OBJECTIVES: To examine ICU prevalence of Ventilator-Associated Pneumonia (VAP) and how prevalence increases with time on a mechanical ventilator (MV). To analyze economic and clinical outcomes associated with VAP controlling for time on MV.

METHODS: Retrospective analysis of patients 18 years+ in the MIMIC IV database. The analysis was based on administrative claims and electronic medical records data. ICU patients with ICD-10 codes were included. VAP was defined by J95.851 ICD-10 diagnosis code. Time on MV was defined by three ICD-10 procedure codes: 5A1935Z, 5A1945Z, 5A1955Z. The relationship between VAP and the following outcomes were examined among patients with MV while controlling for time on MV: length of stay (LOS), ICU LOS, discharge location, and death. Patients that died were excluded from analyzing LOS.

RESULTS: Of the 69,211 adult ICU patients, 6,376 patients had MV, for a MV prevalence of 9.2%. Of those MV patients, 705 had an ICD-10 diagnosis code for VAP, for a VAP prevalence of 11.1%. VAP prevalence was 1.1% for patients with MV <24 hours, 5.3% for patients with MV between 24 and 96 hours, and 39.2% for patients with MV >96 hours (χ²=578, p<0.001). Controlling for time on MV, MV patients with VAP had an average additional LOS of 98 hours (p<0.001) and an average additional ICU LOS of 93 hours (p<0.001) than MV patients without VAP. Paradoxically, MV patients with VAP had lower mortality rates (OR: 0.63, p<0.001), but were substantially more likely to be transferred to a long-term acute care hospital (OR: 1.81, p<0.001).

CONCLUSIONS: This study demonstrated that VAP is a common nosocomial infection and the likelihood of VAP increased substantially with time on MV. Strategies to prevent the occurrence of VAP may reduce overall hospital LOS, ICU LOS, and transfers to long-term acute care.