Introduction
The issue of whether infection with HIV is a risk factor for drug-resistant tuberculosis (TB) has been argued for several years. The question surfaced after the Centers for Disease Control and Prevention (CDC) reported outbreaks of drug-resistant disease amongst HIV-infected people in New York City and Miami [1,2]. Although numerous studies have now been published, the question has not been resolved. Some investigators [3-6] have associated drug-resistant TB with HIV infection. Proposed mechanisms include (i) higher infection rates and more rapid progression of resistant organisms in the immunocompromised host, and (ii) acquired resistance that results from co-infection with HIV and TB. Other investigators have found no association between drug-resistant TB and HIV infection [7,8]. Bloch et al. [9] noted that the epidemiology of drug-resistant TB could not be determined from data on local outbreaks. Bloch et al. [9] therefore evaluated data from the CDC and reported the first survey on drug-resistant TB from a national perspective. At-risk groups were identified, but no data were available at that time on HIV status.
Recently, two additional studies have reported conflicting results. Asch et al. [10] concluded that, in an area with no ongoing outbreak of resistant disease, drug-resistant TB was not more common in HIV-infected patients. However, Gordin et al. [11] found that HIV infection was associated with drug-resistant TB, irrespective of geographic location, prior therapy, age and race.
The studies reported to date are discordant on the association between HIV infection and drug-resistant TB; the question of whether HIV co-infection places a person at increased risk for drug resistance remains largely unanswered. Furthermore, no data have yet been published on the frequency of drug-resistant TB in HIV-serodefined populations in a setting employing directly observed therapy. Resolution of this issue is necessary because it affects considerations of public health policy. If HIV infection is a risk factor for drug resistance and is not dependent on other considerations, then treatment regimens and strategies for these patients must be revised.
This study describes an 8-year study of drug-resistant TB in persons with defined HIV serostatus. Outpatient therapy was based on a policy of universal directly observed therapy.
Materials and methods
All culture-positive TB patients who were reported to the Tarrant County Public Health Department, Fort Worth, Texas between 1 January 1988 and 31 December 1995 were eligible for inclusion in the study. Patients were compared on the basis of HIV status and drug-resistant disease. Patients not tested for HIV infection or without drug susceptibilities were excluded from the comparisons.
The majority of cultures were analyzed by the state laboratory at Austin, Texas. Isolates were characterized for resistance to isoniazid, rifampin, rifabutin, ethambutol, streptomycin, capreomycin, kanamycin, and ethionamide. Testing for pyrazinamide susceptibility was performed if the patient was resistant to isoniazid, rifampin or ethambutol. Initial resistance was determined from the first available susceptibility test. New drug resistance identified on subsequent susceptibility studies was reported as acquired resistance. A drug-resistant case was defined as a culture-positive case with either initial or acquired resistance. Culture specimens were collected at the beginning of treatment and periodically during therapy.
The patient population was characterized according to age, sex, race or ethnicity, residence, foreign-born status, and history of drug use (by any route). Epidemiological data were collected at the first patient interview by a single examiner (S.E.W.); when patients were unavailable for interview, the data were obtained from family interviews and medical records.
Categorical data were compared for statistical significance using a χ2 contingency analysis or Fisher's exact test. In all 2 × 2 contingency analyses, a continuity correction was applied. Statistical significance was based on a two-tailed test with α equal to 0.05. Reported percentages may not total 100% due to rounding.
Results
During the reporting period, 802 culture-positive cases of TB were reported. HIV serology and drug sensitivity data were available for 741 individuals (92.4%). Of these individuals, 646 were HIV-negative, and 95 (12.8%) were HIV-positive. Fifty-nine patients (7.4%) did not have data on HIV serology and two patients (0.2%) did not have susceptibility data. The primary reasons why HIV testing was not performed included patient refusal (39%) and cases reported after death (54%). Excluding patients moving to another county or dying during treatment, 96% of study participants completed therapy.
The characteristics of the patient population are presented in Table 1, which shows the total number of TB cases fitting each population variable with respect to HIV serostatus. Comparisons reaching statistical significance indicate that HIV coinfection was more common among drug users (both injecting and non-injecting), men, middle-aged adults, blacks (non-Hispanic), persons born in the United States and persons living in community-based facilities. History of chronic alcohol use, prior diagnosis of TB and drug resistance did not differ significantly between the two groups. The salient observation on these characteristics was the comparable incidence of initial and acquired drug resistance in the HIV-positive and negative groups. Of the 646 HIV-negative patients, 9% were drug-resistant, compared with 4% of the 95 HIV-positive patients. This difference was not statistically significant.
Analysis of drug-resistant TB cases by risk variable is shown in Table 2. Factors positively associated with resistance included race/ethnicity (Asian), foreign birth, and a prior diagnosis of TB. There was a negative association between drug resistance and illicit drug use. This unusual finding reflects the fact that 59% of resistant cases occurred in foreign-born groups where drug use was uncommon. Sex, age, chronic alcohol use and place of residence were not associated with drug resistance. Most importantly, a positive HIV status was not associated with drug resistance.
Initial drug-resistance patterns by HIV status are shown in Table 3. Single drug resistance occurred in 42 (71%) of the 59 cases. Streptomycin/isoniazid resistance accounted for 48 cases (81%) of initial resistance and was primarily found in foreign-born Asians (35%) and Hispanics (21%). Foreign-born Asians and Hispanics each comprised 12% of the study population. Two episodes of initial resistance to isoniazid and rifampin occurred, and both patients involved were HIV-negative. The HIV-positive group included two cases of streptomycin resistance, one of isoniazid resistance, and one of rifampin and rifabutin resistance.
The 7% of patients without HIV serology were older, non-Hispanic white or black men without classic HIV risk factors. This group included only two cases of resistant disease. Even if both these patients were assumed to be HIV-positive, their inclusion would not have altered the lack of association between HIV infection and drug-resistant TB. The two patients without viable sensitivity data were both HIV-negative, and their exclusion had no impact on the analyses.
Discussion
Universal directly observed therapy has been used by the Tarrant County Public Health Department since October 1986. The department serves a south-central metropolitan area with a population base in excess of 1 million people. A major strength of this study is the completeness of case inclusion and risk variable assessment including HIV serostatus. We concluded that the prevalence of drug-resistant TB is the same in people with or without HIV infection, and that HIV infection is not a risk factor for the development of resistant disease.
Although not statistically significant, drug-resistant TB was actually less common in HIV-positive patients (4 versus 9%). We believe this is explained by noting that resistant disease is associated with Asian descent, foreign birth and prior treatment. These characteristics were uncommon in the HIV-positive group. Excluding patients in these categories from our dataset results in a drug resistance of 4% for both HIV-positive and negative patients.
In a recent study of Los Angeles County, Asch et al. [10] similarly reported lower drug resistance among HIV-infected individuals (4 versus 13%). Although it appears that their results and ours are in concordance, there are important differences. In our TB population, initial resistance occurred in 8% of cases, whereas in their study initial resistance occurred in 11% of cases. Furthermore, we showed the percentage of initial drug resistance based on eight different drugs, whereas their data was based only on isoniazid initial resistance. Had we restricted our data in a similar way, initial resistance in Los Angeles County would be three times greater than in Tarrant County (11 versus 3.6%). The use of universal directly observable therapy for the past 10 years may account for this observation.
New York City occupies a unique although not singular position relative to TB control. Decreased funding occurring concurrently with the HIV epidemic resulted in a public health infrastructure that was susceptible to outbreaks of drug-resistant TB among both HIV-positive and negative persons. This condition, although most evident in New York City, occurred to a lesser extent in other major urban centers in the United States. Brudney and Dobkin [12] reviewed the resurgence of TB in New York City and they made the salient observation that 89% of persons discharged on TB treatment were lost to follow-up and did not complete therapy. Thus, incomplete treatment and not HIV infection may be hypothesized to account for the increase in resistance. Similarly, O'Brien [8] and Reichman [13] have reviewed drug-resistant TB and emphasize that the problem is predominantly due to failure in therapy.
Patients with HIV who become infected with TB are more likely to progress rapidly to active disease [14]. Thus, resistance patterns in HIV-infected individuals more closely reflect the current resistance patterns of the community. By contrast, disease occurring in HIV-negative individuals reflects the resistance patterns at the time of their infection, which was likely to be years to decades earlier. National surveillance data [8,9] indicate that resistance increased from just over 5% in the early 1980s to over 14% in 1991. When earlier studies are reviewed in light of our data, it appears that HIV coinfection does not increase the risk of drug resistance, but rather that the problem of increased drug resistance in these communities is manifested sooner in HIV-infected patients.
Our data provide direct evidence that drug resistance is not an intrinsic feature of HIV infection, as neither initial nor acquired resistance was associated with HIV infection. In populations that have experienced a loss of TB control, published data [6,11] indicate that drug-resistant TB, although more common in HIV-positive persons (16-37%), is also highly prevalent in HIV-negative individuals (17-19%). These resistance rates are two to four times greater than the resistance rates found in Tarrant County where universal directly observed therapy has been used for more than a decade. This adherence-enhancing strategy has already been demonstrated to reduce initial and acquired resistance [15]. These decreases occurred during the same time-frame that resistance was increasing nationally. Chemotherapy of patients is a TB control measure required for the protection of the public, as well as for the treatment of the individual patient. Non-adherence not only leads to increased resistance but also to higher relapse rates, poor therapy completion rates and more disease transmission. The control of drug-resistant TB requires aggressive implementation of currently recommended TB control measures, which include the expanded use of directly observed therapy.
Acknowledgements
The authors thank J. Cowles of the Texas Department of Health for data collection assistance.
References
1. Centers for Disease Control: Nosocomial transmission of multidrug-resistant tuberculosis among HIV infected persons: Florida and New York, 1988-1991. MMWR 1991, 40:585-591.
2. Centers for Disease Control: Transmission of multidrug-resistant tuberculosis among immunocompromised persons in a correctional system: New York, 1991. MMWR 1992, 41:507-509.
3. Pitchenik AE, Burr J, Miller G, et al.: Outbreaks of drug-resistant tuberculosis at AIDS center. Lancet 1990, 336:440-441.
4. Monno l, Carbonara S, Costa D, et al.: Emergence of drug-resistant Mycobacterium tuberculosis in HIV infected patients [letter]. Lancet 1991, 337:852.
5. Nolan CM, Willaims DL, Cave MD, et al.: Evolution of rifampin resistance in human immunodeficiency virus-associated tuberculosis. Am J Respir Crit Care Med 1995, 152:1067-1071.
6. Frieden TR, Sterling T, Pablos-Mendez A, Kilburn JO, Cauthen GM, Dooley SW: The emergence of drug-resistant tuberculosis in New York City. N Engl J Med 1993, 328:521-526.
7. Shafer RW, Chirgwin KD, Glatt AE, Dahdouh MA, Landesman SH, Suster B: HIV prevalence, immunosuppression, and drug resistance in patients with tuberculosis in an area endemic for AIDS. AIDS 1991, 5:399-405.
8. O'Brien RJ: Drug-resistant tuberculosis: etiology, management and prevention. Semin Respir Med 1994, 9:104-112.
9. Bloch AB, Cauthen GM, Onorato IM, et al.: Nationwide survey of drug-resistant tuberculosis in the United States. JAMA 1994, 271:665-671.
10. Asch A, Knowles L, Rai A, Jones BE, Pogoda J, Barnes PF: Relationship of isoniazid resistance to human immunodeficiency virus infection in patients with tuberculosis. Am J Respir Crit Care Med 1996, 153:1708-1710.
11. Gordin FM, Nelson ET, Matts JP, et al.: The impact of human immunodeficiency virus infection on drug-resistant tuberculosis. Am J Respir Crit Care Med 1996, 154:1478-1483.
12. Brudney K, Dobkin J: Resurgent tuberculosis in New York City. Human immunodeficiency virus, homelessness and the decline of tuberculosis control programs. Am J Respir Dis 1991, 144:745-749.
13. Reichman LB: The challenge of drug-resistant tuberculosis. Cleve Clin J Med 1994, 61:431-437.
14. Daley CL, Small PM, Schecter GF, et al.: An outbreak of tuberculosis with accelerated progression among persons infected with the human immunodeficiency virus. N Engl J Med 1992, 326:231-235.
15. Weis SE, Slocum PC, Blais FX, et al.: The effect of directly observed therapy on the rates of drug resistance and relapse in tuberculosis. N Engl J Med 1994, 330:1179-1184.
© Lippincott-Raven Publishers.