rejected", "review_note": "Reject: Body mass index may be altered by physical activity levels, making it a potential mediator rather than a pure confounder."}, {"label": "Smoking status", "rationale": "Smokers tend to be less physically active and have elevated mortality risk, especially from cancer and respiratory diseases.", "edges_to_add": [{"source": "Smoking status", "target": "exposure", "ai_suggested": false}, {"source": "Smoking status", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Changes in physical activity and all-cause mortality in COPD", "url": "https://publications.ersnet.org/content/erj/44/5/1199"}, {"title": "All-Cause Mortality Associated With Physical Activity During Leisure ...", "url": "https://fietsberaad.nl/CROWFietsberaad/media/Kennis/Bestanden/All-cause%20mortality%20associated%20with%20physical%20activity%20during%20leisure%20time,%20work,%20sports%20and%20cycling%20to%20work.pdf?ext=.pdf"}, {"title": "Joint association of smoking and physical activity with mortality in ...", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC9558130"}], "status": "accepted", "review_note": "Accept: Smoking status affects both physical activity engagement and mortality, and is not clearly downstream of physical activity in this cohort."}, {"label": "Hormone therapy use", "rationale": "Postmenopausal hormone therapy use can affect energy levels and activity patterns and influences long-term mortality outcomes.", "edges_to_add": [{"source": "Hormone therapy use", "target": "exposure", "ai_suggested": false}, {"source": "Hormone therapy use", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Physical activity reduces all-cause mortality in patients with cancer: a systematic review and meta-analysis of randomized controlled trials", "url": "https://www.sciencedirect.com/science/article/pii/S0305737226000368"}, {"title": "Menopausal Hormone Therapy and Long-term All-Cause and Cause-Specific Mortality - Wiki Journal Club", "url": "https://www.wikijournalclub.org/wiki/Menopausal_Hormone_Therapy_and_Long-term_All-Cause_and_Cause-Specific_Mortality"}, {"title": "Hormone Replacement Therapy and Cardiovascular Risk | NLA", "url": "https://www.lipid.org/resource/hormone-replacement-therapy-and-cardiovascular-risk"}], "status": "rejected", "review_note": "Reject: Hormone therapy use is likely encompassed by the adjusted 'adjuvant cancer treatments' variable, so it is not an omitted confounder."}, {"label": "Cardiovascular disease", "rationale": "Existing heart disease limits physical capacity and predicts higher all-cause mortality, potentially confounding the activity-mortality link.", "edges_to_add": [{"source": "Cardiovascular disease", "target": "exposure", "ai_suggested": false}, {"source": "Cardiovascular disease", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Frontiers | Effects of physical activity and sedentary behaviors on cardiovascular disease and the risk of all-cause mortality in overweight or obese middle-aged and older adults", "url": "https://www.frontiersin.org/journals/public-health/articles/10.3389/fpubh.2024.1302783/full"}, {"title": "Physical Activity, All-Cause and Cardiovascular Mortality, and Cardiovascular Disease", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC6527136"}, {"title": "New study finds lowest risk of death was among adults who exercised 150-600 minutes/week | American Heart Association", "url": "https://newsroom.heart.org/news/new-study-finds-lowest-risk-of-death-was-among-adults-who-exercised-150-600-minutesweek"}], "status": "rejected", "review_note": "Reject: Cardiovascular disease is already captured within the Charlson Comorbidity Index that was adjusted for, thus not a missing confounder."}], "bias_risks": [{"type": "Healthy user effect", "explanation": "Women who report higher activity may also engage in other health-promoting behaviors (e.g., better diet, adherence to medication) that lower mortality independent of activity.", "severity": "medium"}, {"type": "Measurement error", "explanation": "Physical activity was assessed once at baseline using a self\u2011report questionnaire, likely leading to non\u2011differential misclassification that biases the hazard ratio toward the null.", "severity": "medium"}, {"type": "Residual confounding", "explanation": "Important confounders such as socioeconomic status, BMI, and smoking were not measured, so the observed association may partly reflect these unmeasured factors.", "severity": "high"}, {"type": "Selection bias", "explanation": "Loss to follow\u2011up could differ by health status; sicker survivors may be more likely to discontinue plan membership, distorting the activity\u2011mortality relationship.", "severity": "low"}, {"type": "Immortal time bias", "explanation": "Exposure was defined at cohort entry (baseline) after a minimum 2\u2011year survival period, so there is no evident period where outcome could not occur before exposure classification.", "severity": "low"}], "overall_assessment": "The study controlled for a broad set of clinical and psychosocial confounders, which strengthens the causal interpretation. However, the lack of key socioeconomic, behavioral, and cardiovascular variables, combined with potential measurement error in self\u2011reported activity and possible healthy\u2011user bias, means the observed hazard ratio may be inflated. Consequently, the evidence suggests a protective association but should be interpreted cautiously pending further validation.", "label": "AI-suggested"}, "models": {"extraction": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B", "critique": "nvidia/nemotron-3-super-120b-a12b", "evidence_search": "tavily", "provider": "Nebius Token Factory"}, "review": {"by": "ai", "model": "nvidia/nemotron-3-super-120b-a12b"}, "dataset_license": "CC-BY-4.0"}
{"id": "50eb8cada5", "schema_version": "1.0", "status": "approved", "created_at": "2026-10-28T20:13:19Z", "approved_at": "2026-10-28T21:24:23Z", "paper": {"title": "Consumption of coffee and tea and risk of developing stroke, dementia, and poststroke dementia: A cohort study in the UK Biobank", "authors": "Zhang Y, Yang H, Li S, Li WD, Wang Y.", "year": "2021", "journal": "PLoS medicine", "doi": "10.1371/journal.pmed.1003830", "pmcid": "PMC8594796", "pmid": "34784347", "license": "cc by"}, "study": {"study_design": "cohort", "population": "Adults aged 50 to 74 years from the UK Biobank", "sample_size": "365,682", "follow_up": "median 11.4 years", "exposure": {"name": "Coffee and tea consumption", "definition": "Daily cups of coffee and tea intake assessed at baseline via touchscreen questionnaire"}, "outcome": {"name": "Incident stroke and dementia", "definition": "First hospital-recorded diagnosis of stroke or dementia during follow-up"}, "reported_effect": {"measure": "HR", "value": "0.68", "ci": "0.59-0.79", "adjusted": true}}, "dag": {"nodes": [{"id": "exposure", "label": "Coffee and tea consumption", "role": "exposure", "adjusted": false, "snippet": "Coffee and tea consumption was assessed at baseline using a touchscreen questionnaire", "ai_suggested": false}, {"id": "outcome", "label": "Incident stroke and dementia", "role": "outcome", "adjusted": false, "snippet": "Incident stroke and dementia were ascertained using hospital inpatient records", "ai_suggested": false}, {"id": "sex", "label": "Sex", "role": "confounder", "adjusted": true, "snippet": "sex, age, ethnicity, qualification, income, BMI, physical activity, alcohol status, smoking status, history of cancer, history of diabetes, history of \u2026", "ai_suggested": false}, {"id": "age", "label": "Age", "role": "confounder", "adjusted": true, "snippet": "sex, age, ethnicity, qualification, income, BMI, physical activity, alcohol status, smoking status, history of cancer, history of diabetes, history of \u2026", "ai_suggested": false}, {"id": "ethnicity", "label": "Ethnicity", "role": "confounder", "adjusted": true, "snippet": "sex, age, ethnicity, qualification, income, BMI, physical activity, alcohol status, smoking status, history of cancer, history of diabetes, history of \u2026", "ai_suggested": false}, {"id": "qualification", "label": "Qualification", "role": "confounder", "adjusted": true, "snippet": "sex, age, ethnicity, qualification, income, BMI, physical activity, alcohol status, smoking status, history of cancer, history of diabetes, history of \u2026", "ai_suggested": false}, {"id": "income", "label": "Income", "role": "confounder", "adjusted": true, "snippet": "sex, age, ethnicity, qualification, income, BMI, physical activity, alcohol status, smoking status, history of cancer, history of diabetes, history of \u2026", "ai_suggested": false}, {"id": "bmi", "label": "BMI", "role": "confounder", "adjusted": true, "snippet": "sex, age, ethnicity, qualification, income, BMI, physical activity, alcohol status, smoking status, history of cancer, history of diabetes, history of \u2026", "ai_suggested": false}, {"id": "physical_activity", "label": "Physical activity", "role": "confounder", "adjusted": true, "snippet": "sex, age, ethnicity, qualification, income, BMI, physical activity, alcohol status, smoking status, history of cancer, history of diabetes, history of \u2026", "ai_suggested": false}, {"id": "alcohol_status", "label": "Alcohol status", "role": "confounder", "adjusted": true, "snippet": "sex, age, ethnicity, qualification, income, BMI, physical activity, alcohol status, smoking status, history of cancer, history of diabetes, history of \u2026", "ai_suggested": false}, {"id": "smoking_status", "label": "Smoking status", "role": "confounder", "adjusted": true, "snippet": "sex, age, ethnicity, qualification, income, BMI, physical activity, alcohol status, smoking status, history of cancer, history of diabetes, history of \u2026", "ai_suggested": false}, {"id": "diet_pattern", "label": "Diet pattern", "role": "confounder", "adjusted": true, "snippet": "sex, age, ethnicity, qualification, income, BMI, physical activity, alcohol status, smoking status, history of cancer, history of diabetes, history of \u2026", "ai_suggested": false}, {"id": "cad", "label": "History of cardiovascular arterial disease", "role": "confounder", "adjusted": true, "snippet": "history of cardiovascular arterial disease (CAD), history of diabetes, history of cancer", "ai_suggested": false}, {"id": "diabetes", "label": "History of diabetes", "role": "confounder", "adjusted": true, "snippet": "history of diabetes, history of cardiovascular arterial disease (CAD)", "ai_suggested": false}, {"id": "cancer", "label": "History of cancer", "role": "confounder", "adjusted": true, "snippet": "history of cancer, history of diabetes, history of cardiovascular arterial disease (CAD)", "ai_suggested": false}, {"id": "hdl", "label": "HDL cholesterol", "role": "confounder", "adjusted": true, "snippet": "HDL, LDL, history of cancer, history of diabetes, history of cardiovascular arterial disease (CAD)", "ai_suggested": false}, {"id": "ldl", "label": "LDL cholesterol", "role": "confounder", "adjusted": true, "snippet": "HDL, LDL, history of cancer, history of diabetes, history of cardiovascular arterial disease (CAD)", "ai_suggested": false}, {"id": "hypertension", "label": "Hypertension", "role": "confounder", "adjusted": true, "snippet": "hypertension, history of cancer, history of diabetes, history of cardiovascular arterial disease (CAD)", "ai_suggested": false}, {"id": "baseline_cognition", "label": "Baseline cognition", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "depressive_symptoms", "label": "Depressive symptoms", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "sleep_duration", "label": "Sleep duration", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}], "edges": [{"source": "exposure", "target": "outcome", "ai_suggested": false}, {"source": "sex", "target": "exposure", "ai_suggested": false}, {"source": "sex", "target": "outcome", "ai_suggested": false}, {"source": "age", "target": "exposure", "ai_suggested": false}, {"source": "age", "target": "outcome", "ai_suggested": false}, {"source": "ethnicity", "target": "exposure", "ai_suggested": false}, {"source": "ethnicity", "target": "outcome", "ai_suggested": false}, {"source": "qualification", "target": "exposure", "ai_suggested": false}, {"source": "qualification", "target": "outcome", "ai_suggested": false}, {"source": "income", "target": "exposure", "ai_suggested": false}, {"source": "income", "target": "outcome", "ai_suggested": false}, {"source": "bmi", "target": "exposure", "ai_suggested": false}, {"source": "bmi", "target": "outcome", "ai_suggested": false}, {"source": "physical_activity", "target": "exposure", "ai_suggested": false}, {"source": "physical_activity", "target": "outcome", "ai_suggested": false}, {"source": "alcohol_status", "target": "exposure", "ai_suggested": false}, {"source": "alcohol_status", "target": "outcome", "ai_suggested": false}, {"source": "smoking_status", "target": "exposure", "ai_suggested": false}, {"source": "smoking_status", "target": "outcome", "ai_suggested": false}, {"source": "diet_pattern", "target": "exposure", "ai_suggested": false}, {"source": "diet_pattern", "target": "outcome", "ai_suggested": false}, {"source": "cad", "target": "exposure", "ai_suggested": false}, {"source": "cad", "target": "outcome", "ai_suggested": false}, {"source": "diabetes", "target": "exposure", "ai_suggested": false}, {"source": "diabetes", "target": "outcome", "ai_suggested": false}, {"source": "cancer", "target": "exposure", "ai_suggested": false}, {"source": "cancer", "target": "outcome", "ai_suggested": false}, {"source": "hdl", "target": "exposure", "ai_suggested": false}, {"source": "hdl", "target": "outcome", "ai_suggested": false}, {"source": "ldl", "target": "exposure", "ai_suggested": false}, {"source": "ldl", "target": "outcome", "ai_suggested": false}, {"source": "hypertension", "target": "exposure", "ai_suggested": false}, {"source": "hypertension", "target": "outcome", "ai_suggested": false}, {"source": "baseline_cognition", "target": "exposure", "ai_suggested": true}, {"source": "baseline_cognition", "target": "outcome", "ai_suggested": true}, {"source": "depressive_symptoms", "target": "exposure", "ai_suggested": true}, {"source": "depressive_symptoms", "target": "outcome", "ai_suggested": true}, {"source": "sleep_duration", "target": "exposure", "ai_suggested": true}, {"source": "sleep_duration", "target": "outcome", "ai_suggested": true}]}, "dagitty": "dag {\nexposure [exposure]\noutcome [outcome]\nsex [adjusted]\nage [adjusted]\nethnicity [adjusted]\nqualification [adjusted]\nincome [adjusted]\nbmi [adjusted]\nphysical_activity [adjusted]\nalcohol_status [adjusted]\nsmoking_status [adjusted]\ndiet_pattern [adjusted]\ncad [adjusted]\ndiabetes [adjusted]\ncancer [adjusted]\nhdl [adjusted]\nldl [adjusted]\nhypertension [adjusted]\nbaseline_cognition\ndepressive_symptoms\nsleep_duration\nexposure -> outcome\nsex -> exposure\nsex -> outcome\nage -> exposure\nage -> outcome\nethnicity -> exposure\nethnicity -> outcome\nqualification -> exposure\nqualification -> outcome\nincome -> exposure\nincome -> outcome\nbmi -> exposure\nbmi -> outcome\nphysical_activity -> exposure\nphysical_activity -> outcome\nalcohol_status -> exposure\nalcohol_status -> outcome\nsmoking_status -> exposure\nsmoking_status -> outcome\ndiet_pattern -> exposure\ndiet_pattern -> outcome\ncad -> exposure\ncad -> outcome\ndiabetes -> exposure\ndiabetes -> outcome\ncancer -> exposure\ncancer -> outcome\nhdl -> exposure\nhdl -> outcome\nldl -> exposure\nldl -> outcome\nhypertension -> exposure\nhypertension -> outcome\nbaseline_cognition -> exposure\nbaseline_cognition -> outcome\ndepressive_symptoms -> exposure\ndepressive_symptoms -> outcome\nsleep_duration -> exposure\nsleep_duration -> outcome\n}", "open_backdoors": ["Baseline cognition", "Depressive symptoms", "Sleep duration"], "critique": {"missing_confounders": [{"label": "APOE genotype", "rationale": "The APOE \u03b54 allele increases risk of dementia and may influence coffee/tea metabolism or preference.", "edges_to_add": [{"source": "APOE genotype", "target": "exposure", "ai_suggested": false}, {"source": "APOE genotype", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Frontiers | Moderate tea consumption and dementia-related neuroimaging markers", "url": "https://www.frontiersin.org/journals/neurology/articles/10.3389/fneur.2025.1634621/full"}, {"title": "The Association between Coffee and Tea Consumption at Midlife and Risk of Dementia Later in Life: The HUNT Study", "url": "https://www.mdpi.com/2072-6643/15/11/2469"}, {"title": "Association between coffee and tea consumption and the risk of dementia in individuals with hypertension: a prospective cohort study | Scientific Reports", "url": "https://www.nature.com/articles/s41598-024-71426-y"}], "status": "rejected", "review_note": "Although APOE \u03b54 strongly predicts dementia, there is insufficient evidence that it influences coffee/tea consumption in UK Biobank adults, so it is not a plausible common cause."}, {"label": "Baseline cognition", "rationale": "Lower baseline cognitive function can affect beverage choices and predicts future dementia.", "edges_to_add": [{"source": "Baseline cognition", "target": "exposure", "ai_suggested": false}, {"source": "Baseline cognition", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Is Drinking Coffee and Tea Associated with Reduced Rates of Stroke and Dementia?", "url": "https://www.medscape.com/viewarticle/drinking-coffee-and-tea-associated-reduced-rates-stroke-and-2021a1002cgn"}, {"title": "Research Suggests Caffeinated Coffee or Tea Could Reduce Dementia Risk | Pharmacy Times", "url": "https://www.pharmacytimes.com/view/research-suggests-caffeinated-coffee-or-tea-could-reduce-dementia-risk"}, {"title": "Association between coffee and tea consumption and the risk of dementia in individuals with hypertension: a prospective cohort study | Scientific Reports", "url": "https://www.nature.com/articles/s41598-024-71426-y"}], "status": "accepted", "review_note": "Lower baseline cognition can shape beverage preferences and independently increase risk of later stroke/dementia, meeting confounder criteria and not already adjusted for."}, {"label": "Depressive symptoms", "rationale": "Depression is linked to altered coffee/tea intake and independently raises risk of stroke and dementia.", "edges_to_add": [{"source": "Depressive symptoms", "target": "exposure", "ai_suggested": false}, {"source": "Depressive symptoms", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Coffee and Dementia: What the Research Says", "url": "https://creyos.com/blog/coffee-and-dementia"}, {"title": "Coffee and Tea Drinking May Be Associated With Reduced Rates of Stroke and Dementia - Neuroscience News", "url": "https://neurosciencenews.com/caffeine-stroke-dementia-19662"}, {"title": "Drinking Coffee and Tea May Lower Stroke and Dementia Risk - Psychiatry Advisor", "url": "https://www.psychiatryadvisor.com/news/drinking-coffee-and-tea-may-lower-stroke-and-dementia"}], "status": "accepted", "review_note": "Depressive symptoms affect coffee/tea choices and are established risk factors for stroke and dementia, making them a plausible confounder not captured in the adjustment set."}, {"label": "Sleep duration", "rationale": "Both short and long sleep duration influence beverage consumption and are risk factors for stroke and dementia.", "edges_to_add": [{"source": "Sleep duration", "target": "exposure", "ai_suggested": false}, {"source": "Sleep duration", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Tea consumption and risk of incident dementia: A prospective cohort study of 377 592 UK Biobank participants | Translational Psychiatry", "url": "https://www.nature.com/articles/s41398-022-01923-z"}, {"title": "Drinking Coffee and Tea May Lower Stroke and Dementia Risk - Psychiatry Advisor", "url": "https://www.psychiatryadvisor.com/news/drinking-coffee-and-tea-may-lower-stroke-and-dementia"}, {"title": "Drinking both coffee and tea linked to lower risks for stroke and dementia - Harvard Health", "url": "https://www.health.harvard.edu/mind-and-mood/drinking-both-coffee-and-tea-linked-to-lower-risks-for-stroke-and-dementia"}], "status": "accepted", "review_note": "Sleep duration influences both coffee/tea intake (e.g., via fatigue) and stroke/dementia risk, satisfying the confounder conditions and absent from the adjusted list."}], "bias_risks": [{"type": "measurement error", "explanation": "Coffee and tea intake was self\u2011reported once at baseline via touchscreen questionnaire, which may not capture long\u2011term or changing consumption patterns, leading to nondifferential misclassification and bias toward the null.", "severity": "medium"}, {"type": "selection bias", "explanation": "UK Biobank volunteers are healthier, more educated, and have lower prevalence of smoking and obesity than the general UK population, potentially inducing a healthy volunteer bias that distorts the exposure\u2013outcome relationship.", "severity": "medium"}, {"type": "residual confounding", "explanation": "Despite adjustment for many covariates, unmeasured factors such as genetic predisposition (APOE), baseline cognitive function, depressive symptoms, and sleep habits may still confound the observed association.", "severity": "medium"}, {"type": "information bias (outcome ascertainment)", "explanation": "Stroke and dementia were identified from hospital records; milder or outpatient\u2011managed events may be missed, causing differential under\u2011ascertainment if related to beverage consumption.", "severity": "low"}, {"type": "loss to follow-up", "explanation": "Although follow\u2011up was high, some participants may have withdrawn or died from unrelated causes, and if dropout is related to both exposure and outcome, it could bias results.", "severity": "low"}], "overall_assessment": "The study controlled for a wide range of demographic, lifestyle, and clinical covariates, making major confounding unlikely. However, residual confounding from unmeasured genetic, cognitive, mental health, and sleep factors remains possible, and the exposure measurement is limited to a single baseline self\u2011report. Combined with the healthy volunteer selection of UK Biobank, these issues suggest that the observed inverse association should be interpreted cautiously rather than as definitive proof of a causal protective effect.", "label": "AI-suggested"}, "models": {"extraction": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B", "critique": "nvidia/nemotron-3-super-120b-a12b", "evidence_search": "tavily", "provider": "Nebius Token Factory"}, "review": {"by": "ai", "model": "nvidia/nemotron-3-super-120b-a12b"}, "dataset_license": "CC-BY-4.0"}
{"id": "0efabd9dcc", "schema_version": "1.0", "status": "approved", "created_at": "2026-10-28T20:17:32Z", "approved_at": "2026-10-28T21:24:38Z", "paper": {"title": "Association between de Quervain syndrome and herpes zoster: a population-based cohort study", "authors": "Hsu CY, Ke DS, Lin CL, Kao CH.", "year": "2021", "journal": "BMJ open", "doi": "10.1136/bmjopen-2020-046891", "pmcid": "PMC8704966", "pmid": "34949602", "license": "cc by-nc"}, "study": {"study_design": "cohort", "population": "Patients newly diagnosed with de Quervain syndrome (DQS) and matched controls without DQS from Taiwan's National Health Insurance Research Database, aged \u226520 years, followed from 2000 to 2013.", "sample_size": "8390 (4195 exposed, 4195 unexposed)", "follow_up": "median 6.68 years", "exposure": {"name": "de Quervain syndrome (DQS)", "definition": "Newly diagnosed DQS (ICD-9-CM code 727.04) identified from claims data"}, "outcome": {"name": "Herpes zoster (HZ) reactivation", "definition": "HZ occurrence (ICD-9-CM code 053) diagnosed during follow-up"}, "reported_effect": {"measure": "HR", "value": "1.30", "ci": "1.16-1.47", "adjusted": true}}, "dag": {"nodes": [{"id": "exposure", "label": "de Quervain syndrome (DQS)", "role": "exposure", "adjusted": false, "snippet": "newly diagnosed with DQS between 2000 and 2012", "ai_suggested": false}, {"id": "outcome", "label": "Herpes zoster (HZ) reactivation", "role": "outcome", "adjusted": false, "snippet": "HZ occurrence (ICD-9-CM code 053)", "ai_suggested": false}, {"id": "age", "label": "Age", "role": "confounder", "adjusted": true, "snippet": "matched by age, sex and index year", "ai_suggested": false}, {"id": "sex", "label": "Sex", "role": "confounder", "adjusted": true, "snippet": "matched by age, sex and index year", "ai_suggested": false}, {"id": "ckd", "label": "Chronic kidney disease", "role": "confounder", "adjusted": true, "snippet": "comorbidities of CKD, diabetes, CAD and depression", "ai_suggested": false}, {"id": "diabetes", "label": "Diabetes", "role": "confounder", "adjusted": true, "snippet": "comorbidities of CKD, diabetes, CAD and depression", "ai_suggested": false}, {"id": "cad", "label": "Coronary artery disease", "role": "confounder", "adjusted": true, "snippet": "comorbidities of CKD, diabetes, CAD and depression", "ai_suggested": false}, {"id": "depression", "label": "Depression", "role": "confounder", "adjusted": true, "snippet": "comorbidities of CKD, diabetes, CAD and depression", "ai_suggested": false}, {"id": "obesity", "label": "Obesity", "role": "confounder", "adjusted": false, "snippet": "percentages of patients with obesity, CAD and depression were higher in the DQS cohort", "ai_suggested": false}, {"id": "occupational_manual_labor", "label": "Occupational manual labor", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "systemic_corticosteroid_use", "label": "Systemic corticosteroid use", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "current_smoking", "label": "Current smoking", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "low_socioeconomic_status", "label": "Low socioeconomic status", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "hiv_infection", "label": "HIV infection", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}], "edges": [{"source": "exposure", "target": "outcome", "ai_suggested": false}, {"source": "age", "target": "exposure", "ai_suggested": false}, {"source": "age", "target": "outcome", "ai_suggested": false}, {"source": "sex", "target": "exposure", "ai_suggested": false}, {"source": "sex", "target": "outcome", "ai_suggested": false}, {"source": "ckd", "target": "exposure", "ai_suggested": false}, {"source": "ckd", "target": "outcome", "ai_suggested": false}, {"source": "diabetes", "target": "exposure", "ai_suggested": false}, {"source": "diabetes", "target": "outcome", "ai_suggested": false}, {"source": "cad", "target": "exposure", "ai_suggested": false}, {"source": "cad", "target": "outcome", "ai_suggested": false}, {"source": "depression", "target": "exposure", "ai_suggested": false}, {"source": "depression", "target": "outcome", "ai_suggested": false}, {"source": "obesity", "target": "exposure", "ai_suggested": false}, {"source": "obesity", "target": "outcome", "ai_suggested": false}, {"source": "occupational_manual_labor", "target": "exposure", "ai_suggested": true}, {"source": "occupational_manual_labor", "target": "outcome", "ai_suggested": true}, {"source": "systemic_corticosteroid_use", "target": "exposure", "ai_suggested": true}, {"source": "systemic_corticosteroid_use", "target": "outcome", "ai_suggested": true}, {"source": "current_smoking", "target": "exposure", "ai_suggested": true}, {"source": "current_smoking", "target": "outcome", "ai_suggested": true}, {"source": "low_socioeconomic_status", "target": "exposure", "ai_suggested": true}, {"source": "low_socioeconomic_status", "target": "outcome", "ai_suggested": true}, {"source": "hiv_infection", "target": "exposure", "ai_suggested": true}, {"source": "hiv_infection", "target": "outcome", "ai_suggested": true}]}, "dagitty": "dag {\nexposure [exposure]\noutcome [outcome]\nage [adjusted]\nsex [adjusted]\nckd [adjusted]\ndiabetes [adjusted]\ncad [adjusted]\ndepression [adjusted]\nobesity\noccupational_manual_labor\nsystemic_corticosteroid_use\ncurrent_smoking\nlow_socioeconomic_status\nhiv_infection\nexposure -> outcome\nage -> exposure\nage -> outcome\nsex -> exposure\nsex -> outcome\nckd -> exposure\nckd -> outcome\ndiabetes -> exposure\ndiabetes -> outcome\ncad -> exposure\ncad -> outcome\ndepression -> exposure\ndepression -> outcome\nobesity -> exposure\nobesity -> outcome\noccupational_manual_labor -> exposure\noccupational_manual_labor -> outcome\nsystemic_corticosteroid_use -> exposure\nsystemic_corticosteroid_use -> outcome\ncurrent_smoking -> exposure\ncurrent_smoking -> outcome\nlow_socioeconomic_status -> exposure\nlow_socioeconomic_status -> outcome\nhiv_infection -> exposure\nhiv_infection -> outcome\n}", "open_backdoors": ["Obesity", "Occupational manual labor", "Systemic corticosteroid use", "Current smoking", "Low socioeconomic status", "HIV infection"], "critique": {"missing_confounders": [{"label": "Occupational manual labor", "rationale": "Jobs that require repetitive wrist motions increase the likelihood of developing de Quervain syndrome. Such occupations also involve physical stress and often poorer ergonomic conditions, which can weaken immune defenses and raise susceptibility to herpes zoster reactivation.", "edges_to_add": [{"source": "Occupational manual labor", "target": "exposure", "ai_suggested": false}, {"source": "Occupational manual labor", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Time to Reconsider Occupation Induced De Quervain\u2019s Tenosynovitis: An Updated Review of Risk Factors | Published in Orthopedic Reviews", "url": "https://orthopedicreviews.openmedicalpublishing.org/article/36911-time-to-reconsider-occupation-induced-de-quervain-s-tenosynovitis-an-updated-review-of-risk-factors"}, {"title": "De Quervain Syndrome - Rheumatology and Orthopedics - Merck Manual Professional Edition", "url": "https://www.merckmanuals.com/professional/musculoskeletal-and-connective-tissue-disorders/hand-disorders/de-quervain-syndrome"}], "status": "accepted", "review_note": "Occupational manual labor plausibly increases risk of DQS via repetitive wrist strain and of HZ via physical stress/immune effects, and it is not adjusted for."}, {"label": "Systemic corticosteroid use", "rationale": "Corticosteroid therapy can weaken tendons and predispose to de Quervain syndrome while simultaneously suppressing cell\u2011mediated immunity, thereby increasing the risk of herpes zoster reactivation.", "edges_to_add": [{"source": "Systemic corticosteroid use", "target": "exposure", "ai_suggested": false}, {"source": "Systemic corticosteroid use", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Open access Original research Association between de Quervain ...", "url": "https://bmjopen.bmj.com/content/bmjopen/11/12/e046891.full.pdf"}, {"title": "(PDF) Association between de Quervain syndrome and herpes ...", "url": "https://www.researchgate.net/publication/357288255_Association_between_de_Quervain_syndrome_and_herpes_zoster_a_population-based_cohort_study"}, {"title": "Corticosteroid Use and Risk of Herpes Zoster in a Population ...", "url": "https://www.sciencedirect.com/science/article/pii/S0025619621004766"}], "status": "accepted", "review_note": "Systemic corticosteroid use may predispose to DQS by weakening tendons and increase HZ risk via immunosuppression, and it is not adjusted for."}, {"label": "Current smoking", "rationale": "Smoking impairs tendon healing and has been linked to higher rates of de Quervain syndrome; it also suppresses immune function, making herpes zoster reactivation more likely.", "edges_to_add": [{"source": "Current smoking", "target": "exposure", "ai_suggested": false}, {"source": "Current smoking", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Herpes Zoster: Risk Factors for Occurrence, Complications, and Recurrence with a Focus on Immunocompromised Patients", "url": "https://www.mdpi.com/2079-9721/13/3/71"}, {"title": "De Quervain Syndrome - Rheumatology and Orthopedics - Merck Manual Professional Edition", "url": "https://www.merckmanuals.com/professional/musculoskeletal-and-connective-tissue-disorders/hand-disorders/de-quervain-syndrome"}, {"title": "Time to Reconsider Occupation Induced De Quervain\u2019s Tenosynovitis: An Updated Review of Risk Factors | Published in Orthopedic Reviews", "url": "https://orthopedicreviews.openmedicalpublishing.org/article/36911-time-to-reconsider-occupation-induced-de-quervain-s-tenosynovitis-an-updated-review-of-risk-factors"}], "status": "accepted", "review_note": "Current smoking plausibly raises DQS risk through impaired tendon healing and HZ risk via immune suppression, and it is not adjusted for."}, {"label": "Low socioeconomic status", "rationale": "Lower socioeconomic status is associated with higher prevalence of manual labor jobs that predispose to de Quervain syndrome and with limited access to preventive care and greater psychosocial stress, both of which increase herpes zoster risk.", "edges_to_add": [{"source": "Low socioeconomic status", "target": "exposure", "ai_suggested": false}, {"source": "Low socioeconomic status", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Herpes zoster oticus | PPTX", "url": "https://www.slideshare.net/slideshow/herpes-zoster-oticus-44976734/44976734"}, {"title": "Ogilvie\u2019s Syndrome Following Herpes Zoster Infection: A Comprehensive Review of the Literature", "url": "https://www.cureus.com/articles/318161-ogilvies-syndrome-following-herpes-zoster-infection-a-comprehensive-review-of-the-literature"}], "status": "accepted", "review_note": "Low socioeconomic status influences DQS risk (through manual labor occupations) and HZ risk (via limited healthcare access and stress), and it is not adjusted for."}, {"label": "HIV infection", "rationale": "HIV causes profound immunosuppression, markedly raising the chance of herpes zoster reactivation, and HIV\u2011associated musculoskeletal manifestations can include tenosynovitis resembling de Quervain syndrome.", "edges_to_add": [{"source": "HIV infection", "target": "exposure", "ai_suggested": false}, {"source": "HIV infection", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Herpes zoster oticus | PPTX", "url": "https://www.slideshare.net/slideshow/herpes-zoster-oticus-44976734/44976734"}, {"title": "What causes the reactivation of herpes zoster (shingles)?", "url": "https://www.droracle.ai/articles/653329/what-causes-the-reactivation-of-herpes-zoster-shingles"}, {"title": "Ogilvie\u2019s Syndrome Following Herpes Zoster Infection: A Comprehensive Review of the Literature", "url": "https://www.cureus.com/articles/318161-ogilvies-syndrome-following-herpes-zoster-infection-a-comprehensive-review-of-the-literature"}], "status": "accepted", "review_note": "HIV infection can increase HZ risk through immunosuppression and may raise DQS likelihood via HIV\u2011associated tenosynovitis, and it is not adjusted for."}], "bias_risks": [{"type": "measurement error", "explanation": "Exposure (de Quervain syndrome) was identified solely from ICD\u20119\u2011CM code 727.04 in claims data; coding errors or lack of clinical confirmation could lead to non\u2011differential misclassification, biasing the hazard ratio toward the null.", "severity": "medium"}, {"type": "measurement error", "explanation": "Outcome (herpes zoster) relied on ICD\u20119\u2011CM code 053; miscoding or missed diagnoses (especially atypical presentations) may similarly dilute the true association.", "severity": "medium"}, {"type": "selection bias", "explanation": "Controls were matched 1:1 to cases on age, sex, and index year. Matching on index year\u2014a proxy for healthcare utilization\u2014may induce overmatching if utilization is related to both DQS risk and HZ susceptibility, potentially attenuating the estimated effect.", "severity": "low"}, {"type": "immortal time bias", "explanation": "To be classified as exposed, subjects needed a DQS diagnosis and continuous enrollment prior to that date; this creates a period during which the outcome could not occur, artificially lowering event rates in the exposed group if not properly accounted for.", "severity": "low"}, {"type": "loss to follow-up", "explanation": "The analysis assumes complete follow\u2011up via the NHIR database, but deaths, emigration, or loss of insurance coverage that are not captured could censor subjects differentially by exposure status, introducing bias.", "severity": "low"}], "overall_assessment": "The study adjusted for key demographic and comorbid factors, but important lifestyle and occupational confounders (e.g., manual labor, corticosteroid use, smoking) were not measured, leaving room for residual confounding. Additionally, reliance on claims data introduces measurement error, and the matching design may have induced subtle selection biases. Consequently, while the observed association is suggestive, the evidence for a causal effect of de Quervain syndrome on herpes zoster reactivation remains moderate at best.", "label": "AI-suggested"}, "models": {"extraction": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B", "critique": "nvidia/nemotron-3-super-120b-a12b", "evidence_search": "tavily", "provider": "Nebius Token Factory"}, "review": {"by": "ai", "model": "nvidia/nemotron-3-super-120b-a12b"}, "dataset_license": "CC-BY-4.0"}
{"id": "48e796c60f", "schema_version": "1.0", "status": "approved", "created_at": "2026-10-28T20:16:48Z", "approved_at": "2026-10-28T21:24:52Z", "paper": {"title": "Long-term exposure to particulate air pollution and black carbon in relation to natural and cause-specific mortality: a multicohort study in Sweden", "authors": "Nilsson Sommar J, Andersson EM, Andersson N, Sallsten G, Stockfelt L, Ljungman PL, Segersson D, Eneroth K, Gidhagen L, Molnar P, Wennberg P, Rosengren A, Rizzuto D, Leander K, Lager A, Magnusson PK, Johansson C, Barregard L, Bellander T, Pershagen G, Forsberg B.", "year": "2021", "journal": "BMJ open", "doi": "10.1136/bmjopen-2020-046040", "pmcid": "PMC8438896", "pmid": "34497075", "license": "cc by"}, "study": {"study_design": "cohort", "population": "Swedish adults from three cities (Gothenburg, Stockholm, Ume\u00e5) followed from 1990 to 2011", "sample_size": "68,679", "follow_up": "up to 1 year", "exposure": {"name": "PM10", "definition": "annual mean PM10 concentration at participant's address estimated using dispersion models"}, "outcome": {"name": "natural mortality", "definition": "all-cause mortality determined from the national cause of death register"}, "reported_effect": {"measure": "HR", "value": "1.17", "ci": "1.06-1.31", "adjusted": true}}, "dag": {"nodes": [{"id": "exposure", "label": "PM10 exposure", "role": "exposure", "adjusted": false, "snippet": "annual mean concentrations of PM with aerodynamic diameter \u226410 \u00b5m (PM10)", "ai_suggested": false}, {"id": "outcome", "label": "natural mortality", "role": "outcome", "adjusted": false, "snippet": "cause-specific mortality data were obtained from the national cause of death registry", "ai_suggested": false}, {"id": "sex", "label": "sex", "role": "confounder", "adjusted": true, "snippet": "cohort-specific models were adjusted for sex", "ai_suggested": false}, {"id": "calendar_year", "label": "calendar year", "role": "confounder", "adjusted": true, "snippet": "adjusted for calendar year", "ai_suggested": false}, {"id": "smoking_status", "label": "smoking status", "role": "confounder", "adjusted": true, "snippet": "adjusted for smoking status (current, former, never)", "ai_suggested": false}, {"id": "alcohol_consumption", "label": "alcohol consumption", "role": "confounder", "adjusted": true, "snippet": "adjusted for alcohol consumption", "ai_suggested": false}, {"id": "physical_activity", "label": "physical activity", "role": "confounder", "adjusted": true, "snippet": "adjusted for physical activity", "ai_suggested": false}, {"id": "marital_status", "label": "marital status", "role": "confounder", "adjusted": true, "snippet": "adjusted for marital status", "ai_suggested": false}, {"id": "socioeconomic_index", "label": "socioeconomic index", "role": "confounder", "adjusted": true, "snippet": "adjusted for socioeconomic index by occupation", "ai_suggested": false}, {"id": "education_level", "label": "education level", "role": "confounder", "adjusted": true, "snippet": "adjusted for education level", "ai_suggested": false}, {"id": "area_socioeconomic_status", "label": "area socioeconomic status", "role": "confounder", "adjusted": true, "snippet": "adjusted for area-level socioeconomic status", "ai_suggested": false}, {"id": "bmi", "label": "BMI", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}], "edges": [{"source": "exposure", "target": "outcome", "ai_suggested": false}, {"source": "sex", "target": "exposure", "ai_suggested": false}, {"source": "sex", "target": "outcome", "ai_suggested": false}, {"source": "calendar_year", "target": "exposure", "ai_suggested": false}, {"source": "calendar_year", "target": "outcome", "ai_suggested": false}, {"source": "smoking_status", "target": "exposure", "ai_suggested": false}, {"source": "smoking_status", "target": "outcome", "ai_suggested": false}, {"source": "alcohol_consumption", "target": "exposure", "ai_suggested": false}, {"source": "alcohol_consumption", "target": "outcome", "ai_suggested": false}, {"source": "physical_activity", "target": "exposure", "ai_suggested": false}, {"source": "physical_activity", "target": "outcome", "ai_suggested": false}, {"source": "marital_status", "target": "exposure", "ai_suggested": false}, {"source": "marital_status", "target": "outcome", "ai_suggested": false}, {"source": "socioeconomic_index", "target": "exposure", "ai_suggested": false}, {"source": "socioeconomic_index", "target": "outcome", "ai_suggested": false}, {"source": "education_level", "target": "exposure", "ai_suggested": false}, {"source": "education_level", "target": "outcome", "ai_suggested": false}, {"source": "area_socioeconomic_status", "target": "exposure", "ai_suggested": false}, {"source": "area_socioeconomic_status", "target": "outcome", "ai_suggested": false}, {"source": "bmi", "target": "exposure", "ai_suggested": true}, {"source": "bmi", "target": "outcome", "ai_suggested": true}]}, "dagitty": "dag {\nexposure [exposure]\noutcome [outcome]\nsex [adjusted]\ncalendar_year [adjusted]\nsmoking_status [adjusted]\nalcohol_consumption [adjusted]\nphysical_activity [adjusted]\nmarital_status [adjusted]\nsocioeconomic_index [adjusted]\neducation_level [adjusted]\narea_socioeconomic_status [adjusted]\nbmi\nexposure -> outcome\nsex -> exposure\nsex -> outcome\ncalendar_year -> exposure\ncalendar_year -> outcome\nsmoking_status -> exposure\nsmoking_status -> outcome\nalcohol_consumption -> exposure\nalcohol_consumption -> outcome\nphysical_activity -> exposure\nphysical_activity -> outcome\nmarital_status -> exposure\nmarital_status -> outcome\nsocioeconomic_index -> exposure\nsocioeconomic_index -> outcome\neducation_level -> exposure\neducation_level -> outcome\narea_socioeconomic_status -> exposure\narea_socioeconomic_status -> outcome\nbmi -> exposure\nbmi -> outcome\n}", "open_backdoors": ["BMI"], "critique": {"missing_confounders": [{"label": "BMI", "rationale": "Higher body mass index increases mortality risk and is associated with living in more deprived areas that tend to have higher PM10 levels.", "edges_to_add": [{"source": "BMI", "target": "exposure", "ai_suggested": false}, {"source": "BMI", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Confounding by ill health in the observed association between BMI and mortality: evidence from the HUNT Study using offspring BMI as an instrument", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC6005033"}, {"title": "expert reaction to study looking at BMI and all-cause mortality | Science Media Centre", "url": "https://www.sciencemediacentre.org/expert-reaction-to-study-looking-at-bmi-and-all-cause-mortality"}, {"title": "Estimating Causal Effects of Particulate Matter Regulation on ...", "url": "https://njsanders.human.cornell.edu/Particulates_and_Mortality_Epidemiology.pdf"}], "status": "accepted", "review_note": "BMI influences residential area choice (higher BMI linked to deprived areas with higher PM10) and independently predicts mortality, and it is not adjusted for nor caused by PM10."}, {"label": "Diabetes status", "rationale": "Diabetes elevates mortality risk and individuals with diabetes may reside in neighborhoods with greater traffic-related pollution.", "edges_to_add": [{"source": "Diabetes status", "target": "exposure", "ai_suggested": false}, {"source": "Diabetes status", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Exposure to particulate matters and risk of diabetes-related ...", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC12981974"}, {"title": "Diabetes status and susceptibility to the effects of PM>2.5> ...", "url": "https://profiles.wustl.edu/en/publications/diabetes-status-and-susceptibility-to-the-effects-of-pmsub25sub-e"}, {"title": "Long-term particulate matter exposure and type 2 diabetes ...", "url": "https://www.sciencedirect.com/science/article/pii/S0147651324002884"}], "status": "rejected", "review_note": "PM10 exposure can increase diabetes risk, so diabetes may lie on the causal pathway from exposure to mortality rather than being a pure confounder."}, {"label": "Baseline CVD", "rationale": "Prevalent cardiovascular disease raises mortality and may influence residential choice (e.g., moving to less polluted areas after diagnosis).", "edges_to_add": [{"source": "Baseline CVD", "target": "exposure", "ai_suggested": false}, {"source": "Baseline CVD", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Particulate Matter Pollution and Cardiovascular Disease | Cardiology Grand Rounds", "url": "https://www.youtube.com/watch?v=GDfB-KdYc7g"}, {"title": "Long-term exposure to PM2.5 and cardiovascular disease ...", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC8555193"}, {"title": "Hourly peak PM2.5 concentration associated with ...", "url": "https://meadowscenter.org/wp-content/uploads/2022/04/16871261.pdf"}], "status": "rejected", "review_note": "Prior PM10 exposure contributes to cardiovascular disease, making baseline CVD likely a mediator/consequence of exposure rather than a pure confounder."}, {"label": "Occupational exposure", "rationale": "Job\u2011related airborne pollutant exposure adds to personal PM10 exposure and occupation is linked to mortality risk via socioeconomic and lifestyle factors.", "edges_to_add": [{"source": "Occupational exposure", "target": "exposure", "ai_suggested": false}, {"source": "Occupational exposure", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Long-Term Ambient Multipollutant Exposures and Mortality", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC3040395"}, {"title": "Short-term exposure to particulate matter (PM10 and PM2.5), nitrogen dioxide (NO2), and ozone (O3) and all-cause and cause-specific mortality: Systematic review and meta-analysis", "url": "https://www.sciencedirect.com/science/article/pii/S0160412020318316"}, {"title": "Long-term exposure to NO2 and PM10 and all-cause...", "url": "https://www.ovid.com/00022707-201303000-00006"}], "status": "rejected", "review_note": "Occupational exposure adds to personal PM10 dose and is not a cause of ambient PM10 concentration, thus it does not satisfy the common\u2011cause criterion."}, {"label": "Dietary intake", "rationale": "Poor diet increases mortality and correlates with socioeconomic status and area\u2011level pollution, potentially confounding the PM10\u2013mortality association.", "edges_to_add": [{"source": "Dietary intake", "target": "exposure", "ai_suggested": false}, {"source": "Dietary intake", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "The potential role of dietary patterns in modifying the association between ambient PM2.5 exposure and mortality in elderly Hong Kong Chinese", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC13224638"}, {"title": "Ambient Particulate Matter Air Pollution Exposure and ...", "url": "https://escholarship.org/content/qt3t30d5h7/qt3t30d5h7.pdf"}, {"title": "Estimating Causal Effects of Particulate Matter Regulation on ...", "url": "https://njsanders.human.cornell.edu/Particulates_and_Mortality_Epidemiology.pdf"}], "status": "rejected", "review_note": "Diet does not cause ambient PM10 levels; any association with PM10 is mediated through socioeconomic status, which is already adjusted for."}], "bias_risks": [{"type": "measurement error", "explanation": "PM10 estimates from dispersion models may not capture indoor infiltration or personal activity patterns, leading to nondifferential misclassification of exposure and bias toward the null.", "severity": "medium"}, {"type": "healthy user effect", "explanation": "Residents of lower\u2011pollution areas may also engage in healthier behaviors (e.g., better diet, preventive care) not fully captured by the adjusted lifestyle covariates, potentially inflating the observed association.", "severity": "medium"}, {"type": "selection bias", "explanation": "Loss to follow\u2011up due to migration out of the three cities or emigration may be related to both socioeconomic status (and thus exposure) and mortality risk, threatening internal validity.", "severity": "low"}, {"type": "confounding by unmeasured comorbidities", "explanation": "Baseline diseases such as diabetes, cardiovascular disease, or obesity were not included as covariates and could confound the PM10\u2013mortality relation despite adjustment for smoking, alcohol, physical activity, and SES.", "severity": "medium"}, {"type": "lack of adjustment for green space", "explanation": "Access to parks or vegetation can reduce mortality and may be inversely correlated with traffic\u2011related PM10, leaving a potential source of residual confounding.", "severity": "low"}], "overall_assessment": "The study adjusted for a broad set of demographic, socioeconomic, and behavioral confounders, making the observed HR of 1.17 per 10 \u00b5g/m\u00b3 PM10 plausible. However, residual confounding by unmeasured health conditions (e.g., BMI, diabetes) and measurement error in exposure estimates could bias the effect toward or away from the null. Overall, the evidence supports a causal interpretation but should be interpreted with moderate caution.", "label": "AI-suggested"}, "models": {"extraction": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B", "critique": "nvidia/nemotron-3-super-120b-a12b", "evidence_search": "tavily", "provider": "Nebius Token Factory"}, "review": {"by": "ai", "model": "nvidia/nemotron-3-super-120b-a12b"}, "dataset_license": "CC-BY-4.0"}
{"id": "e6c0c01ebc", "schema_version": "1.0", "status": "approved", "created_at": "2026-10-28T20:15:59Z", "approved_at": "2026-10-28T21:25:11Z", "paper": {"title": "Gabapentin, opioids, and the risk of opioid-related death: A population-based nested case-control study", "authors": "Gomes T, Juurlink DN, Antoniou T, Mamdani MM, Paterson JM, van den Brink W.", "year": "2017", "journal": "PLoS medicine", "doi": "10.1371/journal.pmed.1002396", "pmcid": "PMC5626029", "pmid": "28972983", "license": "cc by"}, "study": {"study_design": "case-control", "population": "Adults aged 15 to 105 years receiving prescription opioids in Ontario, Canada, between August 1, 1997 and December 31, 2013, who were eligible for public drug coverage.", "sample_size": "1,256 cases and 4,619 controls", "follow_up": null, "exposure": {"name": "gabapentin use", "definition": "Concomitant gabapentin use in the 120 days preceding the index date"}, "outcome": {"name": "opioid-related death", "definition": "Death from an opioid-related cause identified via coroner investigation and postmortem toxicology"}, "reported_effect": {"measure": "OR", "value": "1.49", "ci": "1.18-1.88", "adjusted": true}}, "dag": {"nodes": [{"id": "exp_gab", "label": "gabapentin use", "role": "exposure", "adjusted": false, "snippet": "concomitant gabapentin use in the 120 days preceding the index date", "ai_suggested": false}, {"id": "outc_death", "label": "opioid-related death", "role": "outcome", "adjusted": false, "snippet": "death from an opioid-related cause", "ai_suggested": false}, {"id": "conf_age_sex", "label": "age, sex, year of index date", "role": "confounder", "adjusted": true, "snippet": "matched on age, sex, year of index date", "ai_suggested": false}, {"id": "conf_ckd", "label": "history of chronic kidney disease", "role": "confounder", "adjusted": true, "snippet": "history of chronic kidney disease", "ai_suggested": false}, {"id": "conf_opioid_dose", "label": "opioid dose", "role": "confounder", "adjusted": true, "snippet": "adjusted for opioid dose", "ai_suggested": false}, {"id": "conf_meds", "label": "recent medication use", "role": "confounder", "adjusted": true, "snippet": "adjusted for medication use in the prior 120 days", "ai_suggested": false}, {"id": "conf_cancer", "label": "prior cancer diagnosis", "role": "confounder", "adjusted": true, "snippet": "excluded individuals with prior cancer", "ai_suggested": false}, {"id": "conf_alcohol", "label": "alcohol use disorder", "role": "confounder", "adjusted": true, "snippet": "history of alcohol use disorder", "ai_suggested": false}, {"id": "conf_cci", "label": "Charlson Comorbidity Index", "role": "confounder", "adjusted": true, "snippet": "Charlson Comorbidity Index", "ai_suggested": false}, {"id": "conf_lung", "label": "chronic lung disease", "role": "confounder", "adjusted": true, "snippet": "diagnosis of chronic lung disease", "ai_suggested": false}, {"id": "conf_diabetes", "label": "diabetes", "role": "confounder", "adjusted": true, "snippet": "diabetes", "ai_suggested": false}, {"id": "conf_pharmacy", "label": "number of pharmacies", "role": "confounder", "adjusted": true, "snippet": "number of pharmacies from which each person was dispensed opioids", "ai_suggested": false}, {"id": "conf_prescriber", "label": "number of opioid prescribers", "role": "confounder", "adjusted": true, "snippet": "number of opioid prescribers in the past 6 months", "ai_suggested": false}, {"id": "conf_risk_index", "label": "disease risk index", "role": "confounder", "adjusted": true, "snippet": "matched on disease risk index", "ai_suggested": false}, {"id": "med_methadone", "label": "methadone/buprenorphine", "role": "confounder", "adjusted": true, "snippet": "methadone/buprenorphine", "ai_suggested": false}, {"id": "med_ssri", "label": "SSRI antidepressants", "role": "confounder", "adjusted": true, "snippet": "SSRI antidepressants", "ai_suggested": false}, {"id": "med_benzodiazepines", "label": "benzodiazepines", "role": "confounder", "adjusted": true, "snippet": "benzodiazepines", "ai_suggested": false}, {"id": "med_other_psychotropic", "label": "other psychotropic drugs", "role": "confounder", "adjusted": true, "snippet": "other psychotropic drugs/CNS depressants", "ai_suggested": false}, {"id": "psychiatric_illness_severity", "label": "psychiatric illness severity", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "socioeconomic_status", "label": "socioeconomic status", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "underlying_pain_condition", "label": "underlying pain condition", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "illicit_drug_use", "label": "illicit drug use", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "tobacco_use", "label": "tobacco use", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}], "edges": [{"source": "exp_gab", "target": "outc_death", "ai_suggested": false}, {"source": "conf_age_sex", "target": "exp_gab", "ai_suggested": false}, {"source": "conf_age_sex", "target": "outc_death", "ai_suggested": false}, {"source": "conf_ckd", "target": "exp_gab", "ai_suggested": false}, {"source": "conf_ckd", "target": "outc_death", "ai_suggested": false}, {"source": "conf_opioid_dose", "target": "exp_gab", "ai_suggested": false}, {"source": "conf_opioid_dose", "target": "outc_death", "ai_suggested": false}, {"source": "conf_meds", "target": "exp_gab", "ai_suggested": false}, {"source": "conf_meds", "target": "outc_death", "ai_suggested": false}, {"source": "conf_cancer", "target": "exp_gab", "ai_suggested": false}, {"source": "conf_cancer", "target": "outc_death", "ai_suggested": false}, {"source": "conf_alcohol", "target": "exp_gab", "ai_suggested": false}, {"source": "conf_alcohol", "target": "outc_death", "ai_suggested": false}, {"source": "conf_cci", "target": "exp_gab", "ai_suggested": false}, {"source": "conf_cci", "target": "outc_death", "ai_suggested": false}, {"source": "conf_lung", "target": "exp_gab", "ai_suggested": false}, {"source": "conf_lung", "target": "outc_death", "ai_suggested": false}, {"source": "conf_diabetes", "target": "exp_gab", "ai_suggested": false}, {"source": "conf_diabetes", "target": "outc_death", "ai_suggested": false}, {"source": "conf_pharmacy", "target": "exp_gab", "ai_suggested": false}, {"source": "conf_pharmacy", "target": "outc_death", "ai_suggested": false}, {"source": "conf_prescriber", "target": "exp_gab", "ai_suggested": false}, {"source": "conf_prescriber", "target": "outc_death", "ai_suggested": false}, {"source": "conf_risk_index", "target": "exp_gab", "ai_suggested": false}, {"source": "conf_risk_index", "target": "outc_death", "ai_suggested": false}, {"source": "med_methadone", "target": "exp_gab", "ai_suggested": false}, {"source": "med_methadone", "target": "outc_death", "ai_suggested": false}, {"source": "med_ssri", "target": "exp_gab", "ai_suggested": false}, {"source": "med_ssri", "target": "outc_death", "ai_suggested": false}, {"source": "med_benzodiazepines", "target": "exp_gab", "ai_suggested": false}, {"source": "med_benzodiazepines", "target": "outc_death", "ai_suggested": false}, {"source": "med_other_psychotropic", "target": "exp_gab", "ai_suggested": false}, {"source": "med_other_psychotropic", "target": "outc_death", "ai_suggested": false}, {"source": "psychiatric_illness_severity", "target": "exp_gab", "ai_suggested": true}, {"source": "psychiatric_illness_severity", "target": "outc_death", "ai_suggested": true}, {"source": "socioeconomic_status", "target": "exp_gab", "ai_suggested": true}, {"source": "socioeconomic_status", "target": "outc_death", "ai_suggested": true}, {"source": "underlying_pain_condition", "target": "exp_gab", "ai_suggested": true}, {"source": "underlying_pain_condition", "target": "outc_death", "ai_suggested": true}, {"source": "illicit_drug_use", "target": "exp_gab", "ai_suggested": true}, {"source": "illicit_drug_use", "target": "outc_death", "ai_suggested": true}, {"source": "tobacco_use", "target": "exp_gab", "ai_suggested": true}, {"source": "tobacco_use", "target": "outc_death", "ai_suggested": true}]}, "dagitty": "dag {\nexp_gab [exposure]\noutc_death [outcome]\nconf_age_sex [adjusted]\nconf_ckd [adjusted]\nconf_opioid_dose [adjusted]\nconf_meds [adjusted]\nconf_cancer [adjusted]\nconf_alcohol [adjusted]\nconf_cci [adjusted]\nconf_lung [adjusted]\nconf_diabetes [adjusted]\nconf_pharmacy [adjusted]\nconf_prescriber [adjusted]\nconf_risk_index [adjusted]\nmed_methadone [adjusted]\nmed_ssri [adjusted]\nmed_benzodiazepines [adjusted]\nmed_other_psychotropic [adjusted]\npsychiatric_illness_severity\nsocioeconomic_status\nunderlying_pain_condition\nillicit_drug_use\ntobacco_use\nexp_gab -> outc_death\nconf_age_sex -> exp_gab\nconf_age_sex -> outc_death\nconf_ckd -> exp_gab\nconf_ckd -> outc_death\nconf_opioid_dose -> exp_gab\nconf_opioid_dose -> outc_death\nconf_meds -> exp_gab\nconf_meds -> outc_death\nconf_cancer -> exp_gab\nconf_cancer -> outc_death\nconf_alcohol -> exp_gab\nconf_alcohol -> outc_death\nconf_cci -> exp_gab\nconf_cci -> outc_death\nconf_lung -> exp_gab\nconf_lung -> outc_death\nconf_diabetes -> exp_gab\nconf_diabetes -> outc_death\nconf_pharmacy -> exp_gab\nconf_pharmacy -> outc_death\nconf_prescriber -> exp_gab\nconf_prescriber -> outc_death\nconf_risk_index -> exp_gab\nconf_risk_index -> outc_death\nmed_methadone -> exp_gab\nmed_methadone -> outc_death\nmed_ssri -> exp_gab\nmed_ssri -> outc_death\nmed_benzodiazepines -> exp_gab\nmed_benzodiazepines -> outc_death\nmed_other_psychotropic -> exp_gab\nmed_other_psychotropic -> outc_death\npsychiatric_illness_severity -> exp_gab\npsychiatric_illness_severity -> outc_death\nsocioeconomic_status -> exp_gab\nsocioeconomic_status -> outc_death\nunderlying_pain_condition -> exp_gab\nunderlying_pain_condition -> outc_death\nillicit_drug_use -> exp_gab\nillicit_drug_use -> outc_death\ntobacco_use -> exp_gab\ntobacco_use -> outc_death\n}", "open_backdoors": ["psychiatric illness severity", "socioeconomic status", "underlying pain condition", "illicit drug use", "tobacco use"], "critique": {"missing_confounders": [{"label": "psychiatric illness severity", "rationale": "Patients with severe depression or anxiety are more likely to receive gabapentin for off\u2011label use and also have a heightened risk of opioid\u2011related death due to suicidal intent or polysubstance use.", "edges_to_add": [{"source": "psychiatric illness severity", "target": "exp_gab", "ai_suggested": false}, {"source": "psychiatric illness severity", "target": "outc_death", "ai_suggested": false}], "evidence": [{"title": "Study: Gabapentin-Related Overdose Deaths Increasing, Associated with Opioid Use | Pharmacy Times", "url": "https://www.pharmacytimes.com/view/study-gabapentin-related-overdose-deaths-increasing-associated-with-opioid-use"}, {"title": "Patients co-prescribed gabapentin and opioids were 49% ...", "url": "https://www.facebook.com/medscape/posts/patients-co-prescribed-gabapentin-and-opioids-were-49-more-likely-to-suffer-an-o/10155961446424189"}, {"title": "Gabapentin, Opioids & Risk of Opioid-Related Death", "url": "https://addiction-medicine.org/2017/10/05/gabapentin-opioids-and-risk-of-opioid-related-death"}], "status": "accepted", "review_note": "Psychiatric illness severity is a plausible common cause of gabapentin prescribing and opioid-related death and is not captured by the adjusted medication proxies."}, {"label": "socioeconomic status", "rationale": "Lower income or education correlates with both higher gabapentin prescribing (reflecting limited access to alternative pain therapies) and increased risk of opioid\u2011related death via unstable housing, reduced healthcare access, and higher substance\u2011use risk.", "edges_to_add": [{"source": "socioeconomic status", "target": "exp_gab", "ai_suggested": false}, {"source": "socioeconomic status", "target": "outc_death", "ai_suggested": false}], "evidence": [{"title": "Study: Gabapentin-Related Overdose Deaths Increasing, Associated with Opioid Use | Pharmacy Times", "url": "https://www.pharmacytimes.com/view/study-gabapentin-related-overdose-deaths-increasing-associated-with-opioid-use"}, {"title": "Notes from the Field: Trends in Gabapentin Detection and Involvement in Drug Overdose Deaths \u2014 23 States and the District of Columbia, 2019\u20132020  | MMWR", "url": "https://www.cdc.gov/mmwr/volumes/71/wr/mm7119a3.htm"}, {"title": "Tales and Trends of Gabapentin - RADARS System", "url": "https://www.radars.org/system/publications/Gabapentin%20symposium%20rmpds%202022%20-%20JCB.pdf"}], "status": "accepted", "review_note": "Socioeconomic status influences both gabapentin prescribing (via limited alternatives) and opioid-related death risk and is not adjusted for in the study."}, {"label": "underlying pain condition", "rationale": "Severe neuropathic or musculoskeletal pain drives clinicians to prescribe both gabapentin and opioids and is itself associated with higher opioid dosing and greater overdose risk.", "edges_to_add": [{"source": "underlying pain condition", "target": "exp_gab", "ai_suggested": false}, {"source": "underlying pain condition", "target": "outc_death", "ai_suggested": false}], "evidence": [{"title": "Mortality after concurrent treatment with gabapentin and opioids in older adults with spine diagnoses", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC11919562"}, {"title": "Frontiers | Risk of major adverse events associated with gabapentinoid and opioid combination therapy: A systematic review and meta-analysis", "url": "https://www.frontiersin.org/journals/pharmacology/articles/10.3389/fphar.2022.1009950/full"}, {"title": "Notes from the Field: Trends in Gabapentin Detection and Involvement in Drug Overdose Deaths \u2014 23 States and the District of Columbia, 2019\u20132020  | MMWR", "url": "https://www.cdc.gov/mmwr/volumes/71/wr/mm7119a3.htm"}], "status": "accepted", "review_note": "Underlying pain condition drives both gabapentin and opioid prescribing and increases overdose risk, and it is not accounted for among the covariates."}, {"label": "illicit drug use", "rationale": "Use of illegal opioids or stimulants increases the likelihood of gabapentin misuse to potentiate euphoric effects and independently raises the probability of fatal overdose.", "edges_to_add": [{"source": "illicit drug use", "target": "exp_gab", "ai_suggested": false}, {"source": "illicit drug use", "target": "outc_death", "ai_suggested": false}], "evidence": [{"title": "Notes from the Field: Trends in Gabapentin Detection and Involvement in Drug Overdose Deaths \u2014 23 States and the District of Columbia, 2019\u20132020  | MMWR", "url": "https://www.cdc.gov/mmwr/volumes/71/wr/mm7119a3.htm"}, {"title": "Gabapentin - HPSJ/MVHP", "url": "https://www.hpsj.com/improving-the-quality-of-care-risks-associated-with-use-of-gabapentin"}, {"title": "Patients co-prescribed gabapentin and opioids were 49% ...", "url": "https://www.facebook.com/medscape/posts/patients-co-prescribed-gabapentin-and-opioids-were-49-more-likely-to-suffer-an-o/10155961446424189"}], "status": "accepted", "review_note": "Illicit drug use precedes and increases both gabapentin misuse and opioid-related death risk, and it is not included in the adjusted set."}, {"label": "tobacco use", "rationale": "Smoking is common among patients with chronic pain leading to gabapentin prescribing and independently exacerbates respiratory depression when combined with opioids, raising death risk.", "edges_to_add": [{"source": "tobacco use", "target": "exp_gab", "ai_suggested": false}, {"source": "tobacco use", "target": "outc_death", "ai_suggested": false}], "evidence": [{"title": "Notes from the Field: Trends in Gabapentin Detection and Involvement in Drug Overdose Deaths \u2014 23 States and the District of Columbia, 2019\u20132020  | MMWR", "url": "https://www.cdc.gov/mmwr/volumes/71/wr/mm7119a3.htm"}, {"title": "Patients co-prescribed gabapentin and opioids were 49% ...", "url": "https://www.facebook.com/medscape/posts/patients-co-prescribed-gabapentin-and-opioids-were-49-more-likely-to-suffer-an-o/10155961446424189"}, {"title": "Mortality with concurrent treatment with gabapentin and ...", "url": "https://www.medrxiv.org/content/10.1101/2024.04.26.24306460v1.full-text"}], "status": "accepted", "review_note": "Tobacco use is associated with chronic pain leading to gabapentin use and independently raises opioid overdose risk, and it is not adjusted for (chronic lung disease is an incomplete proxy)."}], "bias_risks": [{"type": "confounding by indication", "explanation": "Gabapentin is often prescribed for neuropathic pain or anxiety, which also indicates higher opioid dosing and greater vulnerability to overdose; residual confounding may persist despite adjustment for opioid dose and comorbidity indices.", "severity": "medium"}, {"type": "selection bias (prevalence-incidence bias)", "explanation": "Controls were selected from opioid users alive at the index date; if gabapentin influences short\u2011term mortality, those who die early are excluded as controls, potentially inflating the observed association.", "severity": "medium"}, {"type": "measurement error (exposure misclassification)", "explanation": "Gabapentin exposure was ascertained from pharmacy dispensary records, which may not reflect actual ingestion; non\u2011adherence or over\u2011the\u2011counter use could dilute or exaggerate the true effect.", "severity": "low"}, {"type": "residual confounding (unmeasured pain severity)", "explanation": "Although the model adjusted for opioid dose and comorbidity scores, granular data on pain intensity or specific pain diagnoses were unavailable, leaving room for confounding by the underlying pain condition.", "severity": "medium"}, {"type": "immortal time bias", "explanation": "The 120\u2011day exposure window requires survival to receive gabapentin; cases who died before accumulating sufficient exposure time are automatically classified as unexposed, biasing the odds ratio away from the null.", "severity": "low"}], "overall_assessment": "The study adjusted for a broad set of demographic, clinical, and medication covariates, but important confounders such as psychiatric illness severity, socioeconomic status, pain severity, illicit drug use, and tobacco use were not measured. Additionally, selection biases related to the requirement of surviving the exposure window and potential exposure misclassification may further distort the estimate. Consequently, while the observed association is suggestive, confidence in a causal interpretation is limited.", "label": "AI-suggested"}, "models": {"extraction": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B", "critique": "nvidia/nemotron-3-super-120b-a12b", "evidence_search": "tavily", "provider": "Nebius Token Factory"}, "review": {"by": "ai", "model": "nvidia/nemotron-3-super-120b-a12b"}, "dataset_license": "CC-BY-4.0"}
{"id": "03d4b0d3bf", "schema_version": "1.0", "status": "approved", "created_at": "2026-10-28T20:15:12Z", "approved_at": "2026-10-28T21:25:24Z", "paper": {"title": "Varenicline and Risk of Self-Harm: A Nested Case-Control Study", "authors": "Tadrous M, Martins D, Yao Z, Mamdani MM, Juurlink DN, Gomes T, Antoniou T.", "year": "2016", "journal": "PloS one", "doi": "10.1371/journal.pone.0163681", "pmcid": "PMC5035025", "pmid": "27662654", "license": "cc by"}, "study": {"study_design": "case-control", "population": "Ontario adults aged 18 years and older with publicly funded drug coverage receiving bupropion or varenicline between April 1, 2011 and March 31, 2015", "sample_size": "331 cases and 5,346 matched-controls", "follow_up": "90 days after treatment", "exposure": {"name": "varenicline", "definition": "prescription of varenicline for smoking cessation identified in the Ontario Drug Benefit database"}, "outcome": {"name": "suicide or self-harm", "definition": "hospitalization or emergency department visit for suicide or non-fatal self-harm within 90 days of treatment"}, "reported_effect": {"measure": "OR", "value": "1.15", "ci": "0.71-1.87", "adjusted": true}}, "dag": {"nodes": [{"id": "exposure", "label": "Varenicline use", "role": "exposure", "adjusted": false, "snippet": "prescription of varenicline for smoking cessation", "ai_suggested": false}, {"id": "outcome", "label": "Suicide or self-harm", "role": "outcome", "adjusted": false, "snippet": "hospitalization or emergency department visit for suicide or non-fatal self-harm within 90 days", "ai_suggested": false}, {"id": "age", "label": "Age", "role": "confounder", "adjusted": true, "snippet": "matched on age (within 1 year)", "ai_suggested": false}, {"id": "sex", "label": "Sex", "role": "confounder", "adjusted": true, "snippet": "matched on sex", "ai_suggested": false}, {"id": "history_self_harm", "label": "History of self-harm", "role": "confounder", "adjusted": true, "snippet": "matched on history of self-harm", "ai_suggested": false}, {"id": "psychotropic_meds", "label": "Use of selected psychotropic medications", "role": "confounder", "adjusted": true, "snippet": "matched on use of selected psychotropic medications", "ai_suggested": false}, {"id": "alcohol_abuse", "label": "Alcohol abuse", "role": "confounder", "adjusted": true, "snippet": "matched on alcohol abuse", "ai_suggested": false}, {"id": "mental_health_admission", "label": "Admission to mental health unit", "role": "confounder", "adjusted": true, "snippet": "matched on admission to a mental health unit", "ai_suggested": false}, {"id": "ses", "label": "Socioeconomic status", "role": "confounder", "adjusted": true, "snippet": "adjusted for socioeconomic status", "ai_suggested": false}, {"id": "charlson_comorbidity", "label": "Charlson comorbidity score", "role": "confounder", "adjusted": true, "snippet": "adjusted for Charlson comorbidity score", "ai_suggested": false}, {"id": "physician_visits", "label": "Number of physician visits", "role": "confounder", "adjusted": true, "snippet": "adjusted for number of physician visits in the past year", "ai_suggested": false}, {"id": "ed_visits", "label": "Number of ED visits", "role": "confounder", "adjusted": true, "snippet": "adjusted for number of ED visits in the past year", "ai_suggested": false}, {"id": "psychiatrist_visits", "label": "Number of psychiatrist visits", "role": "confounder", "adjusted": true, "snippet": "adjusted for number of psychiatrist visits in the past year", "ai_suggested": false}, {"id": "past_psychotropic_use", "label": "Past psychotropic use", "role": "confounder", "adjusted": true, "snippet": "adjusted for past psychotropic use in previous 6 months", "ai_suggested": false}, {"id": "psychiatric_comorbidities", "label": "Psychiatric comorbidities", "role": "confounder", "adjusted": true, "snippet": "adjusted for psychiatric comorbidities", "ai_suggested": false}, {"id": "total_drugs_dispensed", "label": "Total number of drugs dispensed", "role": "confounder", "adjusted": true, "snippet": "adjusted for total number of drugs dispensed in the past 180 days", "ai_suggested": false}, {"id": "prior_quit_attempts", "label": "Prior quit attempts", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "recent_life_stressors", "label": "Recent life stressors", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "illicit_substance_use", "label": "Illicit substance use", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}], "edges": [{"source": "exposure", "target": "outcome", "ai_suggested": false}, {"source": "age", "target": "exposure", "ai_suggested": false}, {"source": "age", "target": "outcome", "ai_suggested": false}, {"source": "sex", "target": "exposure", "ai_suggested": false}, {"source": "sex", "target": "outcome", "ai_suggested": false}, {"source": "history_self_harm", "target": "exposure", "ai_suggested": false}, {"source": "history_self_harm", "target": "outcome", "ai_suggested": false}, {"source": "psychotropic_meds", "target": "exposure", "ai_suggested": false}, {"source": "psychotropic_meds", "target": "outcome", "ai_suggested": false}, {"source": "alcohol_abuse", "target": "exposure", "ai_suggested": false}, {"source": "alcohol_abuse", "target": "outcome", "ai_suggested": false}, {"source": "mental_health_admission", "target": "exposure", "ai_suggested": false}, {"source": "mental_health_admission", "target": "outcome", "ai_suggested": false}, {"source": "ses", "target": "exposure", "ai_suggested": false}, {"source": "ses", "target": "outcome", "ai_suggested": false}, {"source": "charlson_comorbidity", "target": "exposure", "ai_suggested": false}, {"source": "charlson_comorbidity", "target": "outcome", "ai_suggested": false}, {"source": "physician_visits", "target": "exposure", "ai_suggested": false}, {"source": "physician_visits", "target": "outcome", "ai_suggested": false}, {"source": "ed_visits", "target": "exposure", "ai_suggested": false}, {"source": "ed_visits", "target": "outcome", "ai_suggested": false}, {"source": "psychiatrist_visits", "target": "exposure", "ai_suggested": false}, {"source": "psychiatrist_visits", "target": "outcome", "ai_suggested": false}, {"source": "past_psychotropic_use", "target": "exposure", "ai_suggested": false}, {"source": "past_psychotropic_use", "target": "outcome", "ai_suggested": false}, {"source": "psychiatric_comorbidities", "target": "exposure", "ai_suggested": false}, {"source": "psychiatric_comorbidities", "target": "outcome", "ai_suggested": false}, {"source": "total_drugs_dispensed", "target": "exposure", "ai_suggested": false}, {"source": "total_drugs_dispensed", "target": "outcome", "ai_suggested": false}, {"source": "prior_quit_attempts", "target": "exposure", "ai_suggested": true}, {"source": "prior_quit_attempts", "target": "outcome", "ai_suggested": true}, {"source": "recent_life_stressors", "target": "exposure", "ai_suggested": true}, {"source": "recent_life_stressors", "target": "outcome", "ai_suggested": true}, {"source": "illicit_substance_use", "target": "exposure", "ai_suggested": true}, {"source": "illicit_substance_use", "target": "outcome", "ai_suggested": true}]}, "dagitty": "dag {\nexposure [exposure]\noutcome [outcome]\nage [adjusted]\nsex [adjusted]\nhistory_self_harm [adjusted]\npsychotropic_meds [adjusted]\nalcohol_abuse [adjusted]\nmental_health_admission [adjusted]\nses [adjusted]\ncharlson_comorbidity [adjusted]\nphysician_visits [adjusted]\ned_visits [adjusted]\npsychiatrist_visits [adjusted]\npast_psychotropic_use [adjusted]\npsychiatric_comorbidities [adjusted]\ntotal_drugs_dispensed [adjusted]\nprior_quit_attempts\nrecent_life_stressors\nillicit_substance_use\nexposure -> outcome\nage -> exposure\nage -> outcome\nsex -> exposure\nsex -> outcome\nhistory_self_harm -> exposure\nhistory_self_harm -> outcome\npsychotropic_meds -> exposure\npsychotropic_meds -> outcome\nalcohol_abuse -> exposure\nalcohol_abuse -> outcome\nmental_health_admission -> exposure\nmental_health_admission -> outcome\nses -> exposure\nses -> outcome\ncharlson_comorbidity -> exposure\ncharlson_comorbidity -> outcome\nphysician_visits -> exposure\nphysician_visits -> outcome\ned_visits -> exposure\ned_visits -> outcome\npsychiatrist_visits -> exposure\npsychiatrist_visits -> outcome\npast_psychotropic_use -> exposure\npast_psychotropic_use -> outcome\npsychiatric_comorbidities -> exposure\npsychiatric_comorbidities -> outcome\ntotal_drugs_dispensed -> exposure\ntotal_drugs_dispensed -> outcome\nprior_quit_attempts -> exposure\nprior_quit_attempts -> outcome\nrecent_life_stressors -> exposure\nrecent_life_stressors -> outcome\nillicit_substance_use -> exposure\nillicit_substance_use -> outcome\n}", "open_backdoors": ["Prior quit attempts", "Recent life stressors", "Illicit substance use"], "critique": {"missing_confounders": [{"label": "Depression severity", "rationale": "More severe depression increases the likelihood of being prescribed varenicline (due to higher addiction burden) and independently raises the risk of suicide or self\u2011harm.", "edges_to_add": [{"source": "Depression severity", "target": "exposure", "ai_suggested": false}, {"source": "Depression severity", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Suicidal Behavior and Depression in Smoking Cessation Treatments | PLOS One", "url": "https://journals.plos.org/plosone/article?id=10.1371%2Fjournal.pone.0027016"}, {"title": "Two Smoking Cessation Drugs Not Linked to Increased Risk of Depression or Suicide | RT", "url": "https://respiratory-therapy.com/products-treatment/pharmaceuticals/us-pharmaceuticals/two-smoking-cessation-drugs-not-linked-to-increased-risk-of-depression-or-suicide"}, {"title": "Risk of neuropsychiatric and cardiovascular adverse events following treatment with varenicline and nicotine replacement therapy in the UK Clinical Practice Research Datalink: a case\u2013cross\u2010over study ", "url": "https://onlinelibrary.wiley.com/doi/10.1111/add.15338"}], "status": "rejected", "review_note": "Depression severity may be altered by varenicline (neuropsychiatric side effects), making it a potential mediator/collider, and is only partially captured by psychiatric comorbidities."}, {"label": "Nicotine dependence", "rationale": "Higher daily cigarette consumption predicts both receipt of varenicline (clinicians may favor it for heavy smokers) and elevated suicidality via withdrawal\u2011related mood dysregulation.", "edges_to_add": [{"source": "Nicotine dependence", "target": "exposure", "ai_suggested": false}, {"source": "Nicotine dependence", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Varenicline and suicidal behaviour: a cohort study based on data from the General Practice Research Database", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC2755726"}, {"title": "Varenicline Use and Risk for Self-Harm and Suicide | NEJM Clinician", "url": "https://clinician.nejm.org/varenicline-use-risk-self-harm-suicide-JW200910270000001"}, {"title": "Risk of neuropsychiatric and cardiovascular adverse events following treatment with varenicline and nicotine replacement therapy in the UK Clinical Practice Research Datalink: a case\u2013cross\u2010over study", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC8246946"}], "status": "rejected", "review_note": "Nicotine dependence is affected by varenicline exposure (the drug reduces dependence), so adjusting would block a mediating pathway."}, {"label": "Prior quit attempts", "rationale": "A greater number of previous quit attempts signals stronger addiction and motivation, influencing prescribing decisions and is associated with higher baseline risk of self\u2011harm.", "edges_to_add": [{"source": "Prior quit attempts", "target": "exposure", "ai_suggested": false}, {"source": "Prior quit attempts", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Varenicline and suicidal behaviour: a cohort study based on data from the General Practice Research Database", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC2755726"}, {"title": "Smoking Cessation Drug Not Linked To An Increased Risk Of Self Harm Or Depression | ScienceDaily", "url": "https://www.sciencedaily.com/releases/2009/10/091001191701.htm"}, {"title": "Are Smoking Cessation Treatments Associated with Suicidality Risk? An Overview", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC4830638"}], "status": "accepted", "review_note": "Prior quit attempts precede varenicline prescribing, influence both exposure and outcome, and are not captured by the adjusted covariates."}, {"label": "Recent life stressors", "rationale": "Events such as job loss, bereavement, or relationship breakup increase both the likelihood of seeking smoking\u2011cessation pharmacotherapy and the risk of suicidal behaviour.", "edges_to_add": [{"source": "Recent life stressors", "target": "exposure", "ai_suggested": false}, {"source": "Recent life stressors", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Varenicline and suicidal behaviour: a cohort study based on data from the General Practice Research Database", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC2755726"}, {"title": "Smoking cessation drug not linked to an increased risk of self harm or depression", "url": "https://medicalxpress.com/news/2009-10-cessation-drug-linked-depression.html"}, {"title": "Risk of neuropsychiatric and cardiovascular adverse events following treatment with varenicline and nicotine replacement therapy in the UK Clinical Practice Research Datalink: a case\u2013cross\u2010over study", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC8246946"}], "status": "accepted", "review_note": "Recent life stressors occur before exposure, affect both likelihood of receiving varenicline and suicide risk, and are not adjusted for."}, {"label": "Illicit substance use", "rationale": "Use of illicit drugs is linked to higher depression and impulsivity, may affect clinician choice of cessation medication, and independently raises suicide/self\u2011harm risk.", "edges_to_add": [{"source": "Illicit substance use", "target": "exposure", "ai_suggested": false}, {"source": "Illicit substance use", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Risk of neuropsychiatric and cardiovascular adverse events following treatment with varenicline and nicotine replacement therapy in the UK Clinical Practice Research Datalink: a case\u2013cross\u2010over study", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC8246946"}, {"title": "Two Smoking Cessation Drugs Not Linked to Increased Risk of Depression or Suicide | RT", "url": "https://respiratory-therapy.com/products-treatment/pharmaceuticals/us-pharmaceuticals/two-smoking-cessation-drugs-not-linked-to-increased-risk-of-depression-or-suicide"}, {"title": "Chantix Unsuitable for First-Line Smoking Cessation Use | Atrium Health Wake Forest Baptist", "url": "https://newsroom.wakehealth.edu/news-releases/2011/11/chantix-unsuitable-for-first-line-smoking-cessation-use"}], "status": "accepted", "review_note": "Illicit substance use precedes exposure, influences prescribing and self\u2011harm risk, and is not represented in the adjusted list."}], "bias_risks": [{"type": "confounding by indication", "explanation": "Varenicline may be preferentially prescribed to patients with heavier smoking histories, prior failed quit attempts, or untreated depression, which themselves increase suicide risk, creating channeling bias not fully captured by the matched variables.", "severity": "high"}, {"type": "measurement error", "explanation": "Outcome ascertainment relied on hospital or ED visits for suicide/self\u2011harm; events managed outpatient or not presenting for care are missed, leading to differential under\u2011ascertainment if drug users differ in healthcare utilization.", "severity": "medium"}, {"type": "immortal time bias", "explanation": "Exposure was defined as having a prescription record; individuals must survive and remain drug\u2011covered until the first prescription, creating a period during which the outcome cannot occur, potentially biasing the OR toward null.", "severity": "medium"}, {"type": "selection bias", "explanation": "Matching on numerous variables (including psychotropic medication use and mental\u2011health admissions) may induce overmatching, removing part of the true exposure effect and distorting the exposure\u2013outcome association.", "severity": "low"}, {"type": "healthy user effect", "explanation": "Patients who successfully initiate and adhere to smoking\u2011cessation therapy may be healthier and more motivated, which could lower their baseline suicide risk independent of the drug, biasing the estimate downward.", "severity": "low"}], "overall_assessment": "The study\u2019s adjusted odds ratio of 1.15 with a wide confidence interval crossing unity suggests no clear evidence of increased suicide/self\u2011harm risk with varenicline versus bupropion. However, important confounders such as depression severity, nicotine dependence, and recent life stressors were not measured, and biases like confounding by indication and potential immortal time bias remain plausible. Consequently, the causal claim that varenicline does not increase suicide risk should be interpreted cautiously.", "label": "AI-suggested"}, "models": {"extraction": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B", "critique": "nvidia/nemotron-3-super-120b-a12b", "evidence_search": "tavily", "provider": "Nebius Token Factory"}, "review": {"by": "ai", "model": "nvidia/nemotron-3-super-120b-a12b"}, "dataset_license": "CC-BY-4.0"}
{"id": "e986586ca2", "schema_version": "1.0", "status": "approved", "created_at": "2026-10-28T20:14:13Z", "approved_at": "2026-10-28T21:25:37Z", "paper": {"title": "Use of sugar in coffee and tea and long-term risk of mortality in older adult Danish men: 32 years of follow-up from a prospective cohort study", "authors": "Treskes RW, Clausen J, Marott JL, Jensen GB, Holtermann A, Gyntelberg F, Jensen MT.", "year": "2023", "journal": "PloS one", "doi": "10.1371/journal.pone.0292882", "pmcid": "PMC10584177", "pmid": "37851689", "license": "cc by"}, "study": {"study_design": "cohort", "population": "Danish men aged 63\u00b15 years from the Copenhagen Male Study cohort, followed from 1985 to 2017", "sample_size": "2923 (1007 exposed)", "follow_up": "32 years", "exposure": {"name": "Sugar addition to tea or coffee", "definition": "Self-reported addition of sugar to coffee or tea (yes/no)"}, "outcome": {"name": "All-cause mortality", "definition": "Death from any cause identified via Danish national registries"}, "reported_effect": {"measure": "HR", "value": "1.06", "ci": "0.98-1.16", "adjusted": true}}, "dag": {"nodes": [{"id": "exposure", "label": "Sugar addition", "role": "exposure", "adjusted": false, "snippet": "adding sugar to coffee or tea (yes/no)", "ai_suggested": false}, {"id": "outcome", "label": "All-cause mortality", "role": "outcome", "adjusted": false, "snippet": "all-cause mortality assessed through national registries", "ai_suggested": false}, {"id": "age", "label": "Age", "role": "confounder", "adjusted": true, "snippet": "Age, smoking status, daily alcohol intake, systolic blood pressure, body mass index, number of cups of coffee and/or tea consumed \u2026", "ai_suggested": false}, {"id": "smoking_status", "label": "Smoking status", "role": "confounder", "adjusted": true, "snippet": "Age, smoking status, daily alcohol intake, systolic blood pressure, body mass index, number of cups of coffee and/or tea consumed \u2026", "ai_suggested": false}, {"id": "alcohol_intake", "label": "Daily alcohol intake", "role": "confounder", "adjusted": true, "snippet": "Age, smoking status, daily alcohol intake, systolic blood pressure, body mass index, number of cups of coffee and/or tea consumed \u2026", "ai_suggested": false}, {"id": "bmi", "label": "Body mass index", "role": "confounder", "adjusted": true, "snippet": "Age, smoking status, daily alcohol intake, systolic blood pressure, body mass index, number of cups of coffee and/or tea consumed \u2026", "ai_suggested": false}, {"id": "systolic_bp", "label": "Systolic blood pressure", "role": "confounder", "adjusted": true, "snippet": "Age, smoking status, daily alcohol intake, systolic blood pressure, body mass index, number of cups of coffee and/or tea consumed \u2026", "ai_suggested": false}, {"id": "socioeconomic_status", "label": "Socioeconomic status", "role": "confounder", "adjusted": true, "snippet": "Age, smoking status, daily alcohol intake, systolic blood pressure, body mass index, number of cups of coffee and/or tea consumed \u2026", "ai_suggested": false}, {"id": "cups_per_day", "label": "Cups of coffee/tea per day", "role": "confounder", "adjusted": true, "snippet": "number of cups of coffee and/or tea consumed per day", "ai_suggested": false}, {"id": "physical_activity", "label": "Physical activity", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "self_rated_health", "label": "Self-rated health", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "medication_use", "label": "Medication use", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "depressive_symptoms", "label": "Depressive symptoms", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}], "edges": [{"source": "exposure", "target": "outcome", "ai_suggested": false}, {"source": "age", "target": "exposure", "ai_suggested": false}, {"source": "age", "target": "outcome", "ai_suggested": false}, {"source": "smoking_status", "target": "exposure", "ai_suggested": false}, {"source": "smoking_status", "target": "outcome", "ai_suggested": false}, {"source": "alcohol_intake", "target": "exposure", "ai_suggested": false}, {"source": "alcohol_intake", "target": "outcome", "ai_suggested": false}, {"source": "bmi", "target": "exposure", "ai_suggested": false}, {"source": "bmi", "target": "outcome", "ai_suggested": false}, {"source": "systolic_bp", "target": "exposure", "ai_suggested": false}, {"source": "systolic_bp", "target": "outcome", "ai_suggested": false}, {"source": "socioeconomic_status", "target": "exposure", "ai_suggested": false}, {"source": "socioeconomic_status", "target": "outcome", "ai_suggested": false}, {"source": "cups_per_day", "target": "exposure", "ai_suggested": false}, {"source": "cups_per_day", "target": "outcome", "ai_suggested": false}, {"source": "physical_activity", "target": "exposure", "ai_suggested": true}, {"source": "physical_activity", "target": "outcome", "ai_suggested": true}, {"source": "self_rated_health", "target": "exposure", "ai_suggested": true}, {"source": "self_rated_health", "target": "outcome", "ai_suggested": true}, {"source": "medication_use", "target": "exposure", "ai_suggested": true}, {"source": "medication_use", "target": "outcome", "ai_suggested": true}, {"source": "depressive_symptoms", "target": "exposure", "ai_suggested": true}, {"source": "depressive_symptoms", "target": "outcome", "ai_suggested": true}]}, "dagitty": "dag {\nexposure [exposure]\noutcome [outcome]\nage [adjusted]\nsmoking_status [adjusted]\nalcohol_intake [adjusted]\nbmi [adjusted]\nsystolic_bp [adjusted]\nsocioeconomic_status [adjusted]\ncups_per_day [adjusted]\nphysical_activity\nself_rated_health\nmedication_use\ndepressive_symptoms\nexposure -> outcome\nage -> exposure\nage -> outcome\nsmoking_status -> exposure\nsmoking_status -> outcome\nalcohol_intake -> exposure\nalcohol_intake -> outcome\nbmi -> exposure\nbmi -> outcome\nsystolic_bp -> exposure\nsystolic_bp -> outcome\nsocioeconomic_status -> exposure\nsocioeconomic_status -> outcome\ncups_per_day -> exposure\ncups_per_day -> outcome\nphysical_activity -> exposure\nphysical_activity -> outcome\nself_rated_health -> exposure\nself_rated_health -> outcome\nmedication_use -> exposure\nmedication_use -> outcome\ndepressive_symptoms -> exposure\ndepressive_symptoms -> outcome\n}", "open_backdoors": ["Physical activity", "Self-rated health", "Medication use", "Depressive symptoms"], "critique": {"missing_confounders": [{"label": "Physical activity", "rationale": "Less active men may prefer sweeter beverages and also have higher mortality risk.", "edges_to_add": [{"source": "Physical activity", "target": "exposure", "ai_suggested": false}, {"source": "Physical activity", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Physical activity combined with tea consumption could further reduce all-cause and cancer-specific mortality | Scientific Reports", "url": "https://www.nature.com/articles/s41598-024-73962-z"}, {"title": "Consumption of coffee and tea with all-cause and cause-specific mortality: a prospective cohort study | BMC Medicine | Springer Nature Link", "url": "https://link.springer.com/article/10.1186/s12916-022-02636-2"}, {"title": "Coffee, Green Tea, and Type 2 Diabetes Mortality Rates | Cardiometabolic Health Congress", "url": "https://www.cardiometabolichealth.org/coffee-green-tea-and-type-2-diabetes-mortality-rates"}], "status": "accepted", "review_note": "Physical activity influences both the likelihood of adding sugar to beverages (via taste or lifestyle habits) and all-cause mortality risk, and it is not adjusted for nor caused by the exposure."}, {"label": "Total dietary sugar", "rationale": "Higher overall sugar intake from foods correlates with adding sugar to drinks and independently increases mortality risk.", "edges_to_add": [{"source": "Total dietary sugar", "target": "exposure", "ai_suggested": false}, {"source": "Total dietary sugar", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Frontiers | The associations of unsweetened, sugar-sweetened, and artificially sweetened tea consumption with all-cause and cause-specific mortality in 195,361 UK Biobank participants: a large prospec", "url": "https://www.frontiersin.org/journals/nutrition/articles/10.3389/fnut.2025.1649279/full"}, {"title": "Adding sugar to coffee, tea may not significantly impact health", "url": "https://resources.healthgrades.com/pro/adding-sugar-to-coffee-tea-may-not-significantly-impact-health"}, {"title": "Consumption of coffee and tea with all-cause and cause-specific mortality: a prospective cohort study | BMC Medicine | Springer Nature Link", "url": "https://link.springer.com/article/10.1186/s12916-022-02636-2"}], "status": "rejected", "review_note": "Total dietary sugar includes the sugar added to tea/coffee itself, making it downstream of the exposure; adjusting would constitute overadjustment bias."}, {"label": "Self-rated health", "rationale": "Poor self-rated health may lead individuals to avoid sugar (or seek comfort via sweetened drinks) and predicts mortality.", "edges_to_add": [{"source": "Self-rated health", "target": "exposure", "ai_suggested": false}, {"source": "Self-rated health", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Frontiers | Coffee and tea consumption and cardiovascular disease and all-cause and cause-specific mortality in individuals with diabetes mellitus: a meta-analysis of prospective observational studies", "url": "https://www.frontiersin.org/journals/nutrition/articles/10.3389/fnut.2025.1570644/full"}, {"title": "Coffee Consumption and Mortality among United States Adults: A Prospective Cohort Study", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC12308089"}, {"title": "Coffee, Green Tea, and Type 2 Diabetes Mortality Rates | Cardiometabolic Health Congress", "url": "https://www.cardiometabolichealth.org/coffee-green-tea-and-type-2-diabetes-mortality-rates"}], "status": "accepted", "review_note": "Poor self-rated health can affect both the decision to add sugar (e.g., avoidance or comfort-seeking) and mortality risk, and it is not already accounted for in the adjustment set."}, {"label": "Medication use", "rationale": "Use of medications for chronic conditions may influence dietary habits (including sugar addition) and is associated with mortality.", "edges_to_add": [{"source": "Medication use", "target": "exposure", "ai_suggested": false}, {"source": "Medication use", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Consumption of coffee and tea with all-cause and cause-specific mortality: a prospective cohort study | BMC Medicine | Springer Nature Link", "url": "https://link.springer.com/article/10.1186/s12916-022-02636-2"}, {"title": "sweetened-and-unsweetened-coffee.pdf", "url": "https://md.teyit.org/file/sweetened-and-unsweetened-coffee.pdf"}, {"title": "Coffee, Green Tea, and Type 2 Diabetes Mortality Rates | Cardiometabolic Health Congress", "url": "https://www.cardiometabolichealth.org/coffee-green-tea-and-type-2-diabetes-mortality-rates"}], "status": "accepted", "review_note": "Medication use reflects underlying chronic conditions that may shape dietary habits (including sugar addition) and predict mortality, and it is not adjusted for nor caused by the exposure."}, {"label": "Depressive symptoms", "rationale": "Depression can increase preference for sweet tastes and is linked to higher mortality.", "edges_to_add": [{"source": "Depressive symptoms", "target": "exposure", "ai_suggested": false}, {"source": "Depressive symptoms", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Consumption of coffee and tea with all-cause and cause-specific mortality: a prospective cohort study", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC9673438"}, {"title": "Does adding sugar to coffee and tea impact your lifespan? 32-year ...", "url": "https://www.news-medical.net/news/20231022/Does-adding-sugar-to-coffee-and-tea-impact-your-lifespan-32-year-study-reveals-surprising-results.aspx"}, {"title": "sweetened-and-unsweetened-coffee.pdf - Teyit", "url": "https://md.teyit.org/file/sweetened-and-unsweetened-coffee.pdf"}], "status": "accepted", "review_note": "Depressive symptoms can increase preference for sweetened drinks and independently raise mortality risk, satisfying the confounder criteria and not being adjusted for."}], "bias_risks": [{"type": "measurement error", "explanation": "Sugar addition was self\u2011reported once at baseline, likely subject to recall and social desirability bias, leading to non\u2011differential misclassification of exposure.", "severity": "medium"}, {"type": "selection bias", "explanation": "The cohort excluded men with prevalent cardiovascular disease, cancer, or diabetes at enrollment, creating a healthier volunteer sample that may limit generalizability and induce bias if the exposure\u2011outcome relation differs in the excluded group.", "severity": "medium"}, {"type": "healthy user effect", "explanation": "Men who add sugar may be less health\u2011conscious overall, and this unmeasured health\u2011consciousness could confound the observed null association.", "severity": "low"}, {"type": "residual confounding", "explanation": "Important lifestyle factors such as physical activity, total dietary sugar intake, and mental health were not measured, leaving open paths for confounding despite adjustment for the listed covariates.", "severity": "medium"}, {"type": "loss to follow-up", "explanation": "Although vital status was ascertained via national registries, a small proportion of participants may have emigrated or had incomplete registry linkage, potentially biasing results if related to both sugar use and mortality.", "severity": "low"}], "overall_assessment": "The study\u2019s hazard ratio is close to null with a confidence interval that includes no effect, and while several key confounders were controlled, important lifestyle and health\u2011related variables were not measured. Measurement error in exposure assessment and selection of a relatively healthy volunteer sample further weaken causal inference. Consequently, the evidence for a causal effect of adding sugar to tea or coffee on mortality is weak and should be interpreted cautiously.", "label": "AI-suggested"}, "models": {"extraction": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B", "critique": "nvidia/nemotron-3-super-120b-a12b", "evidence_search": "tavily", "provider": "Nebius Token Factory"}, "review": {"by": "ai", "model": "nvidia/nemotron-3-super-120b-a12b"}, "dataset_license": "CC-BY-4.0"}
{"id": "2f50ed8247", "schema_version": "1.0", "status": "approved", "created_at": "2026-10-28T20:10:09Z", "approved_at": "2026-10-28T21:25:58Z", "paper": {"title": "Prospective association between ultra-processed food consumption and incident depressive symptoms in the French NutriNet-Sant\u00e9 cohort", "authors": "Adjibade M, Julia C, All\u00e8s B, Touvier M, Lemogne C, Srour B, Hercberg S, Galan P, Assmann KE, Kesse-Guyot E.", "year": "2019", "journal": "BMC medicine", "doi": "10.1186/s12916-019-1312-y", "pmcid": "PMC6463641", "pmid": "30982472", "license": "cc by"}, "study": {"study_design": "cohort", "population": "Adults aged 18\u201386 years from the NutriNet-Sant\u00e9 cohort in France, followed prospectively", "sample_size": "26,730", "follow_up": "mean 5.4 years", "exposure": {"name": "UPF proportion", "definition": "Proportion of ultra-processed foods (%UPF) in the diet, calculated from 24-h dietary records and classified using the NOVA system"}, "outcome": {"name": "Incident depressive symptoms", "definition": "New onset of depressive symptoms identified by CES-D score \u2265 17 in men and \u2265 23 in women at any follow-up assessment"}, "reported_effect": {"measure": "HR", "value": "1.21", "ci": "1.15-1.27", "adjusted": true}}, "dag": {"nodes": [{"id": "exposure", "label": "Ultra-processed food proportion (%UPF)", "role": "exposure", "adjusted": false, "snippet": "proportion of ultra-processed foods (%UPF) in the diet", "ai_suggested": false}, {"id": "outcome", "label": "Incident depressive symptoms", "role": "outcome", "adjusted": false, "snippet": "incident cases of depressive symptoms", "ai_suggested": false}, {"id": "age", "label": "Age", "role": "confounder", "adjusted": true, "snippet": "Age was used as the primary time scale variable", "ai_suggested": false}, {"id": "sex", "label": "Sex", "role": "confounder", "adjusted": true, "snippet": "The first model was adjusted for age, sex, and BMI", "ai_suggested": false}, {"id": "bmi", "label": "Body mass index", "role": "confounder", "adjusted": true, "snippet": "adjusted for sociodemographic characteristics, body mass index, and lifestyle factors", "ai_suggested": false}, {"id": "marital_status", "label": "Marital status", "role": "confounder", "adjusted": true, "snippet": "adjusted for marital status, educational level, occupational categories, monthly household income per consumption unit, residential area, energy intake without alcohol, \u2026", "ai_suggested": false}, {"id": "education", "label": "Educational level", "role": "confounder", "adjusted": true, "snippet": "adjusted for marital status, educational level, occupational categories, monthly household income per consumption unit, residential area, energy intake without alcohol, \u2026", "ai_suggested": false}, {"id": "occupation", "label": "Occupational category", "role": "confounder", "adjusted": true, "snippet": "adjusted for marital status, educational level, occupational categories, monthly household income per consumption unit, residential area, energy intake without alcohol, \u2026", "ai_suggested": false}, {"id": "income", "label": "Monthly household income", "role": "confounder", "adjusted": true, "snippet": "adjusted for marital status, educational level, occupational categories, monthly household income per consumption unit, residential area, energy intake without alcohol, \u2026", "ai_suggested": false}, {"id": "smoking", "label": "Smoking status", "role": "confounder", "adjusted": true, "snippet": "adjusted for marital status, educational level, occupational categories, monthly household income per consumption unit, residential area, energy intake without alcohol, \u2026", "ai_suggested": false}, {"id": "physical_activity", "label": "Physical activity", "role": "confounder", "adjusted": true, "snippet": "adjusted for marital status, educational level, occupational categories, monthly household income per consumption unit, residential area, energy intake without alcohol, \u2026", "ai_suggested": false}, {"id": "dietary_patterns", "label": "Dietary pattern scores", "role": "confounder", "adjusted": true, "snippet": "Five additional models were also performed to account for (a) PCA-extracted dietary patterns and intake of carbohydrates, lipids, and salt", "ai_suggested": false}, {"id": "energy_intake", "label": "Energy intake without alcohol", "role": "confounder", "adjusted": true, "snippet": "adjusted for marital status, educational level, occupational categories, monthly household income per consumption unit, residential area, energy intake without alcohol, \u2026", "ai_suggested": false}, {"id": "alcohol_consumption", "label": "Alcohol consumption", "role": "confounder", "adjusted": true, "snippet": "adjusted for marital status, educational level, occupational categories, monthly household income per consumption unit, residential area, energy intake without alcohol, \u2026", "ai_suggested": false}, {"id": "number_of_records", "label": "Number of 24-h dietary records", "role": "confounder", "adjusted": true, "snippet": "adjusted for marital status, educational level, occupational categories, monthly household income per consumption unit, residential area, energy intake without alcohol, \u2026", "ai_suggested": false}, {"id": "baseline_cesd_score", "label": "Baseline CES-D score", "role": "confounder", "adjusted": true, "snippet": "adjusted for baseline CES-D score (continuous variable), and use of antidepressants during the follow-up", "ai_suggested": false}, {"id": "use_of_antidepressants", "label": "Use of antidepressants", "role": "confounder", "adjusted": true, "snippet": "adjusted for baseline CES-D score (continuous variable), and use of antidepressants during the follow-up", "ai_suggested": false}, {"id": "cds_score", "label": "Cognitive Difficulties Scale score", "role": "confounder", "adjusted": true, "snippet": "adjusted for baseline CES-D score (continuous variable), and use of antidepressants during the follow-up and CDS score (continuous variable)", "ai_suggested": false}, {"id": "social_isolation", "label": "Social isolation", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "family_depression", "label": "Family depression", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "neuroticism", "label": "Neuroticism", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}], "edges": [{"source": "exposure", "target": "outcome", "ai_suggested": false}, {"source": "age", "target": "exposure", "ai_suggested": false}, {"source": "age", "target": "outcome", "ai_suggested": false}, {"source": "sex", "target": "exposure", "ai_suggested": false}, {"source": "sex", "target": "outcome", "ai_suggested": false}, {"source": "bmi", "target": "exposure", "ai_suggested": false}, {"source": "bmi", "target": "outcome", "ai_suggested": false}, {"source": "marital_status", "target": "exposure", "ai_suggested": false}, {"source": "marital_status", "target": "outcome", "ai_suggested": false}, {"source": "education", "target": "exposure", "ai_suggested": false}, {"source": "education", "target": "outcome", "ai_suggested": false}, {"source": "occupation", "target": "exposure", "ai_suggested": false}, {"source": "occupation", "target": "outcome", "ai_suggested": false}, {"source": "income", "target": "exposure", "ai_suggested": false}, {"source": "income", "target": "outcome", "ai_suggested": false}, {"source": "smoking", "target": "exposure", "ai_suggested": false}, {"source": "smoking", "target": "outcome", "ai_suggested": false}, {"source": "physical_activity", "target": "exposure", "ai_suggested": false}, {"source": "physical_activity", "target": "outcome", "ai_suggested": false}, {"source": "dietary_patterns", "target": "exposure", "ai_suggested": false}, {"source": "dietary_patterns", "target": "outcome", "ai_suggested": false}, {"source": "energy_intake", "target": "exposure", "ai_suggested": false}, {"source": "energy_intake", "target": "outcome", "ai_suggested": false}, {"source": "alcohol_consumption", "target": "exposure", "ai_suggested": false}, {"source": "alcohol_consumption", "target": "outcome", "ai_suggested": false}, {"source": "number_of_records", "target": "exposure", "ai_suggested": false}, {"source": "number_of_records", "target": "outcome", "ai_suggested": false}, {"source": "baseline_cesd_score", "target": "exposure", "ai_suggested": false}, {"source": "baseline_cesd_score", "target": "outcome", "ai_suggested": false}, {"source": "use_of_antidepressants", "target": "exposure", "ai_suggested": false}, {"source": "use_of_antidepressants", "target": "outcome", "ai_suggested": false}, {"source": "cds_score", "target": "exposure", "ai_suggested": false}, {"source": "cds_score", "target": "outcome", "ai_suggested": false}, {"source": "social_isolation", "target": "exposure", "ai_suggested": true}, {"source": "social_isolation", "target": "outcome", "ai_suggested": true}, {"source": "family_depression", "target": "exposure", "ai_suggested": true}, {"source": "family_depression", "target": "outcome", "ai_suggested": true}, {"source": "neuroticism", "target": "exposure", "ai_suggested": true}, {"source": "neuroticism", "target": "outcome", "ai_suggested": true}]}, "dagitty": "dag {\nexposure [exposure]\noutcome [outcome]\nage [adjusted]\nsex [adjusted]\nbmi [adjusted]\nmarital_status [adjusted]\neducation [adjusted]\noccupation [adjusted]\nincome [adjusted]\nsmoking [adjusted]\nphysical_activity [adjusted]\ndietary_patterns [adjusted]\nenergy_intake [adjusted]\nalcohol_consumption [adjusted]\nnumber_of_records [adjusted]\nbaseline_cesd_score [adjusted]\nuse_of_antidepressants [adjusted]\ncds_score [adjusted]\nsocial_isolation\nfamily_depression\nneuroticism\nexposure -> outcome\nage -> exposure\nage -> outcome\nsex -> exposure\nsex -> outcome\nbmi -> exposure\nbmi -> outcome\nmarital_status -> exposure\nmarital_status -> outcome\neducation -> exposure\neducation -> outcome\noccupation -> exposure\noccupation -> outcome\nincome -> exposure\nincome -> outcome\nsmoking -> exposure\nsmoking -> outcome\nphysical_activity -> exposure\nphysical_activity -> outcome\ndietary_patterns -> exposure\ndietary_patterns -> outcome\nenergy_intake -> exposure\nenergy_intake -> outcome\nalcohol_consumption -> exposure\nalcohol_consumption -> outcome\nnumber_of_records -> exposure\nnumber_of_records -> outcome\nbaseline_cesd_score -> exposure\nbaseline_cesd_score -> outcome\nuse_of_antidepressants -> exposure\nuse_of_antidepressants -> outcome\ncds_score -> exposure\ncds_score -> outcome\nsocial_isolation -> exposure\nsocial_isolation -> outcome\nfamily_depression -> exposure\nfamily_depression -> outcome\nneuroticism -> exposure\nneuroticism -> outcome\n}", "open_backdoors": ["Social isolation", "Family depression", "Neuroticism"], "critique": {"missing_confounders": [{"label": "Sleep deprivation", "rationale": "Poor sleep increases the risk of developing depressive symptoms and may lead individuals to choose convenient, ultra\u2011processed foods due to fatigue and altered appetite regulation.", "edges_to_add": [{"source": "Sleep deprivation", "target": "exposure", "ai_suggested": false}, {"source": "Sleep deprivation", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Sleep deprivation makes us less happy, more anxious", "url": "https://www.apa.org/news/press/releases/2023/12/sleep-deprivation-anxious"}, {"title": "Effect of Inadequate Sleep on Frequent Mental Distress", "url": "https://www.cdc.gov/pcd/issues/2021/20_0573.htm"}, {"title": "How sleep deprivation can harm your health - Harvard Health", "url": "https://www.health.harvard.edu/sleep/how-sleep-deprivation-can-harm-your-health"}], "status": "rejected", "review_note": "Sleep deprivation may be a consequence of high UPF intake (diet affects sleep quality) and thus adjusting for it could block a mediating pathway rather than control confounding."}, {"label": "Social isolation", "rationale": "Loneliness and lack of social support are strong predictors of depression and can increase reliance on ready\u2011to\u2011eat ultra\u2011processed meals when cooking for others is less frequent.", "edges_to_add": [{"source": "Social isolation", "target": "exposure", "ai_suggested": false}, {"source": "Social isolation", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Ultra-Processed Food Is Positively Associated With Depressive ...", "url": "https://www.frontiersin.org/journals/nutrition/articles/10.3389/fnut.2020.600449/full"}, {"title": "Relationship between 39 dietary markers of ultra-processed food ...", "url": "https://link.springer.com/article/10.1007/s00394-026-04084-7"}, {"title": "CE corner: The risks of social isolation", "url": "https://www.apa.org/monitor/2019/05/ce-corner-isolation"}], "status": "accepted", "review_note": "Social isolation predicts both greater reliance on ultra\u2011processed foods and incident depressive symptoms and is not captured by the adjusted covariates (marital status alone does not measure broader loneliness)."}, {"label": "Chronic illness", "rationale": "Conditions such as cardiovascular disease or diabetes raise depression risk and often necessitate diets higher in processed foods due to physical limitations or dietary recommendations.", "edges_to_add": [{"source": "Chronic illness", "target": "exposure", "ai_suggested": false}, {"source": "Chronic illness", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "The risk associated with ultra-processed food intake on depressive symptoms and mental health in older adults: a target trial emulation | BMC Medicine | Springer Nature Link", "url": "https://link.springer.com/article/10.1186/s12916-025-04002-4"}, {"title": "The Science of Ultra-Processed Foods and Mental Health", "url": "https://www.news-medical.net/health/The-Science-of-Ultra-Processed-Foods-and-Mental-Health.aspx"}, {"title": "Ultra-processed foods: increased risk of depression and anxiety", "url": "https://usrtk.org/ultra-processed-foods/depression-anxiety"}], "status": "rejected", "review_note": "Chronic illnesses such as cardiovascular disease or diabetes are likely downstream effects of high UPF consumption, so adjusting for them would introduce bias by controlling for a mediator."}, {"label": "Family depression", "rationale": "A family history of depression reflects genetic and shared environmental susceptibility to depression and may correlate with familial dietary patterns that include higher ultra\u2011processed food intake.", "edges_to_add": [{"source": "Family depression", "target": "exposure", "ai_suggested": false}, {"source": "Family depression", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "The risk associated with ultra-processed food intake on depressive symptoms and mental health in older adults: a target trial emulation | BMC Medicine | Springer Nature Link", "url": "https://link.springer.com/article/10.1186/s12916-025-04002-4"}, {"title": "Frontiers | Ultra-Processed Food Is Positively Associated With Depressive Symptoms Among United States Adults", "url": "https://www.frontiersin.org/journals/nutrition/articles/10.3389/fnut.2020.600449/full"}, {"title": "The Science of Ultra-Processed Foods and Mental Health", "url": "https://www.news-medical.net/health/The-Science-of-Ultra-Processed-Foods-and-Mental-Health.aspx"}], "status": "accepted", "review_note": "Family history of depression is a pre\u2011existing common cause of both depressive symptoms and familial dietary patterns (including UPF) and is not already adjusted for."}, {"label": "Neuroticism", "rationale": "High neuroticism predisposes individuals to depressive symptoms and is associated with emotional eating of ultra\u2011processed foods as a coping mechanism.", "edges_to_add": [{"source": "Neuroticism", "target": "exposure", "ai_suggested": false}, {"source": "Neuroticism", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Frontiers | Ultra-Processed Food Is Positively Associated With Depressive Symptoms Among United States Adults", "url": "https://www.frontiersin.org/journals/nutrition/articles/10.3389/fnut.2020.600449/full"}, {"title": "The Science of Ultra-Processed Foods and Mental Health", "url": "https://www.news-medical.net/health/The-Science-of-Ultra-Processed-Foods-and-Mental-Health.aspx"}, {"title": "Association Between Ultra-Processed Food Consumption and Risk of Developing Depression in Adults: A Systematic Review - European Medical Journal", "url": "https://www.emjreviews.com/gastroenterology/article/association-between-ultra-processed-food-consumption-and-risk-of-developing-depression-in-adults-a-systematic-review-j050123"}], "status": "accepted", "review_note": "Neuroticism is a stable personality trait that elevates depression risk and promotes emotional eating of ultra\u2011processed foods, satisfying confounder criteria and absent from the adjustment set."}], "bias_risks": [], "overall_assessment": "", "label": "AI-suggested"}, "models": {"extraction": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B", "critique": "nvidia/nemotron-3-super-120b-a12b", "evidence_search": "tavily", "provider": "Nebius Token Factory"}, "review": {"by": "ai", "model": "nvidia/nemotron-3-super-120b-a12b"}, "dataset_license": "CC-BY-4.0"}
{"id": "53def092ed", "schema_version": "1.0", "status": "approved", "created_at": "2026-10-28T20:08:24Z", "approved_at": "2026-10-28T21:26:21Z", "paper": {"title": "Duration and dosing of Proton Pump Inhibitors associated with high incidence of chronic kidney disease in population-based cohort", "authors": "Rodr\u00edguez-Poncelas A, Barcel\u00f3 MA, Saez M, Coll-de-Tuero G.", "year": "2018", "journal": "PloS one", "doi": "10.1371/journal.pone.0204231", "pmcid": "PMC6192564", "pmid": "30332411", "license": "cc by"}, "study": {"study_design": "cohort", "population": "people aged 15 years or over in the Catalan public healthcare system (La Selva Interior region, Spain) between 2005 and 2012", "sample_size": "46,541", "follow_up": "up to 7 years (2005\u20132012)", "exposure": {"name": "PPI use", "definition": "prescription of proton pump inhibitors recorded in medical records, categorized by dose and duration"}, "outcome": {"name": "incident CKD", "definition": "eGFR < 60 ml/min/1.73 m\u00b2 and/or UACR \u2265 30 mg/g in two or more determinations at least 3 months apart"}, "reported_effect": {"measure": "HR", "value": "1.18", "ci": "1.04-1.51", "adjusted": true}}, "dag": {"nodes": [{"id": "ppi_use", "label": "PPI use", "role": "exposure", "adjusted": false, "snippet": "PPI use was measured during the follow-up by recording the prescription", "ai_suggested": false}, {"id": "incident_ckd", "label": "Incident CKD", "role": "outcome", "adjusted": false, "snippet": "Incident CKD was defined as an estimated glomerular filtration rate < 60 ml/min/1.73 m\u00b2 and/or urinary albumin level to creatinine \u2026", "ai_suggested": false}, {"id": "age", "label": "Age", "role": "confounder", "adjusted": true, "snippet": "We categorized the age variable (0 being under 70 years\u2014reference category)", "ai_suggested": false}, {"id": "sex", "label": "Sex", "role": "confounder", "adjusted": true, "snippet": "sex (0 male\u2014reference category -, 1 female)", "ai_suggested": false}, {"id": "hypertension", "label": "Hypertension", "role": "confounder", "adjusted": true, "snippet": "diagnosed hypertension or those receiving treatment with anti-hypertensive medication", "ai_suggested": false}, {"id": "diabetes", "label": "Diabetes", "role": "confounder", "adjusted": true, "snippet": "type II diabetes diagnosed by ADA criteria or treated with antidiabetic medication", "ai_suggested": false}, {"id": "obesity", "label": "Obesity", "role": "confounder", "adjusted": true, "snippet": "body mass index > 30 kg/m\u00b2", "ai_suggested": false}, {"id": "smoking", "label": "Smoking status", "role": "confounder", "adjusted": true, "snippet": "smoking status (0 non-smoker\u2014reference category, 1 smoker, 2 former smoker)", "ai_suggested": false}, {"id": "antihypertensive_treatment", "label": "Antihypertensive treatment", "role": "confounder", "adjusted": true, "snippet": "use of antihypertensive medication", "ai_suggested": false}, {"id": "nsaiids", "label": "NSAIDs", "role": "confounder", "adjusted": true, "snippet": "use of NSAIDs", "ai_suggested": false}, {"id": "high_dose_ppi", "label": "High dose PPI", "role": "effect_modifier", "adjusted": true, "snippet": "High doses of PPI increased the risk of incident CKD", "ai_suggested": false}, {"id": "baseline_egfr", "label": "Baseline eGFR", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "socioeconomic_status", "label": "Socioeconomic status", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "gi_indication_for_ppi", "label": "GI indication for PPI", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "alcohol_consumption", "label": "Alcohol consumption", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "other_nephrotoxic_drugs", "label": "Other nephrotoxic drugs", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}], "edges": [{"source": "ppi_use", "target": "incident_ckd", "ai_suggested": false}, {"source": "age", "target": "ppi_use", "ai_suggested": false}, {"source": "age", "target": "incident_ckd", "ai_suggested": false}, {"source": "sex", "target": "ppi_use", "ai_suggested": false}, {"source": "sex", "target": "incident_ckd", "ai_suggested": false}, {"source": "hypertension", "target": "ppi_use", "ai_suggested": false}, {"source": "hypertension", "target": "incident_ckd", "ai_suggested": false}, {"source": "diabetes", "target": "ppi_use", "ai_suggested": false}, {"source": "diabetes", "target": "incident_ckd", "ai_suggested": false}, {"source": "obesity", "target": "ppi_use", "ai_suggested": false}, {"source": "obesity", "target": "incident_ckd", "ai_suggested": false}, {"source": "smoking", "target": "ppi_use", "ai_suggested": false}, {"source": "smoking", "target": "incident_ckd", "ai_suggested": false}, {"source": "antihypertensive_treatment", "target": "ppi_use", "ai_suggested": false}, {"source": "antihypertensive_treatment", "target": "incident_ckd", "ai_suggested": false}, {"source": "nsaiids", "target": "ppi_use", "ai_suggested": false}, {"source": "nsaiids", "target": "incident_ckd", "ai_suggested": false}, {"source": "high_dose_ppi", "target": "incident_ckd", "ai_suggested": false}, {"source": "baseline_egfr", "target": "ppi_use", "ai_suggested": true}, {"source": "baseline_egfr", "target": "incident_ckd", "ai_suggested": true}, {"source": "socioeconomic_status", "target": "ppi_use", "ai_suggested": true}, {"source": "socioeconomic_status", "target": "incident_ckd", "ai_suggested": true}, {"source": "gi_indication_for_ppi", "target": "ppi_use", "ai_suggested": true}, {"source": "gi_indication_for_ppi", "target": "incident_ckd", "ai_suggested": true}, {"source": "alcohol_consumption", "target": "ppi_use", "ai_suggested": true}, {"source": "alcohol_consumption", "target": "incident_ckd", "ai_suggested": true}, {"source": "other_nephrotoxic_drugs", "target": "ppi_use", "ai_suggested": true}, {"source": "other_nephrotoxic_drugs", "target": "incident_ckd", "ai_suggested": true}]}, "dagitty": "dag {\nppi_use [exposure]\nincident_ckd [outcome]\nage [adjusted]\nsex [adjusted]\nhypertension [adjusted]\ndiabetes [adjusted]\nobesity [adjusted]\nsmoking [adjusted]\nantihypertensive_treatment [adjusted]\nnsaiids [adjusted]\nhigh_dose_ppi [adjusted]\nbaseline_egfr\nsocioeconomic_status\ngi_indication_for_ppi\nalcohol_consumption\nother_nephrotoxic_drugs\nppi_use -> incident_ckd\nage -> ppi_use\nage -> incident_ckd\nsex -> ppi_use\nsex -> incident_ckd\nhypertension -> ppi_use\nhypertension -> incident_ckd\ndiabetes -> ppi_use\ndiabetes -> incident_ckd\nobesity -> ppi_use\nobesity -> incident_ckd\nsmoking -> ppi_use\nsmoking -> incident_ckd\nantihypertensive_treatment -> ppi_use\nantihypertensive_treatment -> incident_ckd\nnsaiids -> ppi_use\nnsaiids -> incident_ckd\nhigh_dose_ppi -> incident_ckd\nbaseline_egfr -> ppi_use\nbaseline_egfr -> incident_ckd\nsocioeconomic_status -> ppi_use\nsocioeconomic_status -> incident_ckd\ngi_indication_for_ppi -> ppi_use\ngi_indication_for_ppi -> incident_ckd\nalcohol_consumption -> ppi_use\nalcohol_consumption -> incident_ckd\nother_nephrotoxic_drugs -> ppi_use\nother_nephrotoxic_drugs -> incident_ckd\n}", "open_backdoors": ["Baseline eGFR", "Socioeconomic status", "GI indication for PPI", "Alcohol consumption", "Other nephrotoxic drugs"], "critique": {"missing_confounders": [{"label": "Baseline eGFR", "rationale": "Lower baseline kidney function increases the likelihood of receiving a PPI (due to dose adjustments or comorbid illness) and strongly predicts incident CKD.", "edges_to_add": [{"source": "Baseline eGFR", "target": "ppi_use", "ai_suggested": false}, {"source": "Baseline eGFR", "target": "incident_ckd", "ai_suggested": false}], "evidence": [{"title": "Kidney function decline associated with proton pump inhibitors: results from the ELSA-Brasil cohort | BMC Nephrology | Springer Nature Link", "url": "https://link.springer.com/article/10.1186/s12882-023-03300-4"}, {"title": "The Assessment of the Association of Proton Pump Inhibitor Usage with Chronic Kidney Disease Progression through a Process Mining Approach", "url": "https://www.mdpi.com/2227-9059/12/6/1362"}, {"title": "Proton pump inhibitors and chronic kidney disease risk: a comparative study with histamine-2 receptor antagonists", "url": "https://www.nature.com/articles/s41598-023-48430-9"}], "status": "accepted", "review_note": "Baseline eGFR influences clinicians' decisions to prescribe PPIs (e.g., dose adjustments) and independently predicts future CKD incidence, satisfying the confounding criteria."}, {"label": "Socioeconomic status", "rationale": "Lower SES is associated with higher prevalence of conditions prompting PPI use and with greater risk of CKD through limited healthcare access and higher burden of risk factors.", "edges_to_add": [{"source": "Socioeconomic status", "target": "ppi_use", "ai_suggested": false}, {"source": "Socioeconomic status", "target": "incident_ckd", "ai_suggested": false}], "evidence": [{"title": "Proton Pump Inhibitor Use and Risk of Chronic Kidney Disease", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC4772730"}, {"title": "Proton Pump Inhibitors May Increase Risk for Kidney Disease", "url": "https://www.medscape.com/viewarticle/857060"}, {"title": "Proton Pump Inhibitor Use and the Risk of Chronic Kidney Disease", "url": "https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/2481157"}], "status": "accepted", "review_note": "Lower socioeconomic status increases likelihood of conditions prompting PPI use and independently raises CKD risk through limited healthcare access and higher burden of risk factors, making it a plausible common cause."}, {"label": "GI indication for PPI", "rationale": "Conditions such as GERD or peptic ulcer disease lead to PPI prescribing and may independently increase CKD risk via inflammation or NSAID use.", "edges_to_add": [{"source": "GI indication for PPI", "target": "ppi_use", "ai_suggested": false}, {"source": "GI indication for PPI", "target": "incident_ckd", "ai_suggested": false}], "evidence": [{"title": "Proton Pump Inhibitors and Kidney Disease\u2014GI Upset for the Nephrologist?", "url": "https://www.sciencedirect.com/science/article/pii/S2468024917300050"}, {"title": "Frontiers | Effect of proton pump inhibitors on the risk of chronic kidney disease: A propensity score-based overlap weight analysis using the United Kingdom Biobank", "url": "https://www.frontiersin.org/journals/pharmacology/articles/10.3389/fphar.2022.949699/full"}, {"title": "do proton pump inhibitors cause chronic kidney disease? - PMC", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC5837711"}], "status": "accepted", "review_note": "GI indications such as GERD or peptic ulcer disease drive PPI prescribing and may independently elevate CKD risk via associated NSAID use or inflammation, fulfilling the confounder criteria."}, {"label": "Alcohol consumption", "rationale": "Heavy alcohol use can cause gastritis leading to PPI use and is also a direct risk factor for kidney injury.", "edges_to_add": [{"source": "Alcohol consumption", "target": "ppi_use", "ai_suggested": false}, {"source": "Alcohol consumption", "target": "incident_ckd", "ai_suggested": false}], "evidence": [{"title": "Proton pump inhibitors and chronic kidney disease risk: a comparative study with histamine-2 receptor antagonists", "url": "https://www.nature.com/articles/s41598-023-48430-9"}, {"title": "Proton pump inhibitors use and the risk of chronic kidney disease", "url": "https://www.authorea.com/doi/10.22541/au.170668890.07689514"}, {"title": "Kidney function decline associated with proton pump inhibitors: results from the ELSA-Brasil cohort | BMC Nephrology | Springer Nature Link", "url": "https://link.springer.com/article/10.1186/s12882-023-03300-4"}], "status": "accepted", "review_note": "Heavy alcohol use can provoke gastritis leading to PPI prescription and directly contributes to kidney injury, acting as a common cause of exposure and outcome."}, {"label": "Other nephrotoxic drugs", "rationale": "Use of drugs like loop diuretics, NSAIDs beyond those captured, or certain antibiotics can increase both PPI prescribing (for GI protection) and CKD risk.", "edges_to_add": [{"source": "Other nephrotoxic drugs", "target": "ppi_use", "ai_suggested": false}, {"source": "Other nephrotoxic drugs", "target": "incident_ckd", "ai_suggested": false}], "evidence": [{"title": "Proton Pump Inhibitor Use and Risk of Chronic Kidney Disease", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC4772730"}, {"title": "Proton Pump Inhibitor Use and the Risk of Chronic Kidney ...", "url": "https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/2481157"}, {"title": "Proton Pump Inhibitors and Kidney Disease\u2014GI Upset for the Nephrologist?", "url": "https://www.sciencedirect.com/science/article/pii/S2468024917300050"}], "status": "accepted", "review_note": "Use of additional nephrotoxic drugs (e.g., loop diuretics, certain antibiotics) prompts PPI co\u2011prescribing for GI protection and independently raises CKD risk, meeting the confounder definition."}], "bias_risks": [{"type": "Confounding by indication", "explanation": "PPIs are prescribed for gastrointestinal disorders that may themselves be linked to CKD through NSAID use or chronic inflammation, yet the DAG does not include GI indication as a confounder.", "severity": "high"}, {"type": "Measurement error", "explanation": "PPI exposure was ascertained from prescription records only, missing over\u2011the\u2011counter use and non\u2011adherence, which could dilute or bias the estimated association.", "severity": "medium"}, {"type": "Immortal time bias", "explanation": "The analysis classified PPI use at baseline and during follow\u2011up without properly accounting for time\u2011varying exposure, potentially creating a period where participants must survive to be classified as exposed.", "severity": "medium"}, {"type": "Selection bias", "explanation": "Incident CKD required two abnormal measurements spaced \u22653 months apart; individuals with early rapid decline or who died before the second measurement were excluded, potentially biasing the sample toward healthier survivors.", "severity": "low"}, {"type": "Loss to follow\u2011up", "explanation": "Participants who moved out of the Catalan health system or had incomplete records were censored; if loss is related to both PPI use and kidney function, the HR could be biased.", "severity": "low"}], "overall_assessment": "While the study adjusted for many conventional confounders, important variables such as baseline kidney function, socioeconomic status, and gastrointestinal indications for PPI use were not accounted for, leaving room for substantial residual confounding. Additionally, measurement of PPI exposure relied solely on prescriptions, potentially misclassifying true use, and the analytic approach may be susceptible to immortal time and selection biases. Consequently, the observed association should be interpreted cautiously; it suggests a possible link but does not firmly establish a causal effect of PPIs on incident CKD.", "label": "AI-suggested"}, "models": {"extraction": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B", "critique": "nvidia/nemotron-3-super-120b-a12b", "evidence_search": "tavily", "provider": "Nebius Token Factory"}, "review": {"by": "ai", "model": "nvidia/nemotron-3-super-120b-a12b"}, "dataset_license": "CC-BY-4.0"}
{"id": "04f7a5ae60", "schema_version": "1.0", "status": "approved", "created_at": "2026-10-28T20:06:57Z", "approved_at": "2026-10-28T21:26:48Z", "paper": {"title": "The association between use of proton-pump inhibitors and excess mortality after kidney transplantation: A cohort study", "authors": "Douwes RM, Gomes-Neto AW, Eisenga MF, Van Loon E, Schutten JC, Gans ROB, Naesens M, van den Berg E, Sprangers B, Berger SP, Navis G, Blokzijl H, Meijers B, Bakker SJL, Kuypers D.", "year": "2020", "journal": "PLoS medicine", "doi": "10.1371/journal.pmed.1003140", "pmcid": "PMC7295199", "pmid": "32542023", "license": "cc by"}, "study": {"study_design": "cohort", "population": "Kidney transplant recipients from University Medical Center Groningen and University Hospitals Leuven, November 2008 to March 2011", "sample_size": "703", "follow_up": "median 8.2 years", "exposure": {"name": "PPI use", "definition": "chronic daily PPI use for at least 3 months before and after baseline visit"}, "outcome": {"name": "all-cause mortality", "definition": "death from any cause during follow-up"}, "reported_effect": {"measure": "HR", "value": "1.68", "ci": "1.21-2.33", "adjusted": true}}, "dag": {"nodes": [{"id": "ppi", "label": "PPI use", "role": "exposure", "adjusted": false, "snippet": "chronic PPI users defined as daily use for at least 3 months before and after baseline", "ai_suggested": false}, {"id": "mortality", "label": "All-cause mortality", "role": "outcome", "adjusted": false, "snippet": "194 KTRs died during follow-up", "ai_suggested": false}, {"id": "age", "label": "Age", "role": "confounder", "adjusted": true, "snippet": "adjusted for age, sex, BMI, eGFR, proteinuria, time since transplantation", "ai_suggested": false}, {"id": "sex", "label": "Sex", "role": "confounder", "adjusted": true, "snippet": "adjusted for age, sex, BMI, eGFR, proteinuria, time since transplantation", "ai_suggested": false}, {"id": "bmi", "label": "Body mass index", "role": "confounder", "adjusted": true, "snippet": "adjusted for age, sex, BMI, eGFR, proteinuria, time since transplantation", "ai_suggested": false}, {"id": "egfr", "label": "Estimated glomerular filtration rate", "role": "confounder", "adjusted": true, "snippet": "adjusted for age, sex, BMI, eGFR, proteinuria, time since transplantation", "ai_suggested": false}, {"id": "proteinuria", "label": "Proteinuria", "role": "confounder", "adjusted": true, "snippet": "adjusted for age, sex, BMI, eGFR, proteinuria, time since transplantation", "ai_suggested": false}, {"id": "time_since_transplant", "label": "Time since transplantation", "role": "confounder", "adjusted": true, "snippet": "adjusted for age, sex, BMI, eGFR, proteinuria, time since transplantation", "ai_suggested": false}, {"id": "deceased_donor", "label": "Deceased donor transplant", "role": "confounder", "adjusted": true, "snippet": "adjusted for age, sex, BMI, eGFR, proteinuria, time since transplantation", "ai_suggested": false}, {"id": "primary_renal_disease", "label": "Primary renal disease", "role": "confounder", "adjusted": true, "snippet": "adjusted for age, sex, BMI, eGFR, proteinuria, time since transplantation", "ai_suggested": false}, {"id": "donor_age", "label": "Donor age", "role": "confounder", "adjusted": true, "snippet": "adjusted for donor age, donor sex, donor height, donor weight, donor serum creatinine, HLA mismatches, induction therapy", "ai_suggested": false}, {"id": "hla_mismatches", "label": "HLA mismatches", "role": "confounder", "adjusted": true, "snippet": "adjusted for donor age, donor sex, donor height, donor weight, donor serum creatinine, HLA mismatches, induction therapy", "ai_suggested": false}, {"id": "induction_therapy", "label": "Induction therapy", "role": "confounder", "adjusted": true, "snippet": "adjusted for donor age, donor sex, donor height, donor weight, donor serum creatinine, HLA mismatches, induction therapy", "ai_suggested": false}, {"id": "alcohol_use", "label": "Alcohol use", "role": "confounder", "adjusted": true, "snippet": "adjusted for lifestyle parameters (alcohol use and smoking behavior)", "ai_suggested": false}, {"id": "smoking", "label": "Smoking", "role": "confounder", "adjusted": true, "snippet": "adjusted for lifestyle parameters (alcohol use and smoking behavior)", "ai_suggested": false}, {"id": "antihypertensive", "label": "Antihypertensive drugs", "role": "confounder", "adjusted": true, "snippet": "adjusted for medication use (antihypertensive drugs, platelet inhibitors, vitamin K antagonists, proliferation inhibitors, and calcineurin inhibitors)", "ai_suggested": false}, {"id": "platelet_inhibitors", "label": "Platelet inhibitors", "role": "confounder", "adjusted": true, "snippet": "adjusted for medication use (antihypertensive drugs, platelet inhibitors, vitamin K antagonists, proliferation inhibitors, and calcineurin inhibitors)", "ai_suggested": false}, {"id": "vitamin_k_antagonists", "label": "Vitamin K antagonists", "role": "confounder", "adjusted": true, "snippet": "adjusted for medication use (antihypertensive drugs, platelet inhibitors, vitamin K antagonists, proliferation inhibitors, and calcineurin inhibitors)", "ai_suggested": false}, {"id": "cnis", "label": "Calcineurin inhibitors", "role": "confounder", "adjusted": true, "snippet": "adjusted for medication use (antihypertensive drugs, platelet inhibitors, vitamin K antagonists, proliferation inhibitors, and calcineurin inhibitors)", "ai_suggested": false}, {"id": "diabetes", "label": "Diabetes", "role": "confounder", "adjusted": true, "snippet": "adjusted for comorbidities (diabetes and history of cardiovascular disease)", "ai_suggested": false}, {"id": "cardiovascular_disease", "label": "Cardiovascular disease", "role": "confounder", "adjusted": true, "snippet": "adjusted for comorbidities (diabetes and history of cardiovascular disease)", "ai_suggested": false}, {"id": "plasma_magnesium", "label": "Plasma magnesium", "role": "mediator", "adjusted": true, "snippet": "adjusted for plasma magnesium and serum iron as potential mediators", "ai_suggested": false}, {"id": "serum_iron", "label": "Serum iron", "role": "mediator", "adjusted": true, "snippet": "adjusted for plasma magnesium and serum iron as potential mediators", "ai_suggested": false}, {"id": "peptic_ulcer", "label": "peptic ulcer", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "frailty", "label": "frailty", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "nsaid_use", "label": "nsaid_use", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "albumin", "label": "albumin", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}], "edges": [{"source": "ppi", "target": "mortality", "ai_suggested": false}, {"source": "age", "target": "ppi", "ai_suggested": false}, {"source": "age", "target": "mortality", "ai_suggested": false}, {"source": "sex", "target": "ppi", "ai_suggested": false}, {"source": "sex", "target": "mortality", "ai_suggested": false}, {"source": "bmi", "target": "ppi", "ai_suggested": false}, {"source": "bmi", "target": "mortality", "ai_suggested": false}, {"source": "egfr", "target": "ppi", "ai_suggested": false}, {"source": "egfr", "target": "mortality", "ai_suggested": false}, {"source": "proteinuria", "target": "ppi", "ai_suggested": false}, {"source": "proteinuria", "target": "mortality", "ai_suggested": false}, {"source": "time_since_transplant", "target": "ppi", "ai_suggested": false}, {"source": "time_since_transplant", "target": "mortality", "ai_suggested": false}, {"source": "deceased_donor", "target": "ppi", "ai_suggested": false}, {"source": "deceased_donor", "target": "mortality", "ai_suggested": false}, {"source": "primary_renal_disease", "target": "ppi", "ai_suggested": false}, {"source": "primary_renal_disease", "target": "mortality", "ai_suggested": false}, {"source": "donor_age", "target": "ppi", "ai_suggested": false}, {"source": "donor_age", "target": "mortality", "ai_suggested": false}, {"source": "hla_mismatches", "target": "ppi", "ai_suggested": false}, {"source": "hla_mismatches", "target": "mortality", "ai_suggested": false}, {"source": "induction_therapy", "target": "ppi", "ai_suggested": false}, {"source": "induction_therapy", "target": "mortality", "ai_suggested": false}, {"source": "alcohol_use", "target": "ppi", "ai_suggested": false}, {"source": "alcohol_use", "target": "mortality", "ai_suggested": false}, {"source": "smoking", "target": "ppi", "ai_suggested": false}, {"source": "smoking", "target": "mortality", "ai_suggested": false}, {"source": "antihypertensive", "target": "ppi", "ai_suggested": false}, {"source": "antihypertensive", "target": "mortality", "ai_suggested": false}, {"source": "platelet_inhibitors", "target": "ppi", "ai_suggested": false}, {"source": "platelet_inhibitors", "target": "mortality", "ai_suggested": false}, {"source": "vitamin_k_antagonists", "target": "ppi", "ai_suggested": false}, {"source": "vitamin_k_antagonists", "target": "mortality", "ai_suggested": false}, {"source": "cnis", "target": "ppi", "ai_suggested": false}, {"source": "cnis", "target": "mortality", "ai_suggested": false}, {"source": "diabetes", "target": "ppi", "ai_suggested": false}, {"source": "diabetes", "target": "mortality", "ai_suggested": false}, {"source": "cardiovascular_disease", "target": "ppi", "ai_suggested": false}, {"source": "cardiovascular_disease", "target": "mortality", "ai_suggested": false}, {"source": "plasma_magnesium", "target": "ppi", "ai_suggested": false}, {"source": "plasma_magnesium", "target": "mortality", "ai_suggested": false}, {"source": "serum_iron", "target": "ppi", "ai_suggested": false}, {"source": "serum_iron", "target": "mortality", "ai_suggested": false}, {"source": "peptic_ulcer", "target": "ppi", "ai_suggested": true}, {"source": "peptic_ulcer", "target": "mortality", "ai_suggested": true}, {"source": "frailty", "target": "ppi", "ai_suggested": true}, {"source": "frailty", "target": "mortality", "ai_suggested": true}, {"source": "nsaid_use", "target": "ppi", "ai_suggested": true}, {"source": "nsaid_use", "target": "mortality", "ai_suggested": true}, {"source": "albumin", "target": "ppi", "ai_suggested": true}, {"source": "albumin", "target": "mortality", "ai_suggested": true}]}, "dagitty": "dag {\nppi [exposure]\nmortality [outcome]\nage [adjusted]\nsex [adjusted]\nbmi [adjusted]\negfr [adjusted]\nproteinuria [adjusted]\ntime_since_transplant [adjusted]\ndeceased_donor [adjusted]\nprimary_renal_disease [adjusted]\ndonor_age [adjusted]\nhla_mismatches [adjusted]\ninduction_therapy [adjusted]\nalcohol_use [adjusted]\nsmoking [adjusted]\nantihypertensive [adjusted]\nplatelet_inhibitors [adjusted]\nvitamin_k_antagonists [adjusted]\ncnis [adjusted]\ndiabetes [adjusted]\ncardiovascular_disease [adjusted]\nplasma_magnesium [adjusted]\nserum_iron [adjusted]\npeptic_ulcer\nfrailty\nnsaid_use\nalbumin\nppi -> mortality\nage -> ppi\nage -> mortality\nsex -> ppi\nsex -> mortality\nbmi -> ppi\nbmi -> mortality\negfr -> ppi\negfr -> mortality\nproteinuria -> ppi\nproteinuria -> mortality\ntime_since_transplant -> ppi\ntime_since_transplant -> mortality\ndeceased_donor -> ppi\ndeceased_donor -> mortality\nprimary_renal_disease -> ppi\nprimary_renal_disease -> mortality\ndonor_age -> ppi\ndonor_age -> mortality\nhla_mismatches -> ppi\nhla_mismatches -> mortality\ninduction_therapy -> ppi\ninduction_therapy -> mortality\nalcohol_use -> ppi\nalcohol_use -> mortality\nsmoking -> ppi\nsmoking -> mortality\nantihypertensive -> ppi\nantihypertensive -> mortality\nplatelet_inhibitors -> ppi\nplatelet_inhibitors -> mortality\nvitamin_k_antagonists -> ppi\nvitamin_k_antagonists -> mortality\ncnis -> ppi\ncnis -> mortality\ndiabetes -> ppi\ndiabetes -> mortality\ncardiovascular_disease -> ppi\ncardiovascular_disease -> mortality\nplasma_magnesium -> ppi\nplasma_magnesium -> mortality\nserum_iron -> ppi\nserum_iron -> mortality\npeptic_ulcer -> ppi\npeptic_ulcer -> mortality\nfrailty -> ppi\nfrailty -> mortality\nnsaid_use -> ppi\nnsaid_use -> mortality\nalbumin -> ppi\nalbumin -> mortality\n}", "open_backdoors": ["peptic ulcer", "frailty", "nsaid_use", "albumin"], "critique": {"missing_confounders": [{"label": "peptic ulcer", "rationale": "Patients with a prior peptic ulcer are more likely to be prescribed PPIs for ulcer prevention and also have higher mortality due to ulcer complications or comorbid conditions.", "edges_to_add": [{"source": "peptic ulcer", "target": "ppi", "ai_suggested": false}, {"source": "peptic ulcer", "target": "mortality", "ai_suggested": false}], "evidence": [{"title": "Association of Proton Pump Inhibitor Use With All-Cause ...", "url": "https://www.gastrojournal.org/article/S0016-5085(22)00729-6/fulltext"}, {"title": "Proton Pump Inhibitors Not Linked to Increased All-Cause, ...", "url": "https://www.gastroenterologyadvisor.com/news/proton-pump-inhibitors-not-linked-to-increased-all-cause-major-cause-mortality-risk"}, {"title": "Association of Proton Pump Inhibitor Use With All-cause and ...", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC9509450"}], "status": "accepted", "review_note": "Prior peptic ulcer increases the likelihood of chronic PPI use and mortality, is not adjusted for, and is not caused by PPI use (it generally precedes exposure)."}, {"label": "charlson", "rationale": "A higher Charlson comorbidity index indicates greater overall disease burden, leading to both increased PPI prescribing (for GI protection) and elevated mortality risk.", "edges_to_add": [{"source": "charlson", "target": "ppi", "ai_suggested": false}, {"source": "charlson", "target": "mortality", "ai_suggested": false}], "evidence": [{"title": "Association of Proton Pump Inhibitor Use With All-cause and ...", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC9509450"}, {"title": "Study: 4 causes of mortality were associated with PPI use", "url": "https://www.facebook.com/medscape/posts/study-4-causes-of-mortality-were-associated-with-ppi-use-circulatory-system-geni/10157940769169189"}], "status": "rejected", "review_note": "The Charlson comorbidity index is a composite of comorbidities already adjusted for individually (e.g., diabetes, cardiovascular disease), thus serving as a close proxy."}, {"label": "frailty", "rationale": "Frailer transplant recipients often receive more medications, including PPIs, and have reduced physiological reserve, increasing mortality risk.", "edges_to_add": [{"source": "frailty", "target": "ppi", "ai_suggested": false}, {"source": "frailty", "target": "mortality", "ai_suggested": false}], "evidence": [{"title": "Proton Pump Inhibitors Not Linked to Increased All-Cause, ...", "url": "https://www.gastroenterologyadvisor.com/news/proton-pump-inhibitors-not-linked-to-increased-all-cause-major-cause-mortality-risk"}, {"title": "Use of Proton Pump Inhibitors and Mortality Among ...", "url": "https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/486915"}, {"title": "Association of Proton Pump Inhibitor Use With All-cause and ...", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC9509450"}], "status": "accepted", "review_note": "Frailty raises both PPI prescribing and mortality risk, is not captured by the adjusted covariates, and is not induced by PPI use."}, {"label": "nsaid_use", "rationale": "NSAID use prompts prophylactic PPI prescribing to prevent gastrointestinal injury, while NSAIDs themselves increase mortality risk via bleeding, cardiovascular, or renal events.", "edges_to_add": [{"source": "nsaid_use", "target": "ppi", "ai_suggested": false}, {"source": "nsaid_use", "target": "mortality", "ai_suggested": false}], "evidence": [{"title": "Association of Proton Pump Inhibitor Use With All-cause and Cause-specific Mortality", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC9509450"}, {"title": "Proton pump inhibitors and risk of all\u2010cause and ...", "url": "https://researchonline.lshtm.ac.uk/id/eprint/4659251/1/bcp.14728.pdf"}, {"title": "Association of Proton Pump Inhibitor Use With All-Cause ...", "url": "https://www.gastrojournal.org/article/S0016-5085(22)00729-6/fulltext"}], "status": "accepted", "review_note": "NSAID use leads to prophylactic PPI prescribing and higher mortality, is not adjusted for, and is not caused by PPI use."}, {"label": "albumin", "rationale": "Low serum albumin reflects malnutrition/inflammation, which is associated with both higher likelihood of PPI use (due to illness) and higher mortality.", "edges_to_add": [{"source": "albumin", "target": "ppi", "ai_suggested": false}, {"source": "albumin", "target": "mortality", "ai_suggested": false}], "evidence": [{"title": "Association of Proton Pump Inhibitor Use With All-cause and Cause-specific Mortality", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC9509450"}, {"title": "Proton Pump Inhibitors Not Linked to Increased All-Cause, ...", "url": "https://www.gastroenterologyadvisor.com/news/proton-pump-inhibitors-not-linked-to-increased-all-cause-major-cause-mortality-risk"}, {"title": "Proton Pump Inhibitor Usage and the Risk of Mortality in ...", "url": "https://www.sciencedirect.com/science/article/pii/S2468024917304369"}], "status": "accepted", "review_note": "Low serum albumin reflects malnutrition/inflammation that increases PPI use and mortality, is not adjusted for, and is not caused by PPI use."}], "bias_risks": [{"type": "confounding by indication", "explanation": "Patients with gastrointestinal symptoms or ulcer history are more likely to receive PPIs and also have higher mortality, creating confounding if not fully captured.", "severity": "high"}, {"type": "immortal time bias", "explanation": "Exposure required PPI use for at least 3 months before and after baseline; person-time before meeting this criterion was misclassified as exposed, potentially biasing the HR away from the null.", "severity": "medium"}, {"type": "measurement error", "explanation": "PPI use was ascertained from self\u2011report or pharmacy records, which may not reflect actual ingestion, leading to nondifferential misclassification that could attenuate or exaggerate the association.", "severity": "medium"}, {"type": "confounding by indication", "explanation": "Unmeasured intensity of immunosuppression (e.g., mycophenolate dose, corticosteroid dose) may influence both PPI prescribing (for GI protection) and mortality.", "severity": "medium"}, {"type": "loss to follow-up", "explanation": "If sicker participants dropped out differentially, the observed association could be biased, although follow\u2011up was relatively complete in this cohort.", "severity": "low"}], "overall_assessment": "While the authors adjusted for a broad set of demographic, clinical, and treatment variables, important confounders such as ulcer history, comorbidity burden, frailty, NSAID use, and nutritional status were not accounted for. Additionally, the exposure definition may introduce immortal time bias, and PPI measurement could be imperfect. Consequently, the observed association should be interpreted cautiously; it suggests a correlation but does not firmly establish a causal effect of PPIs on mortality in kidney transplant recipients.", "label": "AI-suggested"}, "models": {"extraction": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B", "critique": "nvidia/nemotron-3-super-120b-a12b", "evidence_search": "tavily", "provider": "Nebius Token Factory"}, "review": {"by": "ai", "model": "nvidia/nemotron-3-super-120b-a12b"}, "dataset_license": "CC-BY-4.0"}
{"id": "c29b8b42a5", "schema_version": "1.0", "status": "approved", "created_at": "2026-10-28T20:12:00Z", "approved_at": "2026-10-28T21:42:01Z", "paper": {"title": "Association of Total Daily Physical Activity and Fragmented Physical Activity With Mortality in Older Adults", "authors": "Wanigatunga AA, Di J, Zipunnikov V, Urbanek JK, Kuo PL, Simonsick EM, Ferrucci L, Schrack JA.", "year": "2019", "journal": "JAMA network open", "doi": "10.1001/jamanetworkopen.2019.12352", "pmcid": "PMC6777397", "pmid": "31577355", "license": "cc by"}, "study": {"study_design": "cohort", "population": "older adults aged 65 years and older from the Baltimore Longitudinal Study of Aging", "sample_size": "548 (11.1% died)", "follow_up": "mean 4.4 years", "exposure": {"name": "activity fragmentation", "definition": "active-to-sedentary transition probability derived from accelerometer data"}, "outcome": {"name": "all-cause mortality", "definition": "death from any cause identified via National Death Index and physician adjudication"}, "reported_effect": {"measure": "HR", "value": "1.49", "ci": "1.02-2.19", "adjusted": true}}, "dag": {"nodes": [{"id": "exposure", "label": "activity fragmentation", "role": "exposure", "adjusted": false, "snippet": "more fragmented physical activity patterns were associated with greater mortality risk", "ai_suggested": false}, {"id": "outcome", "label": "all-cause mortality", "role": "outcome", "adjusted": false, "snippet": "61 participants (11.1%) died", "ai_suggested": false}, {"id": "age", "label": "age", "role": "confounder", "adjusted": true, "snippet": "adjusted for age, sex, race/ethnicity, body mass index, smoking history, employment, self-reported health, grip strength, usual gait speed, comorbidities, and \u2026", "ai_suggested": false}, {"id": "sex", "label": "sex", "role": "confounder", "adjusted": true, "snippet": "adjusted for age, sex, race/ethnicity, body mass index, smoking history, employment, self-reported health, grip strength, usual gait speed, comorbidities, and \u2026", "ai_suggested": false}, {"id": "race_ethnicity", "label": "race/ethnicity", "role": "confounder", "adjusted": true, "snippet": "adjusted for age, sex, race/ethnicity, body mass index, smoking history, employment, self-reported health, grip strength, usual gait speed, comorbidities, and \u2026", "ai_suggested": false}, {"id": "bmi", "label": "body mass index", "role": "confounder", "adjusted": true, "snippet": "adjusted for age, sex, race/ethnicity, body mass index, smoking history, employment, self-reported health, grip strength, usual gait speed, comorbidities, and \u2026", "ai_suggested": false}, {"id": "smoking", "label": "smoking history", "role": "confounder", "adjusted": true, "snippet": "adjusted for age, sex, race/ethnicity, body mass index, smoking history, employment, self-reported health, grip strength, usual gait speed, comorbidities, and \u2026", "ai_suggested": false}, {"id": "employment", "label": "employment status", "role": "confounder", "adjusted": true, "snippet": "adjusted for age, sex, race/ethnicity, body mass index, smoking history, employment, self-reported health, grip strength, usual gait speed, comorbidities, and \u2026", "ai_suggested": false}, {"id": "self_reported_health", "label": "self-reported health", "role": "confounder", "adjusted": true, "snippet": "adjusted for age, sex, race/ethnicity, body mass index, smoking history, employment, self-reported health, grip strength, usual gait speed, comorbidities, and \u2026", "ai_suggested": false}, {"id": "grip_strength", "label": "grip strength", "role": "confounder", "adjusted": true, "snippet": "adjusted for age, sex, race/ethnicity, body mass index, smoking history, employment, self-reported health, grip strength, usual gait speed, comorbidities, and \u2026", "ai_suggested": false}, {"id": "gait_speed", "label": "usual gait speed", "role": "confounder", "adjusted": true, "snippet": "adjusted for age, sex, race/ethnicity, body mass index, smoking history, employment, self-reported health, grip strength, usual gait speed, comorbidities, and \u2026", "ai_suggested": false}, {"id": "comorbidities", "label": "comorbidities", "role": "confounder", "adjusted": true, "snippet": "adjusted for age, sex, race/ethnicity, body mass index, smoking history, employment, self-reported health, grip strength, usual gait speed, comorbidities, and \u2026", "ai_suggested": false}, {"id": "device_wear_time", "label": "device wear time", "role": "confounder", "adjusted": true, "snippet": "adjusted for age, sex, race/ethnicity, body mass index, smoking history, employment, self-reported health, grip strength, usual gait speed, comorbidities, and \u2026", "ai_suggested": false}, {"id": "short_bouts", "label": "activity bouts <5 minutes", "role": "mediator", "adjusted": false, "snippet": "more frequently engaging in activity bouts lasting less than 5 minutes was associated with greater mortality risk", "ai_suggested": false}, {"id": "frailty_index", "label": "frailty index", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "cognitive_function", "label": "cognitive function", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "depressive_symptoms", "label": "depressive symptoms", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "education_level", "label": "education level", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}], "edges": [{"source": "exposure", "target": "outcome", "ai_suggested": false}, {"source": "age", "target": "exposure", "ai_suggested": false}, {"source": "age", "target": "outcome", "ai_suggested": false}, {"source": "sex", "target": "exposure", "ai_suggested": false}, {"source": "sex", "target": "outcome", "ai_suggested": false}, {"source": "race_ethnicity", "target": "exposure", "ai_suggested": false}, {"source": "race_ethnicity", "target": "outcome", "ai_suggested": false}, {"source": "bmi", "target": "exposure", "ai_suggested": false}, {"source": "bmi", "target": "outcome", "ai_suggested": false}, {"source": "smoking", "target": "exposure", "ai_suggested": false}, {"source": "smoking", "target": "outcome", "ai_suggested": false}, {"source": "employment", "target": "exposure", "ai_suggested": false}, {"source": "employment", "target": "outcome", "ai_suggested": false}, {"source": "self_reported_health", "target": "exposure", "ai_suggested": false}, {"source": "self_reported_health", "target": "outcome", "ai_suggested": false}, {"source": "grip_strength", "target": "exposure", "ai_suggested": false}, {"source": "grip_strength", "target": "outcome", "ai_suggested": false}, {"source": "gait_speed", "target": "exposure", "ai_suggested": false}, {"source": "gait_speed", "target": "outcome", "ai_suggested": false}, {"source": "comorbidities", "target": "exposure", "ai_suggested": false}, {"source": "comorbidities", "target": "outcome", "ai_suggested": false}, {"source": "device_wear_time", "target": "exposure", "ai_suggested": false}, {"source": "device_wear_time", "target": "outcome", "ai_suggested": false}, {"source": "exposure", "target": "short_bouts", "ai_suggested": false}, {"source": "short_bouts", "target": "outcome", "ai_suggested": false}, {"source": "frailty_index", "target": "exposure", "ai_suggested": true}, {"source": "frailty_index", "target": "outcome", "ai_suggested": true}, {"source": "cognitive_function", "target": "exposure", "ai_suggested": true}, {"source": "cognitive_function", "target": "outcome", "ai_suggested": true}, {"source": "depressive_symptoms", "target": "exposure", "ai_suggested": true}, {"source": "depressive_symptoms", "target": "outcome", "ai_suggested": true}, {"source": "education_level", "target": "exposure", "ai_suggested": true}, {"source": "education_level", "target": "outcome", "ai_suggested": true}]}, "dagitty": "dag {\nexposure [exposure]\noutcome [outcome]\nage [adjusted]\nsex [adjusted]\nrace_ethnicity [adjusted]\nbmi [adjusted]\nsmoking [adjusted]\nemployment [adjusted]\nself_reported_health [adjusted]\ngrip_strength [adjusted]\ngait_speed [adjusted]\ncomorbidities [adjusted]\ndevice_wear_time [adjusted]\nshort_bouts\nfrailty_index\ncognitive_function\ndepressive_symptoms\neducation_level\nexposure -> outcome\nage -> exposure\nage -> outcome\nsex -> exposure\nsex -> outcome\nrace_ethnicity -> exposure\nrace_ethnicity -> outcome\nbmi -> exposure\nbmi -> outcome\nsmoking -> exposure\nsmoking -> outcome\nemployment -> exposure\nemployment -> outcome\nself_reported_health -> exposure\nself_reported_health -> outcome\ngrip_strength -> exposure\ngrip_strength -> outcome\ngait_speed -> exposure\ngait_speed -> outcome\ncomorbidities -> exposure\ncomorbidities -> outcome\ndevice_wear_time -> exposure\ndevice_wear_time -> outcome\nexposure -> short_bouts\nshort_bouts -> outcome\nfrailty_index -> exposure\nfrailty_index -> outcome\ncognitive_function -> exposure\ncognitive_function -> outcome\ndepressive_symptoms -> exposure\ndepressive_symptoms -> outcome\neducation_level -> exposure\neducation_level -> outcome\n}", "open_backdoors": ["frailty index", "cognitive function", "depressive symptoms", "education level"], "critique": {"missing_confounders": [{"label": "frailty index", "rationale": "Frailty reflects decreased physiological reserve and predicts both higher activity fragmentation and mortality risk.", "edges_to_add": [{"source": "frailty index", "target": "exposure", "ai_suggested": false}, {"source": "frailty index", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "High Fragmented Physical Activity as an Early Risk ...", "url": "https://karger.com/ger/article/69/3/370/835249/High-Fragmented-Physical-Activity-as-an-Early-Risk"}, {"title": "Frailty and physical activity: are all physical activity variables ...", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC12351540"}, {"title": "Physical Activity Fragmentation and All-Cause Mortality in ...", "url": "https://www.sciencedirect.com/science/article/pii/S2590059526000658"}], "status": "accepted", "review_note": "Frailty index is a plausible common cause of activity fragmentation and mortality and is not fully captured by the adjusted covariates such as comorbidities, grip strength, or gait speed."}, {"label": "cognitive function", "rationale": "Cognitive impairment can lead to more fragmented activity patterns and is independently associated with increased mortality.", "edges_to_add": [{"source": "cognitive function", "target": "exposure", "ai_suggested": false}, {"source": "cognitive function", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Association between cognitive function and all-cause mortality among US older adults with cardiovascular-kidney-metabolic syndrome | BMC Geriatrics | Springer Nature Link", "url": "https://link.springer.com/article/10.1186/s12877-026-07055-z"}, {"title": "\"The association of physical activity fragmentation with all-cause mort\" by Mauro F F Mediano, Yejin Mok et al.", "url": "https://scholarlyexchange.childrensmercy.org/papers/6407"}, {"title": "Activity Intensity and All-Cause Mortality Following Fall Injury Among Older Adults: Results from a 12-Year National Survey", "url": "https://www.mdpi.com/2227-9032/13/19/2530"}], "status": "accepted", "review_note": "Cognitive function can influence both activity fragmentation patterns and mortality risk and is not accounted for by any of the adjusted variables."}, {"label": "depressive symptoms", "rationale": "Depression is linked to altered physical activity patterns and higher mortality, potentially confounding the observed association.", "edges_to_add": [{"source": "depressive symptoms", "target": "exposure", "ai_suggested": false}, {"source": "depressive symptoms", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Frontiers | Depressive symptoms and all-cause mortality among middle-aged and older people in China and associations with chronic diseases", "url": "https://www.frontiersin.org/journals/public-health/articles/10.3389/fpubh.2024.1381273/full"}, {"title": "Association Between Depression and All-Cause Mortality : A Cohort Study Based on NHANES Data", "url": "https://www.medrxiv.org/content/10.1101/2025.04.29.25326650v1.full"}, {"title": "Association between physical activity and all-cause mortality in patients with depression: a prospective cohort study based on NHANES data", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC12040813"}], "status": "accepted", "review_note": "Depressive symptoms may lead to more fragmented activity and higher mortality and are not directly adjusted for in the model."}, {"label": "education level", "rationale": "Lower education correlates with less structured daily activity and higher mortality risk, acting as a socioeconomic confounder.", "edges_to_add": [{"source": "education level", "target": "exposure", "ai_suggested": false}, {"source": "education level", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "The association of physical activity fragmentation with all-cause ...", "url": "https://www.sciencedirect.com/science/article/pii/S2667193X25000067"}, {"title": "Activity Intensity and All-Cause Mortality Following Fall Injury Among Older Adults: Results from a 12-Year National Survey", "url": "https://www.mdpi.com/2227-9032/13/19/2530"}, {"title": "Low educational attainment is associated with higher all-cause and cardiovascular mortality in the United States adult population | BMC Public Health | Springer Nature Link", "url": "https://link.springer.com/article/10.1186/s12889-023-15621-y"}], "status": "accepted", "review_note": "Education level affects both activity fragmentation and mortality and is not included among the adjusted covariates (employment status does not serve as a close proxy)."}, {"label": "cardiovascular medication", "rationale": "Use of cardiovascular drugs indicates underlying heart disease, which may increase activity fragmentation and mortality.", "edges_to_add": [{"source": "cardiovascular medication", "target": "exposure", "ai_suggested": false}, {"source": "cardiovascular medication", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "The association of physical activity fragmentation with all-cause ...", "url": "https://scholarlyexchange.childrensmercy.org/cgi/viewcontent.cgi?article=7369&context=papers"}, {"title": "Physical Activity Fragmentation and All-Cause Mortality in ...", "url": "https://www.kidneymedicinejournal.org/article/S2590-0595(26)00065-8/fulltext"}, {"title": "Activity Fragmentation and All-Cause Mortality in United States Adults", "url": "https://www.sciencedirect.com/science/article/pii/S1047279726002917"}], "status": "rejected", "review_note": "Cardiovascular medication use is likely a proxy for comorbidities, which is already adjusted for, so adjusting for it would not add new confounding control."}], "bias_risks": [{"type": "measurement error", "explanation": "Activity fragmentation derived from accelerometer data may be misclassified due to non\u2011wear periods or algorithm thresholds, biasing the exposure\u2013outcome estimate.", "severity": "medium"}, {"type": "confounding by indication", "explanation": "Residual confounding by unmeasured frailty or subclinical disease may remain despite adjustment for comorbidities and health status.", "severity": "high"}, {"type": "selection bias", "explanation": "The sample consists of well\u2011functioning volunteers from the Baltimore Longitudinal Study of Aging, limiting generalizability and potentially inducing a healthy\u2011user effect.", "severity": "medium"}, {"type": "reverse causation", "explanation": "Early degenerative processes could increase fragmentation before death, and baseline health measures may not fully capture this temporal direction.", "severity": "medium"}, {"type": "loss to follow-up", "explanation": "Although mortality ascertainment is complete, occasional missing accelerometer data could lead to differential exclusion of frailer participants.", "severity": "low"}], "overall_assessment": "While the study adjusted for a broad set of demographic, behavioral, and clinical covariates, important confounders such as frailty, cognitive function, and depressive symptoms were not measured, leaving room for residual confounding. Additionally, measurement error in the exposure and selection of a highly functioning cohort further weaken the causal interpretation. Consequently, the observed association between activity fragmentation and mortality should be viewed as suggestive rather than definitive evidence of a causal effect.", "label": "AI-suggested"}, "models": {"extraction": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B", "critique": "nvidia/nemotron-3-super-120b-a12b", "evidence_search": "tavily", "provider": "Nebius Token Factory"}, "review": {"by": "human", "model": null}, "dataset_license": "CC-BY-4.0"}
{"id": "6e1609603b", "schema_version": "1.0", "status": "approved", "created_at": "2026-10-28T20:08:55Z", "approved_at": "2026-10-28T20:23:12Z", "paper": {"title": "Ultra-processed food intake in association with BMI change and risk of overweight and obesity: A\u00a0prospective analysis of the French NutriNet-Sant\u00e9 cohort", "authors": "Beslay M, Srour B, M\u00e9jean C, All\u00e8s B, Fiolet T, Debras C, Chazelas E, Deschasaux M, Wendeu-Foyet MG, Hercberg S, Galan P, Monteiro CA, Deschamps V, Calixto Andrade G, Kesse-Guyot E, Julia C, Touvier M.", "year": "2020", "journal": "PLoS medicine", "doi": "10.1371/journal.pmed.1003256", "pmcid": "PMC7451582", "pmid": "32853224", "license": "cc by"}, "study": {"study_design": "cohort", "population": "Adults aged 18\u201373.3 years from the French NutriNet-Sant\u00e9 cohort, 2009\u20132019", "sample_size": "110,260", "follow_up": "median 4.1 years", "exposure": {"name": "UPF proportion", "definition": "percentage of ultra-processed foods in the diet, measured as weight ratio of UPF to total diet"}, "outcome": {"name": "BMI change", "definition": "absolute change in body mass index over time, assessed via repeated anthropometric questionnaires"}, "reported_effect": {"measure": "HR", "value": "1.11", "ci": "1.08-1.14", "adjusted": true}}, "dag": {"nodes": [{"id": "exposure", "label": "UPF proportion", "role": "exposure", "adjusted": false, "snippet": "proportion of UPF in the diet", "ai_suggested": false}, {"id": "outcome", "label": "BMI change", "role": "outcome", "adjusted": false, "snippet": "absolute change in BMI over time", "ai_suggested": false}, {"id": "age", "label": "Age", "role": "confounder", "adjusted": true, "snippet": "adjusted for age", "ai_suggested": false}, {"id": "sex", "label": "Sex", "role": "confounder", "adjusted": true, "snippet": "adjusted for sex", "ai_suggested": false}, {"id": "education", "label": "Educational level", "role": "confounder", "adjusted": true, "snippet": "adjusted for educational level", "ai_suggested": false}, {"id": "marital_status", "label": "Marital status", "role": "confounder", "adjusted": true, "snippet": "adjusted for marital status", "ai_suggested": false}, {"id": "smoking", "label": "Smoking status", "role": "confounder", "adjusted": true, "snippet": "adjusted for smoking status", "ai_suggested": false}, {"id": "physical_activity", "label": "Physical activity level", "role": "confounder", "adjusted": true, "snippet": "adjusted for physical activity level", "ai_suggested": false}, {"id": "energy_intake", "label": "Energy intake", "role": "confounder", "adjusted": true, "snippet": "adjusted for energy intake", "ai_suggested": false}, {"id": "alcohol_intake", "label": "Alcohol intake", "role": "confounder", "adjusted": true, "snippet": "adjusted for alcohol intake", "ai_suggested": false}, {"id": "diet_quality", "label": "Dietary quality indicators", "role": "confounder", "adjusted": true, "snippet": "adjusted for sugar, fibre, sodium, and SFA intake", "ai_suggested": false}, {"id": "dietary_patterns", "label": "Dietary patterns", "role": "confounder", "adjusted": true, "snippet": "adjusted for dietary patterns", "ai_suggested": false}, {"id": "household_income", "label": "Household income", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "sleep_duration", "label": "Sleep duration", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "baseline_bmi", "label": "Baseline BMI", "role": "confounder", "adjusted": true, "snippet": "Analyses were tested with and without adjustment for baseline BMI.", "ai_suggested": false}], "edges": [{"source": "exposure", "target": "outcome", "ai_suggested": false}, {"source": "age", "target": "exposure", "ai_suggested": false}, {"source": "age", "target": "outcome", "ai_suggested": false}, {"source": "sex", "target": "exposure", "ai_suggested": false}, {"source": "sex", "target": "outcome", "ai_suggested": false}, {"source": "education", "target": "exposure", "ai_suggested": false}, {"source": "education", "target": "outcome", "ai_suggested": false}, {"source": "marital_status", "target": "exposure", "ai_suggested": false}, {"source": "marital_status", "target": "outcome", "ai_suggested": false}, {"source": "smoking", "target": "exposure", "ai_suggested": false}, {"source": "smoking", "target": "outcome", "ai_suggested": false}, {"source": "physical_activity", "target": "exposure", "ai_suggested": false}, {"source": "physical_activity", "target": "outcome", "ai_suggested": false}, {"source": "energy_intake", "target": "exposure", "ai_suggested": false}, {"source": "energy_intake", "target": "outcome", "ai_suggested": false}, {"source": "alcohol_intake", "target": "exposure", "ai_suggested": false}, {"source": "alcohol_intake", "target": "outcome", "ai_suggested": false}, {"source": "diet_quality", "target": "exposure", "ai_suggested": false}, {"source": "diet_quality", "target": "outcome", "ai_suggested": false}, {"source": "dietary_patterns", "target": "exposure", "ai_suggested": false}, {"source": "dietary_patterns", "target": "outcome", "ai_suggested": false}, {"source": "household_income", "target": "exposure", "ai_suggested": true}, {"source": "household_income", "target": "outcome", "ai_suggested": true}, {"source": "sleep_duration", "target": "exposure", "ai_suggested": true}, {"source": "baseline_bmi", "target": "exposure", "ai_suggested": false}, {"source": "baseline_bmi", "target": "outcome", "ai_suggested": false}]}, "dagitty": "dag {\nexposure [exposure]\noutcome [outcome]\nage [adjusted]\nsex [adjusted]\neducation [adjusted]\nmarital_status [adjusted]\nsmoking [adjusted]\nphysical_activity [adjusted]\nenergy_intake [adjusted]\nalcohol_intake [adjusted]\ndiet_quality [adjusted]\ndietary_patterns [adjusted]\nhousehold_income\nsleep_duration\nbaseline_bmi [adjusted]\nexposure -> outcome\nage -> exposure\nage -> outcome\nsex -> exposure\nsex -> outcome\neducation -> exposure\neducation -> outcome\nmarital_status -> exposure\nmarital_status -> outcome\nsmoking -> exposure\nsmoking -> outcome\nphysical_activity -> exposure\nphysical_activity -> outcome\nenergy_intake -> exposure\nenergy_intake -> outcome\nalcohol_intake -> exposure\nalcohol_intake -> outcome\ndiet_quality -> exposure\ndiet_quality -> outcome\ndietary_patterns -> exposure\ndietary_patterns -> outcome\nhousehold_income -> exposure\nhousehold_income -> outcome\nsleep_duration -> exposure\nbaseline_bmi -> exposure\nbaseline_bmi -> outcome\n}", "open_backdoors": ["Household income"], "critique": {"missing_confounders": [{"label": "Household income", "rationale": "Income affects food purchasing power and thus the proportion of ultra\u2011processed foods consumed, while also being related to BMI change through access to healthier lifestyles and healthcare.", "edges_to_add": [{"source": "Household income", "target": "exposure", "ai_suggested": false}, {"source": "Household income", "target": "outcome", "ai_suggested": false}], "evidence": [{"title": "Ultra-processed Foods, Weight Gain, and Co-morbidity Risk | Current Obesity Reports | Springer Nature Link", "url": "https://link.springer.com/article/10.1007/s13679-021-00460-y"}, {"title": "Ultra-Processed Food Consumption and Obesity: A Narrative Review of Their Association and Potential Mechanisms", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC11799601"}, {"title": "New study reveals stark disparities in ultra-processed food purchases among U.S. households - Global Food Research Program", "url": "https://www.globalfoodresearchprogram.org/new-study-reveals-stark-disparities-in-ultra-processed-food-purchases-among-u-s-households"}], "status": "accepted", "review_note": ""}, {"label": "Sleep duration", "rationale": "Short or poor sleep is linked to increased appetite for ultra\u2011processed foods and independently predicts weight gain, making it a common cause of exposure and outcome.", "edges_to_add": [{"source": "Sleep duration", "target": "exposure", "ai_suggested": false}, {"source": "Sleep duration", "target": "outake", "ai_suggested": false}], "evidence": [{"title": "Sleep Duration and Timing in Relation to BMI and Readiness ...", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC12288046"}, {"title": "Frontiers | Association between sleep behaviors and adiposity indices among U.S. adults: a cross-sectional study", "url": "https://www.frontiersin.org/journals/nutrition/articles/10.3389/fnut.2025.1526422/full"}, {"title": "The association between sleep duration and detailed ...", "url": "https://www.repository.cam.ac.uk/bitstreams/79afda44-6010-436b-be77-bba2d9715ecd/download"}], "status": "accepted", "review_note": ""}], "bias_risks": [{"type": "selection bias", "explanation": "The NutriNet-Sant\u00e9 cohort consists of volunteers interested in nutrition, potentially leading to a healthier\u2011than\u2011average sample and limiting generalizability.", "severity": "medium"}, {"type": "measurement error", "explanation": "Ultra\u2011processed food classification relies on self\u2011reported 24\u2011hour dietary records, which may misclassify processing level and introduce non\u2011differential error in the exposure.", "severity": "medium"}, {"type": "loss to follow-up", "explanation": "Over the 10\u2011year follow\u2011up, participants who dropped out may differ in both UPF consumption and weight trajectory, biasing the estimated association if not random.", "severity": "low"}], "overall_assessment": "While the study controlled for a broad set of demographic and lifestyle confounders, important variables such as baseline BMI, socioeconomic status beyond education, and sleep duration were not accounted for, leaving room for residual confounding. Additionally, selection bias from the volunteer cohort and measurement error in dietary assessment could attenuate or exaggerate the observed association. Overall, the evidence for a causal effect of UPF consumption on BMI change is suggestive but should be interpreted cautiously pending further validation.", "label": "AI-suggested"}, "models": {"extraction": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B", "critique": "nvidia/nemotron-3-super-120b-a12b", "evidence_search": "tavily", "provider": "Nebius Token Factory"}, "review": {"by": "human", "model": null}, "dataset_license": "CC-BY-4.0"}
{"id": "3447637879", "schema_version": "1.0", "status": "approved", "created_at": "2026-10-28T20:07:39Z", "approved_at": "2026-10-28T21:26:38Z", "paper": {"title": "Proton pump inhibitor use and risk of hip fracture in patients with type 2 diabetes", "authors": "Chou YS, Jiang HJ, Chen CH, Ho PS, Lee TC.", "year": "2020", "journal": "Scientific reports", "doi": "10.1038/s41598-020-70712-9", "pmcid": "PMC7443131", "pmid": "32826941", "license": "cc by"}, "study": {"study_design": "cohort", "population": "Patients aged \u226518 years with newly diagnosed T2DM in Taiwan between 2000 and 2008", "sample_size": "44,431 (exposed) and 177,364 (unexposed) after matching", "follow_up": "up to 5 years", "exposure": {"name": "PPI use", "definition": "Long-term PPI use defined as DDD \u2265 28 for at least 3 years"}, "outcome": {"name": "Hip fracture", "definition": "New diagnosis of HFx (ICD-9-CM codes 820\u2013820.9) with internal fixation or hemiarthroplasty"}, "reported_effect": {"measure": "HR", "value": "1.41", "ci": "1.29-1.54", "adjusted": true}}, "dag": {"nodes": [{"id": "ppi", "label": "PPI use", "role": "exposure", "adjusted": false, "snippet": "PPI user group showed an increased risk of HFx with an adjusted hazard ratio of 1.41", "ai_suggested": false}, {"id": "hip_fracture", "label": "Hip fracture", "role": "outcome", "adjusted": false, "snippet": "Hip fracture risk was evaluated using ICD-9-CM codes 820\u2013820.9", "ai_suggested": false}, {"id": "age", "label": "Age", "role": "confounder", "adjusted": true, "snippet": "propensity score matching was performed to minimize selection bias by balancing baseline characteristics, including age and gender", "ai_suggested": false}, {"id": "gender", "label": "Gender", "role": "confounder", "adjusted": true, "snippet": "propensity score matching was performed to minimize selection bias by balancing baseline characteristics, including age and gender", "ai_suggested": false}, {"id": "cci", "label": "Charlson comorbidity index", "role": "confounder", "adjusted": true, "snippet": "Charlson comorbidity index (CCI) scores were used to assess the severity of comorbidities", "ai_suggested": false}, {"id": "hypertension", "label": "Hypertension", "role": "confounder", "adjusted": true, "snippet": "All the covariables, including comorbidities and medication listed in Supplementary Table were adjusted for", "ai_suggested": false}, {"id": "stroke", "label": "Stroke", "role": "confounder", "adjusted": true, "snippet": "stroke and chronic renal failure patients with T2DM showed an increased risk of HFx", "ai_suggested": false}, {"id": "chronic_renal_failure", "label": "Chronic renal failure", "role": "confounder", "adjusted": true, "snippet": "stroke and chronic renal failure patients with T2DM showed an increased risk of HFx", "ai_suggested": false}, {"id": "use_of_anticoagulants", "label": "Use of anticoagulants", "role": "confounder", "adjusted": true, "snippet": "use of anticoagulants was associated with an increased risk of HFx after adjustment", "ai_suggested": false}, {"id": "use_of_sedatives", "label": "Use of sedatives", "role": "confounder", "adjusted": true, "snippet": "use of sedatives was associated with an increased risk of HFx after adjustment", "ai_suggested": false}, {"id": "use_of_statins", "label": "Use of statins", "role": "confounder", "adjusted": true, "snippet": "Patients who used statins had a decreased risk of HFx with an adjusted HR of 0.61", "ai_suggested": false}, {"id": "glucocorticoid_use", "label": "Glucocorticoid use", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}, {"id": "smoking_status", "label": "Smoking status", "role": "confounder", "adjusted": false, "snippet": "", "ai_suggested": true}], "edges": [{"source": "ppi", "target": "hip_fracture", "ai_suggested": false}, {"source": "age", "target": "ppi", "ai_suggested": false}, {"source": "age", "target": "hip_fracture", "ai_suggested": false}, {"source": "gender", "target": "ppi", "ai_suggested": false}, {"source": "gender", "target": "hip_fracture", "ai_suggested": false}, {"source": "cci", "target": "ppi", "ai_suggested": false}, {"source": "cci", "target": "hip_fracture", "ai_suggested": false}, {"source": "hypertension", "target": "ppi", "ai_suggested": false}, {"source": "hypertension", "target": "hip_fracture", "ai_suggested": false}, {"source": "stroke", "target": "ppi", "ai_suggested": false}, {"source": "stroke", "target": "hip_fracture", "ai_suggested": false}, {"source": "chronic_renal_failure", "target": "ppi", "ai_suggested": false}, {"source": "chronic_renal_failure", "target": "hip_fracture", "ai_suggested": false}, {"source": "use_of_anticoagulants", "target": "ppi", "ai_suggested": false}, {"source": "use_of_anticoagulants", "target": "hip_fracture", "ai_suggested": false}, {"source": "use_of_sedatives", "target": "ppi", "ai_suggested": false}, {"source": "use_of_sedatives", "target": "hip_fracture", "ai_suggested": false}, {"source": "use_of_statins", "target": "ppi", "ai_suggested": false}, {"source": "use_of_statins", "target": "hip_fracture", "ai_suggested": false}, {"source": "glucocorticoid_use", "target": "ppi", "ai_suggested": true}, {"source": "glucocorticoid_use", "target": "hip_fracture", "ai_suggested": true}, {"source": "smoking_status", "target": "ppi", "ai_suggested": true}, {"source": "smoking_status", "target": "hip_fracture", "ai_suggested": true}]}, "dagitty": "dag {\nppi [exposure]\nhip_fracture [outcome]\nage [adjusted]\ngender [adjusted]\ncci [adjusted]\nhypertension [adjusted]\nstroke [adjusted]\nchronic_renal_failure [adjusted]\nuse_of_anticoagulants [adjusted]\nuse_of_sedatives [adjusted]\nuse_of_statins [adjusted]\nglucocorticoid_use\nsmoking_status\nppi -> hip_fracture\nage -> ppi\nage -> hip_fracture\ngender -> ppi\ngender -> hip_fracture\ncci -> ppi\ncci -> hip_fracture\nhypertension -> ppi\nhypertension -> hip_fracture\nstroke -> ppi\nstroke -> hip_fracture\nchronic_renal_failure -> ppi\nchronic_renal_failure -> hip_fracture\nuse_of_anticoagulants -> ppi\nuse_of_anticoagulants -> hip_fracture\nuse_of_sedatives -> ppi\nuse_of_sedatives -> hip_fracture\nuse_of_statins -> ppi\nuse_of_statins -> hip_fracture\nglucocorticoid_use -> ppi\nglucocorticoid_use -> hip_fracture\nsmoking_status -> ppi\nsmoking_status -> hip_fracture\n}", "open_backdoors": ["Glucocorticoid use", "Smoking status"], "critique": {"missing_confounders": [{"label": "Glucocorticoid use", "rationale": "Glucocorticoids increase fracture risk and are often co\u2011prescribed with PPIs for gastrointestinal protection, creating a common cause of both exposure and outcome.", "edges_to_add": [{"source": "Glucocorticoid use", "target": "ppi", "ai_suggested": false}, {"source": "Glucocorticoid use", "target": "hip_fracture", "ai_suggested": false}], "evidence": [{"title": "Glucocorticoid - Wikipedia", "url": "https://en.wikipedia.org/wiki/Glucocorticoid"}, {"title": "A General Introduction to Glucocorticoid Biology", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC6621919"}, {"title": "Frontiers | A General Introduction to Glucocorticoid Biology", "url": "https://www.frontiersin.org/journals/immunology/articles/10.3389/fimmu.2019.01545/full"}], "status": "accepted", "review_note": "Glucocorticoids increase fracture risk and are commonly co\u2011prescribed with PPIs for gastrointestinal protection, making them a common cause not adjusted for."}, {"label": "Fall history", "rationale": "Prior falls raise hip fracture likelihood and may lead to greater PPI use due to pain\u2011related gastrointestinal symptoms or prophylaxis.", "edges_to_add": [{"source": "Fall history", "target": "ppi", "ai_suggested": false}, {"source": "Fall history", "target": "hip_fracture", "ai_suggested": false}], "evidence": [{"title": "PPI Use Is Linked to a 30% Higher Hip Fracture Risk | The Dr Kumar Discovery", "url": "https://drkumardiscovery.com/posts/proton-pump-inhibitors-risk-fractures-metaanalysis-11-international"}, {"title": "Proton pump inhibitor use and risk of hip fractures in ...", "url": "https://pubmed.ncbi.nlm.nih.gov/18657011"}, {"title": "Proton Pump Inhibitors and Fractures in Adults: A Critical Appraisal and Review of the Literature", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC7822697"}], "status": "rejected", "review_note": "PPI use may increase fall risk (e.g., via hypomagnesemia or B12 deficiency), so fall history could be downstream of exposure rather than a pure confounder."}, {"label": "Body mass index", "rationale": "Low BMI is a risk factor for osteoporosis and fractures, while higher BMI may increase gastro\u2011oesophageal reflux and thus PPI prescribing.", "edges_to_add": [{"source": "Body mass index", "target": "ppi", "ai_suggested": false}, {"source": "Body mass index", "target": "hip_fracture", "ai_suggested": false}], "evidence": [{"title": "PPI Use Is Linked to a 30% Higher Hip Fracture Risk | The Dr Kumar Discovery", "url": "https://drkumardiscovery.com/posts/proton-pump-inhibitors-risk-fractures-metaanalysis-11-international"}, {"title": "Proton Pump Inhibitor Use, Hip Fracture, and Change in ...", "url": "https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/415901"}, {"title": "Proton Pump Inhibitor Use, Hip Fracture and Change in Bone ...", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC4240017"}], "status": "rejected", "review_note": "Long\u2011term PPI use can alter body weight (e.g., through symptom relief changing intake), so BMI may be affected by the exposure."}, {"label": "Smoking status", "rationale": "Smoking adversely affects bone density and is associated with gastrointestinal conditions that prompt PPI use.", "edges_to_add": [{"source": "Smoking status", "target": "ppi", "ai_suggested": false}, {"source": "Smoking status", "target": "hip_fracture", "ai_suggested": false}], "evidence": [{"title": "Proton pump inhibitor use is associated with hip fracture development: a nationwide population-based cohort study", "url": "https://kjim.org/journal/view.php?number=170256"}, {"title": "Smoking, PPI Use, and Hip Fracture: What's the Bottom Line?", "url": "https://www.medscape.com/viewarticle/762478"}, {"title": "Hip Fracture Risk May Be Higher With Long-Term PPI Use | MDedge", "url": "https://www.mdedge.com/ehospitalistnews/article/31534/gastroenterology/hip-fracture-risk-may-be-higher-long-term-ppi-use"}], "status": "accepted", "review_note": "Smoking reduces bone density and increases gastrointestinal disease prompting PPI use, acting as a common cause not captured by the adjusted covariates."}, {"label": "Vitamin D deficiency", "rationale": "Deficiency impairs bone mineralization and may coexist with conditions prompting PPI therapy (e.g., malabsorption).", "edges_to_add": [{"source": "Vitamin D deficiency", "target": "ppi", "ai_suggested": false}, {"source": "Vitamin D deficiency", "target": "hip_fracture", "ai_suggested": false}], "evidence": [{"title": "Proton Pump Inhibitor Therapy and Hip Fracture Risk", "url": "https://www.jordangeller.com/s/Artice-7.pdf"}, {"title": "PPI Use Is Linked to a 30% Higher Hip Fracture Risk | The Dr Kumar Discovery", "url": "https://drkumardiscovery.com/posts/proton-pump-inhibitors-risk-fractures-metaanalysis-11-international"}, {"title": "Proton Pump Inhibitors and Fractures in Adults: A Critical Appraisal and Review of the Literature", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC7822697"}], "status": "rejected", "review_note": "PPI therapy can impair vitamin D absorption/metabolism, making deficiency potentially a mediator rather than an independent cause."}], "bias_risks": [{"type": "immortal time bias", "explanation": "Exposure required \u22653 years of PPI use; subjects had to survive and remain event\u2011free during this period before being classified as exposed, potentially inflating the hazard ratio.", "severity": "high"}, {"type": "confounding by indication", "explanation": "PPIs are prescribed for GERD, peptic ulcer disease, or dyspepsia, conditions that may themselves increase fall risk or be markers of frailty linked to fracture.", "severity": "medium"}, {"type": "measurement error", "explanation": "PPI use was derived from pharmacy claims; non\u2011adherence or over\u2011the\u2011counter use leads to misclassification, likely biasing the estimate toward the null.", "severity": "medium"}, {"type": "selection bias", "explanation": "Excluding patients with prior PPI use or prior hip fracture creates a collider if those exclusions are related to both exposure propensity and fracture susceptibility.", "severity": "low"}, {"type": "residual confounding", "explanation": "Propensity score matching did not capture lifestyle factors such as physical activity, dietary calcium intake, or fall\u2011prevention practices, leaving open paths for confounding.", "severity": "medium"}], "overall_assessment": "The study\u2019s adjusted HR of 1.41 suggests a modest association, but immortal time bias and unmeasured confounding (e.g., glucocorticoids, fall history) could substantially distort this estimate. While the authors controlled for many comorbidities and used propensity matching, the remaining biases mean the causal claim should be interpreted cautiously.", "label": "AI-suggested"}, "models": {"extraction": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B", "critique": "nvidia/nemotron-3-super-120b-a12b", "evidence_search": "tavily", "provider": "Nebius Token Factory"}, "review": {"by": "ai", "model": "nvidia/nemotron-3-super-120b-a12b"}, "dataset_license": "CC-BY-4.0"}