Sandbox/HIV-TB

Editor-In-Chief: C. Michael Gibson, M.S., M.D. [1]; Associate Editor(s)-in-Chief: Alejandro Lemor, M.D. [2]

HIV Coinfection with Tuberculosis

• Among HIV-infected persons, the presentation of active TB disease is influenced by the degree of immunodeficiency.

Clinical Manifestations

Persons with LTBI are, by definition, asymptomatic. Among HIV-infected persons, the presentation of active TB disease is influenced by the degree of immunodeficiency. In addition, early after initiating ART in severely immunosuppressed patients, previously unrecognized subclinical TB can be unmasked by reconstitution of the immune system.

In HIV-infected patients without pronounced immunodeficiency (e.g., CD4+ count >350 cells/µL), HIV-related TB clinically resembles TB among HIV-uninfected persons. The majority of patients have disease limited to the lungs, and common chest radiographic manifestations include upper lobe fibronodular infiltrates with or without cavitation (289). However, extrapulmonary disease is more common in HIV-infected persons than in HIV-uninfected persons, regardless of CD4+ counts, although clinical manifestations are not substantially different from those described in HIV-uninfected persons. TB must be ruled out in diseases of every organ (290) but especially those related to CNS or meningeal symptoms in which early TB treatment is essential to improve outcomes (291,292).

In advanced HIV disease, the chest radiographic findings of pulmonary TB are markedly different compared with those among patients with less severe immunosuppression. Lower lobe, middle lobe, interstitial, and miliary infiltrates are common and cavitation is less common (285,289,293). Marked mediastinal lymphadenopathy also can be found. Even with normal chest radiographs, patients with HIV infection and pulmonary TB might have acid fast bacilli (AFB)-positive sputum smear and culture results.

With increasing degrees of immunodeficiency, extrapulmonary TB (e.g., lymphadenitis, pleuritis, pericarditis, and meningitis), with or without pulmonary involvement, is more common, and found in the majority of TB patients with CD4+ counts <200 cells/µL. Among such patients, TB can be a severe systemic disease with high fevers, rapid progression, and sepsis syndrome.

Histopathologic findings also are affected by the degree of immunodeficiency. Patients with relatively intact immune function have typical granulomatous inflammation associated with TB disease. With progressive immunodeficiency, granulomas become poorly formed or can be completely absent (286).

In severely immunodeficient patients with a high mycobacterial load, TB disease may be subclinical or oligo-symptomatic. After initiation of ART, immune reconstitution might unmask active TB, resulting in pronounced inflammatory reactions at the sites of infection (294--298). This type of IRIS can manifest as early as 7 days after starting ART. Signs and symptoms include fever; weight loss; and signs of local inflammatory reactions such as lymphadenitis, pulmonary consolidation, infiltrates, nodules, and effusions. Histologically, a vigorous granulomatous reaction, with or without caseation, but with suppuration, necrotising inflammation, and AFB might be evident; cultures of this material are almost invariably positive for M. tuberculosis.

Diagnosis

Diagnosis of Latent Tuberculosis Infection (LTBI)

All persons should be tested for LTBI at the time of HIV diagnosis regardless of their TB risk category (AII). Persons with negative diagnostic tests for LTBI, advanced HIV infection (CD4+ count <200 cells/µL), and without indications for initiating empiric LTBI treatment should be re-tested for LTBI once they start ART and attain a CD4+ count >200 cells/µL (AIII). In general, annual testing for LTBI is recommended for HIV-infected persons who are or remain in a "high-risk" category for repeated or ongoing exposure to persons with active TB, i.e., persons who are or who have been incarcerated, live in congregate settings, are active drug users, or have other sociodemographic risk factors for TB (AIII). All HIV-infected persons with a positive diagnostic test for LTBI should undergo chest radiography and clinical evaluation to rule out active TB (AI).

Diagnosis of LTBI can be accomplished with one of two approaches. The tuberculin skin test (TST), placed by the Mantoux method, is considered positive in HIV-infected persons if induration of >5 mm is demonstrated 48--72 hours after the intradermal placement of 0.1 mL purified protein derivative (PPD). Recently, new in vitro assays that detect IFN-γ release in response to M. tuberculosis-specific peptides have been developed for diagnosing LTBI (299). Given the high risk for progression to active disease in HIV-infected persons, any HIV-infected person with reactivity on any of the current LTBI diagnostic tests should be considered infected with M. tuberculosis (Figure 1) (300).

Evidence suggests that the IGRAs have more consistent and higher specificity (92%--97%) compared with TST (56%--95%), better correlation with surrogate measures of exposure to M. tuberculosis, and less cross reactivity because of Bacillus Calmette-Guérin (BCG) vaccination or other nontuberculous mycobacteria exposure than the TST (299,301). Three IGRAs are FDA-approved and available in the United States: the QuantiFERON®-TB Gold and the QuantiFERON®-TB Gold In-Tube (Cellestis Limited), and the TSPOT™.TB test (Oxford Immunotec) is awaiting final FDA approval (Table 10). For both the TST and IGRAs, however, HIV-related immunosuppression might be associated with false-negative results (302). The frequency of false-negative and indeterminate IGRA results increases with advancing immunodeficiency (303,304).

Results from comparative studies of TST and IGRAs in HIV-infected patients indicate that concordance between the tests is not complete (305,306). The TST remains useful for diagnosing LTBI, particularly for patients who have not been vaccinated for BCG and in settings with cost constraints. The optimal application of IGRAs in HIV-infected persons will be better defined when the results of ongoing studies become available (301). IGRAs might be used in combination with TST to improve sensitivity and specificity for detection of LTBI (301).

Fibrotic lesions consistent with TB might be incidentally noted on a chest radiograph obtained for other reasons. Persons with fibrotic lesions should undergo diagnostic testing for LTBI and be evaluated for active disease. Unless the patient has a known history of prior adequate treatment for active TB, sputum samples for AFB smear and culture should be obtained even if the patient is asymptomatic. HIV-infected persons with CD4+ counts <200 cells/µL who have fibrotic lesions consistent with TB on a chest radiograph and no prior history of treatment should be considered as having M. tuberculosis infection irrespective of the results of LTBI diagnostic tests. In situations with moderate-to-high suspicion of active TB regardless of the results of LTBI tests, empiric treatment for active TB should be initiated while awaiting the results of further diagnostic tests (AII).

Diagnosis of Active Tuberculosis

The evaluation of suspected HIV-related TB should include a chest radiograph regardless of the possible anatomic site of disease. Sputum samples for AFB smear and culture should be obtained from patients with pulmonary symptoms and chest radiographic abnormalities. A normal chest radiograph does not exclude the possibility of active pulmonary TB and when suspicion for disease is high, sputum samples should still be obtained (289,307). Obtaining three unique specimens, preferably in the morning of different days, increases the yield for both smear and culture (308). TST and IGRAs should not be relied upon for the diagnosis of TB disease. Approximately one fourth of HIV-infected persons with pulmonary TB disease have false-negative results (301).

HIV serostatus does not affect the yield from sputum smear and culture examinations; positive smear results are more common in cavitary pulmonary disease (309). The yield of AFB smear and culture of specimens from extrapulmonary sites is greater among patients with advanced immunodeficiency compared with HIV-uninfected adults (310-312). Nucleic acid amplification (NAA) tests, also called "direct amplification tests," can be applied directly to clinical specimens such as sputum and help to evaluate persons with a positive AFB smear. A positive NAA result in an AFB smear-positive patient likely reflects active TB. In persons with AFB smear-negative sputum or extrapulmonary disease, however, NAA tests have lower sensitivity and negative predictive value and should be used and interpreted with caution (299).

For patients with signs of extrapulmonary TB, needle aspiration or tissue biopsy of skin lesions, lymph nodes, or pleural or pericardial fluid should be performed. Mycobacterial blood cultures might be helpful for patients with signs of disseminated disease or worsening immunodeficiency.

A positive AFB smear result in any specimen (sputum, needle aspirate, tissue biopsy) represents some form of mycobacterial disease but does not always represent TB. Because TB is the most virulent mycobacterial pathogen and can be spread from person to person, patients with smear-positive results should be considered to have TB disease until definitive mycobacterial species identification is made. Automated liquid media culture systems might indicate growth of M. tuberculosis within 1--3 weeks compared with 3--8 weeks in solid media.

Drug-susceptibility testing and adjustment of the treatment regimen based on results are critical to the successful treatment of patients with TB and to curbing transmission of drug-resistant M. tuberculosis. For all patients with TB disease, testing for susceptibility to first-line agents (isoniazid [INH], rifampin [RIF], ethambutol [EMB], and pyrazinamide [PZA]) should be performed, regardless of the source of the specimen. Drug susceptibility tests should be repeated if sputum cultures remain positive for M. tuberculosis after 3 months of treatment or become positive after 1 month or longer of negative cultures (308). Second-line drug susceptibility testing should be performed only in reference laboratories and should be limited to specimens from patients who 1) have had previous therapy, 2) are contacts of patients with drug-resistant TB disease, 3) have demonstrated resistance to RIF or to other first-line drugs, 4) have positive cultures after 3 months of treatment, or, 5) are from regions with a high prevalence of multiple drug-resistant (MDR) or extensively drug-resistant (XDR) TB (313).

Molecular beacons, phage-based assays, and line probe assays are three methods for rapidly detecting the presence of drug resistance, specifically to INH and RIF. These assays are expensive, require sophisticated laboratory support, need further study, and are not yet FDA-approved for use in the United States. Published results on the performance of the two assays suitable for direct use on samples, the INNO-LiPA Rif.TB kit (Innogenetics, Gent, Belgium) and FASTPlaque-TB (Biotec Laboratories Ltd., Ipswich, United Kingdom), have been inconsistent. Until results of ongoing validation and field testing of these rapid tests are available, conventional laboratory methods for culture and susceptibility testing should be pursued on all suspect clinical specimens.

Preventing Exposure

HIV-infected persons should be advised that time spent in congregate settings or other environments identified as possible sites of TB transmission (e.g., correctional facilities, homeless shelters, nursing homes) might increase the likelihood of contracting M. tuberculosis infection (BIII) (277). Factors known to increase contagiousness include anatomical site of TB disease (pulmonary or laryngeal), AFB smear-positive sputum, cavities evident on chest radiograph, and aerosolization by coughing or singing. HIV-infected patients who have pulmonary or laryngeal TB are, on average, as contagious as patients who are not HIV-infected. Exposure to patients with known TB, but who have AFB smear-negative sputum results, poses a lower but not nonexistent risk for M. tuberculosis transmission (276,314).

In health-care facilities and other environments with a high risk for transmission, all patients with known or presumed infectious TB should be physically separated from other patients, but especially from those with HIV infection (AII) (276). A patient with infectious TB returning to a congregate living setting or to any setting in which susceptible persons might be exposed should be receiving or should have completed treatment and have three consecutive negative AFB smear results from good quality sputum samples collected >8 hours apart (with one specimen collected during the early morning), be on adequate treatment for >2 weeks, and demonstrate clinical improvement before being considered noninfectious (AIII) (276,280,315). Certain specialists recommend that patients with MDR-TB have a negative sputum culture before returning to a congregate setting.

Treatment of LTBI is effective in reducing TB incidence among populations who reside in areas of low, medium, and high TB transmission (316--318). All possible strategies should be pursued to ensure that HIV-infected persons with risk factors for TB are tested for M. tuberculosis infection and those with LTBI receive and complete a course of LTBI treatment (AII) (319). Persons infected with HIV should be treated presumptively for LTBI when the history of TB exposure is substantial, regardless of the results of diagnostic testing for LTBI (BII) (276,282,283). Use of BCG vaccine is not recommended as a means to control TB in the United States because of the unproven efficacy of the vaccine in the U.S. population and the success of other measures in reducing TB incidence (319). BCG vaccination for HIV-infected persons is contraindicated because of its potential to cause disseminated disease (EII).

Preventing Disease (Treatment of LTBI)

All HIV-infected persons with suspected LTBI or who have symptoms indicating TB should promptly undergo chest radiography and clinical evaluation to rule out active TB regardless of the results of diagnostic tests for LTBI (320).

HIV-infected persons, regardless of age, should be treated for LTBI if they have no evidence of active TB and exhibit the following characteristics: 1) a positive diagnostic test for LTBI and no prior history of treatment for active or latent TB (AI); 2) a negative diagnostic test for LTBI but are close contacts of persons with infectious pulmonary TB (AII); and 3) a history of untreated or inadequately treated healed TB (i.e., old fibrotic lesions on chest radiography) regardless of diagnostic tests for LTBI (AII) (321).

The efficacy of LTBI treatment has not been documented for HIV-infected persons with negative diagnostic tests for LTBI without known exposure to M. tuberculosis. Persons from groups or geographic areas with a high prevalence of M. tuberculosis infection might be at increased risk for primary or reactivation TB and, in this situation, decisions to treat for LTBI must include consideration of CD4+ count and other factors (BIII) (282,283,320).

Treatment options for LTBI include INH daily (AII) or twice weekly (BII) for 9 months (282,283,322). Results from a randomized clinical trial comparing INH daily therapy for 9 months with 12 doses of once-weekly INH-rifapentine are pending (323). Because of an increased risk for fatal and severe hepatotoxicity, a 2-month regimen of daily RIF and PZA is not recommended for LTBI treatment regardless of HIV status (DI) (277). HIV-infected persons receiving INH should receive pyridoxine (BIII) to minimize the risk for developing peripheral neuropathy. Alternatives for persons who cannot take INH or who have been exposed to a known INH-resistant index patient include either RIF or rifabutin alone for 4 months (BIII). Decisions to use a regimen containing either RIF or rifabutin should be made after considering potential drug interactions (see the section on ART in the Management of TB Disease). For persons exposed to INH- and/or RIF-resistant TB, decisions to treat with one or two drugs other than INH, RIF, or rifabutin should be based on the relative risk for exposure to organisms broadly resistant to other antimycobacterial drugs and should be made in consultation with public health authorities (AII). Directly observed therapy (DOT) should be used with intermittent dosing regimens (AI) when otherwise feasible to maximize regimen-completion rates (282,283).

No evidence suggests that LTBI treatment should be continued beyond the recommended duration in persons with HIV infection. Therefore, LTBI treatment should be discontinued after completing the appropriate number of doses (AII).

Treatment of Disease

Considering the variability of yield from smear microscopy and NAA tests, empiric treatment should be initiated and continued in HIV-infected persons in whom TB is suspected until all diagnostic work-up (smears, cultures, or other identification results) is complete (AII). When active TB is diagnosed or suspected, a multi-drug anti-TB treatment regimen should be started immediately (AI). This approach promotes rapid killing of tubercle bacilli, prevents the emergence of drug resistance, and decreases the period of contagion (48). DOT is recommended for all patients with HIV-related TB (AII). Likelihood of treatment success is further enhanced by DOT with support for other social and medical needs of HIV-infected patients (BII) (enhanced DOT) (48). A treatment plan should be based on completion of the total number of recommended doses ingested rather than the duration of treatment administration (AIII) (277). The following text summarizes both duration-based and total number-based dosing recommendations.

Recommendations for anti-TB treatment regimens in HIV-infected adults follow the same principles as for adults without HIV infection (AI) (48). Treatment of drug-susceptible TB disease should include a 6-month regimen with an initial phase of INH, RIF or rifabutin, PZA, and EMB administered for 2 months, followed by INH and RIF (or rifabutin) for 4 additional months (AI). When drug-susceptibility testing confirms the absence of resistance to INH, RIF, and PZA, EMB may be discontinued before 2 months of treatment have been completed (AI) (277). For patients with cavitary lung disease and cultures positive for M. tuberculosis after completion of 2 months of therapy, treatment should be extended with INH and RIF for an additional 3 months for a total of 9 months (AII). All HIV-infected patients treated with INH should receive pyridoxine supplementation (BIII). For patients with extrapulmonary TB, a 6- to 9-month regimen (2 months of INH, RIF, PZA, and EMB followed by 4--7 months of INH and RIF) is recommended (AII). Exceptions to the recommendation for a 6- to 9-month regimen for extrapulmonary TB include CNS disease (tuberculoma or meningitis) and bone and joint TB, for which many experts recommend 9--12 months (AII) (277). Adjuvant corticosteroids should be added when treating CNS and pericardial disease (AII). Treatment with corticosteroids should be started intravenously as early as possible with change to oral therapy individualized according to clinical improvement (Table 3). Recommended corticosteroid regimens are dexamethasone 0.3--0.4 mg/kg tapered over 6--8 weeks (324) or prednisone 1 mg/kg for 3 weeks, then tapered for 3--5 weeks.

The optimal way to prevent relapse has not been determined. How the CD4+ count relates to likelihood of treatment failure and relapse remains uncertain. Some recent observational studies suggest that 9 months of therapy result in a lower rate of relapse than shorter or 6-month anti-TB regimens (325--327). While awaiting definitive results of randomized comparisons of treatment duration in HIV-infected patients with TB disease, 6 months of therapy are probably adequate for the majority of patients, but prolonged therapy (up to 9 months) is recommended (as in HIV-uninfected patients) for patients with a delayed response to therapy, with cavitary disease on chest radiograph, and for those with extrapulmonary or CNS disease (BII) (48).

Intermittent dosing (i.e., twice or thrice weekly) facilitates DOT by decreasing the number of encounters required, might provide more effective peak serum concentrations, and is preferable to complete the regimen. For HIV-infected patients, the initial 8-week phase of therapy should be administered daily by DOT (7 days per week for 56 doses or 5 days per week for 40 doses) (AII) or 3 times weekly by DOT for 24 doses (BII) (277). Because twice-weekly administration of the continuation phase of therapy is associated with an increased risk for relapse with acquired rifamycin-resistant M. tuberculosis strains (277,328--330), for patients with CD4+ counts <100 cells/µL the continuation phase of either 4 months or 7 months should be administered either daily or three times weekly by DOT (AIII). Twice-weekly continuation-phase therapy may be considered in patients with CD4+ counts >100 cells/µL (CIII). Once-weekly administration of INH-rifapentine in the continuation phase should not be used for any patient with HIV infection (EI).

Monitoring and Adverse Events

Monitoring of LTBI Treatment

All patients with a diagnosis of LTBI should be counseled about risk for TB, adherence to treatment regimens, benefits and risks of treatment, interactions with other drugs, and an optimal follow-up plan. HIV-infected patients receiving treatment for LTBI also should have baseline laboratory testing, including an evaluation of hepatic function (serum aspartate aminotransferase [AST], bilirubin, and alkaline phosphatase) for patients treated with INH and a complete blood count and platelet count for patients taking RIF or rifabutin (282,283).

Patients being treated for LTBI should be monitored at least monthly with a history and physical assessment designed to detect hepatitis and neuropathy. Patients should be advised to stop treatment and promptly seek medical evaluation if symptoms suggesting hepatitis occur, such as nausea, vomiting, jaundice, or dark urine. Clinicians in all settings should consider dispensing no more than a 1-month supply of medication (331,332). Routine laboratory monitoring is indicated in HIV-infected patients with abnormal baseline liver-function tests, with chronic liver disease, or in those receiving treatment with ART (282,283,333).

Monitoring of Active TB Disease Treatment

A baseline evaluation and monthly follow-up consisting of clinical, bacteriologic, and periodic laboratory and radiographic evaluations are essential to ensure treatment success. Clinical history and baseline tests to evaluate hepatic function (AST, bilirubin, and alkaline phosphatase), renal function (serum creatinine), complete blood count (including platelet count), and CD4+ counts are recommended for all patients (277). HIV-infected patients being treated for active TB should have a clinic-based evaluation at least monthly. For patients with extrapulmonary TB, the frequency and types of evaluations will depend on the sites involved and the ease with which specimens can be obtained. For patients with pulmonary TB, at least one sputum specimen for AFB smear and mycobacterial culture should be obtained monthly until two consecutive specimens are culture negative. Sputum specimens should be obtained after 8 weeks of treatment to inform clinical decision-making about the duration of the continuation phase. For patients with positive AFB smears at initiation of treatment, follow-up smears may be obtained at more frequent intervals (e.g., every 2 weeks until two consecutive specimens are negative) to provide an early assessment of the treatment response (277).

For patients with positive M. tuberculosis cultures after 3 months of treatment, drug-susceptibility tests should be repeated on newly acquired sputum specimens. Patients with positive M. tuberculosis cultures after 4 months of treatment should be considered as treatment failures and managed accordingly (44). At each visit, patients should be questioned about adherence and possible adverse effects of anti-TB medications; those taking EMB should be asked about blurred vision or scotomata and tested for visual acuity and color discrimination. Routine laboratory monitoring during treatment, even when baseline laboratory abnormalities are not present, could be considered (333).

In HIV-infected persons with active TB, serum concentrations of the first-line anti-TB drugs are frequently lower than published normal ranges (334). However, routine drug-level monitoring is not recommended (277). For those with a slow response to treatment, drug concentration measurements might provide objective information on which to base modifications of treatment (335).

Management of Common Adverse Events

Although the reported frequency of anti-TB drug-related toxicity in patients with HIV infection varies, for most adverse events, rates are not different than for HIV-uninfected patients (333,336--339). Because alternative drugs often have less efficacy and more toxicities than first-line anti-TB drugs and diagnosing a drug reaction and determining the responsible agent can be difficult, the first-line drugs (especially INH, RIF, or rifabutin) should not be stopped permanently without strong evidence that the specific anti-TB drug was the cause of the reaction. In such situations, consultation with a specialist in treating LTBI or TB in persons with HIV infection is recommended (48).

Gastrointestinal reactions are common with many of the anti-TB medications (340). If gastrointestinal symptoms occur, AST and bilirubin should be measured, and if the AST level is less than three times the upper limit of normal (ULN) or the baseline for the patient, the symptoms are assumed not to be caused by hepatic toxicity. Typically, gastrointestinal symptoms should be managed without discontinuing TB medications; initial approaches should include either changing the hour of administration or administering drugs with food.

Skin rashes are common with all of the anti-TB drugs. If rash is minor, affects a limited area, or causes pruritis, antihistamines should be administered for symptomatic relief and all anti-TB medications continued. If the rash is severe, all TB medications should be stopped until the rash is substantially improved, and TB drugs restarted one by one at intervals of 2--3 days. RIF or rifabutin should be restarted first (because they are least likely to cause rash and their role in treatment is critical). If the rash recurs, the last drug added should be stopped. If a petechial rash thought to be caused by thrombocytopenia occurs, the RIF or rifabutin should be stopped permanently (341). If a generalized rash associated with either fever or mucous membrane involvement occurs, all drugs should be stopped immediately, the patient should be switched to alternative anti-TB agents, and LTBI or TB treatment should be managed in consultation with a specialist.

Fever in an HIV-infected patient who has been receiving effective TB therapy for several weeks might represent drug fever, a paradoxical reaction, or IRIS (342). If superinfection or worsening TB is excluded as a potential cause, all TB drugs should be stopped. Once the fever has resolved, the general guidelines described for restarting/stopping drugs in the presence of a rash should be followed.

An increase in AST concentration occurs in approximately 20% of patients treated with the standard four-drug anti-TB regimen (343). Drug-induced liver injury can be caused by INH, rifamycins, or PZA and is defined as an AST elevation to >3 times the ULN in the presence of symptoms, or >5 times the ULN in the absence of symptoms (344). In addition to AST elevation, disproportionate increases in bilirubin and alkaline phosphatase occur occasionally. This latter pattern is more consistent with rifamycin hepatotoxicity than with INH or PZA hepatotoxicity. In most patients, asymptomatic aminotransferase elevations resolve spontaneously.

In the absence of symptoms, elevations of AST <3 times ULN should not prompt changes of TB therapy, but the frequency of clinical and laboratory monitoring should be increased. If AST levels are >5 times the ULN regardless of symptoms, >3 times the ULN with symptoms, or if a significant increase in bilirubin and/or alkaline phosphatase occurs, hepatotoxic drugs should be stopped, and the patient should be evaluated immediately. For any substantial new transaminase or bilirubin elevation, serologic testing for hepatitis A, B, and C should be performed and the patient questioned regarding symptoms suggestive of biliary tract disease and exposures to alcohol and other hepatotoxins.

If anti-TB drugs must be stopped for hepatotoxicity, substituting >3 nonhepatotoxic anti-TB drugs is prudent until the specific cause of hepatotoxicity can be determined and an alternative longer-term regimen constructed. The suspected anti-TB medications should be restarted one at a time after the AST level returns to <2 times the ULN or to near baseline for patients with pre-existing abnormalities. Because the rifamycins are a critical part of the TB regimen and are less likely to cause hepatotoxicity than INH or PZA (343), they should be restarted first. If no increase in AST occurs after 1 week, INH may be restarted. PZA may be restarted 1 week after INH if AST does not increase. If symptoms recur or AST increases, the last drug added should be stopped. If RIF and INH are tolerated and hepatitis was severe, PZA should be assumed responsible and should be discontinued. In this last circumstance, depending on the number of doses of PZA taken, severity of disease, and bacteriological status, therapy might be extended to 9 months with RIF and INH alone.

For HIV-infected patients on LTBI therapy who have hepatotoxicity, most of the general guidelines described for restarting/stopping drugs for patients on therapy for active TB apply. The ultimate decision regarding resumption of therapy with the same or a different agent for LTBI treatment should be made after weighing the risk for additional hepatic injury against the benefit of preventing progression to TB disease (333) and always in consultation with an expert in treating LTBI in persons with HIV infection.

ART in the Management of TB Disease

The treatment of TB can be complicated by drug interactions with the rifamycins and overlapping toxicities associated with antiretrovirals (ARVs) and anti-TB drugs when therapy for both HIV and TB infections is concomitantly administered. Both RIF and rifabutin induce CYP3A enzymes, and although rifabutin is not as potent an inducer as RIF, it is a substrate, leading to drug interactions with the PIs and non-nucleoside reverse transcriptase inhibitors (NNRTIs) when these agents are concomitantly administered with the rifamycins; such administration might result in increased metabolism and suboptimal levels of ARVs (345).

Compared with PI-based regimens, NNRTI-based regimens have fewer interactions with RIF-based TB therapy (346). Rifabutin is an alternative to RIF and can be administered with PIs or NNRTIs with appropriate dose adjustments (346). Concomitant use of RIF with ritonavir-boosted PIs has been shown to result in subtherapeutic levels of the PI. Use of ritonavir-boosted saquinavir with RIF was associated with a high incidence of hepatotoxicity in a pharmacokinetic study using healthy volunteers (347). RIF should not be used in patients on PI-based regimens, with or without ritonavir-boosting (EII). For patients undergoing treatment for active TB, starting ART with either an efavirenz- or nevirapine-based regimen is preferred because these NNRTIs have fewer interactions with RIF (BII); dosage adjustments for these NNRTIs might be needed for persons weighing more than 60 kg (BII) (348,349). Delavirdine should not be used with either RIF or rifabutin (350).

If rifabutin is used in place of RIF, dosage reduction is required with boosted-PI regimens. Efavirenz decreases the levels of rifabutin, and the dose of the latter might have to be increased. Nevirapine does not affect the levels of rifabutin sufficiently to merit adjustment of the rifabutin dose. Underdosing of ARVs or rifabutin can result in selection of HIV drug-resistant mutants or acquired rifamycin resistance, respectively, whereas overdosing of rifabutin might result in dose-related toxicities such as neutropenia and uveitis. Because interpatient variations in the degree of enzyme induction or inhibition can occur, the use of therapeutic drug monitoring for levels of rifabutin, PIs, or NNRTIs might help to adjust dosing for individual patients.

HIV nucleos(tide) analogs and the fusion inhibitor enfuvirtide are not affected by the CYP enzymes and can be used with the rifamycins. Results of ongoing drug-drug interaction studies predict that the combination of RIF (and possibly rifabutin) will result in decreased levels of maraviroc, raltegravir, and elvitegravir. Until data are available to guide dose adjustment, these drugs in combination should be avoided or used with extreme caution. Available NNRTIs and PIs do not have clinically significant drug interactions with other first- and second-line anti-TB drugs; thus, when rifamycins cannot be administered because of toxicity or resistance (MDR or XDR M. tuberculosis strains), ART regimens should be selected on the basis of other factors appropriate to the patient (AIII).

Optimal Timing of Initiation of ART in ART-Naïve Patients with Active TB

For ART-naïve, HIV-infected persons who are diagnosed with active TB, anti-TB treatment must be started immediately (AIII). The optimal timing of initiation of ART in this setting is not clear. Options include simultaneous TB and ART or treatment of TB first with delay of ART by several weeks to months. A positive aspect of starting both regimens simultaneously is the possible prevention of progressive HIV disease and reduction in morbidity or mortality associated with TB or other OIs. A negative of this approach is the possibility of overlapping toxicities, drug interactions, a high pill burden, and the possibility of developing IRIS or a paradoxical reaction. These factors must be weighed carefully when choosing the best time to start ART in individual patients.

Several randomized clinical trials are under way to address the optimal timing of initiation of ART in persons being treated for active TB, but the results will not be known for several years. Pending these results, certain specialists determine when to start ART based on the immunologic status of the patients (339,351,352). For patients with a CD4+ count <100 cells/µL, ART should be started after >2 weeks of TB treatment (BII) to reduce confusion about overlapping toxicities, drug interactions (339), and the occurrence of paradoxical reactions or IRIS (353). For persons with a CD4+ count of 100--200 cells/µL, certain specialists would delay ART until the end of the 2-month intensive phase of anti-TB treatment (BII). In those with a sustained CD4+ count >200 cells/µL, ART could be started during the anti-TB maintenance phase and for those with a CD4+ count >350 cells/µL, after finishing anti-TB treatment (BII). In one study, paradoxical reactions occurred in almost all HIV-infected patients with TB and a CD4+ count <100 cells/µL who started ART within the first 30 days of TB therapy (222). However, other studies suggest this approach might prevent HIV disease progression or death (222,339,353). In a small, prospective, nonrandomized study of 49 HIV-infected patients from Brazil (348) treated with a RIF-based anti-TB regimen and efavirenz-based ART, morbidity and side effects of medications in patients who started ART 3 weeks after initiation of TB treatment were reduced, compared with those who started ART and anti-TB treatment simultaneously. Furthermore, simultaneous anti-TB and anti-HIV treatment did not reduce overall mortality.

When TB occurs in patients already on ART, treatment for TB must be started immediately (AIII), and ART should be modified to reduce the risk for drug interactions and maintain virologic suppression. When TB occurs in the setting of virologic failure, ART drug-resistance testing should be performed and a new ART regimen constructed to achieve virologic suppression and avoid drug interactions with the anti-TB regimen administered (AIII).