Significant and Sustained Reduction in Hospital-Onset Candidemia in VA Hospitals
An 18-year retrospective cohort study of 130 hospitals in the US Veterans Health Administration system revealed a significant reduction in cases of hospital-onset candidemia. But what caused the drop?
The rate of hospital-onset candidemia infections in the US Veterans Health Administration (VHA) system significantly declined from 2000 to 2017, while the rate of non-hospital-onset cases declined at a much slower rate.
Investigators with the University of Iowa and the Iowa City VA Health Care System seeking to determine the incidence of candidemia at 130 hospitals in the VHA system conducted a retrospective cohort study from January 2000 to December 2017. They presented their findings in an oral session at the European Congress of Clinical Microbiology and Infectious Diseases (ECCMID 2019).
Using data from electronic medical records, the research team identified patients with positive blood cultures for Candida spp to calculate the number of unique patients per month. The cases were then divided into either hospital-onset (first positive blood culture >48 hours after admission) or non-hospital-onset. Investigators used joinpoint regression models to identify temporal trends and monitor statistically significant changes.
A total of 15,763 unique patients produced 31,370 blood cultures that tested positive for Candida species. Eighty-five percent of those were identified as hospital-onset.
Joinpoint regression models indicated that the incidence rate of hospital-onset cases of candidemia was increasing until about the mid-2000s, when it began to fall. Two statistically significant reductions in slope were identified, 1 in 8/2004 [95% Confidence Interval (CI): 2/2003-4/2005] and another in 9/2007 (95% CI: 9/2006-6/2009). Investigators identified no statistically significant change in slope for non-hospital-onset cases.
The average monthly incidence rates were 2.79 per 10,000 patient-days [Interquintile Range (IQR): 1.71-3.69] for hospital-onset, and 0.49 per 100,000 patient-month (IQR: 0.33-0.63) for non-hospital-onset.
The significant and sustained reduction in cases of hospital-onset candidemia in the VHA system throughout the study period will likely have a positive impact on the rate of bloodstream infections caused by Candida species, which are associated with significant mortality and morbidity.
"These results strongly suggest that introduction of catheter-related bloodstream infection prevention bundle through [US Centers for Disease Control and Prevention] and other [health care-associated infection] prevention efforts had positive influences collectively," Michihiko Goto, MD, MSCI, assistant professor of internal medicine and infectious diseases at the University of Iowa Health Care and an investigator on the study, told Contagion®. "Our findings support the continued implementation of HAI bundles and provide important basis to future cost-effectiveness studies."
The study, “Nation-wide temporal trends of hospital- and non-hospital-onset candidemia incidence: an 18-year cohort within the US Veterans Health Administration System,” was presented in an oral session on Saturday, April 13, 2019, at ECCMID 2019 in Amsterdam, the Netherlands.