Stephen Evans

Reporting of Noninferiority and Equivalence Randomized Trials: An Extension of the CONSORT Statement

The CONSORT (Consolidated Standards of Reporting Trials) Statement, which includes a checklist and a flow diagram, is a guideline developed to help authors improve the reporting of the findings from randomized controlled trials. It was updated most recently in 2010. Its primary focus is on individually randomized trials with 2 parallel groups that assess the possible superiority of one treatment compared with another. The CONSORT Statement has been extended to other trial designs such as cluster randomization, and recommendations for noninferiority and equivalence trials were made in 2006. In this article, we present an updated extension of the CONSORT checklist for reporting noninferiority and equivalence trials, based on the 2010 version of the CONSORT Statement and the 2008 CONSORT Statement for the reporting of abstracts, and provide illustrative examples and explanations for those items that differ from the main 2010 CONSORT checklist. The intent is to improve reporting of noninferiority and equivalence trials, enabling readers to assess the reliability of their results and conclusions.

Interpretation of the evidence for the efficacy and safety of statin therapy

This Review is intended to help clinicians, patients, and the public make informed decisions about statin therapy for the prevention of heart attacks and strokes. It explains how the evidence that is available from randomised controlled trials yields reliable information about both the efficacy and safety of statin therapy. In addition, it discusses how claims that statins commonly cause adverse effects reflect a failure to recognise the limitations of other sources of evidence about the effects of treatment. Large-scale evidence from randomised trials shows that statin therapy reduces the risk of major vascular events (ie, coronary deaths or myocardial infarctions, strokes, and coronary revascularisation procedures) by about one-quarter for each mmol/L reduction in LDL cholesterol during each year (after the first) that it continues to be taken. The absolute benefits of statin therapy depend on an individuals absolute risk of occlusive vascular events and the absolute reduction in LDL cholesterol that is achieved. For example, lowering LDL cholesterol by 2 mmol/L (77 mg/dL) with an effective low-cost statin regimen (eg, atorvastatin 40 mg daily, costing about £2 per month) for 5 years in 10 000 patients would typically prevent major vascular events from occurring in about 1000 patients (ie, 10% absolute benefit) with pre-existing occlusive vascular disease (secondary prevention) and in 500 patients (ie, 5% absolute benefit) who are at increased risk but have not yet had a vascular event (primary prevention). Statin therapy has been shown to reduce vascular disease risk during each year it continues to be taken, so larger absolute benefits would accrue with more prolonged therapy, and these benefits persist long term. The only serious adverse events that have been shown to be caused by long-term statin therapy-ie, adverse effects of the statin-are myopathy (defined as muscle pain or weakness combined with large increases in blood concentrations of creatine kinase), new-onset diabetes mellitus, and, probably, haemorrhagic stroke. Typically, treatment of 10 000 patients for 5 years with an effective regimen (eg, atorvastatin 40 mg daily) would cause about 5 cases of myopathy (one of which might progress, if the statin therapy is not stopped, to the more severe condition of rhabdomyolysis), 50-100 new cases of diabetes, and 5-10 haemorrhagic strokes. However, any adverse impact of these side-effects on major vascular events has already been taken into account in the estimates of the absolute benefits. Statin therapy may cause symptomatic adverse events (eg, muscle pain or weakness) in up to about 50-100 patients (ie, 0·5-1·0% absolute harm) per 10 000 treated for 5 years. However, placebo-controlled randomised trials have shown definitively that almost all of the symptomatic adverse events that are attributed to statin therapy in routine practice are not actually caused by it (ie, they represent misattribution). The large-scale evidence available from randomised trials also indicates that it is unlikely that large absolute excesses in other serious adverse events still await discovery. Consequently, any further findings that emerge about the effects of statin therapy would not be expected to alter materially the balance of benefits and harms. It is, therefore, of concern that exaggerated claims about side-effect rates with statin therapy may be responsible for its under-use among individuals at increased risk of cardiovascular events. For, whereas the rare cases of myopathy and any muscle-related symptoms that are attributed to statin therapy generally resolve rapidly when treatment is stopped, the heart attacks or strokes that may occur if statin therapy is stopped unnecessarily can be devastating.

Quantitative signal detection using spontaneous ADR reporting

Quantitative methods are increasingly used to analyse spontaneous reports. We describe the core concepts behind the most common methods, the proportional reporting ratio (PRR), reporting odds ratio (ROR), information component (IC) and empirical Bayes geometric mean (EBGM). We discuss the role of Bayesian shrinkage in screening spontaneous reports, the importance of changes over time in screening the properties of the measures. Additionally we discuss three major areas of controversy and ongoing research: stratification, method evaluation and implementation. Finally we give some suggestions as to where emerging research is likely to lead. Copyright

Antidepressant treatment and the risk of fatal and non-fatal self harm in first episode depression: nested case-control study

Abstract Objective To compare the risk of non-fatal self harm and suicide in patients taking selective serotonin reuptake inhibitors (SSRIs) with that of patients taking tricyclic antidepressants, as well as between different SSRIs and different tricyclic antidepressants. Design Nested case-control study. Setting Primary care in the United Kingdom. Participants 146 095 individuals with a first prescription of an antidepressant for depression. Main outcome measures Suicide and non-fatal self harm. Results 1968 cases of non-fatal self harm and 69 suicides occurred. The overall adjusted odds ratio of non-fatal self harm was 0.99 (95% confidence interval 0.86 to 1.14) and that of suicide 0.57 (0.26 to 1.25) in people prescribed SSRIs compared with those prescribed tricyclic antidepressants. We found little evidence that associations differed over time since starting or stopping treatment. We found some evidence that risks of non-fatal self harm in people prescribed SSRIs compared with those prescribed tricyclic antidepressants differed by age group (interaction P = 0.02). The adjusted odds ratio of non-fatal self harm for people prescribed SSRIs compared with users of tricylic antidepressants for those aged 18 or younger was 1.59 (1.01 to 2.50), but no association was apparent in other age groups. No suicides occurred in those aged 18 or younger currently or recently prescribed tricyclic antidepressants or SSRIs. Conclusion We found no evidence that the risk of suicide or non-fatal self harm in adults prescribed SSRIs was greater than in those prescribed tricyclic antidepressants. We found some weak evidence of an increased risk of non-fatal self harm for current SSRI use among those aged 18 or younger. However, preferential prescribing of SSRIs to patients at higher risk of suicidal behaviour cannot be ruled out.

Intergenerational studies of human birthweight from the 1958 birth cohort. 1. Evidence for a multigenerational effect

Objective To investigate possible multigenerational influences on birthweight.

Effect of statins on a wide range of health outcomes: a cohort study validated by comparison with randomized trials

AIMS To assess the effect of statins on a range of health outcomes. METHODS We undertook a population-based cohort study to assess the effect of statins on a range of health outcomes using a propensity score-based method to control for differences between people prescribed and not prescribed statins. We validated our design by comparing our results for vascular outcomes with the effects established in large randomized trials. The study was based on the United Kingdom Health Improvement Network database that includes the computerized medical records of over four and a half million patients. RESULTS People who initiated treatment with a statin (n = 129,288) were compared with a matched sample of 600,241 people who did not initiate treatment, with a median follow-up period of 4.4 years. Statin use was not associated with an effect on a wide range of outcomes, including infections, fractures, venous thromboembolism, gastrointestinal haemorrhage, or on specific eye, neurological or autoimmune diseases. A protective effect against dementia was observed (hazard ratio 0.80, 99% confidence interval 0.68, 0.95). There was no effect on the risk of cancer even after > or =8 years of follow-up. The effect sizes for statins on vascular end-points and mortality were comparable to those observed in large randomized trials, suggesting bias and confounding had been well controlled for. CONCLUSIONS We found little evidence to support wide-ranging effects of statins on health outcomes beyond their established beneficial effect on vascular disease.

Novel statistical tools for monitoring the safety of marketed drugs.

Robust tools for monitoring the safety of marketed therapeutic products are of paramount importance to public health. In recent years, innovative statistical approaches have been developed to screen large post‐marketing safety databases for adverse events (AEs) that occur with disproportionate frequency. These methods, known variously as quantitative signal detection, disproportionality analysis, or safety data mining, facilitate the identification of new safety issues or possible harmful effects of a product. In this article, we describe the statistical concepts behind these methods, as well as their practical application to monitoring the safety of pharmaceutical products using spontaneous AE reports. We also provide examples of how these tools can be used to identify novel drug interactions and demographic risk factors for adverse drug reactions. Challenges, controversies, and frontiers for future research are discussed.

BMI and risk of dementia in two million people over two decades: a retrospective cohort study

BACKGROUND Dementia and obesity are increasingly important public health issues. Obesity in middle age has been proposed to lead to dementia in old age. We investigated the association between BMI and risk of dementia. METHODS For this retrospective cohort study, we used a cohort of 1,958,191 individuals derived from the United Kingdom Clinical Practice Research Datalink (CPRD) which included people aged 40 years or older in whom BMI was recorded between 1992 and 2007. Follow-up was until the practices final data collection date, patient death or transfer out of practice, or first record of dementia (whichever occurred first). People with a previous record of dementia were excluded. We used Poisson regression to calculate incidence rates of dementia for each BMI category. FINDINGS Our cohort of 1,958,191 people from UK general practices had a median age at baseline of 55 years (IQR 45-66) and a median follow-up of 9·1 years (IQR 6·3-12·6). Dementia occurred in 45,507 people, at a rate of 2·4 cases per 1000 person-years. Compared with people of a healthy weight, underweight people (BMI <20 kg/m(2)) had a 34% higher (95% CI 29-38) risk of dementia. Furthermore, the incidence of dementia continued to fall for every increasing BMI category, with very obese people (BMI >40 kg/m(2)) having a 29% lower (95% CI 22-36) dementia risk than people of a healthy weight. These patterns persisted throughout two decades of follow-up, after adjustment for potential confounders and allowance for the J-shape association of BMI with mortality. INTERPRETATION Being underweight in middle age and old age carries an increased risk of dementia over two decades. Our results contradict the hypothesis that obesity in middle age could increase the risk of dementia in old age. The reasons for and public health consequences of these findings need further investigation. FUNDING None.

Pregnancy as a Major Determinant for Discontinuation of Antidepressants: An Analysis of Data From The Health Improvement Network

BACKGROUND Potential adverse effects of antidepressants during pregnancy have caused concern about their use. There are, however, very limited detailed data on patterns of antidepressant prescribing in pregnancy. OBJECTIVE To examine secular trends in prescribing during pregnancy, to assess whether pregnancy is a major determinant for stopping antidepressants, and to identify characteristics of those who stopped antidepressants during pregnancy. METHOD In this cohort study, we obtained data on 114,999 pregnant women (median age at delivery, 30.5 years [interquartile range, 26-34 years]) who had a live birth between 1992 and 2006 and 22,677 nonpregnant women from The Health Improvement Network primary care database, one of the largest sources of continuous anonymized primary care data in the United Kingdom and broadly representative of UK general practice. This database includes information on age, sex, medical diagnosis and symptoms, health promotion activities, referrals to secondary care, and prescriptions for each registered individual. The database also holds information about social deprivation as measured using quintiles of the Townsend score. We used Cox regression analysis to compare time to last prescription in pregnant versus nonpregnant women and to identify characteristics of those women who stopped antidepressants during pregnancy. RESULTS Antidepressant prescribing in pregnancy increased nearly 4-fold from 1992 to 2006 (relative risk = 3.87; 95% CI, 1.73-8.66; P < .001). Since 2001, approximately 3% of the cohort received antidepressants at some stage during pregnancy. Selective serotonin reuptake inhibitors accounted for approximately 80% of the prescribed antidepressants. Antidepressants were more likely to be stopped in pregnant than in nonpregnant women, in particular during the first 6 weeks of pregnancy (hazard ratio = 5.19; 95% CI, 4.85-5.56; P < .001). Only 10% of women treated before pregnancy still received antidepressants at the start of the third trimester. In contrast, 35% of nonpregnant women were still treated after a similar time period. CONCLUSIONS Although antidepressant prescribing in pregnancy increased nearly 4-fold from 1992 to 2006, pregnancy was a major determinant of cessation of antidepressant medication, and most women did not receive further antidepressant prescriptions beyond 6 weeks of gestation. This finding may be explained by concerns about potential adverse effects of the medications, even though these concerns need to be balanced against the potential harm of inadequate treatment of depression during pregnancy.

Are these data real? Statistical methods for the detection of data fabrication in clinical trials

Abstract Objectives To test the application of statistical methods to detect data fabrication in a clinical trial. Setting Data from two clinical trials: a trial of a dietary intervention for cardiovascular disease and a trial of a drug intervention for the same problem. Outcome measures Baseline comparisons of means and variances of cardiovascular risk factors; digit preference overall and its pattern by group. Results In the dietary intervention trial, variances for 16 of the 22 variables available at baseline were significally different, and 10 significant differences were seen in means for these variables. Some of these P values were extraordinarily small. Distributions of the final recorded digit were significantly different between the intervention and the control group at baseline for 14/22 variables in the dietary trial. In the drug trial, only five variables were available, and no significant differences between the groups for baseline values in means or variances or digit preference were seen. Conclusions Several statistical features of the data from the dietary trial are so strongly suggestive of data fabrication that no other explanation is likely.