To the Editor—Mycobacterium chimaera (M. chimaera) is an opportunistic environmental mycobacterium belonging to the M. avium–M. intracellulare complex. Transmission of M. chimaera from contaminated heater-cooler unit (HCU) water tanks to patients is a risk during open-heart surgery. ^{Reference Sax, Bloemberg and Hasse1} Specifically, investigations have revealed that the contaminated HCU devices (Stockert 3T) came from one particular manufacturing facility, LivaNova in Munich, Germany, and were a likely source for cardiothoracic surgery–related severe M. chimaera infections that occurred in Switzerland, Germany, The Netherlands, and the United Kingdom. ^{Reference Haller, Höller and Jacobshagen2} The importance of HCUs lies in their ability to regulate the body temperature of a patient during cardiac surgery. However, evidence suggests that the airborne transmission of aerosolized bacteria from the water tanks was responsible for these infections. ^{Reference Haller, Höller and Jacobshagen2} Currently, the extent of global outbreaks is unknown, but the burden of invasive M. chimaera was estimated to be 156–282 cases per year in 10 countries where most cardiac-valve replacements are performed. ^{Reference Sommerstein, Hasse, Marschall, Sax, Genoni, Schlegel, Widmer and Chimaera Taskforce3} We investigated the possible presence of M. chimaera in HCUs in the Cardiothoracic Department, Latvian Centre of Cardiology of Pauls Stradins Clinical University Hospital, Latvia. The water of HCUs was sampled before they underwent the routine disinfection procedure in May, 2017. Samples were cultured on Bactec MGIT 960 system (Becton Dickinson, Heidelberg, Germany). The growth of mycobacteria was observed in 2 samples. DNA was isolated and the M. chimaera was identified using a GenoType NTM-DR version 1.0 kit (Hain Lifescience, Nehren, Germany). Both isolates, designated LV-2017-1-HCU and LV-2017-2-HCU, were subjected to whole-genome sequencing with 150× coverage on an Ion Proton System (Thermo Fisher Scientific, Waltham, MA). For the bioinformatics analysis, reads in samples produced in this study were aligned to the M. chimaera reference genome JCM_14737 (ENA accession no. PRJNA324238) using the bwa MEM algorithm. Variants were called and marked based on the published criteria ^{Reference Svensson, Jensen, Rasmussen, Folkvardsen, Norman and Lillebaek4} using BCFtools. Sequences of both isolates were compared to strains ZUERICH-1 (DSM 101591) and ZUERICH-2 (DSM 101592), representatives of the major M. chimaera groups 1 and 2. ^{Reference van Ingen, Kohl and Kranzer5}

For the phylogenetic analysis, raw sequencing reads of the isolates LV-2017-1-HCU and LV-2017-2-HCU were mapped against the M. chimaera type strain FI-01069 (DSM 44623) ^{Reference Tortoli, Rindi and Garcia6} and compared with strain ZUERICH-1 and 127 other publicly available data sets of M. chimaera isolates belonging to subgroup 1.1 and previously collected in Denmark, Germany, The Netherlands, the United Kindgom, Ireland, the United States, Australia, and New Zealand (available in the European Nucleotide Archive; http://www.ebi.ac.uk/ena). ^{Reference Svensson, Jensen, Rasmussen, Folkvardsen, Norman and Lillebaek4,Reference van Ingen, Kohl and Kranzer5,Reference Williamson, Howden and Stinear7} The analysis was performed using CSI Phylogeny 1.4 software, Center of Genomic Epidemiology (https://cge.cbs.dtu.dk/services/CSIPhylogeny/), with default settings but including the reference in the final tree. ^{Reference Kaas, Leekitcharoenphon, Aarestrup and Lund8} Molecular phylogenetic analysis was performed using the maximum likelihood method. The phylogenetic tree was built from 176 SNP positions of 130 M. chimaera isolates and was visualized in MEGA6.

Results

In total, 2,555,838 and 1,544,268 sequencing reads were generated for samples LV-2017-1-HCU and LV-2017-2-HCU, respectively. The sequencing reads were deposited in the European Nucleotide Archive (ENA) under study numbers ERS3734298 and ERS3734299.

Molecular phylogenetic analysis revealed that both Latvian HCU‘s samples were closely related to the strain ZUERICH-1 (Fig. 1A). Specific SNP signatures ^{Reference van Ingen, Kohl and Kranzer5} for subgroup 1.1. (substitutions of guanine (G) by adenine (A) at positions 113,518 and 209,278 of the DSM 44623T genome—GenBank accession no. LQOO00000000) were found in both Latvian isolates. This result suggested that they belong to the subgroup 1.1 which contains most M. chimaera isolates from water systems or exhaust air of LivaNova HCUs in clinical use, isolates from HCUs sampled at the LivaNova production site, and isolates from related patients in different countries. ^{Reference van Ingen, Kohl and Kranzer5}

Fig. 1. (A) Molecular phylogenetic analysis of the isolates in the present study using the maximum likelihood method. Both samples were closely related to the strain ZUERICH-1. (B) Maximum likelihood phylogenetic tree built from 176 SNP positions of 130 M. chimaera isolates mapped to the genome of M. chimaera type strain FI-01069. The 2 isolates in the present study are indicated by black circles. Codes of the samples belonging to group 1.1. correspond to those used by van Ingen et al (2017). Note. AU, Australia; D, Germany; NL, The Netherlands; NZ, New Zealand; LV, Latvia; UK, United Kingdom; CH, Switzerland; and DK, Denmark.

During the phylogenetic analysis of the isolates LV-2017-1-HCU and LV-2017-2-HCU, the sequencing coverage was 92.3% and 78% of the reference genome, respectively. The percentage of the reference genome covered by all isolates was 43.8%. As expected, the isolates within the subgroup 1.1 showed little diversity, with a median pairwise distance of 2 SNPs (range, 0–40). ^{Reference van Ingen, Kohl and Kranzer5} A median pairwise distance for the Latvian LV-2017-1-HCU and LV-2017-2-HCU isolates were 32 (range, 31–40) and 4 (range, 3–12) SNPs, respectively. The results showed that M. chimaera sequences from HCU in Latvia genetically clustered with the isolates from HCUs in The Netherlands, the United States, and Germany (samples 110, 128, 180, 187, 2015-22-15-01), as well as isolates from patients in The Netherlands (sample 198) and USA (samples 2015-22-63, 2015-22-79, 2015-22-80) (Fig. 1B).