Clinical, Radiological, and Microbiological Characteristics of Mycobacterium simiae Infection in 97 Patients

ABSTRACT

Mycobacterium simiae is a rare species of slow-growing nontuberculous mycobacteria (NTM). From 2002 to 2017, we conducted a retrospective study that included all patients with NTM-positive respiratory samples detected in two university hospitals of the French overseas department of Reunion Island. We recorded the prevalence of M. simiae in this cohort, as well as the clinical, radiological, and microbiological features of patients with at least 1 sample positive for M. simiae. In our cohort, 97 patients (15.1%) were positive for M. simiae. Twenty-one patients (21.6%) met the American Thoracic Society (ATS) criteria for infection. M. simiae infection was associated with bronchiectasis, micronodular lesions, and weight loss. Antibiotic susceptibility testing was performed for 60 patients, and the isolates were found to have low susceptibility to antibiotics, except for amikacin, fluoroquinolones, and clarithromycin. Treatment failed for 4 of the 8 patients treated for M. simiae infection. Here, we describe a specific cluster corresponding to a large cohort of patients with M. simiae, a rare nontuberculous mycobacterium associated with low pathogenicity and poor susceptibility to antibiotics.

INTRODUCTION

Nontuberculous mycobacteria (NTM) constitute a large group of more than 150 pathogens (1). Mycobacterium simiae is a rare species of NTM that was first detected in monkeys in 1965 (2). A few geographic clusters have since been identified, namely, in Arizona, Cuba, and Israel (3, 4), suggesting environmental contamination. M. simiae was detected in local tap water in some clusters, and a pseudo-outbreak caused by water contamination was reported by El Sahly et al. (5).

When small series of patients with clinical M. simiae isolates are considered, pathogenicity appears to be low (4, 6, 7). Some authors have reported very rare cases of pulmonary infections (8–11) and lymphadenitis (12). Most of those infections occurred in immunocompromised patients with AIDS (10, 13, 14) or patients with predisposing conditions, including children (12), elderly patients, and patients with diabetes mellitus (11).

M. simiae is common in the French overseas department of Reunion Island. We conducted a retrospective study to determine the prevalence of M. simiae on the island, and we recorded the clinical, radiological, and microbiological features of all patients who were positive for M. simiae.

RESULTS

During the study period, 202 samples from 97 patients tested positive for M. simiae. This represented a prevalence of 15.1% in our cohort of patients with respiratory samples that were positive for NTM. One of the 97 patients also had a lymph node that was positive for M. simiae; no other patients had an extrapulmonary sample that was positive for M. simiae. Other mycobacteria were found in 22 patients (22.7%), including Mycobacterium intracellulare in 6 patients, Mycobacterium fortuitum in 4 patients, Mycobacterium abscessus in 4 patients, Mycobacterium gordonae in 4 patients, undefined mycobacterial species in 4 patients, Mycobacterium tuberculosis in 2 patients, and Mycobacterium xenopi in 1 patient.

The M. simiae isolates were cultured from sputum samples (57.8%), gastric aspirate samples (14.4%), and bronchoalveolar lavage fluid or bronchial aspirate samples obtained through fiber-optic procedures (45.3%) (Table 1). Positive bronchoalveolar lavage fluid or bronchial aspirate samples were significantly more common among patients who met the American Thoracic Society (ATS) criteria for infection than among those who did not (71.4% versus 38.2%; P = 0.008). There was a median of 6 respiratory samples (range, 2 to 13 samples) for patients who met the ATS criteria for M. simiae infection, compared with a median of 4.5 samples (range, 2 to 13 samples) for patients who did not (P = 0.06). The number of positive respiratory samples was significantly higher for patients who met the ATS criteria for infection than for those who did not (median of 2 positive samples [range, 2 to 4 positive samples] versus 1 positive sample [range, 1 to 2 positive samples]; P = 0.0002).

TABLE 1.

Microbiological data for patients with respiratory samples positive for M. simiae

Parametera	Total (n = 97)	ATS criteria for infection:		P
		Not met (n = 76)	Met (n = 21)
No. of respiratory samples tested for NTM (per patient) (median [IQR])
Total	5 (2–8)	4.5 (2–7)	6 (2–13)	0.06
Positive for M. simiae	1 (1–2)	1 (1–2)	2 (2–4)	0.0002
Type of positive respiratory sample (no. [%])
Sputum	56 (57.8)	44 (57.9)	13 (61.9)	0.22
Gastric aspirate	14 (14.4)	10 (13.1)	4 (19.0)	0.62
BAL fluid or bronchial aspirate	44 (45.3)	29 (38.2)	15 (71.4)	0.008

^aNTM, nontuberculous mycobacteria; ATS, American Thoracic Society; IQR, interquartile range; BAL, bronchoalveolar lavage.

Drug susceptibility testing was performed for 60 patients, using a single isolate per patient. Most M. simiae strains were found to be susceptible to amikacin, moxifloxacin, ciprofloxacin, and clarithromycin (96%, 92%, 87%, and 100% of tested patients, respectively). The strains were found to be resistant to rifampin, ethambutol, and isoniazid for 85%, 89%, and 92% of tested patients, respectively.

Patient characteristics are presented in Table 2. Of the 97 patients who tested positive for M. simiae, 21 (21.6%) met the ATS criteria for NTM infection with M. simiae. Cough and/or sputum (68%), fever and/or sweat (17.5%), and weight loss (48.4%) were the most commonly reported symptoms. Weight loss and fever and/or sweat were significantly more common among patients who met the ATS criteria for NTM infection (Tables 3 and 4). The most common radiological findings were micronodular or “tree-in-bud” lesions (50.5% of patients) and cavitary lesions (25.8% of patients). Micronodules were significantly more common among patients who met the ATS criteria (Tables 3 and 4). Microbiological ATS criteria were met by 64.9% of patients (Table 3).

TABLE 2.

Characteristics of patients with clinical samples positive for M. simiae

Parametera	Total (n = 97)	ATS criteria for infection:		P
		Not met (n = 76)	Met (n = 21)
Age (median [IQR]) (yr)	61 (50–74)	65 (50–74.3)	57 (51–66)	0.3
Male (no. [%])	51 (52.7)	42 (55.3)	9 (39.1)	0.32
Body mass index (median [IQR]) (kg/m2)	19 (16.5–22.1)	19.5 (17–23)	18.8 (16.5–20.9)	0.42
Respiratory diseases (no. [%])
Non-CF bronchiectasis	48 (49.5)	33 (43.4)	15 (71.4)	0.04
CF	6 (6.2)	5 (6.6)	1 (4.8)	0.73
COPD	24 (24.7)	20 (26.3)	4 (19)	0.42
Previous tuberculosis	15 (15.5)	11 (14.5)	4 (19.0)	0.67
Lung neoplasia	9 (9.3)	9 (11.8)	0 (0)	0.09
Interstitial pneumonia	5 (5.1)	5 (6.6)	0 (0)	0.5
Other diseases (no. [%])
Other neoplasia	5 (5.1)	4 (5.3)	1 (4.8)	0.85
HIV infection	4 (4.1)	2 (2.6)	2 (9.5)	0.44
Other immunodeficiency	5 (5.1)	5 (6.6)	0 (0)	0.32
Diabetes mellitus	11 (11.3)	10 (13.1)	1 (4.8)	0.27
Renal insufficiency	6 (6.2)	4 (5.3)	2 (9.5)	0.48
Gastroesophageal refluxb	8 (8.2)	6 (7.9)	2 (9.5)	0.85
Alcohol abuse	16 (16.5)	14 (18.4)	2 (9.5)	0.3
Lung function
FEV1 (median [IQR]) (liters)	1.60 (1.11–2.30)	1.55 (1.05–2.42)	1.82 (1.50–2.20)	0.54
FEV1 (median [IQR]) (%)	72 (53–90)	70 (52–90)	62 (64–81)	0.64
FVC (median [IQR]) (liters)	2.42 (1.72–3.25)	2.33 (1.71–3.63)	2.55 (1.93–2.70)	0.93
FVC (median [IQR]) (%)	81 (62–92)	76 (52–92)	84.5 (73–90)	0.33
Long-term treatment (no. [%])
Inhaled corticosteroid	25 (25.8)	19 (25)	6 (28.6)	0.84
Systemic corticosteroidc	11 (11.3)	10 (13.1)	1 (4.8)	0.3
Long-term macrolide	19 (19.6)	14 (18.4)	5 (23.8)	0.6

^aIQR, interquartile range; ATS, American Thoracic Society; CF, cystic fibrosis; COPD, chronic obstructive pulmonary disease; FEV1, forced expiratory volume in 1 s; FVC, forced vital capacity.

^bGastroesophageal reflux was diagnosed based on evocative symptoms and/or antacid treatment.

^cSystemic corticosteroid use was defined as more than 10 mg/day of prednisone (or equivalent) for more than 2 weeks.

TABLE 3.

Clinical, radiological, and microbiological criteria for M. simiae infection, as defined by the ATS

Criteria assesseda	No. (%)			P
	Total (n = 97)	ATS criteria for infection:
		Not met (n = 76)	Met (n = 21)
Clinical ATS criteria	84 (86.6)	63 (82.3)	21 (100)	0.06
Fever/sweating	17 (17.5)	10 (13.1)	7 (33.3)	0.04
Cough/sputum	66 (68.0)	49 (64.5)	17 (80.9)	0.2
Weight lossb	47 (48.4)	32 (42.1)	15 (71.4)	0.02
Hemoptysis	3 (3.1)	2 (2.6)	1 (4.8)	0.53
Radiological ATS criteria	74 (76.3)	55 (72.4)	21 (100)	0.04
Nodules of <1 cm	49 (50.5)	35 (46)	14 (66.7)	0.08
Nodules of >1 cm	9 (9.3)	8 (10.5)	1 (4.8)	0.42
Consolidation	30 (30.9)	24 (31.6)	6 (28.6)	0.76
Cavitary lesions	25 (25.8)	17 (22.4)	8 (38.1)	0.14
Microbiological ATS criteria	63 (64.9)	44 (57.9)	21 (100)	0.006

^aATS, American Thoracic Society.

^bWeight loss was defined as a weight reduction of more than 5 kg or 5% of the normal body weight in the previous 6 months.

TABLE 4.

Multivariate analysis of risk factors meeting ATS criteria for M. simiae infection

Variablea	Adjusted odds ratio (95% CI)b	P
Non-CF bronchiectasis	4.11 (0.98–17.24)	0.05
Weight loss	8.13 (2.04–32.26)	0.003
Nodules of <1 cm	−4.29 (1.01–18.18)	0.05
Lung neoplasia		0.97
Fever/sweating	2.91 (0.82–16.1)	0.12
Cavitary lesions	0.83 (0.51–6.65)	0.36

^aCF, cystic fibrosis; CI, confidence interval.

^bThe Hosmer-Lemeshow goodness-of-fit test P value was 0.979. The Nagelkerke and Cox-Snell R² values were 0.388 and 0.254, respectively. Odds ratio for lung neoplasia is not applicable after step backward selection.

Twelve patients started drug therapy, including 8 who were treated for M. simiae infections (8.2%). Two patients were treated for pulmonary tuberculosis, and 2 others were treated for M. intracellulare or M. xenopi infections. The clinical, radiological, and microbiological characteristics of the patients treated for M. simiae infections are reported in Table 5, along with their treatment regimens and clinical outcomes. Among the 21 patients who met the ATS criteria for infection, 7 of 8 treated patients had at least 1 positive smear and 7 of 13 untreated patients had positive smears. The only treated patient with negative smears (patient 2) had a good outcome.

TABLE 5.

Description of M. simiae-infected patients who required antibiotic treatment

Patient no.	Age (yr)	Gendera	Underlying conditions	Clinical findings	Radiological findings	No. of positive microbiological samples/total no. of samples	Treatment regimen	Duration	Outcome
1	61	F	Bronchiectasis, GORD	Hemoptysis, cough/sputum, weight loss	Micronodular	11/13 (2 sputum, 4 GA, 7 BA/BAL fluid)	Amikacin, ethambutol, clarithromycin, moxifloxacin	1 mo (hepatitis)	Treatment failure; death (4 yr later)
2	49	M	HIV (79 CD4+ cells/μl)	Cough	Cavitary with aspergilloma	2/24 (1 sputum, 1 BA)	Ethambutol, azithromycin-clarithromycin, ansatipine	4 mo	No relapse; alive (10 yr later)
3	44	M	HIV (9 CD4+ cells/μl)	Weight loss	Micronodular, adenopathy	2/5 (1 BA, 1 lymph node)	Ethambutol, clarithromycin	In progress	Alive (5 mo later)
4	52	F	Bronchiectasis, GORD	Cough/sputum	Cavitary, nodular	6/16 (4 sputum, 2 GA)	Azithromycin, ethambutol, moxifloxacin, clofazimine, plus surgical resection	15 mo	No relapse with M. simiae, secondary infection with M. intracellulare; alive (3 yr later)
5	67	M	Bronchiectasis, renal insufficiency	Cough/sputum	Cavitary, micronodular	4/19 (3 GA, 1 BA)	Clarithromycin, moxifloxacin, ethambutol, clofazimine	18 mo	No relapse; alive (2 yr later)
6	49	F	Bronchiectasis	Cough/sputum, weight loss, sweat	Micronodular	7/19 (7 sputum)	Clarithromycin, moxifloxacin, clofazimine, amikacin (1 mo), then azithromycin, moxifloxacin, clofazimine, imipenem-cefoxitin	12 + 5 mo	No relapse with M. simiae, secondary infection with M. abscessus; alive (3 yr later)
7	65	F	Bronchiectasis	Cough/sputum	Micronodular, cavitary	2/2 (2 sputum)	Amikacin, moxifloxacin, rifabutin, azithromycin	12 mo	Relapse after 1 yr; alive (6 yr later)
8	69	F	Bronchiectasis, COPD, GORD	Cough/sputum, weight loss, fever	Micronodular, cavitary	3/4 (2 BA, 1 BAL fluid)	Clarithromycin, moxifloxacin, ethambutol, clofazimine, amikacin	8 mo	Treatment failure; alive +(1 yr later)

^aF, female; M, male; GORD, gastroesophageal reflux disease; BAL, bronchoalveolar lavage; BA, bronchial aspirate; GA, gastric aspirate; COPD, chronic obstructive pulmonary disease; HIV, human immunodeficiency virus.

DISCUSSION

To the best of our knowledge, this is the largest cohort of patients with clinical isolates of M. simiae reported in the literature to date (4, 6, 7, 15). Reunion Island is a French overseas department located in the Indian Ocean, with a population estimated in 2016 to be more than 850,000 (Institut National de la Statistique et des Etudes Economiques, https://www.insee.fr/fr/statistiques/2018922). Medical care in Reunion Island is similar to that offered in metropolitan France. The incidence of tuberculosis in Reunion Island is similar to that in metropolitan France, being around 8 cases/100,000 inhabitants per year. Except for a few private laboratories that send respiratory samples to metropolitan France for NTM identification, most mycobacterial samples are referred to one of the two university laboratories on the island (St. Pierre and St. Denis). Moreover, even if patients are initially tested in private laboratories, the majority of those found to be infected with mycobacteria are referred to one of the two university hospital centers in Reunion Island. Although our study was not exhaustive, we estimated a high incidence of NTM in Reunion Island, i.e., approximately 5 cases/100,000 inhabitants per year, about 5 times the incidence observed in metropolitan France. The cystic fibrosis prevalence in Reunion Island is similar to that in metropolitan France, but the NTM prevalence in cystic fibrosis patients is 3 times higher in Reunion Island than it is in metropolitan France (16). M. simiae was common in our cohort, accounting for 15.1% of all patients with NTM isolates, a rate 15 times higher than that observed in metropolitan France (17). We hypothesize that the tropical climate, with high water temperatures, explains the high incidence of NTM, and especially of M. simiae, in Reunion Island (18). NTM species have indeed been found to be common in other tropical regions, such as in Hawaii, as reported by Adjemian et al. (19). High vapor pressure and household proximity to water have also been reported to be associated with NTM infections among cystic fibrosis patients (20–22). Another explanatory factor for the high incidence of NTM may be contamination of water supplies with M. simiae, a finding that was previously reported in a study conducted in Texas (5).

Underlying diseases were common among our patients with M. simiae. The main general predisposing factor was immunodepression, with HIV infection affecting 4.1% of patients who were positive for M. simiae and 25% of patients who were treated for M. simiae infection. Other general predisposing factors were diabetes mellitus and alcohol abuse. Unfortunately, tobacco status and occupational exposure were not clearly reported in our data sources. Wood smoke exposure was common in Reunion Island until the early 1980s, due to cooking habits, but it is rarely notified or quantified in medical reports. Non-cystic fibrosis bronchiectasis was the most prevalent local predisposing factor, affecting 49.5% of our cohort patients and 71.4% of those who met the ATS criteria for infection. In multivariate analysis, the rate of bronchiectasis was significantly higher in patients who met the ATS criteria for M. simiae infection. Our conclusions are limited, however, as bronchiectasis diagnosis and NTM isolation were often performed simultaneously, making it difficult to determine whether bronchiectasis was a predisposing factor for or a consequence of M. simiae infection. Preexisting bronchiectasis was documented for only 2 of the 6 patients treated for M. simiae infection.

Our study confirms the low pathogenicity of M. simiae in humans; 21.6% of our cohort patients met the ATS criteria for infection, but only 8.2% (8 patients) received treatment. In the Texas study, only 6 (10%) of the 62 patients with M. simiae isolates had definite infections (4). Similarly, Rynkiewicz et al. (23) found in their study that only 2 of 23 patients with M. simiae isolates (9%) had true M. simiae infections. In a study conducted in the Netherlands, van Ingen et al. (7) found 28 patients with clinical isolates of M. simiae, among whom only 8 (21%) met the ATS criteria for infection and only 6 received treatment (7). While the appropriateness of gastric aspiration for the diagnosis of M. simiae pulmonary infection may be debated, it did not affect the proportion of infected patients in our cohort.

As reported elsewhere (7, 15), clinical signs of infection among our cohort patients were mostly respiratory symptoms such as productive cough. Our study confirms the findings of van Ingen et al. (7), with patients who lost weight being significantly more likely to meet the ATS diagnostic criteria.

Micronodular lesions were the most common radiological findings and were associated with ATS criteria for M. simiae in the multivariate analysis. Cavitary lesions were present in 25.8% of our cohort patients, compared with studies that reported cavitary lesions in 3% to 15% of patients (11, 15). Unlike in the study by van Ingen et al. (7), cavitary lesions were not associated with ATS criteria for infection in our cohort. We found cavitary lesions in 5 of the 8 treated patients, suggesting more cavitary lesions with M. simiae disease than with Mycobacterium avium complex (MAC) species.

In vitro drug susceptibility testing showed a natural multidrug resistance profile. As in the study by van Ingen et al. (7), we found most strains to be resistant to rifampin, ethambutol, and isoniazid. However, unlike van Ingen et al. (7), we found susceptibility to amikacin, fluoroquinolones, and clarithromycin to be significant. While previous studies showed M. simiae strains to be susceptible to clofazimine and cycloserine (7, 24), susceptibility to these antibiotic agents was not tested in our study.

The treatment regimens administered to our patients included macrolides, rifampin, ethambutol, moxifloxacin, clofazimine, and amikacin. Treatment decisions were more frequent for infected patients with positive smears. Responses to treatment were poor, as 3 of 7 patients who completed the treatment regimen experienced failure or relapse. These results, however, are better than those reported in the literature; van Ingen et al. (7) found that treatment was successful for only 1 of 3 patients in their study, and 14 of 18 patients with AIDS in the Texas study by Al-Abdely et al. (25) died within 6 months, with no response to antibiotics.

This study has several limitations. The retrospective nature of the analysis is clearly a weakness. Despite being the largest study to date, the small numbers of evaluated patients and events might have resulted in a lack of power. The small number of respiratory samples from patients who did not meet the ATS criteria for infection is another limitation of our study, as it suggests that some patients were misclassified due to missing microbiological data.

The data reported here reflect common medical practice, which may also be viewed as a limitation. The conclusions of this double-center study cannot be extrapolated to institutions outside Reunion Island, especially with regard to the choice of antibiotics. While single-center studies are useful to reveal local ecological patterns, larger studies can help to uncover regional and global trends that are not apparent in smaller studies.

In Reunion Island, M. simiae isolates are common but are associated with low pathogenicity. In vitro susceptibility to antibiotics is low, and clinical responses to treatment are poor. Further studies are needed to explore the impact of the environment on NTM isolates in this specific region.

MATERIALS AND METHODS

Patients and samples.

We conducted a retrospective study including all patients whose respiratory samples tested positive for NTM between January 2002 and July 2017, in the microbiological laboratories of the two university hospitals of Reunion Island (St. Denis and St. Pierre). Among patients with NTM isolates, only those with 1 or more isolates positive for M. simiae were included in the study.

All patient data (clinical information, biological findings, and radiological examination results) were recorded in the informatics databases of the two centers from 2002 to 2017. The results of out-of-hospital examinations were also included in the data collection when findings were found in the archived files or reported in clinical observations. We retrospectively reported the following: clinical findings including demographic data, predisposing conditions, and clinical symptoms of NTM infection, according to the ATS criteria (1); radiological signs of NTM infection meeting the ATS criteria (1) (radiologist interpretations of computed tomographic scans were included in the patient files and retrospectively reviewed by two pulmonologists); microbiological findings including the number, type, and result of respiratory samples and drug susceptibility; and, for treated patients, the treatment regimen, duration of treatment, and outcome.

Patient isolates were submitted to the local laboratory or were referred to the Centre National de Référence des Mycobactéries in Paris for identification and drug susceptibility testing of Mycobacterium isolates. Drug susceptibility testing was performed with the agar dilution method, using 7H9 broth and the SLOMYCO Sensititre panel.

Statistical analysis.

Results were expressed as the total number and percentage for categorical variables and as the median and interquartile range (IQR)s for continuous variables. Continuous variables were compared using the Mann-Whitney U test. Categorical variables were compared using the chi-square test or Fisher's exact test, as appropriate. After bivariate analysis, risk factors for M. simiae infection, as defined by the ATS, were entered into a multivariate logistic regression analysis using backward selection with P values of <0.15. Collinearity between independent factors was investigated; when collinearity was identified, the most clinically relevant factor was chosen for use in the multivariate model. P values of <0.05 were considered significant. Analyses were performed using SAS statistical software (version 8.2; SAS Institute, Cary, NC, USA).

Ethics approval.

This study was approved by the Comité d'Evaluation des Protocoles de Recherche Observationnelle of the Société de Pneumologie de Langue Française. Data collection was declared to the Commission Nationale de l'Informatique et des Libertés.