Epidemiological information on the abuse of opioid agonist-antagonists is scanty. Most of the available information relates to pentazocine, the 'oldest' of this class of drugs, and it appears that a sporadic episode of its intravenous abuse in several cities in the United States, is subsiding: reformulation of pentazocine tablets to contain naloxone may have contributed to the decline in this abuse. Apart from a few reports of buprenorphine abuse, there are only occasional accounts of abuse of other drugs in this class. However, the lack of data does not necessarily mean that abuse does not occur and it is essential that epidemiologically sound investigations should be made to explore the existence and extent of any such problem. The interpretation of epidemiological data relating to agonist-antagonist abuse is discussed.