Risk of Hemolytic Uremic Syndrome Related to Treatment of Escherichia coli O157 Infection with Different Antimicrobial Classes

Abstract

Treatment of Shiga toxin-producing Escherichia coli O157 (O157) diarrhea with antimicrobials might alter the risk of hemolytic uremic syndrome (HUS). However, full characterization of which antimicrobials might affect risk is lacking, particularly among adults. To inform clinical management, we conducted a case-control study of residents of the FoodNet surveillance areas with O157 diarrhea during a 4-year period to assess antimicrobial class-specific associations with HUS among persons with O157 diarrhea. We collected data from medical records and patient interviews. We measured associations between treatment with agents in specific antimicrobial classes during the first week of diarrhea and development of HUS, adjusting for age and illness severity. We enrolled 1308 patients; 102 (7.8%) developed confirmed HUS. Antimicrobial treatment varied by age: <5 years (12.6%), 5-14 (11.5%), 15-39 (45.4%), ≥40 (53.4%). Persons treated with a β-lactam had higher odds of developing HUS (OR 2.80, CI 1.14-6.89). None of the few persons treated with a macrolide developed HUS, but the protective association was not statistically significant. Exposure to "any antimicrobial" was not associated with increased odds of HUS. Our findings confirm the risk of β-lactams among children with O157 diarrhea and extends it to adults. We observed a high frequency of inappropriate antimicrobial treatment among adults. Our data suggest that antimicrobial classes differ in the magnitude of risk for persons with O157 diarrhea.

Keywords: Escherichia coli O157; diarrhea; hemolytic uremic syndrome.

Figure 1

Diagram of study enrollment. An infection was excluded if the patient (or guardian) did not speak English or Spanish, did not report diarrhea, or could not be contacted by 10 telephone attempts within 45 days after collection of the specimen that yielded O157. The 747 persons who were not enrolled were similar to the 1329 enrolled patients with respect to age and sex: among those enrolled, median age was 15 years (IQR 5–40) and 52% were female; among those not enrolled, median age was 16 years (IQR 7–36) and 53% were female.

Figure 2

Percentage of patients treated with antimicrobials during the first 7 days of diarrhea caused by STEC O157, by age quartile and antimicrobial class. Panel (a) displays antimicrobial treatment documented in medical records. Panel (b) displays antimicrobial treatment documented in medical records or reported by the patient (or parent/guardian). Data on macrolides is not shown because only 20 patients received macrolides (documented among 12 (0.9%), reported among 8 (0.6%) more); documented macrolide exposure was greatest among children < 5 years old (1.3%) and lowest among adults ≥ 40 years old (0.3%). Abbreviation: SXT, trimethoprim-sulfamethoxazole.