Globally, tuberculosis (TB) remains an immense cause of morbidity and mortality. Approximately one-third of the population, more than 2 billion persons, harbor latent infection with Mycobacterium tuberculosis, and 9 million new cases develop annually. Because of delayed diagnosis, lack of access to medications, or nonadherence to prescribed regimens, approximately 1.5 million people die annually of TB.
The progressive appearance of drug-resistant forms of TB has been ominous. Circa 1990, cases of “multi drug resistant tuberculosis” (MDR-TB) were recognized in New York City and Miami, Florida. Strains of TB resistant to the two major drugs, isoniazid (INH) and rifampin (RIF), were associated with hospital-based outbreaks, primarily among persons with AIDS. Similar, highly lethal outbreaks were subsequently recognized around the globe, and by 2007, the World Health Organization estimated that nearly 500,000 new cases of MDR-TB were occurring yearly. More ominously, in 2005 in South Africa, an even more resistant epidemic was noted, extensive drug-resistant TB (XDR- TB). These strains had evolved from MDR-TB and entailed resistance not only to INH and RIF but also to two other major classes of anti-TB drugs, the ﬂuoroquinolones and the second-line injectables such as amikacin, kanamycin, and capreomycin. Cases of XDR-TB have now been reported from 47 nations on all of the continents.
|DISSEMINATION OF TUBERCULOSIS|
HIV/AIDS has profoundly accelerated case rates and deaths over the past quarter century. The coincidence of TB and AIDS has been most pronounced in sub-Saharan Africa but is increasingly problematic in China, India, Russia, and the former Soviet Republics.
Failure of the TB vaccine, bacillus Calmette-Guérin (BCG), has been a major element of the ongoing epidemic. Although BCG given in infancy does protect against some of the severe forms of childhood TB (meningeal, spinal, or disseminated disease), the vaccine has not curtailed adult pulmonary TB, the vector of ongoing airborne transmission.
In the United States, TB case rates have declined mostly because of the widening use of directly observed therapy (DOT). This practice has increased treatment completion, reduced transmission, and dramatically reduced acquired drug resistance. In 1992 in the United States, when DOT was used in less than 20% of cases, there were nearly 27,000 cases for a rate of 10.4 per 100,000 population, and there were more than 400 MDR-TB cases. By 2007, there were fewer than 14,000 cases, and the rate had fallen to 4.4 per 100K. Notably, 58% of these cases were among foreign-born individuals, the majority of whom brought this infection from their country of origin. Fewer than 100 MDR-TB cases were seen in 2007, mainly among foreign-born individuals who arrived with preformed drug resistance.
TB is primarily spread by patients with sputum smear–positive pulmonary disease. Coughing produces small particles, which ﬂoat in the air and undergo dehydration, forming “droplet nuclei” in the range of 1 m size. These can reach the alveoli and avoid clearance by the mucociliary escalator of the airways. Transmission occurs exclusively indoors, where these particles are concentrated and protected against ultraviolet irradiation (see Plate 4-93).
|EVOLUTION OF TUBERCLE|
After they have been taken up by naïve alveolar macrophages, the bacilli replicate and before inhibitory cellular immunity can limit their growth, there is wide-spread dissemination (see Plates 4-93 and 4-94). In most instances, host defenses prevail. The infected loci involute, and the only detectable manifestation of the encounter is a reactive tuberculin skin test result. However, in various sites, viable tubercle bacilli may persist for years or decades with the potential to cause “reactivation” TB, either in the lungs or extrapulmonary tissue. Among those with compromised immunity (e.g., very young individuals, people with AIDS, patients on immunosuppressive therapy), however, there may be progressive disseminated disease appearing within a few months of exposure.
|INITIAL (PRIMARY) TUBERCULOUS COMPLEX|
Pulmonary disease is the commonest presentation of TB (see Plates 4-95 to 4-98). Among newly infected infants and children, there is a distinctive pattern referred to as primary TB. These young patients commonly react to the new infection with exuberant lymphadenopathy in the peribronchial, hilar, or paratracheal nodes, draining the region of the lung where the new infection occurred. The original parenchymal focus may be very small or not visible on routine chest radio- graphs by this time. Over months or years, this primary lesion may calcify, leaving a “Ghon lesion.” Cavitation rarely occurs, and there are very limited numbers of tubercle bacilli in respiratory secretions. Among such patients, many of whom cannot cooperate with sputum collection, gastric aspirates or bronchoscopy may be required to isolate the TB organisms.
The histopathology of tuberculous lesions classically involves granulomas with central caseous (“cheeselike”) necrosis. Bacilli, if seen, are typically at the margins of the lymphocyte-macrophage palisades and the necrotic debris. As the disease advances, the lesions erode into the airways, allowing expulsion of the bacilli into the environment. However, the critical component of transmission is the formation of cavities. In the environment of the cavity wall, tubercle bacilli undergo logarithmic extracellular replication. Patients with cavitary disease 6 to 108 bacilli per milliliter of um.
|EXTENSIVE CAVITARY DISEASE|
Another aspect of the large population of rapidly multiplying organisms in cavities is the likelihood of spawning drug-resistant mutants. Therapeutically, this places a great premium on the initial intensive phase of treatment (see below).
Pleuritis may occur if the primary parenchymal lesion is close enough to the visceral pleura to induce inﬂammation of the mesothelial surface. In some cases, there is typical sharp chest pain, but in others, an asymptomatic effusion is observed. The pleural effusion may be accessed by aspiration; classically, it is a lymphocyte-rich exudate. Acid-fast bacilli are rarely seen on stain, and culture results are positive in fewer than half of the cases. Biopsy is the most useful study.
Miliary TB is manifested in the lungs by bilateral ﬁne nodular opacities, predominantly in the dependent zones, where the bacilli have been deposited hematogenously (see Plate 4-98). Disseminated TB is seen among patients with impaired immunity, including those at the extremes of age.
TUBERCULIN SKIN TESTING AND INTERFERON RELEASE ASSAYS
Historically, reactivity to intradermally injected tuberculoprotein has been used in the diagnosis of TB (see Plate 4-99). “Old tuberculin,” originally prepared by Koch, was supplanted by puriﬁed protein derivative (PPD). The test results were measured by the amount of induration at 48 to 96 hours after placement. However, the tuberculin skin test (TST) was limited by insensitivity and nonspeciﬁcity; 15% to 25% of “normal hosts” with TB did not have a signiﬁcant response to PPD, and prior BCG vaccination or nontuberculous mycobacterial infections may induce signiﬁcant reactions to PPD.
The primary role of the TST has been identifying latent TB infection (LTBI). Because it has been the only test that deﬁnes LTBI, its sensitivity is unknown. However, the lack of speciﬁcity has clearly been problematic, particularly among high-risk immigrants who have received BCG.
Recent studies have identiﬁed two antigens that are nearly speciﬁc to M. tuberculosis: culture ﬁltrate protein- 10 (CFP-10) and early secretory antigen-6 (ESAT-6). Two new commercial tests have been developed that measure ex vivo production by whole blood of interferon-(IFN-) on exposure to CFP-10 and ESAT-6 (Quantiferon or T. Spot-TB). Overall, these IFN-release assays (IGRAs) appear to be slightly more sensitive and signiﬁcantly more speciﬁc than the TST. Other advantages are that they require only a single encounter, are more objective in interpretation, and do not “boost” reactivity as do serial TSTs. Conversely, the IGRAs are substantially more expensive, technically demanding, and require proximity to a laboratory.
The routine chest radiograph, posteroanterior (PA) and lateral (LAT) views, has been the traditional test for pulmonary, pleural, and miliary TB. The most common sites for disease area the posterior aspects of the upper lobes and the superior segments of the lower lobes. Common ﬁndings include ﬁbronodular stranding extending up from the hilum, retraction cephalad of the pulmonary artery, thickening of the apical pleural cap and (with advanced disease) cavitation. Among HIV- negative “normal” hosts, roughly two-thirds have these typical ﬁndings.
Among immunocompromised patients (people with AIDS, those who have undergone organ transplantation, those with renal failure, and recipients of tumor necrosis factor– [TNF-] inhibitors), the radiographic presentation may be quite atypical, including lower zone pneumonic inﬁltrates, prominent lymphadenopathy, huge pleural effusions, and diffuse opacities that begin as tiny, discrete (“millet seed”) shadows but may progress to a conﬂuent airspace ﬁlling process or acute respiratory distress syndrome.
Computed tomography (CT) scans may reveal important diagnostic ﬁndings, including cavities obscured by osseous or soft tissue structures in the apices, grossly enlarged lymphadenopathy with hypodense centers typical of AIDS-TB, or early miliary shadowing not visible on routine chest radiography.
Pleural effusions were typically assessed by performing lateral decubitus views that led the effusions to collect in the dependent zones. Although decubitus views are still popular, ultrasound studies allow for more thorough characterization of the pleural process and increase the yield of aspiration, biopsy, and drainage procedures.
Sputum studies include nucleic acid ampliﬁcation (NAA) and culture (see Plates 4-100 and 4-101). In much of the world, unprocessed sputum is smeared on a glass slide, stained with the Ziehl-Neelsen or Kinyoun technique, and examined by a technician at high-power magniﬁcation under a light microscope. This system is insensitive, identifying fewer than 50% of patients whose sputum would be culture positive in more sophisticated laboratories. This system has been justiﬁed by identifying the cases most likely to transmit TB. However, it is diminishingly acceptable given its insensitivity and incapacity for drug susceptibility testing.
In modern laboratories, sputum undergoes decontamination (so the media are not overgrown by microorganisms other than mycobacteria), digestion (to release the mycobacteria from the proteinaceous matrix of the mucus), and concentration (in a centrifuge that is refrigerated so that the heat generated by high-speed centrifugation does not kill too many of the mycobacteria). At the end of this three-step process, the supernatant is decanted, and the pellet is subjected to microscopy, NAA, and culture. Fluorescent microscopy using auramine-O stains the mycobacteria bright yellow on a black background, which allows the reading at 40 magniﬁcation.
NAA may be performed on uncultivated bacilli to establish the species as M. tuberculosis. This is particularly valuable in communities with a signiﬁcant prevalence of pulmonary disease because of nontuberculous mycobacteria (NTM; see below). Early species identiﬁcation allows optimal selection of the drug regimen (TB or NTM), efﬁcient use of isolation facilities, and appropriate initiation of contact investigation (see below).
Culture remains the central element in the diagnosis of TB. Because microscopy results are positive in only 40% to 60% of cases ultimately proven to be pulmonary TB, culture enhances the sensitivity of sputum study and conﬁrms the species and facilitates drug susceptibility testing.
Solid media such as Löwenstein-Jensen remain in use in many parts of the world. However, such techniques are slow in yielding results, typically requiring 4 to 6 weeks. Liquid media have replaced or supplemented the solid media in many industrialized nations, offering culture results, speciation, and preliminary susceptibility results in 1 to 2 weeks (see below).
Drug susceptibility testing (DST) has become a critical need in the era of epidemic drug resistance. There are a number of variables in DST. Historically, it was common to perform the culture, and when growth was observed and the species conﬁrmed, to take a subculture and perform DST. This “indirect” method required 6 to 8 weeks on average.
DST in the latter 20th century generally used the “proportionality” method. Equal aliquots were put on drug-free clear agar, and “control,” and others were placed on agar with various concentrations of the drugs. Then the number of colonies on the drug-containing media was compared with the control count.
More recently, techniques using liquid media containing different quantities of the relevant drugs determined the minimum inhibitory concentration (MIC) of the drugs. MIC techniques have either used the cut points derived from the proportionality method (see above) or have calculated presumed efﬁcacy by comparing the maximum concentration (Cmax) derived from pharmacokinetic studies with the MIC of a given drug.
However, clinical and public health issues have created pressure to identify MDR-TB (or XDR) in a much shorter period because of high rates of mortality among patients with AIDS and continued nosocomial transmission to health care workers and other patients on hospital units.
Two methods have been used to facilitate early recognition of resistance: accelerated microbiologic methods and molecular probes for chromosomal markers of resistance. The microscopic observation of drug susceptibility (MODS) technique uses multiple wells of control and drug-containing media. The wells are examined regularly for the presence of corded bacilli. This technique has been reported to yield results in as little as 7 to 10 days for multiple drugs.
By contrast, techniques to probe for chromosomal mutations known to be associated with resistance appear to be faster and simpler. Although the mutations related to resistance for many TB drugs are known, the one most reliable and clinically meaningful is that for rifamycin resistance, the RIF polymerase B (rpoB) mutation. Because resistance to the rifamycins is the keystone to MDR- or XDR-TB, using this mutation to triage patients appears very attractive.
TREATMENT OF TUBERCULOSIS
Modern regimens use four drugs: RIF, INH, pyrazinamide (PZA), and ethambutol (EMB). The current standard duration is 6 months. The RIF, INH, PZA, and EMB are given for the ﬁrst 2 months, and the RIF and INH are given for the last 4 months. Current guidelines, however, indicate that if the sputum culture taken after the ﬁrst 2 months is positive, the RIF and INH should be extended to a total of 9 months to lessen the risk of relapse.
Because nonadherence or noncompliance with the prescribed regimen was shown to be associated with delayed conversion to sputum negativity, higher rates of failure or relapse and—most importantly acquired drug resistance, directly observed therapy (DOT) has become the normative practice in the United States.
Under the typical DOT program, at the time of diagnosis, patients are served notice that they must make themselves regularly available for supervised treatment or will face conﬁnement. Patients cannot be compelled to take medications, but if they are recalcitrant, they can be placed under enforced isolation.
To facilitate supervised therapy, intermittent (less than daily) regimens may be used. One model, based on studies from Hong Kong, used thrice-weekly treatment throughout. Another, ﬁrst used in Denver, begins with 2 weeks of four drugs daily, switches to 6 weeks of four drugs twice weekly, and concludes with 18 weeks of RIF-INH twice weekly. Of note, the twice-weekly schedule has been found inadequate in the setting of AIDS; either thrice-weekly or daily (ﬁve times per week) is indicated.
MDR-TB or XDR-TB
As already noted, MDR (resistance at least to RIF and INH) and XDR [(resistance at least to RIF, INH, one of ﬂuoroquinolone agents, and one of the second-line injectables [amikacin, kanamycin, or capreomycin]) are threatening global TB control.
TB drug resistance is based on chromosomal mutations, not transferable resistance factors. The mutations are not induced by therapy; rather, failure to take an adequate number of drugs in correct dosages allows the resistant mutation to escape and by selection become the dominant strain. Drug-resistant epidemics are ampliﬁed by making drugs available without the infra- structure to ensure reliable therapy. After “primary” resistance is established in a patient, he or she can transmit the resistant strain to others (i.e., “transmitted” or “secondary” resistance).
Treatment of patients with highly resistant TB is challenging because the second-line drugs (SLDs) are less efﬁcacious, more toxic, and more expensive than the ﬁrst-line drugs. Plus the duration of therapy required to cure increases from 6 months to 24 months. Historically, drug-resistant TB was observed to be less readily transmissible and less virulent (likely to progress to active disease) among contacts. However, over the past 20 years, Beijing strains have tended to replace the older European strains, and these organisms have been shown to be readily transmissible and highly virulent.
Optimal management of drug-resistant cases entails access to in vitro susceptibility testing, including SLDs, experience with the clinical nuances of SLDs and ideally access to resectional surgery. Among the agents used are the injectables (streptomycin, amikacin, kanamycin, or capreomycin), the ﬂuoroquinolones (cipro-, levo-, or moxiﬂoxacin), cycloserine, ethionamide, paraaminosalicylic acid (PAS), clofazimine, linezolid, clarithromycin, and amoxicillin/clavulanate. Comprehensive discussions of the treatment of highly resistant TB are beyond the scope of this chapter.
Treatment of Latent Tuberculosis Infection
As noted above, reactivation of latent foci or infection typically acquired months, years, or even decades earlier is the commonest pathway in the pathogenesis of TB. Recognizing this pattern, the United States Public Health Service conducted a series of trials to determine whether giving INH to those with latent TB would reduce their risk of developing active disease. Brieﬂy, these placebo-controlled trials, which involved roughly 72,000 diverse subjects, showed approximately 60% to 70% protection. In the group followed longest,
Alaskan villagers, the protection extended over 19 years. Because BCG vaccination would substantially confound the utility of the TST, the sole means of identifying latent infection, vaccination has not been used in the United States.
In the United States, treatment of latent TB infection (TLTI) focuses on identifying individuals or groups historically at high risk for TB and, based on reactivity to the TST or IGRA, administering “preventive therapy”: (1) contacts to cases of communicable TB who may be presumed to have been recently infected; (2) those with HIV infection or AIDS; (3) health care workers with recent conversion of their TST or IGRA; (4) immigrants from high-risk regions; (5) individuals with radiographic abnormalities consistent with healed TB; and (6) other immunocompromised subjects, including those with chronic renal failure, organ trans- plantation, recipients of tumor necrosis factor inhibitors, or other agents (including high-dose steroids or cytotoxic agents).
The current guidelines from the American Thoracic Society (ATS) and Centers for Disease Control and Prevention (CDC) offer three TLTI regimens: (1) INH for 9 months (ﬁrst choice), (2) INH for 6 months (acceptable), or (c) RIF for 4 months. The 9-month INH regimen is rated highest by evidence-based analysis; 6 months of treatment is deemed acceptable. The 4 months of RIF option is a reﬂection of the efﬁcacy of RIF in accelerated sterilization in the treatment of active disease and its superior activity in the murine model. TLTI with INH or RIF substantially lessens the number of viable bacilli in those with latent infection and thereby prevents (or delays) reactivation.
The side effects and toxicity of INH include rare but serious hepatitis. Roughly 10% to 20% of those started on INH will have modest, asymptomatic elevations in their hepatic transaminase values, more common with increased age. A total of 1% to 3% of those receiving INH may develop symptomatic hepatitis, which requires discontinuation. Lethal liver failure occurs with continued use of INH in the setting of worsening liver chemistries and symptoms. Current ATS and CDC guidelines indicate that symptom surveillance is generally sufﬁcient to prevent serious liver damage.
RIF is less likely to cause serious hepatitis and does not have the central nervous system effects (e.g., lethargy, decreased concentration) experienced by some patients taking INH. The major issue with RIF is the potential for drug interactions with a variety of agents, including antiretroviral (ARV) agents, oral contraceptives, warfarin, and a variety of other meds (see the Physicians’ Desk Reference). It should not be used in people with HIV/AIDS because of the risk of acquired monoresistance.
For contacts with MDR-TB cases, there is no unanimity for TLTI. A 6-month course of a ﬂuoroquinolone (levoﬂoxacin or moxiﬂoxacin) may be appropriate. There are in practical options for TLTI in contacts with XDR-TB.
Treating Tuberculosis in Persons with AIDS
There are four unique aspects of managing TB in persons with AIDS:
1. The clinical presentation may be atypical, including chest radiography. In the absence of cavitation (rare in AIDS), sputum smears are less readily smear positive. Diagnosis may be established by unusual modalities such as blood culture, lymph node biopsy, and culture or bone marrow biopsy and culture.
2. In persons with AIDS with advanced malnutrition and infectious enteritis, there may be suboptimal TB drug absorption; pharmacokinetic studies may be helpful.
3. There may be signiﬁcant drug interaction between the rifamycins and the ARV agents. In general, RIF has such a profound effect in accelerating the elimination of the ARVs that they cannot be used together. Instead, rifabutin, which has about 30% of the effect on the cytochrome P450 system, should be used. (See the CDC’s website for updates on TB therapy and ARVs.)
4. When ARV therapy is begun and the CD4 lymphocyte population increases, there may be an ampliﬁed immune response to the TB. This immune reconstitution inﬂammatory syndrome (IRIS) may result in organ-speciﬁc ﬂares or exaggerated constitutional symptoms. Recent experience indicates that despite this risk, ARV be begun within 2 to 3 weeks of commencing TB therapy. If serious problems arise, corticosteroids may ameliorate the IRIS.
PREVENTING NOSOCOMIAL TRANSMISSION OF TUBERCULOSIS
The MDR-TB epidemic was highly instructive regarding institutional safety. In retrospect, health care authorities had confused the early bactericidal effects of modern regimens with adequate infection control measures. With drug-susceptible TB, the number of viable bacilli in the sputum falls by two to three logs in
the ﬁrst week of treatment; additionally, the patients’ cough frequency decreases abruptly. By contrast, with MDR-TB, the bacillary population and cough do not diminish under standard therapy. In the New York City MDR-TB outbreak in the early 1990s, molecular epidemiology showed that 80% of the cases were nosocomially transmitted.
This recognition led to several protective measures: (1) placing all suspected cases in negative-pressure isolation rooms, (2) establishing six or more air changes per hour in these rooms, (3) having all health care workers potentially exposed to wear ﬁt-tested N95 respirators, and (4) using ultraviolet germicidal irradiation in patient rooms.
These measures appear to have signiﬁcantly reduced the risk of nosocomial transmission. Nonetheless, health care workers are still required to have annual testing, either TSTs or IGRAs.