Differentiating colonization from infection can be difficult Nontuberculous mycobacterial infections: Diagnosis and treatment key words: Pulmonary infections, Nontuberculous mycobacteria, Mycobacterium avium complex, Rapidly growing mycobacteria

January 1, 2007

abstract: Pulmonary disease caused by nontuberculous mycobacteria (NTM) can be challenging to diagnose and manage. Patients typically present with nonspecific symptoms, such as cough and fever, and they often have underlying lung disease, which further complicates both diagnosis and treatment. To avoid treating pseudoinfection, the diagnosis should be based on a combination of the history and results of physical examination, radiographic imaging, and smears and cultures of at least 3 sputum samples. Occasionally, it is necessary to perform bronchoalveolar lavage or obtain tissue via transbronchial or open lung biopsy for histopathology and to assess for tissue invasion. Treatment involves a long course of often costly multiple antimycobacterial drugs. However, treatment with the second-generation macrolides, clarithromycin and azithromycin, has significantly improved cure rates for specific NTM infections. (J Respir Dis. 2007;28(1):7-18)

Nontuberculous mycobacteria (NTM) are important potential causes of pulmonary disease. Since NTM are not spread person-to-person, they do not present the public health threat of tuberculosis; they can, however, cause significant morbidity and even mortality. NTM are diverse, causing both pulmonary and nonpulmonary disease in a wide variety of hosts.

NTM pulmonary infections can be very difficult to manage and even more difficult to cure, and the incidence is reported to be increasing. In this article, we focus on NTM pulmonary infections in HIV-negative patients, discussing the causes, diagnostic assessment, and current and future therapies.

Background

NTM comprise over 100 species of environmental, commensal, and potentially pathogenic bacteria. They are diverse in their microbiologic properties and their ability to cause infection in humans. Runyon1 classified NTM in the late 1950s based on criteria of growth, morphology, and pigmentation. However, the Runyon Classification System is rarely used today (except as a tool for the microbiologist) because of its lack of clinical utility. The clinical presentation of NTM infection and antimicrobial susceptibilities of pathogenic NTM vary by species; it is therefore more relevant to identify clinical isolates to the species level.

Although relatively few NTM are known to result in disease, the incidence of infections has been increasing. In 1979 to 1980, one third of all mycobacterial isolates in the United States were NTM. However, in 1991 to 1992, there were more cases of Mycobacterium avium complex (MAC) infection than there were of Mycobacterium tuberculosis infection.2

For the past 2 decades, published reports and public health surveillance data have suggested an increase in the prevalence of NTM disease.3,4 This increase probably has a multifactorial cause that includes the rise of disseminated MAC infection in patients with AIDS, an increased awareness of the diagnosis among clinicians, and improved diagnostic laboratory techniques. Unlike tuberculosis, infection with NTM is not a reportable disease in the United States; exact incidence rates are therefore unknown, making NTM surveillance difficult.

Pulmonary infections with NTM are challenging to diagnose, in part because of difficulties in distinguishing infection from colonization. They are difficult to treat because of the prolonged length of treatment required, the need for multiple antimycobacterial drugs, increasing rates of antibiotic resistance, drug interactions, and the tendency of patients with pulmonary NTM disease to have underlying lung pathology.

Because of their ubiquitous nature, NTM may colonize the airways without causing disease, thereby complicating the clinical picture. Mycobacteria are quite common in the environment and are readily found in water and soil; it has been reported that 38% of drinking water distribution systems in the United States are colonized with NTM.5

The NTM that most commonly cause pulmonary disease, in decreasing order, are MAC, Mycobacterium kansasii, and rapidly growing mycobacteria (RGM). Species that occasionally cause pulmonary disease include Mycobacterium malmoense, Mycobacterium xenopi, Mycobacterium szulgai, Mycobacterium simiae, Mycobacterium celatum, Mycobacterium asiaticum, Mycobacterium shimodii, Mycobacterium haemophilum, and Mycobacterium smegmatis.

Although diverse, NTM do share some attributes. They are common environmental colonizers that show no evidence of person-to-person transmission; they tend to produce chronic lung infection; and the infections that result from NTM, like all mycobacterial infections, must be treated with multidrug therapy.

Because of the difficulty in diagnosing true NTM pulmonary infection, diagnostic criteria were developed by the American Thoracic Society (ATS) to evaluate patients with suspected NTM pulmonary disease (Table 1).2 These criteria are most specific for the diagnosis of infection with MAC, M kansasii, and Mycobacterium abscessus, but they may also be helpful in the diagnosis of pulmonary infections with other NTM species.

MAC

The most common NTM to cause pulmonary disease are members of MAC, which includes M avium and Mycobacterium intracellulare. Older estimates suggested a prevalence of 1.1 per 100,000 persons in the United States, with the highest numbers in the southeastern states.2 Since the late 1990s, the prevalence of MAC infection appears to have increased. However, like other NTM infections, MAC infection is not a reportable disease and no current reliable estimates are available.

MAC is very common in the environment, colonizing soil and particularly water, which are the likely sources of human pulmonary colonization and infection. Pulmonary MAC infection is acquired by inhalation of aerosolized mycobacteria. It is a relatively avirulent microbe in an immunocompetent host, predominantly affecting patients with previously damaged lungs or chest wall defects.

Impaired pulmonary clearance mechanisms seem to play a role in leading from colonization to infection. Patients infected with MAC frequently have underlying lung disease resulting from chronic obstructive pulmonary disease (COPD), bronchiectasis, prior tuberculosis, bronchogenic carcinoma, chronic aspiration, or pneumoconiosis. One study reported that more than half of patients with pulmonary MAC disease had scoliosis and 27% had pectus excavatum.6

Clinically, MAC infection produces nonspecific symptoms, including chronic cough with variable sputum production, fatigue, malaise, and dyspnea. Fever, hemoptysis, and weight loss are seen less frequently. The classic presentation of pulmonary MAC disease is a subacute illness in middle-aged to elderly white men who are smokers with underlying lung disease (Table 2). Chest radiographs typically show nodular or cavitary infiltrates, predominantly in the upper lobes (Figure 1). A CT scan may reveal bronchiectasis. Pulmonary MAC infection more often leads to bilateral disease and cavitates more frequently than does tuberculosis.

Another presentation of pulmonary MAC disease has been described in predominantly middle-aged to elderly, nonsmoking white women with no underlying lung pathology. This presentation is often referred to as "Lady Windermere syndrome," implying that the pathogenesis is related to habitual cough suppression.7 Patients report chronic cough, sometimes for years, without constitutional symptoms, such as fever and weight loss. Chest radiographs may not show changes early in the disease, which often leads to a delay in diagnosis. Later in the clinical course, characteristic radiographic changes include persistent lingular or right middle lobe nodular infiltrates and bronchiectasis (Figure 2).

A third pulmonary syndrome associated with MAC, commonly referred to as "hot tub lung disease," is seen in people who inhale aerosolized MAC from pools of heated water colonized with MAC. This syndrome is considered to be a form of hypersensitivity pneumonitis; patients present with clinical and radiographic features that are very similar to those of hypersensitivity pneumonitis from other causes. Patients mainly complain of dyspnea and dry cough. Imaging reveals a combination of diffuse, bilateral alveolar or nodular infiltrates and ground-glass changes.8

MAC is also frequently isolated from patients with cystic fibrosis. Whether MAC contributes to progressive bronchiectasis or simply colonizes the damaged airways in these patients is unclear. One study suggests that progressive changes on high-resolution CT scans correlate with clinical decline in patients who have cystic fibrosis and, therefore, high-resolution CT can be an adjunctive tool in determining whether antimycobacterial therapy should be initiated.9

Less commonly, MAC can present as a solitary pulmonary nodule with no symptoms.10 Diagnosis in patients who have this presentation is usually made after surgical resection or biopsy for suspected malignancy. A tissue sample should be sent for culture and histopathologic examination, with special stains to look for acid-fast bacilli and the presence of granulomatous inflammation.

The diagnosis of pulmonary MAC disease is challenging because of the ubiquity of the organism in the environment, indolent nature of the disease, nonspecific symptoms, and lack of a single definitive test to differentiate colonization from active infection. MAC frequently colonizes the airways of patients with chronic lung diseases that may be predisposing factors for MAC infection; it is often difficult to differentiate MAC infection from these diseases. MAC can also be an environmental contaminant of sputum cultures in the laboratory. ATS criteria should be used to assist the clinician in differentiating between true MAC pulmonary infection and pseudoinfection.

As with all mycobacterial infections, pulmonary MAC disease requires multidrug therapy. With the increased availability of the extended-spectrum macrolides, clarithromycin and azithromycin have become the cornerstones of MAC treatment. Macrolide-containing regimens have significantly improved cure rates to up to 90%; this is in contrast to the sputum conversion rates of only 50% and the 20% to 30% relapse rates that occurred in the pre-macrolide era.11,12

Although these newer macrolides have potent anti-MAC activity, monotherapy leads to the development of resistance and therapeu- tic failure; it is therefore not rec- ommended. Combination therapy should include a newer macrolide, ethambutol, and a rifamycin (Table 3). In addition, intermittent streptomycin is recommended for the first 2 months of therapy in severe disease.

The appropriate duration of therapy has not been well established. In the pre-macrolide era, prolonged courses with a minimum of 18 to 24 months of therapy were considered standard. However, initial studies using clarithromycin-based regimens reported good clinical response with few relapses using 12 months of continued therapy beyond sputum conversion to negativity.11 Based on those studies, we recommend a minimum of 12 months of therapy after sputum cultures become negative; as with all NTM infections, however, treatment courses should be individualized.

All previously untreated MAC isolates are presumed to be susceptible to macrolides, and baseline drug susceptibility testing is not routinely recommended. Resistance to clarithromycin can develop and pose a significant challenge to effective therapy. Patients who continue to have positive sputum cultures after 6 months of therapy with a macrolide-containing regimen should be evaluated for macrolide resistance.

A clarithromycin minimal inhibitory concentration (MIC) of 8 µg/mL or less should be considered susceptible, and an MIC of 32 µg/mL or greater should be considered resistant. All clarithromycin-resistant isolates should be considered re- sistant to azithromycin. Susceptibility testing for other antimicrobial agents, such as rifamycins, ethambutol, amikacin, and fluoroquinolones, has not been shown to correlate with clinical response and is therefore not recommended.2

There is no consensus regarding the optimal regimen in patients with macrolide resistance or intolerance. We recommend a 4-drug combination of a rifamycin; ethambutol; a fluoroquinolone; and an aminoglycoside, preferably streptomycin. Among the fluoroquinolones, moxifloxacin appears to have superior activity in vitro compared with ciprofloxacin or levofloxacin, although clinical data are lacking.13 Surgery may be considered in localized, cavitary disease or when medical therapy has failed.

Hot tub lung disease is managed in a manner similar to that of other forms of hypersensitivity pneumonitis. Avoidance of sources of exposure and a short course of corticosteroids are usually effective in controlling symptoms. Rarely, a brief course of adjunctive anti-MAC therapy may be considered.

Patients with NTM infection should be assessed by obtaining monthly sputum specimens for cultures while they are receiving therapy, and periodically thereafter, to monitor for relapses. Patients should be counseled regarding the importance of compliance with medications to prevent resistance. In addition, close follow-up is required to evaluate for drug-related adverse effects and to minimize drug interactions.

Despite the expansion of available therapy for MAC and other NTM infections, the need for a long course of multidrug therapy, which is often poorly tolerated, makes the management of these infections potentially complex. Patients may benefit from consultation with a subspecialist and/or treatment at specialized centers.

M kansasii

The second most frequent NTM to cause pulmonary disease in the United States is M kansasii, a slow-growing mycobacterium. M kansasii is found in an inverted-T distribution from California to Florida and from Texas to Illinois.14 Water systems in urban settings are the main reservoir. In one hospital, the organism was isolated from 57% of faucets.15 Patients who are susceptible to M kansasii tend to be older men, often with underlying lung disease. However, as many as 50% of patients have no known risk factors for infection.16

Clinically, patients present with symptoms and signs suggestive of tuberculosis, including fever, chest pain, cough with or without hemoptysis, night sweats, and weight loss. The diagnosis is based on standard ATS criteria. However, a culture positive for M kansasii, compared with other NTM, is more likely to represent infection than colonization.17

Radiographs show cavitary lesions, typically in the upper lobes, in 54% to 95% of cases. Many have described the cavities as having thinner walls than those seen in cases of tuberculosis (Figure 3).16 Lower lobe disease and normal radiographs are rare.

M kansasii has in vitro susceptibility to isoniazid, rifampin, ethambutol, ethionamide, streptomycin, and clarithromycin. M kansasii has sometimes been noted to be resistant to isoniazid and streptomycin; however, this is based on susceptibility testing of M tuberculosis. Therefore, these results should not be taken into consideration when making treatment decisions about M kansasii.

Regimens that include rifampin have lowered relapse rates from 10%18 to about 1%19; they should be used whenever possible. Administration of isoniazid, rifampin, and ethambutol for 18 months is the standard regimen for treatment. However, patients should be treated until sputum cultures have been negative for 12 months. Intermittent clarithromycin in combination with ethambutol and rifampin has been used, but this regimen is based on uncontrolled data and cannot be recommended.20

RGM

The first case of lung disease caused by an RGM was reported in 1933, in a young woman with achalasia who presented with pulmonary infiltrates. There are more than 20 species of saprophytic RGM; the 3 most clinically significant--Mycobacterium fortuitum,Mycobacterium chelonae, and M abscessus--are classified collectively in the M fortuitum complex. This complex is composed of the M fortuitum group and the M chelonae group, a separation that is important for diagnosis and treatment.

These 2 groups show significant differences in pathogenicity and antimicrobial susceptibility; for example, M chelonae is more resistant to antibiotics. About 80% of RGM pulmonary infections are caused by M abscessus.21 Other RGM that rarely cause pulmonary disease include Mycobacterium immunogenum,Mycobacterium mucogenicum, and Mycobacterium goodii.

Like other NTM, RGM are ubiquitous in the environment and have been recovered readily from water, including treated drinking water, and from soil. Colonization of water supplies in the hospital setting is common, leading to both nosocomial infection and contamination with pseudoinfection. In the United States, most cases have occurred in southern coastal states, particularly in the southeast.

True RGM pulmonary infection is often found in older, otherwise healthy nonsmoking women. However, it also frequently occurs in those with underlying lung disease. There is also a predisposition to infection in those with an esophageal motility disorder, such as achalasia. M abscessus is a relatively pathogenic organism, so obtaining a single culture of M abscessus has a higher positive predictive value for true infection than does obtaining a culture positive for other RGM.

Griffith and colleagues21 conducted the largest review to date of the clinical presentation of RGM pulmonary disease, reviewing 154 cases over a 15-year period. Most patients had never smoked (66%) and were female (65%), with a mean age of 60 years. The most common presenting symptom was cough (96%) with sputum production, and approximately one third had hemoptysis.

Other common features of RGM infection on presentation include fever, weight loss, and dyspnea, although constitutional symptoms tend to be less frequent and severe than those seen with tuberculosis. The vast majority of patients experience an indolent clinical course, with a slow progression of symptoms over months to years.

Typical findings on chest radiographs are interstitial, mixed interstitial-alveolar, and reticulonodular infiltrates. Cavitation occurs in only about 15% of cases. Disease often involves the upper lobes, particularly in patients with previous upper lobe mycobacterial infection or bronchiectasis. It is frequently multilobar, involving more than 3 lobes in more than 50% of cases. In about 75% of cases there is bilateral pulmonary disease.

RGM are resistant to the typical antituberculosis agents, such as isoniazid and rifampin. It is recommended that susceptibility testing be performed on all clinically relevant isolates, using the following panel of antibiotics: clarithromycin, ciprofloxacin, doxycycline, cefoxitin, amikacin, imipenem, linezolid, and a sulfonamide. M fortuitum tends to be more susceptible than M abscessus to the oral agents, making the management of M fortuitum infection relatively less complicated and increasing the potential for cure.

Patients with M fortuitum pulmonary infection should be treated with a minimum of 2 oral agents. In general, treatment for 6 to 12 months is adequate for achieving a cure, depending on the severity of illness, presence of underlying immune compromise, and time to resolution of symptoms.

However, M abscessus tends to be susceptible only to parenteral agents, such as amikacin and imipenem, or newer macrolides. Monotherapy with an oral macrolide is insufficient and may lead to resistance.

The requirement for an additional parenteral antibiotic makes long-term treatment difficult. Inhaled aminoglycosides, particularly amikacin, have been used, but this practice is without empiric support. As a result of these limitations, clinicians often intermittently treat symptomatic exacerbations with short-course, multidrug parenteral regimens, never achieving a cure.

If the organism is susceptible, a macrolide may be combined with an oral fluoroquinolone to treat M abscessus infection. Although M chelonae has a similar susceptibility profile, linezolid inhibits some isolates in vitro22 and may present an option for oral therapy for disease caused by this species. However, there is little clinical experience with this antibiotic in treating RGM infection. Extended courses of therapy with linezolid are expensive and commonly cause cytopenias, which may limit its use.

M malmoense

First isolated in Malmo, Sweden, M malmoense is the second most common pulmonary NTM in Scandinavia. It is fairly uncommon in the United States, occurring mostly in the southern Atlantic states. Risk factors include male gender and underlying lung disease, usually as a result of tobacco use or previous tuberculosis. Clinically, M malmoense infection resembles active tuberculosis. In one study, 74% of patients with M malmoense infection had cavitary disease.23

M malmoense is sensitive to rifampin, ethambutol, and streptomycin; resistance to isoniazid has been reported. Treatment is usually an 18-month course of isoniazid (300 mg once daily), rifampin (600 mg once daily), and ethambutol (25 mg/kg for 2 months, followed by 15 mg/kg)--although 4-drug regimens have been used. The newer macrolide agents are likely very useful, but clinical data are lacking.

M xenopi

This organism was first isolated from skin granulomas of an African toad (Xenopus laevis) in 1959; however, it was not identified as a potential human pathogen until 1965, when it was identified in the sputum of a man with underlying COPD and new respiratory symptoms. Like other NTM, M xenopi is ubiquitous in the environment; its primary niche is water.

This organism has been cultured from the water supplies of hospitals as well as from domestic tap water and showerheads. It is an obligate thermophile and tends to colonize systems that include hot-water heaters. An association with water in natural reservoirs has also been shown, because case clusters tend to occur near coastal regions. Although M xenopi is one of the most common NTM pathogens in the coastal United Kingdom and in Ontario,24 it is infrequently isolated in the United States.

In the 1990s, the number of reported M xenopi pulmonary infections increased in the United States.25,26 Some experts have proposed that this is attributable to the increased number of immunocompromised patients, such as those with HIV/AIDS or solid organ transplant recipients. There is a strong temporal association between the growing number of immunocompromised patients and the rise of M xenopi. However, some researchers have proposed that the increased incidence simply reflected the heightened surveillance and the use of more sensitive laboratory techniques.25

M xenopi infection tends to occur in patients who are immunosuppressed, either because of systemic immune compromise or an alteration of local host defense mechanisms in the respiratory tract. M xenopi has low pathogenicity. In non-HIV-infected patients, respiratory infection typically occurs in older persons with underlying lung disease. Other predisposing conditions include diabetes mellitus, alcoholism, and malignancy.

Like other NTM, M xenopi frequently causes pseudoinfection. This can occur as a result of respiratory colonization; the colonization of hospital water supplies or of diagnostic instruments, such as bronchoscopes27; and laboratory contamination. In most clinical scenarios with positive cultures from sputum, M xenopi is considered a nonpathogen. However, it can be difficult to distinguish true pulmonary disease from pseudoinfection. In addition, those who are predisposed to colonization may have significant signs, symptoms, and radiographic changes, all of which are caused solely by progression of underlying pulmonary disease.

Patients infected with M xenopi present with a subacute illness that is often very similar to active pulmonary tuberculosis. Classic symptoms include fever, night sweats, weight loss, dyspnea, and productive cough. Hemoptysis may be present. The onset of symptoms is insidious, and symptoms may wax and wane over several months.

The typical radiographic presentation also mimics pulmonary tuberculosis, appearing as apical cavitary disease. Nodular infiltrates have been reported as well.28 Patients with HIV infection are less likely to demonstrate radiographic abnormalities, particularly cavitary disease.29

Some microbiology laboratories are unfamiliar with the laboratory characteristics of M xenopi because of its relatively rare occurrence in the United States. M xenopi is most often confused with MAC because of their biochemical and morphologic similarities. Unlike M avium, however, M xenopi is an obligate thermophile that grows optimally at temperatures between 42°C (107.6°F) and 45°C (113°F).30 If a specimen is smear-positive but demonstrates no growth, the laboratory should also incubate the sample at the higher temperature. Another distinguishing characteristic of M xenopi is its 16S ribosomal RNA gene sequence, which does not allow MAC-specific probes to detect M xenopi.

Because of the frequency of pseudoinfection, a single positive culture of M xenopi with a low organism load from a sputum or bronchoalveolar lavage fluid specimen is not diagnostic. Instead, ATS criteria for the diagnosis of pulmonary NTM infection should be used. If tissue is obtained, M xenopi reveals acid-fast bacilli with necrotizing and nonnecrotizing granulomas on histologic examination.

Randomized controlled trials of antimicrobial therapy for M xenopi infection are lacking. Although optimal therapy has not been established, several multidrug regimens have been used with variable success. In the pre-macrolide era, standard antituberculosis therapy with isoniazid, rifampin, and ethambutol was used most often; however, isoniazid is probably not an important component of therapy.

In vitro data and clinical experience suggest that adding a newer macrolide, such as clarithromycin, may be effective. Most patients should be treated with a combination of a macrolide, rifampin or rifabutin, and ethambutol. Some experts recommend adding streptomycin to this regimen, although its contribution is unclear. Some fluoroquinolones show promise in vitro against M xenopi but are not currently considered to be part of standard therapy. On occasion, surgical resection has been shown to be a useful adjunct to antimicrobial therapy, particularly with refractory, single-lobe disease.

References:

REFERENCES

1.

Runyon EH. Anonymous mycobacteria in pulmonary disease.

Med Clin North Am

. 1959;43: 273-290.

2.

American Thoracic Society. Diagnosis and treatment of disease caused by nontuberculous mycobacteria. This official statement of the American Thoracic Society was approved by the Board of Directors, March 1997. Medical Section of the American Lung Association.

Am J Respir Crit Care Med

. 1997;156(2 pt 2):S1-S25.

3.

Field SK, Cowie RL. Lung disease due to the more common nontuberculous mycobacteria.

Chest

. 2006;129:1653-1672.

4.

Butler WR, Crawford JT. Nontuberculous mycobacteria reported to the public health laboratory information system by state public health laboratories, United States, 1993-1996, July 1999. Centers for Disease Control and Prevention. Available at: www.cdc.gov/nchstp/tb/Laboratory_Services/NontuberculousMycobacteria.pdf. Accessed October 30, 2006.

5.

Vaerewijck MJ, Huys G, Palomino JC, et al. Mycobacteria in drinking water distribution systems: ecology and significance for human health.

FEMS Microbiol Rev.

2005;29:911-934.

6.

Iseman MD, Buschman DL, Ackerson LM. Pectus excavatum and scoliosis. Thoracic anomalies associated with pulmonary disease caused by

Mycobacterium avium

complex.

Am Rev Respir Dis

. 1991;144:914-916.

7.

Prince DS, Peterson DD, Steiner RM, et al. Infection with

Mycobacterium avium

complex in patients without predisposing conditions.

N Engl J Med.

1989;321:863-868.

8.

Marras TK, Wallace RJ Jr, Koth LL, et al. Hypersensitivity pneumonitis reaction to

Mycobacterium avium

in household water.

Chest

. 2005;127:664-671.

9.

Olivier KN, Weber DJ, Lee JH, et al. Nontuberculous mycobacteria. II: nested-cohort study of impact on cystic fibrosis lung disease.

Am J Respir Crit Care Med

. 2003;167:835-840.

10.

Kobashi Y, Fukuda M, Yoshida K, et al. Four cases of pulmonary

Mycobacterium avium intracellulare

complex presenting as a solitary pulmonary nodule and a review of other cases in Japan.

Respirology

. 2006;11:317-321.

11.

Wallace RJ Jr, Brown BA, Griffith DE, et al. Clarithromycin regimens for pulmonary

Mycobacterium avium

complex. The first 50 patients.

Am J Respir Crit Care Med

. 1996;153(6 pt 1):1766-1772.

12.

Griffith DE, Brown BA, Girard WM, et al. Azithromycin-containing regimens for treatment of

Mycobacterium avium

complex lung disease.

Clin Infect Dis.

2001;32:1547-1553.

13.

Gillespie SH, Billington O. Activity of moxifloxacin against mycobacteria.

J Antimicrob Chemother

. 1999;44:393-395.

14.

Good RC, Snider DE Jr. Isolation of nontuberculous mycobacteria in the United States, 1980.

J Infect Dis

. 1982;146:829-833.

15.

McSwiggan DA, Collins CH. The isolation of

M kansasii

and

M xenopi

from water systems.

Tubercle

. 1974;55:291-297.

16.

Shitrit D, Baum GL, Priess R, et al. Pulmonary

Mycobacterium kansasii

infection in Israel, 1999-2004: clinical features, drug susceptibility, and outcome.

Chest

. 2006;129:771-776.

17.

Bloch KC, Zwerling L, Pletcher MJ, et al. Incidence and clinical implications of isolation of

Mycobacterium kansasii:

results of a 5-year, population-based study.

Ann Intern Med

. 1998;129:698-704.

18.

Pezzia W, Raleigh JW, Bailey MC, et al. Treatment of pulmonary disease due to

Mycobacterium kansasii:

recent experience with rifampin.

Rev Infect Dis

. 1981;3:1035-1039.

19.

Ahn CH, Lowell JR, Ahn SS, et al. Chemotherapy for pulmonary disease due to

Mycobacterium kansasii:

efficacies of some individual drugs.

Rev Infect Dis

. 1981;3:1028-1034.

20.

Griffith DE, Brown-Elliott BA, Wallace RJ Jr. Thrice-weekly clarithromycin-containing regimen for treatment of

Mycobacterium kansasii

lung disease: results of a preliminary study.

Clin Infect Dis

. 2003;37:1178-1182.

21.

Griffith DE, Girard WM, Wallace RJ Jr. Clinical features of pulmonary disease caused by rapidly growing mycobacteria. An analysis of 154 patients.

Am Rev Respir Dis

. 1993;147:1271-1278.

22.

Wallace RJ Jr, Brown-Elliott BA, Ward SC, et al. Activities of linezolid against rapidly growing mycobacteria.

Antimicrob Agents Chemother

. 2001; 45:764-767.

23.

Pulmonary disease caused by

M malmoense

in HIV negative patients: 5-year follow-up of patients receiving standardized treatment.

Eur Respir J

. 2003;21:478-482.

24.

Falkinham JO 3rd. Epidemiology of infection by nontuberculous mycobacteria.

Clin Microbiol Rev

. 1996;9:177-215.

25.

Donnabella V, Salazar-Schicchi J, Bonk S, et al. Increasing incidence of

Mycobacterium xenopi

at Bellevue hospital: an emerging pathogen or a product of improved laboratory methods?

Chest

. 2000;118:1365-1370.

26.

Faress JA, McKinney LA, Semaan MT, et al.

Mycobacterium xenopi

pneumonia in the southeastern United States.

South Med J

. 2003;96:596-599.

27.

Bennett SN, Peterson DE, Johnson DR, et al. Bronchoscopy-associated

Mycobacterium xenopi

pseudoinfections.

Am J Respir Crit Care Med

. 1994;150:245-250.

28.

Costrini AM, Mahler DA, Gross WM, et al. Clinical and roentgenographic features of nosocomial pulmonary disease due to

Mycobacterium xenopi

.

Am Rev Respir Dis

. 1981;123:104-109.

29.

Juffermans NP, Verbon A, Danner SA, et al.

Mycobacterium xenopi

in HIV-infected patients: an emerging pathogen.

AIDS

. 1998;12:1661-1666.

30.

Marx CE, Fan K, Morris AJ, et al. Laboratory and clinical evaluation of

Mycobacterium xenopi

isolates.

Diagn Microbiol Infect Dis

. 1995;21:195-202.

31.

Wallace RJ Jr, Brown-Elliott BA, Crist CJ, et al. Comparison of the in vitro activity of the glycylcycline tigecycline (formerly GAR-936) with those of tetracycline, minocycline, and doxycycline against isolates of nontuberculous mycobacteria.

Antimicrob Agents Chemother

. 2002;46:3164-3167.