The aim of this prospective pharmacovigilance study was to assess the incidence and the preventability of adverse drug reactions (ADRs) leading to hospital admissions. All patients admitted to the Toulouse University Hospital through the medical admission ward during four non-consecutive weeks were included in the study. Characteristics of patients admitted for a suspected ADR (cases) were compared with those admitted for other reasons (controls). All cases were reviewed by both a pharmacologist and an emergency medicine specialist. Among a total of 671 admissions, 44 ADRs were identified. The incidence of hospital admissions for ADRs was 6.1 per 100 admissions [4.4-8.3]. Cases were exposed to a higher number of drugs (3.6 vs. 1.7; p < 0.001). Some classes of drugs were significantly associated with a higher risk of ADRs: antineoplastic, anti-infectious or musculoskeletal drugs. Haematological, metabolic and electrolytic, liver/gastrointestinal or cutaneous disorders were causes of admissions significantly more frequently related to ADRs. Using Imbs's preventability scale, 65 per cent of ADRs were categorized as 'definitely unpreventable', 26 per cent 'potentially preventable' and 9 per cent 'definitely preventable'. These results underline the frequency of ADRs leading to hospitalization, with 35 per cent which are more than likely preventable. Further studies are needed to validate this preventability scale in order to obtain an easier, more reliable and more reproducible tool.