Nontuberculous Mycobacterial Infections in a French Hospital: A 12-Year Retrospective Study

Background Nontuberculous mycobacteria (NTM) are environmental organisms associated with a range of infections. Reports of NTM epidemiology are mainly focused on pulmonary infections and isolations, and extrapulmonary infections are less frequently described. Methods We conducted a retrospective study of NTM infections at the Bordeaux University Hospital, France, between January 2002 and December 2013. We used the microbiologic component of the American Thoracic Society/Infectious Diseases Society of America's pulmonary NTM disease criteria to define cases of pulmonary NTM, and patients with isolates from a normally sterile site were classified as having extrapulmonary disease. Results In our setting, 170 patients were included. Pulmonary cases predominated (54.1%), followed by skin and soft tissue infections (22.9%), disseminated cases (10.6%), lymphadenitis (7.7%), bone and joint infections (2.9%) and the remaining 1.8% catheter-related infections. Overall, 16 NTM species were isolated. Mycobacterium avium (31.8%) and M. intracellulare (20%) were the most common species identified, followed by M. marinum (13.5%), M. kansasii (10.6%), M. xenopi (9.4%), rapidly growing mycobacteria (9.4%) and other slowly growing mycobacteria (5.3%). In general, NTM isolates were largely prevalent in people older than 50 (62.4%); patients aged 1–10 year-old exclusively yielded M. avium from lymph nodes, almost cases having being diagnosed after 2007. Among the 121 patients with complete follow-up, 78 (64.5%), 24 (19.8%), and 19 (15.7%) were cured, experienced relapse, or died, respectively. Conclusion In our study, extrapulmonary NTM infections represented almost half of cases, consisting mainly in skin and soft tissue infections. The increase lymphadenitis cases in children after 2007 could be linked to the cessation of mandatory BCG vaccination in France. We observed similar cure rates (64%) between pulmonary and extrapulmonary infections.


Introduction
Nontuberculous mycobacteria (NTM) are environmental organisms found in soil and water throughout the world [1]. Unlike Mycobacterium tuberculosis, which is a genuine pathogen, the largely environmental NTM have been often associated with conditions of host impaired immunity, such as primary immunodeficiency, Human Immunodeficiency Virus (HIV) infection or the use of immunosuppressive medication. Increasingly, NTM are isolated in immunocompetent individuals, with or without pre-existing structural lung damage. NTM are a group of over 150 species, but, due to recent advancement of molecular techniques, novel species are being described [2]. They are considered opportunistic pathogens, and several species are associated with human disease. Pulmonary infections are the most frequent, followed by skin/soft tissue infections, lymphadenitis in children, or disseminated infections in severely immunocompromised patients. NTM are an important cause of morbidity and mortality, often on the form of progressive lung disease. Although significant differences in geographic of NTM have been observed [3], Mycobacterium avium and Mycobacterium intracellulare, species belonging to Mycobacterium avium complex (MAC), are the most frequently reported. Other important human NTM pathogens included slowly growing mycobacteria (SGM), such as Mycobacterium kansasii, Mycobacterium xenopi, Mycobacterium malmoense, Mycobacterium marinum and rapidly growing mycobacteria (RGM) such as Mycobacterium abscessus, Mycobacterium chelonae and Mycobacterium fortuitum [4].
NTM infections are rare (1.0 to 1.8 cases per 100 000 persons) [4], even though their numbers increased during the last 20 years [5]. The diagnosis of mycobacterial infections remains difficult, relying on clinical, bacteriological and radiological criteria, especially for pulmonary disease [4].
We conducted an epidemiological and clinical retrospective study on all NTM infections diagnosed at the Bordeaux University Hospital (France) over a 12-year period. We aimed to describe microbiological and clinical characteristics, and patient outcomes.

Methods
The database of the bacteriological laboratory at the Bordeaux University Hospital was used to identify patients with isolation of NTM between January 2002 and December 2013. Only patients followed up at the Bordeaux University Hospital were included in the study. Demographic (gender, age) and clinical data (site of infection, underlying disease, impaired immunity, symptoms, radiographic results, treatment and outcome) were obtained from clinical records.
Diagnosis criteria for NTM lung disease were those of the American Thoracic Society (ATS)/Infectious Diseases Society of America (IDSA) recommendations [4). Patients with cystic fibrosis were excluded. Radiographic abnormalities (nodular, infiltrative or cavitary) were analyzed by chest radiography and/or CT scanning.
We categorized extrapulmonary NTM disease cases as: skin and soft tissue; disseminated; lymphadenitis; bone and joint or catheter-related infections. We defined a case of skin and soft-tissue disease as one or more isolates from tissue samples labeled as "furuncles", "abscesses", "skin", "wound", or "synovial fluid". We defined a disseminated case as NTM isolation from blood, bone marrow or other sterile site body fluids such as spinal site. Case of lymphadenitis was defined as isolation of NTM from a lymph node biopsy or aspirate. The category of bone and joint was used for positive bone biopsy or joint fluid cultures. Catheterrelated infection was defined as isolation of NTM from normally sterile material.
Microscopic examination of the samples was performed by Zielh-Neelsen staining, and specimens were inoculated onto a Löwenstein-Jensen and a Coletsos sold media (Bio-Rad, Marnes La coquette, France) and onto a liquid culture media using the Bactec MGIT 960 mycobacterial detection systems (Becton Dickinson, USA). NTM identification was based on molecular characterization. According to morphology of the bacilli on Ziehl-Neelsen, the appearance and pigmentation of colonies on solid media and the growth rate, we first used hybridization with DNA probes (AccuProbe 1 ) available for M. avium complex, M. avium, M. intracellulare, M. gordonae or M. kansasii when we suspected one this species. If the result was negative or if we suspected another species, we used the GenoType Mycobacterium CM/AS line probe assays (Hain Lifesciences) as recommended by the manufacturer. If all these tests were inconclusive, 16S DNA sequencing was realized.
Statistical analysis was performed using the chi-square test (or Fischer's exact test when appropriate) for categorical variables and the Krustal-Wallis test for continuous variables. A pvalue of <0.05 was considered as significant.
This study was conducted according to the principles expressed in the Declaration of Helsinki. The study was conducted in accordance with the guidelines of the "Direction de la Recherche Clinique et de l'Innovation", the research board of Bordeaux University hospital, Bordeaux, France. Informed consent was waived because this study only used the data gained from clinical practice. All patient data were anonymously reported, with no possibility of connecting the isolates and specimens to individual patients.

Results
A total of 198 patients were retrospectively selected. Twenty-eight were excluded because of the lack clinical data. Overall, 170 patients were included in the study. We identified 92 (54.1%) and 78 (45.9%) patients with pulmonary and extrapulmonary infections, respectively (Table 1). There were 104 men (61.2%) and 66 women (38.8%) with a median age of 55.5 years . Patients with pulmonary infections were significantly older than those with extrapulmonary infections (61.5 vs 52 years, p = 0.001). Seventy-six patients (45.2%) had immunodeficiency of whom 36 (21.4%) were HIV-positive. During the study period 16 different mycobacterial species were identified ( Table 2). No patient was infected by more than one NTM species.

Pulmonary infections
The median age of infected patients was 61.5 years   (Table 1). NTM isolates were prevalent in people older than 60 years (n = 49, 53.3%). Structural lung damage predisposing to NTM pulmonary infection was reported in 42 cases (13 patients had chronic obstructive pulmonary disease, 16 had bronchectiasis, 6 had emphysema, 11 had a history of pulmonary tuberculosis infection and 5 had other causes of chronic pulmonary disease). Thirty five patients (38%) reported smoking actively at the time of their infection. Interestingly, 52.2% (48/92) and 54.3% (50/92) of the patients have no detected underlying lung disease or immunodeficiency, respectively.
The main characteristics of patients with M. avium, M. intracellulare, M. xenopi or M. kansasii infections are presented in Table 3. Median age was >60 years for M. avium and M.   and M. xenopi . Among the 20 patients treated for M. avium pulmonary disease, 13 were considered as cured, four relapsed, one was lost to follow-up, and two died. Similarly, among the 20 subjects treated for M. intracellulare pulmonary infection, 12 cured, two relapsed, four were lost to follow up and two died. Among patients treated for M. xenopi, five cured whereas five relapsed; two were lost to follow-up and two died. For M. kansasii, eight patients were cured after treatment, three were lost to followup and one died; none relapsed.

Extrapulmonary infections
We identified 78 patients with extrapulmonary manifestations (45.9%). Skin and soft-tissue infections were the most common, accounting for 50% (n = 39) of these manifestations in the 12-years period. Almost all specimens received for microbial diagnosis were skin biopsies, the remaining being synovial biopsies (S2 Table). Smear was positive in only three cases. Most infections were due to M. marinum (59%, 23/39), followed by RGM (15.4%, 6/39), M. intracellulare (12.8%, 5/39), M. avium (5.1%, 2/39), and M. terrae, M. xenopi, M. szulgaï isolated once. Histological examination, realized in 31 cases, showed granulomatous inflammation with or without caseous necrosis in 21 cases (S2 Table). Among the 23 patients with M. marinum, infection was limited to a skin disease on the limbs for 18 patients but spread to deeper structures, resulting in tenosynovitis, for five patients. Overall, 17 received treatment, one did not after medical decision and no information about treatment was available for five patients. Monotherapy was prescribed for 11 patients, mostly with clarithromycin (n = 9), azithromycin (n = 1) and doxycycline (n = 1). Five patients received two antibiotics and one was treated with three drugs. Median duration of antibiotherapy was 3 months [1][2][3][4][5][6][7][8][9][10][11][12]. Surgery was performed only for patients with tenosynovytis, in association to bi-or-tri-antibiotherapy. Clinical course, obtained for 11 treated patients, revealed cure for 10 of them and relapse for only one. Among the 10 patients with SGM other than M. marinum skin and soft-tissue infections, five had tenosynovitis (3 with M. intracellulare, one with M. terrae and one with M. xenopi); the five remaining, of whom two were immunosuppressed, presented nodular lesions. Nine patients received antibiotics (from one to four drugs), with a median duration of eight months ; clinical cure was observed for four patients, relapse in three cases and two were lost to follow-up. The tenosynovitis with M. terrae was diagnosed in a farmer, first in 2004 after tenosynovectomy, and then in 2008 and 2010 because of relapse requiring new surgery. Among the six patients with infection caused by a RGM, three were immunocompetent and skin lesions (nodules, cutaneous necrosis) appeared after piercing or cosmetic surgery; the three others received immunosuppressive treatments and presented multiples nodules on the upper or lower limbs. All, except one, were treated with one or two antibiotics, with a median duration of five months [3][4][5][6] and experienced clinical cure.
Disseminated cases were 10.6% of the total NTM cases. All patients were immunosuppressed, of whom 83.3% (15/18) were HIV-positive. The median number of specimens received per patient was 5 [1][2][3][4][5][6][7][8][9] with a median of two positive specimens [1][2][3][4][5][6][7][8]. Most infections were due to M. avium (72.2%, 13/18) (S2 Table). The disseminated infection with M. chelonae was reported in a man who had a tracheotomy post-radiation since seven years and who was hospitalized for pace-maker exteriorization following fall; all specimens (bronchial aspirates, sample of the lodge of the pace-maker and blood culture) grew with M. chelonae. Fourteen patients received a combination of two or three antibiotics from one to 22 months (median: 7.5 months). Five were cured, two relapsed, three were lost to follow-up and four died. Three patients were not treated because death occurred shortly after the discovery of the MNT and one patient was lost to follow-up. Deaths were attributed to the NTM infection only in one case (M. chelonae disseminated infection), to the underlying disease in seven cases and to another cause in one case.
Thirteen cases of lymphadenitis were identified (S2 Table). They mostly occurred in children (n = 8) aged between 1 and 4 years, of whom six were girls. Regarding BCG vaccination status, six children were not vaccinated and status was unrecorded for two. Locations were submandibular (n = 6) or parotid (n = 1); one patient had two distinct lesions (submandibular and preauricular). None had constitutional symptoms. At diagnosis, painless and firm lymphadenopathy was observed in 72% of cases, skin changes associated with violaceus coloration in 14% and cutaneous fistulae in 14%. Direct smear was positive in only two cases and histological examination was contributive in four cases. All cases were due to M. avium. Five children were treated by surgery and antibiotics, one exclusively by surgery, one by antibiotherapy and one received no treatment; two patients presented relapse. Among the five cases that occurred in adults, four were HIV-positive. Location was mainly cervical (60%). Direct smear was positive in only one case and histological results were not specific. M. avium was identified in four subjects and M. kansasii in one. Four patients were treated with a combination of two or three antibiotics during a median duration of 11 months [9][10][11][12][13][14][15][16][17][18][19][20][21]; two patients died because of underlying disease, one relapsed, one observed clinical cure and one was lost to follow-up.
Bone and joint infections were diagnosed in five cases (2.9% of all NTM cases) (S2 Table). All patients were immunocompetent, except one who was HIV-positive. Direct smear was negative in all cases. Three infections were due to RGM: a septic arthritis with M. abscessus after needle injection of corticosteroids, a bone infection of a knee with M. chelonae after a ligamentoplasty and bone infection of a foot with M. chelonae. These patients were treated during 4 or 7 months and cure was obtained for all of them. A prosthetic joint infection was due to M. intracellulare and successfully treated after one year of antibiotherapy. The last case, a bone infection of the elbow caused by M. kansasii, relapsed despite appropriate treatment.
Two port infections due to M. chelonae were identified, and data obtained for one revealed cured after three months of antibiotherapy. One case of cardiac device-related infective endocarditis due to M. peregrinum was recorded; the patient was successfully treated after surgery and three years of antibiotherapy.

Discussion
Similar to other reports, NTM isolations from pulmonary sites occurred most frequently [6,7]. Pre-existing chronic lung diseases, especially chronic obstructive pulmonary disease, asthma and bronchiectasis, are the main risk factors for NTM pulmonary infection [4]. However, consistent with our findings, a considerable proportion of patients with this disease have no detected underlying lung disease or immunodeficiency [8].
In the present study, the species distribution of NTM isolated in pulmonary disease was close to that reported earlier in France [9] and more recently in an inventory study of NTM in the European Union, which included France [3]. MAC was the commonest group of cultured organisms, as reported worldwide [3]. While studies showed that M. avium was the predominant subspecies recovered from human biospecimens [3,10], we found a similar number of M. avium and M. intracellulare isolates in our cohort. M. xenopi, followed by M. kansasii and M. malmoense, is more frequently encountered than is M. abscessus in non-cystic fibrosis patients [3].
In Europe, the manifestations of the MAC pulmonary infection vary in the majority of cases. In the Netherlands, Van Ingen et al. reported that patients with MAC pulmonary disease more often had cavitary than nodular bronchiectatic disease [11]. In contrast, in France, among patients with MAC pulmonary infection, there were lower rates of fibrocavitary than of nodular bronchectatic disease [9,10], which is similar to the results of recent surveys in the USA and South Asia [12][13][14][15] and to our results. It is of interest to note that patients infected with M.avium or M. intracellulare have similar radiological characteristics in our study. It also appears that infections with M. kansasii, M. malmoense or M. xenopi were more frequently associated with fibrocavitary disease [8].
Similarly to previous reports [6], we didn't found any association between NTM pulmonary infection and gender, although M. kansasii appeared to be more common in men (81.2%, 13/16). In contrast, some authors described highest prevalence of MAC infection in women [9].
Previous investigations described increase rates of NTM disease in the elderly [6,7,16]. We found that the age distribution differed among isolation sites. We reported a rise isolation prevalence of pulmonary disease in the oldest (60 years and over] group. For extrapulmonary NTM infections, the highest rates occurred in the "40-59 years" and "60 years and older". Interestingly, in our cohort, although pulmonary disease was predominant, we observed a high rate of extrapulmonary manifestations (45.9%). This is more than that described in literature, with prevalence around 20-25% [6,7,17]. Among them, skin and soft tissue infections were the most encountered. Overall, 70% of cutaneous isolates were M. marinum or RGM, which has also been noted by others [18]. M. marinum was isolated exclusively from skin samples, reflecting the ability of the mycobacterial species to infect and localize in different body sites [19]. M. marinum human infections are generally cutaneous, although can become deeper in some cases resulting in tenosynovitis, arthritis and osteomyelitis [20]. Antimicrobial susceptibility testing is not recommended as M. marinum is usually susceptible to antimicrobials used for treatment, including rifampicin, ethambutol, doxycycline-minocycline, co-trimoxazole and clarithromycin [21]. Testing is however recommended for those that fail to respond to treatment after several months and with positive cultures. As previously described [22], the median duration of antibiotic therapy was three months in our study and recovery was obtained in 90% of the cases. Because M. marinum infection is usually localized, single therapy is widely used. Aubry et al. reported that the duration of antibiotics was significantly longer for patients with deeper structure infections than for patients with infections limited to the skin and soft tissue [22], and that surgery was indeed associated with deep structure infection.
As for pulmonary infections, MAC was the commonest group of cultured organisms from extrapulmonary sites [16], mainly responsible of disseminated cases and lymphadenitis. Disseminated mycobacterial infections developed in immunocompromised people, most being HIV-positive (83.3% in our work) [19]. Inside the MAC, M. avium was isolated in all cases, except one for which M. intracellulare was identified, which is consistent with the current literature [19,23].
In our study, lymphadenitis is the clinical manifestation of NTM infection of children. It occurred in immunocompetent children less than 5 years old, which is consistent with other reports [23]. M. avium, the sole species identified in the eight cases of our cohort, is the main cause of these infections [24]. M. malmoense is frequently described in the United Kingdom and Sweden, M. haemophilum in Israel and The Netherlands [24], and M. lentiflavum in Spain [24,25]. It is of interest to note that none of the eight children was vaccinated with the BCG and that all cases were diagnosed after 2007, year of the cessation of compulsory BCG vaccination in France, suggesting possible cross protection by BCG to mycobacteriosis.
Our therapeutic approach was in accordance with IDSA guidelines [4]. Duration of treatment varies according to the species and localization. Treatment duration of 12 months after sputum culture conversion is generally recommended for most NTM pulmonary disease [4]. The cornerstones of most anti-NTM drug regimens are macrolides, but there are exceptions such as M. kansasii.
In our cohort, deaths of patients with NTM pulmonary disease were not attributed to NTM infection, but to underlying disease (six cases) or another cause (three cases). This result is of interest as recent follow-up studies have indicated that pulmonary NTM findings carry a high mortality rate, attributed to the NTM infection, ranging from 24 to 69% within 5 years, [26][27][28][29][30][31][32]. The mortality among patients with atypical mycobacteria differs per species. In a population-based study from Denmark, M. xenopi was associated to poorer prognosis than MAC [27]. In accordance with these results, in a retrospective study from Helsinki, pulmonary MAC patients had significantly lower risk of death as compared to patients with pulmonary infection of other SGM and as compared to patients with pulmonary RGM [32]. The treatment success rate is higher (70-85%) in patients with non-cavitary nodular bronchiectasis MAC lung disease than in those with cavitary MAC lung disease [33]. The result obtained in our study could be explained by the high number of treated patients (72/92, 78.3%), the use of recommended regimens for MAC and M. kansasii, and a highest number of patients with nodular and infiltrates presentation than that fibrocavitary form.
The use of more recent drugs such as tigecycline may have allowed a better control of RGM infection. In our study, tigecycline was included in the initial treatment regimen in four patients with extra-pulmonary infection, in combination to one or more antibiotics, and complete resolution was observed after treatment courses of four to seven months. Tigecycline has consistently demonstrated excellent activity in vitro against clinical isolates of both tetracycline susceptible and tetracycline-resistant M. abscessus and M. chelonae, but few reports of clinical experience with tigecycline for the treatment of RGM infections are found in the medical literature. Recently, Wallace et al. reported the efficacy of tigecycline as part of an antibiotic regimen for salvage treatment of M. abscessus and M. chelonae infections in a large cohort of 52 patients [34].
This study has several limitations. First, we did not collect data outside of hospital. Second, incomplete reporting could undermine the accuracy of our estimates. Finally, due to the retrospective nature of the study accounts for the lack of information is missing regarding patient follow-up in 28% of the cases or sometimes in the justification for the absence treatment.
In conclusion, the present study, although not representing a population-based investigation, provides a snapshot of the prevalent NTM species in our setting. Extrapulmonary NTM infections represented almost half of cases, consisting mainly in skin and soft tissue infections. Clinicians should be aware of the increase lymphadenitis cases in children observed after 2007, which could be linked to the cessation of compulsory BCG vaccination in France. We observed similar cure rates (64%) between pulmonary and extrapulmonary infections.