The use of cocaine by heroin-dependent individuals, or by patients in methadone or buprenorphine maintenance treatment, is substantial and has negative consequences on health, social adjustment and outcome of opioid-addiction treatment. The pharmacological reasons for cocaine use in opioid-dependent individuals, however, are poorly understood and little is known about the patterns of heroin and cocaine co-use. We reviewed anecdotal evidence suggesting that cocaine is co-used with opioid drugs in a variety of different patterns, to achieve different goals. Clinical and preclinical experimental evidence indicates that the simultaneous administration of cocaine and heroin (i.e. 'speedball') does not induce a novel set of subjective effects, nor is it more reinforcing than either drug alone, especially when the doses of heroin and cocaine are high. There is mixed evidence that the subjective effects of cocaine are enhanced in individuals dependent on opioids, although it is clear that cocaine can alleviate the severity of symptoms of withdrawal from opioids. We also reviewed preclinical studies investigating possible neurobiological interactions between opioids and cocaine, but the results of these studies have been difficult to interpret mainly because the neurochemical mechanisms mediating the motivational effects of cocaine are modified by dependence on, and withdrawal from, opioid drugs. Our analysis encourages further systematic investigation of cocaine use patterns among opioid-dependent individuals and in laboratory animals. Once clearly identified, pharmacological and neuroanatomical methods can be employed in self-administering laboratory animals to uncover the neurobiological correlates of specific patterns of co-use.