Spontaneous Reporting of Adverse Drug Reactions

Drug regulatory authorities in many developed countries have established spontaneous adverse reaction reporting schemes to monitor drug safety [1]. The United Kingdom yellowcard system, organised by the Committee on the Safety of Medicines (CSM), is one of the oldest and the most successful of these national schemes. With typical Anglo-Saxon self-deprecation however, both the medical profession and the lay press have tended to emphasise the limitations of the yellow-card scheme at the expense of its achievements. Spontaneous reporting schemes do indeed have substantial limitations [2]. Firstly, they rely on doctors reporting suspected (and not necessarily proven) adverse reactions. Few adverse reactions, however, are uniquely iatrogenic and an efficient spontaneous reporting scheme depends on doctors' professional skill and judgement in associating their patient's diseases with the drug(s) they have taken. Such schemes are therefore inherently more likely to detect those adverse reactions which occur soon after the start of treatment, and which produce syndromes commonly having an iatrogenic basis such as anaphylactoid reactions, dystonia-dyskinesias and various dermatoses. Conversely, reactions which have a long latency, or which are expressed as conditions only rarely having a recognised iatrogenic basis, may remain unrecognised by this technique. The delayed recognition of the oculomucocutaneous syndrome with practolol is a clear example of this. Secondly, in the United Kingdom (as elsewhere), only a small proportion of even serious adverse drug reactions are reported to the CSM [3]. Underreporting not only results in a substantial loss of useful scientific data, but may also lead to serious bias since reporting rates are conditioned both by manufacturers' promotional claims and by adverse publicity in the medical literature or the media. Thirdly, reporting rates tend to decline with time after a product has been marketed. To some extent this is inevitable: whilst the CSM asks doctors to report all suspected adverse reactions for newly (within three years) marketed drugs, it asks to be notified of only serious reactions to older ones since it would be unnecessarily burdensome (and of little real value) for doctors to continue indefinitely reporting bradycardia with digoxin, nausea with morphine, or dyspepsia with aspirin. Unfortunately, reporting rates of even serious reactions fall off with time. Thus, Inman [4] showed in 1977 that only five out of 44 (11 per cent) of patients dying from blood dyscrasias associated with phenylbutazone or oxyphenbutazone were reported to the CSM; and between 1964 and 1985 there have been only 192 reports [5] of gastrointestinal bleeding or perforation with aspirin! The number of adverse reaction reports for a particular drug will be a function of its usage, as well as its inherent toxicity. The evaluation of spontaneous reporting rates must therefore take into account the size of the exposed population. In the United Kingdom, reporting rates are