Figure 1

Figure 1. Onset of clinical signs in 85 reported cases of infection with nontuberculous mycobacteria associated with healthcare and aesthetic procedures, by year, France, January 2012‒June 2020. Blue line indicates linear fit...

For the study period, 71 reports of extrapulmonary NTM infections related to HAI surgical, medical, or aesthetic procedures were included to give a total of 85 original cases, a mean of 10 cases/year for complete years (i.e., 2012–2019) (Figure 1). The regression identified an increasing trend of cases per year of onset of clinical signs during the study period after excluding incomplete years (p<0.01). This increase was observed particularly during the period 2016–2019 (i.e., after the M. chimaera heater–cooler unit [HCU] outbreak) but was caused by addition of these cases because regression analysis without them showed a similar significant trend (p<0.01). Among the 85 cases, 36 (42%) were found in the CNR-MyRMA strain database and the e-SIN database. The CNR-MyRMA database contained 30 additional cases, and the e-SIN database contained 19 additional cases.

The reports were received from 46 HCFs throughout France. Twenty-nine of the HCFs sent only 1 report during the study period, and 16 HCFs sent >1 (10 HCFs made 2 reports, 3 HCFs made 3 reports, 2 HCFs made 4 reports, and 1 HCF made 5 reports). Most (90%, 64/71) of the reports concerned an individual case, and 7 reports concerned clusters (2 cases in 4 reports, 3 cases in 1 report, and 5 cases in 2 reports).

For most (69%, 59/85) of cases, the infection was acquired inside the HCF, whereas 26 cases (31%) were acquired outside the reporting HCF. These infections acquired outside the HCF were imported either from another HCF (n = 14) or from a non–hospital-based medical or aesthetic practice (n = 12).

More women than men had cases reported (M:F = 31:46); sex was not reported for 8 case-patients. The median case age was 54 years (range 4–86 years; age was not reported for 12 cases). One third of the cases concerned immunosuppressed patients (33%, 23/70 cases; 15 cases did not report this information). NTM infections resulted from surgical procedures (56%, 48/85 cases), other invasive procedures (33%, 28/85 cases), and noninvasive procedures (11%, 9/85 cases). Cardiovascular surgery (n = 14), orthopedic surgery (n = 11), plastic surgery (e.g., breast surgery, (n = 13), face-lift (n = 2), abdominoplasty (n = 1), or capillary implant (n = 1), and catheter-associated infections (n = 17) comprised most of the surgical and other invasive procedure cases reported. The NTM infections concerned mainly skin and soft tissues (36%, 31/85), intravascular catheters (20%, 17/85), bones and joints (18%, 15/85), and arterial/cardiac (15%, 13/85) (Table 1).

Overall, 14 NTM species were isolated: 10 rapidly growing NTM species and 4 slowly growing NTM species. Rapidly growing NTM were M. chelonae (n = 30), M. fortuitum complex (n = 24) (17 cases with M. fortuitum, 2 with M. mageritense, 2 with M. porcinum, 1 with M. senegalense, and 2 strains for which the exact species could not be determined), M. abscessus (n = 14), M. mucogenicum (n = 5), M. neoaurum (n = 2), M. fuerthensis (n = 1), and M. wolinskyi (n = 1). Slowly growing NTM species were M. chimaera (n = 4), M. avium (n = 2), M. lentiflavum (n = 1), and M. marinum (n = 1). For 2 cases of endocarditis, direct examination of the valve samples identified acid-fast bacilli after Ziehl-Neelsen staining, but culture results were negative.

We determined the hypothesized incubation time (i.e., time between the onset of clinical signs and most probable contamination date, which is most often the date of the procedure) for 50 cases (Table 1). Incubation time was shorter for rapidly growing NTM (median time 34 days, n = 40) than for slowly growing NTM (median time 549 days, n = 5).

We suspected that medical devices were related to the infection for 80% of cases (68/85), but a medical device vigilance report was performed for only 21% of those (14/68 cases). Medical devices comprised implantable devices (e.g., breast prosthesis, artificial heart valve and vascular prosthesis, knee or hip prosthesis) for 50% of case-patients that had a medical device (34/68 cases), invasive devices (e.g., catheter and implantable port, dialysis device, endoscopy device, infiltration device in orthopedic surgery, liposuction cannula, mesotherapy, and tattoo injection equipment) (37%, 25/68 cases), and noninvasive devices (e.g., cardiopulmonary bypass HCU, contact lens) (13%, 9/68 cases).

For nearly half of all reported cases (47%, 40/85 cases; 30 cases did not report this information), there was a specific investigation of professional practices after the NTM infection was diagnosed to assess the level of compliance with hygiene guidelines, sterilization procedures, and treatment procedures. For more than one fourth of these cases (28%, 11/40), the investigations found failure to comply with infection risk prevention recommendations. Corrective measures were implemented for 45% of the reports (17/38 reports; 33 did not report this information), most often involving increased hygiene vigilance and recommendations to improve practices (n = 7). For 32 reports, there was no active case finding; most of them were isolated cases. For 30 reports, this information was not available. Active case finding was conducted after 9 reports (30 did not report this information) for procedures such as breast reconstruction, heart surgery, gastrointestinal endoscopy, mesotherapy, orthopedic surgery, and tattooing sessions.

Figure 2

Figure 2. Genomic comparison of nontuberculous mycobacteria isolates by using whole-genome sequencing phylogenetic analysis and maximum parsimony trees. A) Mycobacterium chimaera, B) M. fortuitum, C) M. chelonae...

Environmental investigations were undertaken for 42% (30/71) of the reports. Most (27/30) reports involved water sampling from potential sources of the contamination, such as water supply networks (n = 21), HCU (n = 4), dialysis water (n = 1), and swimming pool water (n = 1). Perioperative surfaces (n = 2) and air samples (n = 1) were rarely sampled. Among the 45% of cases (38/85) involving environmental investigations, mycobacteria samples were positive for 18 (7 were not reported). The same NTM species as in the clinical isolate was found for 12 cases (32%, 12/38) after environmental investigations. For 10 cases, the clinical isolate could not be distinguished from the environmental isolate (Table 2; Figure 2): M. chimaera isolates from HCU and heart surgery infection (patient A3) (Figure 2, panel A), isolates from hospital water supply network, M. fortuitum breast infection (patient C1) (Figure 2, panel B), M. chelonae skin and soft tissue infection (patients D1, E1-E2, E3, and F1) (Figure 2, panel C), M. marinum isolates from pool balneotherapy and skin and soft tissue infection (patient I1) (Figure 2, panel D), and M. mucogenicum catheter-associated infection (patients J1, J2, and J3) (Figure 2, panel E).

We also performed genomic comparisons for case-patients suspected of being contaminated by a common source. For 8/10 case-patients, studied isolates had the same pattern (Table 2; Figure 2). In report A, 3 clinical isolates of M. chimaera endocarditis from 2 patients were clustered (A1-A2 and A3) (Figure 2, panel A). These 2 patients were linked to a worldwide outbreak of HCU contamination, as shown in the section comparing clinical (A3) and environmental (A4-A8) isolates. In report H, which concerned M. chelonae catheter-associated infections diagnosed in the same institution (n = 5 cases), clinical isolates were clustered into 2 distinct groups; the first cluster grouped 3 isolates, H1, H4, and H5, from 3 patients, and the second cluster grouped 2 isolates, H2 and H3, from 2 patients (Figure 2, panel C). The presence of 2 clusters in the same HCF suggested that there were 2 sources of contamination, neither of which were found. In report K, the same genotype was found when comparing M. neoaurum isolates from blood cultures of 1 patient with the isolate found during microbiological control testing after a peripheral autologous stem cell transplant (n = 1 case) (Figure 2, panel F). The contamination of the stem cell transplant was attributed to colonization of the catheter used for the cell sampling.

We describe extrapulmonary NTM infections diagnosed after surgical, medical, or aesthetic procedures in France over an 8-year period (2012–2020). To broaden the spontaneous reporting from medical professionals, we sought 2 information sources: the national EWRS for HAI and the national reference NTM strain database. Because only 85 cases were described in 71 reports over 8 years, we might consider that such NTM infections remain rare. However, most of the cases were related to a medical device, a specific procedure, or lack of hygiene practices and might have been preventable.

Our study highlights a slight increase in reported annual numbers of cases during the study period. This increase, particularly during 2016 (14), could be explained by greater global awareness in public health community after invasive infections with M. chimaera associated with HCUs used during cardiac surgery (15) and other published outbreaks (6).

Both infection sites and NTM species isolated from these cases of infections were diverse, emphasizing the opportunistic nature of these pathogens. The most commonly reported infections were skin and soft tissue infections, catheter-related infections, infective endocarditis, and bone and joint infections. When we compared our findings with the major proportion of extrapulmonary NTM infections described in the literature, but not limited to healthcare-associated and aesthetic procedure‒associated infections, we found that skin and soft tissue infections were the most commonly reported infection site (16,17). However, we identified catheter-associated infections, bone and joint infections, or infective endocarditis, which are less described, except for the HCU M. chimaera outbreak (8).

A wide variety of NTM species were responsible for the infections reported in our study; M. chelonae and M. fortuitum were the most common. Isolates of slowly growing NTM species were rare (11%) compared with those found in a review in the United States, in which 50% of all species were in the M. avium complex (16). One possible explanation for this difference is that rapidly growing NTM cultivated on standard bacteriology diagnostic media might constitute an unexpected etiology diagnosis. We should also consider that these rapidly growing species were regularly found in France in the water networks, one of the sources of infection (1).

Most of the cases in our study were linked to surgical, invasive, and noninvasive procedures. In 5 cases, no specific procedures were identified as the cause of infection, even though all patients underwent healthcare procedures such as previous intravenous catheter, or no procedures were identified at all, such as the case of bone and joint infection, which could not have originated from spontaneous infection. As observed in previous reviews, the most commonly reported infections were those associated with aesthetic care (12), particularly breast prostheses (18).

The source of infection was not determined in 9/10 case-patients, despite environmental investigations conducted for ≈50% of the reports. Even when environmental investigations were performed, they occurred months after the suspected contamination because of long incubation times, as in reports B and G (Table 2).

Because a medical device was implicated in most of the NTM infection cases we reported, we believe that these infections can be prevented. The medical device could be contaminated by NTM before its use or can lead to contamination from environmental NTM (5,19). When an environmental cause is identified, the water system is the major environmental source most frequently considered responsible. Water systems, particularly in hospitals, are frequently identified as NTM reservoirs (20,21).

Genomic comparison of NTM isolates can be used to rule out or confirm any hypothesis concerning the origin of the contamination. However, careful analysis of genomic sequence comparisons should be conducted because several factors, such as the reference sequence on which the reads are mapped (epidemic strain or unrelated strain), quality of the sequenced data, coverage of the mapping assembly, number of sequences included in the comparison, and use of de novo assembly, influence SNP analysis. WGS appears to be a suitable tool for the molecular investigation of NTM infections, but might need expert rules and standardization to be used further.

The major limitations of this study concern the lack of completeness of the reported data. There is no specific surveillance system for NTM infection in France, and the 2 databases used for this case series are not exhaustive. The purpose of the EWRS for HAI platform is to improve the management of HAI reporting by HCF, and the CNR-MyRMA receives NTM isolates for patient diagnosis and treatment and genotypic comparison in epidemiologically related cases with environmental analyses when necessary. Therefore, underdeclaration of NTM infection cases in France is probable (22). However, when combined, the 2 databases provide a useful inventory of extrapulmonary NTM infection cases related to surgical, medical, and aesthetic procedures. The genomic comparison of NTM isolates performed by CNR-MyRMA was able to demonstrate the source when an environmental investigation was conducted and clinical and environmental isolates were available. This comparison provides valuable pointers for the future implementation, improvement, and follow-up of certain preventive measures.

Although data in the current study were not exhaustive, reports of NTM infection cases and subsequent microbiological and workplace practice investigations showed that considerable progress has been made in understanding contamination mechanisms during healthcare treatment. Water used in the procedures appeared to be the infection source for 10 cases. This finding is particularly true for heart surgery after the alert issued concerning the global outbreak of M. chimaera endocarditis as a result of contaminated HCUs (8).

Our observations should prompt more stringent recommendations for prompt reporting of NTM infections and provision of clinical and environmental samples for analysis of strains. Better application of these recommendations should improve methods to identify causes of NTM infections and enable their prevention.