Secondary Logo

Journal Logo

ESPID Reports and Reviews

The Dilemma of Antimicrobial Treatment of Shiga Toxin-producing Escherichia coli

Mor, Meirav MD*†; Ashkenazi, Shai MD, MSc†‡§

Author Information
The Pediatric Infectious Disease Journal: September 2014 - Volume 33 - Issue 9 - p 979-981
doi: 10.1097/INF.0000000000000425
  • Free

The recent widespread outbreak of Shiga toxin-producing Escherichia coli (STEC) infection, which resulted in considerable mortality and morbidity—nearly 4000 infected persons, of whom 855 developed hemolytic-uremic syndrome (HUS) and 53 died—underscored again the dilemma of antimicrobial treatment against infections caused by these organisms.1,2 Antibiotics are usually efficacious for treating bacterial infections, although for a minority of infections antibiotics are ineffective; the situation with STEC is very unusual, in that concerns have been raised as to possible detrimental effects of antibiotics.3,4


STEC, also named enterohemorrhagic E. coli (EHEC), have a wide spectrum of clinical presentations. The infection can be asymptomatic, as documented during outbreaks, presenting as mild watery diarrhea with abdominal cramps or causing severe hemorrhagic colitis with bloody diarrhea and emesis. Contrasting with invasive enteric infections, in STEC gastroenteritis, fever is of low grade or absent. Recovery of the intestinal symptoms usually occurs within a few days.

The major problem, however, is that after 4–10 days, about 10% of patients with STEC infection develop HUS, composed typically of microangiopathic hemolytic anemia, thrombocytopenia and acute renal failure. HUS is currently the leading cause of acquired renal failure in children living in developed countries. This sequence of events is related to unique virulence traits of STEC: after attachment to intestinal epithelium, they produce a potent cytotoxin that binds to a specific globotriaosyl ceramide receptor, stops cellular protein synthesis and subsequently causes cell death.2 The toxin is very similar to that produced by Shigella dysenteriae 1 (Shiga bacillus) and was therefore named Shiga toxin (previously Shiga-like toxin). E. coli serotype O157:H7 is the best characterized STEC, but a number of other serotypes of E. coli may produce the toxin and have similar clinical presentations.


The intestinal manifestations of STEC are usually self-limited and not affected significantly by antimicrobial treatment. The controversial issue is the role of antibiotics in the risk of developing the life-threatening complication of HUS (Table 1). In a meta-analysis of the 9 related studies published by February 2001, Safdar et al.5 examined the overall risk of HUS after antibiotic treatment of E. coli O157:H7. The analysis showed that antibiotic administration did not appear to increase the risk of HUS; the pooled odds ratio (OR) was 1.15 [95% confidence interval (CI): 0.79–1.68]. We summarize below the main studies that were published since then.

The Effect of Antimicrobial Therapy on the Outcome of STEC Infections

Studies Suggesting That Antibiotic Treatment Increases the Risk of Developing HUS

The 3 studies that reached this conclusion included patients with infections caused by the E. coli serogroup O157. Smith et al3 followed Minnesota residents with E. coli O157 infection who did (n = 66) or did not (n = 129) subsequently develop HUS, from 1996 to 2002. They found that overall, antibiotic treatment was not associated with the development of HUS. However, after adjustment for illness severity and gender, individuals who developed HUS were more likely to have been treated with bactericidal antibiotics within 3 days (adjusted OR: 12.4; CI: 1.4–110.3) or within 7 days (OR: 18.0; CI: 1.9–170.9) of the onset of diarrhea. The highest risk was for β-lactam administration, especially during the first 3 days of illness; sulfonamides or azoles did not increase the risk of HUS.

In a prospective 5-state cohort study conducted during 1997–2006, Wong et al4 investigated the effect of antibiotic treatment administered within 1 week of infection with E. coli serotype O157:H7 on the risk of HUS. After adjusting for indices of disease severity, they found a relative risk of HUS of 3.62 (CI: 1.23–10.6, P = 0.02). The association with oliguric HUS was especially high. The antibiotics that were administered included trimethoprim-sulfamethoxazole, β-lactams and azithromycin.

Mody et al investigated a large cohort of 1315 patients with STEC O157 infection by active surveillance in 10 US states during 2006–2010.6 They found that adults were more likely than children to be prescribed antibiotics for STEC diarrhea. Among all patients, treatment with any antibiotic was not significantly associated with developing HUS, but specific treatment with β-lactams or sulfonamides was associated. Among children aged 5–14 years, administration of any antibiotics posed a risk (OR: 4.0, CI: 1.3–12).6

Studies Suggesting That Antibiotic Treatment Against STEC Is Beneficial

Two clinical studies examined the role of antibiotic treatment during the STEC O104:H4 infection, an outbreak that originated in Germany and later spread throughout Europe and beyond and affected mainly adults (90% of the gastroenteritis cases, 88% of the HUS cases).1 Menne et al7 performed a retrospective case-control study in 23 hospitals in northern Germany, which included 298 adult patients with E. coli O104:H4-induced HUS, of whom 52 were treated with antibiotics (meropenem, ciprofloxacin or rifaximin). They found that antibiotic treatment reduced the severity of STEC-induced HUS and lowered the mortality rate from 5.2% to 0 (P = 0.029). Concomitantly, the mean duration of fecal excretion of the pathogen was reduced from 22.6 to 14.8 days (P < 0.001).

Geerdes-Fenge et al8 followed 24 patients with STEC O104:H4 infection hospitalized in a single center, of whom 19 developed HUS. Overall, antibiotic treatment did not affect significantly the risk of HUS: 57% in treated patients vs 88% in those not treated. However, some patients were empirically treated with β-lactams (cefotaxime or amoxicillin), to which the isolate was eventually found as resistant, because it produced an extended-spectrum β lactamase. Ciprofloxacin administration, which was microbiologically appropriate for the pathogen, was found to be protective: 40% HUS among treated patients versus 89% among those not treated (P = 0.043).8

As mentioned above, the toxin produced by STEC is very similar to Shiga toxin, which is produced by S. dysenteriae 1, and is also associated with HUS by a similar mechanism. It is therefore relevant to add that a comprehensive study analyzed 378 patients with S. dysenteriae diarrhea, 93% of whom had received antibiotics during the first 96 hours of illness to reduce the duration of fever and diarrhea.9 Only a single patient developed HUS (overall risk of 0.0026, CI: <0.001–0.015), suggesting that antimicrobial treatment of the infection reduced the risk of HUS. Fecal toxin levels fell rapidly and were undetectable within 72–120 hours.9

Mechanisms of Antimicrobial Effects

Antibiotic treatment of STEC infections could play a role in HUS development in several ways. By simply eliminating the pathogen, antibiotics might reduce toxin production and thus reduce the risk of HUS.9 Shortening the duration of fecal excretion of the pathogen might reduce transmission, and thus the risk of infectivity and of the occurrence of new cases.9–11

Biology, however, is more complicated. Although the growth of STEC strains is susceptible to inhibition by specific antibiotics, the bacteria may respond to antibiotic exposure by enhancing the production and/or release of Shiga toxin. The mechanism involves mainly the induction of prophages harboring Shiga toxin-encoding genes, which are silent in the lysogenic state. Stress signals, including some that are induced by antibiotics, enhance toxin transcription, production and release from bacterial cells by phage-mediated lysis.12 This might eventually lead to increased risk of developing HUS.

As suggested by clinical studies, different antibiotics and disparate timing of their administration may confer very distinct and even opposing effects on the risk of HUS development. Several laboratory studies were conducted to investigate this point in vitro, under controlled conditions. McGannon et al12 treated Shiga toxin lysogens with various classes of antibiotics in the presence or absence of phage-sensitive E. coli and measured toxin activity on Vero cells. They found indeed dual effects of antibiotics: growth-inhibitory levels suppress Shiga toxin production; subinhibitory levels of antibiotics that target DNA synthesis (ciprofloxacin, trimethoprim-sulfamethoxazole) increased toxin production, whereas antibiotics that affect the cell wall, transcription or translation did not.12 Interestingly, ciprofloxacin increased Shiga toxin levels even when the growth of E. coli O157:H7 was completely suppressed, whereas azithromycin decreased toxin levels even when bacterial viability remained high.12

Two studies examined in vitro the effects of antibiotics on toxin production by the STEC O104:H4 isolate that caused the severe 2011 outbreak. Bielaszewska et al2 found that subinhibitory concentrations of ciprofloxacin significantly increased Shiga toxin-harboring phage induction and toxin production, whereas fosfomycin and gentamicin exerted a nonsignificant influence on them, and meropenem, azithromycin, rifaximin and tigecycline significantly decreased them. Corogeanu et al13 examined the effects of antibiotics on strains of different STEC serotypes. Interestingly, unlike the common STEC serotype O157:H7, the strain that caused the recent outbreak with serotype O104:H4 did not release Shiga toxin in response to therapeutic concentrations of ciprofloxacin, meropenem, fosfomycin or chloramphenicole.13 In light of these specific strain-dependent antibiotic effects, the authors suggested that in future outbreaks, the response of the specific STEC epidemiologic strain to antibiotics should be rapidly characterized in vitro with regard to the effects of antimicrobial agents on toxin production; agents that decrease, or at least do not increase, toxin production should be evaluated in clinical studies for their effect on the clinical course of the infection and its complications.13

The timing of antibiotic administration might be crucial; as antibiotics do not influence the biologic effects of toxin that has already bound to its tissue receptors, its early administration during the course of the acute gastroenteritis might be more beneficial. The reported studies were not designed to answer this possibility. Antibiotics may theoretically increase the risk of HUS by killing the normal gut flora, especially when the STEC strain is resistant to the antibimicrobial agent administered, as happened with the epidemic strain that originated in Germany in 2011, which contained extended-spectrum β lactamase-encoding genes.1,8

Finally, it is also possible that antibiotic treatment is a confounder but does not have a causative role in the risk of developing HUS. As the studies published are observational in nature, a selection bias is obviously possible, that is, antibiotics might have been administered to the more severely ill patients, with higher risk of being complicated by HUS, independent of the treatment. Controlling for all variables is not always possible. It seems that only a well-designed randomized controlled trial can resolve this controversial issue.


  1. The effect of antimicrobial treatment of STEC infections on the risk of developing HUS is currently inconclusive, yet seems to depend on the specific antibiotics used and the specific pathogen involved.
  2. A prospective randomized controlled study to elucidate this critical issue is recommended, preferably as a multicenter effort.
  3. Meanwhile, antimicrobials are not recommended for patients with suspected STEC infections (ie, during outbreaks), as advised by the World Health Organization.
  4. In sporadic cases of bloody diarrhea, etiologies other than STEC should be of course considered.
  5. In particular, because antimicrobial treatment is highly efficacious against shigellosis—as it significantly reduces the duration of diarrhea, fever, complications and transmissibility—this etiology should be estimated according to the presence of high fever, toxicity and the local epidemiologic data; antibiotics should be considered on an individual basis. Rapid determination of the causative agent, for example, by molecular methods, will help with the decision.


1. Buchholz U, Bernard H, Werber D, et al. German outbreak of Escherichia coli O104:H4 associated with sprouts. N Engl J Med. 2011;365:1763–1770
2. Bielaszewska M, Idelevich EA, Zhang W, et al. Effects of antibiotics on Shiga toxin 2 production and bacteriophage induction by epidemic Escherichia coli O104:H4 strain. Antimicrob Agents Chemother. 2012;56:3277–3282
3. Smith KE, Wilker PR, Reiter PL, et al. Antibiotic treatment of Escherichia coli O157 infection and the risk of hemolytic uremic syndrome, Minnesota. Pediatr Infect Dis J. 2012;31:37–41
4. Wong CS, Mooney JC, Brandt JR, et al. Risk factors for the hemolytic uremic syndrome in children infected with Escherichia coli O157:H7: a multivariable analysis. Clin Infect Dis. 2012;55:33–41
5. Safdar N, Said A, Gangnon RE, et al. Risk of hemolytic uremic syndrome after antibiotic treatment of Escherichia coli O157:H7 enteritis: a meta-analysis. JAMA. 2002;288:996–1001
6. Mody RK, Kendall M, Dunn J, et al. Antibiotic treatment of Shiga toxin-producing Escherichia coli (STEC) O157 varies by age and may increase risk of hemolytic uremic syndrome. Available at:
7. Menne J, Nitschke M, Stingele R, et al.EHEC-HUS consortium. Validation of treatment strategies for enterohaemorrhagic Escherichia coli O104:H4 induced haemolytic uraemic syndrome: case-control study. BMJ. 2012;345:e4565
8. Geerdes-Fenge HF, Löbermann M, Nürnberg M, et al. Ciprofloxacin reduces the risk of hemolytic uremic syndrome in patients with Escherichia coli O104:H4-associated diarrhea. Infection. 2013;41:669–673
9. Bennish ML, Khan WA, Begum M, et al. Low risk of hemolytic uremic syndrome after early effective antimicrobial therapy for Shigella dysenteriae type 1 infection in Bangladesh. Clin Infect Dis. 2006;42:356–362
10. Nitschke M, Sayk F, Härtel C, et al. Association between azithromycin therapy and duration of bacterial shedding among patients with Shiga toxin-producing enteroaggregative Escherichia coli O104:H4. JAMA. 2012;307:1046–1052
11. Ichinohe S, Ichinohe N, Sakuma F. Antimicrobial therapy and shedding time of Shiga toxin-producing Escherichia coli. Scand J Infect Dis. 2008;40:1002–1003
12. McGannon CM, Fuller CA, Weiss AA. Different classes of antibiotics differentially influence shiga toxin production. Antimicrob Agents Chemother. 2010;54:3790–3798
13. Corogeanu D, Willmes R, Wolke M, et al. Therapeutic concentrations of antibiotics inhibit Shiga toxin release from enterohemorrhagic E. coli O104:H4 from the 2011 German outbreak. BMC Microbiol. 2012;12:160–169
© 2014 by Lippincott Williams & Wilkins, Inc.