Adverse drug reaction-related hospitalisations: a nationwide study in The Netherlands.

The incidence of adverse drug reaction (ADR)-related hospitalisations has usually been assessed within hospitals. Because of the variability in results and methodology, it is difficult to extrapolate these results to a national level. To evaluate the incidence and characteristics of ADR-related hospitalisations in The Netherlands in 2001. We conducted a nationwide study of all hospital admissions in 2001. Data were retrieved from a nationwide computer database for hospital discharge records. All acute, non-planned admissions to all Dutch academic and general hospitals in 2001 were included in the study (n = 668 714). From these admissions we selected all hospitalisations that were coded as drug-related, but intended forms of overdose, errors in administration and therapeutic failures were excluded. Hence, we extracted all ADR-related hospitalisations. We compared age, sex and the risk of a fatal outcome between patients admitted with ADRs and patients admitted for other reasons, as well as the most frequent main diagnoses in ADR-related hospitalisations and which drugs most frequently caused the ADRs. In addition, we evaluated to what extent these ADRs were reported to the Netherlands Pharmacovigilance Centre Lareb for spontaneous ADR reporting. In 2001, 12 249 hospitalisations were coded as ADR related. This was 1.83% of all acute hospital admissions in The Netherlands (95% CI 1.80, 1.86). The proportion increased with age from 0.8% (95% CI 0.75, 0.85) in the <18 years group to 3.2% in the >/=80 years group (95% CI 3.08, 3.32). The most frequent ADR-related diagnoses of hospitalisations were bleeding (n = 1048), non-specified 'unintended effect of drug' (n = 438), hypoglycaemia (n = 375) and fever (n = 347). The drugs most commonly associated with ADR-related hospitalisations were anticoagulants (n = 2185), cytostatics and immunosuppressives (n = 1809) and diuretics (n = 979). Six percent of the ADR-related hospitalisations had a fatal outcome (n = 734). Older age and female gender were associated with ADR-related hospitalisations. Only approximately 1% of the coded ADRs causing hospitalisation were reported to our national centre for spontaneous ADR reporting. The proportion of ADR-related hospitalisations is substantial, especially considering the fact that not all ADRs may be recognised or mentioned in discharge letters. Under-reporting of ADRs that result in hospital admission to our national centre for spontaneous ADR reporting was considerable.