Prevalence of Antimicrobial Resistance and Association with Patient Outcomes in a Rural Kenyan Hospital

Data on antimicrobial resistance (AMR) and association with outcomes in resource-variable intensive care units (ICU) are lacking. Data currently available are limited to large, urban centers. We attempted to understand this locally through a dual-purpose, retrospective study. Cohort A consisted of adult and pediatric patients who had blood, urine, or cerebrospinal fluid cultures obtained from 2016 to 2020. A total of 3,013 isolates were used to create the Kijabe Hospital's first antibiogram. Gram-negative organisms were found to be less than 50% susceptible to third-and fourth -generation cephalosporins, 67% susceptible to piperacillin-tazobactam, 87% susceptible to amikacin, and 93% sus-ceptible to meropenem. We then evaluated the association between AMR and clinical characteristics, management, and outcomes among ICU patients (Cohort B). Demographics, vital signs, laboratory results, management data, and outcomes were obtained. Antimicrobial resistance was defined as resistance to one or more antimicrobials. Seventy-six patients were admitted to the ICU with bacteremia during this time. Forty complete paper charts were found for review. Median age was 34 years (interquartile range, 9-51), 26 patients were male (65%), and 28 patients were older than 18 years (70%). Septic shock was the most common diagnosis (n = 22, 55%). Six patients had AMR bacteremia; Escherichia coli was most common (n = 3, 50%). There was not a difference in mortality between patients with AMR versus non-AMR infections (P = 0.54). This study found a prevalence of AMR. There was no association between AMR and outcomes among ICU patients. More studies are needed to understand the impact of AMR in resource-variable settings.