Geographical region and induration size of positive PPD test help identify those at risk
The resurgence of tuberculosis in the United States over the last decade has strong links to the HIV epidemic. While many groups have studied HIV-related TB, most have restricted their focus to one discrete parameter (such as geographical region or HIV risk behavior) or to settings that traditionally have a high prevalence of TB (such as prisons and shelters for the homeless). Researchers with the Pulmonary Complications of HIV Infection Study have now given us a broader picture of TB in HIV-infected individuals.
Markowitz et al. recruited 1,130 patients in six cities (New York, Newark, Detroit, Chicago, San Francisco, and Los Angeles) over a period of two years. Participants were followed for four years (median follow up: 53 months), during which the incidence of TB was noted and risk factors were recorded. The median absolute CD4 count in the cohort was 410 cells/mm3.
During the observation period 31 participants developed TB. Of these, 16 had pulmonary involvement only, 7 had extrapulmonary involvement only, and 8 had both. Median survival time after diagnosis of TB was 23.6 months. Of the demographic factors examined (HIV transmission behavior, ethnic group, and region), the strongest association was with region. Residents of cities in the eastern U.S. had a relative risk of 5.3 of developing TB compared to participants who lived in the Midwest or West. The investigators diagnosed 22 cases of TB in East Coast cities, versus 6 and 3 cases, respectively, among residents of western and midwestern cities.
Other findings were of interest. Not surprisingly, positive PPD results were associated with a greater risk of TB. The induration-size cutoff point for a positive PPD result in HIV-infected persons with significantly compromised immune function has been a matter of some debate. This study offers further evidence that an induration of 5 mm or greater may be a realistic threshold for positive PPD results in this population. This guideline is consistent with current C.D.C. recommendations.
Tuberculosis developed in 9/70 patients with indurations of at least 10 mm, 2/24 patients with indurations between 5 and 9 mm, and 0/56 patients with indurations less than 5 mm. The researchers note that TB did develop in some patients with negative PPD results; this was more likely to occur in patients with CD4 counts less than 200 cells/mm3. In addition, PPD-negative participants who developed TB were more likely to have negative results of tests for exposure to mumps.
Six months of prophylaxis produces no benefit
Persons who are HIV-positive are 100 times more likely to develop tuberculosis than uninfected individuals, chiefly as a result of the reactivation of latent TB infection, and HIV infection accounts for most of the increased incidence of tuberculosis seen in the United States in the last decade. Unfortunately, the clinical tool used to diagnose latent M. tuberculosis infection, the PPD tuberculin skin test, has low sensitivity in this population because HIV-infected individuals have a high rate of anergy. For this reason the C.D.C. suggested in 1991 that preventive therapy be given to all HIV-positive persons who have anergy and belong to groups, such as the homeless, in which the prevalence of TB infection exceeds 10%. The agent of choice for TB prophylaxis is isoniazid, which has been used successfully for more than 40 years to prevent the reactivation of latent tuberculosis.
Several small studies have shown that isoniazid therapy markedly reduces the incidence of TB in HIV-infected individuals with positive PPD tests. One such study, conducted in New York City, found an infection rate of 9.7 cases per 100 person-years in subjects coinfected with HIV and TB who got no treatment. By contrast, none of the coinfected subjects who got isoniazid developed tuberculosis.
Preventing the spread of tuberculosis is an especially important public-health issue because TB, in both its drug-susceptible and drug-resistant forms, is the only HIV-related opportunistic infection that poses a threat to the general population (see "The Challenge of Multidrug-Resistant Tuberculosis" in this issue). This concern, coupled with the grim fact that TB is the number one killer of HIV-infected people worldwide, prompted investigators associated with the Terry Beirn Community Program for Clinical Research on AIDS to conduct a multicenter, double-blind, placebo-controlled trial that provided six months of isoniazid prophylaxis to HIV-infected patients with anergy who were at high risk for TB.
Between November of 1991 and January of 1994, 517 subjects were enrolled in this study; 260 were randomized to isoniazid, 257 to placebo. The median CD4 count in the study population was 240 cells/mm3 (range: 100 to 417 cells/mm3), and 23% of the patients had C.D.C.-defined AIDS. The two cohorts were similar in baseline characteristics and in completion and discontinuation rates (which were 63% and 9%, respectively). Confirmed TB developed in 3 of the 260 patients assigned to isoniazid, and in 6 of the 257 patients assigned to placebo. Significantly, all of these cases occurred six months or more after the study drug had been discontinued.
Because of the low incidence of TB seen in both groups, investigators concluded that there was no significant benefit associated with isoniazid prophylaxis in this population -- which, as the authors point out, is highly representative of the subpopulation of HIV-positive patients thought to be at high risk for tuberculosis.
Gordin FM, Matts JP, Miller C, Brown LS, Hafner R, 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: 315-20.
Dr. Robert M. Jasmer, co-author of "The Challenge of Multidrug-Resistant Tuberculosis," comments:
Our ability to diagnose latent TB in HIV-positive patients is hampered by their reduced responsiveness to the only currently available clinical measurement of latent infection, the PPD skin test. It was once thought that we might be able to increase the sensitivity of the PPD test in HIV-infected individuals by conducting the test in two stages, with the first test serving as a "booster" to the second, but this did not significantly increase our ability to detect infection. As a result, we remain unable to identify a significant proportion of those coinfected with HIV and TB.
This clinical conundrum led the C.D.C. to propose that practitioners who treat people with HIV disease consider the use of isoniazid as prophylaxis against disease in seropositive individuals likely to have been exposed to tuberculosis. As a number of studies have shown, high-risk patients can benefit from isoniazid prophylaxis, but these benefits must be balanced against the toxic effects of the drug. Long-term use of isoniazid has been associated with hepatotoxicity -- and, in rare instances, with death -- so its administration, especially for the 12-month period that is currently recommended for prophylaxis, is not without risk.
Given the modest incidence of TB seen in both groups studied by Gordin et al., it seems wisest, for now, to restrict the use of isoniazid prophylaxis to individuals who have recently come into close contact with someone who has active TB.
C.D.C. data can help clinicians identify vulnerable patient populations
Since 1993 the C.D.C. has been collecting drug- susceptibility data on patients diagnosed with tuberculosis. Over the past four years the agency has been able to obtain susceptibility results for 89% of all patients with culture-positive TB, 92% of foreign-born patients, and 89% of patients coinfected with TB and HIV.
First the good news: 97% of patients from whom drug-susceptibility results were obtained had isolates that were susceptible to at least two of the three first-line anti-TB agents (i.e., isoniazid, rifampin, and ethambutol). Resistance to isoniazid remained relatively stable over the last four years, and the incidence of multidrug-resistant tuberculosis (MDR-TB) actually declined slightly.
Now the bad news: Resistance to first-line drugs was significantly higher among HIV-positive individuals than among those uninfected with the virus. Infection rates were also higher among the foreign-born and among residents of New York City -- confirming that these factors are indeed risk factors for active tuberculosis infection.
The C.D.C. notes that persons born in Mexico, Vietnam, China, Haiti, India, South Korea, and the Philippines -- all countries where tuberculosis is endemic -- accounted for 66% of the cases of drug-resistant TB seen in foreign nationals. Clinicians who treat HIV-infected patients from these countries should maintain a high index of suspicion that these patients are coinfected with TB, and whenever such patients present with symptoms suggesting active tuberculosis, appropriate tests should be conducted. Common presenting symptoms include: fever, night sweats, cough, dyspnea, and such constitutional symptoms as weight loss, anorexia, and chills.
New York City accounted for a staggering percentage of all cases of drug-resistant TB reported to the C.D.C.: fully 31% of all HIV-positive MDR-TB patients came from Manhattan or its boroughs, as did 14% of all HIV-negative MDR-TB patients and 10% of those whose serostatus was unknown. Accordingly, the C.D.C. warns that HIV-infected patients from New York City should be regarded as being at extremely high risk for drug-resistant TB. This caution also applies to residents of New York State and the District of Columbia, because the incidence of MDR-TB is more than twice as high in these three areas than it is elsewhere in the nation.
Moore M, Onorato IM, McCray E, Castro KG. Trends in drug-resistant tuberculosis in the United States, 1993-1996. JAMA 1997; 278: 833-7.
Dr. Robert M. Jasmer, co-author of "The Challenge of Multidrug-Resistant Tuberculosis," comments:
The emergence of multidrug-resistant strains of tuberculosis has vastly complicated the clinical management of TB in the setting of HIV infection. Disease progression is often rapid in coinfected persons, and the high degree of infectiousness in many of these individuals can accelerate transmission of TB to uninfected partners, care providers, and healthcare professionals. Those who come into close contact with TB patients need to bear this in mind. They should take appropriate steps to reduce the likelihood of transmission, and they should have tuberculin skin tests on a regular basis.
The key to successful treatment of TB is directly-observed therapy. DOT, the acronym given to this rigorous approach to eradicating TB, calls for the administration of all doses of anti-tuberculous medications under the direct supervision of a care provider or health department official. In New York City the MDR-TB case load has fallen by 44% since DOT programs were instituted by the Department of Health. In the District of Columbia, by contrast, cutbacks in funding for DOT programs led to a 36% rise in the number of TB cases.
Several recent reports have suggested that rifampin monoresistance may be unique to HIV-positive individuals, and indeed Moore et al. found that this form of drug resistance was ten times higher in patients infected with HIV than in uninfected individuals. This finding emphasizes the need for clinicians to use all four first-line drugs for initial therapy in all HIV-positive patients with tuberculosis, to prevent the emergence of drug-resistant strains.
Effective, and cost-effective as well
The utility of targeted, culturally-specific HIV prevention programs has become an article of faith among public-health professionals in recent years (see "Effective Intervention Is Possible in the Inner City," Vol. 2, No. 4, pages 83-86, and "Prevention Works!" on pages 87-89 of that same issue). A new study documents the efficacy of this approach. Furthermore, Pinkerton et al. apply a mathematical model to their data, to demonstrate what may seem intuitive: invested prevention dollars save lots of money in averted medical costs.
In this prevention study, 584 gay and bisexual men were recruited in the Pittsburgh area and randomized to two study arms. All participants attended a 60-90 minute lecture on basic virology, transmission of HIV, sequelae of infection, and safer-sex techniques. Half of the study participants also attended a skills-building session that lasted 80 minutes. They received training on how to negotiate safer-sex practices with their partners. The main parameter measured in the study was frequency of condom use by practitioners of anal intercourse. Subjects provided information on their habits at baseline, and then again at intervals of six and 12 months.
At baseline, the rate of condom use among participants randomized to the arm without the skills-building session was slightly higher but statistically insignificant. After six months, however, skills trainees reported condom use with 62% of their partners, while the other participants reported condom use with only 49%. The difference was even more dramatic at 12 months: the condom-use figures were 84% among those who attended the skills-building session, versus 56% among those who did not. The authors postulate that the increases at 12 months are explained by improved negotiation skills.
While these results may be enough to satisfy epidemiologists, budget allocators sometimes need further convincing. Pinkerton et al. provide further arguments in favor of augmented prevention strategies for the latter audience. The costs of the prevention sessions include salaries and benefits for trainers, materials, and indirect costs such as utilities, rent, maintenance, and general administration. The total cost per client was calculated to be $40 and the total costs at $23,360. Assuming a 15% HIV-infection prevalence in the gay community -- a low estimate by most calculations -- the authors assert that 3.05 HIV infections were averted. Using established diagnosis-to-death costs for treating patients with HIV disease, they calculate that their program saved the healthcare system $170,000 in direct medical costs, costs that would have been incurred if those 3.05 infections had occurred.
These results need to be put in some perspective, however. The efficacy of the intervention relies on the accuracy of participants' subjective reports. Some epidemiological parameters had to be estimated. Concerning the financial modeling, start-up costs were not included; the assumption was that community-based organizations are already in place to offer these programs. The authors point out that the cost savings would be lower in populations with HIV prevalence lower than 15%. Nonetheless, this rigorous analysis provides formidable evidence that skills building sessions are a cost-effective component of safer-sex prevention programs when used in appropriate communities.
...but targeted interventions produce results
Some disturbing findings have emerged from an epidemiological study using self-completed questionnaires at gay pride festivals in London over a three-year period. Respondents -- gay men who attended these events from 1993 to 1995 -- were asked to record the number of sexual partners they had had in the previous year, and if they had engaged in anal intercourse, with or without condoms. Sample sizes over the three years of the study were 1,620, 1,735, and 1,168, respectively.
Although the most prevalent sexual behavior was reported to be protected anal intercourse with more than one partner, a very substantial percentage of men -- approximately one-third -- indicated that they practiced unprotected anal intercourse. The proportion of men engaging in this behavior, which carries the greatest risk for HIV transmission, was essentially unchanged over the three years of the study, despite an increase in HIV prevention programs in England during the same period.
The authors entertain two possible explanations for this seeming paradox. First, they observe that even though 80% of new HIV infections in England occur among men who have sex with men, only 12% of the prevention resources in that country are targeted to this population. Alternatively, or perhaps additionally, the authors suggest that the public-health initiatives which are in place may not be working. While this study was uncontrolled, the findings command attention and invite further investigation. They underscore the importance of allocating scarce public resources in ways that are most likely to stem the tide of HIV transmission, and they remind us that we must evaluate the efficacy of risk-reduction programs on a regular basis.
By contrast, Peterson et al. present results of a prevention technique that successfully targeted bisexual and homosexual African-American men. In this study, 318 individuals were recruited and randomized to the following arms: a) participation in three small group HIV prevention training sessions, each lasting three hours, b) one three-hour session, or c) no HIV prevention training.
Participants in the small group sessions were trained in strategies for reducing the risks of HIV transmission associated with sexual activity and needle-sharing. These sessions also taught participants how to negotiate with partners who wanted to engage in unprotected sex, and it encouraged assertiveness and self-protection.
While the percentage of men practicing unprotected anal intercourse remained essentially unchanged in the control group, the percentage engaging in this practice fell from 47% to 38% in the group that attended one training session -- and from 46% to 20% in the group that attended three such sessions -- over the 18 months of the study. These findings add to a growing body of evidence that appropriately targeted intervention programs really do work. Moreover, the work of Peterson et al. reaffirms the importance of equipping at-risk men and women with good negotiating skills, so that they, and not their partners, can set the ground rules for sexual encounters. If the data from Great Britain reveal the shortcomings of broad-based prevention programs, the data from this country confirm that highly targeted programs can have a positive impact on HIV transmission rates.
Hickson FCI, Reid DS, Davies PM, Weatherburn P, Beardsell S, Keogh PG. No aggregate change in homosexual HIV risk behaviour among gay men attending gay pride festivals, United Kingdom, 1993-1995. AIDS 1997; 10: 771-774.
...but only when indirect risks are also addressed
Further evidence of the effectiveness of targeted HIV prevention techniques comes from a four-year study among injection drug users in Chicago. Conducted from 1988 to 1992, the Community Outreach Intervention Project recruited 850 active IDUs who were not in drug-treatment programs. Of the original sample, 25% (209) were HIV-positive. The study focused on the remaining 641 individuals, tabulating the subsequent rates of new HIV infections and the effects of street-based outreach prevention interventions. Former IDUs were engaged to instruct participants on safe needle-use and safe sexual behavior. Needle exchange was not part of the study.
Following baseline interviews, six confidential sessions were held with the participants, who completed questionnaires to reflect their level of adherence to safer behaviors. The rate of HIV seroconversion decreased steadily over the course of the study -- from 8.4 per 100 person-years at the first follow-up evaluation to 2.4 per 100 person-years at the sixth follow-up, four years later.
The success of these techniques is all the more striking in view of the fact that the study participants were not in treatment for their addiction; fully 80% were still using at the end of the four-year period. Although the absence of controls makes it difficult to interpret these data, the authors assert that their model deserves serious consideration by public-health officials. The researchers acknowledge that it is impossible to eliminate confounding factors in a study group like this one, but they feel confident that the significant behavior changes they observed were a result of the instruction these IDUs received through the program.
At the same time, new information is coming to light that complicates prevention strategies for IDUs. To date, education initiatives have largely focused on the importance of clean needles; other potential routes of transmission in this population have not been adequately studied or addressed. "Indirect" routes of transmission in this setting include sharing water to clean syringes or mix drugs, mixing drugs in the same containers, reusing cotton filters, and sharing the drug solution itself. Shared water, cottons, and "works" have all been shown to harbor HIV, and thus they represent infection risks to unsuspecting IDUs -- including those who are careful to use clean needles.
Researchers in Denver interviewed 585 IDUs to ascertain the prevalence of these collateral infection risks. Fully 72% of the sample had exposed themselves to the risk of HIV infection through such indirect routes in the previous 30 days. Approximately 40% of the study participants engaged only in these indirectly risky behaviors, while 33% reported both indirect and direct risk behaviors. The remaining 28% avoided both direct and indirect risks. Reassuringly, researchers found that participants who had participated in community intervention programs were more likely to be in the third category. However, prior drug treatment was not associated with avoiding either direct or indirect risks.
The Denver data sound an alarm about the deficiencies of HIV prevention campaigns for the IDU community that focus exclusively on the importance of clean needles. Many of these individuals may be unaware of the indirect risks involved in sharing paraphernalia other than needles.
Weibel WW, Jimenez A, Johnson W, Ouellet L, Jovanovic B, Lampinen T, et al. Risk behavior and HIV seroincidence among out-of-treatment injection drug users: a four-year prospective study. JAIDS 1997; 12: 282-89.
Back to the October 1997 HIV Newsline contents page.
This article was provided by San Francisco General Hospital. It is a part of the publication HIV Newsline.