2307. Risk factors for Cytomegalovirus (CMV) reactivation and Association with Clinical Outcomes in Critically Ill Adults with Sepsis: A Pooled Analysis of Prospective Studies

Abstract

Abstract Background CMV reactivation in seropositive, non-immunosuppressed adults with sepsis has been associated with worse clinical outcomes. To inform rational design of interventional trials determining whether CMV prevention improves outcomes, it is critical to identify the independence and strength of the association of CMV reactivation measures with clinically-relevant endpoints. Identification of patient factors associated with CMV reactivation would allow for optimization of the study population. Methods We performed a secondary pooled analysis of two prospective cohorts with sepsis: an observational cohort of ICU patients (n = 40) and the placebo cohort from a randomized, double-blind trial of ganciclovir to prevent CMV reactivation in acute critical illness (n = 66). Personnel blinded to the PCR results assessed clinical variables; CMV DNAemia was measured by quantitative plasma PCR twice weekly. Multivariable modeling using logistic and linear methods was used to examine the associations of CMV with clinical outcomes and between baseline patient factors and measures of CMV reactivation (adjusted for age, race, gender, transfusion status, study cohort, and APACHE score). Results CMV reactivation occurred at any level in 38/106 (36%), at >100 IU/mL in 25/106 (24%), and at >1,000 IU/mL in 14/106 (13%). In a multivariate model, CMV reactivation at any level, >100 IU/mL, or >1,000 IU/mL was associated with fewer days alive and not requiring ventilation: mean difference of −3.5 days ([95% CI −7.0, 0], P = 0.057), −5.1 days ([−8.9, −1.2], P = 0.012), and −6.1 days ([−10.9, −1.2], P = 0.016), respectively. Multiple measures of CMV reactivation were associated with other clinically-relevant outcomes, even after adjustment for baseline factors (Table 1). The association of APACHE score with CMV reactivation measures was inconsistent and with small effect size. We did not identify other patient variables associated with subsequent CMV reactivation. Conclusion CMV reactivation in seropositive adults with sepsis is independently and quantitatively associated with clinically-important outcomes, including death or continued hospitalization by day 28, ventilator-, ICU-, and hospital-free days. These effect sizes provide key data to inform design parameters of future interventional trials. Disclosures All authors: No reported disclosures.