Trouble at the Pittsburgh VA

The Pittsburgh VA has had a recent outbreak of nosocomial legionellosis. Based upon press reports, there have been at least five confirmed cases with one death. There are also references in the media to 24 additional cases reported beginning in January of 2011 (8 of which were felt to be community onset, and 16 unknown onset—likely in the “possible nosocomial” category). CDC personnel traveled there to assist in the investigation, and full details are likely to emerge eventually. 

There is a highly charged back-story here. Reading the press reports, one immediately notices some very pointed critical comments from two internationally recognized Legionella experts. In 2006, there was an acrimonious split between the Pittsburgh VA and Victor Yu and Janet Stout (you can read their side of that story here--click through to some of the e-mails to get a flavor for just how nasty this episode was). This public dispute ended with the destruction of a massive organism bank that included thousands of Legionella strains, leading to a letter and petition being published in Clinical Infectious Diseases. And now that the Pittsburgh VA is knee-deep in Legionella without them, these two colleagues are not holding back. 

We’ve had our own history with Legionella at Iowa, and this outbreak in Pittsburgh may reiterate several important lessons we learned long ago: (1) never assume that Legionella has been eradicated from a water system, it is only suppressed to levels that cannot be detected, and will re-emerge when given the opportunity, (2) copper-silver and chlorine dioxide suppression systems work, but only if they are carefully maintained and levels of the active agent(s) are monitored (not just centrally, but also at distal sites), (3) installing a suppression system doesn’t obviate the need for regular water testing in facility that has had nosocomial cases of legionellosis, (4) there is no way to definitively determine the source of a legionellosis case if cultures are not performed on patient samples (the urinary antigen test doesn’t provide an organism for typing), and (5) all legionellosis cases that fall into the “possible nosocomial” category (onset between 2-10 days after admission) should be assumed to be nosocomial, and an appropriate investigation begun to assess for a source in the hospital water supply.