William R. Jarvis
Tuberculosis (TB) is a major cause of morbidity and mortality throughout the world with an estimated 8 million people contracting disease and 3 million dying annually. Worldwide, the prevalence of TB is increasing, in part due to the human immunodeficiency virus (HIV) epidemic. In the mid-1980s, the incidence of TB reversed its downward trend in the United States and, possibly secondary to the HIV epidemic and increased immigration from countries with a high incidence of TB, began to increase [1,2,3,4,5,6]. In the 1990s, the increasing trend continued, hospital outbreaks occurred, and additional federal and state TB resources and funding were provided. Subsequently, the increasing trend has been reversed.
As Mycobacterium tuberculosis emerged in the late 1980s and early 1990s, an increase in the number of Mycobacterium tuberculosis outbreaks, particularly multidrug-resistant strains (MDR-TB; i.e., resistant to at least two first-line antituberculous agents), began to be reported in U.S. healthcare facilities [7].
Pathogen and Pathogenesis
TB is caused by bacteria in the M. tuberculosis complex, of which M. tuberculosis is the most important [2,4]. The primary mode of transmission is by airborne droplet nuclei, which may be produced when people with pulmonary or laryngeal tuberculosis cough, sneeze, speak, or sing. In addition, aerosols can be produced and transmission occurs during manipulation or irrigation of TB lesions or when tissues or secretions from a person with TB are being obtained or processed. The actual sizes of the airborne droplet nuclei are not known but are estimated to be between 1–5µ such particles are light enough to be carried long distances by ambient air currents. When such airborne droplet nuclei are inhaled, they may traverse the body's defenses and become implanted in the respiratory bronchiole or alveoli. Despite capture by macrophages, the bacilli may replicate within the cell or be carried directly or by macrophages through the lymphatics to regional lymph nodes or the bloodstream, or they may disseminate. Cell-mediated immunity usually limits further multiplication or spread, and lesions resolve. The risk of infection with M. tuberculosis varies, depending on the concentration of droplet nuclei in the air and the duration of exposure to the contaminated air. After inhalation, the bacilli remain viable; this condition of latent M. tuberculosis infection is asymptomatic and is not contagious. In immunocompetent individuals, a tuberculin skin test (TST) applied 2 to 10 weeks after initial infection is positive; this is secondary to hypersensitivity to M. tuberculosis cell wall components. In addition, in immunocompetent individuals, a blood assay for M. tuberculosis (BAMT) such as the QuantiFERON-TB Gold test (QFT-G) will be positive [8,9]. In ~10% of infected immunocompetent individuals, the organism begins to multiply and causes active disease in their lifetime. In ~5% of these individuals, this occurs within the first 1 to 2 years after infection, and in the remaining 5%, disease occurs >2 years after infection. Those with immunocompromising conditions are at a >10% risk for development of active TB annually.
Tuberculosis in the United States
From 1953 through 1984, the number of people in the United States each year with active TB decreased by ~6% per year, from 84,304 to 22,201 [3]. However, from 1985 through 1992, the reported number of people with TB increased by ~20% to 26,673 [4]. Thus, from 1985 through 1992, ~52,000 excess episodes of TB were reported above what would have been expected based on the 1953 to 1984 trend. In addition, the proportion of people infected with MDR-TB increased. Factors contributing to this increase
P.540
in TB include the HIV epidemic, the increase in the number of people immigrating from countries with a high TB prevalence, adverse social conditions, and inadequate infection control practices in healthcare facilities, all of which facilitate transmission. The combination of an increasing TB prevalence, HIV-infected people, and MDR-TB resulted in an increase in hospitalization of people with infectious TB and an increased risk of M. tuberculosis transmission in healthcare facilities to both patients (nosocomial acquisition) and healthcare workers (HCWs) (occupational acquisition). With increased emphasis in hospital- and community-based TB control programs, the reported incidence of TB decreased by 5.2% in 1993 and by another 3.7% in 1994 [5,6]. In the late 1990s and the early 2000s, there has been a reversal of the previous upward trend. In 2001, 15,989 active TB patients were reported to the Centers for Disease Control and Prevention (CDC) representing a 2% decrease from 2000 and a 40% decrease since 1992.
Of the 25,287 people reported with tuberculosis in the United States in 1993, 65% were male, 71.5% were racial and ethnic minorities, 38% were 25 to 44 years of age, and 29% were born in a country other than the United States. In 1993, 64% of the reported episodes of tuberculosis were reported from seven states: California, New York, Texas, Florida, Illinois, New Jersey, and Georgia. In 2001, 62% (9,943) of TB patients were men and 38% (6,045) were women. The rate of TB for men (7.1 per 100,000) was almost double that of women (4.2 per 100,000). The TB rates were dramatically higher for Asians/Pacific Islanders (32.7 per 100,000), African Americans (13.8 per 100,000), Hispanics (11.9 per 100,000), and American Indians/Alaska natives (11.0 per 100,000) than for whites (1.6 per 100,000). There were 4,796 African Americans reported as having TB disease; 3,357 whites; 4,001 Hispanics; 3,552 Asians/Pacific Islanders; and 233 American Indians/Alaskan natives. Factors in reversing the increasing trend in TB in the United States include the successful implementation of recommended TB control programs, including directly observed therapy and improved implementation of hospital infection control practices [10].
Diagnosis of Tuberculosis
Prevention of transmission of M. tuberculosis depends on a high index of suspicion of TB and knowledge of the clinical signs and symptoms and the appropriate diagnostic workup. The initial symptoms of TB may be nonspecific and insidious in onset and include fatigue, anorexia, fever, chills, myalgia, sweating, and weight loss. Pulmonary TB is the most common form and the one usually associated with transmission in healthcare facilities. In pulmonary TB, the nonspecific symptoms are accompanied by an insidious cough that gradually progresses over a period of weeks to months to a productive cough with mucopurulent sputum or hemoptysis. Hoarseness or a sore throat suggests possible laryngeal involvement; laryngeal TB usually is accompanied by pulmonary disease, large numbers of acid-fast bacilli (AFB) in sputum, and a high degree of contagion. The infectivity of the patient with TB depends on the site of infection, the presence of AFB-positive sputum, the number of organisms expelled into the air, the presence of pulmonary cavitary or endobronchial lesions, the duration of effective therapy, and whether the person remains in isolation or covers his or her mouth when coughing. Extrapulmonary TB without pulmonary or laryngeal involvement usually is not contagious. However, irrigation or manipulation of TB lesions can produce infectious aerosols and transmit M. tuberculosis. The appropriate diagnostic workup of a patient with signs or symptoms consistent with TB should include a chest radiograph. With initial infection, parenchymal infiltration with ipsilateral lymph node enlargement may be seen whereas in reactivation, apical or posterior segment lesions in the upper lobes are most common. In all patients, particularly those who are immunocompromised, such as those with HIV infection, the chest radiograph findings may range from normal to miliary in appearance [11]. In those with possible pulmonary TB, respiratory secretions should be obtained promptly for AFB smears and culture; in pediatric patients, gastric aspirates may be needed for those who do not cough. Specimens should be transported to the laboratory promptly and the most rapid laboratory diagnostic methods should be used including fluorescent microscopy and radiometric methods for species identification and drug susceptibility testing [12]. Because the AFB smear is the most rapid presumptive diagnostic test and provides an important assessment of the patient's infectivity, AFB smears using fluorescent microscopy should be promptly performed. The development of genetic probes facilitates the rapid identification of M. tuberculosis (seeChapter 10). The initial isolate from people with TB and subsequent isolates from any person not responding to appropriate treatment should be tested for anti-TB drug susceptibility. Anti-TB drug susceptibility results provide the clinician important information for choosing an effective therapeutic regime. Furthermore, the use of genetic probes and nucleic acid amplification techniques can further shorten the time required to determine the species identification. Data show that from 1992 to 1995, the proportion of U.S. hospital laboratories in which the recommended methods were used for AFB smears, primary culture, M. tuberculosis identification, and anti-TB drug susceptibility testing has increased [13].
During the initial evaluation, either a TST or QFT-G test should be performed. If a TST is performed, it should be applied to the surface of the patient's forearm using the Mantoux method and read at 48 to 72 hours. Despite the limitations of the TST (<100% sensitivity and specificity; positive predictive value correlates with the TB prevalence), it remains the most practical screening method for identifying people with M. tuberculosis infection
P.541
(i.e., latent TB infection [LTBI]) in most countries. Only the presence or absence of and the extent of induration (not erythema) transverse to the long axis of the forearm should be measured.
The cut point used for defining a positive TST varies depending on the risk of M. tuberculosis exposure. For people with HIV infection, people in close contact with a person with infectious TB or with a chest radiograph consistent with pulmonary TB, a TST reaction of ≥5 mm is considered positive [14,15,16,17,18]. For people not meeting these criteria but who have other risk factors for TB, a reaction of ≥10 mm is considered positive. These factors include people born in countries with a high prevalence of TB; medically underserved populations with a high prevalence of TB; people with conditions predisposing to TB (e.g., gastrectomy, jejunoileal bypass, ≥10% below ideal body weight, chronic renal insufficiency, immunosuppressive therapy, some malignancies, diabetes mellitus, gastrectomy, or silicosis); residents of long-term care facilities; or other people identified as at high risk for TB. For others, including those who have received the bacille Calmette-Guerin (BCG) vaccine, a reaction of ≥5 mm is considered positive.
The QFT-Gold is a blood test in an ex vivo assay that measures the release of interferon-γ in whole blood in response to stimulation by antigens that are more specific for M. tuberculosis than tuberculin purified protein derivative. This test does not cause boosting when repeated. If one uses the QFT-Gold test rather than the TST, it should be used in all circumstances in which the TST is currently used, including sequential surveillance testing for infection control [8,9]. Negative tests should be interpreted with caution, particularly in those recently exposed to TB or who are immunocompromised. One must work closely with the laboratory to ensure that the blood specimen arrives at the qualified testing laboratory within 12 hours of being obtained (incubation must occur while the blood cells are viable). A single negative QFT-Gold test is sufficient evidence that the person probably is not infected with M. tuberculosis. A person with a positive test does not need to be retested. If doing serial QFT-Gold testing, conversion from negative to positive should be considered in a person with newly diagnosed TB infection.
Periodic TST or QFT-G tests are valuable methods for identifying new M. tuberculosis infection in those exposed to people with infectious TB. For people <35 years of age and most HCWs, an increase in TST induration of ≥10 mm within 2 years should be considered a TST conversion. For people ≥35 years or HCWs with infrequent TB exposure, an increase in TST induration of ≥15 mm in 2 years should be considered a TST conversion. Those with TST conversions should be promptly evaluated for active TB disease and placed on preventive therapy (if active disease is ruled out).
Repeated TST of noninfected individuals does not lead to hypersensitivity to tuberculin. However, in individuals with delayed hypersensitivity to tuberculin from either past infection with mycobacteria or BCG vaccination, reactivity may wane over time. In these individuals, the TST may stimulate the hypersensitivity and result in an initial negative TST followed by a subsequent positive TST when tested from 1 week to 1 year or longer after the initial TST (i.e., the “booster phenomenon”). To avoid the booster phenomenon and reduce the likelihood of interpreting a booster reaction as representing recent infection, TSTs should be conducted periodically with the first test being a two-step test. In the two-step test, the individual first has a TST. If negative, a second TST is performed 1 to 3 weeks later. If the second test is positive, it is probably a booster reaction. Other factors that can alter the individual's ability to react to tuberculin include (1) host factors, such as concomitant infection (e.g., viral or bacterial infections), metabolic abnormalities (e.g., chronic renal failure), nutritional factors (e.g., severe protein calorie malnutrition), lymphoid organ abnormalities (e.g., malignancy, sarcoidosis), immunosuppressive agents, age (e.g., newborns or elderly), and stress (e.g., surgery, burns); (2) factors related to the tuberculin used, such as improper storage, improper dilution, chemical degradation, or contamination; (3) factors related to the TST administration method (e.g., injection of too little antigen or delay in administration after drawing up the tuberculin); or (4) factors related to reading the TST, such as an inexperienced reader or misreading. In these situations, use of the QFT-G may be an attractive alternative, particularly in those who have received BCG or have tuberculin sensitivity [8,9].
Once a presumptive diagnosis of TB is established based on clinical findings plus or minus radiologic and laboratory findings, the patient should be placed on appropriate empiric therapy that can be modified once drug susceptibility results are available [14,15,16,17,18]. If the results of anti-TB susceptibility testing indicate resistance to agents being used or the patient fails to respond to the empiric therapy, the therapeutic regime should be modified.
Several factors influence whether M. tuberculosis is transmitted to either patients or HCWs in a hospital or other healthcare facility. These include the likelihood that they will be exposed to M. tuberculosis, that they will become infected, and that the infection will progress to disease. The risk of exposure in the healthcare setting is influenced by the number of infectious TB patients admitted and the infection control practices implemented in the healthcare facility, including the methods for identifying infectious patients, the type of isolation in which infectious patients are placed, the type of respiratory protection used by the HCW, and the environmental and engineering controls used. Several studies have shown that if the number of TB patients admitted or cared for is low, the risk of exposure to M. tuberculosis
P.542
is low [19,20,21,22,23]. The risk that HCWs or patients will become infected is influenced by the concentration of airborne droplet nuclei they are exposed to, the type and duration of the exposure, and the efficacy of the infection control measures implemented; although immunocompromising conditions increase the risk of disease given infection, they do not appear to increase the risk of infection. The risk that, once infected, a person will progress rapidly from infection to disease depends on the prevalence of conditions in the HCW or patient populations that decrease immunity (e.g., HIV infection, malignancy, or diabetes).
Surveillance for M. tuberculosis Infection/Disease in Healthcare Workers
There is no national surveillance system for M. tuberculosis infection/disease in U.S. HCWs. Before 1993, the national TB surveillance system did not collect occupational data on those reported with TB. In data reported since 1993, HCWs accounted for 0 to 6% of those reported with TB; however, this system does not provide further investigative data to determine whether the TB was occupationally or community acquired [5]. TST conversion data (LTBI) are not collected in this or any other national surveillance system. In studies conducted at individual institutions, the incidence of TB infection in HCWs ranged from 0.1% to 10.0% [17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44] (Table 32-1).
Several national surveys show that in hospitals with HCW TST programs, overall TST conversion rates range from 0.33% to 5.50% [19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53] (Table 32-1). Several studies have shown that HCWs originating from foreign countries or with a history of BCG have a higher prevalence of TST positivity [43,44,50,51,52,53]. Furthermore, in some regions of the country, HCWs have a high prevalence of TST positivity at the time of employment [14,15,16,42,50,51,52,53,54]. For these reasons, it has been recommended that a two-step TST be performed at the time of employment followed by periodic TST based on the risk of exposure [9,14,50,51,52,53,54,55,56] or a BAMT such as the QFT-Gold test [8,9].
Few studies have assessed the risk of TB disease in HCWs. Although U.S. HCWs in general are at low risk of TST conversion or TB, selected HCW groups—those with the closest contact with infectious TB patients or those working in hospitals with M. tuberculosis outbreaks—may be at greater risk [57]. In two surveys of physicians affiliated with medical schools, active TB appeared to be more common than in the community population [38,58]. Barrett-Conner found that 3.5% of physicians at one medical center had been treated for active TB, 75.0% of active disease occurred within 10 years after beginning medical school, and 62.0% occurred after infection after the beginning of medical school [33]. In the 1938 to 1981 graduates of the University of Illinois Medical School, active TB was more common than in the general population, and >66% of the episodes occurred ≤6 years after graduation [56]. Twelve studies, most based on either registries or questionnaires, have assessed the risk of TB disease in HCWs: in five, the risk was estimated to be greater than in the general population; in three, the risk was estimated as less than in the general population; and in four, the risk was not compared with that of the general population [23]. In more recent data from New York state, 2.5% of TB cases in 1994 and 2.0% of cases in 2002 were HCWs; 50.0% of HCWs with TB were foreign born [59].
Nosocomial M. tuberculosis Transmission
Before 1989, reports of nosocomial M. tuberculosis outbreaks were infrequent [58,59,60,61,62,63,64,65]. The sparsity of such reports may reflect either infrequent transmission or the failure to detect such transmission. Only 5%–10% of immunocompetent people exposed to an infectious TB patient contract TB in their lifetime, and such development would occur months to years after the exposure. Because the patient may not be readmitted to the same facility, the possibility of nosocomial acquisition might not be considered; such disease would not be linked to a prior hospitalization, and nosocomial acquisition would not be documented. Thus, unless there is a large group of immunocompetent or immunocompromised patients with newly diagnosed TB or patients infected with a strain of M. tuberculosis with an unusual antimicrobial susceptibility pattern, nosocomial transmission may not be suspected, detected, or reported. Hence, nosocomial patient-to-patient transmission of M. tuberculosis may occur more frequently than recognized or reported.
In contrast, occupational acquisition of M. tuberculosis infection and disease has been a well-documented risk to U.S. HCWs since the early 1930s. In the mid- to late 1930s, Israel et al. followed a cohort of 637 nursing students during their hospital training [63]. At study entry, 360 (57%) nurses were TST negative and 277 (43%) were TST positive. During their training, 100% of the TST-negative nursing students had TST conversions, and active TB developed in 68/637 (11%). Additional studies in the 1930s and 1940s documented high (79% to 85%) TST conversion rates among nurses [63,64,65,66]. Furthermore, during this period, studies documented that the risk of active TB disease among medical school students was >3 times higher than that of the general population [66]. Risk factors for disease among the medical students included exposure to unrecognized TB patients, treating patients in the men's TB ward, and presence in the autopsy room.
Despite these studies suggesting both nosocomial and occupational transmission of M. tuberculosis, efforts to introduce appropriate infection control interventions were slow. Many hospitals opened TB wards, sanitaria (or entire hospitals for TB patients) were common, and, in many hospitals, obtaining routine chest radiographs at the time of patient admission to identify the previously unrecognized infectious TB patient was initiated. Concomitantly, increased suspicion of TB by clinicians,
P.543
improvements in diagnostic and therapeutic modalities, and the identification of effective preventive anti-TB therapy resulted in a dramatic reduction in the incidence of TB in the United States from the mid-1940s until the mid-1980s [1].
|
TABLE 32-1 |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Despite the fact that numerous studies have documented that HCWs are at increased risk of occupational acquisition of M. tuberculosis, no formal surveillance system for HCW TST conversion or disease has been established. Thus, even today there are few specific data on the risk of acquisition of TB among HCWs in the United States and limited data on HCWs in the developing world where the risk may be even greater because of larger numbers of patients with TB; the presence of large, open wards; and minimal or absent infection control precautions. A questionnaire administered to graduates of the University of Illinois School of Medicine from 1938 to 1981 found that in most years, the incidence of TB in the graduates exceeded that in the general population [58]. In another survey of medical school–affiliated physicians in California, 3.5% of physicians had been treated for TB; 75.0% of active disease had occurred <10 years after beginning medical school [33]. However, data from the North Carolina TB control program
P.544
show that in 1983 and 1984, hospital personnel had similar or lower rates of TB than the general population [39]. More recently, surveys have shown that many hospitals do not have active HCW TST programs, that nurses and administrative personnel are more likely to be included in the TST program than are physicians, that HCW TST positivity rates at hire can be high and are related most likely to community TB prevalence, and that HCW (primarily nursing staff) TST conversion rates range from 0% to 27% [19,20,21,67,68].
Nosocomial M. tuberculosis in Human Immunodeficiency Virus–Infected Patients
TB is a common infection in HIV-infected patients [14,15,16,17,69,70] (see Chapter 42). Most of these episodes of TB are thought to be due to reactivation of latent M. tuberculosisinfection. In 1985, the incidence of TB began to rise in the United States; it is unclear how much of this rise was due to HIV-infected people who are at increased risk for both reactivation of latent infection and new, primary infection [1]. Nevertheless, between 1985 and 1992, ~52,000 excess episodes of TB occurred above what would have been expected in the United States had the downward trend in TB been maintained; much of this increase was thought to be in HIV-infected people [1]. Thus, the interaction of the HIV and TB outbreaks has resulted in a major public health challenge. In the late 1980s, nosocomial outbreaks of multidrug-resistant M. tuberculosis (MDR-TB) began to occur in U.S. hospitals [7]. In 1989, Dooley et al. investigated an outbreak of TB among patients on an HIV ward of a hospital in Puerto Rico [71]. They found that HIV ward patients with exposure to an infectious TB patient were 11 times more likely to contract TB than were HIV ward patients without such an exposure. In addition, nurses (i.e., those with patient care responsibilities) working on the HIV or internal medicine wards were significantly more likely to have a positive TST than clerical staff (i.e., those without direct patient care responsibilities) working on other wards at that hospital. This investigation documented that HIV-infected patients sharing a room with an HIV-infected patient with infectious TB were at increased risk of nosocomial acquisition of M. tuberculosis, that the incubation period was most consistent with primary infection, that the risk for development of active disease given exposure to an infectious TB patient was high, and that the duration from M. tuberculosis infection to disease (i.e., incubation period) was shortened in HIV-infected patients. These findings were consistent with those from another outbreak investigation that documented increased risk of nosocomial transmission of M. tuberculosis among hospitalized, HIV-infected patients [72].
Nosocomial M. tuberculosis Outbreaks
As with HCW TB infection or disease, there is no national surveillance or reporting system specifically for nosocomial M. tuberculosis outbreaks. Thus, our knowledge of such outbreaks is based on published reports. From 1960 through 1996, more than 20 nosocomial outbreaks in the United States were reported [2,38,48,61,62,71,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87]. All of the reported outbreaks involved adults and airborne transmission. These outbreaks primarily occurred at acute care general medical–surgical facilities, but outbreaks at a hospice and a health department clinic also have been reported [75,76]. In the healthcare facility, outbreaks have involved the emergency department, inpatient medical wards, inpatient HIV wards, inpatient renal transplant ward, inpatient prison ward, intensive care units, surgical or radiology suites, the outpatient HIV clinic, and an autopsy suite. Most of the reports in the 1960s to 1980s involved patient-to-HCW M. tuberculosis transmission, and patient-to-patient transmission was infrequently detected or reported. In contrast, most of the reports in the 1980s and 1990s involved immunocompromised patients and have included both patient-to-patient and patient-to-HCW M. tuberculosis transmission. Since the late 1990s, additional reports of nosocomial TB outbreaks have been reported from a large number of international settings.
Nosocomial Multidrug-Resistant M. tuberculosis Outbreaks
In the late 1980s and early 1990s, numerous nosocomial TB outbreaks occurred in the United States caused by strains of M. tuberculosis resistant to ≥2 anti-TB agents, most commonly isoniazid and rifampin. In 1990, Edlin et al. investigated an outbreak of MDR-TB among acquired immunodeficiency syndrome (AIDS) patients at a New York City hospital [79]. From 1989 through 1990, 18 AIDS patients had acquired infections with M. tuberculosis strains resistant to isoniazid and streptomycin; in contrast, only 3 patients had had such infections in the preceding 3 years. Compared with 30 AIDS patients with TB caused by drug-susceptible strains of M. tuberculosis, the 18 MDR-TB AIDS patients were significantly more likely to be homosexual men, to have had AIDS for a longer period, or to have been hospitalized at the outbreak hospital within the 6 months preceding diagnosis of their TB. Furthermore, the MDR-TB patients were significantly more likely than AIDS patients infected with drug-susceptible strains of M. tuberculosis to have been hospitalized during their exposure period on the same ward at the same time as another patient with infectious MDR-TB. Compared with a group of AIDS patients infected with drug-susceptible strains of M. tuberculosis and similar durations of hospitalization, the MDR-TB AIDS patients were more likely to have occupied rooms closer to the room of an infectious MDR-TB patient. Restriction fragment length polymorphism (RFLP) analysis of the strains of 16 MDR-TB AIDS patients showed that 13 had an identical pattern. Thus, this investigation provided both epidemiologic and laboratory data supporting nosocomial acquisition of M. tuberculosis by these MDR-TB AIDS patients. The attack rate was 21/346 (6.1%) among all AIDS
P.545
patients and 18/189 (9.5%) among AIDS patients hospitalized no more than two rooms away from an infectious MDR-TB AIDS patient's room. The estimated period from exposure to the development of active disease (i.e., incubation period) was estimated at 1.5 to 6.0 months. An environmental evaluation showed that only 1/16 patients' rooms had negative pressure. A TST survey conducted at the time of the investigation identified TST conversions (from documented negative to positive) in 11/60 (18.3%) HCWs; those with a follow-up TST >2 years before their baseline negative TST had a TST conversion rate of 9/31 (29.0%), whereas those with a follow-up TST of ≤2 years after their negative baseline had a TST conversion rate of 2/29 (6.9%) [88]. The M. tuberculosis strain from the one HCW with active MDR-TB had an RFLP pattern identical to that of the MDR-TB AIDS patient's strain.
At about the same time as the aforementioned outbreak, another MDR-TB outbreak, caused by a strain of M. tuberculosis resistant to isoniazid and rifampin, was occurring at a hospital in Florida [80,81,87]. From 1988 through 1990, 25 MDR-TB patients were identified among HIV-infected patients admitted to the hospital's HIV ward. Compared with HIV ward patients with TB caused by drug-susceptible strains, the MDR-TB patients were more likely to have had an opportunistic infection before being diagnosed with TB, to have been exposed to an AFB sputum smear-positive MDR-TB patient during their hospitalization preceding the diagnosis of TB, to have failed to respond to anti-TB therapy, or to have died. MDR-TB patients remained AFB sputum smear positive for a significantly longer proportion of their hospitalization than did HIV-infected patients with TB caused by drug-susceptible strains 375/860 (44%) vs. 197/1445 (14%) person-days. Exposure to AFB sputum smear-positive, culture-positive patients was significantly more likely to result in transmission of M. tuberculosis than was exposure to AFB sputum smear-negative, culture-positive patients. Exposures to infectious MDR-TB patients occurred both on the HIV ward and the HIV outpatient clinic. Patients who attended the HIV clinic to receive aerosolized pentamidine were at higher risk for development of MDR-TB than were those attending the clinic but not receiving aerosolized pentamidine. All of the available 13 MDR-TB patient M. tuberculosis isolates had one of two RFLP patterns. HCWs on the HIV ward and clinic were significantly more likely than HCWs on a comparison ward to have a TST conversion during the study period (13/39 vs. 0/15). There was a strong correlation between risk for HCW M. tuberculosisinfection and the number of days that an AFB sputum smear-positive MDR-TB patient was hospitalized on the HIV ward. Six of the 23 AFB isolation rooms were found to have positive pressure. In the HIV clinic, the aerosolized pentamidine administration rooms had positive pressure compared with the treatment room, which was positive pressure relative to the discharge waiting room; also, air from the patient treatment rooms was recirculated back into the clinic.
From 1990 through 1992, the CDC conducted nine nosocomial MDR-TB outbreak investigations [7,43,57,79,80,81,82,83,84,85,87,88,89] (Table 32-2). These outbreaks all occurred between 1988 and 1992 and involved 8 to 42 (median, 18) patients at each facility. Seven of nine outbreaks occurred in New York state, with six of these in New York City hospitals. In all of the outbreaks, the M. tuberculosis strain transmitted was resistant to isoniazid; in eight outbreaks, the strain also was resistant to rifampin. Depending on the outbreak, the strains were also resistant to streptomycin, ethambutol, ethionamide, kanamycin, or rifabutin. In two outbreaks, some of the infecting strains were resistant to seven anti-TB agents [43,57]. The proportion of patients in these outbreaks who had HIV infections ranged from 12.5% to 100.0%; most of these outbreaks occurred either on HIV wards or on wards where most of the patients admitted had HIV infection. Mortality ranged from 12.5% to 93.0% (median, 83.0%). The interval from TB diagnosis until death ranged from 4 to 16 (median, 4) weeks. Subsequent nosocomial TB outbreaks in the United States and throughout the world have had similar findings.
Risk Factors for Nosocomial Transmission of Drug-Susceptible or Multidrug-Resistant M. tuberculosis
A wide variety of factors was responsible for patient-to-patient or patient-to-HCW M. tuberculosis transmission in these outbreaks (Table 32-3). These included factors affecting the likelihood of exposure, of infection given exposure, and of active disease given infection. Patient risk factors included having HIV or AIDS, prior hospitalization at the outbreak hospital, the close proximity of AFB sputum smear-positive (i.e., infectious) TB patients and patients with HIV infection or AIDS, or exposure to an infectious TB patient either because of close proximity of the rooms or exposures outside the patient's room. Other risk factors included delayed recognition of TB in the patient because of nonclassic signs, symptoms, or chest radiograph; low index of suspicion by physicians; delayed recognition of infecting MDR-TB strains because of delayed laboratory identification or communication to the clinicians; and delayed institution of effective antituberculosis therapy. In addition, a number of inadequate infection control practices were identified, including delayed initiation of appropriate AFB isolation; isolation rooms without at least six air changes per hour, negative pressure, or air exhausted to the outside; failure to isolate TB patients until they were no longer infectious; lapses in AFB isolation such as allowing infectious patients in isolation to leave their rooms without wearing a mask or for nonmedical reasons (e.g., to attend group social events, to walk the halls, go to common bathrooms, or visit the lounge or television areas); leaving AFB isolation room doors open; inadequate duration of AFB isolation; and inadequate precautions during aerosol-generating procedures such as sputum
P.546
P.547
induction or aerosolization of pentamidine. Last, delay in institution of effective therapy contributed to prolonged infectiousness and increased the risk of transmission of MDR-TB strains.
|
TABLE 32-2 |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
TABLE 32-3 |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
In these and other MDR-TB outbreaks, the failure to rapidly identify and appropriately isolate infectious patients combined with the prolonged infectiousness of these patients secondary to delays in diagnosis and treatment led to exposure of other patients and HCWs. In each outbreak, the major risk factor for MDR-TB or for a TST conversion was exposure to an infectious MDR-TB patient. A low index of suspicion for TB in HIV-infected patients with pulmonary symptoms and the fact that many of the MDR-TB patients did not present with classic TB signs, symptoms, or chest radiographs (i.e., cavitary or miliary patterns) may have contributed to delayed identification of infectious MDR-TB patients [13,85,86]. In some instances, neither TSTs nor AFB sputum smears were performed on potentially infectious patients. Most MDR-TB patients had abnormal chest radiographs; however, they usually had interstitial patterns rather than classic miliary or cavitary patterns. When cultures were obtained, the results of the cultures or antituberculosis susceptibility testing were not available for a median of 7 weeks because the most rapid methods were not used; in some instances, the results were not available for 6 months. These delays in diagnosis resulted in delays in recognition of MDR-TB, in instituting effective anti-TB therapy, and in appropriate patient isolation; in turn, these lapses resulted in prolonged periods of exposure of infectious TB patients to other patients and HCWs. Although in most of the outbreaks infectious TB patients were the source, outbreaks have been associated with an HCW with infectious TB and with irrigation or manipulation of an undiagnosed tuberculous skin abscess or ulcer [57,75,76,78,85,86].
Molecular Typing of Multidrug-Resistant M. tuberculosis Outbreak Isolates
A critical element in confirming the epidemiologic evidence of nosocomial M. tuberculosis transmission in each of the MDR-TB outbreak investigations was the molecular typing of the infecting strains (see Chapter 10). In the early 1990s, a new method to type M. tuberculosis isolates, RFLP, was developed [90,91]. Thus, in each MDR-TB outbreak, available isolates from patients and HCWs with active TB disease were obtained and RFLP typed. Although occasionally a similar strain was identified at more than one facility, particularly in New York City where patients may have either had contact in or outside of the hospital, in most instances, one or more unique strains was documented to have been transmitted within each facility (see Table 32-2). In one instance, use of a newly developed polymerase chain reaction–based RFLP method was necessary because the source-case isolate was no longer viable [87,91]. The combination of epidemiologic and molecular typing data conclusively proved that MDR-TB strains were being transmitted within these facilities. Future application of RFLP and other molecular typing techniques should facilitate further elucidation of nosocomial and community transmission of M. tuberculosis. On the other-hand, depending only on RFLP typing to confirm an outbreak without supporting epidemiologic data can lead to erroneous conclusions.
P.548
Risk for Tuberculosis Infection in Healthcare Workers
In each outbreak, occupational acquisition of TB was suspected. However, conclusive documentation was difficult because in most instances, the HCWs had infection but not disease. Furthermore, at a number of the outbreak hospitals, the HCW TST program was inadequate; often, the HCWs had not had either a baseline two-step TST or a routine TST within the past 1 to 2 years [80,82,86,88]. At one MDR-TB outbreak hospital, 18% of HCWs in whom a TST was applied during the investigation had a positive TST; MDR-TB developed in one HCW, and the infecting isolate had the same RFLP pattern as that recovered from MDR-TB outbreak patients [79,88]. In one investigation, HCWs with or without TST conversions on the outbreak ward were similar in age, race, sex, duration of employment on the outbreak ward, occupation, and shift worked [80]. However, compared with HCWs on control wards where TB patients were not admitted, outbreak ward and clinic HCWs were at significantly greater risk for TST conversion. Risk for TST conversion was associated with exposure to MDR-TB patients who were AFB sputum smear positive rather than to drug-susceptible TB patients; HCW infection risk was significantly lower (although not zero) when they had exposure to AFB sputum smear-negative, culture-positive patients. In three MDR-TB outbreak hospitals, adequate TST data were available to assess HCW TST conversion risk; at these institutions, 22% to 50% of HCWs had TST conversions during the outbreak periods [79,80,82,83,84,85,86]. At the time of these investigations, at least 16 HCWs had contracted active MDR-TB; 7 were HIV positive, 7 were HIV negative, and 2 were of unknown HIV status [85]. Five of the 16 (31.2%) HCWs died; 80.0% were known to be HIV infected. In many if not most instances, HCWs were not wearing the 1990 CDC TB guideline-recommended respiratory protection (i.e., particulate [dust-mist, dust-fume-mist, or high-efficiency particulate air filter [HEPA]] respirators); often, the HCWs either wore no or improperly wore respiratory protection [53].
Tuberculosis Infection Control Programs in the United States
The aforementioned nosocomial MDR-TB outbreaks raised considerable concern about the status of TB infection control programs at U.S. hospitals. Several surveys have been conducted to assess these programs. In 1992, the American Hospital Association (AHA) in collaboration with the CDC conducted a survey of all U.S. municipal, Veterans Administration, and university hospitals, and a 20% random sample of private hospitals [67]. Responses were received from 763 (71%) of the 1,076 surveyed hospitals; of the 763 respondents, 178 (25%) hospitals in 39 states admitted MDR-TB patients. Among the 763 hospitals, the number of AFB isolation rooms meeting 1990 CDC TB guidelines recommendations ranged from 0–60 (median, 7); 219 (29%) hospitals had no AFB isolation rooms meeting the CDC criteria. HCW TST programs varied widely. Fifteen (2%) reported nosocomial transmission of M. tuberculosis to patients, and 91 (13%) reported TB transmission to HCWs.
In 1993, the Society of Healthcare Epidemiology of America (SHEA) in collaboration with the CDC conducted a survey of the SHEA membership, most of whom are at medical school–affiliated teaching hospitals [19,20]. From 1989 through 1992, the number of SHEA member hospitals admitting MDR-TB patients increased from 10/166 (10%) to 49/166 (30%). During this period, the median HCW TST positivity rate at the time of hire increased from 0.54% to 0.81%; in contrast, the median HCW TST conversion rate remained stable at ~0.34% during the period (range, 0.35% to 0.33%). AFB isolation rooms meeting CDC 1990 TB guideline recommendations were reported at 113/181 (62%) hospitals responding to the question. During the study period, the proportion of 191 respondent hospitals in which surgical submicron masks or dust-mist or dust-fume-mist respirators were used increased from 9 (5%) to 85 (43%).
In another survey, conducted in 1993 by the Association for Professionals in Infection Control and Epidemiology (APIC) in collaboration with the CDC and covering the same period, most of the 1,494 respondents were smaller community hospitals, and the proportion of hospitals admitting drug-susceptible tuberculosis or MDR-TB patients increased from 46.4% to 56.6% and 0.8% to 4.5%, respectively [21]. From 1989 through 1992, the HCW pooled mean TST positivity rate at hire rose from 0.95% to 1.14%, and the TST conversion rate increased from 0.4% to 0.5%. In 1992, rooms compliant with the CDC 1990 TB guideline recommendations for AFB isolation were reported at 66% of the hospitals, and at 64% of the hospitals' HCWs still used surgical masks for respiratory protection; from 1989 through 1992, the number of hospitals in which particulate respirators were used increased from 0.4% to 13.8%.
Nosocomial transmission of M. tuberculosis is a particular concern in emergency departments where a wide variety of patients, many of whom have fever and cough, seeks care. Moran et al. found that among patients with active pulmonary TB at an urban emergency department, TB was often not suspected, and isolation of these infectious patients was delayed [92]. In 1993, Moran et al. conducted a survey of infection control practices in emergency departments at a sample of the hospitals responding to the AHA/CDC survey [93]. Of the 446 emergency departments surveyed, 298 returned completed questionnaires. The proportion of emergency departments in which TB patients were seen daily, weekly, monthly, or less frequently was 12.6%, 17.2%, 23.3%, and 46.9%, respectively. The proportion in which TB isolation rooms meeting CDC 1990
P.549
TB guideline recommendations were available in triage or waiting rooms or in the emergency department itself was 1.7% and 19.6%, respectively. One or more HCWs had TST conversions in 16.1% of the surveyed emergency departments in 1991 and 26.9% in 1992. Prevention and control of M. tuberculosis in emergency departments depends on (1) education of the staff about the high-risk groups for and signs and symptoms of TB so that they will suspect and promptly identify infectious patients; (2) ensuring that emergency departments have adequate isolation facilities for infectious TB patients; (3) rapid isolation of screened patients who may have infectious TB; and (4) a comprehensive HCW TST program.
The microbiology laboratory plays a critical role in the rapid identification of infectious TB patients (see Chapter 10). However, the microbiologic methods used in many laboratories are not the most rapid, and communication between the laboratories and the clinicians often is inadequate [13,19,20,21,67,94]. In one survey of U.S. hospital-based laboratories, rapid methods were used for AFB microscopy in 47%, primary culture in 72%, M. tuberculosis identification in 38%, and drug susceptibility testing in 13% of the laboratories surveyed [67]. Approximately 46% of hospitals surveyed and an estimated 30% of laboratories at all U.S. hospitals with ≥100 beds performed the minimal number of mycobacterial cultures deemed necessary to maintain competence [12,13]. Fortunately, since 1992, there has been a significant improvement in compliance with recommended methods for AFB smear, culture, and antimicrobial susceptibility testing [13].
These outbreak and survey data show that the emergence of MDR-TB as a major public health problem was associated with the incomplete implementation of the CDC TB guideline recommendations [9,55,56]. Many hospitals do not have the recommended patient isolation facilities, HCW TST or respiratory protection programs, or laboratory processing practices. Although the APIC/CDC survey shows that as many as 50% of community hospitals do not admit TB patients routinely and few admit MDR-TB patients, the SHEA/CDC and AHA/CDC surveys show that urban, larger hospitals and those with a medical school affiliation routinely admit TB or MDR-TB patients [19,20,21,67]. Data from the SHEA/CDC and APIC/CDC surveys show that there continues to be improvement in the degree of implementation of CDC TB guideline recommendations since the initial AHA/CDC survey. Many, if not most, of the inadequate TB infection control practices reported in the United States exist in healthcare facilities throughout the world. More recently, Manangan et al. in another survey of U.S. hospitals confirmed that from 1992 through 1996, there was continued implementation of the 1994 CDC TB Guideline recommendations and a decreased in HCW TST conversion [95]. In a 1995 survey in Belgium, Ronveaux et al. found that many of the conditions facilitating M. tuberculosis transmission (failure to detect and isolate patients, mixing of HAV and TB patients in the same unit, HCWs not wearing appropriate respiratory protection, and failing to monitor HCW TSTs) were common and that TB infection control practices needed improvement. [96]
Efficacy of the CDC Tuberculosis Guidelines Recommendations
The nosocomial TB and especially the MDR-TB outbreaks raised concern in the infectious disease, infection control, occupational medicine, and industrial hygiene communities about the effectiveness of the CDC TB guideline recommendations [9,55,56]. To assess the efficacy of these control measures, follow-up investigations were conducted at three of the hospitals where MDR-TB outbreaks had occurred and at one where a drug-susceptible M. tuberculosis outbreak occurred. In each MDR-TB hospital, a wide variety of infection control measures, similar to those in the CDC 1990 and 1994 TB guideline, was implemented (Tables 32-4 and 32-5). Because same-ward exposure to an infectious MDR-TB patient was identified as the most significant risk factor in the initial MDR-TB outbreak investigations, this was assessed as a measure of nosocomial or occupational acquisition of TB in the follow-up studies. In the first follow-up investigation at an MDR-TB outbreak hospital in New York City, Maloney et al. documented that the proportion of patients with MDR-TB strains decreased, the proportion of MDR-TB patients with same-ward exposures decreased, and the TST conversion rates of HCWs assigned to the outbreak wards were lower in the follow-up or intervention period than in the outbreak period [97]. In the second follow-up investigation, Wenger et al. showed that, after implementation of control measures on the MDR-TB outbreak HIV ward, no episodes of MDR-TB could be traced to contact with infectious MDR-TB patients and that HCW TST conversions were terminated [98]. In the third follow-up study, Stroud et al. documented that, after implementation of recommended CDC TB infection control measures, the MDR-TB attack rate for AIDS patients decreased from 19/216 (8.8%) to 5/193 (2.6%) [99]. Blumberg et al. conducted a fourth follow-up investigation at a hospital in which a drug-susceptible M. tuberculosis outbreak had occurred [86,100]. Their study documented that, after the hospital introduced mostly administrative controls—more rapid identification of infectious TB patients, isolation of infectious patients in rooms with increased air exhaust, and improved HCW respiratory protection—a decrease occurred in the number of exposures to infectious TB patients (4.4 to 0.6 per month), cumulative number of days per month that infectious TB patients were not in isolation (35.4 to 3.3), and HCW TST conversions (3.3% to 0.4%) [100].
Despite the fact that none of these interventions was a randomized controlled trial or that each of the individual components was independently assessed, each of these studies shows that with more complete implementation of administrative and engineering/environmental controls
P.550
P.551
and use of respiratory protective devices for HCWs, M. tuberculosis outbreaks can be terminated and further M. tuberculosis transmission from patient to patient or patient to HCW can be either terminated or reduced to background TB rates seen on wards where TB patients are not admitted routinely [89]. These data document that the CDC TB guidelines work if they are fully implemented.
|
TABLE 32-4 |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
TABLE 32-5 |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
CDC Tuberculosis Guidelines
In 1990, the CDC published Guidelines for Preventing the Transmission of Tuberculosis in Health Care Settings, with Special Focus on HIV-Related Issues [53]. In 1994, CDC publishedGuidelines for Preventing the Transmission of Tuberculosis in Health-Care Facilities, 1994 [56]. Most recently, the 1994 guideline has been updated, in Guidelines for Preventing the Transmission of Mycobacterium tuberculosis in Health-Care Settings, 2005 [9]. These guidelines provide the elements of an infection control program to prevent the transmission ofM. tuberculosis in healthcare facilities. Emphasis is placed on methods to prevent the generation of infectious airborne droplet nuclei, early identification of infectious TB patients, preventing the spread of M. tuberculosis through source controls, air disinfection to reduce microbial contamination of the air, disinfection and sterilization, and surveillance for TB infection or disease in HCWs. Recommendations for patient AFB isolation rooms included having six or more air changes per hour, negative pressure in relation to other rooms or corridors, and exhausting the air from the room directly to the outside. The guidelines also recommend the use of particulate respirators as respiratory protection for HCWs.
As a result of the nosocomial TB outbreaks and in response to requests for changes or clarifications in the CDC 1990 TB guidelines, the CDC revised these guidelines and published theGuidelines for Preventing the Transmission of Mycobacterium tuberculosis in Health-Care Facilities, 1994 and then updated them in 2005 [9,56] (Table 32-6). The 1994 and 2005 CDC TB guidelines include a recommendations section followed by extensive supplement material. The recommendations include sections on assignment of responsibility, risk assessment, development of the TB infection control program, and periodic risk assessment; identifying, evaluating, and initiating treatment for patients who may have active TB; management of patients in ambulatory care settings or emergency departments who may have active TB; management of hospitalized patients who have confirmed or suspected TB; engineering control recommendations; HCW protection; cough-inducing and aerosol-generating procedures; education and training of HCWs; HCW counseling, screening, and evaluation; problem evaluation; coordination with the Public Health Department; and additional considerations for selected areas in healthcare facilities and other healthcare settings. The supplements include discussion of determining the infectiousness of a TB patient, diagnosis and treatment of both latent infection and active TB disease, engineering controls (i.e., ventilation and ultraviolet germicidal irradiation), respiratory protection, and decontamination (i.e., cleaning, disinfecting, and sterilizing of patient care equipment).
The major differences in the 2005 Guideline compared to the 1994 Guideline are:
P.552
P.553
|
TABLE 32-6 |
|||
|
Hierarchy of Controls
The 1994 and 2005 CDC TB guidelines emphasize the importance of understanding the hierarchy of TB control measures; these include administrative controls to identify infectious TB patients promptly and to reduce HCW and patient exposure to such patients, engineering controls to reduce the generation and spread of airborne droplet nuclei, and respiratory protective devices to protect HCWs from inhalation of infectious airborne droplet nuclei. Although each of these measures was discussed in the 1990 CDC TB guidelines, the 1994 CDC TB guidelines highlight the importance of these measures. The first and most critical TB control measure is the use of administrative controls to reduce the risk of patient or HCW exposure to people with infectious TB. Compliance with this recommendation requires the development of effective written policies and protocols and implementation of these into practice to ensure that clinicians maintain a high index of suspicion for TB in patients with respiratory symptoms and conduct an appropriate diagnostic workup of patients suspected of having TB; that rapid laboratory techniques (fluorescence stains for AFB smears, radiometric methods for species identification and susceptibility testing, or genetic probes for species identification) are used; that laboratory results are rapidly communicated to clinicians; and that infectious TB patients are promptly isolated to minimize exposure to other patients or unprotected HCWs [9,55,56].
Administrative source controls are essential for preventing M. tuberculosis transmission. A critical element of any TB control program is education of HCWs about the epidemiology of TB, the importance of suspecting TB particularly in patients with HIV infection, the importance of HCW TST and respiratory protection programs, the need for clinicians to know the susceptibility of prevalent M. tuberculosis strains in the hospital and community, the importance of rapidly and appropriately isolating TB patients, and the importance of initiating an effective treatment regimen in infectious TB patients.
The next level of the hierarchy is the use of engineering/environmental controls to prevent the generation and
P.554
spread of airborne droplet nuclei and to reduce the concentration of these infectious particles. These include (1) direct source control using local exhaust ventilation, (2) controlling the direction of air flow to prevent contamination of air in adjacent areas, (3) dilution and removal of contaminated air by general ventilation, and (4) air disinfection by air filtration or ultraviolet germicidal irradiation.
The third and last level of the hierarchy is the use of personal respiratory protective devices by HCWs in areas where the risk of exposure to infectious droplet nuclei and resultant occupational acquisition of M. tuberculosis is suspected to be higher than normal (e.g., AFB isolation rooms and where cough-inducing or aerosol-producing procedures are performed).
The outbreak investigations, the follow-up studies at outbreak hospitals, and the surveys have not documented the independent importance of these different measures. Furthermore, because the control measures recommended in the CDC TB guidelines are intended to be implemented as a group, neglecting any one measure could lead to the failure of the others to eliminate or reduce nosocomial transmission of M. tuberculosis. Most of the nosocomial or occupational acquisition of TB results from exposure to an unsuspected or undetected infectious TB patient.
Risk Assessment
Another important element of the 1994 and 2005 CDC TB guidelines is the introduction of the concept of conducting a risk assessment. The purpose of the risk assessment is to identify healthcare facilities or areas within those facilities in which the risk of exposure to infectious TB patients and subsequent patient-to-patient or patient-to-HCW M. tuberculosis transmission is minimal, very low, low, intermediate, or high. In this way, the TB control program can be individualized for a specific area or hospital so that high- or low-risk facilities would not need to comply with the same recommendations; greater flexibility could be given both between and within institutions. This risk assessment is used to determine the frequency of HCW TST, repeat risk assessment, and ventilation evaluation and to suggest supplemental engineering/environmental interventions.
Tuberculosis Control Program
The basic elements of the TB control program are applicable to all types of healthcare facilities (Table 32-6). These elements emphasize the following:
Respiratory Protection
The 1990 CDC TB guidelines recommended particulate respirators, which include dust-mist, dust-fume-mist, or HEPA filter respirators, for use as respiratory protection against TB. However, the 1994 CDC TB guidelines established new criteria for respiratory protection: (1) the ability to filter particles 1 µm in size (in the unloaded state) with a filter efficiency of 95% or better (i.e., filter leakage <5%), given flow rates of up to 50 liters/minute, (2) the ability to be qualitatively or quantitatively fit tested in a reliable way, (3) the ability to be adequately fit checked before each use in a reliable way, and (4) the ability to fit HCWs with different facial sizes and characteristics, which can usually be met by the availability of at least three sizes of respirator. The criteria were based on the characteristics thought to be most desirable in a respirator and on the in-use experience at drug-susceptible or MDR-TB outbreak hospitals where patient-to-HCW M. tuberculosis transmission was terminated by using submicron masks or dust-mist respirators with implementation of the 1990 CDC TB guidelines recommendations [97,98,99,100,101].
When the 1990 CDC TB guidelines first recommended particulate respirators to protect HCWs from occupational acquisition of M. tuberculosis, it became clear that the use of respirators was for protection of HCWs from patient infections, not vice versa. This area then came under the jurisdiction of the U.S. Occupational Safety and Health Administration (OSHA) [101,102]. Thereafter, no matter which respirator was recommended, the law required an HCW respirator training, education, and fit testing program. Also, OSHA requires that, in the United States when HCWs wear a respiratory protective device to protect them from disease, they must use a National Institute for Occupational Safety and Health (NIOSH)–certified respirator. Therefore, in 1992, OSHA asked NIOSH to assess and recommend which respirator would be required to protect HCWs from occupational M. tuberculosisacquisition. Because of a lack
P.555
of data about the concentration of M. tuberculosis in airborne droplet nuclei, what the minimal infectious dose is, the exact size and size distribution of airborne droplet nuclei particles, the potential of mortality given infection, the risk of toxic reaction from anti-TB prophylactic or therapeutic agents, and the NIOSH/OSHA interest in providing a zero-risk environment, NIOSH recommended a powered air-purifying respirator (PAPR) with a HEPA filter for moderate exposures and a positive-pressure, air-line respirator with a tight-fitting half-mask respirator for high-risk exposures [92]. By law, NIOSH may not consider either cost or practicality in making its recommendation. During 1992 and 1993, various OSHA regions made respirator recommendations ranging from dust-mist respirators to PAPRs. Subsequent to the publication of the 1994 TB Guideline, NIOSH recommended that a respirator providing protection equal to or greater than an N-95 (i.e., that it would be certified to filter at least 95% of particles of ≥1 µm) was required.
To understand the complexity involved in arriving at an appropriate respirator recommendation for HCWs to prevent occupational acquisition of M. tuberculosis, it is important to understand the relationship between OSHA and NIOSH and the laws specific to the United States [101]. OSHA requires that any respiratory protective device used to protect HCWs must be NIOSH certified [103]. Before 1995, the existing NIOSH respirator certification process did not adequately test the efficacy of dust-mist or dust-fume-mist respirators against low-concentration aerosols in the size range of M. tuberculosis droplet nuclei (best estimated at 1 to 5 µm). Subsequently, studies documented that there was wide variability in penetration of dust-mist or dust-fume-mist respirators when challenged by either 1-µm particles or M. chelonae aerosols [104]. Although some manufacturers' dust-mist or dust-fume-mist respirators prevented ≥95% of these particles from penetrating, other manufacturers' respirators allowed up to 40% penetration. Thus, in contrast to HEPA filter respirators (equivalent to the current N-95 respirators), all of which are certified to prevent penetration of <0.03% of 0.3 µm particles, the efficacy of dust-mist and dust-fume-mist respirators varied widely, was not certified by NIOSH, and thus could not be ensured. For this reason, the 1994 CDC TB guidelines indicated that although some dust-mist or dust-fume-mist respirators would meet the new guidelines criteria, only HEPA-filtered (now N-95) respirators always meet these criteria and were certified by NIOSH to do so [56,101,105].
In 1996, NIOSH implemented a new respirator certification process that tests respirators with a 0.3 µm test particle and designates them into different classes of respirator based on 99.95%, 99.00%, or 95.00% filtration efficacy [106]. All three classes of respirators surpass the minimum filter criteria recommended in the 1994 CDC TB guidelines (i.e., the ability to filter 95% of 1 µm particles). These 95% filtration efficiency respirators, designated N-, P-, or R-95s, are available for use to protect HCWs from occupational exposure to M. tuberculosis. Because of the 1994 TB guidelines recommendations and this new NIOSH respirator certification process, a greater variety of respirators and respirator sizes has become available at lower cost to protect HCWs from M. tuberculosis. Follow-up data at several of the MDR-TB outbreak hospitals show that use of submicron surgical masks or dust-mist respirators by HCWs together with implementation of recommendations similar to those in the 1990 CDC TB guidelines terminated patient-to-HCW M. tuberculosis transmission on outbreak wards [97,98,99,100,101]. Furthermore, survey data have shown that at healthcare facilities with >6 TB patients or with >200 beds, respirators with submicron filter capability reduce the risk of HCW TST conversion [19,20]. In addition, studies show that some dust-mist or dust-fume-mist respirators filter >95% of particles with a mean size of 0.8 µm, smaller than the estimated particle size of M. tuberculosis droplet nuclei [105]. The new NIOSH respirator certification process permits the use of respirators similar to those shown to be effective in outbreak settings and reduces some of the controversy surrounding the use of respirators in healthcare facilities to protect HCWs from M. tuberculosis.
Respirator education and fit-test programs are integral parts of the U.S.-required HCW respiratory protection program. These are mandated by OSHA for any HCW potentially exposed to M. tuberculosis. Factors to consider when selecting a respirator include face-seal leakage and the ability to fit test and fit check the respirator. HCWs with facial hair (e.g., beards) will not be able to get an adequate face seal with the new N-95 or older particulate respirators. For these HCWs, particularly those performing very high-risk procedures on infectious TB patients, PARPs may be an alternative. In the United States, filtering face-piece respirators used for protection against M. tuberculosis must be selected from those approved by CDC/NIOSH under the provisions of 42 CFR 84 (www.cdc.gov/niosh/npptl/part84.pdf) [107]. A listing of CDC/NIOSH-approved disposable particulate respirators (filtering face pieces) is available at www.cdc.gov/niosh/npptl/topics/respirators/disp_part. On October 17, 1997, OSHA published a proposed standard for occupational exposure to M. tuberculosis. On December 31, 2003, OSHA announced the termination of rulemaking for a TB standard. Previous OSHA policy permitted the use of any Part 84 particulate filter respirator for protection against infection with M. tuberculosis. Respirator usage for TB had been regulated by OSHA under CFR Title 29, Part 1910.139 (29 CFR 1910.139) and compliance policy directive (CPL) 2.106 (Enforcement Procedures and Scheduling for Occupational Exposure to Tuberculosis). In addition, the 1994 and 2005 CDCTB guidelines recommend annual fit testing for HCWs. However, in 2004, the U.S. Congress passed the Fiscal Year (FY) 2005 Omnibus Spending Bill, which was signed by the President George W. Bush, prohibiting OSHA from enforcing the annual fit testing mandate for occupational exposure to M. tuberculosis in U.S. healthcare facilities for FY 2005.
P.556
Healthcare Worker Education/Problem Evaluation
Another new section in the 1994 (and expanded in 2005) CDC TB guidelines is on education and training of HCWs. All HCWs with possible M. tuberculosis exposure should be taught the epidemiology of TB; the potential for occupational exposure; the principles of TB infection control; the importance of routine and periodic HCW TST or, new to the 2005 guideline, QTF-G testing; the principles of preventive therapy; the importance of seeking evaluation if the HCW has TB symptoms; and the higher risk of disease given infection in those who are immunocompromised, particularly those infected with HIV. In addition, counseling of HCWs and options for voluntary work reassignment should be provided. The guidelines also emphasize the importance of the two-step Mantoux TST or QFT-G for all HCWs at the time of hire and that periodic TST or QFT-G should be performed based on the HCW's risk assessment. Furthermore, a section on evaluating problems provides guidance on investigation of TST conversions or active disease in HCWs or possible patient-to-patient TB transmission. In one nosocomial TB outbreak, an HCW was the source. Both the HCW's lack of recognition of symptoms as consistent with TB and the occupational health department's delay in recognition of TB in other HCWs with TST conversions and abnormal chest radiographs contributed to the outbreak [86,100].
Guideline Supplements
The 1994 and 2005 Guidelines have supplements that provide extensive details on the rationale, methods, and guidance for implementation of various elements of the TB control program. For example, the supplement on diagnosis and treatment provides a table on TST or blood assay for M. tuberculosis interpretation, treatment options, and drug dosages. The supplement on engineering controls discusses the basis for current science on ventilation and use of ultraviolet germicidal irradiation, and the supplement on respiratory protection discusses the factors to consider when choosing a respirator, including the key components of a respirator training program and the advantages and disadvantages of the various respirator types.
Cost of Tuberculosis Control Measures
Much of the debate about recommendations for the control of M. tuberculosis transmission in healthcare facilities has involved concerns about the cost of these measures. To facilitate identification of the extent of the infection control program that would be necessary at low-risk vs. high-risk hospitals, the 1994 and 2005 TB guidelines recommend conducting a risk assessment. Restricting the admission of infectious TB patients to selected areas of the facility may significantly reduce the cost of the control program by limiting the area requiring environmental/engineering modifications and the number of HCWs requiring respirator education and fit testing and frequent TST monitoring. The major TB infection control measure costs are related to environmental/engineering modifications of existing older facilities to improve the number of air exchanges or to make the room have negative pressure. Although these costs can be substantial, the modifications are long term, and the costs can be amortized over the life of the equipment. Recent estimates of the cost of such environmental changes at MDR-TB outbreak hospitals varied widely (median, $229,000; range, $70,000–$559,000), depending on the extent and type of modifications required [108].
Most of the concerns about costs of the TB control program have centered around the cost of the respirators and respirator education and fit testing program. Several authors have estimated the cost of respirators or respirator education and fit testing programs based on the number of HCWs at the facility. Nettleman et al. estimated that the use of HEPA-filtered respirators would cost $7 million per case of TB prevented and $100 million per life saved [109]. Adal et al. estimated that at $7.51 to $9.08 per HEPA-filtered respirator, it would cost $1.3 to $18.5 million to prevent one episode of occupational acquisition of M. tuberculosis [110]. Both of these estimates come from institutions with a very low TB incidence. Furthermore, these estimates made a number of assumptions, such as single rather than repeated use of the respirator, admission of TB patients throughout the facility, and inclusion of all HCWs at the institution in the respiratory protection program, which may have artificially inflated the estimates. Kellerman et al. found that actual costs of respirators and respiratory education and fit testing programs were substantially less than these estimates at several of the MDR-TB hospitals; the median annual cost of respirators at four MDR-TB outbreak hospitals was $109,000 (range, $70,000–$223,000), and the median annual cost of respirator education and fit testing programs was $10,000 (range, $3,700–$19,700) [108].
Bacille Calmette-Guerin (BCG) Vaccination
Because of concern about occupational acquisition of M. tuberculosis by HCWs, the use of BCG vaccination as a protective measure has been questioned. Although in many other countries BCG has been given routinely to all infants for >60 years, routine BCG administration in any age group has never been implemented in the United States. In 1988, the U.S. Advisory Committee on Immunization Practices (ACIP) removed the HCW category from the list of people for whom BCG vaccination should be considered. The MDR-TB outbreaks in the late 1980s and 1990s resulted in a revisiting of this issue. Meta-analyses of the efficacy of BCG found that the data were inadequate (methodologic flaws) to evaluate the effectiveness of BCG in protecting HCWs [111,112]. Furthermore, the situation in the United States would be different from that in countries in which BCG is given in
P.557
infancy. In the United States, if protecting HCWs were the goal, such vaccination would have to be given to adults. No data are available on how effective BCG administered only in adulthood would be. Furthermore, BCG is unlikely to provide protection for immunocompromised HCWs and might lead to disseminated BCG disease. Many of the countries in which BCG is routinely administered have a very high prevalence of TB in their populations, suggesting that BCG does not prevent pulmonary disease although it may decrease the risk of disseminated disease, particularly in the very young. In such countries with HIV and TB outbreaks, the HCWs still may be at risk for occupational acquisition of M. tuberculosis. Few of these countries have either HCW TB infection or disease surveillance programs; thus, there are few data on current rates of TB in BCG-vaccinated HCW populations in areas with high prevalence of HIV or TB, or both. As a result of these and other factors, the ACIP and the American Committee for the Elimination of Tuberculosis decided to maintain their recommendation that use of BCG should be made on an individual basis but that it is not recommended in low-risk areas and in healthcare facilities with effective infection control programs [113].
Pediatric Settings
There has been similar controversy about the need for the CDC TB guidelines recommendations in pediatric settings. Many believe that because most pediatric patients have primary disease, poor or absent cough, and infrequent bronchial and laryngeal TB, transmission is unlikely [114]. Although the risk for such transmission may be less, it is not zero, as illustrated by the nosocomial outbreaks caused by adults visiting children at pediatric facilities [115] or outbreaks in neonatal units as described by Lee et al. [116]. Furthermore, there are few data to document the lack of risk to HCWs in these settings. One survey of pediatric facilities showed that of 158 hospital respondents, 62 (40%) had a TB infection control policy specific for children, 147 (93%) reported admitting <6 tuberculosis patients, and 104 (66%) reported admitting no TB patients [117]. At 141/154 (92%) hospitals, a TB isolation policy existed; most isolated pediatric patients with positive AFB smears (139/143; 97%), evidence of cavitary TB on chest radiograph (141/145; 97%), or an AFB-positive gastric aspirate (107/140; 76%). At 10/154 (6.5%) facilities, clusters of >2 HCWs with similar occupational exposures had TST conversions in the preceding 5 years. Thus, like adults, pediatric patients should be evaluated at the time of admission for possible TB. Those with possible TB should be evaluated for potential infectiousness (i.e., positive sputum AFB smears). Those considered infectious should be placed in appropriate TB isolation. Recently, Berkowitz et al. have developed a decision analysis for initiation of TB prophylaxis in newborns in a nursery [118].
Nosocomial Transmission of M. tuberculosis in International Settings
With the worldwide epidemics of HIV and TB, many countries throughout the world are admitting large numbers of patients with HIV infection, TB, or HIV infection and TB to their healthcare facilities. In many of these countries, overall infection control programs are minimal, and TB infection control programs are nonexistent; such facilities often have no HCW TST programs, no TB isolation rooms meeting the CDC-recommended criteria, and minimal laboratory diagnostics for TB (i.e., no rapid AFB smear, culture, or susceptibility capability). In such facilities, M. tuberculosis transmission from patient to patient and patient to HCW undoubtedly occurs. It is incumbent on such facilities to conduct a risk assessment and determine the risk of such transmission to both their HCWs and patients. In these situations, simple control measures, such as enhanced administrative controls, HCW respiratory protection, separation of infectious patients, and improved ventilation or ultraviolet germicidal irradiation may be useful in reducing the risk of M. tuberculosistransmission.
References
P.558
P.559
P.560