LAUSR

Background Gram-negative rod bacteremia (GNRB) is a leading cause of morbidity and mortality among allogeneic hematopoietic cell transplant (HCT) recipients. Historically, GNRB has been treated with inpatient intravenous antibiotic therapies. However, outpatient antibiotic delivery has been shown to lower costs, decrease complications and may reduce rates of hospital-acquired multidrug-resistant organism (MDRO) infections. Despite such advantages, there are no data assessing trends in outpatient management of GNRB among allogeneic HCT recipients nor on the impact changing resistance profiles have on such trends. We estimated the linear time-trend, shape of trend, and average proportion of treatment time spent outpatient for allogeneic HCT recipients with GNRB. Methods First 100 day post-HCT data were retrospectively examined for adult patients transplanted at a single institution between 2007 and 2017. We defined GNRB as blood culture confirmed GNR, limited our analysis to first infections, and presumed antibiotic duration lasted 14 days. We estimated the direction and magnitude of the time trend and tested for an association between calendar year and proportion of time spent outpatient using linear analysis with a continuous year variable. We accounted for a priori selected confounders (graft vs. host disease (GVHD), neutropenia, conditioning regimen, location of GNRB diagnosis, polymicrobial culture, MDRO (Stenotrophomonas maltophilia isolates or resistance to ³1 agents in ³3 categories: cephalosporins, beta-lactamase inhibitors, carbapenems, aminoglycosides, and fluoroquinolones), and age at diagnosis) using adjusted linear analysis and estimated the shape of the time-trend using a linear model with a natural spline for time. We tested if MDRO modified the outpatient temporal trend using a likelihood ratio test. Results Of the 2,165 allogeneic HCT recipients included, 252 (11.6%) experienced 1 or more GNRB with available resistance data. Twenty-four (9.5%) analyzed infections were polymicrobial and 75 (29.8%) were MDR. On average, patients received 59.2% of their antibiotic treatment in an outpatient setting. We observed a crude and adjusted per year decline in the proportion of treatment time spent outpatient (crude: -2.6% [95% CI: -1.0, -4.2%]; adjusted: -1.1% [-2.5, 0.2%]); the decline was non-linear (Figure 1). MDR, GVHD, neutropenia, and being inpatient at diagnosis were associated with a smaller proportion of treatment time spent outpatient and MDR status was found to modify the association between calendar year and proportion of time spent outpatient (p-value: 0.04; Figure 2). Conclusion The proportion of GNRB antibiotic treatment time allogeneic HCT recipients spent in the outpatient setting declined between 2007 and 2017. However, this decline was explained by severity of illness and infection complexity factors, such as MDR and GVHD.