No Clinical Benefit of Empirical Antimicrobial Therapy for Pediatric Diarrhea in a High-Usage, High-Resistance Setting

Abstract

Background: Pediatric diarrheal disease presents a major public health burden in low- to middle-income countries. The clinical benefits of empirical antimicrobial treatment for diarrhea are unclear in settings that lack reliable diagnostics and have high antimicrobial resistance (AMR).

Methods: We conducted a prospective multicenter cross-sectional study of pediatric patients hospitalized with diarrhea containing blood and/or mucus in Ho Chi Minh City, Vietnam. Clinical parameters, including disease outcome and treatment, were measured. Shigella, nontyphoidal Salmonella (NTS), and Campylobacter were isolated from fecal samples, and their antimicrobial susceptibility profiles were determined. Statistical analyses, comprising log-rank tests and accelerated failure time models, were performed to assess the effect of antimicrobials on disease outcome.

Results: Among 3166 recruited participants (median age 10 months; interquartile range, 6.5-16.7 months), one-third (1096 of 3166) had bloody diarrhea, and 25% (793 of 3166) were culture positive for Shigella, NTS, or Campylobacter. More than 85% of patients (2697 of 3166) were treated with antimicrobials; fluoroquinolones were the most commonly administered antimicrobials. AMR was highly prevalent among the isolated bacteria, including resistance against fluoroquinolones and third-generation cephalosporins. Antimicrobial treatment and multidrug resistance status of the infecting pathogens were found to have no significant effect on outcome. Antimicrobial treatment was significantly associated with an increase in the duration of hospitalization with particular groups of diarrheal diseases.

Conclusions: In a setting with high antimicrobial usage and high AMR, our results imply a lack of clinical benefit for treating diarrhea with antimicrobials; adequately powered randomized controlled trials are required to assess the role of antimicrobials for diarrhea.

Keywords: Campylobacter; Shigella; antimicrobial resistance; disease outcome; fluoroquinolones; multidrug resistance; nontyphoidal Salmonella; pediatric diarrhea.

Figure 1.

The antimicrobial resistance profiles for isolated Campylobacter spp. (A), nontyphoidal Salmonella spp. (B), and Shigella spp. (C), showing antimicrobial susceptibility and multidrug resistance (MDR), defined as nonsusceptibility to ≥1 agent in ≥3 antimicrobial categories). Bar graphs shows proportion of organisms exhibiting nonsusceptibility (dark gray) to nalidixic acid (NAL), ciprofloxacin (CIP), ceftriaxone (CRO), ceftazidime (CAZ), amoxicillin-clavulanic acid (AMC), ampicillin (AMP), trimethoprim-sulfamethoxazole (SXT), azithromycin (AZM), chloramphenicol (CHL), amikacin (AMK), gentamicin (GEN), erythromycin (ERY), clindamycin (CLI), and IPM ( ).

Figure 2.

Effect of antimicrobial treatment on clinical outcome. A, B, Kaplan-Meier curves show days in the hospital for diarrheal children treated with antimicrobials (A) or specifically, fluoroquinolones (FLQs) (B). C, D, Effect of antimicrobial usage on the length of hospital stay by different diarrheal types (C) and blood C-reactive protein (CRP) concentration (5 mg/L cutoff) (D). Statistical comparisons for categorical variables were conducted using the Kruskal-Wallis test, where *.05 < P < .01, and *** P < .001. Log-rank tests were used to compare Kaplan-Meier curves for length of hospital stay between groups.

Figure 3.

Effect of antimicrobial resistance on clinical outcome. A, B, Kaplan-Meier curves for length of hospital stay in diarrheal children treated with either antimicrobials or fluoroquinolones (FLQs) and stratified by multidrug resistance (MDR) (A) or FLQs resistance (B) profile of the isolated bacteria. C, D, Effect of MDR (C) and FLQ resistance (D) on the length of hospital stay in diarrheal children infected with Campylobacter, nontyphoidal Salmonella, and Shigella, while being treated with antimicrobials or FLQs, respectively. Statistical comparisons for categorical variables were conducted using the Kruskal-Wallis test. Log-rank tests were used to compare Kaplan-Meier curves for length of hospital stay between groups.