Mycobacterium kansasii

BACKGROUND

Mycobacterium kansasii was first isolated in 1953 (1). The species was initially characterized by the formation of yellow colonies when exposed to light, a phenomenon resulting from the deposition of beta carotene and later termed photochromogenicity (2, 3). In his classification of atypical mycobacteria, Runyon divided nontuberculous mycobacteria into four groups based on growth rate and pigmentation. M. kansasii was classified into group I, along with other photochromogens such as M. marinum (3).

Soon after Runyon’s characterization, M. kansasii was recognized as one of the most frequently isolated nontuberculous mycobacteria in the United States and Europe (4–6). In a recent study examining data from 62 laboratories across 30 countries, M. kansasii was the sixth most commonly isolated nontuberculous mycobacterium (7). The organism has been known primarily as a pulmonary pathogen, but cases of extrapulmonary and disseminated disease have also been described. Most authors consider this to be the most pathogenic nontuberculous mycobacterium, with the majority of culture-positive cases presenting with clinical disease, although colonization does occur (8–12).

In contrast to other common nontuberculous mycobacteria, M. kansasii is infrequently isolated from natural water sources or soil (13). The major reservoir appears to be tap water. Infection is likely acquired through the aerosol route, with low infectivity in regions of endemicity. Human-to-human transmission is not thought to occur despite two case reports of familial clustering (14, 15). Clustering appears to reflect a shared environment, hereditary predispositions, or susceptibility rather than human-to-human transmission (16, 17).

EPIDEMIOLOGY

True epidemiological estimates of M. kansasii disease rates are difficult to determine. Unlike the case with Mycobacterium tuberculosis, a positive M. kansasii culture is not reportable, which makes unbiased population-based measurements difficult. Moreover, laboratory-based rate estimates may be misleading, as M. kansasii growth can represent host colonization or laboratory contamination. Laboratory estimates may better represent true community rates than is the case with other nontuberculous mycobacteria, however, as positive M. kansasii cultures more often reflect disease (9).

Estimates of M. kansasii incidence vary widely by region and have varied over time. In North America, rates are highest in the central and southern United States, giving rise to an “inverted T” distribution (18). In Europe, high rates have been consistently reported in England, Spain, and the Czech Republic (19–24). Likewise, Japan, South Africa, and Brazil have reported high rates of infection (6), leading some to suggest links with industrialization. Accordingly, several publications have noted an urban predominance or an association with mining practices (4, 20, 22, 25, 26). No systematic sampling of rural and underdeveloped regions has been performed to buttress these claims.

Temporal changes in M. kansasii rates are more difficult to discern from the published literature. Longitudinal prevalence studies have yielded conflicting results, likely representing marked regional variability (6, 9, 25, 27–30). Disease rates have also been influenced by the human immunodeficiency virus (HIV) epidemic, as infections are more common in HIV-infected individuals. The annual incidence of M. kansasii infection in people living with HIV may be as high as 532 per 100,000 population, compared with 0.5 per 100,000 population in a survey of 44 U.S. states prior to the HIV epidemic (18, 31). Unlike other nontuberculous mycobacteria, such as Mycobacterium xenopi and Mycobacterium avium, there has been no clear increase in M. kansasii prevalence over time. In fact, data from some centers suggest decreasing prevalence (6, 9).

MICROBIOLOGY

M. kansasii is a slow-growing, acid-fast, photochromogenic mycobacterium. The organism grows readily in several media, including Bactec broth, Middlebrook 7H10 agar, and Lowenstein-Jensen agar, but it may take up to 6 weeks to grow. As a slowly growing mycobacterium, M. kansasii takes more than 7 days to form colonies (32). Culture colonies appear smooth or rough, with yellow pigmentation and characteristic photochromogenicity. With prolonged incubation, cultures turn reddish-orange. On microscopy, M. kansasii appears longer and broader than M. tuberculosis, often with a beaded or cross-barred appearance with Ziehl-Neelsen or Kinyoun staining (Fig. 1). Species can be identified through characteristic culture findings, biochemical testing, or high-performance liquid chromatography. More recently, species-specific DNA probes have been developed, and they are in commercial use (33–35). Real-time PCRs are currently of interest due to their considerably faster detection times than those of conventional PCR methods (36, 37).

Through molecular analysis, five subtypes of M. kansasii have been identified (38, 39). Subtype I is responsible for most human infections in Europe, the United States, and Japan. Clinical isolates appear to be highly clonal, with the same genotypes present in most human infections (32). This may point to highly conserved virulence factors and may complicate efforts to discriminate infecting strains.

CLINICAL CHARACTERISTICS

M. kansasii infection often presents with a clinical syndrome indistinguishable from that of M. tuberculosis (40). Most patients present with pulmonary disease that is symptomatic for several months prior to diagnosis (41). The most common presenting symptoms are cough, chest pain, nonmassive hemoptysis, and dyspnea (41, 42). Clinically significant disease is discovered incidentally on radiology in about 20% of cases, but these patients are usually symptomatic (41–45). Disseminated disease is an uncommon presentation in HIV-negative patients and is usually associated with immunosuppression (46). Patients present over a wide range of ages, with peak prevalence in the fifth to sixth decade of life and with male predominance. Ethnic, racial, and socioeconomic discrepancies have not been consistently identified (47).

The majority of patients with M. kansasii pulmonary disease have underlying pulmonary comorbidities, such as smoking, chronic obstructive pulmonary disease, bronchiectasis, and prior or concurrent M. tuberculosis infection. In patients with pneumoconioses, such as silicosis, M. kansasii is the most common cause of nontuberculous mycobacterial infection. Other common comorbidities include alcohol abuse, HIV, and malignancies (42, 44, 46, 48, 49). Defects in cell-mediated immunity in HIV-negative patients with M. kansasii infection have not been clearly identified (41).

The apparent similarities between M. tuberculosis and M. kansasii have prompted several clinical and radiological comparisons (44, 45, 50–55). Results have been inconsistent, with similar clinical presentations and no reliable distinctions in presenting signs or symptoms. Patients infected with M. kansasii are more likely to have underlying lung disease such as chronic obstructive pulmonary disease, while patients with M. tuberculosis disease may have more immunocompromising comorbidities (44, 45). Given their similarities in clinical presentation, physicians should rely on local epidemiology and prompt microbiological confirmation to distinguish these infections.

HIV-M. kansasii coinfection is often associated with advanced immunocompromised status, with an average CD4 count of <50/mm3 in most studies (48, 56–64). In a study by Marras and Daley (11), 92% of culture-positive patients had clinically important disease, compared to an estimated 50% in HIV-negative populations (31). M. kansasii infection can present earlier, with CD4 counts exceeding 400/mm3, particularly when pulmonary comorbidities are present (65). Disseminated disease is more common at lower CD4 counts and is considered an AIDS-defining diagnosis (63, 66). Accordingly, rates of infection and disseminated disease appear to have decreased since the introduction of highly active antiretroviral therapy (HAART).

EXTRAPULMONARY DISEASE

M. kansasii extrapulmonary disease is infrequent in most regions. In a large study by Kaustová et al., only 0.6% of M. kansasii-infected patients presented with extrapulmonary manifestations (59). Surveys in Great Britain, however, noted higher rates, with 8 to 9% of M. kansasii infections presenting with extrapulmonary disease (21, 23). Common sites of extrapulmonary disease include the lymph nodes, skin, and musculoskeletal and genitourinary systems.

Compared with other nontuberculous mycobacteria, M. kansasii is a relatively rare cause of lymphadenopathy. Clinically, it presents with painless swelling, without associated constitutional symptoms. In contrast, skin and soft tissue infections can present in a variety of ways (67, 68). Generally, immunocompetent patients present with localized lesions and a history of dermatological disease, corticosteroid injection, or skin injury, while immunosuppressed patients present with more diffuse disease or abscess formation.

Musculoskeletal manifestations include tenosynovitis and monoarticular septic arthritis. These are often associated with recent trauma or corticosteroid use, while approximately half of patients have underlying systemic disease (69). Patients with septic arthritis often have mild symptoms such as stiffness in the wrist, knee, or elbow for several months. Arthrocentesis is often ineffective for diagnosis, as the majority of samples will have negative smear and cultures. Synovial biopsy is usually necessary for diagnosis.

RADIOLOGY

Classically, the radiological findings of M. kansasii pulmonary disease in HIV-negative patients are similar to those of reactivation pulmonary tuberculosis, with unilateral, upper lobe airspace opacification and cavitary disease; however, noncavitary disease does exist among individuals with M. kansasii infections (70). Although bilateral disease or lower lobe involvement is commonly observed (45, 51, 52, 55), less common manifestations of M. kansasii pulmonary disease include pleural thickening, pleural effusions, nodules, and hilar adenopathy. Fewer than 10% of patients have normal chest X rays (51, 52). There are no reliable radiological findings that differentiate M. kansasii from M. tuberculosis or other nontuberculous mycobacteria (53, 54).

HIV- and M. kansasii-coinfected persons are less likely to present with cavitary disease. When present, this finding is associated with high mortality (50, 58). Hilar lymphadenopathy and miliary patterns are more common in HIV-positive populations, particularly those with declining CD4 counts (71).

DIAGNOSIS

The signs and symptoms of M. kansasii disease are variable and nonspecific. Patients often present in middle age with vague pulmonary or constitutional symptoms and a history of comorbid lung disease, making clinical and radiological evaluation difficult. Moreover, there are no reliable indicators distinguishing M. kansasii disease from tuberculosis and other mycobacterial infections. Culture results along with a thorough clinical and radiological evaluation are essential for accurate diagnosis.

Recent American Thoracic Society/Infectious Disease Society of America (ATS/IDSA) guidelines recommend a minimal radiological evaluation with chest X ray (or computed tomography in the absence of cavitation), combined with positive cultures and clinical exclusion of other diagnoses (40). Cultures are considered positive when two consecutive positive sputum cultures, one positive culture from bronchoscopy specimens, or one positive sputum culture with compatible pathology is present. Although relaxed since the 1997 guidelines, these microbiological criteria remain somewhat stringent. The ATS/IDSA guidelines apply to all nontuberculous mycobacteria and reflect the need to differentiate disease from contamination and colonization. M. kansasii is rarely a colonizing organism and can rapidly progress in immunocompromised patients. This has led some authors to consider a lower diagnostic threshold for patients with positive M. kansasii cultures, particularly in people living with HIV (65).

INTERFERON GAMMA RELEASE ASSAYS

Interferon gamma release assays (IGRAs) have been identified as sensitive and specific tools for diagnosing M. tuberculosis infection (72). The specificity of IGRAs for M. tuberculosis may be reduced by M. kansasii infection, as M. kansasii encodes CFP-10 and ESAT-6, two antigens targeted by IGRAs (73). Some authors have suggested that IGRAs may prove useful for rapid diagnosis of M. kansasii infection. Indeed, a publication demonstrated the potential utility of IGRAs for smear-positive, PCR probe-negative patients with suspected M. kansasii disease (74). More recently, however, the rate of IGRA positivity was found to be relatively low among M. kansasii patients without past TB infection, at only 18.8% to 20%, depending on the IGRA used (75). Furthermore, in the face of rapid, highly sensitive and specific DNA probes, the utility of IGRA in this context is questionable.

TREATMENT OF HIV-NEGATIVE PATIENTS

M. kansasii infection in HIV-negative patients typically responds well to medical therapy. With susceptible organisms, sputum is usually cleared within 4 months of therapy, and most patients have low rates of relapse. Perhaps reflecting this success, there is limited high-quality evidence to guide the treatment of M. kansasii disease. There are few prospective studies and only one published randomized controlled trial (76–78).

The British Thoracic Society (BTS) recommends daily two-drug therapy with rifampin and ethambutol for 9 months (79). Updated BTS guidelines are in preparation and due for publication in 2017. ATS/IDSA guidelines recommend daily therapy with isoniazid, rifampin, and ethambutol for fully sensitive disease (40) (Table 1). Alternative medications that are active against M. kansasii include streptomycin, clarithromycin, amikacin, ethionamide, sulfamethoxazole, rifabutin, linezolid, and fluoroquinolones. Studies have supported the use of streptomycin as a supplement to the first 3 months of isoniazid, ethambutol, and rifampin in 12-month regimens, but this is not routinely recommended (80, 81).

Prior to the introduction of rifampin in 1968, retrospective series showed high levels of treatment failure and relapse rates close to 10% (82). Surgical interventions were also common, with 20 to 54% of patients subjected to surgery (82, 83). The introduction of rifampin-enhanced sputum clearance decreased surgical interventions and reduced relapse rates (59, 84, 85). In this context, the value of so-called companion drugs, such as isoniazid and ethambutol, is less clear. These drugs may not enhance efficacy but may instead act to prevent the emergence of rifampin resistance (71). In particular, the efficacy of isoniazid has come into question. In a BTS trial examining the efficacy of rifampin-ethambutol, early empiric treatment with isoniazid did not improve outcomes (77). Moreover, isoniazid resistance does not appear to affect outcomes in rifampin-sensitive strains (47).

Given the importance of rifampin to current regimens, ATS/IDSA guidelines recommend routine susceptibility testing of rifampin alone (40). Only when rifampin resistance is present should extended susceptibility testing be employed. Reported rates of rifampin resistance are generally low, with most series reporting rates less than 5% (43, 44, 86). Acquired resistance has been reported and should be of particular concern in noncompliant patients (87). In contrast, depending on the cutoff used, rates of isoniazid resistance exceed 70% in many series (22, 62), although as with other nontuberculous mycobacteria, in vitro sensitivities do not always correlate with clinical response.

More recently, clarithromycin and fluoroquinolones have been promoted as effective therapies, although there is some concern about rising resistance to the latter, including ciprofloxacin and moxifloxacin (88). Nonetheless, the efficacy of clarithromycin has been demonstrated in three published series. In a prospective study of 18 patients with M. kansasii disease, Griffith et al. demonstrated the efficacy of intermittent therapy with clarithromycin, ethambutol, and rifampin; most patients cleared their sputum within 2 months, and there were no cases of relapse despite an average treatment duration of 13 months (76). Shitrit et al. treated 62 patients with rifampin, ethambutol, and clarithromycin until sputum cultures were negative for 12 months. All patients survived with no cases of relapse (45, 89). A study by Davies et al. in 2012 noted high rates of treatment-limiting side effects associated with 3-drug regimens, which included, but were not limited to, the use of clarithromycin combined with ethambutol and rifampin, with a 10% relapse rate (19). Consequently, risk assessment of triple therapy, as well as the efficacy of isoniazid in comparison to clarithromycin, requires further examination. Despite the associated adverse effects, clarithromycin appears to be a preferable alternative to isoniazid as a companion drug. Fluoroquinolones, on the other hand, appear quite effective in vitro, but reliable patient data are lacking.

Due to the organism’s resistance to pyrazinamide (PZA), this drug is not recommended for the treatment of M kansasii infections. Reduced pyrazinamidase activity in M. kansasii is suggested to be one factor generating resistance to PZA (90–92). Although high pyrazinoic acid (POA) efflux rates among mycobacteria are generally known to increase PZA resistance (92–95), this does not seem to be the case with M. kansasii, which has evidence of a weak POA efflux mechanism (90, 91). Current research on POA efflux in relation to PZA resistance exists mostly in the context of M. tuberculosis; while the mechanisms of action of PZA are rather similar in M. tuberculosis and M. kansasii (90), further investigation with respect to M. kansasii is necessary.

TREATMENT OF EXTRAPULMONARY DISEASE

There are limited data on treatment of extrapulmonary disease. Generally, a first episode of M. kansasii lymphadenopathy is best treated with surgical excision and close clinical follow-up. With recurrent disease, surgical excision followed by medical therapy should be considered. When treating septic arthritis, one should consider surgical debridement, along with 12 to 18 months of antimycobacterial therapy (69). In terms of other sites of infection, there are limited data to guide treatment; standard therapy for 12 months would seem reasonable in most contexts.

TREATMENT IN PEOPLE LIVING WITH HIV

Prior to the introduction of HAART, the prognosis of HIV-associated M. kansasii infection was poor, with a median survival of 12 months (31). The majority of deaths were not attributed to M. kansasii infection and were likely secondary to advanced HIV disease. Survival has improved since the introduction of HAART (8, 61–63). M. kansasii patients, however, tend to present with advanced immunosuppression, so 1-year mortality remains high (62, 63). Current evidence supports the use of standard antimycobacterial regimens for people with HIV-M. kansasii coinfection. Choosing an appropriate regimen, however, can be complicated by drug interactions with HAART, as several rifamycin derivatives interact with most protease inhibitors and nonnucleoside reverse transcriptase inhibitors. This can result in subtherapeutic drug levels which may, in turn, facilitate the development of HIV drug resistance. Rifabutin, a rifamycin derivative and less potent cytochrome inducer, has shown promise based on in vitro data. Limited in vivo data on 12 HAART patients show that rifabutin may be an acceptable alternative (62).

Clarithromycin appears to be an attractive alternative to rifamycin derivatives. In a series of 38 patients with M. kansasii-HIV coinfection, clarithromycin improved average survival from 2 to 10 months (63). However, five of the clarithromycin-treated patients in this series also received fluoroquinolones. Recommendations on HIV-antimycobacterial drug interactions can be found on the Centers for Disease Control and Prevention website (http://www.cdc.gov). At present, there is no evidence to support M. kansasii prophylaxis (40).

TREATMENT OF RIFAMPIN-RESISTANT DISEASE

Therapy for rifampin-resistant M. kansasii disease should be tailored to sensitivity profiles, with at least three drugs as part of the regimen (Table 1). Wallace et al. evaluated therapy in rifampin-resistant disease (86). Patients with acquired rifampin resistance were treated with daily high-dose ethambutol, isoniazid, sulfamethoxazole, and pyridoxine combined with aminoglycoside therapy. This regimen was relatively effective, with 90% sputum conversion, in those completing therapy; however, 38% of patients discontinued therapy prior to completing the regimen. Given the potential toxicities, particularly with aminoglycoside therapy, clarithromycin and/or moxifloxacin therapy could be considered as alternatives. ATS/IDSA guidelines recommend a macrolide, such as clarithromycin, moxifloxacin, and a third agent based on patient in vitro susceptibility (40).

DURATION OF THERAPY

In patients with susceptible strains taking a rifampin-containing regimen, sputum usually converts within the first 4 months of therapy, although lengthy treatment regimens are still required to prevent relapse. Several studies have examined short-course chemotherapy in M. kansasii disease, with variable results (43, 47, 59, 62, 77, 78). In six series involving 329 patients receiving 6 to 12 months of a rifampin-containing regimen, the proportion with relapse was 7% (Table 2) (43, 44, 47, 62, 77, 78). This high proportion with relapse was driven by data from a BTS trial that treated 154 patients with 9 months of rifampin and ethambutol. The relapse rate in the BTS trial was 10%, while in 176 patients treated for over 12 months with a rifampin-containing regimen, there were no cases of relapse (43, 47, 76, 78, 89). Further investigation is needed, but the current published data support the most recent ATS/IDSA recommendations, which require 12 months of sputum smear negativity (96).