However, the number of placebo-controlled or double-blind studies conducted are few even though such study approaches offer higher levels of evidence than observation studies in clinical cohorts. In addition, alcohol and SUDs are often unequivocal exclusion criterion in studies on therapeutic efficacy (Swann et al. 2005). In the past, lithium preparations and valproate (antiepileptic) were used most often for the treatment of bipolar disorders, carbamazepine (antiepileptic) is less often prescribed. Lithium has been shown to be less effective with comorbid patients (overview by Cerullo and Strakowski, 2007). Admittedly, the use of antiepileptics is also limited in these patients. Carbamazepine and valproate, for example, can cause induction of liver transaminases (ALAT, ASAT, γGT) and in rare cases liver failure. Patients with severe alcohol dependence often already suffer from liver function disorders.

Recently, controlled clinical studies were published. A placebo-controlled double- blind study on 59 individuals with comorbid bipolar and AUDs was conducted for over 24 weeks (Salloum et al. 2005). Besides improvement of the affective symptoms, less drinking days and lower consumption were shown in the valproate group, which in turn correlated with the plasma valproate levels. Thus, valproate remains as one of the possible pharmacological strategies to treat comorbid patients. Yet another complementary strategy is the use of antipsychotics in comorbid patients. The effectiveness of aripiprazole was initially tested in a relatively small group of 18 comorbid individuals with bipolar or schizo-affective disorders with an AUD for more than 12 weeks (Brown et al. 2005). Although all affective symptoms improved, results on alcohol consumption remained unsatisfactory. However, craving for alcohol and the purchase of alcoholic beverages decreased at this time. The same study group then used quetiapine in dosages more than 600 mg/d for over 12 weeks on 115 comorbid patients in another study (Brown et al. 2008), 102 of these patients could be examined at least once prospectively. Other than depressive symptoms, neither the manic symptoms nor characteristics of alcohol dependence showed any improvements.

A subsequent study combined both pharmacological treatments mentioned above. Bipolar patients dependent on alcohol were treated with either lithium or valproate combined with quetiapine in flexible dosage (300–800 mg/d) or placebo (Stedman et al. 2010). Three hundred sixty-two patients were randomized for treatment of 12 weeks (n = 176 with quetiapine as add on, n = 178 with placebos). No significant effect on the characteristics of alcohol dependence could be observed.

Naltrexone (NTX), which has been shown to be effective in relapse prevention of alcohol dependence especially in the USA, was examined in three studies. Initial investigations on comorbid patients with NTX alone or in combination with disulfiram reported favorable effects on several characteristics of consumption, e.g., amount of alcohol consumed (Petrakis et al. 2005) as well as affective symptoms, craving, and consumption days (Brown et al. 2006a, b).

Naltrexone was then examined by the latter research group for its effectiveness in comorbid individuals (Brown et al. 2009). Fifty affected individuals with comorbid bipolar and AUD were examined with 50 mg/d NTX (as an add-on to mood stabilizers) or placebo for over 12 weeks. Effects of the additional NTX on drinking days, craving, and liver enzymes, however, showed only a statistical trend.

The results of pharmacologic strategies on the treatment of comorbid persons remain conflicting. The best results were reported with valproate but these were no longer seen when valproate was combined with quetiapine. Other substances such as aripiprazole or NTX were tested only in small studies or the effectiveness showed only a statistical trend.