Correlations between prescriptions and drugs taken in self-poisoning. Implications for prescribers and drug regulation

Abstract

Objective: To compare prescription data for Australia with drugs ingested in self-poisoning and suicide, to determine which drugs are over-represented.

Design: Comparison of data on drugs taken in self-poisoning admissions and suicides with Australian prescription drug dispensing data from the Drug Utilization SubCommittee (DUSC).

Setting: Newcastle and Lake Macquarie, Australia, 1989-1992.

Subjects: Between July 1989 and June 1992, 1159 prescription drugs were taken in overdose. Eighty-three drug-related deaths were investigated by the coroner between 1987 and 1992. On 48 occasions a prescription drug was the primary cause of death.

Results: Drugs over-represented in self-poisoning (relative to Australian prescriptions) included not only those prescribed for psychiatric conditions (antidepressants, neuroleptics and lithium), but also benzodiazepines, barbiturates and other anticonvulsants. The highest odds ratios for death when adjusted for prescription numbers were for short-acting barbiturates (523.7; 95% confidence interval [CI], 207-1322), chloral hydrate (58.1; 95% CI, 18.1-187), colchicine (27.9; 95% CI, 3.8-202), dextropropoxyphene (20.8; 95% CI, 8.8-48.9), tricyclic antidepressants (13.3; 95% CI, 7.2-24.5) and anticonvulsants (11.6; 95% CI, 4.1-32.2).

Conclusions: Short-acting barbiturates, chloral hydrate and dextropropoxyphene have little or no clinical advantage over alternatives and excessive toxicity in overdose. They should be removed from the market. The toxicity of anticonvulsants and colchicine should be considered when they are prescribed, and smaller amounts per prescription may be advisable for high risk patients using these and other toxic drugs.