For Debate – the COMBINE study

Read this before you are on the receiving end of a Vivitrex sales pitch:

The monumental study on combined pharmacotherapies and behavioural interventions for alcohol dependence (COMBINE) was set up ‘in order to determine if improvements in treatment outcome for alcohol dependence can be achieved by combining pharmacotherapy and behavioral interventions’ [1]. When the main results were published in May 2006 [2], it was clear that no such combining effect had been detected in the trial. Neither had the combination of the two pharmacological interventions—naltrexone and acamprosate—nor the combination of any or both of these with a behavioural intervention [combined behavioural intervention (CBI)], improved the outcome. At the same time, and in line with the results from other multi-site studies, all the nine different treatment interventions (combinations) showed a substantial reduction in drinking, including those who had received placebo pills. In general, the commentaries on these results, as well as the conclusions from the COMBINE study group, underlined one recommendation for clinical practice: support for the use of naltrexone in primary care settings. However, the empirical basis for this recommendation appears to be somewhat weak.

Given the fact that the magnitude of difference between the two pharmacological interventions is quite small at the end of treatment (i.e. 16 weeks), and basically disappears at the 1-year follow-up, it is possible to claim that the COMBINE study provides somewhat weak support for preferential use of either. In the same way, neither naltrexone nor acamprosate provide a better outcome than non-pharmacological interventions. In comparison with a group that received CBI and placebo (within the context of medical management), naltrexone in combination with CBI did not improve outcomes. Because the mechanisms behind the improvement of the CBI/placebo group must be mind-mediated and not of a neurochemical character, these findings indicate that there is an absence of any neurochemical effect associated with naltrexone which is additive to the psychological mechanisms at work in the CBI/placebo intervention. Although it is not logically necessary that specific pharmacological and placebo/psychological effects have to be additive [3], it does seem likely that they are. Why would a placebo produce a non-pharmacological effect while a pharmacological intervention, indistinguishable from the placebo, would not?

Along the same lines, the lack of empirical support for acamprosate is reported as surprising, but not discussed with any reference to the possibility that the mechanisms of action of acamprosate might not be there or, if there, not of any importance beyond the placebo effect. Given that researchers within the COMBINE study group have pointed previously to the fact that: ‘Despite its widespread use . . . and proven efficacy, the precise mechanism of action is unknown’ ([7], p. 60), it would have been interesting to see an elaboration of this theme in light of the lack of support for acamprosate in the COMBINE study.

One thought on “For Debate – the COMBINE study

  1. I think the research is framed the wrong way. If you are doing something to help manage acute and post-acute withdrawal symptoms, the measure of success is not whether they are still sober in six months. The measure of success is about whether they are comfortable enough in their skin to come back the next day, or to sit in group without throwing up. Whether or not they are still sober at some future point in time will result from a hundred variables which the research couldn't possibly isolate. We need to ask the right questions to get the right answers. (please visit me at!)

Comments are closed.