Using the Adverse Reactions Register to study the effects of age and sex on adverse drug reactions.

Many countries maintain a register of reports of adverse reactions to drugs. Although reports are made voluntarily by doctors and dentists these registers contain much information that may be useful in pharmaco-epidemiological studies. We show how the epidemiology of adverse drug reactions (ADRs) can be studied using data from such a register, together with estimates of the numbers of prescriptions of the drug, categorized by age and sex of the patient. Reporting rates of ADRs are compared for different ages and sexes by logistic analysis, using extrapyramidal reactions to metoclopramide as an example. Highly significant effects of sex and age on reporting rates are found, with young women being most at risk. We consider possible alternative explanations of this finding, and show that relative overdosage does not account for the differences between the age and sex groups. Other possible errors in the data are also considered unlikely to alter these conclusions; these include differential bias in reporting ADRs for various age-sex groups, and errors in the prescription data. Studies based on the adverse reactions registers are useful for formulating hypotheses which can be tested in a prospective study.