To the Editor: Claiming "promise off-label for treating alcohol use disorder" in a review about "evidenced-based pharmacologic treatment" is an oxymoron, as illustrated by 2 examples that were not critically appraised in the review. (1)
First, nalmefene is not approved in the United States. The European Medicines Agency granted approval on a retrospective subgroup analysis of 2 randomized controlled 6-month studies with only one-quarter of the patients and major attrition biases. The European Medicines Agency accepted a reduction in drinking as an end point based on a white paper issued by the very same authors having links of interest with Lundbeck, the drug company. (2)
Second, the results of baclofen's first pivotal trial (NCTO1738282) were negative. The results of the second trial (NCT01604330) were published as positive despite major limitations, including major changes to the initial protocol and a 2-fold increase in mortality with baclofen. (3) Mortality was not statistically significant, the series being short, but a large-scale French pharmacoepidemiologic study of patients with alcohol use disorder (n= 165,334) retrieved a dose-dependent increase in mortality with baclofen vs acamprosate and naltrexone (see reference 8 in the article by Wallach et al (4)). Accordingly, the Special Scientific Committee for the assessment of baclofen in alcohol use disorder concluded that the benefit to harm ratio was negative. Because the director of the French medicines agency overturned this conclusion and the drug company bypassed the usual European process for marketing approval, an unusual choice, France is the only country worldwide where baclofen has marketing approval. Exception in medicine never benefits patients.
As a general rule, off-label use is about caveats, not promises: (1) exposing patients to the unnecessary risk of many adverse events, including mortality, for an all-too-often uncertain efficacy, (5) and (2) hindering the development of evidence-based medicine and therapeutic innovation.
Could the promotion of off-label use through studies with nonvalidated surrogates in short series of highly selected patients who are not representative of the real-life challenges be a strategy to avoid robust assessment of benefit to harm ratios on relevant clinical outcomes by the medicines agency? Coincidentally, in 50 ongoing trials registered on ClinicalTrials.gov for alcohol use disorder, most drugs are repurposed (ie, off-label), series of patients are small (median n=123), durations are short (12 months), and only one-quarter of trials fulfill Food and Drug Administration--recommended responder outcomes (including abstinence or no heavy drinking days). (4) Nonvalidated surrogates, such...