Methods: This retrospective analysis occurred at a large, eight-campus tertiary community hospital, where CDT results were obtained using a two-step algorithm consisting of enzyme immunosorbent assay (EIA) and polymerase chain reaction (PCR). CDT results were classified as hospital-onset (HO-CDT) if they were obtained on or after hospital day 4 as per CDC LabID definitions. We merged and analyzed laxative administrative data and CDT results between September and December of 2014.
Results: A total of 3,234 CDT tests were obtained from 2,543 unique patients, of which 387 (12.0%) were positive, and 149 (4.6%) were classified as HO-CDT. Multiple tests within 7 days were done on 282 (11.1%) patients. Among HO-CDT, the sensitivity of EIA testing was 44% (66/149). During this time period, 234 CDT tests were ordered on patients that were on laxatives ≥24 hours, of which 19 (7.8%) were classified HO-CDT. Among 203 CDT stool tests collected ≥24 hours from the time of order, 15 (7.4%) were classified HO-CDT. Sensitivity of EIA testing of patients on laxatives and those among delayed collection was 16% and 40%, respectively. Assuming all patients on laxatives and those with delayed stool collection represents colonization, we estimate that HO-CDT cases reported overestimate the true burden of HO-CDI cases by over 20%.
Conclusion: Positive HO-CDT results arising from testing delays and laxative use may suggest an overestimate of true HO-CDI cases as suggested by low pretest probability and low EIA sensitivity. Additional clinical studies should be done to validate these estimates. Processes to discourage provider testing for CDT in cases where the pretest probability is low should be considered.
J. Logan, None
V. Hsu, None
See more of: Poster Abstract Session