Surgical site infections by atypical mycobacteria: prevalence and species characterization using MALDI-TOF and molecular LCD chip array

Surgical site infection (SSI) is one of the most common forms of post-operative complications and hospital-acquired infections [1, 2]. It adversely affects patients by causing delayed wound healing, increased length of hospital stay, prolonged use of antibiotics and increased mortality [3].

For proper management of infectious diseases, it is essential to accurately identify the causative pathogens. The surveillance data of the last two decades suggest a change in the profile of microorganisms causing SSIs, with available reports on the occurrence of SSIs by mycobacteria [4]. M. tuberculosis (MTB) attracts most global concern; however, infections caused by non-tuberculous mycobacteria (NTM) exist and may be overlooked due to limitations of conventional diagnostics in reaching accurate identification. Several reports on infections by NTM have been lately presented due to the achieved advance in diagnostic technologies [4‐6]. Atypical NTM have existed in nature over the years and reported in history as causative agents for different types of infections including post-operative infections [7]. It is postulated that contaminated water or surgical instruments might be possible sources of post-operative infection [6]. It has been reported that rapid grower NTM are responsible for most of atypical mycobacterial cutaneous and soft tissue surgical infections, where environmental contamination is usually the source [8, 9]. Clinical diagnosis of mycobacterial infections is considered challenging, and suspicion is usually built on late, mild, however persistent clinical presentation with poor response to antibiotics [10, 11]. Missed diagnosis may lead to inappropriate treatment and unfavorable emergence of antimicrobial resistance; therefore, appropriate laboratory diagnosis is considered essential [12, 13]. Moreover, in mycobacterial infections, it is crucial to differentiate atypical mycobacteria from M. tuberculosis due to their different therapeutic regimens [5], which necessitates the provision of rapid and accurate laboratory diagnostic methods for proper identification [12, 13].

Traditionally, the identification of NTM was carried out by conventional phenotypic biochemical tests. However, these methods were laborious and unable to identify a wide spectrum of species [14]. In this context, many researchers employed the advanced developed technologies such as matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) and molecular DNA hybridization LCD chip array in identification of mycobacteria offering more rapid and accurate diagnosis [15, 16]. To this end, the present work aimed to study the prevalence of atypical mycobacteria among patients with SSI and characterize their species using MALDI-TOF assay and DNA hybridization LCD chip array.

Out of a total 192 processed pus samples of patients suspected with mycobacterial infections by charged physicians, MGIT culture showed positive mycobacterial growth in 11 samples (5.7%). Compared to the gold standard mycobacterial culture, ZN smear showed true positive results in seven samples (smear positive/culture positive); however, four samples had false-negative smear results (smear negative/culture positive) and no samples showed false-positive smears. ZN smear showed a sensitivity of 63.6% (95% confidence interval: 0.316–0.87) and specificity of 100% (95% confidence interval: 0.97–1.00). The results of mycobacterial laboratory workup of pus samples (n = 192) are illustrated as shown in Fig. 1.

The phenotypic characteristics of the grown isolates and the immunochromatography assay divided the isolated mycobacteria into M. tuberculosis (n = 7) and NTM (n = 4), with prevalence of 3.64% and 2.1% of 192 suspected pus samples, respectively. NTM accounted for 36.4% of mycobacterial isolates: three were classified as rapid growers and one as a slow grower (Table 1).

The species of the four NTM isolates were characterized by MALDI-TOF MS and LCD array as described in Table 1. Using MALDI-TOF MS, NTM isolates were identified as M. abscessus (high confidence), M. porcinum (high), M. gordonae (low), and M. bacteremicum (low). The LCD array was able to characterize the species of three out of four NTM isolates, but one isolate failed to be identified. One NTM species was identified as M. abscessus in agreement with MADI-TOF identification, while the two other NTM were broadly classified as one belonging to broad II and the other belonging to broad III. These isolates could not be identified to species level, as many mycobacterial species deliver PCR fragments of equal or similar size that evolved in a group known as “Broad”.

Patients infection by NTM (n = 4) had a history of open abdominal operations except one who had laparoscopic surgery and another who had mesotherapy injection. Three patients were male smokers and two of them were diabetic, while no one had a history of HIV infection (Table 1).

Atypical NTM is not of lesser concern than M. tuberculosis in causing post-operative infections [21]. To the best of our knowledge, data on the prevalence and types of NTM among SSI patients in our country are still limited. In an effort to bridge this gap, we tried to highlight the importance of considering atypical NTM among post-operative mycobacterial infections and avoid being mistaken for MTB. Moreover, there is a need to emphasize the role of employing MALDI-TOF and PCR-based array technologies in improving mycobacterial identification and encourage their adoption in developing countries according to the available financial resources.

In the present study, mycobacterial isolates were grown with prevalence of 3.64% and 2.1% for MTB and NTM, respectively. The NTM rate was consistent with that reported by a study in India (1.2%) [22]. Another Indian study reported higher NTM rates (10.9%), which likely occurred in another Brazilian study that confirmed NTM in 144 out 1051 suspected surgical mycobacterial infections [4, 21]. Variable rates might be due to different geographical areas, study population, types of samples and methods of detection. Of note, most of the previously published reports on NTM were among pulmonary or extra-pulmonary infections other than SSI and limited publications were available for NTM among SSI, apart from some studies on outbreaks, case reports or case series [6, 8]. In our study, NTM were isolated from four patients who underwent open operation, mesotherapy injection or laparoscopic procedure, mostly suggested to have originated from contaminated water or equipment [10, 23, 24].

In terms of laboratory diagnosis, ZN smear is considered a simple and cost-effective method, but of low-test sensitivity [25]. This was supported by ZN sensitivity of our study (63.6%) which well matched with that reported by another study (66.9%) [25]. In our study, we used immunochromatography BD MGIT^™ assay that was encouraged by several studies for initial differentiation of NTM from M. tuberculosis [17, 26]. MALDI-TOF and LCD array were used for further characterization of the four recovered NTM isolates from pus culture. MALDI-TOF managed to identify four NTM species, where confident species identification was high in two isolates, however low in the other two isolates. This was as guided by the manufacturer; however, no global cutoff has been established for the species-level identification of the entire range of mycobacteria [18, 27‐30]. In our study, the LCD array agreed with MALDI-TOF in characterizing one NTM isolate (M. abscessus); however, two NTM were broadly identified as Broad II and Broad III classes and one NTM was undetected. This could be explained by the many mycobacterial species delivering PCR fragments of similar size, thus evolving into a group known as “Broad” [31].

M. abscessus in our study was previously reported by several studies as a clinically important pathogen [10, 32]. M. porcinum was first described in pigs, while in human beings it was reported as a cause of cellulitis, central catheter infections, and pneumonitis [33, 34]. M. bacteremicum had previous reports of challenged diagnosis by VITEK MS, but was successfully detected by 16srRNA [35]. This species was previously reported from post-laparotomy infections [36]. M. gordanae was previously detected in post-traumatic as well as respiratory samples among immunosuppressed patients [37, 38].

In the last decade, few advanced technologies have evolved to overcome the flaws in conventional methods including being time consuming and only able to characterize a narrow spectrum of species [39, 40]. MALDI-TOF MS has been witnessed as a rapid and reliable tool for NTM identification, which includes a wide mycobacterial spectra in software databases, as many as 149 different Mycobacterium species [19]. However, some challenges were reported related to the extraction protocol, cost of the device and the need for a primary bacterial isolate [18, 19, 41‐43]. LCD microarray is a rapid and reliable hybridization method with species-specific oligonucleotide probes that may reach a 100% sensitivity in mycobacterial identification [20, 44‐47]. LCD microarray might not outperform MALDI-TOF due to a narrower spectrum of NTM species identification, besides its high cost; however, it has the privilege of enabling mycobacterial detection and identification directly from the sample, unlike MALDI-TOF [20, 31].

Of note, from our experience in work, the LCD array device is less expensive than MALDI-TOF as a diagnostic instrument; however, it has a higher running cost regarding testing a sample. After all, every healthcare setting shall evaluate the cost/benefit aspect in choosing the optimum diagnostic method to adopt in the laboratory and avoid underreporting due to defective diagnosis.

Clinicians should be aware of NTM in the diagnosis of mycobacterial surgical site infections. The present study detected four NTM isolates with a prevalence of 2.1% among surgical site infections submitted for mycobacterial culture. M. abscessus, M. bacteremicum, M. porcinum, and M. gordanae were characterized by MALDI-TOF. LCD array agreed with MALDI-TOF in identifying M. abscessus. Both assays play a vital role in improving mycobacterial identification, but require continuous update to expand the identification spectrum of mycobacterial species.