I am most grateful for the comments on my paper [1]. By and large, it appears that Bühringer & Pfeiffer-Gerschel [2] and Orford [3] are in agreement with my general line of argument concerning what we can learn from the Combining Medications and Behavioral Interventions (COMBINE) study [4], whereas Kalant [5] raises several objections. While Kalant disagrees with my conclusion that—given the relatively small differences both at the end of treatment and at the 1-year follow-up—the COMBINE study provides rather weak support for the preferential use of either treatment (i.e. naltrexone and acamprosate), he also brings in an unpublished re-analysis provided by one of the researchers in the COMBINE research group [B. J. Mason]. In this new analysis the inclusion of the group that received only combined behavioral intervention (CBI)/no pill (which was not included in the original analysis) brings about a result where all other treatment groups were better than CBI/no pill but did not differ from each other. Kalant brings forward these results as support for his argument: ‘Therefore it cannot be concluded that acamprosate was less effective’. That might be true, but the results from the new analysis seem to fit well with my conclusion that the COMBINE study cannot be regarded as providing strong empirical support for the use of either naltrexone or acamprosate, as the analysis also shows that there is no difference between the treatment groups that received placebo and those given the specific pharmacological interventions.

Kalant finds my critique of the COMBINE study to be a ‘sweeping rejection of pharmacotherapy’. This stands out as based upon a misinterpretation of my text. I have argued that the support for naltrexone (and acamprosate) in the COMBINE study is weak, but never extended this conclusion towards a rejection of the possibilities of pharmacotherapy effects in general. I have also argued that the COMBINE study can be seen as additive to the findings of such major studies as Project MATCH [6] and the United Kingdom Alcohol Treatment Trial (UKATT) [7] study in the respect that it can be interpreted as supportive of a common factor (or contextual) model, which points up the necessity of breaking new ground in the field of addiction treatment, and particularly so in the case of treatment mechanisms.

The major difference between Kalant's position and my own (and, as I understand it, also Orford's and Bühringer's & Pfeiffer-Gerschel's) can be found in our evaluations of the general state of the addiction treatment field. Whereas Kalant finds good reason to dig deeper in the spot upon which he is already standing, I find the accumulated research on treatment outcome to be a good reason to look for opportunities in new places (or, as Orford puts it, I have joined the queue of dissenters). Given his professional background, it is not surprising that Kalant is more interested in the role of genetics and the neurochemical aspects of addiction than the outcomes from psychosocial interventions.