Genomic Epidemiology of Mycobacterium abscessus on the Island of Montréal Is Not Suggestive of Health Care-Associated Person-to-Person Transmission

Abstract

Background: Mycobacterium abscessus complex (MABC), an opportunistic nontuberculous mycobacteria, can lead to poor clinical outcomes in pulmonary infections. Conflicting data exist on person-to-person transmission of MABC within and across health care facilities. To investigate further, a comprehensive retrospective study across 5 health care institutions on the Island of Montréal was undertaken.

Methods: We analyzed the genomes of 221 MABC isolates obtained from 115 individuals (2010-2018) to identify possible links. Genetic similarity, defined as ≤25 single-nucleotide polymorphisms (SNPs), was investigated through a blinded epidemiological inquiry.

Results: Bioinformatics analyses identified 28 sequence types, including globally observed dominant circulating clones (DCCs). Further analysis revealed 210 isolate pairs within the SNP threshold. Among these pairs, there was 1 possible laboratory contamination where isolates from different patients processed in the same laboratory differed by only 2 SNPs. There were 37 isolate pairs from patients who had provided specimens from the same hospital; however, epidemiological analysis found no evidence of health care-associated person-to-person transmission between these patients. Additionally, pangenome analysis showed higher discriminatory power than core genome analysis for examining genomic similarity.

Conclusions: Genomics alone is insufficient to establish MABC transmission, particularly considering the genetic similarity and wide distribution of DCCs, although pangenome analysis has the potential to add further insight. Our findings indicate that MABC infections in Montréal are unlikely attributable to health care-associated person-to-person transmission.

Keywords: Mycobacterium abscessus; antimicrobial resistance; cystic fibrosis; genomic epidemiology; whole-genome sequencing.

Figure 1.

Summary of Mycobacterium abscessus subspecies isolates collected in Montréal and Québec, Canada. Zero single-nucleotide polymorphism (SNP) distances on the y-axis were omitted due to the logarithmic scale. A, Subspecies assignment of genomes by collection date from the 5 Montréal hospitals. B, Pairwise SNP distances of genomes sequenced from the same person grouped by subspecies. C, Pairwise SNP distances of genomes sequenced from different individuals stratified by subspecies. The boxplots display the median pairwise SNP distance (central line), the interquartile range (IQR, represented by the box), and the whiskers extend to 1.5 times the IQR from the box. Points beyond the whiskers represent outliers.

Figure 2.

Core genome maximum phylogenetic tree of Mycobacterium abscessus subspecies midpoint rooted using an ultrafast bootstrap replicates of 10 000 and the TPM3u + F + ASC + R4 best fit model. Phylogeny was inferred from recombinant masked variant positions. Tree tips are labeled with case identity numbers and tip colors denote CF status. PubMLST STs are shown in colored strips labelled with the ST. Grey strips are regions where a ST could not be assigned by PubMLST. A, M. abscessus subsp. abscessus (n = 145) inferred phylogenetic tree from 42 663 single-nucleotide polymorphisms. B, M. abscessus subsp. massiliense (n = 58) inferred phylogenetic tree from 95 529 single-nucleotide polymorphisms. Abbreviations: CF, cystic fibrosis; DCC, dominant circulating clone; M, Island of Montréal hospitals; QC, other Québec hospitals; ST, sequence type.

Figure 3.

Tanglegram of the core genome 95 529 single-nucleotide polymorphism-based phylogenetic tree with the accessory genome phylogenetic tree inferred from pangenome analyses using presence and absence of 4000 accessory genes of Mycobacterium massiliense (n = 58). Tree tips are colored red for individuals with cystic fibrosis and black for those without cystic fibrosis. Lines between trees are colored based on isolates from the same individual. Robinson-Foulds distance between both trees is estimated at 0.37. Abbreviations: M, Island of Montréal hospitals; QC, other Québec hospitals.