The role of trauma in drug use disorders

Abstract

The essential components of the self-medication hypothesis are: (i) the causal role of distress—psychoactive substances are used to relieve negative affect—and (ii) pharmacological specificity—the choice of drug for its action in relieving a specific type of distress. Darke's review of the evidence on heroine dependence and psychoactive substances, in general [1], concludes that the second component is unsupported by the available evidence, given the prevailing pattern of polydrug use. The conclusion on the first essential component—the causal role of distress—is that it is securely established. A critical part of the evidence on the etiologic role of distress in drug use disorder comes from the large literature on the heavy burden of trauma and childhood maltreatment in the lives of drug users. It is this claim that needs further reappraisal. Recent epidemiological evidence casts doubt on the role of trauma in the etiology of substance use disorders. We reported recently that people who experienced traumatic events but did not develop post-traumatic stress disorder (PTSD) are not at increased risk of nicotine dependence, alcohol or drug use disorder 2, 3. Importantly, this finding emerged in prospective data and retrospective life-time data, using statistical approaches that take into account the temporal order between trauma and onset of substance use disorder. In several community samples, we found that only a small minority of trauma victims—those with PTSD resulting from the trauma—were at increased risk of drug use disorders 2, 3 and emerging features of drug use disorder 4, compared to people who have not experienced trauma. The majority of trauma victims does not develop PTSD and is not at elevated risk of substance use disorders. This pattern also characterized the relationship of trauma with the subsequent onset of major depression and anxiety disorders 5. One possible interpretation is that only those trauma victims who have developed PTSD are distressed and are motivated to turn to drugs to self-medicate their distressing PTSD symptoms. A competing interpretation of the evidence that should be considered is that the minority of trauma victims who have developed PTSD had pre-existing susceptibilities that account for their PTSD as well as their drug use disorders. The role of predispositions in PTSD, primarily family history of psychopathology, neuroticism, cognitive ability and pre-existing maladjustment, is well documented in the literature 6-10. Darke's review relies heavily upon evidence of high rates of childhood sexual and physical abuse reported by people with drug use disorders. Childhood sexual and physical abuse occurs in the context of pervasive family pathology, and it is nearly impossible to isolate the effect of childhood abuse from other childhood adversities. Furthermore, evidence of history of child abuse among people with drug use disorder suffers from two critical methodological flaws. First, with very few exceptions, the evidence comes from retrospective accounts and is thus subject to recall bias, a limitation that threatens the validity of casual inferences. People with drug use disorders and associated psychiatric disturbances might be more likely to recall negative experiences, whereas those with no psychiatric disturbances might be more likely to forget and less likely to attribute casual meaning to objectively similar events 11-13. A reporting bias associated with psychiatric status is likely to influence reports of childhood maltreatment and lead to an apparent (but spurious) association between drug disorders and childhood maltreatment. Secondly, the evidence fails to take into account major sources of confounding, primarily heritability and genetic factors that might be implicated in childhood maltreatment as well as subsequent drug use disorder 14. In sum, the evidence on trauma and child abuse, on which the weight of the argument for the etiologic role of distress relies, is challenged by alternative explanations and severe methodological limitations. None.