Reliable assessment of the incidence of childhood autoimmune hemolytic anemia.

BACKGROUND: Childhood autoimmune hemolytic anemia (AIHA) is a rare and severe disease characterized by hemolysis and positive direct antiglobulin test (DAT). Few epidemiologic indicators are available for the pediatric population. The objective of our study was to reliably estimate the number of AIHA cases in the French Aquitaine region and the incidence of AIHA in patients under 18 years old. PROCEDURE: In this retrospective study, the capture-recapture method and log-linear model were used for the period 2000-2008 in the Aquitaine region from the following three data sources for the diagnosis of AIHA: the OBS'CEREVANCE database cohort, positive DAT collected from the regional blood bank database, and the French medico-economic information system. RESULTS: A list of 281 different patients was obtained after cross-matching the three databases; 44 AIHA cases were identified in the period 2000-2008; and the total number of cases was estimated to be 48 (95% confidence interval [CI]: 45-55). The calculated incidence of the disease was 0.81/100,000 children under 18 years old per year (95% CI 0.76-0.92). CONCLUSION: Accurate methods are required for estimating the incidence of AIHA in children. Capture-recapture analysis corrects underreporting and provides optimal completeness. This study highlights a possible under diagnosis of this potentially severe disease in various pediatric settings. AIHA incidence may now be compared with the incidences of other hematological diseases and used for clinical or research purposes.

Fibrates and risk of cancer in tissues with high PPAR-α concentration: a nested case-control study.

BACKGROUND: Fibrates are lipid-lowering agents that act as peroxisome proliferator-activated receptor (PPAR)-α agonists. They have been associated with cancers in experimental models, but data in humans are rare, and among published studies none has investigated cancers in tissues with high PPAR-α concentrations. METHODS: A nested case-control study was performed in a French population-based healthcare database. Adults aged ≥45 years, and free of cancer for 3 years, were followed for 5 years for incident cases of melanoma, non-melanoma skin cancers, thyroid, pancreas, bladder, and kidney cancers. Cases were matched with up to ten controls for age, sex, and diseases that could increase the risk of cancers. Conditional logistic models, adjusted for drug-use as potential confounders, were used to estimate the risk (odds ratio [OR]) of cancers of interest (and individual cancers) associated with cumulative exposure to fibrates (defined daily doses [DDD]). RESULTS: Among the 147,338 eligible subjects, 3,331 (2.3 %) cases of studied cancers were identified. Only use of fibrates >550 DDDs was associated with an increased risk (OR 1.26; 95 % CI 1.12-1.42), and similar results were found for statins (≥1,460 DDDs; OR 1.15; 95 % CI 1.03-1.28). When considering cancers individually, the association was significant for non-melanoma skin-cancer (OR 1.35; 95 % CI 1.14-1.60), and close to significance for bladder cancer (OR 1.26; 95 % CI 0.96-1.64); similar associations with the use of statins were found. CONCLUSIONS: The associations found between fibrate exposure and cancers of tissues with high PPAR-α concentrations were most likely related to residual confounding as similar associations were found for statins.

Validity of a traffic air pollutant dispersion model to assess exposure to fine particles.

INTRODUCTION: Fine particles (PM(2.5)) are an important component of air pollution. Epidemiological studies have shown health effects due to ambient air particles, particularly allergies in children. Since the main difficulty is to determine exposure to such pollution, traffic air pollutant (TAP) dispersions models have been developed to improve the estimation of individual exposure levels. One such model, the ExTra index, has been validated for nitrogen oxide concentrations but not for other pollutants. The purpose of this study was to assess the validity of the ExTra index to assess PM(2.5) exposure. METHODS: We compared PM(2.5) concentrations calculated by the ExTra index to reference measures (passive samplers situated under the covered part of the playground), in 15 schools in Bordeaux, in 2000. First, we collected the input data required by the ExTra index: background and local pollution depending on traffic, meteorology and topography. Second, the ExTra index was calculated for each school. Statistical analysis consisted of a graphic description; then, we calculated an intraclass correlation coefficient. RESULTS: Concentrations calculated with the ExTra index and the reference method were similar. The ExTra index underestimated exposure by 2.2 microg m(-3) on average compared to the reference method. The intraclass correlation coefficient was 0.85 and its 95% confidence interval was [0.62; 0.95]. CONCLUSIONS: The results suggest that the ExTra index provides an assessment of PM(2.5) exposure similar to that of the reference method. Although caution is required in interpreting these results owing to the small number of sites, the ExTra index could be a useful epidemiological tool for reconstructing individual exposure, an important challenge in epidemiology.