Clinical characteristics and drug susceptibility profiles of Mycobacterium abscessus complex infection at a medical school in Thailand

Objectives This study investigated the differences in epidemiological and clinical data, and antimicrobial susceptibilities among different subspecies of Mycobacterium abscessus complex (MABSC) clinical isolates at a medical school in Thailand. Methods A total of 143 MABSC clinical isolates recovered from 74 patients were genotypically analyzed for erm(41), rrl, and rrs mutations, and antimicrobial susceptibilities were determined using a broth microdilution method. Patient characteristics and clinical outcomes were reviewed from the medical records. Results Seventy-four patients were infected with 28/74 (37.8%) M. abscessus subspecies abscessus (MAB), 43/74 (58.1%) M. abscessus subsp. massiliense (MMA), and 3/74 (4.1%) M. abscessus subsp. bolletii (MBO). The clinical findings and outcomes were generally indistinguishable between the three subspecies. All three subspecies of MABSC clinical isolates exhibited high resistance rates to ciprofloxacin, doxycycline, moxifloxacin, TMP/SMX, and tobramycin. MAB had the highest resistance rates to clarithromycin (27.8%, 20/72) and amikacin (6.9%, 5/72) compared to MBO and MMA, with p < 0.001 and p = 0.004, respectively. In addition, the rough morphotype was significantly associated with resistance to amikacin (8.9%, 5/56), clarithromycin (26.8%, 15/56), and imipenem (76.8%, 43/56) (p < 0.001), whereas the smooth morphotype was resistant to linezolid (57.1%, 48/84) (p = 0.002). In addition, T28 of erm(41), rrl (A2058C/G and A2059C/G), and rrs (A1408G) mutations were detected in 87.4% (125/143), 16.1% (23/143), and 9.1% (13/143) of MABSC isolates, respectively. Conclusions Three MABSC subspecies caused a variety of infections in patients with different underlying comorbidities. The drug susceptibility patterns of the recent circulating MABSC strains in Thailand were different among the three MABSC subspecies and two morphotypes. Supplementary Information The online version contains supplementary material available at 10.1186/s12941-023-00637-4.


Background
Mycobacterium abscessus complex (MABSC) is one of the most common rapidly growing mycobacteria (RGM) recovered from clinical specimens and is resistant to a variety of antimicrobial agents [1].In Thailand, RGM, especially MABSC, have emerged as a causative agent of recalcitrant infections in patients with preexisting chronic lung diseases and people living with human immunodeficiency virus (HIV), as reported worldwide.Moreover, MABSC is associated with localized or disseminated infection in patients with adult-onset antiinterferon-gamma (IFN-γ) autoantibody syndrome, which is prevalent in Southeast and East Asia [2,3].
This study aimed to investigate the differences in epidemiological and clinical data as well as comprehensive drug susceptibility patterns among the three subspecies of recent circulating MABSC clinical isolates in Thailand.In addition, the correlations between clinical and microbiological characteristics were determined.

Study population and Mycobacterium abscessus complex clinical isolates
MABSC clinical isolates were obtained from patients at King Chulalongkorn Memorial Hospital, Bangkok, Thailand, between January 2018 and December 2021.The patient data regarding epidemiology, clinical presentations, microbiology, and clinical outcomes were analyzed from the medical records (Table 1).Treatment outcomes of nontuberculous mycobacterial pulmonary disease (NTM-PD) were determined using the Nontuberculous Mycobacterium Network European Trials Group (NTM-NET) group consensus statement including culture conversion, microbiological cure, clinical cure, treatment failure, recurrence, relapse, reinfection, and death [15], which was modified for extrapulmonary infections but excluded culture conversion.A total of 232 MABSC clinical isolates were initially included in the study.However, 17 MABSC isolates were excluded due to bacterial contamination (n = 8), no growth after subculture (n = 3), and being mixed with other NTM (n = 5) or M. tuberculosis (n = 1).Finally, a total of 143 MABSC isolates, including single and sequential isolates from 74 patients diagnosed with MABSC infections according to clinical, radiological, and microbiological criteria (≥ 2 positive sputa or endotracheal aspirates with the same MABSC subspecies isolated, ≥ 1 positive BAL or sterile site) as described in the ATS/ERS/ESCMID/IDSA clinical practice guideline [16,17], were included in this study.The Institutional Review Board (IRB) of the Faculty of Medicine, Chulalongkorn University approved this study with IRB No. 541/63 (COA No. 967/2021).

Mycobacterial isolation and culture
MABSC clinical isolates were recovered from clinical specimens (Table 2) using the sodium hydroxide-N-acetyl-L-cysteine-sodium citrate method [18].The MABSC isolates were stored at -80 °C until use.The frozen stock of MABSC clinical isolates, M. abscessus ATCC19977, and M. peregrinum ATCC700686 were subcultured on Lowenstein-Jensen (LJ) medium at 35 °C for 3 to 5 days for the colony morphology study, drug susceptibility, and molecular testing.

Genotypic assays
The MABSC subspecies and mutations associated with drug resistance were analyzed using the GenoType NTM-DR assay (Hain Lifescience) [19] (n = 135) and gene sequencing (n = 8).For the GenoType NTM-DR assay, nucleic acid was extracted from mycobacterial colonies using the Genolyse kit, and further PCR amplification and reverse hybridization followed the manufacturer's protocol.
For next-generation sequencing, genomic DNA was extracted from culture materials using the cetyltrimethylammonium bromide (CTAB)-sodium chloride extraction method [20].Library preparations were constructed using the NEBNext® Ultra™DNA Library Prep Kit for Illu-mina® following the manufacturer's protocol.NGS was carried out using a 2 × 150 paired-end (PE) configuration by the NovaSeq instrument.FastQC and Trimmomatic V0.32 were used for quality checking and trimming the sequencing reads.Adapters and low-quality sequences (Phred score < 30) were removed.The filtered reads were mapped to the rrl, erm(41), and rrs genes from the reference genome of M. abscessus ATCC19977 (Accession no.: CU458896).Single nucleotide mutations were identified using VarScan with a minimum variant allele frequency criterion of 0.01.Drug resistance prediction and variant identification were performed using the pipelines from a previous report [12].

Drug susceptibility testing
The MICs of 16 antimicrobial agents were examined using Sensititre™ Myco RAPMYCOI (Thermo Fisher Scientific) and a broth microdilution method that were performed and interpreted according to the Clinical and Laboratory Standards Institute (CLSI) guidelines M24, Ed3 [21] and M24S [22] (Table S1).For drugs without addressed breakpoints, the MIC values were reported and calculated for the MIC 50 and MIC 90 .Clofazimine (Sigma-Aldrich) was dissolved in dimethyl sulfoxide (DMSO) and dispensed into cation-adjusted Mueller-Hinton broth (CAMHB) using 2-fold serial dilutions (Table S1).The MIC plates were incubated at 30 ºC for 3-5 days until a positive control had sufficient growth and reexamined to observe an ICR on day 14 of incubation [21].The colonies of M. abscessus ATCC19977 and M. peregrinum ATCC700686 were included as quality controls.

Statistical analysis
The results are expressed as the median and interquartile range (IQR) for age, and as the number and percentage (%) for categorical variables.Data were compared using Pearson's X 2 or Fisher's exact test for categorical variables.A two-sided p-value of < 0.05 was considered statistically significant for the analyses.Evaluation of the positive predictive value (PPV) and negative predictive value (NPV) of molecular detection of T28 was performed using a 95% confidence interval (95% CI) (p < 0.05).All analyses were performed using IBM SPSS version 28 statistical software.

Patient characteristics and clinical outcomes
The clinical characteristics and MABSC subspecies of 74 patients with MABSC infection are shown in Moreover, medical procedures or exposure to natural water sources mediated SSTIs (n = 6) and lymphadenitis (n = 1) in patients without specific underlying diseases or immunocompromised status.Procedures related to SSTIs include platelet-rich plasma injection into the knee joint, multiple drug injections into the groin, unsterile acupuncture, cosmetic surgery, and postpsoas abscess surgery.One patient with dental implantation developed lymphadenitis.In addition, half of the patients (4/8, 50%) with disseminated diseases had adult-onset IFN-γ autoantibody syndrome.

Genotypes associated with clarithromycin and amikacin resistance
Mutations in the erm(41), rrl, and rrs genes were investigated (Table 4 and Figure S2).T28 of erm(41) was detected in 87.4% (125/143) of MABSC isolates which were MAB (43.2%, 54/125), MBO (5.6%, 7/125), and MMA (51.2%, 64/125).A phenotypic ICR was not observed for three of the MAB and all MMA that had T28.C28 was detected in 25% (18/72) of the MAB that was susceptible to clarithromycin.None of C28 was observed in erm(41) of MBO and MMA.For MAB and MBO, the PPV and NPV of molecular detection of T28 conferring an ICR were 95.1% (95% CI, 84.1-98.6%)and 100.0%, respectively.Acquired clarithromycin resistance conferred by rrl mutations at positions 2058 and/or 2059 (A2058C, A2058G, A2059C, and A2059G) was detected in 16 MAB and one MBO isolate that harbor rrl mutations but were susceptible to clarithromycin on day 3 of incubation.However, the MBO with A2058G and A2059G became clarithromycin resistant on day 14 of incubation.For amikacin resistance, the rrs mutation (A1408G) was detected in 13 MAB isolates recovered from one patient at different time points.These isolates had variable amikacin susceptibilities, which were susceptible (n = 1), intermediate (n = 7), and resistant (n = 5).Overall agreement between the GenoType NTM-DR assay (rrl and rrs) and a broth microdilution method (intermediate and resistant results considered as R) of clarithromycin and amikacin susceptibility was 98.6% (141/143).

Discussion
In this study, the MABSC clinical isolates were identified to the subspecies level, and their susceptibilities to 16 antimicrobial agents were determined.In addition, the clinical characteristics of the patients were investigated.In Thailand, this information is still limited, as only a few studies have reported the prevalence and drug susceptibility profiles of MABSC [7,23].MABSC is the most common RGM isolated from clinical specimens at our hospital.True MABSC pulmonary infections or NTM-PD can be diagnosed using clinical, radiological, and microbiological criteria [16,17] and differentiated from specimen contamination [24] or colonization in the  respiratory tract.For the diagnosis of extrapulmonary NTM disease, a single positive specimen from a sterile site, body fluid, or tissue biopsy is sufficient [25].These criteria were applied for the diagnosis of MABSC infection in this study.Most patients with MABSC infections or colonization at our hospital had comorbidities such as immunocompromised status and bronchiectasis, which are associated with NTM infections [13,26].In addition, the rate of infections related to medical procedures has increased.
To treat NTM-PD infections, a combination of antimicrobial drugs including oral (azithromycin or clarithromycin, clofazimine, linezolid) and intravenous drugs (amikacin, imipenem or cefoxitin, tigecycline), is recommended and applied for extrapulmonary diseases.However, the optimal duration of treatment has not been universally established [16,17].Drug susceptibilities should be determined for clinically significant MABSC isolates.However, MABSC drug susceptibilities may not be the only key factor that influences the clinical outcomes of infected patients.In this study, some patients with favorable outcomes were infected by ICR-MABSC but responded to macrolide-based regimens.However, certain patients who were infected with clarithromycin/ amikacin-susceptible MABSC strains experienced treatment failure or death (Table S2).Therefore, other factors, such as drugs used in the regimen, duration of use, severity of the disease, or patient comorbidities [27], could impact the clinical outcomes for the individual patient.In this study, the MABSC subspecies were not statistically associated with clinical outcomes.This could be due to a low number of patients (n = 31) from loss to follow-up, one of the limitations of this study.
The major MABSC subspecies that caused infections and were recovered from clinical specimens in our hospital were MAB followed by MMA.MBO was a relatively rare pathogen, consistent with findings in other countries.In addition, the drug susceptibility patterns of each MABSC subspecies and clinical isolates from various regions and countries can be different [5][6][7][8][9].Therefore, the MABSC subspecies could be one of the important factors in the selection of an optimal and effective therapeutic regimen for patient management.In this study, MAB was associated with clarithromycin and amikacin resistance.However, the subspeciation analysis of MABSC is not available in most routine diagnostic laboratories.Therefore, the drug susceptibility patterns of the clinical isolates of MABSC without subspeciation are shown in Table 3. Amikacin was still the most effective drug against MABSC, and most MABSC strains were highly resistant to several drugs, including clarithromycin (day 14 of incubation), ciprofloxacin, doxycycline, moxifloxacin, TMP/SMX, and tobramycin, as previously reported [5,6,9].A high intermediate rate was observed for imipenem and cefoxitin, consistent with the findings from previous studies [8,9,28].For drugs without addressed breakpoints, clofazimine and tigecycline could be potential drugs to effectively treat MABSC infections due to their low MIC 50 and MIC 90 values, consistent with a study from the United States [8].On the other hand, amoxicillin/clavulanic acid, cefepime, ceftriaxone, and minocycline may be ineffective in inhibiting the pathogens, so they have not been recommended for MABSC treatment [16,17].
At present, molecular testing has become an important tool for rapidly detecting the gene mutations associated with clarithromycin and amikacin resistance.In this study, discrepant results between phenotypic and genotypic (GenoType NTM-DR) drug susceptibility results were observed, as previously reported [7,29].This indicated that clarithromycin and amikacin resistance might involve other mechanisms apart from those mutations detected in the erm(41), rrl, and rrs genes.In addition, despite the presence of T28, three MAB isolates did not express ICR.These findings suggest a point mutation, especially nonsynonymous or nonsense mutation, leading to a truncated Erm protein [30].Therefore, sequencing analysis of erm(41) should be performed.
For colony morphology, MABSC can switch from a smooth to a rough morphotype, which has been shown to be more invasive and associated with poor clinical outcomes [13,14].Our findings show that the rough morphotype was significantly associated with amikacin, clarithromycin, and imipenem resistance, while the smooth morphotype was associated with linezolid resistance.However, the correlations between MABSC morphotype and drug susceptibility are still limited and lacking clarity.A French study reported high MICs of imipenem and cefoxitin in the rough morphotype [31], which was consistent with our findings for imipenem.However, previous studies showed that the MABSC morphotype did not significantly impact antimicrobial susceptibility [32,33].Therefore, future genetic analysis of glycolipid (GLP) synthesis or transport genes [34] of MABSC clinical isolates should be performed to confirm their true morphotypes and should study their correlation with drug susceptibility or clinical outcomes.
This study has many strengths, as it is the first study to investigate the clinical and microbiological associations of patients with MABSC infections in Thailand.However, there are some limitations in this study.First, this study included MABSC isolates from both treated and nontreated patients, which could affect the drug susceptibilities from the selection of resistant strains in treated patients.Second, the number of patients and MBO isolates was low.Third, the discrepant results between the phenotypic and genotypic susceptibilities of clarithromycin and amikacin were not further investigated using other methods, such as gene sequencing analysis.Last, this was a single-center study, and most patients came from Central Thailand.None of the patients living with HIV who were particularly vulnerable to NTM infection were recruited for this study.Therefore, these limitations could impact the statistical analysis and might not represent all MABSC isolates from the Thai population and patients living with HIV.Future studies will be conducted with additional patients and MABSC isolates, as well as genetic analysis of the genes associated with MABSC morphotypes and drug susceptibility.

Conclusions
This study demonstrated the differences in the clinical and microbiological data of patients with MABSC infections caused by the three different subspecies and two morphotypes of MABSC.The findings of this study could be useful for the selection of antimicrobial regimens and the treatment of patients with MABSC infections.

Table 5
Summary of antimicrobial susceptibility of M. abscessus complex (MABSC) based on morphotypes