Water System Monitoring Retards Legionnaires' Disease

PITTSBURGH, Aug. 23 -- Hospital-acquired Legionnaires' disease can be predicted by monitoring water systems for high levels of the bacteria that cause it, researchers here said.

In a study testing the efficacy of routine environmental testing, clinical legionellosis cases were found only in hospitals extensively colonized with Legionella pneumophila serogroup 1 in more than 30% of potable water taps.

These findings from a prospective surveillance study of 20 hospitals in 13 states argue for routine water system monitoring, wrote Victor L. Yu, M.D., of the University of Pittsburgh, and colleagues, in the July issue of the journal Infection Control and Hospital Epidemiology.

Although hospital-acquired Legionella pneumonia accounts for up to 45% of all cases and has twice the fatality rate of community-acquired cases (28% versus 14%), the CDC recommends clinical surveillance without routine surveillance of water systems except in transplant units.

The researchers collaborated with the Allegheny County Health Department on an alternative prevention strategy that led to the virtual disappearance of this infection in Pittsburgh, they said.

Their strategy called for routine testing of hospital water for Legionella species. When found, physicians were to routinely request Legionella diagnostic tests for patients with hospital-acquired pneumonia even in the absence of known cases of Legionnaires' disease. Hospitals found to be highly colonized could disinfect their water systems.

"A powerful argument against routine environmental surveillance for Legionella has been the supposedly ill-defined relationship between the presence of Legionella species in hospital water systems and the risk for legionellosis," Dr. Yu and colleagues wrote.

To provide such evidence, the researchers conducted a study monitoring Legionella colonization in the water systems of 20 U.S. hospitals that had never had a known outbreak of hospital-acquired Legionella pneumonia.

Each hospital collected swabs and water samples from faucets around the hospital as well as from hot water tanks five times from 2000 to 2002. These were sent to a central laboratory for testing along with samples from clinical surveillance.

Only 25% of hospitals tested negative for Legionella species at all time points. The majority of hospitals (70%) had water systems colonized with L. pneumophila, and 45% had L. anisa.

Six of these 14 hospitals (43%) had high levels of colonization with L. pneumophila, defined as more than 30% of water outlets testing positive.

Overall, 633 patients at 12 hospitals were evaluated for Legionella pneumonia using sputum and urine samples collected during clinical surveillance.

Of these 12 hospitals, 10 had L. pneumophila in the water system and four identified cases of legionellosis (40%). All four were among the five highly colonized hospitals.

At these four hospitals, six patients total were diagnosed with Legionella pneumonia, all in a single monoclonal antibody subtype group that matched the bacteria in the water system of their respective hospitals. Half the patients died.

One hospital was highly colonized (83% of tested taps) with L. pneumophila serogroup 1 bacteria but no reported no clinical cases, yielding a less than 100% positive predictive value from routine environmental monitoring.

However, it may not be evidence against the utility of routine water system screening, the researchers cautioned.

This hospital had a strain known to be less virulent. Furthermore, the municipal water supply for the hospital switched from chlorine to monochloramine-- an active Legionella disinfectant -- for water treatment during the study, after which the hospital's colonization rate dropped to 0%.

"The lack of identified cases may have been influenced by changes in the degree of colonization of the hospital water system caused by the change in municipal water system," they noted.

The study puts to rest the argument that Legionella in water systems is not well correlated to risk for legionellosis, they concluded.

The results provide the strongest evidence so far that "determination of the status of Legionella colonization is useful in evaluating the risk for hospital-acquired Legionella pneumonia," and "that environmental surveillance for Legionella should be part of a proactive strategy for prevention," they wrote.