Combined pharmacological and psychosocial interventions for alcohol use disorder

Background Alcohol use disorder (AUD) is a mental disorder, characterised by a strong desire to consume alcohol, and impaired control on alcohol use, with devastating consequences. Many people with AUD do not respond to psychosocial and pharmacological interventions when provided alone. Combining these interventions may improve the response expanding targets or creating synergy, though evidence remains limited. Objectives To assess the effects of combined pharmacological and psychosocial interventions for the treatment of AUD in adults. Search methods We searched CENTRAL, MEDLINE, Embase, three other databases, and two trials registers in November 2023, without language restrictions. Selection criteria We included randomised controlled trials (RCTs) comparing combinations of pharmacological and psychosocial interventions with pharmacological or psychosocial interventions alone, or no intervention/treatment as usual (TAU) in adults with AUD. Our primary outcomes were continuous abstinent participants, frequency of use (measured as heavy drinkers, percentages of abstinent days, heavy drinking days), amount of use (number of drinks per drinking day), adverse events, serious adverse events, dropouts from treatment and dropouts due to adverse events. Data collection and analysis We assessed risk of bias with RoB 1, performed random-effects meta-analyses, and evaluated evidence certainty using GRADE. Main results We included 21 RCTs (4746 participants). The most studied pharmacological and psychosocial interventions were naltrexone (81.0%) and cognitive behavioural therapy (42.9%), respectively. Most participants were men (74%), aged about 44 years, with AUD, without comorbid mental disorders or other substance use disorders; 15 RCTs detoxified participants before treatment. We judged 28.5% of the studies at low risk of bias for random sequence generation, allocation concealment, performance bias for objective and subjective outcomes, and detection bias for subjective outcomes; all the studies at low risk of bias for detection bias for objective outcomes; 85.7% at low risk of attrition bias; 14.2% at low risk of reporting bias. 1) Compared to psychosocial intervention alone, adding pharmacotherapy probably (moderate-certainty evidence) reduces: a) heavy drinkers above the clinically meaningful threshold (MID) of 2%; absolute difference (AD) -10%, 95% confidence interval (CI) -18% to -2%; risk ratio (RR) 0.86, 95% CI 0.76 to 0.97; 8 studies, 1609 participants). It may (low-certainty evidence) increase: b) continuous abstinence (MID: 5%; AD 5%, 95% CI 1% to 11%; RR 1.17, 95% CI 1.02 to 1.34; 6 studies, 1184 participants). It probably (moderate-certainty evidence) has trivial or no effect on: c) abstinent days (MID: 8%; mean difference (MD) 4.16, 95% CI 1.24 to 7.08; 10 studies, 2227 participants), d) risk of serious adverse events (MID: 1%; AD -2%, 95% CI -3% to 0%; RR 0.20, 95% CI 0.03 to 1.12; 4 studies; 524 participants), e) participants who did not complete treatment (MID: 10%; AD -3%, 95% CI -5% to 0%; RR 0.89, 95% CI 0.79 to 1.01; 15 studies, 3021 participants), and f) those who did not complete due to adverse events (MID: 5%; AD 2%, 95% CI 0% to 5%; RR 1.91, 95% CI 1.04 to 3.52; 8 studies, 1572 participants). It may (low-certainty evidence) have trivial or no effect on: g) heavy drinking days (MID: 5%; MD -3.49, 95% CI -8.68 to 1.70; 4 studies, 470 participants), h) drinks per drinking day (MID: 1 drink; MD -0.57, 95% CI -1.16 to 0.01; 7 studies, 805 participants), and i) participants who reported adverse events (MID: 30%; AD 17%, 95% CI -5% to 46%; RR 1.25, 95% CI 0.93 to 1.68; 4 studies, 508 participants). 2) Compared to pharmacological intervention alone, we are uncertain (very low-certainty evidence) whether adding psychosocial interventions reduces: a) drinks per drinking day (MID: 1 drink; MD -2.40, 95% CI -3.98 to -0.82; 1 study, 241 participants). It may (low-certainty evidence) have trivial or no effect on: b) abstinent days (MID: 8%; MD -1.18, 95% CI -4.42 to 2.07; 2 studies, 1158 participants), and c) participants who did not complete treatment (MID: 10%; AD 1%, 95% CI -10 to 14%; RR 0.98, 95% CI 0.65 to 1.47; 3 studies, 1246 participants). We are uncertain (very low-certainty evidence) whether it has trivial or no effect on: d) continuous abstinent participants (MID: 5%; AD 3%, 95% CI -5% to 18%; RR 1.22, 95% CI 0.62 to 2.40; 1 study, 241 participants), e) heavy drinkers (MID: 2%; AD 2%, 95% CI -4% to 8%; RR 1.03, 95% CI 0.94 to 1.12; 1 study, 917 participants), and f) participants who did not complete treatment due to adverse events (MID: 5%; AD -1%, 95% CI -3% to 6%; RR 0.61, 95% CI 0.14 to 2.72; 2 studies, 1165 participants). 3) Compared to TAU, we are uncertain (very low-certainty evidence) whether combined interventions have trivial or no effect on: a) heavy drinkers (MID: 2%; AD -5%, 95% CI -13% to 2%; RR 0.93, 95% CI 0.83 to 1.03; 1 study, 616 participants), b) abstinent days (MID: 8%; MD 3.43, 95% CI -1.32 to 8.18; 1 study, 616 participants), c) participants who did not complete treatment (MID: 10%; AD 0%, 95% CI -10% to 15%; RR 0.98, 95% CI 0.58 to 1.65; 2 studies, 696 participants), and d) participants who did not complete due to adverse events (MID: 5%; AD 3%, 95% CI 0% to 15%; RR 2.97, 95% CI 0.70 to 12.67; 1 study, 616 participants). Certainty of evidence ranged from moderate to very low; certainty was downgraded mainly due to bias risk from the included studies and imprecise estimates. Authors' conclusions As implications for practice, our findings indicate that adding pharmacological to psychosocial interventions is safe and helps people with AUD recover. Conclusions are based on low- to moderate-certainty evidence. Due to few studies and very low-certainty evidence, benefits of adding psychosocial to pharmacological interventions or comparing combinations to TAU are less clear. As implications for research, further studies should investigate the effects of combinations compared to pharmacotherapy or TAU.