Clinical and Microbiological Differences between Mycobacterium abscessus and Mycobacterium massiliense Lung Diseases

Abstract

In recent years, many novel nontuberculous mycobacterial species have been discovered through genetic analysis. Mycobacterium massiliense and M. bolletii have recently been identified as species separate from M. abscessus. However, little is known regarding their clinical and microbiological differences in Japan. We performed a molecular identification of stored M. abscessus clinical isolates for further identification. We compared clinical characteristics, radiological findings, microbiological findings, and treatment outcomes among patients with M. abscessus and M. massiliense lung diseases. An analysis of 102 previous isolates of M. abscessus identified 72 (71%) M. abscessus, 27 (26%) M. massiliense, and 3 (3%) M. bolletii isolates. Clinical and radiological findings were indistinguishable between the M. abscessus and M. massiliense groups. Forty-two (58%) patients with M. abscessus and 20 (74%) patients with M. massiliense infections received antimicrobial treatment. Both the M. abscessus and M. massiliense groups showed a high level of resistance to all antimicrobials, except for clarithromycin, kanamycin, and amikacin. However, resistance to clarithromycin was more frequently observed in the M. abscessus than in the M. massiliense group (16% and 4%, respectively; P = 0.145). Moreover, the level of resistance to imipenem was significantly lower in M. abscessus isolates than in M. massiliense isolates (19% and 48%, respectively; P = 0.007). The proportions of radiological improvement, sputum smear conversion to negativity, and negative culture conversion during the follow-up period were higher in patients with M. massiliense infections than in those with M. abscessus infections. Patients with M. massiliense infections responded more favorably to antimicrobial therapy than those with M. abscessus infections.

INTRODUCTION

Mycobacterium species are common causes of pulmonary infections in both humans and animals (14). Although members of the Mycobacterium tuberculosis complex cause the majority of pulmonary infections worldwide, many nontuberculous mycobacteria (NTM) can cause similar infections (13, 20). In recent years, many novel NTM species have been discovered through the increased application of genetic investigation tools; detailed genetic characterizations have helped define new taxonomic groupings (17, 29). Recently, two new M. abscessus-related species, M. massiliense and M. bolletii, were identified, which were previously grouped with M. abscessus (1, 3). The rate of isolation of these two species has been increasing in Japan. However, very little is known about the natural epidemiology and pathogenicity of M. massiliense and M. bolletii outside outbreak situations. One report found that the ratio of M. abscessus to all NTM is much higher in South Korea (19) than in other countries, including Japan.

Here, we aimed to evaluate the epidemiology, clinical and radiological spectrum, treatments, drug susceptibility, and outcome of M. abscessus and M. massiliense lung diseases during therapy in Japan.

MATERIALS AND METHODS

Study population.

We retrospectively reviewed the medical records of patients initially diagnosed with M. abscessus lung disease according to the 2007 American Thoracic Society/Infectious Diseases Society of America (ATS/IDSA) guidelines (16) between January 1990 and December 2010 at 12 hospitals or institutions in Japan. These M. abscessus species were thereafter identified as M. abscessus, M. massiliense, and M. bolletii. Clinical, radiological, microbiological, management, and outcome data were collected from medical files. Permission was obtained from the institutional review board committee of Hokkaido Social Insurance Hospital (approval number 2011-11). Informed consent was waived because of the retrospective nature of the study.

Microbiological examination.

Sputum or bronchoalveolar lavage fluid was used for smears and mycobacterial cultures according to standard methods (5). Any processed specimens that remained were stored at 2°C to 8°C for the duration of culturing in the study to allow the retesting of the specimens that showed a discrepancy in results between culture growth and preliminary identification by a DNA-DNA hybridization (DDH) mycobacterium kit (Kyokuto Pharmaceutical Industrial Co., Tokyo, Japan). Samples were cultured by using the Bactec MGIT 960 system or Ogawa solid medium, and the isolates were stored at −30°C to −80°C. M. abscessus species were preliminarily identified by microplate DDH technology using the DDH mycobacterium kit (21) at each hospital or institution. We collected all frozen isolates for multisequencing and susceptibility testing. Frozen isolates were recultured by using the MGIT 960 system and Ogawa solid medium and checked for contamination by growth on p-nitrobenzoic acid agar medium.

Further differentiation among M. abscessus species was performed at the Department of Mycobacteriology, Leprosy Research Centre, National Institute of Infectious Disease, and the Kobe Institute of Health. Sequences of clinical isolates which were previously identified as M. abscessus by DDH were compared with the reference M. abscessus (JCM 15300^T), M. massiliense (JCM 13569^T), and M. bolletii (JCM 15297^T) strains. The majority of the 16S rRNA gene, partial aspects of the hsp65 and rpoB genes, and the 16S-23S rRNA internal transcribed spacer (ITS) region were amplified by PCR using AmpliTaq Gold DNA polymerase (Applied Biosystems, Foster City, CA) and primers described previously (22). The PCR products were sequenced with the BigDye Terminator v3.1 cycle sequencing kit (Applied Biosystems) on an ABI Prism 310 genetic analyzer (Applied Biosystems). Sequences were analyzed for their similarity to sequences in the GenBank database by using BLAST (http://www.ncbi.nlm.nih.gov/BLAST).

Antimycobacterial susceptibility testing was performed at the National Hospital Organization Kinki-Chuo Chest Medical Center (Osaka, Japan). Susceptibility was determined by MIC methods using the Etest and broth microdilution.

Etest.

Isolates of rapidly growing mycobacteria (RGM) were mailed on Ogawa slant medium and subcultured on 5% sheep blood agar at 35°C in ambient air for 72 h. Bacterial suspensions were prepared in cation-adjusted Mueller-Hinton broth (Eiken Chemical, Tokyo, Japan) to a 1.0 McFarland standard and plated onto cation-adjusted Mueller-Hinton agar. Etest strips (Sysmex bioMérieux, Tokyo, Japan) were placed onto Mueller-Hinton agar (Becton Dickinson, Fukushima, Japan) according to the manufacturer's instructions, and the results were read after 72 h of incubation. Final concentration ranges were 0.016 to 256 μg/ml for isoniazid (INH) and ethambutol (EB); 0.002 to 32 μg/ml for rifampin (RFP), ciprofloxacin (CPFX), and moxifloxacin (MFLX); and 0.064 to 1,024 μg/ml for streptomycin (SM). Susceptibility was evaluated according to Clinical and Laboratory Standards Institute (CLSI) breakpoint recommendations (8, 37) and those proposed previously by Woods et al. (36), Wallace et al. (33), Yang et al. (38), and Swenson et al. (27, 28).

Broth microdilution.

Serial double dilutions of clarithromycin (CAM), kanamycin (KM), amikacin (AMK), and imipenem (IPM) were prepared at a concentration range of 0.015 to 64 μg/ml according to CLSI recommendations (37). Briefly, pure colonies were cultured in 7H9 broth with 0.2% glycerol in a tube for 3 to 5 days, vigorously vortexed, and then adjusted to a density equivalent of 0.5 on the McFarland scale. Bacterial suspensions in cation-adjusted Mueller-Hinton broth were transferred into the wells of dry microdilution plates containing the antimicrobial agents (Eiken Chemical). The inoculated plates were placed into plastic bags, incubated at 35°C in ambient air, and read after 72 h. The MIC was defined as the lowest concentration of drug that inhibited visible growth. Susceptibility was evaluated according to CLSI breakpoint recommendations (8, 37) and those proposed previously by Shen et al. (24).

Data analysis.

The results are expressed as ranges and means or as numbers of patients. Categorical variables were analyzed by using the χ² or Fisher's exact test. Continuous variables were analyzed by using the Mann-Whitney U test or t test. All P values are two sided; a P value of <0.05 indicates statistical significance. Statistical analyses were performed by using Excel 2011 (Microsoft) with the add-in software Statcel 3 (OMS Publishing Inc., Saitama, Japan).

RESULTS

Identification of clinical isolates.

Clinical isolates from 102 patients previously diagnosed with M. abscessus lung disease by DDH were identified by sequence analysis of the 16S rRNA, hsp65, rpoB, and ITS genes. Seventy-two (71%) isolates were identified as M. abscessus, 27 (26%) were identified as M. massiliense, and 3 (3%) were identified as M. bolletii.

Patient characteristics.

We compared clinical characteristics and treatment outcomes among the 72 patients with M. abscessus and 27 patients with M. massiliense infections. Patients with M. bolletii infection were not included for further study because of their small number.

Table 1 summarizes the baseline characteristics of the patients. No significant differences were found between the M. abscessus and M. massiliense groups in any of the baseline characteristics, including demographic data, underlying conditions, and respiratory symptoms. The distributions of radiographic disease types were similar in both groups, except for bronchiectasis. Bronchiectasis was significantly more frequent in the M. abscessus group (73%; 40 of 55) than in the M. massiliense group (43%; 10 of 23) (P = 0.014). None of these patients tested positive for HIV.

Table 1.

Clinical characteristics of patients with Mycobacterium abscessus and M. massiliense lung disease

Characteristica	Value for group		P value
	M. abscessus (n = 72)	M. massiliense (n = 27)
Age (yr) [range (mean)]	27–94 (68)	44–84 (67)	0.637
No. of males/no. of females	26/45	13/13	0.233
Mean body mass index (kg/m2)	19.5	18.8	0.699
No. of patients with symptom
Cough	25	9	0.931
Sputum	22	9	0.759
Hemoptysis	14	8	0.262
Fever	8	5	0.905
Dyspnea	4	5	0.057
No. of patients with underlying disease
Previous pulmonary tuberculosis	12	7	0.297
NTM (MAC)	15	4	0.516
Mycosis	4	2	0.798
Interstitial pneumonia	4	2	0.803
COPD	3	2	0.879
Diabetes mellitus	3	3	0.196
Steroid use	3	3	0.196
Malignancy	2	1	0.823
No. of smokers/no. of nonsmokers	33/11	12/8	0.223
No. of patients who consumed alcohol/no. of patients who did not consume alcohol	30/6	11/3	0.490
No. of patients with/no. of patients without radiological finding of:
Cavitation	28/28	14/9	0.379
Centrilobular lesion	38/19	14/11	0.355
Infiltration	35/16	12/11	0.173
Bronchiectasis	40/15	10/13	0.014
Pleural effusion	2/49	2/22	0.383
No. of patients with positive AFB smear	50	21	0.363

^aAbbreviations: NTM, nontuberculous mycobacteria; MAC, Mycobacterium avium complex; COPD, chronic obstructive pulmonary disease; AFB, acid-fast bacillus.

Drug susceptibility.

Table 2 shows the drug susceptibility results for 63 M. abscessus and 23 M. massiliense isolates, and Table S1 in the supplemental material shows the MIC distributions of the strains. The drug was determined to be effective against an isolate if the MIC of the antimicrobial agent was less than the susceptible concentration, as shown in Table 2. Of the parenteral antibiotics, KM and AMK were effective against most M. abscessus isolates (95% and 94%, respectively) and M. massiliense isolates (100% for both), with no difference between the species (P = 0.388 and 0.509, respectively). However, SM was ineffective against both M. abscessus (67%) and M. massiliense (61%) isolates (P = 0.617). Moreover, the rates of IPM drug resistance were significantly higher in M. massiliense than in M. abscessus isolates (48% and 19%, respectively; P = 0.007).

Table 2.

In vitro susceptibilities of 63 Mycobacterium abscessus and 23 M. massiliense isolates to different antimicrobials^a

Drug	Species	MIC (μg/ml) for categorization of susceptibility of:			Reference(s)	% resistant isolates (no. of resistant isolates)	P value
		Susceptible	Intermediate	Resistant
Clarithromycin	M. abscessus	≤2	4	≥8	15	16 (10)	0.145
	M. massiliense					4 (1)
Kanamycinb	M. abscessus	≤16	32	≥64	14, 20	5 (3)	0.388
	M. massiliense					0 (0)
Amikacin	M. abscessus	≤16	32	≥64	15	6 (4)	0.509
	M. massiliense					0 (0)
Imipenem	M. abscessus	≤4	8–16	≥32	15	19 (12)	0.007
	M. massiliense					48 (11)
Moxifloxacin	M. abscessus	≤1	2	≥4	15	92 (58)	0.488
	M. massiliense					96 (22)
Ciprofloxacin	M. abscessus	≤1	2	≥4	15	95 (60)	0.883
	M. massiliense					91 (21)
Isoniazidc	M. abscessus	NA	NA	NA		NA
	M. massiliense					NA
Rifampinb	M. abscessus	≤1	2	≥4	14, 18	97 (61)	0.534
	M. massiliense					100 (23)
Ethambutolc	M. abscessus	NA	NA	NA		NA
	M. massiliense					NA
Streptomycinb	M. abscessus	≤32	NA	≥64	14, 20	67 (42)	0.617
	M. massiliense					61 (14)

^aDrug susceptibility results are shown for 63 patients with M. abscessus and 23 patients with M. massiliense infections. NA, not available.

^bThe CLSI breakpoints (8) for Staphylococcus species have been substituted as the breakpoints of these drugs against M. abscessus and M. massiliense isolates.

^cThe breakpoints for M. abscessus and M. massiliense isolates have not yet been established.

Of the oral antibiotics, CAM was effective against most M. abscessus and M. massiliense isolates (84% and 96%, respectively; P = 0.145). In contrast, MFLX, CPFX, and RFP were ineffective against most M. abscessus (92%, 95%, and 97%, respectively) and M. massiliense (96%, 91%, and 100%, respectively) isolates, with no difference between the species (P = 0.488, 0.883, and 0.534, respectively). We could not interpret the INH and EB susceptibilities because no breakpoints have been established for these drugs.

Antimicrobial treatment and response.

Of the 102 patients, 42 (58%) with M. abscessus, 20 (74%) with M. massiliense, and 2 (67%) with M. bolletii infections received antimicrobial treatment for 3 to 178 months (mean = 33 months), 1 to 122 months (mean = 36 months), and 4 to 68 months (mean = 36 months), respectively. Forty-two patients with M. abscessus lung disease and 20 patients with M. massiliense lung disease were analyzed for treatment response (Table 3). Radiographic improvement rates were lower for patients with M. abscessus infection than for those with M. massiliense infection (29% and 48%, respectively; P = 0.101).

Table 3.

Antimicrobial treatments and treatment responses

Characteristica	Value for group		P value
	M. abscessus (n = 72)	M. massiliense (n = 27)
Treatment
No. of patients with/no. of patients without operation	3/69	3/24	0.201
No. of patients receiving/no. of patients not receiving chemotherapy	42/30	20/7	0.149
No. of patients on chemotherapy regimen of:
M, R, E	8	5
M, C, A	6	5
M	8	0
M, F	6	0
H, R, E	2	2
Other	12	8
Duration of treatment (mo) [range (mean)]	3–178 (33)	1–122 (36)	0.723
Result
No. of patients with/no. of patients without radiological improvement	17/41	12/13	0.101
No. of patients with/no. of patients without sputum smear conversion to negativity	17/38	11/11	0.115
No. of patients with relapse/no. of patients without relapse after sputum smear conversion to negativity	13/7	3/7	0.077
Duration of positive sputum results (mo) [range (mean)]	1–120 (25)	1–62 (18)	0.776

^aAbbreviations: M, macrolides (clarithromycin, erythromycin, and azithromycin); A, aminoglycosides (streptomycin, amikacin, and kanamycin); F, fluoroquinolones (levofloxacin, moxifloxacin, garenoxacin, and gatifloxacin); C, carbapenems (imipenem and meropenem); H, isoniazid; R, rifampin; E, ethambutol.

Microbiological responses also differed between the two groups. The initial sputum conversion rates were lower in patients with M. abscessus infection than in those with M. massiliense infection (31% and 50%, respectively; P = 0.115). The sputum acid-fast bacillus (AFB)-positive relapse rate after the initial conversion to a negative result was higher for patients with M. abscessus infection than for those with M. massiliense infection (65% and 30%, respectively; P = 0.077). Thus, the proportion of patients whose sputum converted and remained culture negative during the follow-up period was lower for patients with M. abscessus infection than for those with M. massiliense infection.

DISCUSSION

The M. abscessus group comprises ubiquitous environmental organisms frequently associated with nosocomial outbreaks and pseudo-outbreaks (18, 32, 34). The increasing availability of gene sequencing has tremendously influenced the taxonomy of bacteria, particularly mycobacteria, with many new species being described every year (30).

A new species related to M. abscessus, M. massiliense, was recently described (3). The species M. massiliense was proposed in 2004 based upon nonconventional phenotypic characterization and genotypic studies of 2 isolates recovered from the sputum and bronchoalveolar fluid of a patient in France (3). Since M. massiliense is closely related to M. abscessus, it is possible that M. massiliense infections have overlapped with M. abscessus infections in previous reports (3). Although several unique phenotypes that differentiate M. massiliense from M. abscessus have been identified, they are difficult to characterize using conventional clinical techniques because of the inability to discriminate between RGM due to their overlapping phenotypic patterns (9, 26). The molecular, biological, and clinical characteristics of these strains will help us to better understand and treat severe infections due to RGM (39). The proper identification of members of the M. abscessus complex has proven beneficial in both therapeutic management and epidemiological studies. M. massiliense is very closely related to M. abscessus but showed different susceptibilities to CAM, and their pathogenic potentials have been demonstrated by infections of immunocompetent and immunocompromised hosts (15, 18).

The proportions of M. massiliense strains among M. abscessus species vary according to geographical distribution. The prevalences of M. massiliense were 28% of 40 patients at the National Institutes of Health in the United States (39), 21% of 39 clinical isolates in the Netherlands (31), 22% of 50 patients with cystic fibrosis in France (23), 55% of 150 patients in South Korea (19), and 26% of 102 patients in Japan. There is currently no explanation for the large difference in the prevalences of M. massiliense between South Korea and Japan, as they are in the same Asian region.

No significant differences were found between the baseline clinical characteristics in the M. abscessus and M. massiliense groups, except for bronchiectasis in the radiological findings. Bronchiectasis was found significantly more frequently in the M. abscessus group than in the M. massiliense group (73% and 43%, respectively; P = 0.014). This result is almost consistent with the results of a Korean study (19).

Importantly, favorable microbiological response rates with similar combinations of antibiotic therapy were much higher for M. massiliense than for M. abscessus lung disease. This could be explained by the differences in CAM resistance. This study demonstrated a high level of resistance to CAM in M. abscessus isolates but not in M. massiliense isolates, indicating that treatment of M. abscessus lung disease may be more difficult. In fact, M. abscessus lung disease has been regarded as a chronic, incurable infection for most patients, given the current antibiotic options (4). The low MIC and absence of CAM resistance (except for one isolate) suggest that M. massiliense lung disease may be treated more effectively with a CAM-based antibiotic regimen. Recent studies showed that some RGM, such as M. abscessus and M. fortuitum, have an erm gene that induces macrolide resistance (4, 12). It is unknown whether other RGM such as M. massiliense have an erm gene. Thus, species-level identification is important because antibiotic susceptibilities and therapies differ significantly depending on the RGM species (4, 16). Since the KM and AMK resistance rates were less than 10% for both the M. abscessus and M. massiliense groups in the present study, KM and AMK could be two key drugs for the treatment of M. abscessus and M. massiliense lung diseases.

The IPM resistance rate was much lower in the present study than in the Korean study (19) (19% and 44% for M. abscessus and 48% and 67% for M. massiliense, respectively). The MFLX, CPFX, and RFP resistance rates were >90% for both the M. abscessus and M. massiliense groups in the present study. These findings are compatible with the fact that the M. abscessus complex has been regarded as being fluoroquinolone resistant (7). However, moderate in vitro activities of some fluoroquinolones against members of the M. abscessus complex have been demonstrated (2, 6, 19, 25). The use of fluoroquinolones as alternative oral agents during combination antibiotic therapy for M. abscessus and M. massiliense infections should be studied further.

The present study showed poor activities of INH and EB in both the M. abscessus and M. massiliense groups (MIC₉₀ for INH of ≥512 μg/ml; MIC₉₀ for EB of ≥32 μg/ml), which are identical to findings reported previously (27). Thus, INH and EB seem to be ineffective against both the M. abscessus and M. massiliense groups. Another study found that MFLX was active against M. abscessus and that a combination of CAM and MFLX was effective against M. abscessus strains in in vitro models (10). Moreover, some M. abscessus isolates are susceptible to the oral drug linezolid (35, 38). However, linezolid was rarely used in our survey to treat M. abscessus species lung disease because of the high cost and moderate to severe side effects in Japan. Thus, further studies are required to evaluate active combinations of oral antibiotics and determine their clinical significance.

The present study has several limitations. First, the retrospective study design necessitates the use of medical records for data collection, leading to variations in each factor. Second, the number of sputum specimens collected over time was relatively small. Had samples been collected more frequently, more conversions to negativity and relapses after conversion may have been found. Third, treatment regimens cannot be optimized based solely on retrospective studies with limited follow-up data. Moreover, the use of IPM, cefoxitin, fluoroquinolone, and linezolid is not permitted for the treatment of NTM diseases under the Japanese social health insurance system. Thus, the combination therapy recommended by the ATS/IDSA (16) has not been applied in most cases. Treatment regimens were decided in practice by physicians. Therefore, it is difficult to evaluate the true treatment response in this study.

In conclusion, we found clinically significant differences between M. abscessus and M. massiliense lung infections in Japan. Treatment responses rates with CAM-based antibiotic therapy were higher for M. massiliense than in M. abscessus lung disease. This difference in treatment responses may be explained by the difference in CAM susceptibilities between the two groups. Prospective clinical trials are needed to clarify these aspects.