PEG PLACEMENT IN PATIENTS WITH VENTRICULOPERITONEAL SHUNTS

To the Editor: The technique of PEG has received significant attention since it was first introduced in 1980. Although it has been shown to be practical, cost-effective, and safe for the long-term non-volitional enteral support of patients with neurologic disease,l little has been reported concerning complications of PEG placement or replacement in patients with ventriculoperitoneal (VP) shunts. In this letter we re­ port a patient who after PEG catheter replacement devel­ oped a bacterial infection of her VP shunt that required shunt hardware removal. A 30-year-old woman with a history of a VP shunt placed at age 14 for congenital aqueductal stenosis underwent PEG placement after a motor vehicle accident left her a quadri­ plegic. A 14F Ross PEG was placed uneventfully with intra­ venously administered prophylactic antibiotics using the standard Ponsky technique. One year later a replacement gastrostomy tube was placed endoscopically without antibi­ otic prophylaxis. A wire snare was looped around the gastrostomy tube bumper, which was then cut externally and removed endoscopically. A 14F replacement tube was placed through the existing gastrocutaneous fistula. The next day the patient developed a temperature of 39° C. Physical ex­ amination was unremarkable. Her WBC count was 14,000 with 90% neutrophils. Her chest x-ray film, urinalysis, and blood cultures were all unremarkable. Over the course of the next several days the patient remained intermittently febrile with temperature elevations as high as 40° C. Her abdomen remained soft and without tenderness. The PEG site was clean and dry. The patient underwent a diagnostic VP shunt aspiration. The cerebrospinal fluid analysis was remarkable for a WBC count of 980 with 76% neutrophils with an ele­ vated protein and a decreased glucose concentration. The gram stain showed gram-positive cocci, and the cerebrospi­ nal fluid culture grew Staphylococcus epidermidis. The complications of PEG placement have been well re­ ported and, although of varying severity, include infection,2 bleeding, tube leakage, aspiration, peritonitis,3 necrotizing fasciitis,4 and gastrocolic fistula. No data suggest an increase in complications in PEG placement in patients with VP shunts,5 although one may suspect that placement difficul­ ties may occur related to the location of the shunt and that infectious complications might be more common in any sub­ group of patients who have co-existent peritoneal catheters. 6 It is now our policy to use intravenously administered pro