Antibiotic prophylaxis for percutaneous endoscopic gastrostomy insertion in patients with non‐malignant disease

SIRS, The report by Saadeddin et al. on single-dose systemic antibiotic prophylaxis for percutaneous endoscopic gastrostomy (PEG) insertion in patients with non-malignant disease is a useful addition to the literature on prevention of infection after PEG placement. However, the high incidence of peristomal (47.4%) and systemic infection (38.3%) in the placebo arm of the trial makes it difficult to determine the relevance of the study to management of infectious complications of PEG insertion in other Institutions. The large discrepancy between the peristomal infection rate in this prospective, randomized trial and the incidence of peristomal sepsis in the same authors’ previous retrospective audit (5%) is likely to reflect the difficulties in definition and quantitation of peristomal sepsis, as much as the methodological problems associated with retrospective analyses. One possible explanation for a high peristomal infection rate, in studies using a scoring system in which the presence of ‘pus’ alone is classified as peristomal infection, is the misclassification of reflux of enteral feed through the stoma as ‘pus’. Therefore, it would be useful to know how many patients were classified as having a peristomal infection on the basis of the presence of ‘pus’ only, with no other signs of PEG-site infection. In the placebo group, nine isolates from peristomal wounds were methicillin-resistant Staphylococcus aureus (MRSA). However, in the antibiotic group no MRSA organisms were cultured from wound swabs. We suggest that this surprising mismatch in MRSA infection/colonization between the two randomized groups does not allow the conclusion to be drawn that antibiotic prophylaxis reduces infection risk, as the antibiotic group with fewer MRSA infections would be expected to fare better than MRSA-positive individuals, regardless of the administration of systemic antimicrobial chemotherapy. Indeed, a previous study has shown nasal carriage of S. aureus was the only significant independent risk factor for subsequent surgical-site infection by any organism in orthopaedic patients. The decision regarding which prophylaxis regimen to use for patients undergoing PEG insertion will be dependent on local microbiological screening. Following our prospective study demonstrating the importance of MRSA colonization for subsequent peristomal infection, we have used a prophylaxis regimen consisting of nasal and topical therapy against MRSA that does not provide cover against Gram-negative organisms. However, to date, no significant PEG-site infection has been attributable to a Gram-negative organism in the Leeds Teaching Hospitals NHS Trust. G. ROBINS* & M. HULL *Academic Unit of Medicine, University of Leeds, and Leeds Institute for Molecular Medicine, Jif Building, St James’s University Hospital, Leeds, North Yorkshire, UK E-mail: medggr@leeds.ac.uk