Abstract
In this issue ofTHE JOURNAL, Ross-Degnan et al1present a fascinating analysis of the effect of the withdrawal of zomepirac sodium from the market. They note that zomepirac prescribers, faced with withdrawal of the drug from the market, switched, not only to other nonsteroidal anti-inflammatory drugs, but surprisingly to propoxyphene (hydrochloride or napsylate) and barbiturate-containing analgesics. These drugs carry risks that arguably are as serious as those of zomepirac. Based on those findings, they suggest that the risks posed by substituted products be considered before making a decision to withdraw a drug from the market. Although an appealing theoretical concept, it seems unlikely that such a suggestion is of any practicable use. To understand the basis for this statement, it may be instructive to review the bases on which drugs have been withdrawn from the American marketplace in the past 30 years. The largest number of drug products