The End of My Addiction

placebo group. ANCOVA also revealed significant effects of treatment and time on both compulsive [ F treatment (1,78) = 4.60, P < 0.05; F time (3,78) = 6.40, P < 0.0005] (Fig. 3, centre panel) and obsessive [ F treatment (1,78) = 5.06, P < 0.05; F time (3,78) = 11.53, P < 0.00005] (Fig. 3, bottom panel) drinking subscales of OCDS, with scores in the baclofen groups constantly lower than those of the placebo group throughout T1–T4.

    Fig. 3. Obsessive Compulsive Drinking Scale (OCDS) total (top panel), OCDS Compulsive Drinking subscale (centre panel) and OCDS Obsessive Drinking subscale (bottom panel) scores in baclofen and placebo groups at T0 (baseline) and over the four weekly visits (T1–T4). Each value is the mean ± SEM of 17 patients in the baclofen group and 11 in the placebo group.
    ANCOVA revealed significant effects of both treatment and time factors on state anxiety [ F treatment (1,78) = 4.62, P < 0.05; F time (3,78) =3.05, P < 0.05] (Fig. 4, top panel), with lower scores in the baclofen than placebo group at T1–T4. In contrast, no significant difference was observed in depression score [ F treatment (1,78) = 0.70, P > 0.05; F time (3,78) =2.28, P > 0.05] (Fig. 4, bottom panel).

    TABLE 1. Main biological markers of alcohol misuse in patients treated with baclofen or placebo at the start (T0) and at the end (T4) of the study
    MCV, mean cell volume; GGT, [.gamma]-glutamyltranspeptidase; AST, aspartate aminotransferase; ALT, alanine aminotransferase. Each value is the mean ± SEM of 17 patients in the baclofen group (with the exception of the GGT data for 16 patients) and 11 patients in the placebo group.
    Table 1 reports values obtained in laboratory investigations before and after baclofen or placebo administration.
    No serious systemic or single-organ event leading to drug cessation was reported and no patient discontinued the drug. Tolerability was fair in all patients; as previously reported (Addolorato et al ., 2000 b ), the most common side-effects were sleepiness (two patients), tiredness (one patient), vertigo (one patient) in the baclofen group and abdominal pain (one patient) in the placebo group, which resolved within 1–2 weeks of drug treatment and did not recur. No patient reported euphoria or other pleasant effects caused by the drug. No subject showed craving for baclofen. At drug discontinuation, neither drug withdrawal syndrome nor side-effect due to drug suspension was observed.

    Fig. 4. State anxiety score, evaluated by the State Anxiety Inventory Test (STAI-Y1), and current depression score, evaluated by the Zung Self-rating Depression Scale, in baclofen and placebo groups at T0 (baseline) and over the four weekly visits (T1–T4). Each value is the mean ± SEM of 17 patients in the baclofen group and 11 in the placebo group.
    Discussion
    Recent preclinical (Colombo et al ., 2000, 2002) and preliminary clinical data (Addolorato et al ., 2000 b , 2002) suggest that the GABA B receptor agonist, baclofen, may be effective in the treatment of patients with alcohol problems. However, to date, no double-blind, randomized placebo-controlled study has been conducted. In spite of the limitation due to the low number of patients evaluated, the results of the present study indicate that administration of relatively low doses of baclofen to alcohol-dependent patients is more effective than placebo in inducing and maintaining abstinence from alcohol (both in terms of number of patients reaching complete abstinence and CAD), reducing alcohol intake, suppressing alcohol craving in both its ‘obsessive’ and ‘compulsive’ features, and reducing state anxiety. Baclofen, however, did not differ from placebo in terms of reduction of current depression.
    In agreement with our previous observation (Addolorato et al ., 2000 b ), abstinence from alcohol or reduction in alcohol intake was achieved within the first week of baclofen treatment and was maintained throughout the treatment period. The increased efficacy of baclofen over placebo may be related to its suppressant effect on craving; indeed, the drug produced a rapid decrease in the ‘compulsive’ and ‘obsessive’ components of craving, as indicated by the immediate reduction in mean score of both OCDS subscales. It is noteworthy that an anti-craving effect of baclofen has already been observed with other substances of abuse, particularly cocaine in cocaine users (Ling et al ., 1998). The anti-craving effect of baclofen may depend on