Prevalence of Mycobacterium lentiflavum in cystic fibrosis patients, France

Between January 2010 and September 2014, respiratory tract specimens (sputum, bronchoalveolar lavages and bronchial aspirates) collected from 354 cystic fibrosis patients (235 adults ≥18 years and 119 children <18 years) with a female/male ratio of 199/155 (56.2 %) were analyzed for mycobacteria (mean of 13.1 collected specimen/patient).

In our series, 25/354 (7.1 %) cystic fibrosis patients (twelve children and thirteen adults) had at least one respiratory tract specimen that yielded NTM, including twelve (48 %) patients with M. abscessus complex mycobacteria, eight (32 %) patients with M. avium complex mycobacteria and six (24 %) patients with M. lentiflavum (Fig. 1); one patient had both M. avium and M. lentiflavum successively isolated during the study. A total of thirteen M. lentiflavum isolates were identified on the basis of 99.6 % ± 0.003 % similarity with the reference M. lentiflavum CIP 105465^T partial rpoB sequence [GenBank:EU109300]. The six M. lentiflavum patients were co-infected by Staphylococcus aureus, Gram-negative bacilli and fungi and were all under azithromycin long-term low-dose prophylaxis (250 mg three times a week). In two patients, M. avium had been previously isolated in 2004 and 2012 from respiratory tract specimens (Table 2).

The female/male sex ratio of M. lentiflavum patients (0/6) significantly differed from that of the NTM positive cohort (199/155; 56.2 %) (p = 0.007, Fisher exact test) and from that of patients with another NTM (11/9; 55 %) (p = 0.02). It further differed from that of M. avium complex patients (5/3; 63 %) (p = 0.03) and that of M. abscessus complex patients (6/6; 50 %) (p = 0.054) (Fig. 2a). The six M. lentiflavum patients were aged 22.2 ± 11.4 y; the patients with another NTM were aged 22.1 ± 15.1 y; the M. avium complex patients were aged 26.5 ± 19.9 y; and the M. abscessus complex patients were aged 19.2 ± 10.9 y (no statistical significance; p = 0.56; ANOVA) (Fig. 2b).

Two M. lentiflavum patients were clinically stable, had only one positive specimen and were classified as ‘colonized’ [12]. The four other M. lentiflavum patients had between two and four positive sputum specimens and in two of them, M. lentiflavum isolation occurred contemporaneously to the decline of their lung function and thus may fulfill the American Thoracic Society’s (ATS) criteria for NTM lung infection [12] (Additional file 2: Figure S1). Thereafter, the forced expiratory volume improved during antibiotic treatment for M. lentiflavum. One of these two infected patients, aged 17, underwent a double lung-transplant because of poor progression of cystic fibrosis, two years after M. lentiflavum infection had been treated with combined rifabutin, clarithromycin and ethambutol for fourteen months and amikacin for one week. At the four-month follow-up after transplantation, microbiological surveys did not yield any further M. lentiflavum from four separate bronchoalveolar lavages.

In this study, we confirmed the identification of M. lentiflavum, a fastidious organism usually isolated over a period of three weeks, using rpoB partial sequencing [10] and a specific real-time PCR assay targeting the M. lentiflavum smpB gene. Indeed, M. lentiflavum shares <96 % similarity regarding the rpoB gene sequence with closely related species, including Mycobacterium stomatepiae DSM 45059, Mycobacterium florentinum DSM 44852, Mycobacterium genavense FI-06288 and Mycobacterium triplex ATCC 700071 [GenBank:HM022213, HM022205, HM022216 and GQ153311]. Moreover, the routinely used 16S-23S rRNA intergenic spacer sequencing [5], hsp65 restriction fragment length polymorphism and sequencing [13], and commercial probes for M. avium complex [7, 8, 13] may not be sufficiently discriminative as, for example, M. lentiflavum shares > 99 % similarity in the 16S rRNA gene sequence with M. simiae [1].

M. lentiflavum has emerged over a five-year period as the third most prevalent NTM isolated from the respiratory tract in our cystic fibrosis cohort. Furthermore, we observed an unexpectedly higher prevalence of patients (6 out of 354; 1.7 %) showing M. lentiflavum isolation than other reported epidemiological surveys. Indeed, only one out of 2912 (0.03 %) French patients [5] and two out of 2970 (0.06 %) American patients [7] were reported in previous studies. In addition, a recent epidemiology survey of NTM isolated in cystic fibrosis patients in Turkey revealed nine M. lentiflavum isolates collected from one young male teenager out of 130 (0.8 %) cystic fibrosis patients [8]. The reason why M. lentiflavum has only been isolated in male cystic fibrosis patients remains unexplained. As reported by Bryant et al. [14], patient-to-patient contamination may be suspected although this type of cross-contamination is very rare and whole genome sequence analyses would probably be required to satisfactorily conclude a phylogenetic link and track transmission events. Moreover, the six patients described here were treated in two distinct centers (adult and pediatric) for cystic fibrosis. Environmental transmission is another hypothesis for such prevalence.

An in-lab contamination hypothesis has been proposed, and eleven non-cystic fibrosis patients yielded M. lentiflavum isolates (out of more than 800 patients (≈1.3 %) who yielded at least one mycobacterial isolate) during the same five-year period. This demonstrates that M. lentiflavum was isolated every two months on average and that the probability of in-lab cross-contamination does exist but remains low.

In order to detect M. lentiflavum rapidly in cystic fibrosis patients, we developed an ‘in-lab’ real-time PCR targeting the M. lentiflavum smpB gene. This real-time PCR proved its ability to identify all M. lentiflavum isolates specifically. Moreover, our preliminary results indicate that this real-time PCR may be used as a first screening step directly performed on heat-inactivated sputum specimens with good sensitivity and 100 % specificity. These results have to be compared with traditional laboratory tools (culture and AFB smears) to clarify its relevance for clinical practice [15] and have to be validated on larger series of prospectively-collected sputum specimens, including from patients who had previously yielded M. lentiflavum in sputum cultures.

M. lentiflavum had been considered to be a harmless organism. However, this interpretation was recently challenged by the publication of a few cases with disseminated M. lentiflavum infections [16‐18]. In one case, hemophagocytic lymphohistiocytosis and disseminated M. lentiflavum infection in a heart-transplanted patient led to the death of this immune-compromised patient within ten days [18].

In the present study, two out of six patients fulfilled the ATS clinical and microbiological criteria for NTM lung disease [12] and had improved respiratory function while receiving specific antibiotic therapy. However, ATS criteria, while they continue to be applied to cystic fibrosis patients, are far from specific for such patients where radiographic findings, which are often associated with NTM, are commonplace irrespective of colonization. Moreover, clinical decline occurs in cystic fibrosis for a multitude of reasons. The antibiotic therapy our patient received is, moreover, active against a wide range of respiratory tract pathogens.

In our series, all patients were receiving long-term azithromycin therapy (Table 2). This use of macrolide in CF patients was shown to be a risk factor for NTM infection, especially with M. abscessus [19], by inhibiting intracellular killing of mycobacteria in macrophage by impairing autophagic and phagosomal degradation [20]. Such mechanisms may have played a role in the increase of M. lentiflavum isolation. However, as M. lentiflavum is usually susceptible to clarithromycin [8], the fact that patients were receiving macrolide and M. lentiflavum developed further supports the theory that this is either an environmental contaminant or a transient colonizer which has not been exposed to macrolide for prolonged periods.