Secondary Logo

Journal Logo

Clinical: Original Papers

Effectiveness of isoniazid chemoprophylaxis for HIV-infected drug users at high risk for active tuberculosis

Gourevitch, Marc N.ab; Hartel, Dianaa; Selwyn, Peter A.c; Schoenbaum, Ellie E.ab; Klein, Robert S.ab

Author Information
  • Free



High rates of active tuberculosis have been described in tuberculin-positive and anergic persons with HIV infection[1-3]. Clinical trials have uniformly suggested a benefit of isoniazid chemoprophylaxis among tuberculin-positive HIV-infected persons[4-8], and its use among such patients has received wide support[9,10]. The effectiveness of isoniazid in such patients outside of the clinical trial setting has been less well described[11-13].

Data from trials of isoniazid among HIV-infected anergic persons are less consistent. In a recent placebo-controlled study[14], isoniazid reduced the tuberculosis risk in HIV-infected anergic patients by more than half, but this reduction did not achieve statistical significance. Another trial[7], conducted in Uganda, suggested a short-term benefit that disappeared as follow-up time elapsed. As a result of these studies, and because anergy test results may vary over time[15], the recommendation [16] that isoniazid chemoprophylaxis be offered routinely to anergic HIV-infected individuals from communities with a high (>10%) prevalence of tuberculosis infection has recently been discarded[10,17]. However, the high risk of tuberculosis among anergic HIV-infected drug users in some regions has been demonstrated repeatedly[1,3,12]. It is therefore important not to overlook whether, in select populations of anergic HIV-infected persons, isoniazid chemoprophylaxis may still be justified.

Drug users are at particularly high risk of tuberculosis infection and disease, and high rates of active tuberculosis associated with HIV infection among tuberculin-positive and anergic drug users have previously been described[2,3,11-13]. As rates of active tuberculosis decline in New York City and elsewhere in the United States[18], resources previously devoted to active case finding and treatment are increasingly being devoted to prevention. We therefore sought to evaluate the effectiveness of providing chemoprophylaxis to drug users at high risk of active tuberculosis who were enrolled in a substance abuse treatment program.

Materials and methods

We studied HIV-seropositive drug users enrolled in a methadone maintenance treatment program with on-site primary care, located in the Bronx, New York City, USA. Enrollees are screened upon admission to the program and annually thereafter for tuberculosis infection. All persons not giving a history of tuberculin reactivity or of active tuberculosis undergo skin testing for tuberculosis infection with 5 tuberculin units of purified protein derivative (PPD) (Mantoux technique). Assessment for cutaneous anergy is also performed, with the Multitest CMI(cell-mediated immunity) device. Skin tests are read 48-72h after placement. HIV-seropositive individuals with 5mm or greater induration are classified as tuberculin reactors (PPD positive). Anergy is defined as the absence of reaction to any of the seven Multitestantigens of 2mm or greater in excess of that produced by the control antigen (glycerin) of the testing device, and by less than a 2mm response to the PPD test. Methadone program enrollees participating in an on-site study of HIV infection in drug users are tested for HIV as part of the study protocol[19]. HIV-infected study enrollees undergo CD4 T lymphoctye subset determination prospectively at 6 month intervals. Patients not enrolled in that study are offered confidential on-site HIV testing. Serum is analysed for HIV infection by enzyme immunoassay, with confirmation by Western blot. Only HIV-seropositive participants are included in the current analysis.

Patients were considered eligible for isoniazid chemoprophylaxis if they: (a) gave a history of tuberculin reactivity but not of having received chemoprophylaxis or treatment of active tuberculosis; (b) were tuberculin reactors when tested in the program; or (c) were anergic. Eligible patients were offered the option of receiving their isoniazid under direct observation together with their daily methadone: nearly all patients agreeing to chemoprophylaxis accepted this form of administration. Patients were prescribed isoniazid (300mg daily) with pyridoxine (50mg daily) for 12 months. On days on which patients were not scheduled to receive methadone at the program (Sundays for all patients, additional days depending on patients‚ pick-up schedules), they were given pre-packaged unit doses of isoniazid and pyridoxine to take home. Seventy-four per cent of program patients came to the clinic at least 5 days per week for their methadone and, by extension, their isoniazid. Dates of chemoprophylaxis were taken from computerized records of isoniazid administration at the methadone dispensing window. Previous analyses demonstrated high rates of adherence (88%) with chemophrophylaxis administered to this patient population in this fashion[20]. Skin test data and HIV serological data were recorded prospectively. All patients with complete skin test and HIV status data from 1985 to 1996 were included.

Active surveillance for tuberculosis was conducted throughout the study period. The methadone program is notified of all hospitalizations of its patients, as formal verification of patients‚ methadone doses is mandatory before inpatient administration of methadone. The hospital charts of reported cases of tuberculosis were reviewed, and confirmed cases entered into a tuberculosis cases database. Cases of tuberculosis reported by patients during research interviews and cases diagnosed at the on-site primary care center were also included in this database. Since 1992, regular cross-matches between the methadone program roster and the New York City Department of Health Tuberculosis Registry further enhanced case ascertainment. All cases of tuberculosis included in this analysis had been confirmed by positive culture for Mycobacterium tuberculosis (except for a single empirically diagnosed case, indicated as such in the Results section).

The incidence of tuberculosis was determined according to skin test status and the extent of chemoprophylaxis. Persons without valid dates or results for HIV, tuberculin or anergy tests were excluded from the analysis. Patients were classified as receiving no chemoprophylaxis, or incomplete (1-11 months) or complete (≥⊇12 months) chemoprophylaxis. Reasons for receiving no or incomplete chemoprophylaxis were determined by chart review. The observation period in which the incidence of active tuberculosis was calculated began with the date of eligibility for isoniazid prophylaxis (the date on which tuberculin-reactor or anergic status was determined or past reactor status declared) and ended with the development of active tuberculosis, the last observed research or clinic contact, or 31 December 1996, whichever was earliest. Rates are expressed as the number of active cases divided by the sum of the person-years of observation. As CD4 cell count determinations were not timed specifically to coincide with the initiation of chemoprophylaxis, the CD4 cell count determination closest in time to the date of the skin test used to determine chemoprophylaxis eligibility was used (median difference 6 months).

Rate ratios and confidence intervals were calculated for all groups except those without observed cases. Mid-P binomial intervals and exact statistical tests were computed for rate ratio data. Otherwise, categorical data were examined using chi-square tests. t-Tests were performed for data with a normal distribution (e.g. age). Non-parametric tests (Wilcoxon) were performed for CD4 cell data. All statistical associations were assessed with the use of two-tailed tests with an alpha error of 0.05 for statistical significance.


Of 1710 persons with confirmed HIV status, 593 (35%) were HIV-seropositive and 1117 (65%) were HIV-seronegative. Of the HIV-seropositive patients, a total of 431 (73%) were potentially eligible for chemoprophylaxis during the study period: 202 (47%) tuberculin reactors and 229 (53%) anergic individuals. Of these eligible individuals, 155 (36%) patients commenced chemoprophylaxis (46% of tuberculin reactors and 28% of anergic individuals). Among isoniazid recipients, 79 (51%) received 1-11 months of treatment (non- completers) and 76 (49%) completed 12 months or more (completers). Among the 79 non-completers, 39 (49%) persons discontinued their isoniazid during the course of the study (13% of tuberculin reactors and 43% of anergic individuals). The most common reasons for discontinuation were patient reports of side-effects and documented hepatotoxicity. An additional 40 persons were lost to follow-up during the course of their isoniazid therapy or were still on isoniazid at the close of the observation period (31 December 1996). Two hundred and seventy six (64%) eligible patients never commenced chemoprophylaxis (non-recipients). The most common reasons for not initiating chemoprophylaxis included patient refusal and physicians‚ concerns about hepatotoxicity among patients with hepatic dysfunction. Selected characteristics of chemoprophylaxis recipients are given in Table 1.

Table 1
Table 1:
Selected characteristics of study participants, by eligibility for and receipt of chemoprophylaxis during the study perioda

Twenty-five cases of tuberculosis, including five previously reported[3], developed among the 431 persons eligible for chemoprophylaxis (5.8%) over 1748 person-years of observation, for an overall case rate of 1.43 per 100 person-years. The completion of 12 months of isoniazid was associated with a reduction in tuberculosis risk among both tuberculin reactors and anergic participants (Table 2). Among tuberculin reactors, the incidence of tuberculosis declined from 2.07 and 3.31/100 person-years among those taking no isoniazid or receiving less than the full 12 month course, respectively, to 0.51/100 person-years among those completing the full 12 months (P=0.05 for completers versus non-users of isoniazid). Grouping together tuberculin reactors receiving an incomplete course or no isoniazid and comparing this group with completers, the receipt of 12 months or more of isoniazid was associated with a reduction in the incidence of tuberculosis among tuberculin reactors from 2.39 to 0.51/100 person-years [rate ratio 0.21, 95% confidence interval (CI) 0.03-0.82]. Among anergic individuals, no cases of tuberculosis developed among completers, compared with 0.77 and 1.44 cases/100 person-years among those receiving less than the full 12 month course or taking no isoniazid, respectively. Grouping together anergic individuals receiving an incomplete course or no isoniazid and comparing this group with anergic completers, the receipt of 12 months or more of isoniazid was associated with a reduction in the incidence of tuberculosis among anergic individuals from 1.31 to 0/100 person-years (rate ratio undefined). These findings were essentially unchanged when the single empirically diagnosed case of tuberculosis was excluded from the analysis.

Table 2
Table 2:
Rates of development of active tuberculosis among HIV-positive drug users, by skin test and isoniazid chemoprophylaxis status

We determined the rate of tuberculosis among HIV-infected tuberculin-negative persons who were not anergic. Three cases developed among 173 such persons over 360 person-years of observation, for a rate of 0.83/100 person-years (Table 2).

To assess whether the observed effect of chemoprophylaxis on tuberculosis rates was caused by preferential chemoprophylaxis of less immunosuppressed participants, we analysed the CD4 T-lymphocyte numbers among those receiving isoniazid. As expected, the median CD4 cell number was lower among anergic patients than among tuberculin reactors (301 versus 382/mm3, P=0.003, Wilcoxon test). Persons receiving isoniazid tended to have lower CD4 cell counts overall than those not receiving isoniazid (Table 2). The higher rates of tuberculosis we observed among persons receiving no isoniazid cannot thus be attributed to more advanced immunosuppression in this group.


Although the efficacy of isoniazid chemoprophylaxis to prevent active tuberculosis has been demonstrated in a variety of settings, its real effectiveness among persons at highest risk of developing active disease is less well characterized. We present data on clinical outcomes of drug users at risk of tuberculosis, demonstrating that a complete course of chemoprophylaxis is associated with a marked reduction in the incidence of tuberculosis among tuberculin-positive and anergic drug users with HIV infection. These findings underscore the benefits drug users can derive when given ready access to medical services and supervised therapy.

The effectiveness of prescribing isoniazid to anergic HIV-infected persons has recently been called into question by data from controlled clinical trials. In one, isoniazid was found to reduce the rate of tuberculosis by half (relative risk 0.48), an effect of similar magnitude to the one we report[14]. However, because of the low number of cases observed, the confidence intervals associated with this reduction did not exclude one, and the authors concluded that isoniazid lacked effectiveness in this setting. In another study[7], conducted in Africa, a high rate of active tuberculosis among the isoniazid arm (2.5/100 person-years) was not substantially lower than the rate among those not receiving isoniazid (3.1/100 person-years). These findings, in the context of reports that the state of cutaneous anergy is not an all-or-nothing phenomenon but rather waxes and wanes in some persons, recently led the Centers for Disease Control and Prevention to step back from its earlier recommendation that chemoprophylaxis be strongly considered for anergic HIV-infected persons from communities with a substantial (>10%) prevalence of tuberculosis infection[17,21]. However, our data support the effectiveness of chemoprophylaxis of HIV-infected anergic persons at high risk of developing active disease.

Three conditions may account for our findings of the effectiveness of chemoprophylaxis among anergic individuals. The relatively high incidence of tuberculosis among anergic persons (1.44/100 person-years), and the long period over which we conducted these observations, strengthened our ability to discern the effect of this intervention. In addition, we provided chemoprophylaxis under direct observation, thus optimizing the effectiveness of this therapy.

Testing for cutaneous anergy has been criticized because the test results are not always reproducible in a given individual over time[15,22]. Persons testing anergic may thus later manifest delayed-type hypersensitivity responsiveness. Persons who are anergic on consecutive determinations are less likely to revert to reactive status[15]. We selected a single determination of cutaneous anergy as defining the anergic state. Despite this relatively loose definition, we observed a treatment effect of chemoprophylaxis. This result suggests that despite the instability of the test, the risk of developing active tuberculosis among HIV-infected persons demonstrating cutaneous anergy on even a single occasion is reduced by isoniazid chemoprophylaxis. To further explore this point, we compared the rate of tuberculosis among HIV-infected tuberculin-negative persons who were and were not anergic (Table 2). In the anergic group, only those who did not receive isoniazid were included. Tuberculosis rates were lower in the non-anergic than in the anergic group (0.83 versus 1.44/100 person-years, respectively), although because of the available sample size the confidence intervals are wide (rate ratio 0.58, 95% CI 0.12-2.12). Anergy testing thus appears to confer additional information about tuberculosis risk beyond that conveyed by a negative tuberculin test, consistent with findings in other populations[15].

The chemoprophylaxis regimen used in our program by HIV-infected persons consists of daily isoniazid (with pyridoxine) for 1 year. Recent data suggest that far shorter courses of chemoprophylaxis (2 months of rifampin with pyrazinamide, daily or twice weekly) are effective among HIV-infected tuberculin reactors[8]. The potential of offering HIV-infected anergic patients a 16-dose course of prophylactic therapy, if proved to be effective, could render treating such patients still more attractive by removing the obstacle of patient and provider reluctance to undertake a year of daily therapy to prevent a complication that may never arise.

Rates of active tuberculosis among HIV-infected drug users in the Bronx, who have not received chemoprophylaxis with isoniazid despite being tuberculin reactors or anergic, have fallen over the past decade. The rates we now report among tuberculin reactors (2.07/100 person-years) and anergic individuals (1.44/100 person-years) are substantially lower than those we previously reported in this study population (7.9/100 person-years in tuberculin reactors [2] and 6.6/100 person-years in anergics [3]). This decline probably reflects the decrease in recent years in the incidence of active tuberculosis in New York City, resulting in less transmission and consequently a lower incidence of active disease after relatively recently acquired infection. It is also possible that the more potent antiretroviral therapy used in recent years has played a role in suppressing the reactivation of latent tuberculosis infection.

Two cases of active tuberculosis developed in HIV-infected tuberculin reactors who had received a full 12 month course of isoniazid chemoprophylaxis. Both patients were immunosuppressed (most recent CD4 cell counts before tuberculosis were 54 and 212 /mm3), and active disease developed 20 and 28 months after the completion of isoniazid. Although we were unable to determine whether these cases represented a failure of chemoprophylaxis or exogenous reinfection, low rates of disease after isoniazid are consistent with the findings of clinical trials of preventative therapy among co-infected persons[8].

The results presented here are limited by the observational nature of the study design, and the modest size and follow-up time of the sample. The ingestion of take-home doses of isoniazid were not directly observed, but the possibility that some such doses were missed would have made it more difficult for us to observe an effect of isoniazid on the incidence of tuberculosis. The higher rate of tuberculosis seen among non-completers was not statistically significant and may reflect a greater risk of active disease among persons who for various reasons were unable to tolerate a full course of chemoprophylaxis. Finally, we have an insufficient number of tuberculosis cases in each of the skin test categories (tuberculin reactors and anergic individuals) to be able to conduct a meaningful analysis of the possible differential case distribution when stratified by periods of varying intensity of local tuberculosis transmission.

Our data support aggressive efforts to provide a complete course of preventative therapy to HIV-infected tuberculin reactors, and lend weight to the findings of others that isoniazid can reduce the rate of tuberculosis in high-risk anergic HIV-infected persons by approximately 50%.


The authors wish to thank Randy Teeter and the staff and participants of the Tuberculosis Study for the time and effort they devoted to this investigation, and Bill Wasserman, MPH, for his assistance in collecting and preparing the data.


1. Markowitz N, Hansen NI, Hopewell PC, et al. Incidence of tuberculosis in the United States among HIV-infected persons.Ann Intern Med 1997, 126:123-132.
2. Selwyn PA, Hartel D, Lewis VA, et al. A prospective study of the risk of tuberculosis among intravenous drug users with human immunodeficiency virus infection.N Engl J Med 1989, 320:545-550.
3. Selwyn PA, Sckell BM, Alcabes P, Friedland GH, Klein RS, Schoenbaum EE. High risk of active tuberculosis in HIV-infected drug users with cutaneous anergy.JAMA 1992, 268:504-509.
4. Pape JW, Jean SS, Ho JL, Hafner A, Johnson WD. Effect of isoniazid prophylaxis on incidence of active tuberculosis and progression of HIV infection.Lancet 1993, 342:268-272.
5. Hawken MP, Meme HK, Elliott LC, et al. Isoniazid preventive therapy for tuberculosis in HIV-1-infected adults: results of a randomized controlled trial.AIDS 1997, 11:875-882.
6. Wadhawan D, Hira S, Mwansa N, Sunkutu R, Adera P, Perine P. Preventive tuberculosis chemotherapy with isoniazid (INH) among patients infected with HIV-1. IXth International Conference on AIDS/IV STD World Congress. Berlin, June 1993 [Abstract PO-B07-1114].
7. Whalen CC, Johnson JL, Okwera A, et al. A trial of three regimens to prevent tuberculosis in Ugandan adults infected with the human immunodeficiency virus.N Engl J Med 1997, 337:801-808.
8. Halsey NA, Coberly JS, Desormeaux J, et al. Randomized trial of isoniazid versus rifampicin and pyrazinamide for prevention of tuberculosis in HIV-1 infection.Lancet 1998, 351:786-792.
9. American Thoracic Society. Treatment of tuberculosis and tuberculosis infection in adults and children. Am J Respir Crit Care Med 1994, 149:1359-1374.
10. Centers for Disease Control and Prevention. 1997 USPHS/IDSA guidelines for the prevention of opportunistic infections in persons infected with human immunodeficiency virus.MMWR 1997, 46(No. RR-12):10-12.
11. Moreno S, Miralles P, Diaz MD, et al. Isoniazid preventive therapy in human immunodeficiency virus-infected persons. Arch Intern Med 1997, 157:1729-1734.
12. Guelar A, Gatell JM, Verdejo J, et al. A prospective study of the risk of tuberculosis among HIV-infected patients.AIDS 1993, 7:1345-1349.
13. Jansà JM, Serrano J, Caylà JA, Vidal R, Ocaña I, Español T. Influence of the human immunodeficiency virus in the incidence of tuberculosis in a cohort of intravenous drug users: effectiveness of anti-tuberculosis chemoprophylaxis. Int J Tuberc Lung Dis 1998, 2:140-146.
14. Gordin FM, Matts JP, Miller C, et al. A controlled trial of isoniazid in persons with anergy and human immunodeficiency virus infection who are at high risk for tuberculosis.N Engl J Med 1997, 337:315-320.
15. Caiaffa WT, Graham NMH, Galai N, Rizzo RT, Nelson KE, Vlahov D. Instability of delayed-type hypersensitivity skin test anergy in human immunodeficiency virus infection.Arch Intern Med 1995, 155:2111-2117.
16. Centers for Disease Control. Purified protein derivative (PPD)-tuberculin anergy and HIV infection: guidelines for anergy testing and management of anergic persons at risk of tuberculosis.MMWR 1991, 40 (No. RR-5):27-33.
17. Centers for Disease Control and Prevention. Anergy skin testing and preventive therapy for HIV-infected persons: revised recommendations.MMWR 1997, 46 (No. RR-15):1-10.
18. Centers for Disease and Prevention. Tuberculosis morbidity - United States, 1996.MMWR 1997, 46:695-700.
19. Schoenbaum EE, Hartel D, Selwyn PA, et al. Risk factors for human immunodeficiency virus infection in intravenous drug users.N Engl J Med 1989, 321:874-879.
20. Gourevitch MN, Wasserman W, Panero MS, Selwyn PA. Successful adherence to observed prophylaxis and treatment of tuberculosis among drug users in a methadone program.J Addict Dis 1996,15:93-104.
21. Centers for Disease Control and Prevention. Prevention and treatment of tuberculosis among patients infected with human immunodeficiency virus: principles of therapy and revised recommendations.MMWR 1998, 47:(No. RR-20):36-43.
22. Chin DP, Osmond D, Page-Shafer K, et al. Reliability of anergy skin testing in persons with HIV infection.Am J Respir Crit Care Med 1996, 153:1982-1984.

Anergy; chemoprophylaxis; drug users; HIV; isoniazid; purified protein derivative; tuberculosis

© 1999 Lippincott Williams & Wilkins, Inc.