Tuberculosis

Jonathan E. Golub

Jacqueline S. Coberly

Richard E. Chaisson

INTRODUCTION

The ancient Greeks called it phthisis, the Romans tabes, and the Hindus rajayakshma; in Victorian England, it was known as consumption.¹ All of these names referred to the wasting illness that is characteristic of the disease we now call tuberculosis (TB). Tuberculosis is a complex communicable disease of humans caused by the tubercle bacilli, a group of genetically related mycobacteria also known collectively as the Mycobacterium tuberculosis complex. The tubercle bacilli are a group of slow-growing mycobacteria that include M. tuberculosis, M. africanum, M. canettii, M. bovis, and M. microti. The first three members of this group are, at least to date, strictly human pathogens.

M. bovis causes illness in a variety of animals as well as in humans.², ³ and ⁴ Similarly, M. microti generally causes disease in rodents, but has been linked retrospectively to infections in llamas, ferrets, and cats. It has also been implicated as the cause of pulmonary tuberculosis in a small number of humans.⁵ With the advent of effective drug treatment in the 1950s and preventive therapy (or chemoprophylaxis, as it was then called) in the 1960s, many in the medical and public health communities, particularly in industrialized countries, assumed tuberculosis was conquered. This hubris led to several decades of neglect by the biomedical community, during which control efforts were ignored or deliberately weakened.⁶ Unfortunately, the economic, social, and public health factors that foster the propagation of tuberculosis had not been eliminated, not even from the industrialized nations. Thus, in the 1980s and 1990s, as the deterioration of control programs coincided with the burgeoning epidemic of human immunodeficiency virus (HIV), tuberculosis rebounded. Numerous outbreaks were seen in the larger cities of the United States that had the highest incidence of HIV.⁷ More serious, however, was the fact that in some areas of the developing world where tuberculosis and HIV are both endemic, the incidence of tuberculosis doubled, and healthcare facilities became overwhelmed by the dual epidemic.⁶

The World Health Organization (WHO) declared tuberculosis to be a global emergency in 1993.⁶ Eighteen years later, considerable progress has been made in managing this infection, but tuberculosis remains a leading cause of premature death in young adults around the world.⁸ Roughly one third of the world’s inhabitants are latently infected with M. tuberculosis. WHO has estimated that 8.8 million people developed tuberculosis and 1.45 million died from it in 2010, representing a significant decline from the 9.4 million cases reported in 2009.⁸ Efforts to promote tuberculosis control have more than doubled since 1993, and the tuberculosis incidence rate has continued to decline annually since 2002 when it peaked at 142 cases per 100,000 population. Moreover, TB mortality has fallen by more than a third since 1990 and all WHO regions, other than the African Region, are on track to halve their 1990 mortality rates by 2015. Control efforts, highlighted by the DOTS (directly observed therapy, short course)/Stop TB Strategy, have saved almost 7 million lives by successfully treating 46 million of 55 million TB patients treated in DOTS programs.⁸ The global success has been driven to a significant extent by China’s overwhelming achievements since 1990. That country has seen an annual 3.4% decline in its TB incidence, a halving of prevalence, and an 80% reduction in TB mortality.⁸

Despite this success, China and India account for one third of the world’s TB cases, with the primary drivers of the epidemic in these countries being social determinants such as tobacco exposure, diabetes, malnutrition, alcohol abuse, and indoor air pollution.⁹ Meanwhile, in Africa, the high prevalence of
latent TB infection and HIV infection, a high prevalence of multidrug-resistant TB, and the complex epidemiology and natural history of tuberculosis continue to make control of the disease on this continent particularly challenging.¹⁰

THE ORGANISM

The family Mycobacteriaceae, of the order Actinomycetales, is composed of a number of slow-growing, acid-fast bacilli. Most are saprophytes— useful inhabitants of soil and water that fix nitrogen and help degrade organic material. Some are pathogens in animals and occasionally cause opportunistic infection in humans.¹¹, ¹² Only four species are highly pathogenic in humans: Mycobacterium leprae (which causes leprosy) and three of the tubercle bacilli. The tubercle bacilli, or Mycobacterium tuberculosis complex, comprise a group of five closely related mycobacteria that cause tuberculosis (Table 18-1). M. tuberculosis, M. africanum, and M. bovis are the most common cause of human tuberculosis, although M. bovis is also known to cause disease in a variety of animal species. The other two members of the complex, M. canettii and M. microti, do cause tuberculosis in humans, albeit infrequently.¹, ³, ⁴ and ⁵, ¹² In fact, M. microti has been identified as a human pathogen only relatively recently.⁵ Although they vary widely by favored host, phenotype, and pathogenicity, the
bacteria that make up the M. tuberculosis complex share more than 90% of their genome and have identical 16S rRNA sequences.², ¹³ With the advent of the HIV epidemic, several other mycobacteria—most notably, M. avium complex (MAC)— have emerged as common opportunistic pathogens and, in people infected with HIV, cause illness that is clinically similar to disseminated tuberculosis.¹, ¹³

Table 18-1 Species of Mycobacteria

	Microbe	Reservoir	Clinical Manifestation
Always pathogenic in man	M. tuberculosis	Man	Pulmonary and disseminated TB
	M. bovis	Cattle, man	TB-like disease
	M. leprae	Man	Leprosy
	M. africanum	Man, monkey	Rarely TB-like pulmonary disease
Potentially pathogenic in man	M. avium complex	Soil, water, birds, fowl, swine, cattle, and environment	Disseminated and pulmonary TB-like disease
	M. canettii	Man, possibly others	Rarely TB-like pulmonary disease
	M. microti	Rodents, llamas, cats, ferrets, and possibly man	Rarely TB-like pulmonary disease
	M. kansasii	Water, cattle	TB-like disease
Uncommon or rarely pathogenic in man	M. genavense	Possibly man and pet birds	Blood-borne disease with AIDS
	M. haemophilum	Unknown	Subcutaneous nodules and ulcers primarily with AIDS
	M. malmoense	Environment, possibly others	Adults: TB-like pulmonary children: Lymphadenitis
	M. marinum	Fish, water	Skin infections
	M. scrofulaceum	Soil, water	Cervical lymphadenitis
	M. simiae	Monkey, water	TB-like pulmonary and disseminated disease with AIDS
	M. szulgai	Unknown	TB-like pulmonary disease
	M. ulcerans	Man, environment	Skin infections (Buruli ulcer)
	M. xenopi	Water, birds	TB-like pulmonary disease
Adapted from Brooks GF et al, editors. Jawetz, Melnick & Adelberg’s Medical Microbiology, 21st Edition. Stamford: Appleton & Lange: 1998; and Niemann S et al. (2000). Two cases of Mycobacterium microti-derived tuberculosis in HIV-negative immunnocomptent patients. Emerging Infectious Diseases 6(5): 539-42; and Pffyer, G et al (1998). Mycobacterium canettii, the smooth variant of M. tuberculosis, isolated from a Swiss patient exposed in Africa. Emerging Infectious Diseases 4(4): 631-4.

The bacteria that make up the M. tuberculosis complex are slender, slightly curved, rod-shaped bacteria averaging 4 by 0.3 µm in size.¹, ³, ¹³ M. tuberculosis is strictly aerobic; in contrast, M. bovis is microaerophilic and adapts more easily to nonpulmonary sites of infection. Like other mycobacteria, the tubercle bacilli have an unusual concentration of high-molecular-weight lipids in their cell wall, accounting for approximately 50% of their dry weight. This high lipid content makes these organisms hydrophobic and resistant to aqueous bactericidal agents and drying. It is also responsible for their acid-fast nature, a characteristic that is essentially synonymous with mycobacteria.³, ¹³

Figure 18-1 Circular map of the chromosome of M. tuberculosis H37Rv. The outer circle shows the scale in Mb, with 0 representing the origin of replication. The first ring from the exterior denotes the positions of stable RNA genes and the direct repeat region; the second ring inward shows the coding sequence by strand; the third ring depicts repetitive DNA; the fourth ring shows the positions of the PPE family members; the fifth ring shows the PE family members; and the sixth ring shows the positions of the PGR5 sequences. The figure was generated with software from DNASTAR. Reprinted by permission of Macmillan Publishers Ltd: Nature. Cole ST, et al. Deciphering the biology of mycobacterium tuberculosis from the complete genome sequence. June 11; 393(6685):538. Copyright © 1998.

Mycobacteria are slow growing and fastidious in culture. Indeed, because M. tuberculosis has a very long generation time (approximately 24 hours), culture is a slow process, often resulting in diagnostic delays and sometimes misdiagnosis.⁵ Traditionally mycobacteria are grown on solid, enriched media, where colonies appear 4 to 6 weeks after inoculation. They can also be grown in liquid culture, where they form characteristic strings that can be seen by light microscopy. Rapid liquid culture systems (e.g., BACTEC) have been adapted for use with mycobacteria and allow identification of organisms in as little as 9-16 days, depending on the concentration of microbes in the specimen being tested.³, ¹⁴ DNA probes speed speciation of organisms following growth. Alternatively, a number of biochemical tests can be used to speciate mycobacteria, though these approaches are time consuming.³, ¹⁵

In 1998, a consortium of scientists deciphered and published the genome map of the H37Rv strain of M. tuberculosis (Figure 18-1). The complete genome was revealed to be 4,411,529 base pairs
long and contains approximately 4000 genes.¹⁶ Since then, advances in genotyping technology have been rapidly applied to studies of the molecular epidemiology of the tubercle bacilli. Genotyping methods have been rapidly adopted, including restriction fragment-length polymorphism (RFLP), polymerase chain reaction (PCR)-based spoligotyping, and profiling of the mycobacterial interspersed repetitive units based on the number and size of the variable number tandem repeats in the genome (MIRU-VNTR). Most recently, the application of whole-genome sequencing analysis has been shown to be optimal, though this technique remains too costly for universal use. These methods vary in sensitivity and specificity, so some caution is needed when interpreting results.¹⁸

Genotyping studies are providing some interesting insights into the epidemiology of tuberculosis. For example, 15 years ago it was dogma that M. tuberculosis was a mutated form of M. bovis; the assumption being that around 7000-4000 BCE, when humans began domesticating animals, they were exposed to M. bovis, which, over time, mutated into a human pathogen.¹⁹ In a recent fingerprinting study, however, the genomes of 100 strains of M. tuberculosis complex were mapped and compared. The genetic lineage developed from these analyses provides evidence that human M. tuberculosis is not a mutation of M. bovis; rather, both bacteria diverged from a common ancestor long before either infected humans², ¹³ (Figure 18-2).

Figure 18-2 Scheme of the proposed evolutionary pathway of the tubercle bacilli illustrating successive loss of DNA in certain lineages (gray boxes). The scheme is based on the presence or absence of conserved deleted regions and on sequence polymorphisms in five selected genes. Note that the distances between certain branches may not correspond to actual phylogenetic differences calculated by other methods. The top five dark arrows indicate that strains are characterized by katG463. CTG (Leu), gyr A95 ACC (Thr), typical for group 1 organisms. The bottom seven arrows indicate the strains belong to group 2 characterized by katG463 CGG (Arg), gyrA95 ACC (Thr). The light gray arrow indicates that strains belong to group 3, characterized by katG463 CGG (Arg), gyrA95 AGC(Ser), as defined by Sreevatsan et al. Reprinted from Brosch, R, AS Pym, SV Gordon, ST Cole. The evolution of mycobacterial pathogenicity: clues from comparative genomics. Trends Microbiol. Sep;9(9):452-8. Copyright © 2001, with permission from Elsevier.

DNA fingerprinting is also being used with traditional field epidemiology to help link index and secondary cases and to distinguish active disease resulting from reactivation versus recent transmission. When distinguishing between reactivation and recent transmission, the presumption is that the genotype of cases due to reactivation will not match the genotype of other cases in the community because the infection was acquired at some distant point in the past. Conversely, in cases involving recent infection, the genotype of the organism should be shared with at least the index case and probably other cases in the community resulting from the same index case. Note, however, that the “orphan” isolates that do not share a genotype with any other organism from the community could still be related to other cases
in the community, but the link cannot be identified with traditional epidemiology; this can happen when the index case is sputum negative, or when the DNA fingerprint of the index is obtained for other reasons. Nevertheless, cluster analysis is fairly reliable and has provided new and interesting epidemiologic information about tuberculosis.

Before fingerprinting was commonly available, it was believed that most infections in immunosuppressed people in low-prevalence areas represented reactivation of latent infection. The populations at risk were often transient, and linking cases epidemiologically was difficult. More recently, DNA fingerprinting studies of community cohorts have identified some surprising clusters and helped guide the shoe-leather epidemiology needed to confirm the linkages suggested by fingerprint evidence.²⁰ Fingerprinting studies have also helped characterize the global distribution of genotypes, clarify patterns of transmission in communities, detect laboratory transmission, and prioritize control activities.²¹, ²², ²³, ²⁴, ²⁵ and ²⁶ A U.S. study helped dispel the dogma that TB can be transmitted only through prolonged contact, and that casual transmission frequently occurs.²⁷ DNA fingerprinting is also being used to investigate the origins of susceptibility to the tuberculosis bacilli and innate host immunity to it.¹⁸, ²⁸, ²⁹

HISTORY

Evolution of the Tubercle Bacilli

Mycobacteria are ancient organisms that probably first appeared more than 1 million years ago in soil and water and gradually adapted to animal hosts during the Paleolithic period^*.¹⁹, ³⁰ Until recently, it was hypothesized that when humans began domesticating cattle, they began to be exposed to M. bovis, which eventually mutated into the human pathogen M. tuberculosis. A recent comparison of the genotype of 100 tubercle bacilli showed, however, that M. bovis, M. tuberculosis, and the other members of the M. tuberculosis complex all diverged from a common ancestor through successive loss of DNA (Figure 18-2).²

Regardless of the evolutionary path of the tubercle bacilli, at some point in early prehistory tuberculosis became pathogenic for humans. Skeletal remains of Neolithic humans with deformities suggestive of tuberculosis have been found in Germany, France, Italy, Denmark, and Jordan and are dated from 8000 to 5000 BCE.¹, ³¹ It cannot be conclusively proved that these deformities are the result of infection with the tubercle bacillus, as the organism cannot be cultured from the bone. Nevertheless, the deformities are strongly suggestive of spinal tuberculosis or Pott’s disease.¹, ¹⁹

Traces of tuberculosis-like disease have also been identified in Egyptian mummies dating from 3500 to 400 BCE.³, ³⁰ One particularly well-preserved mummy of a 5-year-old boy from about 1000 BCE was discovered with the lungs intact. Again, the causative organism could not be conclusively identified, but a diagnosis of tuberculosis was made based on the observation of pleural adhesion, blood in the trachea, and the presence of acid-fast bacilli in the lung tissue.³ Egyptian artifacts from this period also begin to show people with the spinal deformities characteristic of Pott’s disease.³

Based on available evidence, tuberculosis appears to have initially been a sporadic disease in humans. This is difficult to verify, however, because relatively few large collections of skeletal remains of prehistoric human are available for examination. What is known with certainty is that as civilization developed and people began to gather together in ever increasing numbers, reports of tuberculosis became more numerous, leading to the establishment of tuberculosis as an endemic disease of humans by the beginning of the first millennium.¹, ¹⁹

Global Evolution of the Epidemic

By CE 100, tuberculosis was well established in the Mediterranean states and Western Europe. It remained relatively sporadic for centuries, until people began settling in larger communities.¹⁹ The advent of the Industrial Revolution and the great migrations to the cities that followed created an ideal environment for the spread of tuberculosis.³⁰, ³² In 1662, 1 of every 6 northern Europeans had tuberculosis; 100 years later, this figure had doubled.³ Tuberculosis was so common that most of the population became infected, and 25% of all deaths were attributed to tuberculosis.¹⁹ Thomas Sydenham was quoted in 1682 as saying, “Two thirds of those who die of chronic diseases … [are] … killed by phthisis.”³² The epidemic peaked in England in 1780s and in Western Europe by 1800.³⁰

Although sporadic, M. bovis infections were not uncommon in the largely agrarian societies of the New World. However, it was the English and European colonists who are mostly likely responsible for spreading M. tuberculosis throughout the world.

Tuberculosis came to North America on the Mayflower and was well established in the colonies by the early 1700s.¹, ³⁰ The epidemic passed through the United States in a wave, pushing south and west with the spread of industrialization, and by the early 1900s it was endemic in North America.³⁰ Tuberculosis was also spread to South America by colonists, but because the Spanish quarantined consumptives in the 1600s and 1700s, its introduction in that region was slowed.¹ Tuberculosis also spread into Eastern Europe, Asia, and Africa from Western Europe. The epidemics in Russia came in the late 1800s and those in Asia in the early 1900s. Tuberculosis was still largely unknown in Africa at the beginning of the 1900s and spread slowly through the interior with the colonizing Europeans.¹, ¹⁹

Herd immunity to tuberculosis develops slowly because of the long incubation period of the TB organism. Differences in the epidemiology in different parts of the world may be related to when tuberculosis was introduced to the population, with the most serious disease present in Africa where tuberculosis was introduced latest, leaving the least amount of time for herd immunity to develop.¹⁹, ³³

CLINICAL MANIFESTATIONS

As discussed in more detail later, tuberculosis begins with latent infection that can progress to active disease. Latent tuberculosis infection (LTBI) causes no symptoms, and most latently infected people are unaware that they harbor tubercle bacilli. Tuberculosis disease generally affects the lungs and respiratory tract, but can strike nearly any organ system in the body. Both primary and reactivation tuberculosis disease can result in pulmonary or extrapulmonary manifestations. In immunocompetent people, approximately 80% of tuberculosis is pulmonary, whereas extrapulmonary disease is less common. Extrapulmonary disease is much more common in immunodeficient individuals and children.³⁴, ³⁵ A small percentage of immunocompetent patients and many more immunodeficient ones develop both pulmonary and extrapulmonary tuberculosis.³⁵

The onset of active tuberculosis is insidious, and the symptoms can be nonspecific. Pulmonary disease causes symptoms ranging from very mild to severe and can present with productive cough with or without bloody sputum, fatigue, anorexia, weight loss, fever, sweating and/or chills, and chest pain.¹³, ³⁴, ³⁵ Extrapulmonary tuberculosis also causes fatigue and night sweats, but generally other symptoms specifically related to the affected organ system will be prominent.¹³, ³⁴, ³⁵ and ³⁶ The most frequent sites of extrapulmonary infection include the pleura, pericardium, larynx, lymph nodes, skeleton (particularly the spine), genitourinary tract, eyes, meninges, gastrointestinal tract, adrenal glands, and skin.³, ³⁷, ³⁸ and ³⁹ Systemic infection with tubercle bacilli occurs when hematogenous or lymphatic dissemination spreads the organism throughout the body, producing small nodules of infection in essentially every organ. Early researchers named disseminated disease miliary tuberculosis because they thought the tiny nodules resembled grains of millet, particularly when seen by chest radiography.¹³, ³⁵, ⁴⁰ Miliary tuberculosis is especially common in children and in people with immunosuppression.⁷, ³⁴, ⁴¹

DIAGNOSIS

Latent Tuberculosis Infection

Latent tuberculosis infection is asymptomatic; therefore, diagnosis is based on clinical tests that identify signs of infection or immunologic responses to tuberculosis antigens. In the past, tuberculosis infection was often diagnosed radiographically, with calcified lesions interpreted as evidence of infection.³, ⁴² This technique has been shown to lack both sensitivity and specificity and has now been abandoned. Serologic diagnosis of tuberculosis infection has been extensively investigated but also lacks sensitivity because the number of tubercle bacilli in latent tuberculosis is very small, and antibody responses are limited.⁴³ Identification of specific cellular immune responses through the induction of delayed-type hypersensitivity with a tuberculin skin test (TST) remains the most widely used method for diagnosing latent infection.⁴³, ⁴⁴

Purified protein derivative (PPD) of tuberculin is a solution prepared from cultures of tubercle bacilli.⁴⁵, ⁴⁶ It was developed by Robert Koch in 1890 and touted as a cure for tuberculosis. The curative value of tuberculin was soon disproved, but further studies showed that it could be used to identify people with tuberculosis infection.⁴², ⁴⁶, ⁴⁷ The Mantoux method of intracutaneous injection of a standardized dose of PPD has been widely validated, whereas other methods (e.g., the Tine test) have not. Individuals with LTBI develop a zone of induration when a standardized amount of tuberculin is injected intracutaneously, whereas uninfected people react minimally or not at all. The TST result is reported as the width of induration that surrounds the test site observed at a minimum of 48-72 hours following injection.⁴⁸ The test is usually applied to the volar surface of the forearm and read by a trained observer, who measures induration by
marking the edge of the hardened area of skin, either visually or manually, to demarcate the border, and recording the distance across in millimeters. Induration of the injection site is the result of a delayed-type hypersensitivity response in which activated T cells and macrophages migrate to the site of antigen injection and mount a localized cellular immune response. It is essential that at least 48 hours be allowed for this process to mature; earlier readings may produce falsely negative or positive results. Erythema of the site is nonspecific and, except in Japan, is not used to determine test results. Positive results remain measurable for more than one week in most instances.⁴⁹

The performance characteristics of the tuberculin test were standardized in the 1940s in an important study conducted by Carroll Palmer, in which varying concentrations of tuberculin were administered to a group of patients with active tuberculosis and to controls who were unlikely to have ever encountered tuberculosis.⁵⁰ The currently used standard dose of 5 tuberculin units of PPD-S was found to elicit a reaction of greater than 10 mm of induration in almost 98% of tuberculosis patients and in less than 5% of controls. A smaller dose of another preparation (i.e., 2 tuberculin units of RT-23 tuberculin) produces similar results.

Exposure to non-tuberculous mycobacteria may induce cross-reactivity to the tuberculin, which can result in a falsely positive reaction.⁴⁶ Similarly, vaccination against tuberculosis with bacillus Calmette-Guérin (BCG) vaccine can induce a response. The strength and durability of the reaction produced by BCG is less than with tuberculosis infection, and after three to five years it does not generally interfere with PPD testing.³⁴, ⁵¹, ⁵² Skin test results can also be falsely negative when the cellular immune system is impaired. People with active tuberculosis who are coinfected with HIV, have cancer, or have other immunosuppressive illnesses are frequently unable to mount a response to any skin test antigen, so tuberculin results in these people should be interpreted with some care.³⁶, ³⁷, ⁴⁶, ⁵⁰ Moreover, malnutrition and micronutrient deficiencies interfere with immunity and can inhibit response to tuberculin, as can acute viral infections such as measles.¹³, ³⁴, ⁴², ⁵³, ⁵⁴, ⁵⁵ and ⁵⁶ In addition, between 10% and 25% of people with active tuberculosis fail to respond to tuberculin.⁴⁶, ⁵⁷ The size of the tuberculin response is not, however, associated with the patient’s stage of disease, as the size of reaction in active cases and their close contacts is similar.³⁷, ⁴⁶

Table 18-2 Cut-Points for Positive Tuberculin Skin Test (TST)

Category	Induration = TST Positive
Coinfected with HIV Close contacts of known active case People with fibrotic changes consistent with prior tuberculosis People with organ transplants or other immunosuppression	≥5 mm
Recent immigrants (<5 yrs) Injection drug users (IDUs) Prolonged exposure to high-risk congregate settings Mycobacterial laboratory personnel People with clinical conditions that put them at increased risk Infant, children, and adolescent contacts exposed to adults in high-risk categories	≥10 mm
People with no known risk factors for TB, but recommend only targeted skin testing	≥15 mm
Adapted from Centers for Disease Control and Prevention, D. o. T. E. (2005). Tuberculin skin testing, document #250140.

In summary, the size of the reaction to tuberculin varies considerably depending on the individual’s exposure to the tubercle bacilli and other non-tuberculous mycobacteria, BCG vaccination history, and immune status.⁴⁶

Although the sensitivity and specificity of tuberculin testing can be evaluated in people with active tuberculosis, determining the accuracy of the test in people with latent infection is more difficult, as no gold standard for diagnosis exists. In studies of U.S. Navy recruits undertaken in the 1950s and 1960s, a tuberculin reaction of greater than 10 mm was strongly associated with the subsequent risk of tuberculosis. Nonetheless, because some people with negative tests develop tuberculosis and many people with positive tests do not, the interpretation of test results must be modified on the basis of clinical and epidemiologic knowledge. Perhaps more so than with most clinical tests, interpretation of the tuberculin test is highly dependent on prior probability of tuberculosis infection and clinical consequences of misreading the result.

As shown in Table 18-2, the reaction size that is considered positive varies depending on the clinical
status of the person being tested.⁴⁸ For example, people likely to have been recently exposed to tuberculosis, such as household contacts of an active infectious case, have a high prior probability of infection and the cut-off point for a positive test is reduced to 5 mm. Similarly, an HIV-infected person has a very high risk of developing active tuberculosis if infected, so a 5-mm reaction is considered positive because the consequences of misinterpreting the result are severe and because HIV infection suppresses the cellular immune response and can diminish the size of the reaction to PPD.⁴⁹ People who come from an area where tuberculosis is prevalent have a lower risk than household contacts but a higher risk than someone who lives in a low-incidence country; therefore, a 10-mm response is considered positive. For people with a low prior probability of tuberculosis exposure, the cut-off point for a positive test is 15 mm of induration.⁴⁸

In recent years, attempts to modernize the diagnosis of latent tuberculosis have focused on assays that detect the production of interferon-gamma (IFN-γ) by T cells in response to stimulation with mycobacterial antigens, collectively referred to as interferon-gamma release assays (IGRAs). These “indirect tests” do not directly detect the presence of actual TB bacilli, but rather identify an immune response that suggests recent or remote exposure to TB bacilli.⁵⁰, ⁵¹ Such immunodiagnostic assays promise to improve on the tuberculin test by eliminating skin test placement and reading errors, obviating the need for patients to return for interpretation of results, and standardizing the assessment of results with objective measurements in a laboratory. T cells obtained from individuals who have been infected with M. tuberculosis will produce IFN-γ when cocultivated with tuberculous antigens, whereas T cells from uninfected individuals will not. Detection of the produced IFN-γ is possible with either an enzyme-linked immunosorbent assay (ELISA) that measures soluble cytokines or an enzyme-linked immunospot assay (ELISPOT) that stains intracellular IFN-γ in T cells. Two commercial applications of this approach have been developed: an ELISA-based method called Quantiferon-Gold In-Tube (QGIT) and an ELISPOT assay called T SPOT TB. Early versions of these assays relied on PPD-S as the antigen and, therefore, were associated with false-positive results in people who had been vaccinated with BCG or who had non-tuberculous mycobacterial infections.⁵⁰ Second-and third-generation assays, now commercially available in a number of countries, rely on M. tuberculosis-specific antigens from a region of the mycobacterial genome known as RD-1 that is absent in BCG and most non-tuberculous mycobacteria.⁵¹ Studies of these tests show that they have excellent agreement with the tuberculin skin test, but appear to be more specific, as they are less likely to be positive in individuals with a history of BCG vaccination who are otherwise at low risk for tuberculosis infection.⁵¹, ⁵² QGIT is now recommended for determining LTBI status in the United States, and its utility has been verified in a variety of settings.⁵³ However, in low- and middle-income countries, these tests have not been endorsed by WHO, as insufficient data exist to prove their utility in these settings, particularly in those countries with high TB and/or HIV prevalence.⁵⁴

Surprisingly, IGRAs are no more likely to predict risk of developing active tuberculosis than the TST, though data available on follow up for IGRA studies pale in comparison to the wealth of data from TST follow up studies.⁵⁵ While a positive result for either test indicates a doubling of risk for active TB, IGRAs tend to produce fewer false-positive results, so they are a more efficient test in TB control programs that seek to implement preventive therapy.

As the literature for IGRAs has exploded over the past 10 years, several systematic reviews have assessed the assays’ potential usefulness in different settings and populations. These tests have been conclusively shown not to aid in the diagnosis of active TB disease.⁵⁶ Although they represent an enormous step forward in diagnosing latent TB, their use is limited by their high cost, exacting technical requirements (fresh blood must be put into cell culture within several hours of being collected), and lack of prospective validation. Nonetheless, IGRAs are a welcome addition to the diagnostic arsenal and will play a valuable role in both clinical practice and epidemiologic research in the coming years.

Active Tuberculosis Disease

In active tuberculosis, clinical signs and symptoms result from the replication of large numbers of tubercle bacilli and the ensuing inflammatory host response. Diagnosis of active tuberculosis is based on evaluation of epidemiologic assessment of tuberculosis risk, clinical findings and symptoms, and laboratory tests including chest radiographs, tuberculin skin tests, microscopic examination, and culture of tissues such as sputum or biopsy specimens.

The signs and symptoms of active tuberculosis are nonspecific and overlap with a number of other pulmonary and systemic diseases. Fever, sweats, and weight loss are prominent systemic findings and are usually of several weeks’ or months’ duration. Cough is a principal feature of pulmonary tuberculosis
and can be associated with sputum production or hemoptysis. The symptoms of extrapulmonary tuberculosis are highly variable and depend on the specific organ involved.

Chest radiographs are critically important in diagnosing pulmonary tuberculosis. Classically, patients with reactivation tuberculosis have upper-lobe cavitary infiltrates or involvement of the superior segments of the lower lobes, whereas patients with primary tuberculosis have mostly lower or mid-lung infiltrates and hilar adenopathy. Recent studies using molecular epidemiologic techniques, however, have shown that both primary and reactivation tuberculosis can present with chest X-ray findings that are classic for the other form of disease. In the setting of clinical symptoms consistent with tuberculosis and an abnormal chest radiograph, specific diagnostic tests for tuberculosis should be undertaken. With HIV infection and tuberculosis, however, the chest X-ray in some patients may be normal.

Diagnosis of tuberculosis is confirmed by identification of acid-fast bacilli by smear or by isolation of M. tuberculosis in cultures of sputum or other tissues. Zeihl-Nielsen staining is the standard method used globally for microscopic identification of acid-fast bacilli. In this process, a fixed smear is exposed to hot carbol fushin dye for 2 to 3 minutes, rinsed, and decolorized with a dilute acid-alcohol solution. Mycobacteria absorb the carbol fushin dye but resist decolorization because of the high lipid content in their cell walls—hence the name acid-fast bacilli.¹⁵ Fluorescent staining with auraminerhodamine is a more sensitive but more expensive technique. Unfortunately, the number of tubercle bacilli present in sputum may be very low, particularly in noncavitary disease, so direct microscopic observation is a fairly insensitive way to diagnose tuberculosis. Only approximately 60% of culture-confirmed cases of tuberculosis are smear positive. Also, positive microscopy simply proves the presence of acid-fast bacilli, which could include nonpathogenic species of mycobacteria or other acid-fast bacteria such as Nocardia.³⁶, ³⁷

Several types of tests, including nucleic acid amplification assays (e.g., PCR), mass spectrometry or gas-liquid chromatography for tuberculostearic acid, and immunoassays for mycobacterial antigens and antibodies, have been evaluated, but with mixed results. Nucleic acid amplification tests have been approved in the United States for the diagnosis of tuberculosis in patients with positive sputum smears. Both the positive and negative predictive value of these tests are high if the sputum smear is positive, but in patients with negative sputum smear the positive predictive value of nucleic acid amplification is only in the vicinity of 50%. In many parts of the world where tuberculosis is endemic, culture and radiography are unavailable or extremely limited. In these areas, diagnosis relies primarily on clinical history and microscopic examination of sputum. New, simple, rapid diagnostic techniques are desperately needed in these areas and would also be valuable in more industrialized areas. Research in this area is ongoing, and advances have recently been made.

Significant advancements in diagnosing TB have evolved over the past several years, particularly in diagnosing TB and multidrug-resistant TB simultaneously. Line probe assays (LPAs) are a strip test that can detect TB bacteria from smear-positive sputum specimens or culture isolates using PCR amplification, while detecting drug resistance to rifampicin and/or isoniazid simultaneously.⁵⁷ Two LPAs are currently on the market: one made by Innogenetics in Belgium (INNO-LiPA Rif.TB) and another produced by Hain Life Science GmbH in Germany (GenoType MTBDR). Rifampin resistance is detected through mutations in the rpoB gene, while isoniazid resistance is identified through mutations in the katG and inhA genes.⁵⁸, ⁵⁹ When assessing M. tuberculosis isolates from culture, the LPAs’ sensitivity and specificity are consistently high.⁶⁰ With clinical specimens (including sputum) and culture isolates, the sensitivity and specificity for rifampicin resistance are high for both LPAs across multiple subpopulations, at 98.1% and 98.7% for the Belgian and German tests, respectively.⁶¹ Accuracy for isoniazid was considerably lower (84.3%) and varied widely across groups, while specificity was high for all populations (99.5%).

More recently, the Xpert MTB/RIF assay has been called a game-changer in the TB diagnostic world. Xpert MTB/RIF is a fully automated system that allows the operator to perform sample processing, DNA amplification, and detection of M. tuberculosis and of rifampicin resistance in less than 2 hours time with very little hands-on operation.⁶² Boehme et al. evaluated Xpert in a multicountry study and showed that with a single test, 98% of patients with smear-positive and culture-positive TB were correctly identified, while the same proportion of rifampinresistant isololates were detected.⁶² Moreover, 72.5% of smear-negative, culture-positive patients were detected. This assay has created considerable excitement, and many studies investigating the full capacity of this assay are currently under way, including use of Xpert outside of the traditional laboratory setting because of its minimal operator and
laboratory requirements. The test is more costly than traditional diagnostics, so the cost-effectiveness in various settings needs to be evaluated. WHO recently recommended Xpert as the initial diagnostic test in individuals suspected of having HIV-associated TB and has published several documents designed to aid implementation of this diagnostic tool by TB programs.⁷³, ⁶⁴

Ultimately, the diagnosis of tuberculosis involves a synthesis of clinical and laboratory findings. The case definition of tuberculosis used for surveillance purposes accepts the diagnosis if there is a positive culture, a positive acid-fast smear with compatible clinical findings, or a characteristic illness with other evidence suggestive of tuberculosis and an appropriate response to anti-tuberculosis therapy.

THERAPY

History of Therapy

The history of tuberculosis therapy is divided into three eras: the pre-sanatorium, sanatorium, and chemotherapeutic eras. From earliest recognition of tuberculosis as a disease until the middle of the 1800s, therapy for tuberculosis was based on the prevailing medical dogma.⁶⁵ When ill airs were thought to cause tuberculosis, patients were told to move to mild, mountain, or seaside climates. When imbalance of bodily humors was thought to be the cause of all disease, bloodletting was recommended for tuberculosis. Rest or mild exercise and different variations in diet were also recommended at various times. Although most of these treatments did the patient no harm, they also did little to deter the progress of the infection.

In the 1850s, a number of physicians observed that a prolonged rest in quiet, mountainous, rural areas had cured their patients of tuberculosis, and the sanatorium movement was born. The underlying premise of this approach was that clean air combined with rest or mild exercise and good food would stimulate the body to heal itself. Consequently, patients were isolated in rural institutions built solely for the treatment of tuberculosis.⁴², ⁶⁵ The first sanatorium was established by Brehmer in 1854 in the mountains of Germany; as the idea took hold, additional sanatoria were built throughout Europe, the United States, and England.⁴², ⁶⁵ Isolation of tuberculosis cases in sanatoria, although perhaps no more beneficial to patients than extended rest at home, decreased the spread of tuberculosis in the community, contributing to the large decline in tuberculosis incidence seen in the late 1800s and early 1900s in Europe and the United States.⁶⁵ Developments in the basic sciences during the sanatoria movement also contributed to this decline. During this time Koch discovered the causative agent of tuberculosis, and radiographic technology and surgical techniques were developed that greatly enhanced physicians’ ability to diagnose and treat tuberculosis.

Unfortunately, sanatorium care had its limitations, and in the early 1900s tuberculosis was still a major cause of death. In 1938, Rich and Follis showed that sulfanilamide inhibited the growth of M. tuberculosis in guinea pigs, and the search for effective chemotherapeutic agents for tuberculosis began. Dapsone was tested against tuberculosis in 1940, and in 1943 streptomycin was found to have anti-tuberculosis action. The identification of other anti-tuberculosis drugs, including para-aminosalicylic acid (PAS) and isoniazid (INH), soon followed.⁶⁵

The tradition of randomized clinical trials has a prominent place in the history of tuberculosis research. The scarcity of streptomycin in the early 1940s led the British Medical Research Council to perform the first multicenter, randomized, controlled clinical trial to estimate the efficacy of streptomycin against a placebo.⁶⁶ This elegant trial showed the profound efficacy of streptomycin against the tubercle bacilli but also the limitations of single-drug therapy in the treatment of tuberculosis. More trials followed the first rapidly as new drugs were identified, each building on the information provided by earlier work. A series of trials over several decades proved the value of combination therapy for curing tuberculosis and preventing drug resistance: the efficacy of dual therapy with streptomycin and PAS⁶⁷, ⁶⁸; the efficacy of combined therapy using isoniazid⁶⁷; the utility of multidrug therapy in shortening the duration of tuberculosis treatment⁶⁸; the minimum treatment time needed for effective cure of tuberculosis⁶⁵, ⁶⁶ and ⁶⁷; the optimal drug combination for therapy⁶⁷, ⁶⁸, ⁶⁹, ⁷⁰, ⁷¹, ⁷² and ⁷³; the efficacy of intermittent (twice or thrice weekly) treatment⁷¹, ⁷⁴, ⁷⁵ and ⁷⁶; and the efficacy of treatment for tuberculosis in HIV-infected people.⁷⁷, ⁷⁸, ⁷⁹ and ⁸⁰

Current Therapy

The drugs most commonly used in treatment of tuberculosis today and their mode of action are shown in Table 18-3. Because the bacillary population in an infected person consists of actively growing, semi-dormant, and dormant mycobacteria,⁸¹ effective chemotherapy is complex. Some drugs that kill actively growing bacilli cannot kill those in the latent, resting phase. Drug treatment must, therefore,
continue for a minimum of six months to allow the majority of latent organisms to be exposed to the drugs during periods of metabolic activity and to be killed. Unfortunately, this long period of treatment also allows sufficient time for mutant bacilli to emerge that are resistant to the drug being used for treatment. When a single drug is used for treatment of tuberculosis, mutants resistant to that drug rapidly emerge and eventually become the predominant bacilli, and therapy fails. Use of at least two drugs to which the organisms are susceptible reduces the probability of developing drug-resistant microbes to essentially zero.

Table 18-3 First-Line Anti-Tuberculosis Drugs and Their Modes of Action

Agent	Activity	Toxicity
Isoniazid	Bactericidal	Liver, peripheral nerve, hypersensitivity
Rifampin	Bactericidal and sterilizing	Liver, gastrointestinal, discoloration of body fluids, nausea, hematological
Rifapentine	Bactericidal and sterilizing	Liver, gastrointestinal, discoloration of body fluids, nausea, hematological
Pyrazinamide	Sterilizing	Liver, hyperuricemia, gout, malaise, gastrointestinal
Ethambutol	Bacteriostatic (dose dependent)	Liver, optic neuritis, skin
Streptomycin	Bactericidal	Ototoxicity, kidneys
Data from AD Harris and AD Maher. TB/HIV: A Clinical Manual. Copyright 1996, World Health Organization; and RH Alford, Antimycobacterial Agents. Principles and Practice of Infectious Diseases, 3rd Ed., pp. 350-360 , GL Mandell et al. eds., © 1990, Churchill Livingstone.