Routine care

Patients are evaluated for TB when they present with symptoms or through targeted fluoroscopy screening. Diagnosis is confirmed using radiographic, clinical and/or bacteriologic data. All active cases undergo screening for AUDs using the Alcohol Use Disorders Identification Test (AUDIT) (13); HIV screening using ELISA (enzyme-linked immunosorbant assay), and drug susceptibility testing of their infecting M. tuberculosis strain. Individuals are evaluated for alcohol detoxification prior to initiation of TB treatment. Individuals with socio-economic or medical problems that could complicate therapy (including most individuals with AUDs) are admitted to the hospital to start treatment; patients are usually transitioned to ambulatory care once stable. TB treatment is provided free of charge and under direct observation; regimens and outcome criteria follow World Health Organization (WHO) guidelines (14). Alcohol services are provided by the Tomsk Narcology Services, requiring physician or self-referral and payment by the patient; however, a full-time addiction specialist (narcologist) employed at the TB hospital provides free services to inpatients. Standard of care for AUDs includes psychotherapy, disulfiram, and placebo implants; naltrexone is not readily available although it is approved for use in Russia for both alcohol and narcotic dependence (15).

Study protocol

The study design has been described elsewhere (16). Adults starting TB treatment were consecutively referred to the study and evaluated for alcohol abuse or dependence based on Diagnostic and Statistical Manual Fourth Edition (DSM-IV (17)) as measured by the Composite International Diagnostic Interview-Substance Abuse Module (CIDI-SAM (18))(19). Inclusion criteria were as follows:

1. TB Diagnosis and registered for TB therapy with the TOTBS;

2. TB treatment initiation in one of the three study sites (inpatient hospital, polyclinic, day hospital); and

3. Diagnosis with alcohol abuse or dependence based on the CIDI-SAM).

Individuals were excluded if they:

1. Were under 18 years of age;

2. Had liver function tests more than three times upper limit of normal range. The participant could be retested after five days and enrolled if repeat liver function tests were less than three times upper limit of normal range;

3. Reported opioid use in the past month or a positive urine screen for opioids. The participant could be retested after five days and enrolled if the second urine screen was negative;

4. Was pregnant or breastfeeding;

5. Had inadequate understanding of the study after undergoing informed consent;

6. Had any co-occurring other medical or psychiatric condition that would make it impossible for them to comply with the study procedures;

Abstinence prior to enrollment was not required. Those who consented to participate in the study were stratified by assigned TB provider and block-randomized to receive one of four possible non-blinded interventions:

1. Brief Counseling Intervention (BCI) plus treatment as usual (i.e. standard referral to and management by a narcologist);

2. Naltrexone and Brief Behavioral Compliance Enhancement Treatment (BBCET) plus treatment as usual;

3. BCI plus naltrexone-BBCET plus treatment as usual;

4. Treatment as usual alone.

Non-blinded randomization assignment was generated by computer by the study team.

Study interventions

The NIH/NIAAA’s Helping Patients with Alcohol: A Health Practitioner’s Guide (20) was adapted for our study population, described elsewhere (21). BCI consisted of six 10–15 minute discussions delivered monthly by TB physicians embedded in their standard 45–60 minute TB appointments (22–25). Naltrexone was orally administered to patients as a daily single dose of 50 mg for six months, starting at the time of study enrollment, usually within two weeks of TB treatment start. TB staff administered naltrexone together with anti-tuberculosis medications under direct observation and recorded all doses on a naltrexone administration card. Monthly assessment included a structured interview for adverse events, liver and kidney function tests, and pregnancy or opioid test if indicated. TB physicians managed naltrexone side effects as well as treatment interruptions due to detoxification, opioid use (including surgery), and pregnancy.

All participants randomized to naltrexone also received brief adherence counseling adopted from the Brief Behavioral Compliance Enhancement Treatment (BBCET) used in the Topiramate Clinical Protocol (26), with permission from the authors. The BBCET was a focused intervention to enhance naltrexone adherence and promote changes in drinking behavior through goal-setting and naltrexone-specific feedback. BBCET differs from BCI in that it does not use counseling or motivational interviewing. Six 5–10 minute BBCET encounters were delivered monthly by TB physicians during standard follow-up appointments.

Due to the factorial study design, some participants were randomized to receive both BCI and naltrexone-BBCET. We created a BCI/BBCET manual which combined both behavioral interventions in a streamlined fashion, with each encounter lasting approximately 15–25 minutes.

Training and fidelity assessment

TB providers received training on naltrexone, including the mechanism of action, pharmacology, administration, dosing, side effect management, and contraindications. Training and fidelity procedures for behavioral interventions have been described elsewhere (21). Fidelity assessment was conducted throughout the study, via review of audio-taped sessions by our study team with standardized coding for fidelity and adherence. Scoring sheets provided quantitative and qualitative feedback to providers. 98% of BCI encounters reviewed met criteria for procedural adherence; for BBCET, 87% of encounters adhered to protocols, while 13% included counseling such as discussing events or goals not related to adherence or engaging in problem solving.

Data collection and analysis

Baseline and follow-up data were collected by trained TB providers, including chart review for sociodemographic information and co-morbidities. Baseline depression was measured using the Center for Epidemiologic Studies Depression Scale (CES-D) (27). Patients were interviewed about alcohol use with the Time Line Follow Back (TLFB) questionnaire with three-month recall at baseline, three and six months (28). We obtained TB bacteriologic, adherence and outcome data from the TOTBS registry.

We obtained approval from the Partners Healthcare IRB as well as the State Research Center Virology and Biotechnology, Novosibirsk Region and the Siberian State Medical University of the Federal Agency for Health care and Social Development, local IRB in Tomsk, Russia. The trial was registered with clinicaltrials.gov, registration number NCT00675961.

Study outcomes

We assessed both TB- and alcohol-related outcomes. The primary TB endpoint was “favorable outcome,” i.e. cure, completed treatment, while “poor outcomes” included treatment failure, death from any cause, or default. For those without a final outcome (including transfers and those still in treatment), individuals who achieved culture conversion (i.e. at least two consecutive monthly negative cultures) were considered to have a favorable outcome. TB treatment adherence defined as the percent of doses taken as prescribed under direct observation was also considered.

Our primary alcohol outcome was defined as the change in mean number of abstinent days in the last month of study period compared with pre-hospitalization baseline. The secondary alcohol endpoint was: the change in mean number of heavy drinking days (4 drinks per drinking day for women and 5 drinks per drinking day for men) (29) in the last month of the study compared with pre-hospitalization baseline. Additionally, because this was a non-treatment seeking group with long-standing AUDs, we hypothesized that the individuals with prior attempts to quit drinking may be more likely to respond to study interventions. We therefore performed a secondary analysis limited to individuals who reported any prior attempt to quit drinking alcohol.

Baseline characteristics

Baseline characteristics of the study cohort are presented in Table 1. The mean age was 41 years, and 82% were male. Participants often presented with severe tuberculosis, reflected by a low body mass index in 29.1% and cavitary disease in 84.7% of participants. Smoking was almost ubiquitous (92.9%), while only 14 (7.1%) had a concomitant drug use disorder. Most individuals (63.0%) had a diagnosis of alcohol dependence, and consumption was heavy: 27.6% reported nearly daily drinking the past 12 months, with median of 16 standard drinks per drinking day. Depression, assessed only among study participants, was observed in 13.5% of the cohort.

Compared with study participants (Table 1), eligible candidates who declined were less likely to be married or living with a partner (53.8% v. 37.8%, P=0.01), and were less likely to be unemployed (63.8% versus 78.1%, P=0.01). In addition, those who declined tended to have less severe AUDs: they were less likely to have a diagnosis of alcohol dependence (39.7% versus 63.0%, P=<0.001), reported less daily drinking in the past year (18.0% versus 27.6%, P=<0.001) and fewer number of standard drinks (median 8.5 versus 16.0, P=0.03), and had lower average baseline AUDIT scores (16.0 versus 18.3, P=0.03) compared with study participants.

Baseline characteristics displayed in Table 1 did not differ significantly among individuals randomized to naltrexone versus no naltrexone. Individuals assigned to BCI versus no BCI were more likely to have a diagnosis of alcohol dependence (70.1% versus 55.8%, P=0.04), but did not differ significantly in other respects.

Intervention delivery

Participants randomized to naltrexone-BBCET received a median [first and third quartiles, Q1, Q3] of three [1, 5] BBCET encounters and 112 [39, 159] doses of naltrexone (74.7% of the total planned 150 doses). Five participants received no BBCET and an additional eight individuals received neither BBCET nor naltrexone. Those randomized to BCI received a median [Q1, Q3] of three BCI encounters [2, 5], and thirteen received no BCI.

Adverse events

Monthly assessment for adverse events yielded no significant difference in specific and overall (combined) adverse events among individuals assigned to naltrexone versus those not assigned to naltrexone. Likewise, the frequency of adverse events among individuals assigned to BCI versus no BCI did not significantly differ.

This study was supported in part by grants R01 AA016318 from the National Institute of Health (SS) and K24 DA 019855 (SFG) from the National Institute on Drug Abuse. We would also like to thank Sid Atwood Oksana Ponomarenko, Alex Golubkov, Yekaterina Pushkariova, Natalia Morozova-Arliapova, Eugeniy Kotsin, Dmitriy Taran, Natalia Zemlianaya, Aivar Strelis, Sergey Mishustin, Yekaterina Stepanova, and the personnel of the 1^st TB Department in the Tomsk Oblast TB Hospital for their meaningful contributions throughout the study.