[Prevention of fungal infections related to the water supply in French hospitals: proposal for standardization of methods].

OBJECTIVES

The aims of this study were to assess the risk of fungal infections related to the water supply in several hospitals and to clarify the appropriate methodology in order to standardize the technical conditions of the controls and develop guidelines. It was conducted in 10 university hospital centers across the country from February 2004 through March 2005.

METHOD: A preliminary study allowed us to optimize the mycological analysis. The study was conducted under the same conditions as for bacteriological controls: water filtration through a cellulose acetate membrane cultured on agar. Departments with the highest patient risk were selected, including hematology, organ transplantation, and burn units. We selected 98 sites and sampled both water and water-related surfaces at each: three one-liter water samples (the first flow, cold and hot water) and two or three surface samples (inside the tap, pommel of the shower and siphon). At each site, a form was filled to specify its location in the unit, any water treatment (chlorine or other), filtering, and temperature. Water from taps equipped with sterilized filtration was sampled without the filter.

RESULTS

There was a significant difference (p=0.039) in the number of positive cultures between the three types of water sampled: hot water (>50 degrees C) was colonized less often than first flow or cold water. Only 4% of the hot-water samples had positive cultures, compared to the 52% of the cold-water samples. Except in two hospitals with generalized contamination of the water pipes (one with Exophiala spp and the other with Fusarium spp), colonization was usually slight. Cold water was more colonized than hot water, but 79% of the samples yielded fewer than 5CFU/L. Dematiaceous hyphomycetes were isolated; Aspergillus spp were rare. The number of CFU in surface samples (that is, biofilms) was higher (mean=15 CFU per sample) but surfaces were positive less often than water (13% compared with 43% of all water samples). Sampling from siphons was productive more often than from taps (23%), but the molds isolated differed from those in the related water. Relations to bacterial flora and P. aeruginosa were also studied, together with the effects of chemical treatment.

CONCLUSION

Current regulations require only bacteriological survey. The absence of knowledge about the threshold of contamination at which there is a risk of nosocomial invasive fungal infections makes it difficult to impose routine monitoring. Mycological surveys of water are required during hospital renovation, plumbing work, pipe maintenance and when air samples are negative during nosocomial infection investigations.