3.1 Retention, Adverse Events, Missing Data and Blind Assessment
Retention rates did not differ among medication groups (Kaplan-Meier χ2= 2.82, df=3, p=0.42; ), with 104 completers (64.8%). Average duration of participation ranged from 11.1–12.2 wks. Fifty-four participants did not complete the protocol for reasons specified in . Forty-nine instances of adverse events occurred, of which nine were deemed study-related (). All symptoms resolved upon decreasing methadone dose, terminating disulfiram administration, or without intervention in the case of the alcohol-disulfiram reactions. The overall percentage of missing urine samples out of the total possible (i.e., 39 urine samples) ranged from 8.8 (18.6) to 17.0 (27.4) in the placebo and 250 mg disulfiram groups, respectively, and did not differ significantly across groups (χ2=2.75, df=3, p=0.43). Groups did not differ with respect to the percentage of participants reporting whether they thought they were receiving placebo or disulfiram, with 43.9% and 56.1% of the overall sample believing they were receiving placebo and disulfiram, respectively (χ2=4.43, df=3, p=0.22). Similarly, research staff members at the UAMS site reported believing that 50% participants overall were on either placebo or disulfiram and no differences across groups occurred (χ2=3.60, df=3, p=0.31), although several cells had less than the expected cell count and so the accuracy of this statistical result is questionable.
Figure 2 Weekly percentage of participants retained in each of the four treatment groups across the 14-week trial: placebo (open circles), disulfiram at 62.5 mg/day (closed triangles), disulfiram at 125 mg/day (closed squares), and disulfiram at 250 mg/day (closed (more ...)
Summary of study-related adverse events.
3.2 Demographics and Baseline Measures
Of the 158 enrolled into the study proper, six participants did not participate beyond week 2 (). shows the demographic data for the 152 participants who actually received disulfiram or placebo. Groups generally did not differ in terms of subject characteristics, except for race; however, this result was problematic because there were fewer than expected participants in several cells, making the chi-square test unreliable.
Summary of group characteristics
The primary opioid of choice differed across sites, with heroin used significantly more often at VAHCS (F=423.2, df=1, p<0.001) and other opioids used more often at UAMS (F=190.2, df=1, p<0.001). Current alcohol dependence was diagnosed more often at VACHS (24.5% versus 3.5%; χ2=6.3, df=1, p=0.01). However, the proportion of participants enrolled at each site did not differ across treatment groups (see ).
3.3. Opioid Use Treatment Outcomes
The percentage of opioid positive urine screens did not differ across groups during either baseline (wk 2) or the disulfiram phase (wks 3–14) of the study (). Average self-reported heroin use or opioid pill use did not differ across groups at either baseline or during weeks 3–13 of the study. Neither methadone maintenance dose nor the number of missed doses differed across groups during the disulfiram phase of the study.
Summary of Selected Treatment Outcome Measures
3.4. Alcohol Use Treatment Outcomes
During week 2, the mean reported number of alcoholic beverages was less than 1 drink/day and did not differ across groups (). Overall, self-reported alcohol use continued to remain low and did not differ across groups during weeks 3–14. In addition, none of the disulfiram groups showed a change in alcohol use over time relative to the placebo group (p>0.2; data not shown).
3.5 Effects of Disulfiram Dose on Cocaine Use
Cocaine use during the trial, in terms of the percentage of urine samples positive for cocaine and the number of participants with urine samples negative and positive for cocaine, is shown in and , respectively. During week 2, cocaine positive urine screens did not differ across groups, with the percentage of urines positive for cocaine ranging from 57.7 (48) to 66.2(43) in the 62.5 and 125 mg disulfiram groups, respectively (χ2=0.49, df=3, p=0.92; see , left panel). During weeks 2–14, cocaine-positive urine screens increased over time in the 62.5 (t=3.74, df=4806, p=0.0002; slope=0.08) and 125 (t=5.45, df=4806, p<0.0001; slope=0.18) mg disulfiram groups relative to placebo (slope=−.05), but similarly decreased over time in the 250 mg disulfiram and placebo groups (t=−1.33, df=4806, p=0.18; slope=−0.10 and −0.05, respectively; ).
Figure 3 Weekly percentage of urine toxicology screens positive for cocaine (left panel) and weekly self-reported number of dimes of cocaine used (right panel) for the four treatment groups across the 14-week trial: placebo (open circles), disulfiram at 62.5 mg/day (more ...)
The number of participants with urine samples that either tested all negative (neg) or had at least one positive (pos) test, respectively, for cocaine/cocaine metabolite within a given week.
During the course of the trial, the number of participants with all cocaine-free urine samples within a given week essentially did not change over time in the placebo and disulfiram 250 groups, but decreased over time in the disulfiram 62.5 and 125 groups (). The number of participants with at least one cocaine-positive urine sample decreased similarly over time in the placebo, disulfiram 125 and disulfiram 250 groups, but essentially did not change or slightly increased in the disulfiram 62.5 group.
Overall, self-reported cocaine use did not differ across groups at baseline, with the number of dimes used per week ranging from 3.5 (4.9) to 6.6 (9.4) in the placebo and 125 mg disulfiram groups, respectively (χ2=1.52, df=3, p=0.67). Nevertheless, self-reported cocaine use increased over the course of the trial in the 125 mg disulfiram group (t=2.02, df=954.1, p=0.04; slope=0.013), but did not differ over time in the 62.5 (t=1.17, df=1068, p=0.24; slope=−0.009) and 250 (t=−1.28, df=1076, p=0.20; slope=−0.074) mg disulfiram groups relative to placebo (slope=−0.04; , right panel).