Background Proton pump inhibitors are considered safe medications, but their use has been associated with colonization or infection with nosocomial pathogens. We used a mouse model to test the hypothesis that inhibition of gastric acid production by proton pump inhibitor treatment may facilitate establishment of intestinal colonization by nosocomial pathogens such as vancomycin-resistant enterococci (VRE) and Klebsiella pneumoniae.

Methods In vitro studies were performed to examine killing of VRE and K. pneumoniae by acid and physiologic concentrations of nitrite. Mice received subcutaneous pantoprazole or saline every 12 hours for 3 days followed by orogastric inoculation of 100 colony-forming units of VRE or K. pneumoniae in combination with subcutaneous clindamycin. The presence and density of the pathogens in stool were monitored by plating samples onto selective media. Student's t-test was performed to compare the pH of gastric contents from pantoprazole versus saline-treated mice and chi-square test was performed to compare the frequency of acquisition of stool colonization with the pathogens.

Results Exposure to pH 2 buffer or fasting human gastric contents resulted in significant killing of the pathogens and physiological concentrations of nitrite-enhanced killing. In comparison to saline controls, pantoprazole-treated mice had elevated gastric pH (mean ± SD, 4.8 ± 0.87 versus 2.3 ± 0.3; p < .001) and increased frequency of establishment of intestinal colonization with the pathogens (75% versus 25% for VRE, p = .01; 100% versus 40% for K. pneumoniae, p < .001). For all mice that developed colonization, high densities of the pathogens were measured (range 6-9.8 log10CFU/g of stool). Conclusions: Proton pump inhibitor administration facilitates establishment of intestinal colonization by VRE and K. pneumoniae in clindamycin-treated mice. These data suggest that the widespread use of proton pump inhibitors in health care facilities could contribute to the dissemination of nosocomial pathogens.