[French pharmacovigilance database system: examples of utilisation]

The French drug surveillance (pharmacovigilance) system is based on a network of 31 regional centres which receive adverse drug reaction (ADR) reports from health professionals and are drug information centers. Cases are entered into a common database, with causality scores. This database contains large amounts of data, which may be used for pharmaco-epidemiological studies. As an example, all cases in which an antihypertensive drug, suspect or not, was cited were identified. ACE-inhibitor cough was also explored.

Since 1985, > 70,000 case reports have been entered into the database. 63 per cent were reported by specialists, 20 per cent by GPs. 54 per cent came from University Hospitals, 21 per cent from private practice. The most numerous age group was 60 to 69. The overall sex ratio (F/M) was 1.28, the female preponderance being most marked at < 39 and > 70 years of age. 43 per cent took only one drug, 20 per cent two drugs, 13.4 per cent three, and 24 per cent > three drugs. The most frequently reported effects concerned the skin and appendages (15 per cent), general status and central nervous system (9.5 per cent each), platelets, liver, and GI systems (6 per cent each). Outcome was favourable in 74 per cent. Dechallenge was positive in 71 per cent, rechallence in 6 per cent. 3.4 per cent of the patients died; in 2.2 per cent death was related to a reaction. Causality assessment indicated close temporal relationship (C2 or C3) in 69 per cent of cases; in 51 per cent of cases, no other obvious cause was found. 66 per cent of the reactions were labelled when reported. The database could also be used to explore drug utilisation: as an example, we studied the age and sex distribution of reports containing antihypertensive drugs, irrespective of their possible causal role in the reaction. Antihypertensives were mentioned in 14 per cent of the reports. The age distribution was skewed towards greater age, with a maximum of 70 years. F/M was 1.57, with more M use < 20 and 30-59, whereas F were more common between 20-29 and 60 years. beta-blockers were more often associated with patients under 70, whereas above 70 diuretics and centrally acting antihypertensive drugs were more often reported. This could be related to greater use or worse tolerance of these drugs. As an example of the exploration of a specific drug-reaction relationship, we explored the relationship between the use of ACE inhibitors (ACEI) and cough. ACE inhibitors were present in 6 per cent of cases, but in 75 per cent of reports of cough. F/M was 1.29 (NS) for all reports concerning ACEI, 1.28 for cough unrelated to ACEI, 2.1 for cough with ACEI (P < 0.05). Cough was present in 12 per cent of all reports concerning ACEI. There was no clear difference between ACEI for cough or sex ratio; women cough more with ACEI. This does not seem related to greater ACEI use by women or to greater sensitivity of women to cough. The reason for this sex difference remains to be explained. There are large amounts of essentially underutilized data in drug surveillance databases. How they can or should be used remains to be validated.