Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles.
A fundamental error contained in the review by Safdar, et al (JAMA.
2002 Aug 28;288(8):996-1001) is perpetuated in the communication by
Phillips, et al (BMJ 2005;330:409-410 (19 February),
doi:10.1136/bmj.330.7488.409). Specifically, Safdar, et al, included in
their meta-analysis the study by Ikeda, et al. (Clin Nephrol. 1999
Dec;52(6):357-62), and characterized this study as demonstrating that
fosfomycin treatment of Escherichia coli O157:H7 infections diminished the
risk of developing the haemolytic uraemic syndrome (HUS), compared to
using no antibiotics. In fact, only two of the 256 children studied
intensively by Ikeda, et al., were untreated, so the efficacy of
antibiotics in preventing HUS was not at all demonstrated. The only
benefit that was statistically significant emerged only after subgroup
analysis was performed and identified that treatment on day two of illness
was valuable, and only in comparison of fosfomycin to other antibiotics.
Such a subgroup would be nearly impossible to identify in actual clinical
practice; additionally the biologic plausibility of why antibiotic therapy
would help only on one day out of the total illness seems lacking. In
multiple other careful analyses of outbreaks (Bell, BP, et al, Pediatrics.
1997 Jul;100(1):E12; Pavia, AT, et al., J Pediatr. 1990 Apr;116(4):544-51;
Dundas, et al., Clin Infect Dis. 2001 Oct 1;33(7):923-31. Epub 2001 Sep
05; Carter, AO, et al., N Engl J Med. 1987 Dec 10;317(24):1496-500) and
sporadic cases (Wong, CS., et al, N Engl J Med. 2000 Jun 29;342(26):1930-
6), antibiotics appear to increase the risk of this complication when
given to infected patients.
The quality of meta-analyses depends on the interpretation of the analyzed
research. Unfortunately, once published, there is often uncritical
acceptance of a meta-analysis’ accuracy. It is incumbent upon the authors
of meta-analyses and of derivative communications, and of clinicians
making decisions based on these analyses, to evaluate completely and
accurately the original studies on which such exercises are based.
Regarding the patient described by Phillips, et al, we have been impressed
that E. coli O157:H7 infections rarely present with a clinical picture
suggestive of sepsis. None of us has ever seen bacteremia with E. coli
O157:H7 prior to the development of HUS. The nature of the patient’s
“clotting abnormality” (beyond thrombocytopenia) would be helpful to know,
because HUS is rarely accompanied by a consumptive coagulopathy, and the
fibrinogen levels and prothrombin times can often assist in sorting out a
confusing differential diagnosis. We also are not aware of data
demonstrating harm in using antibiotics after HUS is manifest as was done
in the case reported by Phillips, et al. However, we believe that the
available data do suggest that antibiotics administered to patients
infected with E. coli O157:H7 are associated with an increased risk of
developing HUS. Based on the fundamental principle of “First, do no
harm,” we urge against their use.
Competing interests:
None declared
Competing interests:
No competing interests
25 February 2005
Stephanie Dundas
Consultant physician
Marguerite A. Neill, W.T. Andrew Todd, Phillip I. Tarr
Lanarkshire Area Infectious Diseases Unit, Monklands Hospital, ML6 0JS, Scotland
Antibiotics and Escherichia Coli O157:H7 Infections
A fundamental error contained in the review by Safdar, et al (JAMA.
2002 Aug 28;288(8):996-1001) is perpetuated in the communication by
Phillips, et al (BMJ 2005;330:409-410 (19 February),
doi:10.1136/bmj.330.7488.409). Specifically, Safdar, et al, included in
their meta-analysis the study by Ikeda, et al. (Clin Nephrol. 1999
Dec;52(6):357-62), and characterized this study as demonstrating that
fosfomycin treatment of Escherichia coli O157:H7 infections diminished the
risk of developing the haemolytic uraemic syndrome (HUS), compared to
using no antibiotics. In fact, only two of the 256 children studied
intensively by Ikeda, et al., were untreated, so the efficacy of
antibiotics in preventing HUS was not at all demonstrated. The only
benefit that was statistically significant emerged only after subgroup
analysis was performed and identified that treatment on day two of illness
was valuable, and only in comparison of fosfomycin to other antibiotics.
Such a subgroup would be nearly impossible to identify in actual clinical
practice; additionally the biologic plausibility of why antibiotic therapy
would help only on one day out of the total illness seems lacking. In
multiple other careful analyses of outbreaks (Bell, BP, et al, Pediatrics.
1997 Jul;100(1):E12; Pavia, AT, et al., J Pediatr. 1990 Apr;116(4):544-51;
Dundas, et al., Clin Infect Dis. 2001 Oct 1;33(7):923-31. Epub 2001 Sep
05; Carter, AO, et al., N Engl J Med. 1987 Dec 10;317(24):1496-500) and
sporadic cases (Wong, CS., et al, N Engl J Med. 2000 Jun 29;342(26):1930-
6), antibiotics appear to increase the risk of this complication when
given to infected patients.
The quality of meta-analyses depends on the interpretation of the analyzed
research. Unfortunately, once published, there is often uncritical
acceptance of a meta-analysis’ accuracy. It is incumbent upon the authors
of meta-analyses and of derivative communications, and of clinicians
making decisions based on these analyses, to evaluate completely and
accurately the original studies on which such exercises are based.
Regarding the patient described by Phillips, et al, we have been impressed
that E. coli O157:H7 infections rarely present with a clinical picture
suggestive of sepsis. None of us has ever seen bacteremia with E. coli
O157:H7 prior to the development of HUS. The nature of the patient’s
“clotting abnormality” (beyond thrombocytopenia) would be helpful to know,
because HUS is rarely accompanied by a consumptive coagulopathy, and the
fibrinogen levels and prothrombin times can often assist in sorting out a
confusing differential diagnosis. We also are not aware of data
demonstrating harm in using antibiotics after HUS is manifest as was done
in the case reported by Phillips, et al. However, we believe that the
available data do suggest that antibiotics administered to patients
infected with E. coli O157:H7 are associated with an increased risk of
developing HUS. Based on the fundamental principle of “First, do no
harm,” we urge against their use.
Competing interests:
None declared
Competing interests: No competing interests