Table 2.
Summary of the evidence on pharmacotherapies in participants with comorbid stimulant and opioid use disorders
| Outcome | N studies per outcome | Summary of findings by outcome | Strength of Evidence* |
|---|---|---|---|
| Antidepressants (bupropion, desipramine, and fluoxetine) | |||
| Abstinence for ≥3 consecutive weeks | NA | NA | No evidence |
| Use | 1 RCT of bupropion (Poling et al., 2006) | No difference. One unclear-ROB RCT (N=106) reported no difference in use of cocaine. | Insufficient |
| Retention | 10 RCTs: 6 of desipramine (Arndt et al., 1992; Kolar et al., 1992; Kosten et al., 2003; Kosten et al., 1992; O’Brien et al., 1988; Oliveto et al., 1999); 2 of bupropion (Margolin et al., 1995b; Poling et al., 2006); 2 of fluoxetine (Grabowski et al., 1995; Winstanley et al., 2011) |
Favors placebo. A previous SR (Pani et al., 2011) found RR for dropout 1.22 (95% CI 1.05 to 1.41) combining 10 RCTs (N=1,006). | Moderate |
| Harms | 5 RCTs: 3 of desipramine (Arndt et al., 1992; Kolar et al., 1992; Kosten et al., 1992); 1 of bupropion (Margolin et al., 1995b); 1 of fluoxetine (Winstanley et al., 2011) |
Favors placebo. A previous SR (Pani et al., 2011) reported a combined RR of withdrawal due to an adverse event RR of 2.47 (95% CI 1.03 to 5.90; 5 RCTs, N=492). Severe AEs NR. |
Moderate |
| Anticonvulsants (gabapentin, tiagabine, and topiramate) | |||
| Abstinence for ≥3 consecutive weeks | 1 RCT of topiramate (Umbricht et al., 2014) | No difference. 1 low-ROB RCT (N=171) with 4 arms for topiramate vs placebo, +/− contingency management: longest duration of cocaine abstinence (mean weeks ± SE): 3.8 + 0.8 for TOP/CM, 3.7 ± 0.7 for TOP/Non-CM, 4.4 ± 0.7 for P/CM, for 3.5 ± 0.6 P/Non-CM. | Low |
| Use | 1 RCT of topiramate (Umbricht et al., 2014) | No difference. 1 low-ROB RCT (N=171) % of UA that were cocaine-negative: OR: 1.051, 95% CI: 0.6 to 1.84; p = 0.86, | Low |
| Retention | 3 RCTs of 4 medications: 2 of tiagabine (Gonzalez et al., 2007; Gonzalez et al., 2003); 1 of gabapentin (Gonzalez et al., 2007); 1 of topiramate (Umbricht et al., 2014) |
Favors placebo. Worse retention with anticonvulsants in 3 trials (N=292), pooled RR 0.86 (0.76 to 0.97). | Moderate |
| Harms | 3 RCTs (Gonzalez et al., 2007; Gonzalez et al., 2003; Umbricht et al., 2014) |
No difference. Severe AEs: None occurred in 2 RCTs. NR in 1 RCT. Dropouts due to AEs: None occurred in 2 RCTs; 6 vs 7 in 1 trial of topiramate. |
Low |
| Antipsychotics (aripiprazole and risperidone) | |||
| Abstinence for ≥3 consecutive weeks | NA | NA | No evidence |
| Use | 1 RCT of aripiprazole (Moran et al., 2017) | No difference. A high-ROB RCT of aripiprazole (N=18) found no differences in cocaine use. | Insufficient |
| Lapse and relapse | No difference. A high-ROB RCT of aripiprazole (N=18) found no differences in cocaine lapse or relapse. | Insufficient | |
| Insufficient | |||
| Retention | 1 RCT of risperidone (Grabowski et al., 2004) 1 RCT of aripiprazole (Moran et al., 2017) |
No difference. A high-ROB RCT of aripiprazole (N=18) and an unclear-ROB RCT of risperidone (N=96) found no difference in retention between groups. | Insufficient |
| Harms | 1 RCT of aripiprazole (Moran et al., 2017) |
No difference. A high-ROB RCT of aripiprazole found no difference in withdrawal due to AEs between groups. Severe AEs: NR |
Insufficient |
| Dopamine agonists | |||
| Abstinence for ≥3 consecutive weeks | NA | NA | No evidence |
| Use | 4 RCTs: 3 of amantadine (Handelsman et al., 1995; Kolar et al., 1992; Kosten et al., 1992) and 1 of bromocriptine (Handelsman et al., 1997) | No difference. There were no differences in cocaine use in 3 studies of amantadine (2 unclear-ROB and 1 high-ROB, or in 1 study of bromocriptine (unclear-ROB). | Low |
| Retention | 4 RCTs: 3 of amantadine (Handelsman et al., 1995; Kolar et al., 1992; Kosten et al., 1992) and 1 of bromocriptine (Handelsman et al., 1997) | No difference. Meta-analysis combining 3 studies of amantadine and 1 study of bromocriptine found no difference, without significant heterogeneity: RR 0.95, 95% CI 0.84 to 1.07 (N=205). | Low |
| Harms | 4 RCTs: 3 of amantadine (Handelsman et al., 1995; Kolar et al., 1992; Kosten et al., 1992) and 1 of bromocriptine (Handelsman et al., 1997) |
No difference. Withdrawal due to AEs: no difference from placebo in 1 study of bromocriptine and 2 studies of amantadine. Severe AEs: None occurred in 1 study of bromocriptine and 1 study of amantadine. |
Low |
| Medications for Opioid Use Disorder (methadone, buprenorphine, and buprenorphine combined with naloxone) | |||
| Abstinence for ≥3 consecutive weeks | 2 RCTs of methadone vs buprenorphine (Schottenfeld et al., 2005; Schottenfeld et al., 1997) | Methadone vs buprenorphine: Favors methadone. Continuous abstinence more frequent with methadone, combining 2 low-ROB RCTs (N=219): RR 1.85, 95% CI 1.25 to 2.75. | Low |
| 1 RCT of buprenorphine- naloxone vs placebo (Ling et al., 2016) |
Buprenorphine-naloxone vs placebo: No difference. 1 low-ROB RCT (N=302). No significant difference in abstinence (days 25–54). |
||
| Use | 1 RCT of buprenorphine- naloxone vs placebo (Ling et al., 2016) | Favors buprenorphine-naloxone: in a low-ROB RCT (N=302), more UDS(−) at higher dose (16–4 mg buprenorphine-naloxone) OR 1.71, P=0.022. No difference at lower dose (4–1 mg buprenorphine-naloxone) OR 1.09, P=0.105. | Insufficient |
| 1 RCT of naltrexone implant (Tiihonen et al., 2012) |
Favors naltrexone in 1 high-ROB RCT; % of drug free UDS (−): 38% vs 16%; P=0.01 Amphetamine: Week 10 UDS (−): 40% vs 24%; P=0.09 Heroin: Week 10 UDS (−): 52% vs 20%, P = 0.001 |
Insufficient | |
| Retention | 3 RCTs of methadone vs buprenorphine (Oliveto et al., 1999; Schottenfeld et al., 2005; Schottenfeld et al., 1997) |
Methadone vs buprenorphine: No difference. 2 low-ROB and 1 unclear-ROB RCTs (N=309), pooled RR 1.17 (95% CI 0.91 to 1.51). |
Low |
| 1 RCT of buprenorphine- naloxone vs placebo (Ling et al., 2016) |
Buprenorphine-naloxone vs placebo: No difference. 1 low-ROB RCT (N=302), high dose vs low dose vs placebo: 100% vs 98% vs 99%. |
Insufficient | |
| 1 RCT of naltrexone implant (Tiihonen et al., 2012) |
Favors naltrexone in 1 high-ROB RCT: 26/50 (52%) vs 14/50 (28%); RR=1.86 (1.11–3.12) |
Insufficient | |
| Harms | 2 RCTs (Oliveto et al., 1999; Schottenfeld et al., 2005) |
Methadone vs buprenorphine: No difference. Dropouts due to AE: NR Severe AEs: 1 |
Insufficient |
| 1 RCT of naltrexone implant (Tiihonen et al., 2012) | Naltrexone implant vs placebo: No difference. | Insufficient | |
| Medications FDA-approved for other substance use disorders (disulfiram, varenicline) | |||
| Abstinence for ≥3 consecutive weeks | 2 RCTs of disulfiram (George et al., 2000; Schottenfeld et al., 2014) | No difference. Pooled RR 0.89 (95% CI 0.53 to 1.48) based on 1 low-ROB RCT and 1 unclear- ROB RCT(N=207) | Low |
| Use | 1 RCT of varenicline (Poling et al., 2010) | No difference. 1 unclear-ROB RCT (N=31) found no difference in slope over time or between groups (Z = 0.20, p < 0.84). | Insufficient |
| Retention | 6 RCTs of disulfiram (Carroll et al., 2012; George et al., 2000; Kosten et al., 2013; Oliveto et al., 2011; Petrakis et al., 2000; Schottenfeld et al., 2014) | Favors placebo. Significantly worse treatment retention with disulfiram compared to placebo, 6 studies combined (N=605). RR 0.86, 95% CI 0.77 to 0.95. | Moderate |
| 1 RCT of varenicline (Poling et al., 2010) | No difference. 1 unclear-ROB RCT (N=31) found no difference (log rank χ2 = 1.3,p < 0.26) | ||
| Harms | 6 RCTs of disulfiram (Carroll et al., 2012; George et al., 2000; Kosten et al., 2013; Oliveto et al., 2011; Petrakis et al., 2000; Schottenfeld et al., 2014) 1 RCT of varenicline (Poling et al., 2010) |
No difference. | Low |
| Psychostimulants (dexamphetamine, mazindol, and methylphenidate) | |||
| Abstinence for ≥3 consecutive weeks | NA | NA | No evidence |
| Use | 3 RCTs: 2 of mazindol (Margolin et al., 1995a; Margolin et al., 1997) 1 of dexamphetamine (Grabowski et al., 2004) |
No difference. Cocaine-free UDSs occurred more frequently with psychostimulants compared with placebo in 3 Unclear/high ROB RCTs, but the difference did not reach statistical significance (standardized mean difference (SMD) 0.35, 95% CI −0.05 to 0.74; 3 RCTs, N=115). | Low |
| Retention | 4 RCTs: 2 of mazindol (Margolin et al., 1995a; Margolin et al., 1997); 1 of dexamphetamine (Grabowski et al., 2004); 1 of methylphenidate (Dursteler-MacFarland et al., 2013) |
No difference. (RR 0.98, 95% CI 0.71 to 1.36; N=210), although statistical heterogeneity was on the margin of significance (P=0.05; Figure 3). | Low |
| Harms | NA | NA | No evidence |
| Medications for amphetamine/methamphetamine use disorder | |||
| Abstinence for ≥3 consecutive weeks | NA | NA | Insufficient |
| Use | 1 RCT of naltrexone implant (Tiihonen et al., 2012) |
Favors naltrexone in 1 high-ROB RCT; % of drug free UDS (−): 38% vs 16%; P=0.01 Amphetamine: Week 10 UDS (−): 40% vs 24%; P=0.09 Heroin: Week 10 UDS (−): 52% vs 20%, P = 0.001 |
Insufficient |
| Retention | 1 RCT of naltrexone implant (Tiihonen et al., 2012) |
Favors naltrexone in 1 high-ROB RCT: 26/50 (52%) vs 14/50 (28%); RR=1.86 (1.11–3.12) |
Insufficient |
| Harms | 1 RCT of naltrexone implant (Tiihonen et al., 2012) | No difference. | Insufficient |
Abbreviations: AE = adverse event; N = number of; NA = not applicable; P = p-value; RCT = randomized controlled trial; ROB = risk of bias; UDS = urinary drug screen; RR = risk ratio; SMD = standard mean difference; SR = systematic review