INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Persons coinfected with the human immunodeficiency virus (HIV) and Mycobacterium tuberculosis are at much greater risk of developing tuberculosis (TB) than are HIV-seronegative individuals with tuberculous infection (1, 2). It is therefore important to identify coinfected individuals so that preventive therapy with isoniazid (INH) can be initiated. Unfortunately, the ability to identify persons who are infected with M. tuberculosis is limited in the setting of HIV infection because of the high rate of anergy that occurs in this population (3, 4). Because high rates of TB have been reported among anergic, HIV-seropositive injection drug users (IDUs) (5, 6), the Centers for Disease Control and Prevention (CDC) recommend that anergic HIV-seropositive persons who come from a population with more than a 10% prevalence of tuberculous infection be considered for INH preventive therapy (7). However, because the incidence of TB in IDUs will depend on the prevalence of tuberculous infection and on the amount of ongoing transmission of M. tuberculosis in the population, the CDC recommendation may not be appropriate in all areas. In order to determine the incidence of TB among IDUs in San Francisco, including those who were anergic, we conducted a prospective study among persons attending methadone maintenance clinics in the city.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Between 1990 and 1994, we prospectively evaluated IDUs in five methadone maintenance clinics in San Francisco as part of a CDC-sponsored cooperative agreement. All subjects who entered the programs were offered HIV testing and were screened with a tuberculin skin test (TST), using 0.1 ml of purified protein derivative (PPD) (5 tuberculin units) (Connaught Laboratories, Inc., Swiftwater, PA). HIV-seropositive subjects also underwent delayed-type hypersensitivity skin testing with two of the following antigens: mumps (Connaught Laboratories, Inc.); tetanus toxoid (Connaught Laboratories, Inc.); or Candida (Bayer Corporation, Elkhart, IN). The TST was considered positive if the induration was  5 mm in diameter in HIV-seropositive subjects and  10 mm in HIV-seronegative subjects. All tests were administered and read by trained personnel. Anergy was defined as no induration (0 mm) with any of the three antigens. Prevalent cases of TB and persons with a history of TB were excluded from analysis in order to calculate incidence data.

Follow-up was accomplished by reviewing the acquired immune deficiency syndrome (AIDS), TB, and death registries of the City and County of San Francisco. In addition, the methadone clinic records of the study subjects were reviewed to determine the last known contact with the subjects. Patients were classified as alive or dead at the last known follow-up. A person was considered to have TB if he or she met the laboratory culture-confirmed case definition for public health surveillance described by the CDC (8). A chest radiograph was performed on all TST-positive and anergic patients. INH preventive therapy was offered routinely to TST-positive patients regardless of their HIV serostatus. However, anergic subjects were not offered preventive therapy unless the chest film showed parenchymal or hilar calcifications consistent with prior tuberculous infection.

Since 1991, M. tuberculosis isolates from culture-positive cases of TB in San Francisco have undergone IS6110-based restriction-fragment-length polymorphism (RFLP) analysis using standard methodology (9). In our study, isolates from cases occurring within 2 yr of one another, and which had identical RFLP patterns, were considered to have disease as a result of recent transmission.

Incidence rates were calculated from the number of observed TB cases divided by the person-yr of follow-up. Confidence intervals (CIs) for rates and for rate ratios, and tests for differences in rates, were calculated with the assumption of an exponential distribution for the rates. The sample size and time at risk were sufficient to reveal a difference of 5% in incidence rates between the groups within the HIV-seropositive subject population with at least 80% power. Rate ratios were not calculable when the reference rate was zero.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

A total of 1,745 subjects underwent treatment in the methadone maintenance clinics: 529 were not offered HIV testing, 57 refused testing, 27 did not return for the results, 11 had a current or past history of TB, and 12 were lost to follow-up. Thus, 1,109 subjects were evaluated, with a median follow-up of 22 mo (range: 0.07 to 58.0 mo); 338 (30.5%) were HIV seropositive and 771 (69.5%) were HIV seronegative. Study subjects were more likely to be African-American (33% versus 22%, p = 0.001) or male (66% versus 59%, p = 0.003) than those who were not offered HIV testing. Of the 1,109 individuals evaluated, 96 (28.0%) of the 338 HIV-seropositive subjects had a positive TST, compared with 336 of the 771 (44.0%) HIV-negative subjects (p < 0.001). There were 242 (72%) HIV-seropositive subjects with a negative TST, of whom 108 were anergic, 80 were not anergic, and 54 for whom the anergy status was unknown.

During the follow-up period there were eight cases of TB; five of which involved HIV-seropositive (0.73 per 100 person-yr) and three HIV-seronegative subjects (0.18 per 100 person-yr) (p = 0.04). No cases occurred among the anergic individuals. One hundred fourteen (10%) of the study subjects died during the study; 80 (70.2%) were HIV seropositive. Thirty-four (42.5%) of the deaths among the HIV-seropositive individuals occurred in anergic subjects.

Table 1 shows the rates of TB among all subjects, including those who did and those who did not take INH preventive therapy. The rate of TB among the HIV-seropositive, TST-positive subjects who did not take INH was 5.0 per 100 person-yr, compared with 0.39 per 100 person-yr among HIV-seronegative, TST-positive subjects (p = 0.007). Seropositve subjects with a positive TST also had a higher rate of TB than did seropositive anergic subjects (5.0 versus 0 per 100 person-yr; p = 0.005). The difference in the rates of TB in the TST-positive and TST-negative (not anergic) seropositive subjects was of borderline significance (5.0 versus 0.96 per 100 person-yr; p = 0.065). There was no difference in the rates of TB between the seropositive TST-negative (not anergic) and the anergic patients.

The rate of TB among HIV-seropositive, TST-positive patients was significantly lower in those who had previously taken INH preventive therapy (0.61 versus 5.0 per 100 person-yr, p = 0.04). Of the 68 HIV-seropositive, TST-positive subjects who were prescribed INH preventive therapy, 53 (78%) completed the regimen. One person developed TB 2.3 yr after finishing 1 yr of directly observed preventive therapy. Of the 199 HIV-seronegative, TST-positive subjects who were prescribed INH preventive therapy, 147 (74%) completed the regimen. One person developed TB 3 mo after finishing 5 mo of preventive therapy.

It is of note that both HIV-seropositive, TST-negative patients who developed TB did so as a result of recent infection with M. tuberculosis. Both patients were documented by RFLP pattern analysis to be part of clusters of tuberculosis. Case 1, a documented skin test converter, developed TB as part of an outbreak at a residential care facility that was reported previously (10). Case 2 was also part of a cluster, involving 10 people, that occurred over a 2-yr period. This patient had a negative TST at the time of diagnosis. Among the other cases, one of the three incident cases among the HIV- seropositive, TST-positive subjects was also part of a cluster; the patient in this case was the first person identified in a cluster of two cases separated by 15 mo.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This prospective study of IDUs in San Francisco documented high rates of coinfection with HIV-1 and M. tuberculosis. The rate of TB among the HIV-seropositive IDUs with a positive TST who did not take INH preventive therapy was 5.0 per 100 person-yr, which is similar to that in other reports in the literature (1, 2). Despite the high prevalence of tuberculous infection in the group as a whole and the high rate of TB in the HIV-seropositive, TST-positive subjects, there were no cases of TB among the anergic HIV-seropositive individuals. Overall, 30.5% (n = 348) of the subjects were HIV seropositive, and 28% (n = 96) of these individuals had a positive TST. The prevalence of tuberculin positivity was significantly greater among the seronegative individuals, but this difference was probably due to the increased rate of anergy in the seropositive subjects (3). A high rate of tuberculous infection has also been reported by other investigators. Selwyn and colleagues in New York reported that 23% of HIV-seropositive and 20% of HIV-seronegative IDUs had a positive TST (1). MacGregor and coworkers (11) in Philadelphia reported that 32.7% of HIV-seropositive IDUs at a methadone clinic were TST positive. Thus, the high rate of tuberculous infection in our study is in keeping with previous reports from other urban areas in the United States.

The rate of TB among the seropositive, TST-positive subjects in our study who did not take INH preventive therapy was 5.0 per 100 person-yr, which was similar to the 7% per year reported by Selwyn and associates (1). In Spain, the rate of TB among TST-positive subjects was reported to be 10.4 cases per 100 person-yr (2). Importantly, in our study the rate of TB among HIV-seropositive, TST-positive patients was significantly lower among those who had previously taken INH preventive therapy (5.0 versus 0.62; p = 0.04), suggesting a protective effect of this intervention.

In contrast to previously reported studies (5, 6), ours did not demonstrate a high rate of TB among anergic seropositive subjects; no anergic patient developed TB during the 4 yr of the study. Other investigators have reported rates of disease in anergic seropositive persons ranging from 6.6 to 12.4 cases per 100 person-yr (5, 6). These reports led the CDC to recommend that anergic persons who come from a population with more than a 10% prevalence of tuberculous infection be considered for INH preventive therapy (7). Clearly, in areas where high rates of TB have been documented, this may be an appropriate recommendation. However, in San Francisco, the incidence of TB among anergic individuals is too low to justify following this guideline. In addition, data from Baltimore have demonstrated that the rate of TB in IDUs is decreased by providing INH preventive treatment to TST-positive individuals only, and not to anergic patients (12), at least in an area with a low prevalence of TB. The extent of TB transmission in a community must also be considered. In San Francisco, high rates of recent transmission with rapid progression to disease have been documented with IS6110-based RFLP analysis (9). In such areas, the CDCs recommendation of using INH preventive therapy may or may not be effective in preventing disease. For example, if the two TST-negative patients in our study who developed disease had been given INH at their initial screening, it might not have prevented their subsequent infection and progression to disease. The same could be said of the anergic patients, had any of them been exposed to a person with infectious TB.

It is unclear why the anergic patients in our study did not develop TB, since they came from a population with a high prevalence of tuberculous infection. Possibly, they were not infected with M. tuberculosis. Chin and coworkers (13) have shown that the assessment of immune function with mumps and Candida antigens is imprecise at best, and that anergy defined by these antigens may not be permanent. Thus, "anergy" may not have been indicative of reduced cell-mediated immunity in our patients. It is also possible that since TB can occur early in HIV disease, those patients who were most likely to develop TB would already have done so by the time they became anergic. If they were to develop TB it would be because of recent acquisition of infection, such as occurred in our TST-negative patients, but for some reason the anergic subjects did not come into contact with infectious cases of TB. Additionally, simply providing preventive therapy to those anergic subjects who had radiographic evidence consistent with tuberculous infection might have selected out the highest risk group, which may make this one approach to determining which tuberculin-negative or anergic patients should receive preventive therapy.

Had we followed CDC recommendations (7), we would have given 99 anergic IDUs INH preventive therapy and not prevented a single case of TB. Likewise, we would probably not have prevented the development of TB among the TST-negative subjects. We therefore believe that in San Francisco, and probably in other areas, INH preventive therapy should be given only to seropositive patients with a documented positive TST, a history of a positive TST, or recent contact with an infectious case of TB. Whether or not preventive therapy should be provided to tuberculin-negative or anergic patients with radiographic abnormalities suggestive of tuberculous infection requires further study.