Article Text


How should endoscopic accessories be selected: trial or error?
  1. M A NARAIN,
  1. University Hospital Birmingham
  2. Selly Oak Hospital
  3. Raddlebarn Road
  4. Birmingham B29 6JD, UK

    Statistics from

    See article on page 395

    Before being put on the market, medicines undergo stringent evaluation of their efficacy and safety, satisfying the regulatory bodies of individual countries. The costs of these activities are borne by manufacturers or developers who may later be held responsible for subsequent shortcomings. By contrast, medical devices, as for example those used in endoscopic procedures, require no such prior assessment in many countries. In the United Kingdom, since June 1998, all new products must satisfy the requirements of the Medical Devices Agency,1 the UK Competent Authority, to be awarded the CE mark allowing use throughout the European Community. Endoscopists and assistants commonly learn of new products from commercial stands at exhibitions held alongside scientific meetings or from manufacturers' representatives or sales staff. The apparent virtues and novelty of products are exhorted whereas objective clinical evidence of efficacy or safety is usually not available. Phase III development is not a clearly defined prerequisite before device marketing. Why should devices be less well investigated and evaluated than drugs when offered for sale? The initiative and ingenuity of manufacturers in developing and producing new equipment is to be warmly commended, but enthusiasm for sales ahead of rigorous clinical evaluation must be seriously questioned.

    Endoscopists are confronted with an overwhelming range of equipment from competing companies and at present choices are based upon previous experience, recommendations from colleagues and an inspection of the product. This situation is unsatisfactory and devices must be subjected to the same rigorous evaluation as medicines. Who should organise such trials? Ideally, the manufacturers in necessary collaboration with clinicians, obeying the rules of controlled trials with ethical approval. The expense would be considerable but less than in the case of medicines. Moreover the time between development and commercial release would be similarly prolonged—but would this necessarily be a bad thing? Many products currently offered for sale would not reach the market place.

    One alternative approach is for trials to be organised by clinicians, perhaps with commercial sponsorship, for independent publication in peer-reviewed journals. The paper by England et al in this issue (see page 395) is an excellent example of such collaboration, investigating two designs of stents for palliation of malignant obstruction of the bile duct. The study was supported by the equipment manufacturers and the findings were analysed by a professional monitoring group according to European Good Clinical Practice guidelines.

    Bacterial colonisation and encrustation of plastic biliary stents has been shown by in vitro studies to be affected by the surface characteristics of the stent material and in in vivo studies by the structure of the stent used. Much interest has focused on the potential for reducing encrustation by removal of side holes from the stent design—for example, the “fir tree” or Tannenbaum stent. A previous non-randomised study showed significantly improved patency compared with the pig tail design.2 Patency rates for 12F and 10F straight stents are similar whereas 7F and 8F stents are more prone to occlusion.3-5 Use of plastic stents in the palliation of malignant obstructive jaundice has been challenged by the superior performance of large bore expanding metal stents with improved patency rates revealed in four prospective randomised trials.6-9However, at detection, the majority of such strictures are caused by pancreatic malignancy often with a median survival less than the median patency of plastic stents. Many of those who develop recurrent jaundice are not restented with patients' wishes and compliance rather than procedural logistics proposed as the major factors.10Diagnostic uncertainty at the outset reserves a role for an easily removable and exchangeable stent—factors inherent in plastic constructions. Whether the choice of stent affects overall survival is less clear.

    England et al have looked at a remaining issue of whether the 10F Teflon Tannenbaum stent or the more commonly used 10F polyethylene Cotton-Leung stent with side holes, achieved longer patency. In a well planned prospective multicentre study they randomised at deep cannulation 134 patients presenting with a malignant common duct stricture to palliation using either the Tannenbaum or the Cotton-Leung design. All patients received prophylactic antibiotics at stenting. Respective groups of 65 and 69 patients were similar in age, type of malignancy (over 60% pancreatic), length of stricture, and American Society of Anesthesiologists' classification. There was an overall generous use of needle knife papillotomy at 35%, but this was evenly distributed between the groups. No statistical difference was found in the time to restenting or death with jaundice. Median patency for Tannenbaum stents was 181 (95% CI 59 to 303) days and for Cotton-Leung stents 133 (92 to 174) days, and there was no statistical difference in the secondary end point of survival. The Kaplan-Meier analysis incorporated 42% of recruited patients as censored data, consisting of non-jaundiced patients who either died or withdrew from the study. Eight of 14 patients who remained jaundiced after first stenting died jaundiced without restenting. It is difficult to determine whether this poor uptake of restenting would have significantly affected the secondary end point of survival; five of these eight patients had received the Tannenbaum stent. Interestingly, a randomised prospective study comparing three patient groups who underwent either stent exchange for recurrent jaundice, three monthly prophylactic stent exchange or a single procedure with an expanding metal stent, showed no difference in overall survival in the three groups although the symptom-free survival was equally greater in the second two groups (p<0.05).9In the real world, however, the longer lasting stent is probably associated with increased survival, as was observed in a retrospective study, where the significantly increased survival in the metal stented group was associated with a lower percentage of patients dying with untreated occlusive jaundice (9%) when compared with 30% in the plastic stent group (p<0.001).10

    England et al's important work has shown that the more expensive Tannenbaum stent confers no significant advantages over the Cotton-Leung stent in the setting of a malignant biliary stricture where we believe, until adjuvant therapies significantly improve survival, a major role continues to exist for plastic stents. The relative performance of various designs in non-surgical candidates with a retained common duct stone merits further investigation. An independent group such as the Consumers Association, publishers of Which? magazine and the Drug and Therapeutics Bulletin, could become involved with evaluation of medical devices. Reviews of equipment used in many medical situations might be considered but the infrequency of reports in a given field would perhaps delay innovation. Manufacturers must grasp the nettle and justify claims for their products rather than passing responsibility for evaluation entirely to clinicians while reaping the commercial rewards.

    See article on page 395


    View Abstract

    Request permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Linked Articles