Fatal adverse drug reactions (ADRs) are a significant public health issue; it has been estimated that these affect 1 in 350 of hospitalized patients in the United States [1]. Because they are not common, there is a risk that potentially fatal ADRs may not be recognized at the time of drug launch and spontaneous reporting is therefore crucial for early identification. This research was performed to determine the reporting rate of fatal suspected ADRs (SADR) to the Northern Regional Monitoring Centre (RMC) of the Committee on Safety of Medicines (population served 3.1M) and compare the characteristics of fatal with non-fatal SADR reports (‘yellow cards’).

All reports of fatal SADR and a similar number of randomly selected non-fatal SADRs submitted to the Northern RMC between 1986 and 2001 were identified. The age, sex, reporter type, suspected drug and suspected drug reaction were recorded. Causality assessment was evaluated by two independent observers in a randomly-selected sample of 72 fatal and non-fatal reports.

During the study period 275 fatal SADR reports were received, representing 1.7% of all SADR reports. Causality was rated as likely in 58 (81 %) and unlikely in 2 (3%) of the 72 reports assessed. Patients with fatal SADR were significantly older (58 ± 22 years) than those with non-fatal SADR (48 ± 22 years); no significant sex difference was identified. Hospital doctors and hospital pharmacists reported a disproportionately high number of fatal SADR ( χ² = 58.8; p < 0.01) while general practitioners reported most non-fatal episodes ( χ² = 38.8; p < 0.01). Only one fatal SADR was reported by a coroner. The most frequently implicated drug classes associated with fatal SADRs were central nervous system (25.5%) and cardiovascular (16.9%) drugs; a higher proportion of SADR were fatal for drugs used to treat malignant disease and for immunosupression ( p < 0.01), while the converse was true for immunological products, vaccines and gastrointestinal drugs ( p < 0.01). A significantly higher proportion of cardiovascular ( p < 0.01) and respiratory ( p < 0.05) SADRs were fatal. The most common fatal SADRs were sudden death and gastrointestinal haemorrhage. Of the fatal SADRs, 149 (45%) were listed on the drug datasheet or summary of product characteristics at the time of reporting.

Fatal and non-fatal SADR reports appear to differ in many respects although selective reporting bias may account for some of the differences. It is important to encourage wider use of the Yellow Card reporting scheme to detect serious ADR signals at an earlier stage. It is particularly important to encourage reporting from hospitals and coroners if potentially fatal drug safety issues are to be identified rapidly.

1. Lazarou J, et al. JAMA 1998; 279 : 1216.