Treatment, clinical outcomes, and predictors of mortality among a national cohort of hospitalized patients with Stenotrophomonas maltophilia infection.

OBJECTIVES: To analyze treatment, clinical outcomes, and predictors of inpatient mortality in hospitalized patients with Stenotrophomonas maltophilia infection. STUDY DESIGN: Retrospective cohort study. METHODS: We included patients admitted to Veterans Affairs hospitals nationally with S. maltophilia cultures and treatment from 2010 to 2019. We described patient and clinical characteristics, antibiotic treatment, and clinical outcomes. Univariate and multivariable logistic regression were used to evaluate predictors of inpatient mortality. RESULTS: We identified 3891 hospitalized patients treated for an S. maltophilia infection, of which 13.7% died during admission. The most common antibiotic agents were piperacillin/tazobactam (39.7%), sulfamethoxazole/trimethoprim (23.3%), and levofloxacin (23.2%). Combination therapy was used in 16.6% of patients. Independent predictors of inpatient mortality identified in multivariable analysis included the following: presence of current acute respiratory failure (adjusted odds ratio [aOR] 4.74, 95% confidence interval [CI] 3.63-6.19), shock (aOR 3.00, 95% CI 2.31-3.90), acute renal failure (aOR 2.06, 95% CI 1.64-2.60), and septicemia (aOR 1.90, 95% CI 1.49-2.42), age 65 years and older (aOR 2.05, 95% CI 1.07-3.94, reference age 18-49 years), hospital-acquired infection (aOR 1.87, 95% CI 1.48-2.37), Black (aOR 1.58, 95% CI 1.21-2.06) and other races (aOR 1.65, 95% CI 1.41-2.41, reference White), liver disease (aOR 1.51, 95% CI 1.02-2.22), and median Charlson comorbidity score or higher (aOR 1.36, 95% CI 1.08-1.71, reference less than median). Clinical outcomes were similar between patients infected with sulfamethoxazole/trimethoprim-resistant, levofloxacin-resistant, and multidrug-resistant S. maltophilia strains compared to non-resistant strains. CONCLUSIONS: In our national cohort of hospitalized patients with S. maltophilia infection, 13.7% of patients died during admission and several predictors of inpatient mortality were identified. Predictors related to the severity of infection were among the strongest identified. It is important that in severely ill patients presenting to the hospital, S. maltophilia be considered as a cause.

Predictors of potentially suboptimal treatment of urinary tract infections in long-term care facilities.

BACKGROUND: Suboptimal antibiotic treatment of urinary tract infection (UTI) is high in long-term care facilities (LTCFs) and likely varies between facilities. Large-scale evaluations have not been conducted. AIM: To identify facility-level predictors of potentially suboptimal treatment of UTI in Veterans Affairs (VA) LTCFs and to quantify variation across facilities. METHODS: This was a retrospective cohort study of 21,938 residents in 120 VA LTCFs (2013-2018) known as Community Living Centers (CLCs). Potentially suboptimal treatment was assessed from drug choice, dose frequency, and/or treatment duration. To identify facility characteristics predictive of suboptimal UTI treatment, LTCFs with higher and lower rates of suboptimal treatment (≥median, < median) were compared using unconditional logistic regression models. Joinpoint regression models were used to quantify average percentage difference across facilities. Multilevel logistic regression models were used to quantify variation across facilities. FINDINGS: The rate of potentially suboptimal antibiotic treatment varied from 1.7 to 34.2 per 10,000 bed-days across LTCFs. The average percentage difference in rates across facilities was 2.5% (95% confidence interval (CI): 2.4-2.7). The only facility characteristic predictive of suboptimal treatment was the incident rate of UTI per 10,000 bed-days (odds ratio: 4.9; 95% CI: 2.3-10.3). Multilevel models demonstrated that 94% of the variation between facilities was unexplained after controlling for resident and CLC characteristics. The median odds ratio for the full multilevel model was 1.37. CONCLUSION: Potentially suboptimal UTI treatment was variable across VA LTCFs. However, most of the variation across LTCFs was unexplained. Future research should continue to investigate factors that are driving suboptimal antibiotic treatment in LTCFs.