>Tuberculosis, a disease modern medicine once thought it had conquered, is now the leading cause of death among people with HIV worldwide. Until recently, TB was not perceived as a particular threat to HIV-positive individuals in this country -- but that was before the emergence of new strains of tuberculosis, strains that are resistant to one or more of the drugs used to treat this infection. These drug-resistant strains of tuberculosis have resulted in outbreaks of disease that are difficult to treat, rapidly progressive, and exceedingly dangerous.

In the past, a simple skin test, administered in your care provider's office, was sufficient to determine if you had been exposed to TB. (This test, the only currently available clinical measurement of latent TB infection, is known as the tuberculin skin test.) If this skin test produced a firm welt, not much bigger than a pimple, that meant that you had been exposed to tuberculosis. Once active TB had been ruled out, drug therapy was then initiated for latent (inactive) tuberculosis. The problem is that the tuberculin skin test doesn't work very well in people with HIV. Many of them, especially those with more advanced HIV disease, have what is called anergy: their weakened immune systems fail to respond to the skin test even if they have been exposed to TB. As a result, care providers have no way of knowing for sure if a person with HIV has TB infection until that person develops symptoms of active disease.

At one time HIV experts thought that they might be able to increase the sensitivity of the tuberculin test by conducting the test in two stages, with the first test serving as a "booster" to the second, but this did not prove possible. As a result, care providers remain unable to identify a significant proportion of those who are infected with both HIV and TB.

This diagnostic dilemma led the C.D.C. to propose that doctors who treat people with HIV consider using isoniazid, one of the most powerful of the first-line drugs used to treat TB, as prophylaxis against infection in people 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 liver toxicity -- and, in rare instances, death -- so its administration, especially for the 12-month period that is currently recommended for prophylaxis, is not without risk.

"Preventing the spread of tuberculosis is an especially important public-health issue because TB is the only HIV- related opportunistic infection that poses a threat to the general population. Unlike PCP, CMV and MAC, TB can be transmitted to otherwise healthy people who live with, care for, or simply come into intimate contact with infected individuals."

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. Unlike PCP, CMV, and other AIDS-related opportunistic infections, TB can be transmitted to otherwise healthy people who live with, care for, or simply come into intimate contact with infected individuals. This concern, coupled with the grim fact that TB is the number one killer of HIV-infected people around the globe, prompted investigators 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, and 23% of the patients had C.D.C.-defined AIDS. Confirmed TB developed in 3 of the 260 patients assigned to isoniazid, and in 6 of the 257 assigned to placebo. Significantly, all of these cases occurred six months or more after the study drug had been discontinued.

Given the modest incidence of TB seen in both groups, it seems wisest, for now, to give isoniazid as TB prophylaxis only to anergic people with HIV who have recently come into close contact with someone who has active tuberculosis (see "Should You Be on Preventive Therapy Against TB?" in the Newsline section of this issue).

While we do not have a way of telling, with absolute certainty, if an individual who is HIV-positive has been exposed to TB, we do know that certain groups are at higher risk of TB infection than others. The epidemiologic risk factors for tuberculosis include a history of incarceration, homelessness, and substance abuse (alcohol or drugs). Country of origin is another risk factor: patients born in countries with a high prevalence of tuberculosis run an increased risk of infection because of exposure to TB in their homeland prior to immigration. Infection is also more likely to occur in the eastern United States than in the Midwest or Far West.

Moreover, we now know that certain groups are at much higher risk of developing tuberculosis that is resistant to one or more of the drugs that are used to treat this infection. These groups include foreign-born nationals from Mexico, Vietnam, China, Haiti, India, South Korea and the Philippines, and residents of New York City, New York State, and the District of Columbia (see "Drug-Resistant TB: Are You at Risk?" in the Newsline section).

What are the symptoms of multidrug-resistant TB?

Tuberculosis that is resistant to two or more of the first-line drugs used to treat TB -- that is, TB resistant to isoniazid and rifampin, with or without resistance to either pyrazinamide, ethambutol, or streptomycin -- is known as multidrug-resistant tuberculosis, or MDR-TB. It is an especially formidable problem among HIV-infected patients, in whom mortality from MDR-TB can be as high as 80%.

The clinical presentation of MDR-TB does not differ substantially from that of tuberculosis caused by drug-susceptible bacilli. In HIV-infected patients, as in other populations, tuberculosis is primarily a pulmonary infection, although infection outside the lungs occurs in up to 40% of patients infected with both HIV and TB. Tuberculosis can occur at any CD4 cell count, but as the CD4 cell count decreases, the incidence of disseminated tuberculosis increases.

In HIV-infected patients with MDR-TB, the most common presenting symptoms are fever, night sweats, cough, and shortness of breath (Table). These symptoms tend to be present for several weeks to months prior to evaluation, in contrast to the often more acute presentation of bacterial pneumonia. Constitutional symptoms, such as weight loss, anorexia, and chills, are also frequently reported. If you have one or more of these symptoms -- especially in combination with one or more epidemiologic risk factors for TB exposure -- you should consult your doctor immediately.

SYMPTOMS OF MULTIDRUG-RESISTANT TB Symptoms Comments Epidemiologic risk factors: Incarceration, homelessness, substance abuse (alcohol or drugs), contact with someone with TB, birth outside U.S. Patients from countries with a high incidence of TB are more likely to be infected prior to arrival in U.S. Most common presenting symptoms: Fever, night sweats, cough, shortness of breath Constitutional symptoms: Weight loss, anorexia, chills Symptoms present for weeks to months, in contrast to acute presentation of bacterial pneumonia Physical findings: Fever, abnormal breath sounds, generalized or regional swelling of the lymph nodes Other OIs produce similar symptoms, making diagnosis difficult

"Because the symptoms of tuberculosis in people with HIV are so nonspecific, the key to accurate diagnosis is a high degree of suspicion -- and the inclusion of tuberculosis in the differential diagnosis. The single most powerful predictor of the presence of MDR-TB is a history of previous treatment for tuberculosis."

Because the symptoms of tuberculosis in people with HIV are so nonspecific, the key to accurate diagnosis is a high degree of suspicion -- and the inclusion of tuberculosis in the differential diagnosis. The single most powerful predictor of the presence of MDR-TB is a history of previous treatment for tuberculosis. Inadequate therapy, due to either poor prescribing practices by physicians or poor compliance by patients, is the most common route by which resistant organisms are acquired.

Previous treatment is not the only risk factor for MDR-TB, however. Other risk factors include exposure to a patient with MDR-TB, being from a country with a high prevalence of MDR-TB (such as Mexico or the Philippines), and living in a community with greater than 4% resistance to isoniazid. In the U.S., the highest rates of MDR-TB have been found in large urban areas such as New York City, Washington, D.C., and Miami.

Treating multidrug-resistant TB

Only laboratory testing can determine whether a strain of tuberculosis is drug-resistant. These same tests tell your care provider which drugs your particular strain of TB will respond to. With modern techniques for identifying drug-resistant strains, diagnosis of MDR-TB can usually be made within three to six weeks after specimens have been sent to the laboratory. Given the extremely high mortality from MDR-TB in HIV-infected patients, it is imperative that treatment occur under the supervision of physicians who are experts in the management of tuberculosis. Because the treatment of MDR-TB varies depending upon the patient's specific resistance pattern and history of previous treatment regimens, therapy is highly individualized. Several principles do apply to treating MDR-TB in all HIV-infected patients, however.

The single most important principle in the treatment of MDR-TB is to administer at least two drugs to which there is documented susceptibility. The corollary to this principle is: Never administer only one drug to a patient who is failing a TB regimen. In fact, not adhering to this principle is responsible in part for the very existence of strains of MDR-TB, since therapy with only one active drug will inevitably lead to the development of resistant organisms.

It is currently recommended that initial treatment of tuberculosis in all patients consist of all four first-line drugs (isoniazid, rifampin, pyrazinamide, and ethambutol or streptomycin). There are rare exceptions to this rule, but from a practical point of view all HIV-infected persons should be treated with this four-drug regimen until their drug susceptibility data become available.

Because these drugs interact with all four of the currently approved protease inhibitors -- in ways that dramatically diminish the effectiveness of the protease inhibitors -- special precautions must be taken when people with HIV who are taking a protease inhibitor as part of a combination antiretroviral regimen develop active tuberculosis. These precautions are outlined in the special section found below, under the heading "Guidelines for the Treatment of TB in People with HIV."

Another important issue involves the duration of treatment of MDR-TB. Treatment of MDR-TB requires at least nine months of therapy, and often as much as 24 months. As a general rule, if rifampin is one of the drugs to which the patient's TB strain is resistant, treatment will often require up to 24 months of therapy after the patient's sputum culture becomes negative.

In addition, all HIV-infected patients with MDR-TB should be given directly observed therapy, or DOT -- that is, each dose should be taken under the direct supervision of a care provider or health department official. DOT programs have been shown to prolong survival in HIV-infected patients with MDR-TB (see "Aggressive treatment programs defuse a global TB crisis," AIDS Care, Vol. 1., No. 2), and they are now considered to be the standard of care for all patients with MDR-TB.

Summary

Diagnosing and treating MDR-TB in HIV-infected patients can be a significant clinical challenge. The key to diagnosis is to maintain a high level of suspicion, especially in persons with epidemiologic risk factors for tuberculosis, and to order appropriate diagnostic studies whenever TB is suspected. Treatment of MDR-TB should ideally be handled by persons with expertise in the management of tuberculosis. Given the extraordinarily high mortality of this disease, and the risk it presents to those who come into close contact with infected individuals, we cannot afford to make its treatment less than a top priority.

Guidelines for the Treatment of TB in People with HIV All four of the currently approved protease inhibitors interact with the main drugs used to treat tuberculosis infections. These anti-TB agents, the rifamycins -- a family that includes rifampin and rifabutin -- accelerate the metabolism of protease inhibitors, and when these drugs are taken at the same time the rifamycins reduce blood concentrations of the protease inhibitors to ineffective levels. To make matters worse, this drug-drug interaction works in both directions: the protease inhibitors retard the metabolism of rifamycins -- and this can cause blood concentrations of the latter drugs to rise to toxic levels. The insidious interaction of these two drug classes creates a dilemma when patients who are taking protease inhibitors develop TB. Physicians and other healthcare providers recognize that untreated tuberculosis can prove fatal in patients with advanced HIV infection, so prompt and aggressive therapy is required. In theory, this means that protease inhibitor therapy should be discontinued for the length of time a patient is receiving a rifamycin as one component of a multidrug anti-TB regimen. The current recommendation is that this anti-TB treatment last at least six months. As all those who treat HIV disease now recognize, however, any interruption of combination antiretroviral therapy -- even for periods as brief as a few days -- can result in the development of drug-resistant viral strains. Because active TB infection is seen in increasing numbers of people with advanced HIV disease, a growing number of individuals are candidates for treatment with both rifampin and protease inhibitors. The C.D.C., recognizing that the management of these patients is complex, has issued guidelines for the care of individuals who are infected with both HIV and TB. HIV-positive patients with TB who have considered, but have not begun, treatment with a protease inhibitor: If you have been found to have a form of TB that is responsive to drug therapy -- and you have not yet begun antiretroviral therapy with a multidrug combination that includes a protease inhibitor -- the C.D.C. recommends that you complete TB treatment with a regimen that contains rifampin before starting that multidrug regimen. Anti-TB therapy should last at least six months, and treatment should be prolonged if your response is slow or suboptimal. Directly-observed therapy is routinely recommended for the treatment of TB, to ensure a high level of adherence with the standard four-drug regimen. DOT, the acronym given to this approach to eradicating TB, calls for the administration of all doses of anti-TB medications under the direct supervision of a care provider or health department official. During this period patients can continue to take, or can switch to, combination antiretroviral therapies -- such as AZT or d4T plus 3TC and nevirapine -- that do not include protease inhibitors. HIV-positive patients with TB who are on protease inhibitor therapy: The C.D.C. offers three options for patients who develop active TB while they are taking a protease inhibitor as part of their antiretroviral therapy. OPTION #1: Discontinue protease inhibitor therapy until you have completed a six-month course of anti-TB therapy with a regimen that contains rifampin. The C.D.C. recognizes that this may not be regarded as an option by many patients (or their care providers), since any interruption in administration of a prescribed protease inhibitor can lead to the development of drug-resistant viral strains -- mutations that may be resistant not just to the protease inhibitor that the patient was taking but to all drugs in this class. Clearly, discontinuing protease inhibitor therapy poses certain risks. Treating active TB with regimens that do not include rifampin also carries risks, however. The potential consequences of withholding rifampin from patients coinfected with HIV and TB include: prolonged duration of therapy (to at least 18 and perhaps as long as 24 months), increased likelihood of treatment failure and death, and slower conversion of sputum culture to negative -- meaning that patients remain capable of infecting others for longer periods of time. Because persistent TB infection is associated with more rapid progression of HIV disease, the C.D.C. does not recommend treating rifampin-susceptible TB with regimens that do not contain that drug. OPTION #2: To minimize the adverse effects of interrupting protease inhibitor therapy, treat TB with an aggressive, short-term, four-drug regimen (daily isoniazid, rifampin, pyrazinamide, and either ethambutol or streptomycin) for a minimum of two months. Once bacteriologic response has been documented and the results of drug-susceptibility testing are available -- a process that usually takes three months -- treatment may be modified. The continuation phase of therapy, which lasts at least 16 months, consists of isoniazid (15 mg/kg) and ethambutol (50 mg/kg) two times each week. This two-drug anti-TB regimen permits the reintroduction of protease inhibitor therapy. Some experts recommend the addition of streptomycin or pyrazinamide to the patient's anti-TB regimen, if test results show that the infecting organism is susceptible to the third drug. For obvious reasons, OPTION #2 must be ruled out if susceptibility testing establishes that the patient in question has isoniazid-resistant TB. OPTION #3: Continue protease inhibitor therapy with indinavir (800 mg every 8 hours) and administer a four-drug, nine-month anti-TB regimen, substituting rifabutin (150 mg daily) for rifampin. This approach to the management of TB requires exceedingly careful monitoring of all patients, and physicians will want to consider obtaining measurements of blood concentrations of rifabutin at intervals during the nine-month course of therapy. This alternative TB therapy is recommended by the C.D.C. based on what is known about the pharmacokinetics of rifabutin and on limited data from clinical trials -- including one which found that the combination of rifabutin plus indinavir, at the standard doses given above, yielded acceptable blood levels of both drugs. OPTION #3 cannot be recommended for patients taking either saquinavir or ritonavir, because of the strong drug-drug interactions that occur when these protease inhibitors are taken in combination with rifabutin. Patients who are currently taking one of these protease inhibitors may be switched to indinavir and then prescribed rifabutin, but these changes in medication should be made only after consulting an expert on the use of protease inhibitors. What your care provider needs to consider: Although a multicenter international trial has shown that six months of anti-TB therapy with a regimen containing rifabutin was as effective as therapy with a similar regimen containing rifampin, rifabutin is not currently approved by the F.D.A. for the treatment or prevention of TB. It is approved for prevention of MAC, however, and there has been some concern that the coadministration of indinavir and rifabutin might produce drug-drug interactions that would complicate the prophylaxis and treatment of MAC. According to Merck, the makers of indinavir, rifabutin at a daily dose of 150 mg will provide effective MAC prophylaxis in patients taking this protease inhibitor. Because neither OPTION #2 or OPTION #3 has been studied in large clinical trials, the C.D.C. recommends that care providers who treat patients infected with both HIV and TB follow these interim guidelines: On initiation of therapy, perform frequent bacteriologic evaluations to document sputum conversion to culture-negative status -- and, after culture conversion, to detect possible treatment failure. Extend the duration of therapy to at least 18 months for OPTION #2, and to nine months for OPTION #3. Use only indinavir with OPTION #3. Carefully monitor patients for drug toxicity. Use directly-observe-therapy (DOT) throughout the course of anti-TB treatment. Reëvaluate successfully treated patients at regular intervals for at least two years, beginning with an assessment of bacteriologic status six months after therapy ends, and instruct patients to promptly report symptoms suggestive of relapsed tuberculosis.

Robert M. Jasmer, M.D., is a Senior Fellow at the Division of Pulmonary and Critical Care Medicine, University of California, San Francisco.

Laurence Huang, M.D., is Assistant Professor of Medicine, UCSF; Medical Director of the Inpatient AIDS Unit, San Francisco General Hospital; and Chief of the San Francisco General Hospital AIDS Chest Clinic.