Stuart Gitlow is underwhelmed by a Cochrane review of naltrexone for alcohol dependence. In the process of dismantling Cochrane’s conclusion (Naltrexone appears to be an effective and safe strategy in alcoholism treatment), he questions the value of drink counting.
The authors reviewed 50 studies with nearly 7,800 participants. The review summary indicates that more patients who took naltrexone were able to reduce the amount and frequency of drinking than those who took placebo. They continue by noting that one of nine patients taking naltrexone were helped by the drug. However, when one leaves the summary to read the actual results of the review, a different story is portrayed.
Naltrexone reduced the risk of heavy drinking to 83 percent of the risk of the placebo group. That risk ratio is not a particularly strong endorsement of naltrexone, but more importantly, the outcome measure is completely unimportant. No research has ever demonstrated that a reduction in heavy drinking frequency by patients with alcohol dependence results in any improvement in morbidity or mortality.
Naltrexone was noted to reduce drinking days by 4 percent. If an alcohol dependent patient reduces the number of drinking days within a 100-day timeframe from a 50- to 48-day timeframe, would you consider that to be an improvement?
The big finding in the review comes when the authors note that the effects of naltrexone on return to drinking missed statistical significance. That is, the impact of naltrexone failed to show any difference from placebo in terms of leading the alcohol dependent patient to maintain sobriety. Since sobriety, abstinence, and recovery are our goals for the treatment of addictive disease, naltrexone is demonstrated here not to work.