Epidemiologic Investigation

After they identified an initial case of M. abscessus infection in May 2024, hospital IP staff initiated retroactive case finding. The hospital had a previous cluster of 3 cases of M. abscessus (2 positive CSF cultures, 1 positive sputum culture) in 2021, shortly after occupancy of a newly constructed building. Preoccupancy water quality testing was within acceptable ranges but did not include specific testing for M. abscessus. With guidance from MDH and CDC, water samples from the 2021 outbreak were tested at CDC, where NTM was identified, but not M. abscessus.

In the cluster reported in 2024, a total of 10 patients had M. abscessus identified from clinical cultures during March 28, 2023–June 5, 2024. Hospital IP staff identified 1 additional patient in September 2024 and reported that case to MDH, bringing the total to 11 patients. The hospital implemented mitigation strategies after the first 2 cases were identified in 2024, including hospitalwide water flushing, increased point-of-entry supplemental disinfectant (monochloramine) levels, restrictions on use of tap water and ice machines, and staff education and monitoring of sink safety. The hospital also installed PALL point-of-use filters (https://www.pall.com) in sinks and medication rooms in high-risk patient care areas, including intensive care unit patient rooms, operating room (OR) scrub sinks, and cancer centers. The hospital water management plan guided some of the mitigation strategies and was ultimately updated to include more frequent building flushes, residual disinfectant testing, and inclusion of M. abscessus in water microbial testing. All those measures have been shown to reduce risk for patient exposure to contaminated water sources (6). Investigators also collected information regarding timing of clinical culture collection, patient location within the facility, and clinical manifestations (including patient procedures).

Laboratory Testing

Before initiating mitigation strategies, the hospital water management team completed environmental sampling to attempt to identify potential sources of M. abscessus in the hospital. Water sampling included bulk water samples (1 L, mixed hot and cold) and swab specimens from faucets, sinks, and ice machines; testing was conducted by an accredited commercial environmental microbiology laboratory. The team chose locations for water samples on the basis of patient location in the hospital, high-risk patient care areas, and sources implicated in reports of other M. abscessus outbreaks (5). In addition to cultures, supplemental disinfectant levels were measured for faucets inside patient rooms, medication rooms, ice machine water, and OR scrub sinks. Heater-cooler machines have been implicated in other outbreaks of NTM (8–9), and a bulk water sample from a heater-cooler machine (CardioQuip, https://www.cardioquip.com) used during 1 patient’s hospitalization was included; that case-patient was the only one with heater-cooler machine exposure. After mitigation methods were implemented, resampling was conducted to determine effectiveness. For point-of-use filters, sampling was done before and after filter installation.

Whole-Genome Sequencing

Whole-genome sequencing (WGS) was performed at the Wadsworth Laboratory in New York for both clinical and environmental isolates using the NextSeq platform (Illumina, https://www.illumina.com) (10–11). De novo genome assemblies were performed using the Shovill pipeline version 1.1.0, default parameters (https://github.com/tseemann/shovill). No postmitigation isolates were sequenced.

Epidemiologic Investigation

Among 11 patients identified, median age was 40 (range 24–76) years. Five patients underwent surgical procedures; 2 had procedures in OR A, 2 in OR B, and 1 in OR C. No patients had the same procedure; however, the 2 patients in OR A each had cardiac surgery. One patient had a heater-cooler machine used during the cardiac procedure. Two rooms housed multiple affected patients; however, their admissions did not overlap in time. Otherwise, we identified no shared patient ward or unit locations. Six (55%) of the 11 patients were treated with antimicrobial therapy as directed by infectious disease physicians, after positive culture results. One (9%) patient died within 60 days of the first positive culture; that death was not attributed to M. abscessus infection.

Laboratory Testing

Among culture samples from 11 patients, 6 (55%) were from sputum samples, 4 (36%) from blood samples, and 1 (9%) from an abdominal abscess. Before initiation of mitigation strategies, the IP team collected 47 environmental samples from 19 locations throughout the hospital. Nineteen were bulk water samples, and 28 were swab specimens from faucet and sink drains, including ice machines. Testing revealed M. abscessus in the water system at different levels throughout the hospital building, including in ice machines. Among all samples, 18 (38%) grew M. abscessus and 15 (32%) had acid-fast bacilli but not M. abscessus. For samples that grew M. abscessus, density varied: 10–48,000 CFU/swab, or <1–39 CFU/mL for bulk water samples. The sample from the CardioQuip heater-cooler machine did not yield any bacteria. An investigation into various uses of tap water and ice within the hospital did not identify any obvious source of specimen contamination from tap water or ice during specimen collection or processing. However, investigation of case 11, which was identified after mitigation efforts began, revealed ice from restricted ice machines was used to store saline syringes that were used during bronchoalveolar lavage.

After review of clinical and epidemiologic data, we determined that 6 patients had clinical infections and 5 represented likely colonization or specimen contamination. All 5 of those cases involved respiratory tract cultures (Table), thought to be colonization or the result of specimen contamination either because the specimen collection was completed in the emergency department before admission and patients had no previous exposure to the facility or because they had respiratory illness at admission.

The hospital installed point-of-use filters in several locations, guided by the epidemiologic investigation, including high-risk patient areas, ORs, and intensive care units. Water sampling after installation of point-of-use filters showed no bacterial growth. Repeat sampling of sink and faucet sites after flushing and increasing point of entry supplemental disinfectant revealed optimal residual disinfectant but continued detection of M. abscessus to different degrees.

WGS

All 18 environmental samples that grew M. abscessus and 10 available clinical isolates from 10 distinct patients were sent to the Wadsworth Laboratory for WGS. Sequences reported in this article were submitted to the National Center for Biotechnology Information BioProject Database (BioProject accession no. PRJNA1403517). Among the 28 samples sequenced, we identified 5 subclusters, consisting of 10 patient isolates and 17 environmental isolates. Maximum whole-genome single-nucleotide polymorphism (SNP) distance of the core genome across the 28 isolates was 98,252. Within each subcluster, associated subcluster isolates were within 0–24 SNPs, indicating a high degree of relatedness (12–13). Four subclusters each included both patient isolates and environmental isolates (0–15 SNPs), and 1 included only isolates from patients (20–24 SNPs). One patient isolate identically matched an environmental sample from the postanesthesia care unit (PACU) ice machine (0 SNP difference). The patient had been in the PACU after a procedure for <24 hours. Two other patient isolates showed close association with PACU environmental samples. One showed close association (5 SNPs) to a PACU sink water sample; that patient had 3 operations during hospital admission. The other patient, whose isolate was closely associated (11 SNPs) with a PACU sink faucet swab, had not spent any time in the PACU and was ultimately determined to be a likely specimen contamination because the sample was collected in the emergency department before admission. Two patient isolates were not found to be closely associated with any of the environmental samples but did cluster closely (20–24 SNPs) with a patient isolate from the facility in 2021. However, we could not determine any additional epidemiologic link between the 2 isolates from this cluster and the 2021 isolate. One environmental isolate was not closely related (>80 SNPs) to any other environmental or patient isolates.