Patrick A. Murphy
Epidemiology
Incidence Trends
In 1990–1992, the United States was experiencing an increase in tuberculosis incidence. The cause was the human immunodeficiency virus (HIV) epidemic, which had created a large number of susceptible hosts. In addition, the number of cases of tuberculosis (TB) caused by bacilli resistant to most or all standard antituberculosis drugs rapidly increased. These multidrug-resistant (MDR) bacilli were highly lethal for patients with HIV infection, and it was feared that they would also produce untreatable infections in the normal population. Because of these problems, there was a resurgence of interest in the disease, medical and hospital practices were changed, and there was intense followup of diagnosed cases by public health officers.
All this activity has paid off, and the current situation looks better (see American Thoracic Society,http://www.hopkinsbayview.org/PAMreferences). In 2003, there were 14,874 new cases of TB in the United States, well below the 1985 level of 22,201 and the 1992 level of 26,673. Furthermore, 52% of all the cases occurred in people who were born in other countries, and, of those, two-thirds were born in Haiti, Mexico, India, China, Vietnam, or the Philippines. It seems likely that most such cases were acquired abroad and imported. The number of cases in United States natives fell 44% between 1986 and 2003, whereas the incidence in immigrants rose 53%. More than half of all immigrants with TB present within 5 years after entry into the United States. MDR TB is still virtually 100% lethal in the HIV-infected person. However, the incidence of MDR TB nationwide has decreased markedly. In 1993 there were more than 400 cases (2.5% of total) of MDR TB in persons who had never previously been treated for TB. These were presumably primary infections with a resistant organism. In 2003 there were fewer than 120 such cases (0.8%).
This improvement in the national picture can be attributed to many causes. Skin testing for TB has come back as a diagnostic procedure and is mandatory for health care workers. Patients discovered by this method are promptly given isoniazid (INH) preventive therapy. Clinical suspicion for TB by doctors has escalated, and TB is being considered as a possible cause of almost any chest illness and almost any abnormal chest radiograph. Microbiologic methods for diagnosis have improved: polymerase chain reaction (PCR) allows rapid diagnosis of TB from the sputum, and metabolic inhibition tests allow early detection of drug-resistant bacilli. Four-drug therapy for TB has virtually replaced the old two- and three-drug regimens, and therapy is commonly enforced by directly observing the patient take his or her medicines. Finally, hospitals, clinics, and prisons have refined their ventilating systems to reduce airborne transmission of TB. These changes in practice are discussed later in the chapter.
A major improvement in the investigation of cases of TB is the Tuberculosis Genotyping Program funded by the Centers for Disease Control and Prevention (CDC) in 2004. One culture can be submitted for analysis from every case of TB diagnosed in the United States. It used to be thought that most clinical TB in the United States was because of the reactivation of latent TB in persons who were old or ill. Such infections were thought to be an individual, but not a public health problem. Genotyping has shown that 52% of all U.S. cases of TB were associated with at least one other clinical case, and that the average cluster size was six cases. Most of the clusters were in young people, mostly men, with known risk factors for TB such as homelessness, low socioeconomic status, substance abuse,
P.502
unemployment, and HIV infection. They often lived in places where transmission is known to occur, such as shelters and jails. Genotyping did confirm some of the old teachings. For example, a man with newly diagnosed TB, age older than 60 years, born abroad, with none of the risk factors mentioned above, is likely to have a unique isolate. Also, although it is possible to be reinfected with a new strain of TB, the incidence is very low. A study from Capetown, South Africa, with a very high TB transmission rate, showed that 16 of 698 cases of pulmonary TB who were treated and apparently cured subsequently developed a second episode of clinical TB. Of these, 12 were infected with a new strain, which was clearly different by genotyping when compared with the strain that caused the original infection (1). At least 11 of these patients with exogenous reinfection were HIV negative (the twelfth was not tested). Because of the low TB rate in the United States, it is thought that the incidence of reinfection here must be much less than 1%.
The relatively good news about TB in the United States is not applicable to the worldwide situation. There are approximately 8 million new cases of TB every year, with about 1.8 million deaths. The world incidence is rising at 1.8% per year, and in sub-Saharan Africa, it is rising at 6.4% per year. The main driver of the TB epidemic is HIV infection.
Human Immunodeficiency Virus and Tuberculosis
It is estimated that a normal adult newly infected with TB has approximately a 5% chance of developing clinical illness in the first 5 years after infection, and a total lifetime risk of approximately 10%. The other 90% of normal people infected with TB have a positive skin test but never develop clinical illness. In contrast are people who are tuberculin positive as a result of a long-standing infection and who then acquire HIV infection. Such people have an annual incidence of clinical disease of 8% to 10%, with essentially a 100% lifetime incidence (2). Even more devastating are the consequences of being HIV positive and then becoming infected with TB. Because HIV seropositivity is often unknown to or concealed by the HIV-infected person, precise estimates of annual risk in this situation are not available. However, there are enough epidemics of severe primary TB among HIV-infected people exposed to an index case of TB to make it clear that the risk must be considerably higher than the 8% to 10% noted for the situation in which TB is acquired first. The annual risk is almost certainly greater than 50% and may approach 100% (3).
There are two other reasons why HIV infection amplifies TB rates. The first is based on the intensity of exposure needed to become infected. TB is acquired by inhaling dried droplet nuclei, which typically contain one to three organisms. A normal person usually must inhale several hundred such nuclei before one of them successfully establishes a caseous focus. Most cases of TB in normal people occur after long-standing, intense exposure to an infected person living in the same house or working in daily close contact. By contrast, a person infected with HIV acquires TB after inhaling an average of less than 10 infected nuclei, meaning that HIV-infected people often acquire TB after casual contact with an infected person. The other reason why HIV infection amplifies TB rates is that the clinical disease is often severe, and many bacilli are coughed out. Each case of TB in an HIV-infected person is therefore more infectious for other people than is the average case of TB in an immunocompetent person (3).
No matter how immunosuppressed one is, one cannot develop TB infection unless one is exposed to the organism. HIV infection acquired by a young male homosexual in Iowa is highly unlikely to be complicated by TB because there is little chance that he will ever be exposed to a case of pulmonary TB.
Reactivation and Primary Tuberculosis
There are now two distinct epidemiologic profiles of clinical TB. Most sporadic symptomatic cases still arise as reactivation disease in a patient who was infected many years ago and was never treated or was inadequately treated. When the patient becomes old or develops an immunosuppressive illness, the disease reactivates. Such patients are infectious for others, and TB can develop in people closely exposed to them. The response to primary TB infection depends on the kind of person infected. Children are highly susceptible to TB and tend to have clinical symptoms shortly after primary infection. Primary infection in an adult usually causes no symptoms, but the newly infected person becomes tuberculin positive (4). As discussed earlier, primary TB in an HIV-infected person tends to be severe and progressive (3).
Causes
The cause of TB in an immunologically normal person residing in the United States is almost always Mycobacterium tuberculosis. Atypical mycobacteria, such as Mycobacterium kansasii and Mycobacterium avium-intracellulare, and certain fungi, such as Cryptococcus neoformansand Histoplasma capsulatum, may produce disease indistinguishable from TB and should be considered in the differential diagnosis. All of these diseases are more common and more severe in immunosuppressed persons. In foreign countries, bovine TB bacilli can be important infectious agents.
Drug Resistance
In populations who have never been treated with antituberculosis drugs, primary resistance of tubercle bacilli to
P.503
most drugs occurs at a low level (approximately 1% to 3%). Because the mechanisms of resistance are different for every drug, primary resistance to multiple drugs is uncommon. However, during the past 50 years many inner-city patients have been given antituberculosis drugs without close supervision. Such patients are likely to take drugs singly, because of real or perceived side effects from multiple drugs, and to discontinue therapy entirely after symptoms improve. Not only do the bacilli in these patients become resistant to many antituberculosis drugs, but these resistant organisms may then infect other people. In such cases, the TB is “primary” because the infected person has a new infection. Nonetheless, the tubercle bacilli may be resistant to one or more of the standard antituberculosis drugs. Even in 1965, before the HIV epidemic, children in Brooklyn who were younger than 5 years of age and had primary TB had INH-resistant organisms 15% of the time (4).
The HIV epidemic has amplified the problem of drug-resistant tubercle bacilli that already existed in large cities. It is now not uncommon for tubercle bacilli to be resistant to all three standard antituberculosis drugs: INH, rifampin (RIF), and ethambutol (EMB). Some strains have been resistant to all known antituberculosis drugs. Such MDR strains are disastrous for the HIV-infected person because the disease cannot be controlled and is rapidly fatal. MDR strains are also a major public health hazard for immunocompetent people who are exposed to intense concentrations of infectious droplet nuclei. There is no reason to believe that these organisms are any more virulent than average, and presumably a normal person infected with such strains would have a 10% lifetime risk of developing clinical disease. However, if disease developed, it could not be treated with the usual drugs.
Diagnosis
History
When symptomatic, TB almost always presents with signs and symptoms of weeks’ to months’ duration. Almost the only time it presents as acute disease is in rare cases of acute meningitis or TB pneumonia. These acute episodes occur when a tuberculous focus ruptures into the cerebrospinal fluid (CSF) or a bronchus, and large quantities of tuberculous pus are suddenly discharged. The history should be directed toward both defining the symptom complex and determining possible exposure to known sources of disease.
Because TB has multiple presentations, chronic unexplained symptoms in any patient should be considered suspicious. Weight loss (documented over a defined period), fever (particularly in the late evenings), night sweats (to be differentiated from environmentally induced sweats), decreased appetite, and the loss of a sense of well-being are the most important nonspecific symptoms. Persistent cough (usually with sputum production), hemoptysis, and pleuritic chest pain are more specific findings suggestive of pulmonary involvement.
It is important to know whether the patient has previously had TB, has previously been skin-tested for TB (and if so, when and what the results were), and when the patient has had previous chest films (and where they can be obtained).
Possibly significant history also includes the patient's country of origin, any family member or close friend with known TB, any person in school or at work with known disease, and any recent history of travel to a country where TB is common.
Because extrapulmonary TB can occur in any organ (e.g., pleura, lymph nodes, endometrium, kidneys, ureters, bones and joints, skin, meninges, small intestine, peritoneum) or as a disseminated (miliary) form, chronic symptoms and signs in any organ must raise the consideration of TB.
Physical Examination
The physical examination may be entirely normal, even with obvious evidence of pulmonary disease on the chest radiograph. The following findings, when present, may be of considerable help in suggesting the diagnosis: rales localized to the upper posterior chest or auscultatory evidence of pulmonary cavitation (bronchovesicular breathing and whispered pectoriloquy), evidence of pleural effusion, supraclavicular and infraclavicular retraction, lymphadenopathy, evidence of weight loss, and fever. Although rare in the United States, large, matted, nontender cervical lymph nodes (at times with draining sinuses) are almost diagnostic of scrofula, a form of tuberculous adenitis seen primarily in children (that also may be caused by atypical mycobacteria).
Tuberculin Skin Tests
The standardized skin test (5) for evidence of tuberculous infection uses 5 units of Tween-stabilized intermediate-strength purified protein derivative (PPD), which is injected intradermally on the volar skin of the forearm. The use of control tests with “anergy panel” antigens has been abandoned. The PPD test requires intradermal injection, which is a highly skilled procedure. Tine tests and automated methods of introducing antigen into the skin such as the Heaf gun can be used in population surveys. Positive results elicited with these methods should be confirmed with the standard PPD. The PPD reaction should be read at 48 hours. A practical method for determining the diameter of the indurated area is the ballpoint pen method: a line is drawn from a point 1 to 2 cm away from the margin
P.504
of a positive reaction; when the pen tip reaches the margin of the indurated area, definite resistance is felt; this is repeated on the opposite side, and the diameter of the indurated reaction is measured.
The diameter required for a positive test varies according to circumstances.
Cases in which More than 5 mm is Regarded as Positive
Cases in which More than 10 mm is Regarded as Positive
Cases in which More than 15 mm is Regarded as Positive
A PPD conversion requires an increase of 10 mm within 2 years for patients of any age. Note that a change from 8 to 12 mm is not a PPD conversion. The criteria are not foolproof and should be applied with common sense. In particular, it has been suggested that even 2 mm induration may be significant in an HIV-infected patient (6).
The tuberculin skin test is an excellent way of diagnosing recent tuberculous infection in children or young adults. Such people are rarely tuberculin positive in the absence of recently acquired tuberculous infection. The test is also useful for investigating family members or health care personnel who have had known exposure to patients with infectious TB. Because it is neither specific nor sensitive enough to guide a clinical decision, the tuberculin test is not useful for the evaluation of patients with nonspecific symptoms that may be caused by TB. In some populations of elderly patients, up to 50% may be tuberculin positive as a result of long-standing infection that is inactive and has nothing to do with the present illness. Fully 20% of patients who do have active clinical TB are PPD negative. When one adds in other clinical variables such as tumors, steroid therapy, radiation therapy, and organ transplantation, a negative PPD test is often uninterpretable. Also, a negative PPD test may be harmful if it is used to rule out the diagnosis of TB in a patient whose clinical state may be caused by active disease.
A person with a known positive tuberculin skin test should not have it repeated. In such patients, there is a risk of producing a very strong positive reaction characterized by tender induration, axillary adenopathy, temperature elevation, and sloughing of the epidermis after 1 week. If a patient develops this complication, it should be treated with a sterile gauze dressing impregnated with a topical steroid, such as 0.1% triamcinolone.
The CDC recommends that previous Bacillus Calmette-Guérin(BCG)vaccination be disregarded when interpreting the response to a PPD (5). The reason is that a positive PPD reaction induced by BCG tends to wane with time. There is no reliable way of determining whether a positive PPD reaction is caused by real TB or by the BCG vaccine. There is agreement that when a person is PPD tested for the first time, very pronounced reactions (greater than 20 mm) are unlikely to be a result of BCG vaccination. On the other hand, recall of PPD sensitivity induced by repetitive PPD testing in persons vaccinated with BCG is commonly seen. Necrotic tuberculin reactions are particularly common in people who were vaccinated with BCG in childhood, become health care workers, and must be PPD tested every 6 or 12 months. If the vaccination was done 20 to 30 years earlier, the first tuberculin test is commonly negative. However, with repeated testing the immunologic memory revives and the reaction becomes progressively more intense. Not only does this raise unwarranted fears of tuberculous infection, but the reactions are very painful. The CDC regulations do not cover this situation; however, it is prudent not to do more than one tuberculin test in anyone previously vaccinated with BCG.
Persons with a remote tuberculous or atypical mycobacterial infection who have become skin-test negative may, on repeat annual skin testing, develop a positive response because the repeated exposure to antigen reinvigorates the immune response (booster effect) (7). Repeated
P.505
testing with tuberculin skin test antigen will not induce a positive test in an uninfected person. The booster phenomenon may mimic PPD conversion in persons who are repeatedly tested, such as elderly people, in whom loss of PPD reactivity is common, and people such as hospital employees who may be skin tested frequently. The booster effect is detected by administering a second tuberculin test 2 weeks after the first test in persons who initially have a negative response. If the second response is positive, these individuals can be said to have had past infection but are not considered to have a recently acquired infection.
Current practice in hospitals is that all employees not known to be PPD positive are tested on entry. If they are PPD negative on entry and have contact with patients, they are retested annually, or more often if the hospital is in an area where TB is common.
Laboratory Evaluation
Chest Radiograph
Both posteroanterior and lateral views should be obtained. If the standard radiograph is normal but the patient has strong clinical evidence for TB, an apical lordotic view or a computed tomography scan of the chest should also be obtained. The radiologic findings typical of TB (e.g., apical scarring, hilar adenopathy with peripheral infiltrate, upper lobe cavitation, miliary infiltrate) are not specific; however, a negative chest film rules out pulmonary TB (with the rare exception of early miliary disease), making the chest film a very sensitive test.
Cultures and Smears of Sputum
Sputum for smear and culture should be obtained at least three times. A positive sputum smear highly suggests TB (but is not absolutely diagnostic because of the possibility of atypical infection or of contamination), and a positive culture is diagnostic. If sputum is difficult to obtain, one can obtain assisted sputum after the patient has inhaled hypertonic saline aerosol. This procedure is obviously dangerous to personnel and should be done only in a special room with proper precautions for infection control. If clinical features suggest infection outside the lung, smears and cultures of other body fluids such as urine and CSF are appropriate.
PCR tests for M. tuberculosis are approved by the U.S. Food and Drug Administration (FDA) for direct use on sputum. However, the tests are subject to errors due to the presence of inhibitors, and most clinical laboratories are not able to routinely offer PCR tests directly on sputum.
Cultures usually become positive within 3 or 4 weeks. Any tubercle bacillus isolated from a patient should be tested for sensitivity to standard drugs, because it may be necessary to change treatment if the organism shows resistance. The state laboratory will be glad to do this if necessary. In parts of the world with very limited funds, it may be necessary to confine drug susceptibility tests to tubercle bacilli from patients who have failed standard therapy.
Miscellaneous Laboratory Tests
In patients with active TB, the hematocrit value may be normal or low. The anemia caused by TB is normochromic and normocytic, the so-called anemia of chronic disease (see Chapter 55). The white blood cell (WBC) count and differential count are usually normal; occasionally a monocytosis is seen in patients with severe disease. The urine should be tested routinely; if sterile pyuria is found, it is suggestive of renal TB and cultures should be sent for analysis.
Determinations of serum aminotransferases, alkaline phosphatase, and bilirubin may be helpful if disseminated disease or liver involvement is suspected. Also, these values provide baselines in case the patient develops hepatitis caused by antituberculosis drugs. Other procedures, such as thoracocentesis, lumbar puncture, and liver biopsy, are indicated only when specific organ involvement is suspected.
Presumptive Diagnosis
The presumptive diagnosis of active TB can be made when any of the following is found:
The diagnosis of active TB is confirmed by a positive culture from any body fluid or biopsy specimen. All patients with a presumptive or confirmed diagnosis of TB must be reported promptly to the appropriate local health department.
Diagnosis of Tuberculosis in HIV-Infected Patients
In 1993, pulmonary TB in an HIV-positive patient was designated as an acquired immunodeficiency syndrome (AIDS)–defining condition (seeChapter 39). When clinical TB appears before the development of other AIDS-indicator conditions, presentation occurs in a typical fashion, with pulmonary disease predominantly in the apices and often with cavitation. Fever, sweats, cough, anorexia, and wasting are common complaints. When TB appears after AIDS is already established, many patients have progressive primary disease, with extensive lower lobe
P.506
involvement, no cavitation, and hilar adenopathy (3). Extrapulmonary TB is more common in patients with AIDS and may involve lymph nodes, liver, brain, meninges, bone marrow, adrenals, and the genitourinary tract. Aspiration or biopsy of the suspected site of infection should be performed for acid-fast stain and culture. If acid-fast bacilli (AFB) are found in one of these specimens, treatment for M. tuberculosis should be initiated while awaiting culture results. In most of the world, acid-fast organisms seen on smears or biopsies are likely to be M. tuberculosis. In the low-incidence United States, M. avium-intracellulare is more common.
The diagnosis of tuberculosis is important because delays have real consequences in the form of clinical infections of contacts. A schoolboy in California developed TB that was not diagnosed for 12 months (8). He exposed 1,293 people. Among these, there were 12 cases of actual clinical tuberculosis, and 292 PPD conversions.
Management and Course
Treatment
When TB is diagnosed in association with systemic signs or symptoms, the patient should be treated for active disease. Patients with positive tuberculin reactions as the only manifestation of disease are more difficult to treat (see Isoniazid Treatment of Latent Tuberculosis Infection).
Isolation
Any patient with symptomatic TB who is hospitalized should be placed at once in respiratory isolation in a negative-pressure room. Isolation must be instituted as soon as the diagnosis is made or even strongly suspected. One can never prevent a patient walking into an emergency room and exposing a doctor, two or three nurses, a clerk, a radiology technician, and other patients to the risk of TB. But an undiagnosed patient who stays on a medical ward for 2 weeks may easily expose 200 people. Isolation should be maintained until the patient is clearly noninfectious (see Respiratory Isolation for Ambulatory Patients).
In some communities, no hospital has the required isolation rooms and MDR tubercle bacilli are uncommon or unknown. In these circumstances, a patient who has a home can be started on treatment and sent home. The logic is that everyone in the home has already been exposed. The patient is instructed not to leave the house for 2 weeks and not to allow any previously unexposed visitors for the same period. Therapy in the home should be directly observed (see Directly Observed Therapy), and the patient should be allowed to leave the house or have visitors only when clearly improving. It should be noted that this exception to the guidelines does not apply to patients who are homeless or are unlikely to follow directions because of alcoholism or drug addiction. It also does not apply in communities where MDR TB is common.
Drug Therapy for Tuberculosis
The imperatives of the MDR TB situation have resulted in a standard four-drug protocol for treatment of TB (9). There is little room for individual variation, and attempts to use nonstandard methods usually result in telephone calls from the local health department. The pressure is in favor of rendering the patient noninfectious in the shortest possible time. That way, fewer nurses are required to supervise the therapy.
The current standard initial regimen is as follows:
All four of these drugs are given for 2 months, and then the EMB and PZA are discontinued. INH and RIF are continued for another 4 months, making 6 months of treatment in all. When patients are treated on a twice-weekly schedule, the daily dose of drugs is larger (Table 34.1).
After a patient's culture results become available, sensitivities should be checked. If the bacilli are susceptible to all four drugs and if directly observed therapy is used, more than 98% of cases of TB are permanently cured (9).
The enthusiasm for four-drug supervised therapy should not obscure the fact that this regimen is not appropriate for everyone. It is possible to cure TB with two or three drugs, but it takes longer. Patients should receive special consideration if they cannot take all of the components of the standard regimen (10).
There are some special situations where the standard therapy for pulmonary tuberculosis should either be prolonged or changed. The recurrence rate after standard therapy is about 2%, and this value appears to be an irreducible minimum, because prolongation of the regimen does not reduce the recurrence rate. However, if a patient either has a pulmonary cavity, or if the sputum culture is positive after 2 months of therapy, the recurrence rate is about 6%. If there is a cavity and the sputum culture is positive at 2 months, then the recurrence rate is 22%. There is general agreement that such patients should be treated for longer, perhaps for 9 months.
Drug Toxicity
Patients may develop a number of toxicities from antituberculosis medications (Table 34.1).
|
TABLE 34.1 Drugs for the Treatment of Mycobacterial Disease in Adults |
|||||||||||||||||||||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||||||||||||||||||||
P.507
Isoniazid
Hepatic toxicity is the most common adverse reaction; it occurs at a biochemical level in approximately 20% of patients who take the drug, and the incidence of toxicity increases with age and with excessive alcohol intake (11). Pregnant women are at increased risk for INH toxicity. Laboratory evidence of mild injury to the liver is not in itself a reason to stop the drug, however, because in most subjects the aminotransferase level returns to normal while the drug is being continued. If the patient
P.508
develops jaundice or develops fever with elevated liver enzymes, the drug should be stopped. The liver injury is usually reversible and heals without further therapy. However, in some patients (elderly men and particularly those with chronic alcohol-related liver disease), the liver injury can be severe and sometimes fatal. INH liver toxicity most often occurs early in therapy, so that the first 2 to 3 months are the most critical in the detection of adverse drug reactions. Specific guidelines for monitoring for INH hepatitis are discussed later (see Monitoring of Patients Taking INH). Sensory peripheral neuropathy is an uncommon complication of INH therapy that occurs only in patients on an inadequate diet; it can be prevented by taking 25 mg of pyridoxine every day.
Ethambutol
The most serious side effect of EMB is optic neuritis, which causes decreased visual acuity and inability to distinguish the color green. This problem was seen frequently when the drug was given at a dose of 25 mg/kg. It is extremely uncommon at the current recommended daily dose of 15 mg/kg.
Rifampin
Serious allergic complications of RIF therapy, including thrombocytopenia manifested by purpura, petechiae, and hematuria; acute renal failure; and a flu-like syndrome, occur in approximately 1% of patients and necessitate cessation of therapy. There is a modest increase in hepatic toxicity that may be additive to INH toxicity, so patients taking both drugs should be supervised closely. Patients should be warned that RIF will result in an orange-red color in secretions such as urine and saliva and may irreversibly stain contact lenses. RIF accelerates the metabolism of other drugs (Table 34.1) and may necessitate an increase in the dosage of these drugs. Interaction with HIV protease inhibitors is a major problem and requires expert advice.
Special Situations
Patients with nonpulmonary TB can be treated with the 6-month regimen just described. Exceptions are patients with TB meningitis or bone marrow or joint infection. These patients should receive 12 months of treatment (10).
Pregnant women are difficult to treat because of the need to consider the effects of prescribed drugs on the fetus. The standard recommendation, for areas where MDR TB is uncommon, is a three-drug regimen of INH, RIF, and EMB. There is worldwide experience with this regimen, and it is generally regarded as safe. PZA is a category C drug that could be added if necessary; streptomycin is category D (it may cause fetal deafness), and its use is discouraged. For other antituberculosis drugs, one should seek expert advice.
Patients with renal failure are difficult to treat because both EMB and PZA are excreted largely by the kidneys. If renal function is only moderately reduced, this situation can be dealt with by dosage reduction. However, if the patient is on hemodialysis, the calculations become a nightmare. Because the toxicity of EMB is blindness and that of PZA is severe hepatitis, errors are serious. Both INH and RIF can be given in full dosage to people with no renal function, and provided the organisms are sensitive to both antimicrobials, INH plus RIF given together for 9 months cures all forms of TB. It seems better to accept the longer treatment. If there is a real possibility of MDR TB, and a four-drug regimen is necessary, INH, RIF, PZA, and streptomycin might be the safest drugs to prescribe. The hepatotoxicity of PZA is generally reversible, and although ototoxicity of streptomycin is irreversible, the blood level of streptomycin can be measured.
Patients who have HIV infection but are infected with drug-sensitive tubercle bacilli have a good prognosis for cure of their TB. It is usual to treat for 9 months rather than 6 and to be meticulous about follow-up. There is good evidence that patients with HIV as well as TB have higher relapse rates than patients with TB who are immunologically normal (12).
Drug Resistance
If the bacilli prove to be resistant to either INH or RIF, treatment should be modified to a three- or four-drug regimen in which the bacilli are susceptible to all components. Such a regimen should be continued for at least 12 months, and perhaps 18 months in a case of extensive disease.
If the bacilli are resistant to both INH and RIF, the organism is by definition MDR. The patient should be referred for treatment to an expert, either at a university or at the state health department. The patient will be kept in isolation as long as bacilli are found on sputum culture. In order to treat MDR cases, one needs to know the sensitivities to all the drugs that might be useful. It is very important not to provide empiric regimens in the absence of such knowledge, because of the probability that the patient will be treated with only one or two effective drugs. Such treatment usually generates even more resistant bacilli. It is best to wait until the information is available, and pick at least three and preferably four drugs that have activity and can be tolerated. Old drugs such as streptomycin, ethionamide, cycloserine, para-amino salicyclic acid (PAS), and capreomycin may be combined with newer ones such as quinolones and linezolid. If the organism is resistant to rifampin, it is usually but not always resistant to rifampin relatives such as rifabutin and rifapentine. It is generally possible to find some regimen that works. Regimens of six to twelve drugs are not uncommon (13). If the response to the best available therapy remains poor, the ad-dition of γ-interferon infusions can be tried (14). If the
P.509
patient is HIV infected, the outlook is probably hopeless, but there are occasional survivors (15). Details about longitudinal management of patients with drug-resistant TB are given later (see Course in Treated Patients, Drug Resistance).
Additional Therapy
The patient may have symptoms that require special management, such as high fever and toxicity for which corticosteroids may be helpful, or a pleural effusion that needs draining. The patient should be encouraged to eat an adequate diet. If the patient is eating poorly, pyridoxine (25 mg per day) should be taken with INH to prevent peripheral neuropathy. Pyridoxine should also be taken routinely by pregnant patients and by patients with other diseases that can cause peripheral neuropathy (e.g., alcoholism, diabetes, end-stage renal disease).
Course in Treated Patients
Followup Schedule
Physicians in the United States are rarely expected or allowed to supervise the treatment of a patient with TB. Both death rates and recurrence rates are higher for patients treated by physicians (16), and directly observed therapy is so clearly superior for most patients that only completely reliable patients should be treated in a physician's office. Health departments will supervise therapy, test contacts and provide prophylactic INH if necessary, provide free drugs, and perform sputum cultures and radiographs as tests of cure (17). In rural areas of the United States and in other countries, the physician may have to supervise therapy for TB.
After treatment has been initiated, the patient should be seen or contacted at least once a month, chiefly to ensure drug compliance and to monitor for drug side effects. Sputum cultures should be obtained monthly for the first 3 months. At 3 months and between 6 months and 1 year, chest radiographs should be obtained. Sputum culture should be negative after 3 months of therapy, although occasionally nonculturable acid-fast organisms are seen on smear for longer periods. A test-of-cure culture should be done on all patients at 5 or 6 months. Because resolution of pulmonary infiltrates is often slow; the former practice of monthly chest radiography is not warranted. Chest radiographs are most helpful in excluding progression of disease and in documenting the patient's status when TB is cured.
If a patient receiving the standard four-drug therapy still has positive sputum cultures at 3 months, it is worth measuring the serum RIF level 2 hours after RIF has been given. The level should be 8 to 24 µg/mL. Occasionally a patient has a serum RIF level lower than 5 µg/mL, even as low as 1 µg/mL. If such patients are given enough RIF to get the serum RIF into the therapeutic range, the TB improves at a normal rate (18).
If the patient does not have HIV infection, has sensitive bacilli, and follows a recommended and supervised regimen for the prescribed time, then relapse is so rare (less than 2%) that postcure followup is not necessary. However, if the patient has HIV, the relapse rate is higher, 2% to 8%. In such cases, TB followup is worthwhile and can be integrated with HIV treatment (12).
Usual Response
Patients with active TB who comply with therapy have an excellent prognosis. Exceptions are patients with organisms resistant to the usual antituberculosis drugs and patients who develop adverse effects from the antituberculosis therapy. The patient should show some symptomatic improvement within 1 week after initiation of antituberculosis therapy. Improvement is usually indicated by an increased sense of well-being, an increase in appetite, and a decrease in cough, fever, and night sweats; temperature should be normal within 14 days after treatment initiation. Most patients are back to their usual state of health in 1 to 2 months.
The improvement is caused by the prompt antibacterial effects of the drugs, which lead to a decrease in the inflammatory response of the host. After 2 or 3 weeks of therapy, the patient who is improving clinically (decreased cough, absence of fever, improved appetite) can be considered noninfectious. If the patient is known to be infected with an MDR organism, three sputum samples must be negative on smear before the patient can be released from isolation.
Respiratory Isolation for Ambulatory Patients
Health care workers who are the first points of contact in facilities serving patients at risk for TB should be trained to ask questions to detect patients who should be isolated because of signs or symptoms suggestive of TB. Even clerks can ask questions about chronic respiratory symptoms and alert the doctor or nurse. Patients with signs or symptoms suggestive of TB should be evaluated promptly to minimize the time spent in ambulatory care areas. TB precautions should be applied while the diagnostic evaluation is being conducted.
TB precautions in the ambulatory care setting consist of placing the patient in a separate waiting area, apart from other patients and not in an open waiting area, ideally in a room meeting TB isolation requirements, and giving the patient a surgical mask and instruction to keep it on. Patients should also be given tissues and instructed to cover their mouths and noses when coughing or sneezing, if they must remove their mask to facilitate respiratory clearance.
Ventilation in ambulatory care areas serving patients at high risk for TB should be designed and maintained to
P.510
reduce the risk of TB transmission. General use areas (e.g., waiting rooms) and special areas (e.g., treatment or TB isolation rooms in ambulatory areas) should be ventilated in the same manner as described for similar inpatient areas. Enhanced general ventilation or use of air disinfection techniques, such as in-room recirculation of air through high-efficiency particulate air (HEPA) filters or ultraviolet irradiation of the upper room air, may be useful. Ambulatory care settings in which patients with TB are frequently seen should have a negative-pressure room.
In general, patients with suspected or confirmed active TB should be considered infectious if cough is present, they are undergoing cough-inducing procedures, sputum AFB smears are positive, they are not receiving chemotherapy or have just started chemotherapy, or they have a poor clinical or bacteriologic response to chemotherapy. A person with drug-susceptible TB who is receiving adequate chemotherapy and has had a significant clinical and bacteriologic response to therapy (reduction in cough, resolution of fever, and progressively decreasing quantity of bacilli on smear) is probably no longer infectious. There is good data that the number of live tubercle bacilli in sputum falls by 90% after 48 hours of effective therapy, and by 99% after 2 weeks.
Patients with active TB who need to be seen in a clinic should have appointments scheduled to avoid exposing HIV-infected or otherwise severely immunocompromised patients. This could be accomplished by setting aside certain times of the day for appointments for these patients or by having them seen in areas where immunocompromised patients are not treated.
Patients with infectious TB should have at least two negative sputum smears for AFB before being placed in indoor environments that are especially conducive to transmission (e.g., shelters for the homeless) or in settings where highly susceptible people (e.g., those with HIV infection) will be exposed (19).
Drug Resistance
The treatment of patients with drug-resistant tubercle bacilli has become common in certain U.S. cities such as New York and Miami. In other cities, drug resistance to even one antibiotic occurs less than 3% of the time, and MDR strains have rarely been observed. However, the population is so mobile that a patient who has acquired infection in one city may present for treatment in another.
Because MDR strains do not constitute the majority of strains in any city, it is customary to start treatment in all patients with a standard regimen. Exceptionally, the patient may give a clear history of exposure to a known case of MDR TB or may come from an epidemic situation where an MDR organism is known to be responsible. In these cases, the patient can be started at once on the best available regimen.
Patients infected with MDR strains may show a delayed clinical response during the first few weeks of therapy. Because the laboratory may take 8 to 10 weeks to provide sensitivity data on the original isolates, it can be difficult to detect this problem early. Newer tests based on firefly luciferase assays may permit detection of drug-resistant tubercle bacilli in as little as 8 hours, but such assays are not yet commercially available.
The cardinal rules for management of proven or suspected MDR TB are (a) never add one drug to a failing regimen (20); (b) refer the patient to the TB section of the local health department before any changes in regimen are tried; and (c) keep the patient in respiratory isolation until he or she is demonstrably cured. Control of MDR infection with chemotherapy is difficult. Treatment may require empiric regimens containing up to 12 drugs and lasting 18 to 36 months; at times, resectional surgery may be required (13). If a patient has MDR TB complicating HIV infection, the patient usually dies despite taking all available anti-tuberculosis drugs.
Patient's Role in Therapy
In most of the world, TB treatment is initiated and conducted entirely in the ambulatory setting. The patient's role in the successful treatment of TB, daily self-administration of drugs for a period of 6 or 9 months and return for regular followup visits, is crucial. Because poor compliance accounts for most therapeutic failures in the treatment of TB, the most important function of monthly visits is the assessment, reinforcement, and documentation of com-pliance.
The patient should be advised about the communicable nature of TB; this is particularly important until therapy has continued for at least 2 weeks. During those first 2 weeks, patients should avoid intimate contact with others and should cough into tissue, which then should be incinerated or disposed of in closed plastic bags. After 2 weeks, most patients can be considered noncontagious, and activities can be dictated solely by their sense of well-being.
Directly Observed Therapy
Directly observed therapy (DOT) is a response to the present epidemic of uncontrolled drug-resistant TB in some U.S. cities. The preferred DOT regimen is a four-drug, 6-month regimen. All drugs are given twice per week, either on Monday and Thursday or on Tuesday and Friday. In general, the doses of drugs are increased over those given daily, but for each drug there is a maximum dose, (Table 34.1). The patient must go to a treatment center and swallow the medication while a nurse is watching. In cities, it may be possible to have nurses take TB therapy into schools, factories, or prisons where large numbers of patients are congregated. Disabled patients can
P.511
receive supervised therapy in their homes. The regimen is modified as necessary by the results of sensitivity tests on the patient's tubercle bacilli. Patients who do not keep appointments are pursued into their homes by health department personnel and, if necessary, by the police. Some states have laws that permit the incarceration of uncooperative infectious TB patients, and in extreme cases they are applied.
There is no doubt that supervised therapy is effective (17). On the other hand, this approach is an affront to human dignity, and it is not always feasible. Vast numbers of uneducated and even illiterate patients have been treated successfully by traditional methods; most societies simply do not have funds to pay for supervisors; and in rural areas, patients would have to make long journeys to be treated. The problem is that lack of compliance is difficult to predict (see Chapter 4); even such traditional indicators as alcoholism, drug addiction, and psychosis correlate significantly but poorly (21).
Prevention of Tuberculosis
Case Detection among Known Contacts
An integral part of initiation of care in any patient with active TB in the United States is case reporting to the local health authority and investigation of contacts. This entails tuberculin testing of all household and intimate nonhousehold contacts and retesting of nonreactors in 3 months. Reactors are examined by chest radiograph and, if free of active disease, are given chemoprophylaxis with INH (see Isoniazid Treatment of Latent Tuberculosis Infection). With the exception of evaluation of family members, this type of investigation is usually difficult for a physician to carry out alone and should be done by the local city or county health department. Such departments have trained personnel who are available to visit homes and workplaces to detect cases among contacts.
Tuberculin Testing in Prevention
Ideally, the tuberculin skin test status of all people should be determined at some time in their early adult life (see earlier description of technique for skin testing). In almost all school-age children, screening for tuberculin positivity is coordinated with school health programs. In adult populations, a number of factors, such as urban residence, the presence of chronic disease, a history of residence in underdeveloped countries, and health care occupation, increase the importance of periodic tuberculin testing. This is particularly true for people for whom INH would be recommended if the PPD were positive.
Because of the increased risk of contact with unrecognized cases of TB, physicians and hospital personnel who work with populations that have an increased prevalence of TB have an increased chance of acquiring infection. Both for personal protection and because of the risk of transmitting TB to patients, these health care workers should have annual tuberculin testing and should take INH chemoprophylaxis if they convert from negative to positive.
Isoniazid Treatment of Latent Tuberculosis Infection
INH treatment for 6 to 12 months has been shown to prevent 55% to 90% of expected new cases of active TB among groups of people with latent infection who are at increased risk (19). On the basis of accumulated data, recommendations regarding whom to treat for latent TB infection and the duration of treatment have changed somewhat in recent years.
At present, INH prophylaxis (300 mg per day) is recommended for people in the following groups:
In the past, patients were not given INH prophylaxis if they were older than a certain age, variously given as 35 to 55 years. The current recommendation is that persons in the groups listed, who have a TB risk that may be 30 times that of a person who is merely PPD positive, should be treated at any age. However, older persons who are PPD positive and do not have any of these risk factors should be left untreated.
Duration of Treatment
Nine months of INH therapy is recommended for patients with HIV infection or other forms of immunosuppression. It is recommended that persons younger than 18 years of age also receive 9 months of therapy. Other infected persons should receive a minimum of 6 continuous months of therapy, and preferably 9 months. For patients at especially high risk of TB whose compliance is questionable, supervised therapy may be indicated. When resources
P.512
do not permit supervised daily therapy, INH may be given under supervision twice weekly at the dosage of 15 mg/kg (9).
Monitoring of Patients Taking INH
The guidelines recommended by the American Thoracic Society for monitoring patients taking INH are the following (19):
Monitoring by routine laboratory tests (e.g., aminotransferases, serum bilirubin, alkaline phosphatase) is not always useful in predicting hepatic disease in INH recipients and therefore is not recommended routinely. However, in evaluating signs and symptoms such tests are mandatory. Preventive therapy should be reinstituted only if biochemical studies are normal and signs and symptoms are absent.
Because it has been recognized that monthly monitoring may fail to detect an occasional patient with severe hepatitis, monthly assessment of aminotransferase levels is recommended for patients who are in the groups at the highest risk for development of INH hepatitis: those older than 35 years of age, daily drinkers, patients concomitantly taking other potentially hepatotoxic drugs, and patients with a history of liver disease. This measurement would detect the transient aminotransferase elevation that occurs in approximately 20% of subjects taking INH; a cutoff level, such as a concentration three or five times normal, is recommended as the criterion for discon-tinuing INH.
There is no established alternative to INH prophylaxis. If a patient needs prophylactic antituberculous therapy, but cannot tolerate INH, he/she is usually given rifampin and ethambutol for 6 months, but there is no literature supporting this practice.
Specific References
For annotated General References and resources related to this chapter, visit http://www.hopkinsbayview.org/PAMreferences.