Lysergic acid diethylamide (LSD) for alcoholism : meta-analysis of randomized controlled trials, Teri S Krebs and Pål-Ørjan Johansen, 2012

Lysergic acid diethylamide (LSD) for alcoholism : meta-analysis of randomized controlled trials

Teri S Krebs and Pål-Ørjan Johansen

Journal of Psychopharmacology, 2012, 1–9

DOI: 10.1177/0269881112439253



Assessments of lysergic acid diethylamide (LSD) in the treatment of alcoholism have not been based on quantitative meta-analysis. Hence, we performed a meta-analysis of randomized controlled trials in order to evaluate the clinical efficacy of LSD in the treatment of alcoholism. Two reviewers independently extracted the data, pooling the effects using odds ratios (ORs) by a generic inverse variance, random effects model. We identified six eligible trials, including 536 participants. There was evidence for a beneficial effect of LSD on alcohol misuse (OR, 1.96; 95% CI, 1.36–2.84; p = 0.0003). Between-trial heterogeneity for the treatment effects was negligible (I² = 0%). Secondary outcomes, risk of bias and limitations are discussed. A single dose of LSD, in the context of various alcoholism treatment programs, is associated with a decrease in alcohol misuse.

Keywords : Alcoholism, alcohol-related disorders, hallucinogens, meta-analysis, psychedelics, substance-related disorders


Alcohol is said to cause more overall harm than any other drug (Nutt et al., 2010). Alcohol contributes to about 4% of total mortality and about 5% of disability adjusted life-years to the global burden of disease (Rehm et al., 2009). Despite the often extreme individual and social consequences of alcohol misuse, many users find it challenging to stop drinking. Alcoholism, also called alcohol dependence, continues to be difficult to treat, and many patients do not achieve recovery from existing treatments (Schuckit, 2009).

Numerous clinical investigators have claimed that treating alcoholics with individual doses of lysergic acid diethylamide (LSD), in combination with psychosocial interventions, can help to prevent a relapse of alcohol misuse, for example, by eliciting insights into behavioural patterns and generating motivation to build a meaningful sober lifestyle (Dyck, 2008). LSD is wellknown for inducing spectacular and profound effects on the mind (Henderson and Glass, 1994; Passie et al., 2008). It has previously been used in standard treatment programs for alcoholism at many clinics, but, unfortunately, assessments of the clinical value of LSD have not been based on formal systematic review and meta-analysis (Mangini, 1998). Hence, we have performed a quantitative evaluation of the effectiveness of LSD for alcoholism, based on data from randomized controlled clinical trials.


Search strategy and selection criteria

We searched the PubMed and PsycINFO databases (1943–2010), without language restrictions, using the following terms: LSD, lysergic, lysergide, psychedelic*, or hallucinogen*; and alcohol*, addict*, or dependence. We independently inspected the search results by reading the titles and abstracts. We retrieved each potentially relevant publication located in the search and assessed it for inclusion, subsequently examining the reference lists of eligible studies and relevant review articles. We supplemented our search for trials by contacting experts. If publications lacked important information, we attempted to contact study investigators and institutions.

We specified inclusion and exclusion criteria and defined primary and secondary outcomes in the meta-analysis study protocol. We included randomized controlled trials of LSD for alcoholism, in which control condition involved any type of treatment, including doses of up to 50 mcg LSD as an active control. If a trial included multiple randomized treatment arms, all participants in the eligible LSD arms and all participants in the eligible control arms were pooled for analysis. We excluded participants with schizophrenia or psychosis from analysis, as psychosis is recognized as a contraindication for treatment with LSD (Johnson et al., 2008; Passie et al., 2008).

Data extraction
Both reviewers independently extracted data and rated the risk of bias of each included trial. Differences between the reviewers were resolved through discussion. The following were recorded from each trial where available: intervention characteristics (LSD dose, control condition, additional treatments); participant characteristics (number, gender, age, inclusion and exclusion criteria); information given to the participants on the study and the effects of LSD; trial characteristics (publication year, location, funding source); outcomes (primary and secondary outcomes, time of follow-up, method of outcome assessment); evaluation of each domain of the Cochrane risk of bias assessment tool (sequence generation, allocation concealment, blinding, incomplete outcome data, selective outcome reporting) (Higgins and Altman, 2008). Primary outcomes were alcohol misuse, defined as alcohol use or consequences of alcohol use, as systematically measured by interview or self-report at the first reported follow-up. Secondary outcomes were alcohol misuse at short-term (approximately 3 months), medium-term (approximately 6 months) and long-term (approximately 12 months) follow-up. We also extracted data on abstinence, reports of adverse events and any other secondary outcomes.

Data analysis
Categorical data on alcohol misuse were dichotomized into ‘improved’ or ‘not improved’. We counted as ‘improved’ outcome categories indicating clear, substantial improvement in alcohol misuse. Dichotomous and continuous outcome data were pooled using the generic inverse variance method with a random effects model. We calculated the effects of intervention results with estimates of pooled odd ratios (ORs) and 95% confidence intervals (CI) using Review Manager 5.0 (Nordic Cochrane Centre, Cochrane Collaboration). The percentage of outcome heterogeneity attributable to between-trial heterogeneity was assessed by the I2 statistic. Participants lost to follow-up were counted as not improved. In a post hoc analysis of trials with available dichotomized data, we calculated the pooled benefit difference on improvement in alcohol misuse at first follow-up and also calculated the number needed to treat. The benefit difference (also known as the risk difference) for each trial is the percentage of improved patients in the LSD group minus the percentage of improved patients in the control group. The number needed to treat is the inverse of the pooled benefit difference and provides an estimate of the average number of patients needed to be treated with LSD rather than without LSD to achieve one additional patient with improved outcome on alcohol misuse.