Legionella is not the only bug in hospital water capable of threatening patients with deadly infections. Hospital Infection Control & Prevention recently talked to two investigators in the CDC’s vaunted Epidemic Intelligence Service. Investigating two separate waterborne outbreaks, these medical detectives offered some sage advice in the form of an observation and a caveat.
“I think in older structures with older hospital plumbing there is a potential for biofilms to develop in the piping,” says Mark Weng, MD, EIS officer. “There are a lot of older hospitals out there and the risk mitigation needs to span across the whole spectrum of water delivery.”
Though few would argue with that general observation, the caveat comes from another EIS investigator, Kimberly A. Skrobarcek, MD.
“This isn’t just necessarily because the plumbing is old,” she says. “This can happen with any kind of system that is complex in nature, in [buildings with many] floors and any kind of dead spaces that can lead to biofilm formation.”
Water was the common denominator in both outbreaks, but the two investigated distinctly different situations, one dealing with a neonatal ICU and the other with infections following reconstructive breast surgery.
The NICU outbreak was so devastating that two of nine infected infants died and the state health department shut down the unit at a Maryland hospital until corrective measures were in place. The culprit was a common water bug, Pseudomonas aeruginosa, which is the leading cause of serious healthcare-associated infections in NICUs, Weng and colleagues reported at the annual CDC EIS meeting in Atlanta.1
There was a high burden of P. aeruginosa in the facility water, so the CDC investigators focused on possible transmission pathways and control measures.
“The difficulty in our situation was that during this investigation, the NICU had been shut down, so it was hard for us to [observe] hand hygiene and [other infection control measures.],” he said.
From staff interviews, observations as available, and environmental sampling of the NICU, the investigators concluded that the likely factors included inadequate hand hygiene and cleaning of sinks, equipment, and breast milk collection supplies. P. aeruginosa was isolated from stored breastmilk, tap water, a staff bathroom, and breakroom sinks. The isolates from two patients and two sinks in the NICU were indistinguishable by pulsed field gel electrophoresis (PFGE) analysis. A breast milk isolate was closely related to a Pseudomonas isolate from a third patient.
Transmission from contaminated hospital water sources to patients could have occurred via several pathways, they concluded. CDC recommendations for the NICU included improving hand hygiene:
- measures to decrease P. aeruginosa in the hospital water system;
- avoidance of using tap water for cleaning equipment;
- revised cleaning procedures for sinks, patient care, and breast-feeding equipment in the NICU;
- PA surveillance to evaluate the effectiveness of these control measures.
“These were the recommendations we handed over to the state department as they worked to come up with a plan, a path to reopening the NICU,” Weng said. “Since then, I have not heard of any additional cases. These were all pretty sick preterm babies as one might expect from a NICU. I think what it comes down to is making the team really pay attention in the healthcare setting to how water is used in the NICU. I think that was the common denominator. We knew the bacteria was in the water. It kind of makes sense that we ended up getting evidence for multiple potential pathways that were supported through culture and interviews with facility staff.”
SSIs Linked to OR Water
In the other outbreak, water contaminated with nontuberculous mycobacteria (NTM) was implicated in an outbreak among patients undergoing reconstructive breast plastic surgery (BPS) in a South Carolina hospital.2 Skrobarcek’s EIS team conducted a case-control study of patients who underwent BPS from January 2014–October 2016. Case-patients had surgical site infections (SSIs) linked to NTM.
They identified 17 case-patients and 51 controls. Three NTM species were identified among case-patients, with none in controls. Statistically significant risk factors included various surgery types, surgery duration (median of 5.1 hours for case-patients; 2.5 hours for controls), operating room staff number (median of 9 for case-patients; 7 for controls), and overnight admission (16/17 case-patients, 26/51 controls).
“Technically, statistically speaking, the key factor was actually the number of staff within the OR,” Skrobarcek says. “However, that is very highly correlated with a long surgery duration. That makes sense because if you have a long surgery, [healthcare workers] may have to have relief staff come in. If a surgery goes five hours, someone has to go out and take a break at some point. The more people, the more opportunities there are for potential deficiencies in infection control practices.”
Method of Introduction
Still, that raises the question of how the waterborne pathogen was introduced.
“If someone is coming in scrubbed up [they may have] wet hands in the middle of the surgery when the operating site is open,” she explains. “That is also going to create higher risk for any kind of potential changes in air flow and other things that can affect surgical site infections. That would be decreased if everyone is in the OR at the beginning of the surgery and there isn’t a lot of staff coming in and out, with the door opening that can change air flow.”
In hospital water used for handwashing, total bacterial counts were 103–106 colony-forming units per milliliter, exceeding the Environmental Protection Agency’s drinking water standards. Four NTM species grew in 25/37 environmental samples.
“We believe that the source was likely the water in the hospital plumbing system — the distribution system,” Skrobarcek says. “We weren’t able to specifically link anything with an identified patient with the exact genetic typing [matching] what was found in the environment. But this is a common infection that normally comes from the water. There are biofilms and different species living within those biofilms. We were only able to sample a small amount [of water] so it wasn’t surprising that we didn’t find an exact match. But given that the numbers of bacteria were so high within the water throughout the hospital, and within the operating floor and the sinks that were used for scrubbing, that is the likely source.”
The hospital increased the frequency of water chlorination and had filters placed on all faucets and plumbing outlets on the OR floor, she notes.
- Weng MK, Brooks R, Glowicz J, et al. Cluster of Pseudomonas Infections among Neonatal Intensive Care Unit Patients — Maryland. 2016 CDC EIS Conference. Atlanta, May 2-5, 2017.
- Skrobarcek KA, Vasquez A, Margus C, et al. Nontuberculous Mycobacteria Infections Among Breast Plastic Surgery Patients — Hospital A, South Carolina, 2014–2016. CDC EIS Conference. Atlanta, May 2-5, 2017.