Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
Objectivity required?
- David Carvel (5 October 1999)
-
Re: Objectivity required?
- Winfried V Kern (5 October 1999)
-
A selective patient cohort?
- M J Carter (12 October 1999)
-
Antibiotic prophylaxis in percutaneous endoscopic gastrostomy (PEG)
- Arno Dormann (13 October 1999)
-
Percutaneous endoscopic gastrostomy (PEG) and antibiotics
- Brian Scott (22 October 1999)
|
|
|||
|
David Carvel G41 3EH
Send response to journal:
|
I have a few concerns about Preclik et al's study of Co-amoxyclav compared to placebo in percutaneous endoscopic gastrostomy. Why was the trial stopped when (significant) results seemed to be emerging from within this relatively small group of 84 most of whom were elderly, had malignant disease and serious co-morbidity? Was it truly double blinded if during the trial it was discovered certain results were emerging in one arm? If these 84 subjects were from 6 centres (assuming an average of 14 per centre-though not specified) could most cases of infection have occurred in one or two centres? Why was Co-amoxyclav, which is manufactured solely by SmithKline Beecham, chosen? It costs more than twice generic Amoxycillin. The authors themselves admit other comparable antibiotics may be sufficient prophylaxis in theirs' and previous trials. This paper was "in part supported" by SKB. Which part and how much and how does it differ from a pharmaceutical advertisement? David Carvel MRCGP |
|||
|
|
|||
|
Winfried V Kern, Ass. Professor University Hospital, Ulm, Germany
Send response to journal:
|
Ad 1: The trial was stopped after the interim analysis revealed that continuing the trial to the originally planned sample size would be highly unlikely to yield different results, i.e. superiority of verum over placebo; the stoping rules had been defined before interim analysis, and the interim analysis was done by an external consulting statistician. The personnel involved in the trial remained blinded until the final analysis was completed. Ad 2: Would it be ethical to pursue the trial if the trial´s question has been answered? Admittedly, the results are valid for a patient population similar to that included in the study, i.e. elderly patients with comorbidities and - often - malignant underlying diseases. Ad 3: See 1; it was discovered that the OVERALL incidence of infection was higher than presumed; this always raises the question of overestimation of the sample size or of some unforeseen paradoxical results necessitating a more detailed anaylsis. Nobody was aware of the infection rates in the both arms, and the (unblinded) interim anaylsis was done by an external consulting statistician. Ad 4: There was no center effect; this was analysed. Ad 5: The reason to choose amoxicillin/clavulanic acid was to increase the coverage of gram-negative rods and anaerobes. Data from trials assessing preoperative prophylaxis in ENT cancer patients with aminopenicillin plus a ß-lactamase inhibitors and one PEG trial using amoxicillin/clavulanic acid for prophylaxis supported this choice (in our view), whereas there was conflicting data with regard to cephalosporins. If we had used amoxicillin and had no such impressive results, readers would have asked the question: why not amoxicillin/clavulanic acid to be sure that it works or it does not! As discussed in the paper, it is probably adequate to use a cephalosporin. This question, which drug to use and whether the choices are equivalent, is something that others may study. Without the support by SmithKline Beecham we would have been unable to perform the study - it was not performed by the company, but we were grateful to have study drug and placebo provided and to have a grant to ascertain rapid and reliable data management which - again - was not performed by the company. I believe the readers will be able to judge whether this trial is an advertisement or a clinical trial that answers a clinically relevant question. |
|||
|
|
|||
|
M J Carter
Send response to journal:
|
Editor- Preclik et al reported an incidence of 65% peristomal infection in their control group(1). This is considerably higher than previous studies where rates vary from 19-29.4%(2,3,4). This discordance can be explained by considering the indications for gastrostomy insertion. This feeding method is employed in five broad categories of patients- cerebrovascular disease, neurodegenerative conditions, malignancy (primarily ropharyngeal/oesophageal), dementia with anorexia and head injuries. In the previous studies malignancy represented 14.8-21.1%(2,3,4), whilst in Preclik's cohort it is 65%. It has previously been suggested that patients with underlying malignancy are more susceptible to peristomal infection.(2). This may explain the high infection rate in the controls of this study and therefore the benefit of antibiotics in this particular patient sub-group. This, however, cannot necessarily be extrapolated to other indications for gastrostomy insertion. The British Society of Gastroenterology has suggested that this is an area requiring further evaluation(5). We require a similar study encompassing all the conditions for which gastrostomies are inserted in the United Kingdom to confidently resolve this issue. 1 Preclik G, Grune S, Leser HG, Lebherz J, Heldwein W, Machka K et al. A prospective, randomised, double-blind trial of prophylaxis with single dose co-amoxiclav before percutaneous endoscopic gastrostomy. BMJ 1999;319:881-4. 2 Sturgis TM, Yancy W, Cole JC, Proctor DD, Minhas BS, Marcuard SP. Antibiotic prophylaxis in percutaneous endoscopic gastrostomy. Am J Gastroenterol 1996;91:2301-4. 3 Jain NK, Larson DE, Schroeder KW, Burton DD, Cannon KP, Thompson RL et al. Antibiotic prophylaxis for percutaneous endoscopic gastrostomy. A randomised, double-blind clinical trial. Ann Intern Med 1987;107:824-8. 4 Jonas SK, Neimark S, Panwalker AP. Effect of antibiotic prophylaxis in percutaneous endoscopic gastrostomy. Am J Gastroenterol 1985;80:438-41. 5 British Society of Gastroenterology. Antibiotic prophylaxis in gastrointestinal endoscopy. Guidelines in Gastroenterology. Gut(suppl) September 1996. D S Sanders M J Carter |
|||
|
|
|||
|
Arno Dormann
Send response to journal:
|
EDITOR - Preclik et al´s paper is the second recently published to investigate antibiotic prophylaxis in percutaneous endoscopic gastrostomy (PEG) showing effciacy in reducing local infections. In both papers the subgroup of patients with underlying neurological disease is in the minority. These patients are at high risk for systemic infections especially post PEG pneumonia. The authors state that prior selected cancer patients are at high risk of post PEG infections which has not been proven yet. It can not be ruled out that these patients have been under chemo- or radiotherapy, that a concommitant procedure e. g. bouginage was performed or a medication e. g. immunosuppression was given during PEG intervention which might also lead to higher infection rates. In most studies and in clinical practice local infections rarely lead to induction of systemic antibiotic therapy post PEG. Infection is mostly controlled by intense local treatment 2,4. Systemic infections are still the major problem leading to induction of antibiotic therapy. They mostly occure in patients with neurological disease 3. Care has to be taken of these patients who are at high risk for aspiration sometimes induced by a feeding protocol with great bolus application. In our final results of 216 (106 individuals with ceftriaxone prophylaxis) standardized PEG procedures including pre- and postinterventional aftercare 67,2% of the patients had a neurological disease 4. Patients with tumours were scored positive for local infection on day ten postinterventionally without and with antibiotic prophylaxis in 31,0% vs. 3,7% (p<0,05) respectively whereas in patients with neurological disease infection rate was 20,3% vs. 11,4% respectively (p>0,05). Out of 18.5% patients scoring positive for peristomal infection 2.7% (all patients without prophylaxis) received antibiotics for this reason. The total rate of systemic infections was 11,8% without and 1,9% with antibiotic prophylaxis (p<0,05). Patients without and with prophylaxis and neurological disease had pneumonia in 9,5% and 1,9% (p<0,05%) respectively and with tumor disease 1,4% and 0% respectively. All pneumonias occured within 96 hours post-PEG. Preclik et al state that the choice of prophylactic regimen is unlikely to account for differing results. This is data based on local infection but there are no data available for systemic infection in which a prolonged interval of antibiotic coverage or short term therapy could be more effective 2,4. Our data indicate that a long acting substance like ceftriaxone is superior in reducing systemic complications. Arno J. Dormann M.D. Literature 1. Peclik G, Grüne S, Leser HG, Lebherz J, Heldwein W, Machka K, Holstege A, Kern WV. Prospective, randomised, double blind trial of prophylaxis with single dose of co-amoxiclav before percutaneous endoscopic gastrostomy. BMJ 1999; 319: 881-4 2. Gossner L, Keymling J, Hahn EG, Ell C. Antibiotic prophylaxis in percutaneous endoscopic gastrostomy (PEG): a prospective randomized clinical trial. Endoscopy 1999; 2: 119-24 3. Kadakia SC, Sullivan HO, Starnes E. Percutaneous endoscopic gastrostomy or jejuno-stomy and the incidence of aspiration in 79 patients. Am J Surg 1992; 164: 114-8. 4. Dormann AJ, Wiggínghaus B, Risius H, Kleimann F, Kloppenborg A, Grünewald T, Huchzermeyer H. A single dose of ceftriaxone administered 30 minutes before percutaneous endoscopic gastrostomy (PEG) significantly reduces local and systemic infective complications. Am J Gastroenterol 1999; in press 5. Coben RA, Weintraub A, DiMarino AJ, Cohen S. Gastroesophageal reflux during gastrostomy feeding. Gastroenterology 1994; 106: 13-18 |
|||
|
|
|||
|
Brian Scott
Send response to journal:
|
EDITOR - For many of us who, over many years, have frequently inserted PEGs without antibiotic cover, the report by Preclik and colleagues (1) will have caused surprise. In particular, the 47% incidence of non-minor wound infection needs scrutiny. Although without a similar formal prospective study we cannot reliably refute similar findings in our practice, it does seem at great variance with our experience. One therefore has to look at reasons for the likely difference between the experience in those units studied in Germany and our own. Firstly, we would require more details of the practice in those units. For instance, do they, like us, adopt a strict aseptic technique with gloves, gowns and abdominal drapes? Do they, like us, use a narrow gauge tube (such as the 9 French Fresenius 'Freka' tube with an external diameter of 2.9mm) which does not require a scalpel blade incision thus reducing the area of exposed tissue? We would then require a confirmatory study emulating practice in this country before adopting the routine antibiotic prophylaxis recommended by the authors. Brian Scott 1) Preclik G, Grune S, Leser HG, Lebherz J, Heldwein W, Machka K, Holstege A, Kern WV. Prospective, randomised, double blind trial of prophylaxis with single dose of co-amoxiclav before percutaneous endoscopic gastrostomy. BMJ 1999;319:881-4. |
|||