Search

BackgroundVariations in inpatient medical care are typically attributed to system, hospital or patient factors. Little is known about variations at the physician level within hospitals. We described the physician-level variation in clinical outcomes and resource use in general internal medicine (GIM).MethodsThis was an observational study of all emergency admissions to GIM at seven hospitals in Ontario, Canada, over a 5-year period between 2010 and 2015. Physician-level variations in inpatient mortality, hospital length of stay, 30-day readmission and use of ‘advanced imaging’ (CT, MRI or ultrasound scans) were measured. Physicians were categorised into quartiles within each hospital for each outcome and then quartiles were pooled across all hospitals (eg, physicians in the highest quartile at each hospital were grouped together). We report absolute differences between physicians in the highest and lowest quartiles after matching admissions based on propensity scores to account for patient-level variation.ResultsThe sample included 103 085 admissions to 135 attending physicians. After propensity score matching, the difference between physicians in the highest and lowest quartiles for in-hospital mortality was 2.4% (95% CI 0.6% to 4.3%, pConclusionsPatient outcomes and resource use in inpatient medical care varied substantially across physicians in this study. Physician-level variations in length of stay and imaging use were unlikely to be explained by patient factors whereas differences in mortality and readmission should be interpreted with caution and could be explained by unmeasured confounders. Physician-level variations may represent practice differences that highlight quality improvement opportunities.

IntroductionXylitol (or ‘birch sugar’) is a naturally occurring sugar with antibacterial properties that has been used as a natural non-sugar sweetener in chewing gums, confectionery, toothpaste and medicines. In this preventative randomised trial, xylitol will be tested for the prevention of acute otitis media (AOM), a common and costly condition in young children. The primary outcome will be the incidence of AOM. Secondary outcomes will include upper respiratory tract infections (URTIs) and dental caries.Methods and analysisThis study will be a pragmatic, blinded (participant and parents, practitioners and analyst), two-armed superiority, placebo-controlled randomised trial with 1:1 allocation, stratified by clinical site. The trial will be conducted in the 11 primary care group practices participating in the TARGet Kids! research network in Canada. Eligible participants between the ages of 2–4 years will be randomly assigned to the intervention arm of regular xylitol syrup use or the control arm of regular sorbitol use for 6 months. We expect to recruit 236 participants, per treatment arm, to detect a 20% relative risk reduction in AOM episodes. AOM will be identified through chart review. The secondary outcomes of URTIs and dental caries will be identified through monthly phone calls with specified questions.Ethics and disseminationEthics approval from the Research Ethics Boards at the Hospital for Sick Children and St. Michael’s Hospital has been obtained for this study and also for the TARGet Kids! research network. Results will be submitted for publication to a peer-reviewed journal and will be discussed with decision makers.Trial registration numberNCT03055091; Pre-results.

Background Immigrants may be at a higher risk of adverse drug reactions, in that poor language proficiency reduces individuals understanding of drug label instructions. Additionally, there are reports of severe or fatal toxicity due to CYP2D6 ultrarapid hepatic metabolism of codeine to morphine among some ethnic groups, especially those from Eastern Africa. Methods Between 2002 and 2012 we conducted a population-based cohort study among residents of Ontario, Canada. We used administrative health databases that linked immigrants and Canadian-born individuals to both prescription medication use and emergency department visits and hospital admissions. The primary composite outcome was the risk of drug overdose or all-cause mortality within 30 days of codeine prescription, comparing patients from various world regions to Canadian-born individuals. A secondary analysis stratified by codeine dose and ability to speak English and/or French. Results There were 553 504 individuals exclusively prescribed codeine. Relative to an incidence rate of 57.1/100 000 person-days among Canadian-born codeine recipients, those who migrated from various world regions were at a lower risk of drug overdose or death. For example, Eastern Africans had an adjusted HR of 0.60 (95% CI 0.31 to 1.17) on controlling for potential confounders such as age, sex, income and physician visits. Patients unable to speak English or French who were prescribed codeine were at a lower risk of the composite outcome relative to those proficient in either language (adjusted HR 0.63, 95% CI 0.54 to 0.74). Interpretation Overdose and death following the institution of codeine therapy are not more commonly observed among immigrants from world regions with a high prevalence of ultrarapid CYP2D6 status relative to those born in Canada. Lower proficiency in English or French also did not appear to heighten the risk.

BackgroundThe risk of death is elevated in patients taking opioids for chronic non-cancer pain. Respiratory depression is the main cause of death due to opioids and sleep apnoea is an important associated risk factor.MethodsIn chronic pain clinics, we assessed the STOP-Bang questionnaire (a screening tool for sleep apnoea; Snoring, Tiredness, Observed apnoea, high blood Pressure, Body mass index, age, neck circumference and male gender), Epworth Sleepiness Scale, thyromental distance, Mallampati classification, daytime oxyhaemoglobin saturation (SpO2) and calculated daily morphine milligram equivalent (MME) approximations for each participant, and performed an inlaboratory polysomnogram. The primary objective was to determine the predictive factors for sleep apnoea in patients on chronic opioid therapy using multivariable logistic regression models.ResultsOf 332 consented participants, 204 underwent polysomnography, and 120 (58.8%) had sleep apnoea (AHI ≥5) (72% obstructive, 20% central and 8% indeterminate sleep apnoea), with a high prevalence of moderate (23.3%) and severe (30.8%) sleep apnoea. The STOP-Bang questionnaire and SpO2 are predictive factors for sleep apnoea (AHI ≥15) in patients on opioids for chronic pain. For each one-unit increase in the STOP-Bang score, the odds of moderate-to-severe sleep apnoea (AHI ≥15) increased by 70%, and for each 1% SpO2 decrease the odds increased by 33%. For each 10 mg MME increase, the odds of Central Apnoea Index ≥5 increased by 3%, and for each 1% SpO2 decrease the odds increased by 45%.ConclusionIn patients on opioids for chronic pain, the STOP-Bang questionnaire and daytime SpO2 are predictive factors for sleep apnoea, and MME and daytime SpO2 are predictive factors for Central Apnoea Index ≥5.Trial registration numberNCT02513836

IntroductionSchool readiness is a multidimensional construct that includes cognitive, behavioural and emotional aspects of a child’s development. School readiness is strongly associated with a child’s future school success and well-being. The Early Development Instrument (EDI) is a reliable and valid teacher-completed tool for assessing school readiness in children at kindergarten age. A substantial knowledge gap exists in understanding how early child growth, health behaviours, nutrition, cardiometabolic risk and development impact school readiness. The primary objective was to determine if growth patterns, measured by body mass index trajectories in healthy children aged 0–5 years, are associated with school readiness at ages 4–6 years (kindergarten age). Secondary objectives were to determine if other health trajectories, including health behaviours, nutrition, cardiometabolic risk and development, are associated with school readiness at ages 4–6 years. This paper presents the Fit for School Study protocol.Methods and analysisThis is an ongoing prospective cohort study. Parents of children enrolled in the The Applied Health Research Group for Kids (TARGet Kids!) practice-based research network are invited to participate in the Fit for School Study. Child growth, health behaviours, nutrition, cardiometabolic risk and development data are collected annually at health supervision visits and linked to EDI data collected by schools. The primary and secondary analyses will use a two-stage process: (1) latent class growth models will be used to first determine trajectory groups, and (2) generalised linear mixed models will be used to examine the relationship between exposures and EDI results.Ethics and disseminationThe research ethics boards at The Hospital for Sick Children, Unity Health Toronto and McMaster University approved this study, and research ethics approval was obtained from each school board with a student participating in the study. The findings will be presented locally, nationally and internationally and will be published in peer-reviewed journals.Trial registration numberNCT01869530.

ObjectivesCare gaps in asthma may be highly prevalent but are poorly characterised. We sought to prospectively measure adherence to key evidence-based adult asthma practices in primary care, and predictors of these behaviours.DesignOne-year prospective cohort study employing an electronic chart audit.SettingThree family health teams (two academic, one community-based) in Ontario, Canada.Participants884 patients (72.1% female; 46.0±17.5 years old) (4199 total visits; 4.8±4.8 visits/patient) assigned to 23 physicians (65% female; practising for 10.0±8.6 years).Main outcome measuresThe primary outcome was the proportion of visits during which practitioners assessed asthma control according to symptom-based criteria. Secondary outcomes included the proportion of: patients who had asthma control assessed at least once; visits during which a controller medication was initiated or escalated; and patients who received a written asthma action plan. Behavioural predictors were established a priori and tested in a multivariable model.ResultsPrimary outcome: Providers assessed asthma control in 4.9% of visits and 15.4% of patients. Factors influencing assessment included clinic site (p=0.019) and presenting symptom, with providers assessing control more often during visits for asthma symptoms (35.0%) or any respiratory symptoms (18.8%) relative to other visits (1.6%) (pConclusionsMajor gaps in evidence-based asthma practice exist in primary care. Targeted knowledge translation interventions are required to address these gaps, and can be tailored by leveraging the identified behavioural predictors.Trial registration numberNCT01070095; Pre-results.

To describe the contributions of prescribed and non-prescribed opioids to opioid related deaths.Population based cohort study.Ontario, Canada, from 1 January 2013 to 31 December 2016.All Ontarians who died of an opioid related cause.Active opioid prescriptions, defined as those with a duration overlapping the date of death, and recent opioid prescriptions, defined as those dispensed in the 30 and 180 days preceding death. Postmortem toxicology results from the Drug and Drug/Alcohol Related Death database were used to characterise deaths on the basis of presence of prescribed and non-prescribed (that is, diverted or illicit) opioids, overall and stratified by year and age.2833 opioid related deaths occurred. An active opioid prescription on the date of death was relatively common but declined slightly throughout the study period (38.2% (241/631) in 2013 and 32.5% (278/855) in 2016; P for trend=0.03). Older people and women were relatively more likely to have an active opioid prescription at time of death. In 2016, 46% (169/364) of people aged 45-64 had an active opioid prescription compared with only 12% (8/69) among those aged 24 or younger (P for trend0.001). Similarly, 46% (124/272) of women had an active opioid prescription at time of death compared with 26.4% (154/583) of men (P0.001). Among people with active opioid prescriptions at time of death, 37.8% (375/993) also had evidence of a non-prescribed opioid on postmortem toxicology. By 2016, the non-prescribed opioid most commonly identified after death was fentanyl (41%; 47 of 115 cases). Among people without an active opioid prescription at time of death, fentanyl was detected in 20% (78/390) of deaths in 2013, increasing to 47.5% (274/577) by 2016 (P0.001).Prescribed, diverted, and illicit opioids all play an important role in opioid related deaths. Although more than half of all opioid related deaths still involved prescription drugs (either dispensed or diverted) in 2016, the increased rate of deaths involving fentanyl between 2015 and 2016 is concerning and suggests the need for a multifactorial approach to this problem that considers both the prescribed and illicit opioid environments.

Introduction Neurological injury remains the major cause of morbidity and mortality following open aortic arch repair. Systemic hypothermia along with antegrade cerebral perfusion (ACP) is the accepted cerebral protection approach, with axillary artery cannulation being the most common technique used to establish ACP. More recently, innominate artery cannulation has been shown to be a safe and efficacious method for establishing ACP. Inasmuch as there is a lack of high-quality data comparing axillary and innominate artery ACP, we have designed a randomised, multi-centre clinical trial to compare both cerebral perfusion strategies with regards to brain morphological injury using diffusion-weighted MRI (DW-MRI). Methods and analysis 110 patients undergoing elective aortic surgery with repair of the proximal arch requiring an open distal anastamosis will be randomised to either the innominate artery or the axillary artery cannulation strategy for establishing unilateral ACP during systemic circulatory arrest with moderate levels of hypothermia. The primary safety endpoint of this trial is the proportion of patients with new radiologically significant ischaemic lesions found on postoperative DW-MRI compared with preoperative DW-MRI. The primary efficacy endpoint of this trial is the difference in total operative time between the innominate artery and the axillary artery cannulation group. Ethics and dissemination The study protocol and consent forms have been approved by the participating local research ethics boards. Publication of the study results is anticipated in 2018 or 2019. If this study shows that the innominate artery cannulation technique is non-inferior to the axillary artery cannulation technique with regards to brain morphological injury, it will establish the innominate artery cannulation technique as a safe and potentially more efficient method of antegrade cerebral perfusion in aortic surgery. Trial registration number NCT02554032.

NP, RHG and ADP were supported as Clinician Scientists by the Department of Family and Community Medicine at the University of Toronto and at St Michael's Hospital. NP is supported by a fellowship from Physicians Services Incorporated Foundation. KT was supported as a Clinician Scientist by the Department of Family and Community Medicine at the University of Toronto and at Toronto Western Hospital. ML received salary support through a Canada Research Chair and a Michael Smith Foundation for Health Research Scholar Award. Introduction : Cost-related non-adherence to medicines is common in low-income, middle-income and high-income countries such as Canada. Medicine non-adherence is associated with poor health outcomes and increased mortality. This randomised trial will test the impact of a carefully selected list of essential medicines at no charge (compared with usual medicine access) in primary care patients reporting cost-related non-adherence. Methods and analysis : This is an open-label, parallel two-arm, superiority, individually randomised controlled trial conducted in three primary care sites (one urban, two rural) in Ontario, Canada, that was codesigned by a community guidance panel. Adult patients (≥18 years) who report cost-related non-adherence to medicines are eligible to participate in the study. Participants will be randomised to receive free and convenient access to a carefully selected list of 125 essential medicines (based on the WHO's Model List of Essential Medicines) or usual means of medicine access. Care for patients in both groups will otherwise be unchanged. The primary outcome of this trial is adherence to appropriately prescribed medicines. Secondary outcomes include medicine adherence, appropriate prescribing, blood pressure, haemoglobin A1c, low-density lipoprotein cholesterol, patient-oriented outcomes and healthcare costs. All participants will be followed for at least 12 months. Ethics and dissemination : Ethics approval was obtained in all three participating sites. Results of the main trial and secondary outcomes will be submitted for publication in a peer-reviewed journal and discussed with members of the public and decision makers. Trial registration number : NCT02744963. Publisher PDF

Analytical strategies can help deal with potential confounding but readers need to know which strategy is appropriate The previous articles in this series1 2 argued that cohort studies are exposed to selection bias and confounding, and that critical appraisal requires a careful assessment of the study design and the identification of potential confounders. This article describes two analytical strategies—regression and stratification—that can be used to assess and reduce confounding. Some cohort studies match individual participants in the intervention and comparison groups on the basis of confounders, but because matching may be viewed as a special case of stratification we have not discussed it specifically and details are available elsewhere.3 4 Neither of these techniques can eliminate bias related to unmeasured or unknown confounders. Furthermore, both have their own assumptions, advantages, and limitations. Regression uses the data to estimate how confounders are related to the outcome and produces an adjusted estimate of the intervention effect. It is the most commonly used method for reducing confounding in cohort studies. The outcome of interest is the dependent variable, and the measures of baseline characteristics (such as age and sex) and the intervention are independent variables. The choice of method of regression analysis (linear, logistic, proportional hazards, etc) is dictated by the type of dependent variable. For example, if the outcome is binary (such as occurrence of hip fracture), a logistic regression model would be appropriate; in contrast, if the outcome is time to an event (such as time to hip fracture) a proportional hazards model is appropriate. Regression analyses estimate the association of each independent variable with the dependent variable after adjusting for the effects of all the other variables. Because the estimated association between the intervention and outcome variables adjusts …

Although confounding is an important problem of cohort studies, its effects can be minimised to enable valid comparison In cohort studies, who does or does not receive an intervention is determined by practice patterns, personal choice, or policy decisions. This raises the possibility that the intervention and comparison groups may differ in characteristics that affect the study outcome, a problem called selection bias. If these characteristics have independent effects on the observed outcome in each group, they will create differences in outcomes between the groups apart from those related to the interventions being assessed. This effect is known as confounding.1 In the first paper in the series we dealt with the design and use of cohort studies and how to identify selection bias.2 This paper focuses on the definition and assessment of confounders. For a characteristic to be a confounder in a particular study, it must meet two criteria.1 The first is that it must be related to the outcome in terms of prognosis or susceptibility. For example, in the study of the association between antipsychotic use and hip fracture that we considered in the first paper,2 age is known to be related to risk of hip fracture and therefore has the potential to be a confounder. The second criterion that defines a confounder is that the distribution of the characteristic is different in the groups being compared. It can differ in terms of either the mean or the degree of variation or variability in that characteristic. For example, for age to be a confounder in a cohort study, either the average age or the variation in the age in the groups being compared would have to be different. Assessing variation as well as average values is important because groups can have the same average value …

Objective: To compare rates of upper gastrointestinal haemorrhage among elderly patients given selective cyclo-oxygenase-2 (COX 2) inhibitors and non-selective non-steroidal anti-inflammatory drugs (NSAIDs). Design: Observational cohort study. Setting: Administrative data from Ontario, Canada, used from 17 April 2000 to 31 March 2001 to identify population based, NSAID-naive cohorts of patients. Patients: Subjects aged ≥ 66 years who started taking non-selective NSAIDs (n=5391), diclofenac plus misoprostol (n=5087), rofecoxib (n=14 583), or celecoxib (n=18 908) and a randomly selected control cohort not exposed to NSAIDs (n=100 000). Main outcome measures: Rate ratios of hospital admission for upper gastrointestinal haemorrhage in each drug cohort with adjustment for potential confounders. Results: Relative to controls, the multivariate model revealed an increased short term risk of upper gastrointestinal haemorrhage for users of non-selective NSAIDs (adjusted rate ratio 4.0 (95% confidence intervals 2.3 to 6.9)), diclofenac plus misoprostol (3.0 (1.7 to 5.6)), and rofecoxib (1.9 (1.3 to 2.8)) but not celecoxib (1.0 (0.7 to 1.6)). Relative to celecoxib, significantly higher risks of upper gastrointestinal haemorrhage were observed for non-selective NSAIDs (4.4 (2.3 to 8.5)), diclofenac plus misoprostol (3.2 (1.6 to 6.5)), and rofecoxib (1.9 (1.2 to 2.8)). Relative to rofecoxib, non-selective NSAID users were at significantly higher risk of upper gastrointestinal haemorrhage (1.9 (1.0 to 3.5)). Conclusions: This population based observational study found a lower short term risk of upper gastrointestinal haemorrhage for selective COX-2 inhibitors compared with non-selective NSAIDs.

Objective Clarithromycin strongly inhibits enzyme cytochrome P450 3A4, preventing the metabolism of some other drugs, while azithromycin is a weak inhibitor. Accordingly, blood concentrations of other drugs increase with clarithromycin coprescription leading to adverse events. These macrolide antibiotics also differ on other properties that may impact outcomes. In this study, we compared outcomes in two groups of macrolide antibiotic users in the absence of potentially interacting drugs. Design Population-based retrospective cohort study. Setting Ontario, Canada, from 2003 to 2010. Patients Patients (mean 74 years) prescribed clarithromycin (n=52 251) or azithromycin (referent group, n=46 618). Main outcomes The primary outcomes were hospital admission within 30 days of a new antibiotic prescription with any of the 12 conditions examined separately (acute kidney injury, acute myocardial infarction, neuroimaging (proxy for delirium), hypotension, syncope, hyperkalaemia, hyponatraemia, hyperglycaemia, arrhythmia, ischaemic stroke, gastrointestinal bleeding and sepsis). The secondary outcome was mortality. Results The baseline characteristics of the two groups, including patient demographics, comorbid conditions, infection type and prescribing physician specialty, were nearly identical. The median daily dose was 1000 mg for clarithromycin and 300 mg for azithromycin and the median duration of dispensing antibiotics was 10 and 5 days, respectively. There was no difference between the groups in the risk of hospitalisation for any condition studied (relative risk ranged from 0.67 to 1.23). Compared with azithromycin, clarithromycin was associated with a slightly higher risk of all-cause mortality (0.46% vs 0.37%, relative risk 1.25, 95% CI 1.03 to 1.52). Conclusions Clarithromycin can be used to assess drug interactions in population-based studies with azithromycin serving as a control group. However, any differences in mortality observed between the two antibiotic groups in the setting of other drug use may be partially attributable to factors beyond the inhibition of drug metabolising enzymes and transporters, as the difference for this outcome was significant.

Inherited thrombophilias affect over 16% of people and predispose them to venous thromboembolism.1 Pre-eclampsia is associated with occlusion of the placental spiral arteries.2 Thrombophilias may cause thrombosis of placental vessels,3 thereby explaining the link between thrombophilia and pre-eclampsia. 4 5 We tested the hypothesis that women with pre-eclampsia have a higher risk of subsequent venous thromboembolism. The study was approved by our research ethics board and took place in Ontario, Canada, where all hospital services are publicly funded. We used an administrative database that is based on anonymised populations and which records all hospitalisations from 1 April 1988 onwards. All discharges between 1 April 1990 and 1 January 1994 with a primary ICD-9 (international classification of diseases, 9th revision) code of pre-eclampsia, severe pre-eclampsia, eclampsia, or toxaemia were identified (see appendix A on bmj.com). These codes had a sensitivity of 89% (95% confidence interval 78% to 94%) and a …

Objectives To characterise the risk of admission to hospital for hyperkalaemia in elderly patients treated with trimethoprim-sulfamethoxazole in combination with spironolactone. Design Population based nested case-control study. Setting Ontario, Canada, from 1 April 1992 to 1 March 2010. Participants Cases were residents of Ontario aged 66 years or above receiving chronic treatment with spironolactone and admitted to hospital with hyperkalaemia within 14 days of receiving a prescription for either trimethoprim-sulfamethoxazole, amoxicillin, norfloxacin, or nitrofurantoin. Up to four controls for each case were identified from the same cohort, matched on age, sex, and presence or absence of chronic kidney disease and diabetes, and required to have received one of the study antibiotics within 14 days before the case’s index date. Main outcome measures Odds ratio for association between admission to hospital with hyperkalaemia and receipt of a study antibiotic in the preceding 14 days, adjusted for conditions and drugs that may influence risk of hyperkalaemia. Results During the 18 year study period, 6903 admissions for hyperkalaemia were identified, 306 of which occurred within 14 days of antibiotic use. Of these, 248 (81%) cases were matched to 783 controls. 10.8% (17 859/165 754) of spironolactone users received at least one prescription for trimethoprim-sulfamethoxazole. Compared with amoxicillin, prescription of trimethoprim-sulfamethoxazole was associated with a marked increase in the risk of admission to hospital for hyperkalaemia (adjusted odds ratio 12.4, 95% confidence interval 7.1 to 21.6). The population attributable fraction was 59.7%, suggesting that approximately 60% of all cases of hyperkalaemia in older patients taking spironolactone and treated with an antibiotic for a urinary tract infection could be avoided if trimethoprim-sulfamethoxazole was not prescribed. Treatment with nitrofurantoin was also associated with an increase in the risk of hyperkalaemia (adjusted odds ratio 2.4, 1.3 to 4.6), but no such risk was found with norfloxacin (adjusted odds ratio 1.6, 0.8 to 3.4) Conclusions Among older patients receiving spironolactone, treatment with trimethoprim-sulfamethoxazole was associated with a major increase in the risk of admission to hospital for hyperkalaemia. This drug combination should be avoided when possible.

Objective To compare the risk of acute myocardial infarction, heart failure, and death in patients with type 2 diabetes treated with rosiglitazone and pioglitazone. Design Retrospective cohort study. Setting Ontario, Canada. Participants Outpatients aged 66 years and older who were started on rosiglitazone or pioglitazone between 1 April 2002 and 31 March 2008. Main outcome measure Composite of death or hospital admission for either acute myocardial infarction or heart failure. In a secondary analysis, each outcome was also examined individually. Results 39 736 patients who started on either pioglitazone or rosiglitazone were identified. During the six year study period, the composite outcome was reached in 895 (5.3%) of patients taking pioglitazone and 1563 (6.9%) of patients taking rosiglitazone. After extensive adjustment for demographic and clinical factors and drug doses, pioglitazone treated patients had a lower risk of developing the primary outcome than did patients treated with rosiglitazone (adjusted hazard ratio 0.83, 95% confidence interval 0.76 to 0.90). Secondary analyses revealed a lower risk of death (adjusted hazard ratio 0.86, 0.75 to 0.98) and heart failure (0.77, 0.69 to 0.87) with pioglitazone but no significant difference in the risk of acute myocardial infarction (0.95, 0.81 to 1.11). One additional composite outcome would be predicted to occur annually for every 93 patients treated with rosiglitazone rather than pioglitazone. Conclusions Among older patients with diabetes, pioglitazone is associated with a significantly lower risk of heart failure and death than is rosiglitazone. Given that rosiglitazone lacks a distinct clinical advantage over pioglitazone, continued use of rosiglitazone may not be justified.

Objectives: To determine the association between inhibition of serotonin reuptake by antidepressants and upper gastrointestinal bleeding. Design: Retrospective cohort study from population based databases. Setting: Ontario, Canada. Participants: 317 824 elderly people observed for more than 130 000 person years. The patients started taking an antidepressant between 1992 and 1998 and were grouped by how much the drug inhibited serotonin reuptake. Patients were observed until they stopped the drug, had an upper gastrointestinal bleed, or died or the study ended. Main outcome measure: Admission to hospital for acute upper gastrointestinal bleeding. Results: Overall, 974 bleeds were observed, with an overall bleeding rate of 7.3 per 1000 person years. After controlling for age or previous gastrointestinal bleeding, the risk of bleeding significantly increased by 10.7% and 9.8%, respectively, with increasing inhibition of serotonin reuptake. Absolute differences in bleeding between antidepressant groups were greatest for octogenarians (low inhibition of serotonin reuptake, 10.6 bleeds/1000 person years v high inhibition of serotonin reuptake, 14.7 bleeds/1000 person years; number needed to harm 244) and those with previous upper gastrointestinal bleeding (low, 28.6 bleeds/1000 person years v high, 40.3 bleeds/1000 person years; number needed to harm 85). Conclusions: After age or previous upper gastrointestinal bleeding were controlled for, antidepressants with high inhibition of serotonin reuptake increased the risk of upper gastrointestinal bleeding. These increases are clinically important for elderly patients and those with previous gastrointestinal bleeding. What is already known on this topic A case-control study found that the risk of upper gastrointestinal bleeding increases with intake of antidepressants that extensively inhibit serotonin reuptake The study9s validity was questioned because antidepressants were not specifically classified by the extent that they inhibit serotonin reuptake, and absolute differences in bleeding rates between antidepressants were unavailable What this study adds The risk of upper gastrointestinal bleeding in elderly and depressed patients increases with antidepressants having the greatest extent of inhibition of serotonin reuptake This increased risk of bleeding is clinically important for patients with a high risk of bleeding—namely, octogenarians and those with previous upper gastrointestinal bleeding The extent that an antidepressant inhibits serotonin reuptake should be considered when drugs are required for depression in high risk patients