Mycobacterium

Tuberculosis remains a major global public health problem. Based on the recent survey data

of the World Health Organization (WHO) the actual global prevalence of M.

tuberculosis infection is 33%, corresponding to approximately 2.2 billion people. It is

estimated that 9.27 million new cases of tuberculosis occurred in 2007 (139 per 100,000

population). Of these, approximately 44%, or 4.1 million (61 per 100,000), were smear

positive and hence highly infectious cases. In 2007, an estimated 1.8 million people died of

tuberculosis (237). Reducing the burden of tuberculosis largely depends on how rapidly

DOTS (directly observed therapy, short course) programs can be implemented. Among the

prime obstacles for DOTS expansion are shortages of trained staff, lack of political

commitment, and poor laboratory services, together with inadequate patient management.

Of particular concern is the increasing number of multidrug-resistant tuberculosis cases (0.5

million in 2007) and extensively drug-resistant tuberculosis cases (92) as well as the

problem of people coinfected with human immunodeficiency virus (HIV). According to the

WHO, one of four TB deaths is HIV related. In 2007, there were 456,000 deaths and an

estimated 1.37 million new tuberculosis cases among HIV-infected individuals. The total

number of global tuberculosis cases is still increasing in absolute terms as a result of

population growth. Nevertheless, the number of incident cases per capita is falling globally in

all WHO regions. In the Western hemisphere, the number of reported cases is steadily

decreasing (approximate case rate, 6.8/100,000), reflecting the effectiveness of prevention

strategies and control measures implemented by the health authorities, among them the use

of more rapid and efficient laboratory algorithms to detect M. tuberculosis and susceptibility

testing against anti-TB drugs. In this context, the clinical mycobacteriology laboratory plays

a pivotal role.

Apart from M. tuberculosis complex, there is a growing number of NTM species, some of

which are sources of important diseases in humans (4, 52, 163, 195). Thus, rapid and

reliable identification of NTM is mandatory. The level of service and the choice of methods

applied in the clinical mycobacteriology laboratory should be determined by the patient

population served and by the resources available.

TAXONOMY AND DESCRIPTION OF THE GENUS Back to top

The genus Mycobacterium is the only genus in the family Mycobacteriaceae (224) and is

related to other mycolic acid-containing genera. The high G+C content of the DNA

of Mycobacterium species (61 to 71 mol% for all species except M. leprae [55%]) is within

the range of the other mycolic acid-containing genera, i.e., Gordonia(63 to 69

mol%), Tsukamurella (68 to 74 mol%), Nocardia (64 to 72 mol%), and Rhodococcus (63 to

73 mol%) (34).

Mycobacteria are aerobic (though some species are able to grow under a reduced

O2 atmosphere), non-spore-forming (except M. marinum [66]), nonmotile, slightly curved or

straight rods, 0.2 to 0.6 μm by 1.0 to 10 μm, which may branch. Colony morphology varies

among the species, ranging from smooth to rough and from nonpigmented

(nonphotochromogens) to pigmented. Colonies of the latter are regularly or variably yellow,

orange, or, rarely, pink, usually due to carotenoid pigments. Some species require light to

form pigment (photochromogens), while other species form pigment in either the light or the

dark (scotochromogens). Aerial filaments are very rarely formed and never visible without

magnification. Filamentous or myceliumlike growth may sometimes occur but on slight

disturbance easily fragments into rods or coccoid elements (154).

The cell wall peptidoglycolipid contains meso-diaminopimelic acid, alanine, glutamic acid,

glucosamine, muramic acid, arabinose, and galactose. Mycolic acids (number of carbon

atoms ranging from 70 to 90), together with free lipids (e.g., trehalose-6,6′-dimycolate),

provide for a hydrophobic permeability barrier (105, 112). Other important fatty acids are

waxes, phospholipids, mycoserosic, and phthienoic acids. Various patterns of cellular fatty

acids (number of carbon atoms ranging from 10 to 20) are found as well, among which is

tuberculostearic (10-R-methyloctadecanoic) acid, a unique cell component for a number of

aerobic actinomycetes (105).

The high content of complex lipids of the cell wall prevents access of common aniline dyes.

Although not readily stained by Gram’s method, mycobacteria are usually considered gram

positive. When stained with special procedures (e.g., Ziehl-Neelsen staining) mycobacteria

are not easily decolorized, even with acid-alcohol; i.e., they are acid fast. However, acid

fastness can be partly or completely lost at some stage of growth by a proportion of the cells

of some species, particularly the rapidly growing ones.

Compared to other bacteria, the growth of most mycobacterial species is slow, with

generation times of up to ~20 h (for M. ulcerans, up to 36 h) on commonly used media. A

natural division exists between slowly and rapidly growing species of mycobacteria. Slow

growers require more than 7 days to produce colonies on solid media from a dilute inoculum

under ideal culture conditions. Rapid growers, by definition, require less than 7 days when

subcultured on Lowenstein-Jensen (L-J) medium but may also take several weeks to appear

on primary culture from clinical specimens.

NUTRITIONAL REQUIREMENTS AND GROWTH Back to top

Most species adapt readily to growth on relatively simple substrates, using ammonia or

amino acids as nitrogen sources and glycerol as a carbon source in the presence of mineral

salts. A few species (e.g., M. haemophilum and M. genavense) are fastidious and require

supplements such as mycobactin, hemin, or other iron compounds. To date, M. leprae has

not been cultured outside living cells. Growth of mycobacteria is stimulated by carbon

dioxide and by fatty acids, which may be provided in the form of egg yolk or oleic acid, even

though the latter is toxic in higher concentrations (≥1%) and has to be neutralized by

albumin. Optimum temperatures for growth vary widely among species (from <30 to 45°C).

With the genomes of several mycobacterial species deciphered, functional genomics provide

new insights into their physiological and metabolic regulation and relation to virulence (226).

SUSCEPTIBILITY TO PHYSICAL AND CHEMICAL

AGENTS Back to top

Mycobacteria are able to survive for weeks to months on inanimate objects if protected from

sunlight. M. tuberculosis complex, for instance, survives for several months on surfaces or in

soil or cow dung from which other animals may be infected (127). Mycobacteria are easily

killed by heat (>65°C for at least 30 min) and by UV (sun) light, but not by freezing or

desiccation. They are more resistant to acids, alkali, and some chemical disinfectants than

most other non-spore-forming bacteria. Quaternary ammonium compounds,

hexachlorophene, and chlorhexidine are bacteriostatic at best. The concentration of

malachite green in standard acid-fast media (e.g., L-J) was selected to maximize growth of

mycobacteria while inhibiting other microorganisms. Other commonly used agents, such as

ethylene oxide and formaldehyde vapor, as well as disinfectants such as chlorine compounds,

70% ethanol, 2% alkaline glutaraldehyde, peracetic acid, and stabilized hydrogen peroxide

are effective in killing M. tuberculosis. However, agents that are inactivated in the presence

of organic matter (e.g., alcohols) cannot be relied upon to disinfect sputum and other

protein-containing materials. With iodophors, the bactericidal effect depends on the content

of available iodine as well as on the presence of organic matter (14). In their guidelines, the

Centers for Disease Control and Prevention (CDC) and National Institutes of Health (NIH)

suggest intermediate-level disinfectants (25; see also chapter 11).

EPIDEMIOLOGY AND TRANSMISSION Back to top

The genus Mycobacterium includes obligate pathogens, opportunistic pathogens, and

saprophytes. Incapable of replication in the inanimate environment, the major ecological

niche for M. leprae and M. tuberculosis complex are tissues of humans and warm-blooded

animals. M. tuberculosis is carried in airborne particles (droplet nuclei) generated when

patients with pulmonary tuberculosis cough. These particles, 1 to 5 μm in size, are kept

“suspended” by normal air currents. Infection occurs when a susceptible person inhales the

droplet nuclei. Once in the alveoli, the organisms are engulfed by alveolar macrophages.

Usually, the host cell-mediated immune response limits multiplication and spread of M.

tuberculosis. However, some bacilli can remain viable but dormant for many years after

initial infection. Patients latently infected with M. tuberculosisare asymptomatic and not

infectious but usually have a positive tuberculin skin test (TST) or a positive result with one

of the commercially available interferon gamma (IFN-γ) release assays (IGRAs). In general,

persons with a latent infection have a 10% risk during their lifetime for development of

active tuberculosis, whereas patients with HIV infection have a 10 to 15% risk per year for

progression to manifest disease (2).

In contrast, the NTM are free-living mycobacteria, usually found in association with watery

habitats such as lakes, rivers, wet soil, etc. For some human pathogenic NTM species,

e.g., M. ulcerans, M. haemophilum, or M. szulgai, the reservoir has not yet been defined

(4). M. avium complex (MAC), M. genavense, M. kansasii, M. xenopi, M. simiae, M.

gordonae, and some rapidly growing mycobacteria have been recovered from tap water.

Some of them can play a role in nosocomial disease and/or pseudo-outbreaks (4, 81, 221). A

well-known set of other sources for positive cultures are bronchoscopes and related devices.

Organisms isolated representing pseudoinfections include M. tuberculosis (179) and M.

xenopi (11) as well as other NTM. Although not components of the microbiota of humans or

animals, NTM may be isolated as “bystanders” from the skin, upper respiratory tract,

intestinal tract, and genital tract in asymptomatic individuals (52). Due to their ubiquitous

nature, the question of their clinical significance is therefore important but often difficult to

answer (4).

CLINICAL SIGNIFICANCE AND DESCRIPTION OF

SPECIES Back to top

With the advent of molecular techniques for appropriate identification, close to 200

mycobacterial species have now been described, and their number is increasing steadily. This

chapter focuses on the slowly growing mycobacteria only; rapidly growing mycobacteria are

described in chapter 30.

M. tuberculosis Complex

The M. tuberculosis complex includes M. tuberculosis, M. bovis, M. bovis BCG, M. africanum,

M. caprae, M. microti, “M. canettii,” and M. pinnipedii. Although the members of M.

tuberculosis complex are characterized by different phenotypes and mammalian host ranges,

they display a most extreme genetic homogeneity, with ~0.01 to 0.03% synonymous

nucleotide variation only and no significant trace of genetic exchange among them

(18,63, 71). Identification to the species level is not merely an academic exercise but is

justified for epidemiologic, public health, and therapeutic reasons.

M. tuberculosis

In the industrialized world, a higher prevalence of tuberculosis occurs in the medically

underserved ethnic minorities, the urban poor, homeless persons, prison inmates, alcoholics,

intravenous drug users, the elderly in general, foreign-born persons from areas of high

prevalence, and contacts of persons with active tuberculosis. The greatest known risk factor

for progression of latent infection to active tuberculosis is HIV infection. Combined HIV and

tuberculosis infections, especially in combination with drug resistance, have caused

outbreaks in the past with extremely high mortality rates. In addition, the recent emergence

of extensively drug-resistant tuberculosis exerts a dramatic impact on the changing patterns

of global tuberculosis (64). Groups with a higher likelihood of progression also include

individuals with underlying medical conditions, persons who have been infected within the

past 2 years, children ≤4 years old, and persons with fibrotic and cancerous lesions on chest

X rays.

Tuberculosis in adults is a slowly progressive process characterized by chronic inflammation

and caseation and formation of cavities. These foci may rupture into the bronchi, allowing

very large numbers of organisms to spread to other areas of the lungs and to be aerosolized

by coughing, hence infecting other persons. The clinical features of pulmonary tuberculosis

are cough, weight loss, night sweat, low-grade fever, dyspnea, and chest pain.

Extrapulmonary manifestations of M. tuberculosis infection include cervical lymphadenitis,

pleuritis, pericarditis, synovitis, meningitis, and infections of the skin, joints, bones, and

internal organs (85). Unlike the ordinary clinical picture of tuberculosis, pulmonary disease in

AIDS patients often differs in radiologic findings and is usually rapidly progressing. In these

patients, extrapulmonary manifestation and disseminated disease, sometimes even without

the formation of granulomas, are seen more frequently (85).

In culture, colonies of M. tuberculosis are off-white and rough on solid medium (Fig. 1),

although on moist media they may tend to be smoother. The genome of M.

tuberculosis (4,411,529 bp) was deciphered more than 10 years ago (32). Newer studies

have shown that M. tuberculosis has an extremely low level of genetic variation, suggesting

that the entire population of M. tuberculosis resulted from clonal expansion after an

evolutionary bottleneck some 35,000 years ago; i.e., today’s strains constitute just the

visible tip of a much broader progenitor species. Based on very recent molecular data, the M.

tuberculosis genome appears to be a composite assembly resulting from horizontal gene

transfer events predating clonal expansion. The amount of synonymous nucleotide variation

in housekeeping genes suggests that tuberculosis bacilli were contemporaneous with early

hominids in East Africa and coevolved with their human host much longer than previously

thought (71).

M. bovis

M. bovis causes tuberculosis in warm-blooded animals, such as cattle, dogs, cats, pigs,

parrots, badgers, deer, some birds of prey, and also in primates and humans. While M.

bovis caused as much as 25% of cases of human tuberculosis in developed countries in the

late 19th century, the number of cases has dropped to 1 to 2% today (144). Human disease

is very similar to that caused by M. tuberculosis and treated accordingly, except that

pyrazinamide is ineffective due to inherent resistance of M. bovis. Colonies on egg-based

media are small and rounded, with irregular edges and a granular surface; on agar media

colonies are small and flat (224). The genome of M. bovis has a size of 4,345,492 bp with a

G+C content of 65.6%. The sequence is >99.95% identical to that of M. tuberculosis. There

are some deletions in the genome which led to a reduced genome size (65).

M. bovis BCG

In many parts of the world, bacillus Calmette-Guerin (BCG) is still used for vaccine purposes.

The strain was distributed by Calmette in 1924 to laboratories around the world and has

been maintained in vitro by serial passages. Today, there exists a genetically heterogeneous

conglomerate of BCG strains (Fig. 2) (10, 189;www.bcgatlas.org) that predominantly

conform to the properties described for M. bovis, except that they are more attenuated in

virulence. The 4,374,522-bp genome contains nearly 4,000 protein-coding genes, 58 of

which are present in two copies as a result of two independent tandem duplications, DU1 and

DU2. Recently, lesions in genes encoding σ-factors and pleiotrophic transcription regulators,

like PhoRn and Crp, were uncovered in various BCG strains. Together with gene

amplification, these lesions affect gene expression levels, immunogenicity, and possibly

protection against tuberculosis, suggesting that early BCG vaccines may be superior to the

later ones, which are more widely used (17). In rare instances, BCG may disseminate as a

complication of intravesical BCG immunostimulation against bladder cancer (1).

M. africanum

M. africanum causes human tuberculosis in tropical Africa but has also been reported from

other continents such as the United States, mainly in patients who had lived in Africa (42).

The colonies of M. africanumresemble those of M. tuberculosis, and the physiological and

biochemical properties position the organism between M. tuberculosis and M. bovis. Prior to

molecular genetics, the definition of M. africanum was difficult and its validity was questioned

by some authors. Recent genotypic analyses based on variable numbers of tandem repeats

(VNTRs) and other molecular characteristics have set M. africanum clearly apart from other

members of the complex (60, 218). Likewise, Mostowy et al. (134) and Niemann et al. (139)

consider M. africanum to be a unique species within M. tuberculosis complex. Brosch et al.

(18) had reported that isolates of M. tuberculosis do not have the deletion of chromosomal

region 9 (RD 9), in contrast to M. bovis, M. microti,and M. africanum. The distribution of

deleted sequences suggests that M. africanum subtype II isolates are situated among strains

of “modern” M. tuberculosis, while subtype I isolates are heterogenous and constitute two

distinct evolutionary branches within the M. tuberculosis complex. Currently, the genome

(4,389,314 bp) with a G+C content of 65.6% is being sequenced (www.sanger.ac.uk).

M. caprae

M. caprae, formerly called M. tuberculosis subsp. caprae (5) and M.

bovis subsp. caprae (106), was originally described as preferring goats to cattle as hosts. M.

caprae not only is seen in cattle but also accounted for 31% of human tuberculosis cases,

mostly as pulmonary manifestation, in Germany between 1999 and 2001 (106). Easily

recognized by its susceptibility to pyrazinamide, M. caprae thus adds to the agents of human

tuberculosis contracted from animals. Based on mycobacterial interspersed repetitive unit

genotyping, it was demonstrated that M. caprae is closely related to the branches of

classical M. bovis, M. pinnipedii, M. microti, and ancestral M. tuberculosis but stands apart

from modern M. tuberculosis (164).

M. microti

Originally isolated from rodents such as voles and shrews, M. microti causes naturally

acquired tuberculosis in guinea pigs, rabbits, llamas (142), cats, and other warm-blooded

animals. It has recently been identified as the causative agent of tuberculosis in both

immunocompetent and immunosuppressed humans (49, 216). Usually revealing a

characteristic “croissant”-like morphology in stained smears (Fig. 3), the organism normally

fails to grow in culture. At least the vole type of M. microti can easily be recognized upon

spacer oligotyping (see chapter 29), since it contains an exceptionally short genomic direct

repeat region resulting in identical two-spacer sequence reactions (216). Compared to M.

tuberculosis, numerous deletions have been discovered (59), and the whole genome is

currently being sequenced (www.sanger.ac.uk).

M. canettii

This organism was first collected by Georges Canetti in 1969. van Soolingen et al. (215) and

Pfyffer et al. (153) reported M. canettii causing lymphadenitis in a child and generalized

tuberculosis in an HIV-positive patient, respectively. Although its natural reservoir is

unknown, the facts that both patients were exposed in Africa and more cases of cervical

lymphadenitis have been reported from Djibouti (51) supported the hypothesis that M.

canettii might be more abundant on the African continent. In 2002, pulmonary manifestation

of M. canettii was reported in Africa (126), and in 2009, the first case of M.

canettii meningitis recognized in a Sudanese refugee living in the United States was reported

(184).

With its smooth, round, and glossy colonies (Fig. 4), M. canettii differs considerably from all

other members ofM. tuberculosis complex and can even be mistaken for an NTM. However, it

has to be stressed that M. canettiimay also exist as a stable rough morphotype (69),

mimicking M. tuberculosis. As a consequence, adequate molecular techniques have to be

applied to identify the organism at the species level (e.g., line probe assay [Hain Lifescience,

Nehren, Germany], or 65-kDa hsp PCR restriction enzyme analysis). Studies by Brosch et al.

(18) and Marmiesse et al. (116) of 20 regions where insertion-deletion events took place in

the genome of M. tuberculosis suggested that M. canettii diverged first from the rest of M.

tuberculosis complex. Based on VNTR genotyping and analysis of hsp65 gene polymorphism

in 44 strains of M. canettii Fabre et al. (51) confirmed that M. canettii is the most probable

source species of M. tuberculosis complex rather than just another branch of the taxon. In

fact, it is assumed that M. canettii appeared some 2.8 million years ago and may, therefore,

be the ancestor of all members of the present-day M. tuberculosis complex (71).

M. pinnipedii sp. nov.

On the basis of host preference, phenotypic, and genotypic characteristics, M. pinnipedii, a

new member of theM. tuberculosis complex, was defined by Cousins et al. in 2003 (36).

Pinnipeds appear to be the natural host, but the organism is also pathogenic for guinea pigs,

rabbits, and possibly cattle. The organism also affects animals in zoological gardens, e.g.,

camels (Camelus bactrianus) and tapirs (Tapirus indicus [133]). Transmission of M.

pinnipedii infection from sea lions to humans was recently demonstrated by TST and IGRAs

(99). Infections with the “seal bacillus” manifest with granulomatous lesions in lymph nodes,

lungs, pleura, and spleen and are able to disseminate.

M. leprae

As a result of rigorous control programs in many areas of endemicity such as South and

Southeast Asia, Africa, and Latin America, the number of new cases of leprosy (Hansen’s

disease) has declined steadily. At the beginning of 2008, the registered prevalence of leprosy

was 212,802 compared to >763,000 in 2001 (236). Access to diagnosis and treatment with

multidrug therapy (dapsone, rifampin, and clofazimine [234]) remain key elements in the

strategy to eliminate the disease. While most countries where the disease was previously

endemic have now reached elimination (defined as a registered prevalence rate of <1

case/10,000 population), Brazil, Nepal, and Timor-Leste accounted for about 23% of

registered cases at the beginning of 2008 (236). In the past centuries, leprosy occurred on a

large scale also in Europe, in particular in Norway.

Leprosy is a chronic, granulomatous, and debilitating disease (77). Its principal

manifestations include anesthetic skin lesions and peripheral neuropathy with nerve

thickening. Leprosy illustrates a continuous spectrum of disease with very few demonstrable

bacilli (tuberculoid leprosy) to a progressive, widespread, and most severe form of the

disease with massive numbers of organisms due to the absence of cell-mediated immunity

(lepromatous leprosy). The majority of leprosy patients show manifestations between these

two polar forms and are clinically unstable. Medical complications arise from nerve damage

and immune reactions (77).

Shedding from the nose, rather than from skin lesions, is important for transmission, which

results most likely from prolonged and intimate contact with a person with multibacillary

disease. The natural reservoir for M. leprae is not well established, but naturally occurring

infections in the nine-banded armadillo (Dasypus novemcinctus) have been documented in

the southern United States with a prevalence of 0 to 10% in the animal (111).

Together with M. lepraemurium, M. leprae differs from all other mycobacteria in that it

cannot be cultured in vitro. By tradition, the diagnosis of leprosy is essentially a clinical one,

based on finding one or more signs of disease which are supported by the presence of acidfast

bacilli (AFB) on slit skin smears or in skin biopsy specimens. Since leprosy bacilli are

much less acid and alcohol fast than M. tuberculosis, 10% sulfuric acid is preferentially used

as a decolorizer in place of an acid/alcohol solution (Fite-Faraco stain [57]). In the case of

lepromatous disease, nodules and plaques are the preferred sites for biopsies, which will

reveal numerous AFB. Conversely, in patients with tuberculoid leprosy, the rims of lesions

should be biopsied, and there usually only a few or no AFB are found. A number of PCR

assays have been established to conclusively detect the organism (171) and to characterize

the M. leprae genotypes. For instance, multiple-locus VNTR analysis and single-nucleotide

polymorphism typing (175) have been applied to search for leprosy transmission. The

complete gene sequence is 3,268,203 bp in length with a G+C content of 57.8% (33).

Similar to what is observed with M. bovis, the genome of M. leprae has lost >1.1 Mb and

accumulated >1,100 pseudogenes during reductive evolution (33).

Unique clinicopathologic features of two cases of diffuse lepromatous leprosy and

phylogenetic analyses of the genes of 16S rRNA, rpoB, and hsp65 led to the discovery of a

new mycobacterial species, for which the name M. lepromatosis has been proposed (72).

Nontuberculous Mycobacteria Frequently Involved in Human

Disease

The American Thoracic Society (ATS) and the Infectious Diseases Society of America have

recently revised the guidelines for the diagnosis, treatment, and prevention of NTM disease

in HIV-positive and HIV-negative individuals (4).

Slowly Growing Species

M. avium Complex

MAC organisms have been isolated from water, soil, plants, animals, indoor water systems,

hot tubs, and pools. They are important pathogens of poultry and swine but were not

recognized as a cause of human disease until the 1940s. Generally, these organisms are of

low pathogenicity. Single positive specimens with low numbers of AFB are not infrequently

observed in individuals without apparent disease. This complicates the interpretation of

culture results, particularly from specimens of the respiratory tract (4, 52).

Before the advent of AIDS, the most common presentation of MAC infection was pulmonary

disease showing several different clinical patterns, i.e., tuberculosis-like infiltrates, nodular

bronchiectasis, and solitary nodules, as well as diffuse infiltrates in immunocompromised

patients (222). Tuberculosis-like upper lobe fibrocavitary disease due to MAC typically occurs

in white men 45 to 60 years of age who are heavy smokers, many of whom abuse alcohol,

and some of whom have preexisting lung disease. The clinical presentation is similar to that

of tuberculosis. In women, nodular bronchiectasis usually occurs in elderly nonsmoking

individuals with no predisposing disorders of the lungs or immune system other than

associated bronchiectasis (“Lady Windermere syndrome” [44]). These patients usually

present with persistent cough only; this disease tends to have a much slower progression

than cavitary disease. Less frequent are thoracic infections in otherwise healthy children

(55). MAC is also the leading cause of localized mycobacterial lymphadenitis in children,

which is usually unilateral and involves lymph nodes in the submandibular, submaxillary, or

periauricular areas (229). Generalized MAC infections in non-AIDS patients are extremely

rare (87).

In patients (n = 385) with cystic fibrosis, overall prevalence of NTM was 8%, with the most

prominent species being M. abscessus (39%), MAC (21%), and M. gordonae (18%) (158).

While M. abscessus was isolated at all ages (study population, 1 to 24 years), MAC was not

recovered before 15 years.

In conjunction with HIV infection MAC has become the most common environmental NTM

causing disease in humans. Patients with AIDS may present with disseminated or focal

infections (86, 87), mostly when the CD4 count is below 100 cells/mm3. Bacteremia occurs

in nearly all those patients, its magnitude ranging from <1 to 102 CFU/ml. The organism is

found predominantly in circulating monocytes. Almost any organ (e.g., lungs or intestines)

may be involved, with levels of mycobacteria as high as 1010 CFU/g of tissue. Focal infections

commonly involve the lungs or the gastrointestinal tract, occasionally also peripheral lymph

nodes (89).

MAC organisms are well known for their heterogeneous colony morphology. Glossy, whitish

colonies may often occur together with smaller translucent colonies. A third, less frequent

morphology resembles the dry and flat colonies of M. tuberculosis. Some MAC strains may

develop a yellowish pigment with age.

MAC is a very heterogeneous group of AFB comprised of, by classical definition, the two

taxa M. avium and M. intracellulare. Clinically, the former seems to be the more important

pathogen in disseminated disease, while the latter is more often seen in respiratory disease.

As more sophisticated molecular tools have become available in the laboratory, the

taxonomy of MAC has become increasingly complex. New species have been proposed, and

new subspecies have been discovered (206). At present, the species M. avium consists of

three subspecies, i.e., subsp. avium, subsp. silvaticum, and subsp. paratuberculosis (194).

The latter is an obligate pathogen of ruminants (Johne’s disease), while in humans, the

association of the organism with Crohn’s disease seems to be specific. However, its role in

the etiology of the disease remains to be defined (54). Since M.

avium subsp. paratuberculosis is one of the slowest growing mycobacterial species, primary

isolation can take several months, and the medium needs special supplements (206).

Additional strains that are similar to the classical MAC have been described. There are

several important differences at the genetic level that distinguish a newer MAC organism, M.

avium subsp. hominissuis, from M. avium subsp. avium (206). By using IS1245-based

restriction fragment length polymorphism (RFLP), M. aviumisolates from birds have been

identified as M. avium subsp. avium. Since highly variable RFLP patterns were found among

the M. avium isolates that all belonged to the M. avium subsp. hominissuis, a relation to pet

birds in the etiology of lymphadenitis could not be established, i.e., the source of infection

may be environmental (19, 125).

Genetically closely related to MAC organisms, but not typically considered part of the MAC

is M. lepraemurium,the agent of rodent leprosy. This organism cannot be cultured and is

identified by sequencing techniques only (206).

There are other taxa very closely related to MAC that could not be assigned to one or the

other classical species of MAC. For instance, Tortoli et al. (199) proposed to elevate a genetic

variant of MAC (MAC-A) to species rank as M. chimaera sp. nov. (see also reference 180),

the former MAC-X sequevar gave rise to the species M. colombiense (137, 220), and the

scotochromogenic sequevar MAC-Q is now also considered a species, named M.

vulneris (212). Finally, Ben Salah et al. (12) have found clinical MAC isolates that appear to

represent three other new species: M. marseillense, M. timonense, and M.

bouchedurhonense.

M. genavense

M. genavense is a slowly growing NTM that was isolated in 1991 from the blood of an AIDS

patient in Geneva, Switzerland, and was subsequently found in the United States and in

several European countries (37, 52,196). It has been associated with enteritis, genital and

soft tissue infections, and lymphadenitis in HIV-positive and in HIV-negative

immunocompromised individuals. M. genavense causes up to 12.8% of all NTM infections in

AIDS patients (40). These infections are similar to those caused by MAC, except that stool

specimens are more often smear positive in M. genavense infections (145). M. genavense is

also the most common cause of mycobacterial disease in a variety of pet birds, including

parrots and parakeets (84, 113).

Analysis of the 16S rRNA gene sequence indicates that this species is most closely related

to M. simiae.

M. haemophilum

M. haemophilum was first isolated in 1978 from a subcutaneous lesion in a patient with

Hodgkin’s disease (178). Approximately 50% of infections have been in patients with AIDS,

with a relatively large number reported from New York City. The other cases have been in

other immunosuppressed individuals (96, 178), but also in immunocompetent pediatric

patients with localized cervical lymphadenopathy (20, 31) or with a pulmonary nodule (227).

The classical clinical presentation is that of multiple skin nodules in clusters or without a

definite pattern, commonly involving the extremities, and occasionally associated with

abscesses, draining fistulas, cellulitis, endophthalmitis, and osteomyelitis (52, 178).

M. kansasii

In the United States and many other countries, M. kansasii is second to MAC as a cause of

NTM lung disease (4, 52). The organism has been cultured from its major reservoir, tap

water, in municipalities around the world where clinical disease occurs. It is common in mine

workers in both the United Kingdom and South Africa (35) and differs from that due to MAC

in that the response to chemotherapy is much better (4).

Chronic pulmonary disease resembling classical tuberculosis is the most common

manifestation of M. kansasii(183). Extrapulmonary infections are uncommon and include

cervical lymphadenitis in children, cutaneous and soft tissue infections, and musculoskeletal

disease. M. kansasii rarely disseminates, except in patients with severely impaired cellular

immunity (e.g., due to organ transplants or AIDS [4]).

M. kansasii is a photochromogenic species. Studies of the base sequences of the 16S rRNA

gene suggest that phylogenetically, it is closely related to the slowly growing, nonpigmented

species M. gastri. Molecular studies have defined up to seven genotypes of M. kansasii, with

subtypes I and II being the predominant subspecies responsible for human infection

(4, 190).

M. malmoense

The species name M. malmoense is derived from the city of Malmo in Sweden, where the

first strains were isolated from patients in 1977. Disease due to this organism was later

found in other European countries (4) with increasing incidence in Scandinavia (83). It

remains rare in the United States, Canada, and other areas of the world. However, in these

countries M. malmoense infection may be more common than suspected, because it may

require 8 to 12 weeks to isolate some strains, which is longer than many laboratories in

North America hold mycobacterial cultures.

M. malmoense isolates are clinically significant in 70 to 80% of patients (83). Patients

with M. malmoenseinfection are usually adults with chronic, difficult-to-treat pulmonary

disease, mostly middle-aged men with previously documented pneumoconiosis or young

children with cervical lymphadenitis (4, 52). Other extrapulmonary and disseminated

infections have rarely been reported.

M. marinum

M. marinum causes cutaneous infections as a result of trauma to the skin and subsequent

exposure to contaminated freshwater fish tanks (“fish tank granuloma”) or salt water (109).

The disease occurs worldwide. In the United States, it is most common in southern coastal

states. The typical presentation is a single papulonodular lesion confined to one extremity,

usually involving the elbow, knee, foot, toe, or finger. It appears 2 to 3 weeks after

inoculation and, with time, may become verrucous or ulcerated (4, 52). A second type

resembles cutaneous sporotrichosis, in which the primary inoculation is followed by spread

along the lymphatics. More severe complications include tenosynovitis, arthritis, bursitis, and

osteomyelitis. Disseminated infections, including infections in patients with AIDS or persons

under systemic steroid therapy, have been rare (4).

M. marinum is photochromogenic and requires temperatures of 28 to 30°C for primary

isolation. Israeli M. marinum isolates from humans and fish were compared by direct

sequencing of the 16S rRNA and hsp65genes and restriction mapping and amplified fragment

length polymorphism analysis. Surprisingly, significant molecular differences separated all

clinical isolates from the piscine isolates (207). Ghosh et al. (66) observed spores in old

cultures of M. marinum, which upon exposure to fresh medium germinated into vegetative

cells and reappeared again in the stationary phase with endospore formation. With its

genome comprising ~6,636,827 bp (G+C content, 65.8%), M. marinum is genetically very

closely related to M. ulcerans but also toM. tuberculosis, the latter having undergone genome

downsizing and external lateral gene transfer to become a specialized pathogen of humans

and other mammals (188).

M. simiae

M. simiae was first isolated in 1965 from rhesus macaques. More than 400 cases have been

reported from a few geographic areas, including the southwestern United States, Israel, and

the Caribbean (4, 172). The environmental niche is assumed to be aquatic. The majority of

cases relate to HIV-positive patients, involving primarily the lungs and the reticuloendothelial

system. In non-HIV patients, pulmonary manifestations are common, but lymphadenopathy,

skin lesions, genitourinary tract infections, and uveitis also occur (4, 114, 183,211). In

Israeli cystic fibrosis patients, M. simiae was the organism seen most often (40.5%),

followed by M. abscessus (31%) and MAC (14.3%) (108). M. simiae is one of the very few

NTM synthesizing niacin. Unless tested for pigment production under the influence of light, it

may be misidentified as M. tuberculosis by inexperienced observers.

M. szulgai

M. szulgai was first described as a distinct species in 1972 and is rarely recovered from the

environment. Therefore, isolation of this organism is almost always considered clinically

significant. Patients were mainly middle-aged men presenting with chronic pulmonary

disease indistinguishable from tuberculosis (4). The remaining presentations included rare

cases of bursitis, cervical adenitis, tenosynovitis, cutaneous infections, and osteomyelitis

(4, 52). Cases of M. szulgai infection in AIDS patients and disseminated disease in an

immunocompetent patient have been reported as well. Although M. szulgai is closely related

to M. malmoensebased on the 16S rRNA gene sequences, phenotypic distinction between the

two species is easy. M. szulgai is scotochromogenic at 37°C and photochromogenic at 25°C

(98).

M. ulcerans

The frequency of M. ulcerans infection has long been underestimated due to difficulties in

isolating the pathogen. Today, it is the third most frequent mycobacterial disease in humans

after tuberculosis and leprosy. In Africa the disease is known as Buruli ulcer and in Australia

as Bairnsdale ulcer (4, 52, 162, 235). Cases of Buruli ulcer have also become evident in Peru

in the most recent past (70). Closely associated with tropical wetlands, M. ulcerans most

likely proliferates in mud beneath stagnant waters. Evidence for a role of insects in

transmission of this pathogen is growing.

All ages and both sexes are affected, among them many children under 15 years.

Manifestation typically begins as a painless lump under the skin at the site of previous

trauma on the lower extremities. After a few weeks, a shallow ulcer develops at the site of

the lump. M. ulcerans produces a cytotoxin (mycolactone) with immunomodulating

properties that causes necrosis (208). The type of disease ranges from a localized nodule or

ulcer to widespread ulcerative or nonulcerative disease including osteomyelitis. If untreated,

severe limb deformities with contractures and scarring are common. There is growing

evidence that M. ulcerans also produces disease in wild animals such as lizards, opossums,

koala bears, armadillos, rats, mice, and cattle (161).

Failure to cultivate this organism in the past was due to its fastidious, heat-sensitive nature

(temperature optimum, 30°C) as well as to an excessively long generation time (up to 36 h).

The organism often requires several months of incubation to achieve isolation in primary

culture. Molecular techniques that may provide a more rapid result have been developed

(62, 157). Comparative genomic analysis has revealed that M. ulcerans arose from M.

marinum by horizontal gene transfer of a virulence plasmid that carries a cluster of genes for

mycolactone production, followed by reductive evolution (41).

M. xenopi

M. xenopi was first isolated in 1957 from skin lesions on an African toad (Xenopus

laevis), but it was not recognized as a human pathogen until 1965. In some areas such as

Canada and Southeast England, it is second only to MAC as an NTM clinical isolate (4).

Increased isolation of M. xenopi from clinical specimens may also be due to improved

laboratory techniques. With an optimum growth temperature of 45°C, it seems to frequently

occur in hot water systems. Nosocomial infection and pseudoinfection via water storage

tanks in hospitals have also been described.

Most M. xenopi infections occur in the lungs, usually in male adult patients with underlying

lung disease such as chronic obstructive pulmonary disease or bronchiectasis.

Extrapulmonary infections such as septic arthritis, spondylitis, and disseminated disease

have also been described in immunocompromised individuals (4). Skin and soft tissue

manifestations as well as a recent case report of M. xenopi spondylodiscitis in an AIDS

patient not only highlight its potential pathogenic role but also point to the uncertainties in

therapeutic management (124).

Nontuberculous Slowly Growing Mycobacteria That Are Rarely

Recovered or Rarely Cause Human Disease

Several species of slowly growing mycobacteria including M. gordonae, M.

scrofulaceum, and M. terrae complex are frequently recovered but are rarely associated with

human disease. Some of the case reports of infections attributable to these mycobacteria,

especially from the era before the introduction of molecular laboratory techniques, lack

sufficient documentation of identification or disease association. Other species (such as M.

asiaticum or M. shimoidei) are so rarely recovered that most laboratories will never see

them.

M. asiaticum

M. asiaticum was not recognized as a distinct species until 1971. The photochromogenic

organism has since very infrequently been isolated from patients with respiratory disease in

Australia, the United States, and elsewhere (193). Cases of bursitis and tenosynovitis have

been described as well.

M. celatum

First described in 1993, M. celatum has been isolated from diverse geographic areas

(throughout the United States as well as in Finland and Somalia), mostly from respiratory

tract specimens but also from stool and blood. In one series, 32% of the patients from

whom M. celatum was isolated were infected with HIV. M. celatum has also been isolated

from immunocompetent patients (a child with lymphadenitis and an elderly patient with a

fatal pulmonary infection [195]).

M. celatum shares phenotypical characteristics with MAC, M. malmoense, and M.

shimoidei and thus cannot be identified with conventional tests. Within the bacterial

chromosome, M. celatum has two copies of the 16S rRNA gene. Several subtypes (1 to 3)

have been identified by 16S rRNA gene sequencing or RFLP of the gene encoding the 65-

kDa hsp. Due to high similarities of the 16S rRNA gene sequence with that of M.

tuberculosis a few strains have been misidentified as M. tuberculosis complex by a

commercially available DNA probe or as M. xenopi on account of similar biochemical and

cultural features (195).

M. gordonae

M. gordonae is the most commonly encountered “nonpathogenic” species in clinical

mycobacteriology laboratories. This scotochromogenic species is widely distributed in soil and

water. A pseudo-outbreak associated with drinking water in a French hospital underlined the

necessity for proper maintenance of water supply equipment (107). Convincing evidence

that M. gordonae plays a role in disease is difficult to find (4). There are a few reports of

peritonitis in patients undergoing continuous ambulatory peritoneal dialysis and in renal

transplant patients (160). Eckburg et al. (47) have reviewed clinical and chest radiographic

findings among persons with sputum culture positive for M. gordonae and concluded that it is

a nonpathogenic colonizing organism, even among persons with local or general immune

suppression and abnormal chest X-ray findings.

M. scrofulaceum

The name of this species was derived from scrofula, a historical term used to describe

mycobacterial infections of the cervical lymph glands. Until the 1980s, M. scrofulaceum was

the most common cause of mycobacterial cervical lymphadenitis in children. Since then it

has been replaced primarily by MAC (229). Other types of clinical disease are rare. They

include pulmonary disease, conjunctivitis, osteomyelitis, meningitis, granulomatous hepatitis,

and disseminated disease (4, 52). M. scrofulaceum accounted for 14% of the isolates tested

in respiratory specimens collected from South African miners (35) and for ~2% of the

mycobacterial infections in AIDS patients (4).

M. shimoidei

M. shimoidei was first described in 1988 in a case of a Japanese patient with chronic cavitary

lung disease. Only a few clinical cases have been reported since, mainly in Japan and Finland

(120). It is a thermophilic organism growing well at 45°C. Biochemically, the organism is

similar to M. terrae complex, but it can be distinguished by catalase and β-galactosidase

tests. The unique sequence of the 16S rRNA gene and the 16S-23S rRNA gene spacer region

allow unambiguous identification of the organism (104).

M. terrae complex

M. terrae complex consists of three species, M. terrae, M. nonchromogenicum, and M.

triviale. Clinical disease due to M. terrae is generally limited to tenosynovitis of the hand

following local trauma and pulmonary disease (4). M. nonchromogenicum, ubiquitous in the

aquatic environment, has been the cause of bacteremia in an AIDS patient (4). Separation of

the members of the complex, especially M. terrae from M. nonchromogenicum,requires

molecular methods.

New Species of Nontuberculous Mycobacteria

Most of these species have been described within the past 10 years (195). As a consequence

of more sophisticated molecular technologies applied in the clinical mycobacteriology

laboratory, the number of new species is rapidly increasing (Table 1). However, much less is

known about these species and their clinical relevance remains to be elucidated.

Mycobacterium: Laboratory Characteristics of Slowly Growing

EPIDEMIOLOGY AND TRANSMISSION Back to top

Unlike M. tuberculosis, for which humans are the definitive host, most species of NTM are

widely distributed in the environment, and the occurrence of NTM disease is attributed to a

combination of host factors, such as age, body weight, the presence of chronic lung diseases

(such as cystic fibrosis, bronchiectasis, or chronic obstructive pulmonary disease), alterations

of chest structure, and other conditions, along with exposure. Organisms can be found in

samples of soil and water, including both natural and treated water sources. For example, M.

kansasii, M. xenopi, and M. simiae are almost always recovered from municipal water and

rarely, if ever, from other environmental sources.

Furthermore, for NTM there has been no evidence of animal-to-human or human-to-human

transmission, unlike with MTBC. Human disease due to NTM is assumed to be acquired from

environmental sources either directly by inhaling organisms in aerosols or indirectly by

ingesting contaminated food or water, even though the source of infection may not always

be detected (39).

Incidence rates of NTM disease are only estimates since, unlike with tuberculosis (TB), all

NTM are noncommunicable from human-to-human, and therefore numbers of infections are

nonreportable. One publication from 2007 reported that the isolation prevalence of all NTM

species (excluding M. gordonae) in pulmonary disease in Ontario, Canada, increased from

9.1/100,000 in 1997 to 14.1/100,000 by 2003, with a mean annual increase of 8.4%.

Similar increases were noted for individual species. These findings indicate a significant rise

in pulmonary disease caused by NTM in Canada (65). Increasing numbers of NTM isolates

were also reported for several European countries (66).

The most common infections with NTM currently are pulmonary diseases, but skin and soft

tissue, lymphatic, and disseminated infections are also important. The last most often occur

in the setting of advanced human immunodeficiency virus (HIV) disease, but non-HIVinfected

patients can also be affected. MAC is the most commonly isolated pathogenic slowly

growing NTM, but other species of NTM also cause disease. In the United States, M.

kansasii infection is the second most frequently recovered pathogenic species (39).

CLINICAL SIGNIFICANCE Back to top

Because of the presence of multiple species of NTM in the environment and their

opportunistic pathogenic nature, the determination of the clinical significance of the isolation

of these species is based upon multiple factors, including clinical setting, host-specific

factors, species, the pathogenic potential of the organism, the number of positive cultures,

the source of the culture isolate, and quantification of the organisms detected (by smear and

culture) (39). For example, although the incidence of a specific NTM such as M. gordonae in

cultures is high, the pathogenicity of this species is very low, in contrast to that of species

such as MAC and M. kansasii(7).

Unlike with NTM, the laboratory diagnosis of MTBC is the most important finding in a clinical

mycobacteriology laboratory. The finding of this species has vital epidemiologic and public

health consequences. Further details on clinical significance of the MTBC may be found

in chapter 28 in this Manual.

DIRECT EXAMINATION Back to top

Microscopy

One of the first, easiest, and least expensive means of detecting the presence of

mycobacteria in clinical samples has been microscopic examination. Special acid-fast stains,

along with the use of bright field and fluorescence microscopy, are needed for staining of the

organisms since the routine Gram stain is not optimal for staining mycobacteria. Further

discussion of the specific techniques can be found in chapter 28.

Antigen Detection

Antigen detection is not currently performed for direct detection of mycobacteria in

diagnostic laboratories.

Nucleic Acid Detection

Early detection of disease caused by members of the MTBC is essential in controlling

transmission of TB. Direct nucleic acid amplification (NAA) techniques for the detection of

MTBC bacteria are increasingly being used (26,28, 31, 63, 85, 86, 87, 88, 93). These tests

can provide results in as little as 2 hours, enabling the rapid detection of MTBC from clinical

specimens. In addition to “home-brew” in-house polymerase chain reaction (PCR) assays,

several commercially available kits which are based on either classical PCR or alternative,

isothermal amplification techniques are available (Table 3). Only the

Amplified Mycobacterium tuberculosis direct test (AMTD) has been approved by the U.S.

Food and Drug Administration (FDA) at this time for smear-positive and smear-negative

respiratory specimens (16). For nonrespiratory specimens, no FDA-approved test is

available.

Many publications have reported on the diagnostic values of the different NAA tests. The

most widely used tests in these studies were the AMTD or BD ProbeTec assay (BD

Diagnostics, Sparks, MD) (69, 72, 107, 149). Several reviews and meta-analyses have

evaluated these publications to estimate the diagnostic accuracy of NAA tests

(26, 28, 31, 63, 85, 86, 87, 93).

One main finding among these analyses was the observation of a high degree of variability in

accuracy across the studies. Sensitivity values ranged from 36 to 100% (63) or 27 to 100%

(93), whereas specificity was more consistent, with values ranging from 54 to 100% (63) or

91 to 100% (93). Subgroup analyses could not explain the high variability found in the study

results, even when the same test system was used. An important factor influencing the

estimates was the rate of smear-positive samples included in the studies. The specificities

and positive-predictive values of the tests were higher and more consistent than the

sensitivities, mainly from smear-positive specimens. Thus, a positive NAA test result

combined with a high clinical probability provides a rapid diagnosis of TB. In contrast, a

patient can be presumed to be infected with an NTM if a negative NAA result with inhibitors

excluded was obtained from a smear-positive specimen. As stated above, the sensitivities of

the tests are lower, especially for smear-negative specimens. Combined sensitivity values

from several studies differed markedly, depending on whether the smears were positive or

negative: for AMTD2, the sensitivities were 90 to 100% (smear-positive specimens) and 63.6

to 100% (smear-negative specimens); for ProbeTec, 90 to 100% (smear-positive specimens)

and 33.3 to 100% (smear-negative specimens) (93). A negative test result does not rule out

MTBC infection.

Quality control for the NAA test performance should include controls for the presence of

inhibitors for each specimen to rule out false-negative results. To prevent crosscontamination

with amplification products, a strict workflow must be followed, equipment

must remain in its respective area, and an adequate cleaning procedure must be performed.

The use of closed systems, like real-time analysis, has the advantage that tubes containing

amplification products do not need to be reopened, and thus the risk of cross-contamination

is reduced.

In general, commercially available tests are preferred over in-house assays because of the

standardized protocols and better quality control. A meta-analysis for the use of in-house

NAA tests reported a highly heterogeneous estimate of diagnostic accuracy (range of

sensitivity, 9.4 to 100%; range of specificity, 5.6 to 100%) (31).

In response to the increasing demand for NAA testing for TB and recognition of the

importance of prompt laboratory results in TB diagnosis and control, “Updated Guidelines for

the Use of Nucleic Acid Amplification Tests in the Diagnosis of Tuberculosis” have recently

been released by the Centers for Disease Control and Prevention (CDC) (16). The new

recommendations state that NAA testing should be performed on at least one respiratory

specimen from each patient with signs and symptoms of pulmonary TB for whom a diagnosis

of TB is being considered but has not yet been established and for whom the test result

would alter case management or infection control activities. A detailed testing and

interpretation algorithm is proposed. However, culture remains the reference standard for

laboratory confirmation of TB and is required for drug susceptibility testing and genotyping.

When NAA tests are being implemented, performance characteristics, such as sensitivity and

specificity, need to be established for the laboratory by use of well-known samples.

Proficiency should be examined regularly by comparing NAA test results with microbiological

and clinical data.

IDENTIFICATION Back to top

From a global viewpoint, the rapid detection and identification of M. tuberculosis is the most

important task of a clinical mycobacteriology laboratory. However, within the last 30 years,

the number of currently validated NTM species has increased from approximately 30 to more

than 130 species. This development has been paralleled by an increasing incidence of

infections due to NTM of mainly slowly growing species. Thus, clinical mycobacteriology

laboratories today also encounter the challenge of providing unequivocal identifications of

NTM species. The introduction of molecular techniques into the mycobacteriology laboratory

has dramatically accelerated the diagnostic process. MTBC bacteria can be detected rapidly

in clinical specimens by NAA tests. DNA probes for the confirmation of MTBC and a few NTM

species grown on culture media have been available for many years. Historically, species

identification has relied, with few exceptions, on the analysis of a series of phenotypic tests

for which performance was mostly restricted to specialized laboratories. These tests require a

sufficient amount of bacterial cells and several weeks of incubation. As previously discussed,

it is now recognized that most of the newer mycobacterial species cannot be reliably

identified by biochemical and other phenotypic tests. Several species may exhibit convergent

characteristics, and strains of one species may show variability in certain features; such

specifics are unknown since most new species have not been studied in detail. The most

reliable methods for identification of all mycobacterial species today involve molecular

analyses of certain genes. These techniques have the additional advantage that they can be

performed from liquid culture media, which in general enable more rapid growth and a more

sensitive detection of mycobacteria than solid culture media. Mixed cultures with NTM are

not rare, and thus there is a danger of working with cultures in broth only. Thus, all results

obtained by molecular methods should be confirmed, even after the reporting of the results,

by some important phenotypic characteristics, such as growth rate, colony morphology, and

pigmentation (Table 2)

The Clinical and Laboratory Standards Institute (CLSI) recognizes that not all laboratories

have unlimited funds or instrumentation to provide state-of-the-art testing in

mycobacteriology. Laboratories which have access only to probe technology should probe

isolates to rule out MTBC at a minimum and then refer isolates to a reference laboratory for

further testing. If no technology is available, it is advisable to refer the isolate to a reference

laboratory (18).

Phenotypic Methods

Growth Rate

Growth rate is an obvious property that can be observed with the primary solid culture, at

least within some limits (i.e., dependent on appropriate incubation temperature and number

of organisms in the primary specimen). The estimation is not reliably applicable for growth in

liquid culture medium. To perform a standardized growth test, defined suspensions of

mycobacteria are inoculated on solid media and incubated at 30° and 35 to 37°C. Cultures

are observed for growth at 5 to 7 days and weekly thereafter. Mycobacteria can thus be

classified into the slowly and rapidly growing species. Rapidly growing mycobacteria are able

to form visible colonies within 7 days of incubation, whereas slowly growing species require a

longer period of time for colony formation.

Temperature

Mycobacterium species differ in the ability to grow at certain temperatures. For

determination of the preferred growth temperature, solid culture media are inoculated with

defined suspensions of mycobacteria and incubated at various temperatures. For slowly

growing species, the minimum set of temperatures for incubation comprises 30 } 2°C and

35 } 2°C. Most slowly growing species grow well at 35 to 37°C, althoughM. marinum is an

example of a slowly growing pathogen that grows optimally at a lower temperature (30°C)

(especially on primary isolation). Additional media incubated at 22 to 25°C and 42°C may be

necessary for optimal growth of some species, such as M. haemophilum (25°C), M

xenopi (42°C), and M. stomatepiae (22°C). Cultures to determine temperature requirements

of known rapidly growing mycobacteria can be read within 1 week, while slowly growing

mycobacteria require a longer period of incubation.

Colony Morphology

Colony morphology is a phenotypic property of mycobacteria that can easily be determined.

Morphologic characteristics such as size and colony description (flat, raised, etc.) should be

noted, but most significant is the smooth or rough growth form of the colonies. M.

tuberculosis usually grows in rough, nonpigmented colonies on Lowenstein-Jensen slants

(Fig. 1) as well as on agar-based media. In contrast, M. bovis grows in flat and smooth

nonpigmented colonies. The colony morphology may vary when different formulations of

solid media are used. Slowly growing NTM species may also exhibit either rough or smooth

variants or both types.

Back to top

GENOTYPIC IDENTIFICATION OF MYCOBACTERIAL

SPECIES Back to top

Complete Genome Sequences

Complete genomic sequences have been determined for a series of mycobacterial species: M.

tuberculosisstrain H37Rv, M. tuberculosis strain CDC1551, M. bovis, M. leprae, “M.

avium subsp. hominissuis” strain 104, M. avium subsp. paratuberculosis (MAP) strain K-

10, M. marinum, M. smegmatis, and M. ulcerans. Detailed information on all genome

projects either completed or still in progress can be obtained

athttp://www.ncbi.nlm.nih.gov/sites/entrez?Db=

genomeprj&Cmd=Search&TermToSearch=txid 176. The genomic sizes vary markedly among

the species, from 3.3 million bp for M. leprae (21), with approximately 4.3 million bp for M.

bovis (35) and 4.4 million bp for M. tuberculosis (20), up to 6.6 million bp for M.

marinum (115) and almost 7 million bp for M. smegmatis. Comparative genome analyses of

these species repeatedly show that reductive genome evolution has taken place by the

process of the adaptation of an environmental bacterium to an intracellular, parasitic lifestyle

(38, 116). Genome downsizing, accumulation of pseudogenes, the acquisition of foreign

genes that confer fitness advantages, and a high degree of clonality are common

characteristics of those mycobacteria that are adapting to a new, stable environment in a

host. Complete genome analysis may thus help our understanding of the evolution, host

adaptation, and pathobiology of mycobacterial disease and possible targets for antimicrobial

chemotherapy. Sequence comparisons may aid the identification of new diagnostic as well as

vaccine targets.

Gene Analyses

Partial sequence analysis of selected genes is the most reliable method for identification of

mycobacterial species. Sequence variability among species, but homogeneity within species,

is the basic prerequisite for this application. For several genes (16S rRNA gene, internal

transcribed spacer [ITS], hsp65, and rpoB), specific regions within the gene which work well

for species identification have been identified (1, 32, 53, 55, 56, 70,88, 94, 105, 147). The

most important target for identification of mycobacteria is the 16S rRNA gene. One main

advantage of sequence analysis of this gene is the requirement of most journals for

simultaneous updating of public databases at the time of publication of a new species.

However, quality control and updating of public databases has been a major obstacle in the

identification of new species (134). Although the use of public databases for sequence

comparisons is widespread among investigators, the lack of quality control and monitoring of

these databases, which may include base errors, ambiguous base designations, and

incomplete sequences, has contributed to errors in identification. This seems to be

particularly problematic for isolates that were identified in the early molecular era, when

quality sequences were less available (134).

16S rRNA Gene

For routine identification of mycobacteria, sequence analysis of the complete 16S rRNA gene

(approximately 1,500 bp) is not practical and also not necessary. The information content of

a sequence stretch of approximately 600 bp located at the 5′ end of the gene is sufficient for

identification of most species. This sequence can easily be obtained with one analysis run.

Specific primers for amplification and sequencing of the mycobacterial 16S rRNA gene (Table

5) have been described and validated (8, 23, 30, 43, 45, 55, 90, 113,148).

Primers 264 and 247 target genus-specific sequence regions and can be used for the

detection of mycobacteria even in the presence of contaminating bacteria. Furthermore, the

3′-end primers 285 and B9 for the amplification can be used for the sequence reaction.

Either of these can be combined with any of the reverse primers. Both forward primers are

located upstream of hypervariable regions A and B, which enable the species-specific

identification of most mycobacterial species (55). A few species have identical hypervariable

regions A and B or a complete 16S rRNA gene sequence. This includes M. marinum, M.

ulcerans,M. chelonae, M. abscessus, and M. kansasii (sequence variants I and IV) and M.

gastri. For accurate identification of these species, sequence analysis of other genes (e.g.,

ITS, hsp65, and rpoB) is required. All members of the MTBC have identical 16S rRNA gene

sequences and thus cannot be discriminated by this technique, while for some species, minor

intraspecies 16S rRNA gene sequence variants which differ in a few base pairs have been

observed (M. gordonae, M. kansasii, and M. lentiflavum [54, 100, 114]).

Other Target Genes

Sequencing of several other genes has also been used for identification of mycobacterial

species, but the database for these genes is less complete.

23S rRNA Gene

The 23S rRNA gene is also known to contain conserved and variable sequence regions that

enable the specific amplification and species identification of mycobacteria (62, 136).

Variable regions can also be found in the 5′ region. The disadvantage of this target is the

length of the gene, approximately 3,100 bp, which is not readily analyzed for most of

the Mycobacterium species. Thus, public databases contain few 23S rRNA gene sequence

data from mycobacteria.

ITS 1 Gene

Another target is the spacer sequence, which separates the 16S and 23S rRNA genes in the

operon and is denominated ITS 1. The sequence of this fragment comprises only 200 to 330

bp and thus can easily be analyzed. Several sets of primers which enable the amplification of

the complete fragment and an additional sequence analysis are published (Table 6)

(44, 100, 106).

Primers Sp1 and Sp2 allow the genus-specific amplification of the region. Since they target

the start site of the ITS, they are not optimal for analysis of the whole sequence. Primers ITS

1 and ITS 2 are located in the 16S and 23S rRNA gene regions, respectively. Using these

primers, the entire ITS can be amplified and sequenced. However, they are not genus

specific and cannot be used when cultures of mycobacteria are contaminated with other

bacteria.

For the ITS 1 sequence, a high variability, which could be used for species identification, has

been shown (33,34, 44, 76, 105, 106). However, for some species, mainly for rapidly

growing species but also for some slow growers (M. simiae, M. xenopi) two or more

sequence variants have been observed. As with the 16S rRNA gene sequence, M.

marinum and M. ulcerans have identical ITS 1 sequences and thus cannot be differentiated

by this analysis.

hsp65 Gene

One of the first genetic targets used for the differentiation of mycobacteria is an

approximately 440-bp fragment of the hsp65 gene, which codes for the 65-kDa heat shock

protein and is also known as the groEL2gene (94). The amplification of this fragment,

followed by a restriction enzyme digestion using the restriction enzymes BstEII and HaeIII

and an analysis of the obtained digestion products using agarose gel electrophoresis,

provides restriction fragment length polymorphism (RFLP) patterns that are specific for most

species (27, 121). An algorithm for the differentiation of the most important species showing

the apparent molecular sizes of the fragments has been devised. This PCR-restriction

enzyme analysis technique has been used widely, since this is a simple and rapid method

with no need for more sophisticated sequencing techniques. However, technical difficulties,

such as small size differences between the fragments or the incidence of similar or identical

restriction patterns for some species of closely related mycobacteria and, in particular, new

species of mycobacteria, are major problems with using this method. With the greater

availability and decreasing cost of sequencing, sequence analysis of this hsp65 gene

fragment has been increasingly used instead of the PCR-restriction enzyme analysis

(70, 104).

Advantages of sequencing the hsp65 gene rather than the 16S rRNA gene are especially

evident in the identification of closely related species, as the hsp65 gene is much less

conserved than the 16S rRNA gene.hsp65 gene sequencing allows for differentiation of M.

marinum from M. ulcerans, M. gastri from M. kansasii,and, with some restrictions, M.

avium subsp. avium from M. avium subsp. hominissuis.

rpoB Gene

The rpoB gene encodes the β subunit of the bacterial RNA polymerase. For M.

tuberculosis, mutations in a certain region of this gene are known to confer resistance to

rifampin. For NTM species, sequence variability that can be used for species identification has

been shown (1, 53, 60). Several different sequence fragments of the approximately 3,600-

bp gene have been used for amplification and sequence determination. Kim et al. (53)

amplified a fragment comprising 306 bp at positions 1362 to 1668 (referred to as region

2/3), whereas Lee et al. (60) targeted a 360-bp sequence at positions 902 to 1261 (referred

to as region 1/2) (1). In contrast, Adekambi et al. (1) chose a fragment more distant (region

5). Those authors analyzed an approximately 760-bp fragment at positions 2573 to 3337. In

the studies of Kim et al. (53) and Lee et al. (60), type strains of many slowly growing

mycobacteria have been included, whereas detailed analyses using region 5 (1) have been

performed mainly for rapidly growing mycobacteria. Extensive investigations using clinical

isolates of slowly growing mycobacteria are not available for any part of the rpoB gene so

far. Thus, currently, no final recommendation has been made to determine which fragment is

most suitable for the identification of slowly growing mycobacteria.

gyrB Gene

The gyrB gene encodes the B subunit of DNA gyrase (topo isomerase II), an enzyme

essential for bacterial replication. In 2000, Kasai et al. (52) showed single nucleotide

polymorphisms (SNPs) in an approximately 1.2-kbp fragment of the gyrB gene which were

specific for some species of the MTBC. More-detailed analyses confirmed these results and

determined that they could be extended to most members of the MTBC (82).

For M. microti and M. caprae (previously M. bovis subsp. caprae), unique characteristic SNPs

are known. Identical gyrB sequences are shared by M. tuberculosis and “M. canettii,” by M.

africanum and M. pinnipedii, and by M. bovis (synonym, M. bovis subsp. bovis) and BCG.

However, “M. canettii” is rarely observed outside East Africa and M. pinnipedii has so far

been exclusively isolated from seals and sea lions, with one exception of infection in a seal

trainer (24, 143). Thus, these two species are usually not observed in routine clinical

laboratories. In the case of M. bovis and BCG, a definite identification is essential for the

diagnosis. To finally identify the species, additional techniques, such as estimation of the

presence or absence of region of difference 1 (RD1) by PCR analysis, are necessary (120) to

differentiate BCG from M. bovis. Identification of the characteristic SNPs can be performed by

either sequence analysis (Tables 7 and 8) or by restriction enzyme digestion using RsaI,

SacII, or TaqI and an additional agarose gel electrophoretic analysis (82).

TAXONOMY AND DESCRIPTION OF THE AGENTS Back to top

Rapidly growing mycobacteria (RGM) are generally defined as nontuberculous species that

grow within 7 days on laboratory media (9). RGM contain long-chain fatty acids known as

mycolic acids that can be quantitated using chromatographic techniques, such as highperformance

liquid chromatography (HPLC). Before the molecular era, HPLC was used for

identification of species of RGM in most major reference laboratories. However, this method

has been replaced in most laboratories by more definitive molecular identification methods

for more accurate species identification (76).

Currently, there are more than 130 known species of nontuberculous mycobacteria (NTM), of

which 70 are species of RGM. More than one-half of RGM species have been described since

the early 1990s. Since 2007, 13 new species have been added to the list: M. aubagnense, M.

monacense, M. phocaicum, M. setense, M. aromaticivorans, M. crocinum, M.

fluoranthenivorans, M. insubricum, M. llatzerense, M. pallens, M. pyrenivorans, M.

rufum, and M. rutilum (35, 43, 56, 77, 78). The first four species have been associated with

human, animal, or fish disease, while the latter nine species have thus far been considered

environmental nonpathogens. Additionally, most recently, the species M. abscessus has been

subdivided into two subspecies. The former species, M. abscessus, is now M.

abscessus subsp. abscessus, and two previously described species, M. massiliense and M.

bolletii, were determined to compose another subspecies of M. abscessus now known as M.

abscessus subsp. bolletii (39, 43a, 44).

There are currently six major groups or complexes of RGM based on pigmentation and

genetic relatedness (Table 1). Nonpigmented pathogen species now are composed of 11

species within the M. fortuitum group (43, 87, 90) and the former third biovariant complex

(9, 64, 89).

The second group of nonpigmented RGM is the M. chelonae/M. abscessus group, as listed

in Table 1 (1–3, 5, 9, 44). A previously described species, M. salmoniphilum, has recently

been revived and is also considered to be related to this group. Although this species has

been recovered from salmon and trout with disseminated disease, it has not yet been

recovered from humans (92).

A third nonpigmented group, the M. mucogenicum group, currently includes the three

species noted in Table 1(1, 9, 69).

The fourth group, the M. smegmatis group, is currently composed of the two late-pigmenting

species, M. smegmatis (formerly M. smegmatis sensu stricto) and M. goodii (7, 9, 89).

The fifth group of RGM includes the early-pigmented species, which traditionally have been

difficult to identify by conventional (phenotypic) laboratory methods. The only proven

pathogen in this group is M. neoaurum(Table 1) and several newly described species,

including M. canariasense, M. cosmeticum, and M. monacense. There are a number of

previously listed environmental (nonpathogenic) species as well (9, 35, 43, 66, 77, 78).

Current studies based on DNA sequence analysis suggest a sixth group, composed of M.

mageritense and M. wolinskyi (2, 3, 76).

The recent introduction of several new species within the RGM highlights the importance of

molecular identification of these organisms to the species level and questions the

meaningfulness of the current “group” classification, especially within the M. fortuitum group.

However, because previous data and publications use this “group” nomenclature, this

designation is retained in this chapter for ease of discussion (9).

CLINICAL SIGNIFICANCE Back to top

The RGM are opportunistic pathogens that produce disease in a variety of clinical settings.

The three major clinically important species of RGM responsible for approximately 80% of

disease in humans include M. fortuitum, M. chelonae, and M. abscessus (30). Other

potentially pathogenic and clinically significant RGM species have been included in Table

2 (1–5, 7, 9, 28, 43, 52, 56, 63–66, 76–78, 89, 92). RGM are presumed to be common in

the environment but have been most often identified in tap water when associated with

outbreaks of catheter sepsis in bone marrow transplants, wound infections, and associated

pseudo-outbreaks of disease (9). The specific reservoir for M. abscessus chronic lung

infections has yet to be identified.

Skin and Soft Tissue Infections

The most common infection seen with RGM is a posttraumatic wound infection. Patients are

generally healthy, and drug-induced immune suppression results in a minimal increase in

risk for this type of infection. The M. fortuitum group accounts for approximately 60% of

cases of localized cutaneous infections, but any of the more than 30 pathogenic RGM species

listed in Table 2 can cause disease (9, 21, 48, 64).

Traumatic wound infections, especially open fractures, often involve species within

the Mycobacterium fortuitum third biovariant complex (9, 21). More than 75% of the

infections reported from a series of 85 isolates of the M. fortuitum third biovariant complex

from the United States and the Queensland, Australia, state laboratory were associated with

skin, soft tissue, or bone infections (9). The majority of infections occurred 4 to 6 weeks

following puncture wounds or open fractures. Metal puncture wounds (48%) and motor

vehicle accidents (26%) were the most common antecedent injuries, and approximately 40%

of the injury sites involved the foot or leg. Stepping on a nail was the most frequently related

scenario. None of the isolates in this series were studied by molecular techniques that would

identify them as one of the newly described species within the M. fortuitum third biovariant

complex (i.e., M. houstonense, M. boenickei, and M. porcinum).

In a 1989 report (64), approximately 80% of RGM wound isolates related to cardiac surgery

were from seven southern coastal states, including Texas, Louisiana, Georgia, Maryland,

Alabama, Florida, and South Carolina. A second report published in the same year showed

that 92% of 37 identified cases of surgical wound infection following augmentation

mammaplasty were also from patients in southern coastal states, with the majority being in

Texas, Florida, and North Carolina, suggesting that the disease risk was highest in the

southeastern United States (9).

Sporadic cases of localized wound infections following medical or surgical procedures,

including needle injections, can occur with M. chelonae but are less common than those

with M. fortuitum. The clinical picture of posttraumatic wound infection ranges from localized

cellulitis or abscesses to osteomyelitis (9). A 2006 report from a major U.S. clinical referral

center of patients from Minnesota, Wisconsin, Iowa, and South Dakota characterized 63

human immunodeficiency virus-negative patients with RGM infections involving M.

abscessusor M. chelonae (71%) or the M. fortuitum group (29%). Moreover, patients with M.

chelonae or M. abscessususually had multiple (disseminated) cutaneous lesions, in contrast

to those with single (localized) lesions due to M. fortuitum. Most patients with M.

fortuitum had undergone a prior surgical procedure or had experienced a penetrating trauma

at the infected site. Patients with M. chelonae or M. abscessus were older and more likely to

be on some type of immunosuppressive agent (82). Localized or disseminated infections

with M. chelonaemost frequently occur in patients receiving long-term corticosteroids and/or

chemotherapy, organ transplant recipients, patients with rheumatoid arthritis or other

autoimmune disorders, or patients receiving suppressive therapy (82). Immune suppression

in patients with diseases such as AIDS has not been a significant risk factor for development

of localized or disseminated M. chelonae infections (9, 82).

Starting in 2000, an outbreak of furunculosis caused by M. fortuitum on the lower

extremities was described for 32 otherwise healthy patients who were patrons of a nail salon

in California (94). The organism was also cultured from contaminated foot baths and from

the inlet suction screens containing hair and other debris, and shaving the legs prior to the

footbath and pedicure was an identified risk factor (94, 95). Other species, including M.

fortuitum, M. abscessus, and M. mageritense, have subsequently been recovered from cases

in California and Georgia (9, 27, 68, 85). Strains from the footbath and from patients were

identical by DNA strain typing.

Occasionally, M. wolinskyi, M. mageritense, and members of the M. smegmatis group have

been reported from infections following traumatic injury and surgical or medical procedures,

such as cardiac surgery, breast reduction surgery, and face-lift plastic surgery. Cellulitis and

localized abscess are the most common manifestations (7, 9, 30).

Disseminated Cutaneous Disease

Disseminated cutaneous disease due to members of the M. fortuitum group (including M.

fortuitum) is rare even in immunocompromised patients, including those with AIDS

(9, 13, 14, 30).

Disseminated cutaneous disease due to M. chelonae is much more common. It typically

presents as multiple chronic painful red nodules, usually involving the lower extremities

(9, 30). These lesions then drain spontaneously, with the drainage usually being acid fast

bacillus smear positive. Almost all patients are immunosuppressed, usually from low-dose

corticosteroid therapy. Although the disease is presumably a consequence of hematogenous

spread, bacteremia is rarely identified. A portal of entry for the infection is rarely evident

(9, 30). In a series of 100 clinical isolates from skin and soft tissue, Brown-Elliott and

Wallace reported that 53% were from patients with disseminated cutaneous infections (9).

Disseminated cutaneous disease due to M. abscessus occurs rarely but is serious (9). As with

disseminated M. chelonae disease, most cases occur in chronically immunosuppressed

patients receiving corticosteroids, and the disease has no apparent portal of entry. Also, as

with patients with M. chelonae, patients with disseminated cutaneous infection due to M.

abscessus rarely have detectable bacteremia and/or endocarditis but usually present with

multiple draining cutaneous nodules, usually in the lower extremities (9, 30).

A rare type of disseminated infection due to RGM in immunocompetent hosts presenting with

lymphadenopathy has recently been described (12).

Bone and Joint Infections

RGM may also cause bone and joint infections. Like with bacterial disease, osteomyelitis may

follow open bone fractures, puncture wounds, and hematogenous spread from another

source. The most common scenario is an open fracture of the femur, often followed by

orthopedic surgical procedures. The most frequent pathogen recovered in this setting is a

member of the M. fortuitum group, including newly described species, namely, M.

houstonense, M. boenickei, and M. setense (9, 20, 43, 64, 75). The two newly described

species in the M. smegmatis group, M. goodii and M. wolinskyi, have also been associated

with osteomyelitis (7, 9). Bone involvement secondary to a puncture wound is likely the

second major cause of osteomyelitis. Infections most commonly involve members of the M.

fortuitum group (9). M. fortuitum infections in prosthetic knees and joints have also been

reported (20). Vertebral osteomyelitis has also been described (54, 62).

Pulmonary Infections

Chronic lung infections can occur with RGM, most often in nonsmoking older women with

bronchiectasis, and are sometimes associated with M. avium complex (MAC) as well. M.

abscessus is the causative agent in >80% of cases of pulmonary disease due to RGM (30).

Similarities exist between patients with MAC and those with M. abscessus such that a

common pathogenicity or host susceptibility factor may be involved (30). Multiple cultures

of M. abscessus from respiratory samples are usually associated with significant pulmonary

disease.

Patients with cystic fibrosis (CF) may also become infected with both subspecies of M.

abscessus, and this species has been isolated with increasing frequency from the respiratory

tracts of patients with CF (9, 18, 19,36, 51). M. abscessus is the second-most-common

species of NTM recovered in CF patients (after MAC) and may be the most-common species

associated with clinical disease in this setting (51). Patients with CF also have bronchiectasis

in addition to chronic, recurrent airway and parenchymal infections, which may be the

primary risk factors for susceptibility to NTM disease (36, 51).

Other RGM, including M. chelonae, members of the M. smegmatis group, and M.

fortuitum are less frequently associated with pulmonary disease. M. fortuitum has been

reported as a pathogen in half of the cases of chronic aspiration disease secondary to

underlying gastroesophageal disorders such as achalasia (9, 30). Pulmonary disease with M.

fortuitum in the absence of these disorders is rare. Pulmonary disease with M.

chelonae and M. smegmatis has been described for only a few cases, including lipoid

pneumonia (9, 30).

Hypersensitivity pneumonitis among metal grinders in industrial plants working with

contaminated metalworking fluids has been associated with a newly described species of

RGM, M. immunogenum (9, 30, 93). Multiple pseudo-outbreaks associated with this species,

resulting from contaminated automated bronchoscope cleaning machines and metalworking

fluids, have been reported. This species is able to grow and remain viable in degraded

metalworking fluid and is resistant to the routine biocides used for disinfecting the

metalworking fluids (9, 93). However, this species has not yet been reported from open-lung

biopsy specimens from these patients.

Central Nervous System Disease

Central nervous system disease involving RGM is rare, but morbidity and mortality are high.

Most of the reported cases have been associated with M. fortuitum (9, 22, 86).

Corneal Infections (Keratitis)

The number of RGM recovered from ocular infections has been increasing over the last 20

years. A retrospective review of cases of NTM keratitis from 1982 to 1997 at an eye institute

in Florida showed that 19 out of 24 cases were due to RGM (24). A recent study which

identified 113 ophthalmic isolates by molecular methods showed that the most common RGM

were M. chelonae (45%), M. abscessus (42%), and members of the M. fortuitum group (8%)

(8).

Since the early 1990s, other descriptions of epidemic and sporadic ocular infections

associated with RGM, including after keratoplasty and following laser in situ keratomileusis

(LASIK) surgery, have been published (8,9).

Otitis Media

The most common NTM associated with chronic otitis media is M. abscessus. In a 1988

outbreak of 17 cases of otitis media in two ear, nose, and throat clinics, patients presented

with chronic ear drainage, a perforated tympanic membrane, and a prior tympanostomy tube

(9). In another series, 20 of 21 cases of sporadic chronic otitis media (some with associated

mastoiditis) were due to M. abscessus following ear tube placement. Approximately one-half

of the isolates from these cases were aminoglycoside resistant as a result of the patients’

long-term use of aminoglycoside ear drops (9).

Health Care-Associated Infections

Previously, health care-associated disease resulting from RGM has been reported most

commonly with M. fortuitum, M. chelonae, M. abscessus, and M. mucogenicum, although any

species may be involved. Most infections follow contamination with tap water (9, 17, 21, 88).

Types of infections include postsurgical wound infections, catheter sepsis, infections following

hemodialysis, postinjection abscesses, vaccine-related outbreaks, and otitis media following

tympanostomy tube replacement (9, 26, 40). These have been seen as both sporadic cases

and localized outbreaks. Recent outbreaks have involved cosmetic procedures such as

liposuction, liposculpture, acupuncture, and mesotherapy, a procedure comprised of multiple

subcutaneous injections of pharmaceutical or homeopathic medications for cosmetic

purposes (6, 25, 26, 40, 47, 48, 83, 84,97).

Recovery of both subspecies of M. abscessus has been reported from outbreaks of infections

associated with laparoscopic surgeries and cosmetic surgeries in Brazil, the Dominican

Republic, and Korea (25, 40, 83). These and other recent reports suggest that although few

studies have identified these newer subspecies in invasive infections, they have been

misclassified in previous studies (67).

Central-catheter-associated infections are the most common health care-associated infection

due to RGM (9,55, 63). They are the most common cause of RGM bacteremia, but the

disease may also present as local wound drainage as part of an exit site or tunnel infection

(9, 55). Other types of catheters can also lead to infections, including peritoneal catheters,

ventriculoperitoneal shunts, and shunts for hemodialysis (9). The most common species

are M. mucogenicum, M. fortuitum, and M. abscessus. Recently, an outbreak of M.

phocaicum and M. mucogenicum was described for five patients with central venous

catheters in an oncology unit in a Texas hospital (16). This outbreak represents the first

report of clinical isolates of M. phocaicum in a hospital in the United States.

Surgical wound infections due to RGM are a well- recognized clinical entity. In the 1970s and

1980s, these were most commonly associated with augmentation mammaplasty and

coronary artery bypass surgery, and multiple disease outbreaks occurred (9, 60). Infections

following these types of surgery are now less common, although recently a cluster of 12

cases of postaugmentation mammaplasty surgical-site infection due to M. fortuitum and M.

porcinum was reported between 2002 and 2004 in Brazil (60). More often, however, these

types of infections have been replaced by infections following other types of cosmetic

surgeries, such as liposuction, and other types of prosthetic surgeries, such as knee

replacements.

Infections following insertion of prosthetic devices, including prosthetic heart valves, artificial

knees and hips, lens implants, and metal rods inserted into the vertebrae to stabilize bones

following fractures, have also been described (9, 20). Again, M. fortuitum is the most

common pathogen, but any of the pathogenic RGM, including members of the M.

smegmatis group, can be associated with this type of infection (9).

In addition to true outbreaks of infection, numerous health care-associated pseudooutbreaks

have been described. Contaminated or malfunctioning bronchoscopes, automated

endoscope cleaning machines, and contaminated laboratory reagents and ice have been

implicated (9, 31, 33, 42, 88, 91).

RGM are generally resistant to the activities of biocides, such as organomercurials, chlorine,

2% formaldehyde, and alkaline glutaraldehyde, all of which are commonly used disinfectants

(9). A report of the contamination of benzalkonium chloride, a widely used antiseptic

compound, with M. abscessus and of resulting outbreaks of this species in several patients

following steroid injections after skin disinfection emphasizes the limitations of disinfectants

against RGM (74).

COLLECTION, TRANSPORT, AND STORAGE OF

SPECIMENS Back to top

Details of standard methods are included in chapter 28. Transport of species is accomplished

by using leak-proof containers and proper safety protocols. Specimens for detection of RGM

should be delivered to the laboratory in a timely manner by following appropriate shipping

and handling regulations for shipping biological or infectious materials (23).

DIRECT EXAMINATION Back to top

Microscopy

The use of mycobacterial smears is a rapid and reasonably sensitive step in the diagnosis of

RGM disease. Gram staining of colonies showing faintly staining “ghost-like” beaded grampositive

bacilli is often helpful in establishing a diagnosis of mycobacteriosis. Ziehl-Neelsen or

Kinyoun stain may also be useful. However, a smear alone is not sufficient to identify

species. A large study found that NTM, including RGM, are likely to be detected by

fluorochrome staining of specimens, especially from patients at low risk for AIDS in areas in

which lung disease is endemic (23, 98). Further details of the staining procedures are found

in chapter 28.

Nucleic Acid Detection

Currently there are no commercial systems available for direct detection of NTM. An indirect

test could be an acid-fast bacillus smear-positive sample that is negative by the commercial

MTB direct test, but this does not prove the presence of RGM as distinguished from slowly

growing NTM.

ISOLATION PROCEDURES Back to top

Primary isolation of RGM optimally requires culture at 28 to 30°C rather than 35°C,

especially for recovery ofM. chelonae and M. immunogenum (9). Direct examination and

isolation procedures are detailed in chapter 28.

Recent studies of a murine model suggest that colony morphology, such as a smooth or

rough phenotype, may be related to the invasiveness of strains of M. abscessus. The smooth

phenotype has been associated with biofilm production and a lack of infectivity; in contrast,

strains of the rough phenotype do not form biofilms and invade macrophages (29).

IDENTIFICATION OF RGM Back to top

Biochemical Testing

As previously stated, RGM are defined as NTM that grow within 7 days (most species grow

within 3 to 4 days) (9). Until the advent of more-modern molecular techniques, traditional

laboratory identification of RGM was based primarily upon growth rate, pigmentation,

colonial morphology, and a select battery of biochemical tests (9). These standard tests

include arylsulfatase production, tolerance to 5% NaCl, nitrate reductase activity, and iron

uptake. All members of the M. fortuitum group and M. chelonae/M. abscessus group exhibit a

strongly positive arylsulfatase reaction at 3 days. The M. smegmatis group (M.

smegmatis and M. goodii) and M. wolinskyi are similar in growth rate but do not exhibit

arylsulfatase activity at 3 days (9). Approximately 95% of the isolates of M.

smegmatis (sensu stricto) and 80% of M. goodii develop a late (7 to 10 days) yellow-orange

pigmentation (5, 9).

The current proposal for clinical laboratories is that biochemical testing of RGM should be

replaced with molecular methods. Moreover, biochemical testing should be performed only

when a new species is being described as part of a polyphasic identification algorithm.

Supplemental Biochemical Testing: Carbohydrate Utilization

The supplementation of standard biochemical tests with carbohydrate utilization has allowed

more complete and accurate laboratory identification of established species and

discrimination of some (but not all) new species (9). Identification to the species level and

susceptibility testing (see chapter 73) should be performed on isolates of RGM considered to

be clinically significant.

However, as previously stated, molecular testing is the only definitive means of identifying

RGM species, and laboratories should proceed cautiously when identifying these species by

biochemical testing alone.

Antimicrobial Susceptibility Tests for Identification

As discussed above, other adjunctive nonmolecular tests, including antimicrobial

susceptibility tests, have also been utilized for the identification of RGM (9). As a screening

tool, isolates of the M. fortuitum group and the M. chelonae/M. abscessus group can be

differentiated by the use of a polymyxin B disk diffusion method. Generally, isolates of the M.

fortuitum group exhibit a partial or clear zone of inhibition (≥10 mm) around the polymyxin

disk, whereas isolates of the M. chelonae/M. abscessus group show no zone of inhibition (9).

Isolates of the M. fortuitum group are usually susceptible to a broad range of antimicrobials,

including amikacin, quinolones, sulfonamides, linezolid, and imipenem. Most of the isolates of

this group appear to be intrinsically resistant to the macrolides due to the presence of

several related inducible erm genes (9, 49).

Moreover, M. chelonae and M. abscessus also have different antimicrobial susceptibility

patterns. One major difference between the two species is resistance to cefoxitin. By agar

disk diffusion, M. chelonae shows complete resistance to cefoxitin, with no partial or

complete zones of inhibition, in contrast to the partial or complete zones seen with M.

abscessus. The MICS of cefoxitin for isolates of M. chelonae are generally ≥256 μg/ml,

whereas the modal MIC for isolates of M. abscessus is 32 to 64 μg/ml (9). Furthermore,

recent studies have shown that isolates of M. abscessus, but not M. chelonae, have an

inducible erm gene similar to the gene in M. fortuitum, which conveys macrolide resistance

(49, 50).

The MICs of amikacin for isolates of M. abscessus are lower than those for M. chelonae, and

these isolates are resistant to tobramycin, whereas tobramycin is more active than amikacin

against M. chelonae. (Amikacin is the preferred aminoglycoside for treatment of M.

abscessus, while tobramycin is the preferred agent for M. chelonae.) Additionally, isolates

of M. chelonae are more susceptible in vitro to some of the newer antibiotics, including

linezolid and moxifloxacin, than are isolates of M. abscessus (5).

With these differences in susceptibility patterns of the rapidly growing species in mind,

tentative identification of the most commonly encountered species of RGM is possible and

optimal therapeutic regimens can be designed. However, as with other phenotypic tests,

susceptibility testing does not provide definitive species identification. Confirmation of

antimicrobial susceptibility patterns in newer species will require testing larger numbers of

isolates that have been identified to the species level by molecular techniques.

HPLC Identification

HPLC analysis of mycobacterial cell wall mycolic acid content is routinely used in large

reference or state health department laboratories to identify slowly growing isolates of NTM

but has been problematic with RGM (9, 10). HPLC can be helpful for placing RGM into groups

or complexes but is not specific enough to identify most species with a high degree of

accuracy.

Molecular Identification Methods

Nucleic Acid Probes

The INNO LiPA multiplex probe assay (Innogenetics, Ghent, Belgium) is based on the

principle of reverse hybridization (79). Although the assay has not received Food and Drug

Administration (FDA) clearance, it is being used in some U.S. research laboratories. The

assay can identify both rapidly and slowly growing mycobacterial species. Biotinylated DNA

obtained by PCR amplification of the 16S–23S internal transcribed spacer region is hybridized

with specific oligonucleotide probes immobilized as parallel lines on membrane strips. The

main advantage of this system is that a large variety of species may be identified by a single

assay without the need to select an appropriate probe. One limitation of the assay is the

cross-reactivity that may be detected between species of the M. fortuitum group and several

species that are rarely found in clinical samples, such as M. thermoresistibile, M. agri, and M.

alvei (9, 79). Additionally, it failed to differentiate isolates of closely related species, such

as M. chelonae, from M. abscessus. Since the original studies with the INNO-LiPA assay,

however, the system has been improved to include additional probes for the M. fortuitum-M.

peregrinum complex and M. smegmatis.

A similar commercial PCR method which targets the 23S rRNA gene, the GenoType

Mycobacterium assay (Hain Lifescience, GmbH, Nehren, Germany), provides probes for

simultaneous identification of M. chelonae and specific probes for M. peregrinum, M.

fortuitum, and M. phlei. These two systems are widely used in Europe for NTM identification

(57, 59).

Sequence Analysis for Identification of RGM

Nucleic acid sequence analysis has been performed for the identification of mycobacteria for

several years. This identification tool has been useful for the discrimination of most of the

newly described species of RGM (32, 41, 52, 81).

16S rRNA Gene Sequence Analysis

Generally, the identification of mycobacteria, including RGM, focuses on two main

hypervariable domains known as region A and region B, located on the 5′ end of the 16S

rRNA gene. These regions correspond toEscherichia coli positions 129 to 267 and 430 to 500,

respectively. Hypervariable region A, especially, contains most of the species-specific

sequence variations (so-called “signature sequences”) in mycobacterial species, and

sequencing of this region allows taxonomic identification of most mycobacteria, including

many species of RGM (52, 76).

It is important to note that isolates of two major RGM pathogens, M. chelonae and M.

abscessus, require sequencing of sites outside regions A and B, as they are identical in

regions A and B but differ at other 16S rRNA gene sites (in the 3′ region), though only at

four base pairs (32, 41, 52). To ensure accurate identification, sections of at least 300 bp of

quality sequence should be compared between the reference and the query sequences and

cover at least one region of the gene where variations are to be expected. Therefore,

currently, most clinical laboratories sequence between 450 and 480 bp in order to provide an

adequate sequence (53).

In general, members of the genus Mycobacterium are closely related to each other, and

closely related species may differ by only a few base pairs or none at all. For example, M.

goodii, which is phenotypically difficult to distinguish from M. smegmatis except by

susceptibility pattern, has only a four-base difference in its 16S rRNA gene from that of M.

smegmatis (5).

A commercial gene sequencing system, the MicroSeq 500 16S rRNA gene bacterial

sequencing kit (Applied Biosystems, Foster City, CA), analyzes the first 500-bp sequences

and compares the sequences with a commercially prepared database. The use of this

commercial system alone cannot differentiate some major species of RGM, such as M.

chelonae and M. abscessus, which require sequencing of other regions or other genes for

identification (32, 81). This lack of entries makes identification of unknown strains difficult

when there is no exact match in the database. For this reason, clinical laboratories have

supplemented the commercial database with additional sequences from their own or other

libraries, such as RIDOM or GenBank.

A quality-controlled database is indispensable for the evaluation and accurate identification

of unknown strains (81). The laboratorian should also recognize that sequence analysis is an

important component in a polyphasic approach to the identification of unknown strains. While

in some instances, molecule-based identifications without conventional testing may be

adequate, more often there are cases in which the broader picture must be reviewed. Some

investigators suggest that key phenotypic tests, including colonial morphology, pigmentation,

and growth rate analyses, are necessary, especially in the differentiation of closely related

species (81).

The lack of consensus for a standard reporting criterion or cutoff value has been a major

obstacle in the interpretation of sequence data (53, 81). A reporting criterion such as (i)

distinct species, (ii) “related” to a species, or (iii) “most closely related to” a species,

depending upon the amount of sequence difference between the unknown isolate and the

16S rRNA gene database entries (40), has been recommended but not validated (52, 81).

A recent Clinical and Laboratory Standards Institute (CLSI) document (53) has

recommended guidelines for 16S rRNA gene sequencing in order to

identify Mycobacterium sp. in a consistently practical manner. For sequences with 100%

sequence probability, a definite genus and species may be assigned. However, for sequence

probabilities from 99.0 to 99.9%, the document recommends reporting the isolate as “genus,

most closely related to species,” and for isolates with a sequence probability of ≥95% to

98.9%, laboratories should consider reporting as “unable to definitively identify by 16S rRNA

gene sequencing, most closely related to Mycobacterium sp.” (53). We agree that although

100% identity is mandatory for signature sequences, one or a few mismatches at other

positions may be acceptable for identification to the species level (53).

Despite the availability of a commercial sequencing method, sequencing remains a complex

and often cost-prohibitive procedure for a routine clinical laboratory, which also may not

have an adequate volume of isolates to warrant sequencing. Therefore, for these reasons,

the general consensus of opinion is that not all laboratories should attempt to incorporate

sequencing into their laboratory routine. Moreover, requests for sequencing should instead

be sent to a qualified reference laboratory with skill and experience in the method (53, 81).

Sequencing of the hsp65 Gene

Although the 65-kDa heat shock protein gene (hsp65) is highly conserved among species of

mycobacteria, it exhibits greater interspecies and intraspecies polymorphism than the 16S

rRNA gene sequence (58, 71, 72). This variability can be advantageous to the development

of other strategies for the identification of genetically related species of RGM (71, 72). Most

sequencing or restriction fragment polymorphism analyses have utilized a 441-bp sequence

identified by Telenti et al. (72) and often referred to as the Telenti fragment.

Studies based on DNA sequencing have demonstrated interspecies allelic diversity within

RGM. Detailed studies of several RGM species, including M. peregrinum, M. porcinum, M.

senegalense, M. chelonae, and M. abscessus, have shown four to six sequence variants

(sequevars) per species that differ by 4 to 6 nucleotides within the 441-bp Telenti fragment

(46, 72).

Additionally, unlike the 16S rRNA gene sequencing method, the hsp65 sequencing method is

able to differentiate isolates of M. abscessus from M. chelonae (they differ by almost 30 bp in

their 441-bp hsp65sequences, compared to a difference of only 4 bp in their entire 1,500-bp

16S rRNA gene sequences) (46). Unlike with 16S rRNA gene sequencing, with sequencing of

the hsp65 gene, even RGM species with a high degree of 16S rRNA gene similarity, such

as M. fortuitum, M. septicum, M. peregrinum, M. houstonense, and M. senegalense, can be

discriminated as distinct species.

As for other sequencing methods, one limitation of sequencing of the hsp65 gene is that few

or no sequences of newer RGM species are available in databases, and detailed sequencing

of older species (i.e., multiple strains) has not been done, such that only one sequence per

species is generally available. Thus, development of a comprehensive database and in-house

validation are essential (46, 76, 81).

rpoB Sequence Analysis

Initial studies using the rpoB gene for description of species were based upon analysis of a

partial rpoB gene sequence, comprising only about 20% of the entire gene length. Other

investigators have suggested that species identification of a variety of RGM is possible using

a 340- to 360-bp region, but extensive variation may require development of more speciesspecific

probes (1–3, 37, 38, 45).

The utility of rpoB has recently been emphasized in a study comparing the phylogenetic

relationships of therpoB genes of 19 RGM, including the major pathogens in this group, to

several different sequence targets, including 16S rRNA, hsp65, sodA, and recA. All 19 species

showed good discrimination with the rpoB gene (3,19).

Not only has the rpoB sequence been useful for identification of the established species, but

it has also helped to enable the discrimination of species that could not be differentiated by

16S rRNA or the hsp65 gene sequence alone. The rpoB fragment can be sequenced directly

in both directions, allowing identification of most currently recognized species of RGM (2, 3).

Newly described species that are usually differentiated byrpoB include M.

abscessus subsp. bolletii, M. phocaicum, and M. aubagnense (1, 2, 5, 39). However, recent

studies have shown that multilocus sequencing is necessary to identify M.

abscessus subsp. massiliense (43a,44, 100).

Sequence Analysis of Other Gene Targets

Other molecular targets for taxonomic identification, including the 32-kDa protein gene, the

superoxide dismutase (sod) gene, the dnaJ gene, the 16S–23S rRNA internal transcribed

spacer, the secA1 gene, and therecA gene, have been suggested for mycobacterial

identification utilizing either PCR-restriction fragment length polymorphism analysis (PRA) or

direct sequencing (3, 14, 15, 99). However, preliminary data suggest that these gene

sequences are more varied than the hsp65 gene sequence and, to date, have been less

commonly utilized in laboratory identifications of the species of RGM (3, 100, 101).

Moreover, a major limitation for all sequence-based testing is the lack of sufficient databases

(2, 81). Additionally, a multigenic approach for taxonomic evaluation of species has been

widely suggested by investigators and has recently been proposed by the ad hoc committee

for the reevaluation of the species definition in bacteriology (70).

PCR Restriction Enzyme Analysis (PRA)

PRA of the hsp65 gene has become a valuable tool used in the identification of RGM. Using

the nonsequencing method PRA on the hsp65 gene (sequevars) of a species to determine

minor differences from the hsp65 genes of other species rarely involves a restriction site, so

most species have only one PRA pattern. Currently, the 441-bp Telenti fragment of

the hsp65 gene remains the most useful sequence for PRA identification of RGM, although it

has not been evaluated extensively in pigmented RGM and with the newer species and

subspecies of RGM, such as M. phocaicum, M. aubagnense, M. abscessus subsp. bolletii, and

others (9, 71, 72).

The advantages of PRA are that the method of identification does not rely upon growth rate

and nutritional requirements, the equipment is relatively inexpensive, and the results for a

large number of mycobacterial species can be generated rapidly. The disadvantages are that

it requires knowledge of PCR and is a relatively complex procedure that requires extensive

in-house validation, since the method is not approved by the FDA. However, as with all

sequence-based methods of identification, its utility is limited by the availability of an

updated public database. There is no commercial system for hsp65 PRA. As with all

molecular techniques, careful in-house validation with large databases of isolates is essential

for laboratories that perform this method.

Algorithms for the identification of mycobacterial species, including RGM, using PRA of

the hsp65 gene have been proposed (23, 71, 72). Figure 1 shows a PRA gel of the RGM most

commonly encountered in clinical laboratories.

Pyrosequencing

Pyrosequencing technology (Biotage, Uppsala, Sweden) employs nucleic acid sequencing of a

20- to 30-bp segment of hypervariable region A of the 16S rRNA gene. The method is based

on the detection of pyrophosphate during DNA synthesis. During sequencing, visible light

that is proportional to the number of incorporated nucleotides is produced.

The method is not as discriminating as traditional sequencing but is an attractive and less

expensive alternative. Another advantage of the pyrosequencing technology is its commercial

availability. Like other sequencing methods, however, the major limitation is the quality of

the databases for interpretation and comparison of sequences (34). In a recent study of 50

RGM (M. chelonae/M. abscessus and M. mucogenicum),consensus sequences were obtained

for 40 isolates, and results were compared to the results of traditional sequencing. Of 10

isolates of M. fortuitum, three had a sequence which identically matched M. fortuitum/M.

peregrinum, and the remaining 7 isolates matched M. fortuitum (80). To date, this method

has provided reliable and rapid identification of a variety of RGM (80).

A real-time PCR with melting curve analysis that consistently detects and differentiates M.

tuberculosis from NTM has been developed. In a recent study, 20 isolates previously

identified as M. fortuitum were confirmed as M. fortuitum. However, 7 of 24 isolates

previously identified as M. chelonae or M. abscessus were found by pyrosequencing analysis

to have been misidentified by traditional methods (65). The application of these methods is

currently used only in research or in large reference laboratories after a more extensive

evaluation has been done.

TYPING SYSTEMS Back to top

Pulsed-Field Gel Electrophoresis

Pulsed-field gel electrophoresis (PFGE) is the most widely used method for molecular strain

typing of RGM. Although PFGE has never been standardized for RGM, most investigators

concur that small (two- to three-band) differences between isolates indicate that the isolates

are closely related; differences of four to six bands indicate that the strains are possibly

related, and seven or more differences in bands indicate that the isolates are genetically

different (33, 73, 89, 91). Because unrelated strains of most RGM contain highly diverse

PFGE patterns, this technique has been useful in epidemiological investigations. With the

addition of thiourea as a recent modification of the original method, it is now possible to

obtain reliable results by PFGE of all species of RGM, including isolates previously affected by

DNA degradation (102, 103).

Randomly Amplified Polymorphic DNA-PCR

In the randomly amplified polymorphic DNA-PCR (RAPD-PCR) method that uses one arbitrary

primer and low-stringency conditions, the primer hybridizes to both strands of template DNA

where it is matched or partially matched, resulting in strain-specific heterogeneous DNA

products. Zhang and colleagues applied RAPD-PCR or the arbitrarily primed PCR analysis

method to compare strains of M. abscessus (102). They were able to confirm several

previous observations about prior nosocomial RGM outbreaks, including a 1988 epidemic of

otitis media due to aminoglycoside-resistant M. abscessus in children with prior

tympanostomy tubes and an outbreak among cardiac surgery patients (9, 102).

Repetitive-Sequence-Based PCR

Recently, a commercial system, DiversiLab system (BioMerieux, Durham, NC) was developed

for strain typing of organisms, including mycobacteria, using repetitive elements interspersed

throughout the mycobacterial genome. The system electrophoretically separates repetitivesequence-

based PCR amplicons on microfluidic chips to provide computer-generated

readouts. The discriminative power has been reported to equal or exceed that of standard

restriction fragment length polymorphism analysis for some species of mycobacteria, with a

smaller sample size than that used for standard PFGE and in a much more rapid time frame

(11, 101). Limitations of the system include the lack of an extensive established database

and the cost of the system.

Enterobacterial Repetitive Intergenic Consensus PCR

Enterobacterial repetitive consensus sequences are repetitive elements distributed along the

bacterial chromosome at intergenic regions of polycistronic operons or flanking open reading

frames. The method was recently evaluated with isolates of the M. abscessus/M.

chelonae complex and with isolates of M. fortuitum (60,61). Typing of isolates by

enterobacterial repetitive intergenic consensus (ERIC) PCR works in mycobacteria as a RAPDPCR,

because the presence of ERIC repeats has never been demonstrated in

available Mycobacteriumgenomes and amplification with appropriate ERIC primers can occur

in the absence of genuine ERIC sequences (61). In a study of outbreak strains of M.

abscessus in Brazil, ERIC PCR showed higher discriminatory power than PFGE for typing of

strains which had shown smear patterns with PFGE using thiourea (61), although this

method was not as discriminatory when isolates of M. fortuitum were tested (60,61).

SEROLOGIC TESTS Back to top

Serologic classification of mycobacteria was attempted starting in 1925 with MAC isolates.

However, serotyping has not been suitable for routine species identification of mycobacteria,

including RGM, and early studies served to emphasize the complexity of the antigenic

compositions of mycobacteria, as many antigens are shared by more than one species (96).

ANTIMICROBIAL SUSCEPTIBILITIES Back to top

Several different methods have been used for susceptibility testing of RGM for clinical

purposes. These methods include agar disk diffusion, broth microdilution, agar disk elution,

and the Etest. Each method has proven useful, but none of the methods were well

standardized until 2003, with the publication of the CLSI-approved guidelines (98). In the

M24-A document, the CLSI recommended broth microdilution as the “gold standard” for

susceptibility testing of RGM. Nine antimicrobials, including amikacin, cefoxitin, ciprofloxacin,

clarithromycin, doxycycline, linezolid, imipenem, sulfamethoxazole, and tobramycin, have

been recommended for testing, and breakpoints have been established for these agents. A

proposal to publish CLSI guidelines for additional agents, such as moxifloxacin, minocycline,

and trimethoprim-sulfamethoxazole, is under consideration (98).

Briefly, for the broth microdilution method, drug dilutions are prepared using serial twofold

dilutions of cation-adjusted Mueller-Hinton broth. Suspensions of organisms are prepared to

match a 0.5 McFarland turbidity standard. The suspensions are then diluted to a

concentration of approximately 106 CFU/ml. From that suspension, 100 μl is delivered into

the wells of a 96-well microtiter plate, with a final concentration of approximately

104 CFU/well (98). MICs are optimally read after incubation at 30°C for 3 days.

Several specific recommendations about test results have also been made. Tobramycin MICs

should be reported only for isolates of M. chelonae. Any RGM isolate for which the amikacin

MIC is ≥64 μg/ml should be retested and/or sent to a reference laboratory in order to

confirm resistance (although mutational resistance involving the 16S rRNA gene does occur).

Imipenem MICs should not be reported for isolates of the M. chelonae/M. abscessus group

because the results are not reproducible. Also, if the imipenem MIC for any isolate of the M.

fortuitum group is >8 μg/ml, the specimen should be tested again with careful attention paid

to inoculum density and with a maximum incubation time of 3 days because of the instability

of imipenem over time.

The recent finding of the presence of an erm gene that induces macrolide resistance in

isolates of the M. fortuitum group and of M. abscessus but not of M. chelonae has made

changes to the manner in which clarithromycin reporting should be done (49, 50). A proposal

to the CLSI for an initial 3-day reading of MICs followed by a final reading at 14 days unless

the isolate becomes resistant before that time has been made in an attempt to detect

isolates that contain the erm gene. The significance of the finding of these genes has not yet

been assessed in clinical trials.

Another caveat is that the MICs of sulfamethoxazole and trimethoprim-sulfamethoxazole

should be read using 80% inhibition of growth as the susceptibility endpoint, not the 100%

inhibition used for the other antimicrobials. Overinoculation of the MIC panels is often most

obvious with sulfonamides. An inexperienced laboratorian may interpret an isolate as

resistant to sulfonamide when in reality the inoculum was too heavy. Rarely, isolates of

the M. fortuitum group are resistant to sulfonamides. If an isolate in this group is found

resistant, a repeated test with a lower inoculum is warranted (98).

Further details of antimicrobial susceptibility methods and guidance for patient therapy may

be found inchapter 73.

EVALUATION, INTERPRETATION, AND REPORTING OF

RESULTS Back to top

Dramatic taxonomic changes largely attributed to the advent of molecular testing have

occurred over the past 10 to 20 years. Multiple new species have been introduced, and some

former subspecies have attained species status. Of the more than 70 valid species of RGM

currently described, almost half have been described within the past 10 years.

The recommended methods of identifying species are evolving, with a declining interest in

and efficiency of phenotypic testing, including HPLC, and an increasing availability and

accuracy of molecular methods. Phenotypic tests have limited utility (e.g., citrate utilization

to separate M. chelonae from M. abscessus) and are best applied in conjunction with

molecular methods. Currently, molecular methods are preferred and generally are the only

way to identify many newer species or subspecies, such as M. goodii, M.

abscessus subsp. bolletii, M. wolinskyi, and others. Laboratories in which molecular testing is

unavailable should consider referring RGM isolates to a reference laboratory with molecular

capabilities.

The major species of RGM (Table 2) have different levels of virulence in different clinical

settings and different drug susceptibilities. The diagnostic priorities in terms of disease

include bloodstream infection (usually catheter sepsis), disseminated or posttraumatic wound

infection, and pulmonary disease (especially in the setting of bronchiectasis). For

example, M. mucogenicum is a recognized cause of catheter sepsis, but because of its

common presence in tap water, the species is usually considered a contaminant in sputum.

The most common RGM isolated from pulmonary specimens and associated with disease

is M. abscessus. However, other RGM, such as M. fortuitum, which is rarely considered a

pathogen except in the setting of achalasia or lipoid pneumonia, may also be recovered from

respiratory samples. The pathogenic potential of RGM is generally related to clinical findings

(unexplained fever and dimorphic inflammatory lesions, etc.), the immune status of the

patient, the number of positive cultures, the quantity of organisms recovered from smearpositive

samples, and the sources of the recovered species. Some established and newly

described species have been identified from environmental samples but as yet have not been

identified as human or animal pathogens (23). These factors emphasize the need to perform

species-level identification and susceptibility testing on clinically significant isolates of RGM.

When species identification of a clinically significant isolate is not available (i.e., susceptibility

may be finalized prior to the identification of the species), MICs should be reported for

antimicrobials, as recommended by the CLSI. An identification such as “rapidly

growing Mycobacterium sp.” may be acceptable until identification has been performed.

However, the report should include a caveat that tobramycin is validated only for M.

chelonaeand that imipenem is not validated for the M. chelonae/M. abscessus group (98).

Isolates recovered from a single sample are less likely to be significant than those from

multiple samples. Moreover, when MIC results are reported, an interpretation of the isolate

(i.e., as “susceptible,” “intermediate,” or “resistant”) should be given for each antimicrobial

for which there are recommended breakpoints (98). For agents such as tigecycline, for which

no breakpoints have been recommended, an MIC value with a notation that “no CLSI

breakpoints have been established for this species” should accompany the report.

Furthermore, for cultures that remain positive for the same species after 6 months of

appropriate antimicrobial therapy, confirmation of species identification by molecular

methods and repeat antimicrobial susceptibility testing is warranted (98).

Although recent advances in antimicrobial therapies, including those using the new

macrolides, fluoroquinolones, oxazolidinones, and tigecycline, have enhanced the therapeutic

options and the prognosis for RGM disease, there is still a compelling need for the

development of more- efficient, more-effective, and safer oral antimicrobials for treatment.

For example, M. abscessus lung disease is still generally considered incurable with the

currently available antimicrobials. Susceptibility testing of RGM is necessary in order to select

an optimal antimicrobial therapy and to monitor the development of mutational drug

resistance which may occur with the prolonged therapy required for RGM disease.