253: The effect of statins on the risk of first non-fatal myocardial infarction: A population-based observational study using the PGRx information system.

Background Despite demonstrated positive effects in a number of clinical trials, the evidence is lacking as to the impact of statins on the risk of first myocardial infarction (MI) in real life settings. Objectives To assess the impact of real life statin utilization on the risk of first non-fatal MI Methods Case-control methodology using the pharmacoepidemiological information system PGRx. Data on comorbidities, risk factors and medications were obtained from medical records and patient telephone interviews. General practices (n=371) and cardiology centres (n=60) across France were employed in the study. Cases were patients with the first MI ≤ 1 month before the date of recruitment (n=2238). Controls were patients seen by a general practitioner (GP) with no restriction as to the reasons of consultation (n=2238), matched to MI cases on gender, age, frequency of visits to a doctor, date of recruitment and personal history of non-cardiovascular chronic disease. Statin exposure was defined as any utilisation in the two-year prior to date of MI in cases or recruitment date in controls. Adjusted odds ratios (OR) of the risk of first MI was estimated by multiple conditional logistic regression models. Comparative effectiveness and propensity to use of individual statin molecules were assessed. Results The use of statins was associated with a lower MI risk (adjusted OR 0.67 [95% CI 0.56 - 0.79] for current use (within 2 months before the index date) and 0.73 [0.62 0.86] for any use within 24 months). Among individual statins, rosuvastatin was associated with the lowest MI risk (adjusted OR 0.49 [0.35 - 0.68] for any use in 24 months preceding the index date) followed by simvastatin (0.62 [0.46 - 0.84]). Conclusions In this first major population-based observational study we reproduced the results observed in recent meta-analyses accounting for real life compliance and population variability. The results could be of interest and applicable to other industrialised countries as the observed risk reduction was constant across MI risk levels. [ABSTRACT FROM AUTHOR]