Jacques Benichou

A review of adjusted estimators of attributable risk

This paper reviews adjusted methods of estimation of attributable risk (AR), that is methods that allow one to obtain estimates of AR while controlling for other factors. Estimability and basic principles of AR estimation are first considered and the rationale for adjusted AR estimators is discussed. Then, adjusted AR estimators are reviewed focusing on cross-sectional, cohort and case-control studies. Two inconsistent adjusted estimators are briefly commented upon. Next, adjusted estimators based on stratification, namely the weighted-sum and Mantel-Haenszel (MH) approaches, are reviewed and contrasted. It appears that the weighted-sum approach, which allows for full interaction between exposure and adjustment factors, can be affected by small-sample bias. By contrast, the MH approach, which rests on the assumption of no interaction between exposure and adjustment factors may be misleading if interaction between exposure and adjustment factors is present. Model-based adjusted estimators represent a more general and flexible approach that includes both stratification approaches as special cases and offers intermediate options. Bruzzi et al.’s and Greenland and Drescher’s estimators are reviewed and contrasted. Finally, special problems of adjusted estimation are considered, namely estimation from case-cohort data, estimation for risk factors with multiple levels, for multiple risk factors, for recurrent events, estimation of the prevented and preventable fractions, and estimation of the generalized impact fraction. Comments on future directions are presented.

Magnesium sulphate given before very‐preterm birth to protect infant brain: the randomised controlled PREMAG trial*

Objective  To evaluate whether magnesium sulphate (MgSO4) given to women at risk of very‐preterm birth would be neuroprotective in preterm newborns and would prevent neonatal mortality and severe white‐matter injury (WMI).

Variance calculations and confidence intervals for estimates of the attributable risk based on logistic models.

The attributable risk (AR), defined as AR = [Pr(disease) - Pr(disease/no exposure)]/Pr(disease), measures the proportion of disease risk that is attributable to an exposure. Recently Bruzzi et al. (1985, American Journal of Epidemiology 122, 904-914) presented point estimates of AR based on logistic models for case-control data to allow for confounding factors and secondary exposures. To produce confidence intervals, we derived variance estimates for AR under the logistic model and for various designs for sampling controls. Calculations for discrete exposure and confounding factors require covariances between estimates of the risk parameters of the logistic model and the proportions of cases with given levels of exposure and confounding factors. These covariances are estimated from Taylor series expansions applied to implicit functions. Similar calculations for continuous exposures are derived using influence functions. Simulations indicate that those asymptotic procedures yield reliable variance estimates and confidence intervals with near nominal coverage. An example illustrates the usefulness of variance calculations in selecting a logistic model that is neither so simplified as to exhibit systematic lack of fit nor so complicated as to inflate the variance of the estimate of AR.

Estimates of absolute cause-specific risk in cohort studies.

In this paper we study methods for estimating the absolute risk of an event c1 in a time interval [t1, t2], given that the individual is at risk at t1 and given the presence of competing risks. We discuss some advantages of absolute risk for measuring the prognosis of an individual patient and some difficulties of interpretation for comparing two treatment groups. We also discuss the importance of the concept of absolute risk in evaluating public health measures to prevent disease. Variance calculations permit one to gauge the relative importance of random and systematic errors in estimating absolute risk. Efficiency calculations were also performed to determine how much precision is lost in estimating absolute risk with a nonparametric approach or with a flexible piecewise exponential model rather than a simple exponential model, and other calculations indicate the extent of bias that arises with the simple exponential model when that model is invalid. Such calculations suggest that the more flexible models will be useful in practice. Simulations confirm that asymptotic methods yield reliable variance estimates and confidence interval coverages in samples of practical size.

Incidence and Mortality of Bullous Pemphigoid in France

A major increase in the incidence of BP has been recently reported in the United Kingdom. In addition, there are some controversies about the over-mortality of BP patients. The primary objective was to reevaluate the incidence of BP in France as compared with that we estimated 15 years ago. The secondary objective was to assess mortality of BP patients. BP incidence was retrospectively estimated from all BP cases diagnosed between January 2000 and December 2005 in three French regions with a total population of 3.858 million inhabitants. BP mortality was assessed from a prospective cohort accrued during the same time period. A total of 502 incident BP patients (mean age: 82.6±8.8 years) were identified. Overall estimated incidence was 21.7 cases per million persons per year (95%CI:19.8-23.7 cases per million persons per year), which is about 3-fold higher than the incidence that we estimated 15 years ago. In the population aged 70 years or above, BP incidence was 162 cases per million per year (95%CI:147-177 cases per million per year). The overall 1-year survival rate was 62% (95% CI: 56-67%). The risk of death for BP patients was more than six times greater than that for the general population (SMR:6.60; 95%CI:5.47-7.90). The incidence of BP in France has increased 3-fold in the last 15 years. BP is associated with high mortality.

Dietary intakes of ω-6 and ω-3 polyunsaturated fatty acids and the risk of breast cancer

Experimental studies suggest detrimental effects of ω‐6 polyunsaturated fatty acids (PUFA), and beneficial effects of ω‐3 PUFAs on mammary carcinogenesis, possibly in interaction with antioxidants. However, PUFA food sources are diverse in human diets and few epidemiologic studies have examined whether associations between dietary PUFAs and breast cancer risk vary according to food sources or antioxidant intakes. The relationship between individual PUFA intakes estimated from diet history questionnaires and breast cancer risk was examined among 56,007 French women. During 8 years of follow‐up, 1,650 women developed invasive breast cancer. Breast cancer risk was not related to any dietary PUFA overall; however, opposite associations were seen according to food sources, suggesting other potential effects than PUFA per se. Breast cancer risk was inversely associated with α‐linolenic acid (ALA) intake from fruit and vegetables [highest vs. lowest quintile, hazard ratio (HR) 0.74; 95% confidence interval (CI) 0.63, 0.88; p trend < 0.0001], and from vegetable oils (HR 0.83; 95% CI 0.71, 0.97; p trend 0.017). Conversely, breast cancer risk was positively related to ALA intake from nut mixes (p trend 0.004) and processed foods (p trend 0.068), as was total ALA intake among women in the highest quintile of dietary vitamin E (p trend 0.036). A significant interaction was also found between ω‐6 and long‐chain ω‐3 PUFAs, with breast cancer risk inversely related to long‐chain ω‐3 PUFAs in women belonging to the highest quintile of ω‐6 PUFAs (p interaction 0.042). These results emphasize the need to consider food sources, as well as interactions between fatty acids and with antioxidants, when evaluating associations between PUFA intakes and breast cancer risk.

Benzodiazepines and Injurious Falls in Community Dwelling Elders

BackgroundBenzodiazepines are frequently used medications in the elderly, in whom they are associated with an increased risk of falling, with sometimes dire consequences.ObjectiveTo estimate the impact of benzodiazepine-associated injurious falls in a population of elderly persons.MethodA nested case-control study was conducted using data collected during 10 years of follow-up of the French PAQUID (Personnes Agées QUID) community-based cohort. The main outcome measure was the occurrence of an injurious fall, which was defined as a fall resulting in hospitalization, fracture, head trauma or death. Controls (3:1) were frequency-matched to cases. Benzodiazepine exposure was the use of benzodiazepines over the previous 2 weeks reported at the follow-up visit preceding the fall.ResultsBenzodiazepine use was significantly associated with the occurrence of injurious falls, with a significant interaction with age. The adjusted odds ratio for injurious falls in subjects exposed to benzodiazepines was 2.2 (95% CI 1.4, 3.4) in subjects aged ≥80 years and 1.3 (95% CI 0.9, 1.9) in subjects aged <80 years. The population attributable risk for injurious falls in subjects exposed to benzodiazepines was 28.1% (95% CI 16.7, 43.2) for subjects aged ≥80 years. The incidence of injurious falls in subjects aged ≥80 years exposed to benzodiazepines in the PAQUID cohort was 2.8/100 person-years. Over 9% of these falls were fatal. According to these results and to recent population estimates, benzodiazepine use could be held responsible for almost 20 000 injurious falls in subjects aged ≥80 years every year in France, and for nearly 1800 deaths.ConclusionGiven the considerable morbidity and mortality associated with benzodiazepine use and the fact that existing good practice guidelines on benzodiazepines have not been effective in preventing their misuse (possibly because they have not been applied), new methods for limiting use of benzodiazepines in the elderly need to be found.

Gastric involvement in systemic sclerosis: a prospective study

OBJECTIVES:This study aims to assess the prevalence of gastric electrical activity dysfunction with cutaneous electrogastrography (EGG), disturbances of gastric emptying function using radiopaque pellets, and gastric endoscopic abnormalities in patients with systemic sclerosis (SSc). We also investigate for an association between EGG and gastric-emptying data with clinical manifestations and esophageal motor disturbances.METHODS:Fasting and postprandial gastric electrical activity was studied in 22 consecutive patients with SSc (17 with and 5 without clinical gastric manifestations) and 22 age- and sex-matched healthy subjects. Gastric emptying of radiopaque pellets and gastroscopy were also performed in SSc patients.RESULTS:The prevalence of EGG disturbances was as high as 81.82% in SSc patients. SSc patients exhibited, compared with controls, higher median percentage of dominant frequency in bradygastria during the fasting period and lower median values for postprandial electrical power and postprandial to fasting ratio for electrical power. Gastric emptying of radiopaque pellets was delayed in 11 SSc patients, and gastroscopy demonstrated “watermelon stomach” in 3 SSc patients. No correlation was found between the severity of gastric impairment and clinical presentation, SSc duration and subsets, and esophageal manometric impairment.CONCLUSIONS:Our study underlines the high frequency of gastric dysfunction in SSc patients. It suggests the usefulness of EGG in SSc in noninvasively detecting disorders of gastric electrical activity at an early stage and symptomatic patients with gastroparesis (because there was a correlation between values of postprandial to fasting ratio for electrical power of < 1 on EGG and delayed gastric emptying). Gastric-emptying evaluation of radiopaque pellets may further be a helpful method to depict symptomatic SSc patients with gastroparesis. Finally, our series confirms that watermelon stomach diagnosis should be excluded in SSc patients presenting with gastrointestinal hemorrhage or with anemia related to iron deficiency.

Stressful Life Events as a Risk Factor for Inflammatory Bowel Disease Onset: A Population-Based Case–Control Study

BACKGROUND AND AIMS:Stress is often perceived by patients with inflammatory bowel disease (IBD) as the leading cause of their disease. The aim of this study was to assess whether stress, evaluated through life event (LE) occurrence, is associated with IBD onset.METHODS:Incident cases of IBD, including 167 patients with Crohns disease (CD) and 74 with ulcerative colitis (UC), were compared with two control groups, one of 69 patients with acute self-limited colitis (ASLC) and another of 255 blood donors (BDs). Stress was assessed using Paykels self-questionnaire of LEs. Only LEs occurring within 6 months before the onset of symptoms in IBD cases and ASLC controls and before blood donation in BD controls were registered. Anxiety and depression were assessed using Bates and Becks questionnaires, respectively.RESULTS:In univariate analysis, occurrence of LEs was more frequent in the 6-month period prior to diagnosis in CD cases than in UC cases or either control group. After adjustment for depression and anxiety scores as well as other characteristics such as smoking status and sociodemographic features, this association appeared no longer significant. No associations were noted between occurrence of LEs and onset of UC relative to controls.CONCLUSIONS:Despite its separate association with CD, LE occurrence does not appear to be an independent risk factor for IBD onset.

Estimates of the cancer burden in Europe from radioactive fallout from the Chernobyl accident

The Chernobyl accident, which occurred April 26, 1986, resulted in a large release of radionuclides, which were deposited over a very wide area, particularly in Europe. Although an increased risk of thyroid cancer in exposed children has been clearly demonstrated in the most contaminated regions, the impact of the accident on the risk of other cancers as well as elsewhere in Europe is less clear. The objective of the present study was to evaluate the human cancer burden in Europe as a whole from radioactive fallout from the accident. Average country‐ and region‐specific whole‐body and thyroid doses from Chernobyl were estimated using new dosimetric models and radiological data. Numbers of cancer cases and deaths possibly attributable to radiation from Chernobyl were estimated, applying state‐of‐the‐art risk models derived from studies of other irradiated populations. Simultaneously, trends in cancer incidence and mortality were examined over time and by dose level. The risk projections suggest that by now Chernobyl may have caused about 1,000 cases of thyroid cancer and 4,000 cases of other cancers in Europe, representing about 0.01% of all incident cancers since the accident. Models predict that by 2065 about 16,000 (95% UI 3,400–72,000) cases of thyroid cancer and 25,000 (95% UI 11,000–59,000) cases of other cancers may be expected due to radiation from the accident, whereas several hundred million cancer cases are expected from other causes. Although these estimates are subject to considerable uncertainty, they provide an indication of the order of magnitude of the possible impact of the Chernobyl accident. It is unlikely that the cancer burden from the largest radiological accident to date could be detected by monitoring national cancer statistics. Indeed, results of analyses of time trends in cancer incidence and mortality in Europe do not, at present, indicate any increase in cancer rates—other than of thyroid cancer in the most contaminated regions—that can be clearly attributed to radiation from the Chernobyl accident.