Synopsis  

This review emphasizes key issues for reducing the burden of risk inherently involved in all interventional procedures, and the need to ensure that patients are informed sufficiently to be able to participate in decisions about their care. Whilst each patient and each procedure may be unique, the principles of risk minimization and management are universal. Procedures must be done for appropriate indications, by well-trained endoscopists, assisted by expert staff, using the correct equipment. Known risk factors should be reviewed beforehand, and steps taken to accommodate them whenever necessary and possible. Patient consent exists only when there has been a meaningful educational process, adjusted appropriately to the individual patient and clinical situation. Unplanned events should be defined carefully and their occurrence monitored, so that standard methods for continuous quality improvement can be applied.

 Introduction   

Endoscopy ('looking inside') is invasive by definition, and therefore potentially hazardous. Therapeutic interventions add to the risk. The goal of everybody involved in an endoscopy procedure, not least the patient, is to maximize the benefits and to minimize the risks.

This contribution focuses on the potential risks of endoscopy in general. Specific risks of individual procedures (such as ERCP, and various therapeutic maneuvers) are covered in the relevant sections. However, the principles are the same. Endoscopists must understand the risks, the factors which affect them, the ways to reduce them, and how best to educate our patients about them. There have been many reviews of this topic [1–23].

 The contract with the patient; informed consent   

Endoscopy is normally conducted as part of a comprehensive evaluation by a gastroenterologist or by another digestive specialist. It is used mostly electively in the practice environment, or hospital outpatient clinic, but may sometimes be needed elsewhere (e.g. emergency room, intensive care unit). Sometimes, endoscopists offer an 'open access service' where the initial clinical assessment and continuing care is performed by another physician.

Responsibility   

In all of these situations it is the responsibility of endoscopists to assure themselves that the potential benefits of the proposed procedure exceed the potential risks, and to convey that information clearly to their patients. Truly informed consent means that the patient really does understand the potential risks and benefits, as well as the possible limitations and any available alternative approaches [24]. That is our contract with the patient. Signing an 'informed consent form' is a medico-legal requirement in many institutions, but this is nothing more than confirmation of the education process.

Educational materials   

Nothing can replace a detailed discussion between the endoscopist and the patient (and any accompanying persons), but this process can be enhanced with written, video-, or web-based educational materials. Suitable brochures are available from national organizations, and on many websites (e.g. http://www.ddc.musc.edu). They can be adapted for local conditions. Whatever the details of the education process, patients must be given the opportunity to ask questions of their endoscopist (and support staff) immediately before the procedure. The process of informed consent must be clearly documented, and preferably witnessed.

Humanity   

It is appropriate here to emphasize also the importance of simple courtesies and common humanity in dealing with our customers. What is familiar and routine to the endoscopist and staff may be viewed by patients as a major ordeal—especially by those unfortunate enough to experience a significant adverse event.

 What are 'risks' and 'complications'?   

Definitions   

Risks cannot be described, discussed, or monitored unless there is agreement about their definition. This has proved to be a challenge [25,26]. The word 'complication' has unfortunate medico-legal connotations. It suggests that something has 'gone wrong'. It seems preferable to use a more neutral term, like 'unplanned event', which simply indicates that the procedure has not gone completely 'according to plan', a subtle but important difference in emphasis. Some use the term 'adverse events' instead of complications. However, not all unplanned events are actually adverse; occasionally, things may go better than expected. The logic of using 'unplanned event' is that it ties directly in with the informed consent process. The truly informed patient knows what the plan is, and has specific expectations about what will happen before, during, and after the procedure, as well as the likely outcomes, both positive and negative.

Threshold for 'a complication'   

Some unplanned events are relatively trivial, e.g. the need for sedation reversal during or after a procedure, or self-limited bleeding at the site of an endoscopic incision. These events should be adequately documented, so that quality improvement processes can be applied, but they are not significant enough clinically to be called complications. Mostly, patients are unaware of such events. However, there is a level of severity at which an 'unplanned event' does become clinically important, and where 'complication' is a legitimate descriptor. The question is where to place the threshold, i.e at what level of severity does an unplanned event become a complication? We attempted to define that threshold in 1991 at a consensus conference discussing outcomes of ERCP [27]. We stated that 'a complication is an unplanned event, related to the procedure, that requires the patient to be admitted to hospital, or to stay in hospital longer than expected, or to undergo other unplanned interventions'. By this definition, we do not count in our statistics of 'complications' any deviation that occurs during a procedure but which is not obvious to the patient afterwards (e.g. transient bleeding), or one which can be treated on an outpatient basis (e.g. localized thrombophlebitis at an infusion site). This definition is not an attempt to hide such events, merely an effort to provide a common language. This is essential if we are to collect data, for, without data, we cannot study our outcomes or improve them.

Severity   

A complication is an unplanned event that reaches a certain level of severity. But not all complications are of equal importance. They vary from relatively minor episodes (e.g. one day of pancreatitis after ERCP) to those which are life threatening (e.g. perforation). It seems reasonable to stratify the severity of complications by the extent of the disturbance which they cause to the patient. The consensus conference settled on the following stratification of severity:

• Mild; unplanned events requiring hospitalization of 1–3 days.
  
• Moderate; needing 4–9 days in hospital.
  
• Severe; more than 10 days in hospital, or needing surgery, or intensive care.
  
• Fatal; death attributable to the procedure.
  

Attribution   

Use of the terms 'attributable' and 'related to the procedure' (in the earlier general definition) introduces a regrettable element of subjectivity, with the potential for reporting bias. That could be avoided by counting everything adverse that occurs within a certain period (say 30 days, as used to be the case after surgery). However, this would not be appropriate. No one would causally connect a simple diagnostic endoscopy with a myocardial infarction which occurred 29 days afterwards, in an elderly patient with known heart disease, especially if they had undergone multiple other medical interventions in the meantime. However, attribution would seem reasonable if the same event had occurred only two days after the endoscopy, especially if aspirin had been discontinued, or if the patient had experienced transient hypotension or hypoxia. Other issues with unplanned events and complications are whether they cause the procedure to be interrupted or aborted, and whether they leave any permanent sequelae.

Timing of unplanned events   

Most unplanned events occur during or immediately after the procedures. However, they can happen even before the endoscope is introduced (e.g. a reaction to prophylactic antibiotics or other preparation), during the procedure (e.g. transient bradycardia or hypoxia), immediately afterwards (e.g. pain due to perforation), a few hours later (e.g. pancreatitis after ERCP), or can be delayed for several days or weeks (e.g. aspiration pneumonia or delayed bleeding). Some events (e.g. viral transmission) may be so far delayed that the connection is difficult to make, or is missed completely. Keeping track of unplanned events which occur after patients leave the procedure unit is a challenge. Making routine follow-up phone calls is time-consuming, and not completely reliable. Perhaps automatic, computer-generated phone calls or e-mail will be useful in the future.

Direct and indirect events   

When considering complications, most endoscopists think first about the obviously related direct events (such as bleeding and perforation) that occur in organs which are being traversed or treated. However, there are also many indirect events, which occur outside the digestive tract. Cardiopulmonary complications (often related to sedation) are probably the commonest of the unplanned events. Renal, neurological, and musculo-skeletal complications have all occurred. Indirect events are more likely to occur after the patient leaves the procedure unit, and are thus more often overlooked and under-reported.

Data set for unplanned events   

A proposed data set for documentation of unplanned events, their consequences, and their severity is shown in Fig. 1[28].

 General issues of causation and management   

The issues impacting each specific type of event are discussed in detail below. However, there are some general risk factors, and important points to be made about the recognition and management of unplanned events.

Unplanned events may occur because of:

• Poor technical or cognitive performance by the endoscopist or staff.
  
• Patients' 'fitness' or, really, 'unfitness'. This is a combination of, and interaction between the complexity/severity of the presenting complaint and the patients' comorbidities (including allergies).
  
• Non-compliance by patients and their helpers.
  

Technical and cognitive performance   

The skill of the endoscopist (both clinical and technical) is probably the most important single factor determining the likelihood of unplanned events. As for certain surgeries, there is now good evidence that better-trained and more experienced endoscopists have higher success rates and fewer complications [29]. This raises the important and controversial issues of training, assessment of competence, report cards, recredentialling, and regionalization of specialist services [30]. However, endoscopists do not work on their own. Their specialist nursing and technical associates have much responsibility for many aspects of patient safety. The most important areas for them are the preparation of the patient, and the equipment, sedation, and monitoring of patients during and after the procedures.

Fitness for procedures   

Defining a patient's 'fitness' for a procedure is a complex issue. Their 'unfitness', or degree of 'illness' is a combination of the presenting disease and any existing comorbidities (and their possible interactions, which are sometimes complex). As yet, there is no agreed and relatively simple score that can be used to compare the acuity of different patients and populations of patients undergoing endoscopic and related procedures.

ASA score   

The American Society of Anaesthesiology grading system is well validated for operative procedures requiring anaesthesia [31]. It has some correlation with the cardiopulmonary responses to sedation during endoscopy, but other specific issues are more important.

Other risk indices   

Many other risk indices are in use in other areas, e.g. APACHE, POSSUM, Charlson, but none measure the likely risk of endoscopic procedures [32–34]. Developing an overall (and procedure specific) risk score can be done only by collecting the data on large numbers of patients and procedures, and analysing the outcomes. This has been done in certain contexts, e.g. ERCP and sphincterotomy [28,35].

Prompt recognition and management   

The keys to effective management of complications are early recognition and prompt focused action. Delay is dangerous both medically and legally [36]. Patients in pain and distress after procedures should always be examined carefully, and never simply 'reassured' without careful evaluation.

Communication   

Poor communication is the basis for much unhappiness, and many lawsuits. Remember that the truly informed patient (and relatives) have been told already that complications can happen. This is an integral, important part of the consent process. So it is appropriate and correct to address suspected complications in that spirit. 'It looks as if we have a perforation here. We discussed that as a remote possibility beforehand, and I am sorry that it has occurred. Here is what I think we should do.'

Distress   

Your distress is understandable and worthy, and it is important to be sympathetic, but equally important to be professional and matter of fact. Excessive apologies may give the impression that some avoidable mishap has occurred. Never, never, attempt to cover up the facts.

Document   

Document what has happened and communicate widely, with the patient, interested relatives, referring physicians, supervisors, and your Risk Management advisors.

Act promptly   

Get appropriate laboratory studies and radiographs, and consult other experts in the relevant fields, including a surgeon for anything that might remotely require surgical intervention. Sometimes it may be wise to offer to transfer the patient's care to a specialty colleague, or to a larger medical centre, but if this happens, try to keep in touch, and to show continuing interest and concern. Apparent abandonment alienates patients and their relatives, and may lead to initiation of legal action.

Specific unplanned events   

Whilst cardiopulmonary events related to sedation are probably the commonest unplanned events, medico-legal claims arise more often from perforation and from failure to diagnose [37]. It is therefore smart to emphasize both of these risks in the informed consent process.

Failure to diagnose   

Missing, or apparently missing, cancer and precursors is a particular risk in colonoscopy, but can occur in other organs [38]. Endoscopists have been faulted for failing to follow suspicious lesions (e.g. gastric ulcers), even when poor compliance of the patient has been the main issue.

Perforation   

Perforation is the most feared complication of endoscopic procedures, because the consequences may be severe, and because its occurrence suggests (but does not prove) imperfect practice [39–43]. It can occur anywhere that endoscopes travel. It may be caused by the endoscope tip, or by pressure from the shaft in a tight loop, or by therapeutic dilatation or incision.

Risk factors   

The risk of esophageal perforation is greater in the elderly, particularly in the presence of a Zenker's diverticulum. It is also markedly increased during esophageal dilatation, especially in patients with malignancy or achalasia [44]. Perforation of the stomach or duodenum is very unusual in patients without major focal pathology. Retroperitoneal perforation occurs after endoscopic sphincterotomy, particularly when the needle-knife or precut technique is used by inexperienced operators. Perforation of the afferent loop is a possibility during manipulations after Billroth II gastrectomy. Colonic perforation is the most severe complication of colonoscopy. The risk may be increased by focal disease, such as diverticula or tumours.

Recognition   

Early recognition of perforation is essential. The endoscopic view may leave no room for doubt. Pain and distress are the hallmark symptoms. Patients with esophageal perforation may develop subcutaneous emphysema. Perforation at colonoscopy is often associated with dramatic abdominal distension. Tachycardia is common. Steps should be taken immediately to clarify the situation. Plain radiographs usually are diagnostic, but CT scanning is more sensitive and should be carried out quickly if perforation is suspected and standard radiographs are negative or equivocal. The retroperitoneal nature of perforation after endoscopic sphincterotomy may delay recognition and has few specific signs. The patient may appear to have pancreatitis. Severe pain without impressive elevation of serum amylase or lipase is suggestive of perforation, and early CT scan is advised.

Treatment   

Treatment of perforation is sometimes controversial. Careful assessment, review of the available literature, and surgical advice are essential. Whilst prompt surgical intervention might seem to be the obvious solution (especially to surgeons), it is not always necessary. Intra-abdominal perforations are almost always treated surgically, although a few selected cases have been managed conservatively, and by endoscopic clipping or sewing. Many esophageal and most sphincterotomy perforations have been treated conservatively (nil by mouth, intravenous fluids, antibiotics, and sometimes with targeted drainage tubes). Collaborative management with specialist (surgical) colleagues is strongly recommended whatever strategy is proposed, with frequent review—and much communication with the patient and anxious relatives.

Bleeding   

Bleeding may occur due to endoscopic manipulations (e.g. biopsy, polypectomy, or sphincterotomy), or can arise during procedures from existing lesions (e.g. an ulcer with a visible vessel), or occasionally due to retching (Mallory Weiss tear). The risk of bleeding is a balance between the technique of the endoscopist, the specific clinical lesion carried by the patient, and by the coagulation status.

Risk factors   

Patients with poor coagulation and/or portal hypertension are clearly at increased risk. Coagulopathy should be normalized or improved wherever possible, preferably to achieve coagulation parameters like those usually accepted for percutaneous liver biopsy. Anticoagulation should be stopped ahead of time, and (if clinically necessary) replaced temporarily by intravenous heparin for the procedure, and during the phase of early recovery. The use of antiplatelet drugs also increases the risk of bleeding, although the extent is difficult to determine [45]. There is little evidence to support the widely held belief that aspirin and non-steroidal anti-inflammatory drugs (NSAIDS) increase the risk, but many endoscopy units recommend that these agents be discontinued for five days or more before endoscopy (and for a week or so afterwards). It used to be common practice to measure coagulation parameters in all patients scheduled for therapeutic endoscopy. However, this practice is not supported by data, nor by the recent guidelines from the ASGE. A careful personal and family history of bleeding problems is more discriminating [46].

Endoscopic incisions (e.g. for polypectomy or sphincterotomy) should be performed in a controlled manner, to provide sufficient coagulation of local vessels. The addition of epinephrine to a 'saline cushion' may be helpful when resecting sessile polyps. Polyps with large stalks often have large vessels. The risk of bleeding may be reduced by injecting epinephrine or sclerosant into the stalk, or by deploying a detachable loop. The risk of inducing immediate or delayed bleeding from lesions which have themselves recently bled (e.g. ulcers or varices) depends on the size of the vessel and the pressure within it, and on the precise endoscopic technique. Endoscopic ultrasound and Doppler techniques have been used to detect significant submucosal vessels.

Recognition   

Recognition of bleeding is usually obvious, during or after endoscopy. Occasional patients may present later only with anaemia. Delayed bleeding should not be attributed to the treated area without endoscopic confirmation.

Treatment   

Most endoscopically induced bleeding can be treated by standard medical means, with repeat endoscopic assessment and treatment when bleeding is persistent or severe. Expert angiographic investigation and treatment has been used successfully. Surgery is needed very rarely.

Cardiopulmonary and sedation complications   

Cardiopulmonary problems and adverse reactions to sedation are the commonest cause of significant unplanned events attributable to endoscopy [47–66]. The stress of the procedures (and the presence of the tube itself) may provoke dysrhythmias or induce hypotension or hypoxia, especially in patients with established cardiac or pulmonary disease, and in the morbidly obese. Aggressive sedation (with inadequate monitoring) can be disastrous. The cardiopulmonary status of potential patients should be assessed carefully (and improved if possible), and the approach to sedation/anaesthesia should be cautious and intelligent. A history of toleration of past procedures may be helpful. Some patients can accept simple endoscopy procedures without any sedation or analgesia. Others with significant comorbidities, and predictable airway difficulties, are best managed by anaesthesiologists. Endoscopists and staff should be fully trained in sedation, monitoring, and resuscitation, with ready access to all necessary equipment. Cardiological consultation may be appropriate in patients with unusual rhythms (or medications), pacemakers and implantable defibrillators.

Infection   

Endocarditis   

Endoscopy can provoke bacteremia, especially during therapeutic procedures such as dilatation. This may be dangerous in a very small number of patients who are immunocompromised, or who have important cardiovascular lesions. Endoscopy-induced endocarditis is extremely rare (and therefore difficult to study). Hard facts are scarce, and there is certainly room for disagreements in practice [67]. Guidelines concerning the use of antibiotic prophylaxis have been developed by many cardiological and endoscopic authorities [68–71]. Most recommend giving antibiotics to certain 'high risk' patients undergoing the higher risk procedures, but leave much leeway for individual judgement.

Infections   

Infections can be transmitted in the endoscopy unit [72]. Outbreaks of salmonella, shigella, pseudomonas, hepatitis and other infections have been reported [73]. Such patient-to-patient transmission should be prevented by assiduous attention to cleaning and disinfection regimes, and appropriate management of accessories, tissues, and fluids. Patient-to-staff transmission (e.g. of H. pylori) has probably occurred, when staff protection precautions are inadequate. Transmission of tuberculosis, and of prion diseases, is a cause for concern [73,75]. Staff-to-patient transmission is theoretically possible. Prevention of infection in the endoscopy unit clearly requires a comprehensive approach, involving appropriate 'universal precautions', assiduous cleaning and disinfection practice, and continuous monitoring. Selective immunization of staff is relevant.

Instrumentation   

Instrumentation can stir up dormant colonization, e.g. when contrast is injected under pressure into a contaminated biliary tree. The same is theoretically true when endoscopy manipulates other infected areas, such as a small peridiverticular abscess.

Allergic reactions   

Certain patients may react adversely to medications used during endoscopy procedures. Local anaesthetic agents, antibiotics, and iodine-containing contrast materials are most commonly cited. What to do about suspected iodine allergy before ERCP is controversial [76]. Some patients are allergic to latex rubber. Staff can become sensitized to gluteraldehyde. Obviously, a careful history of drug use and allergies is mandatory.

IV site issues   

Blemishes and local phlebitis are common. Spreading infection is very rare.

Miscellaneous and rare events   

There are many other theoretical and rare unplanned events. Patients may sustain injuries during sedation, by falling, or by pressure on nerves. Teeth have been damaged or lost, and more than one shoulder has been dislocated. Electrical injury can occur during diathermy, and there have been colonic explosions. Very rarely, endoscopes have become impacted, e.g. by retroversion up the esophagus, or have caused serious damage to adjacent structures, e.g. avulsion of the spleen.

 Preventing unplanned events   

The responsibility for safety (and any unplanned events) falls ultimately on the endoscopist, and there are very important issues of training, competence, and credentialling [76]. However, it is equally clear that the professionalism of nurses and other staff is crucial. Electronic monitoring of some aspects is helpful, but the personal involvement of educated staff is even more so. Emergency equipment must be readily available. Endoscopists should be trained in advanced cardiac life support techniques.

Endoscopy can never be completely safe. Keeping unplanned events to a minimum should be of the highest priority for the endoscopy team, and for the patient. Understanding the risks and risk factors is fundamental. Careful prospective monitoring of defined events allows continuous quality improvement.

 Outstanding issues and future trends   

Reducing the burden of unplanned events in the future will involve progress in many different areas, including:

• Reducing the use of endoscopy for purely diagnostic purposes, using better patient selection (e.g. by fecal DNA testing), and alternative non-invasive methods (e.g. virtual colonoscopy).
  
• Refining methods for therapeutic endoscopy.
  
• Improving the training of endoscopists and support staff, with objective testing of competence for credentialling, and recredentialling.
  
• Continuous quality improvement, through structured documentation of outcomes, and appropriate feedback.
  
• Making endoscopes easier to clean and disinfect, even sterilize.
  
• Fostering patient empowerment, through meaningful education and the widespread use of report cards.
  

 References  

 1 Arrowsmith, JB, Gerstman, BB, Fleische, DE & Benjamin, SR. Results from the American Society for Gastrointestinal Endoscopy/U.S. Food and Drug Administration collaborative study on complication rates and drug use during gastrointestinal endoscopy. Gastrointest Endosc 1991; 37 (4): 421–7. PubMed

 2 Carey, WD. (1987) Indications, contraindications, and complications of upper gastrointestinal endoscopy. In: Gastrointestinal Endoscopy (eds. MV Sivak). p. 301. W.B. Saunders, Philadelphia.

 3 Chan, MF. Complications of upper gastrointestinal endoscopy. Gastrointest Endosc Clin N Am 1996; 6 (2): 287–303. PubMed

 4 Cooper, GS & Blades, EW. (2000) Indications, Contraindications, and Complications of Upper Gastrointestinal Endoscopy, in Gastroenterologic Endoscopy. In: Gastrointestinal Endoscopy Vol 1 (eds. MV Sivak). pp. 438–54. W.B. Saunders, Philadelphia.

 5 Delvaux, M. The complications of digestive endoscopy. Gastroenterol Clin Biol 1995; 19: B23–32.

 6 Eimiller, A. Complication in endoscopy. Endoscopy 1992; 24 (1–2): 176–84. PubMed

 7 Eisen, GM, Baron, TH, Dominitz, JA, Faigel, DO, Goldstein, JL & Johanson, JF et al. Complications in Upper GI Endoscopy. Gastrointest Endosc 2002; 55 (7): 784–93. PubMed

 8 Freeman, ML, Nelson, DB, Sherman, S, Haber, GB, Herman, ME & Dorsher, PJ et al. Complications of endoscopic biliary sphincterotomy. N Engl J Med 1996; 335: 909–18. PubMed

 9 Froehlich, F, Gonvers, JJ, Vader, JP, Dubois, RW & Burnand, B. Appropriateness of gastrointestinal endoscopy: risk of complications. Endoscopy 1999; 31 (8): 684–6. PubMed

10 Hart, R & Classen, M. Complications of diagnostic gastrointestinal endoscopy. Endoscopy 1990; 22: 229–33. PubMed

11 Kavic, SM & Basson, MD. Complications of endoscopy. Am J Surg 2001; 181 (4): 319–32. PubMed

12 Keeffe, EB. (1991) Endoscopic procedural safety. In: Quality Control in Endoscopy (ed. R Mccloy) pp. 33–45. Springer-Verlag, Berlin.

13 Muhldorfer, SM, Kekos, G, Hahn, G & Ell, C. Complications of therapeutic gastrointestinal endoscopy. Endoscopy 1992; 24: 276–83. PubMed

14 Newcomer, MK & Brazer, SR. Complications of upper gastrointestinal endoscopy and their management. Gastrointest Endosc Clin North Am 1994; 4 (3): 551–70.

15 Quine, MA, Bell, GD, McCloy, RF, Charlton, JE, Devlin, HB & Hopkins, A. A prospective audit of upper gastrointestinal endoscopy in two regions of England: safety, staffing and sedation methods. Gut 1995; 36: 462–7. PubMed

16 Rankin, GB & Sivak, MV. (2000) Indications, contraindications, and complications of colonoscopy. In: Gastroenterologic Endoscopy, Vol. 2 (eds. MV Sivak). pp. 1222–52. W.B. Saunders, Philadelphia.

17 Reiertsen, O, Skjoto, J, Jacobsen, CD & Rosseland, AR. Complications of fiberoptic gastrointestinal endoscopy—five years' experience in a central hospital. Endoscopy 1987; 19 (1): 1–6. PubMed

18 Rothbaum, RJ. Complications of pediatric endoscopy. Gastrointest Endosc Clin N Am 1996; 6 (2): 445–59. PubMed

19 Schauer, PR, Schwesinger, WH, Page, CP, Stewart, RM, Levine, BA & Sirinek, KR. Complications of surgical endoscopy. A decade of experience from a surgical residency training program. Surg Endosc 1997; 11 (1): 8–11. PubMed

20 Schmitt, CM. Procedure-specific outcomes assessment for esophagogastroduodenoscopy. Gastrointest Endosc Clin N Am 1999; 9 (4): 609–24, vi–vii. PubMed

21 Silvis, SE, Nebel, O, Rogers, G, Sugawa, C & Mandelstam, P. Endoscopic complications. Results of the 1974 American Society for Gastrointestinal Endoscopy Survey. JAMA 1976; 235 (9): 928–30. PubMed

22 Taylor, MB. (1992) Complications of upper endoscopy and colonoscopy. In: Gastrointestinal Emergencies (eds. MB Taylor). pp. 551–6. Williams & Wilkins, Baltimore.

23 Zubarik, R, Eisen, G, Mastropietro, C, Lopez, J, Carroll, J, Benjamin, S & Fleischer, DE. Prospective analysis of complications 30 days after outpatient upper endoscopy. Am J Gastroenterol 1999; 94 (6): 1539–80. PubMed PubMed

24 Plumeri, PA. Informed consent for upper gastrointestinal endoscopy. Gastrointest Endosc Clin North Am 1994; 4 (3): 455–62.

25 Cotton, PB. Outcomes of endoscopy procedures: struggling towards definitions. Gastrointest Endosc 1994; 40: 514–18. PubMed

26 Fleischer, DE. Better definition of endoscopic complications and other negative outcomes. Gastrointest Endosc 1994; 40: 511–14. PubMed

27 Cotton, PB, Lehman, G, Vennes, J & Geenen, JE et al. Endoscopic sphincterotomy complications and their management: an attempt at consensus. Gastrointest Endosc 1991; 37: 383–93. PubMed

28 Cotton, PB. Income and outcome metrics for the objective evaluation of ERCP and alternative methods. Gastrointest Endosc 2002; 56 (6): S283–90. PubMed

29 Freeman, ML. Training and competence in gastrointestinal endoscopy. Rev Gastroenterol Disord 2001; 1 (2): 73–86. PubMed

30 Cotton, PB. How many times have you done this procedure, Doctor? Am J Gastroenterol 2002; 97: 522–3. PubMed

31 Keat, AS. The ASA classification of physical status: a recapitulation. Anesthesiology 1978; 49: 233–7. PubMed

32 Knaus, WA, Wagner, DP, Draper, EA, Zimmerman, JE, Bergner, M & Bastos, PG et al. The APACHE III prognostic system: risk prediction of hospital mortality for critically ill hospitalized adults. Chest 1991; 100: 1619–36. PubMed

33 Charlson, ME, Pompei, P, Ales, KL & Mackenzie, CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chron Dis 1987; 40: 373–83. PubMed

34 Prytherch, DR, Whiteley, MS, Higgins, B, Weaver, PC, Prout, WG & Powell, SJ. POSSUM and Portsmouth POSSUM for predicting mortality. Br J Surg 1998; 85: 1217–20. PubMed

35 Freeman, ML, DiSario, JA, Nelson, DB, Fennerty, MB, Lee, JG & Bjorkman, DJ et al. Risk factors for post-ERCP pancreatitis: a prospective, multicenter study. Gastrointest Endosc 2001; 54: 425–34. PubMed

36 American Society for Gastrointestinal Endoscopy ad hoc Risk Management Committee. (2001). Risk management for the GI endoscopist.

37 Gerstenberger, PD & Plumeri, PA. Malpractice claims in gastrointestinal endoscopy: analysis of an insurance industry data base. Gastrointest Endosc 1993; 39: 132–8. PubMed

38 Plumeri, PA. The failure to diagnose colon cancer. A potential medical-legal disaster. Gastrointest Endosc Clin North Am 1993; 3 (4): 749–61.

39 Adamek, HE, Jakobs, R, Dorlars, D, Martin, WR, Kromer, MU & Riemann, JF. Management of esophageal perforations after therapeutic upper gastrointestinal endoscopy. Scand J Gastroenterol 1997; 32: 411–14. PubMed

40 Fernandez, FF, Richter, A, Freudenberg, S, Wendt, K & Manegold, BC. Treatment of endoscopic esophageal perforation. Surg Endosc 1999; 13: 962–6. PubMed

41 Enns, R & Branch, MS. Management of esophageal perforation after therapeutic upper gastrointestinal endoscopy. Gastrointest Endosc 1998; 47 (3): 318–20. PubMed

42 Moghissi, K. Instrumental perforations of the oesophagus. Br J Hosp Med 1988; 39 (3): 231–6. PubMed

43 Quine, MA, Bell, GD, McCloy, RF & Matthews, HR. Prospective audit of perforation rates following upper gastrointestinal endoscopy in two regions of England. Br J Surg 1995; 82: 530–3. PubMed

44 Clouse, RE. Complications of endoscopic gastrointestinal dilation techniques. Gastrointest Endosc Clin N Am 1996; 6 (2): 323–41. PubMed

45 Miller, AM, McGill, D & Bassett, ML. Anticoagulant therapy, anti-platelet agents and gastrointestinal endoscopy. J Gastroenterol Hepatol 1999; 14: 109–13. PubMed

46 Eisen, GM, Baron, TH, Dominitz, JA, Faigel, DO, Goldstein, JL & Johanson, JF et al. Guideline on the management of anticoagulation and antiplatelet therapy for endoscopic procedures. Gastrointest Endosc 2002; 55 (7): 775–9. PubMed

47 Lipper, B, Simon, D & Cerrone, F. Pulmonary aspiration during emergency endoscopy in patients with upper gastrointestinal haemorrhage. Crit Care Med 1991; 19: 330–3. PubMed

48 Benjamin, SB. Complications of conscious sedation. Gastrointest Endosc Clin N Am 1996; 6 (2): 277–86. PubMed

49 Daneshmend, TK, Bell, GD & Logan, RFA. Sedation for upper gastrointestinal endoscopy: results of a nationwide survey. Gut 1991; 32: 12–15. PubMed

50 Eckardt, VF, Kanzler, G, Schmitt, T, Eckardt, AJ & Bernhard, G. Complications and adverse effects of colonoscopy with selective sedation. Gastrointest Endosc 1999; 49 (5): 560–5. PubMed

51 Freeman, ML. Sedation and monitoring for gastrointestinal endoscopy. Gastrointest Endosc Clin N Am 1994; 4 (3): 475–99. PubMed

52 Froehlich, F, Gonvers, JJ & Fried, M. Conscious sedation, clinically relevant complications and monitoring of endoscopy: results of a nationwide survey in Switzerland. Endoscopy 1994; 26 (2): 231–4. PubMed

53 Holm, C & Rosenberg, J. Pulse oximetry and supplemental oxygen during gastrointestinal endoscopy: a critical review. Endoscopy 1996; 28 (8): 703–11. PubMed

54 Keeffe, EB & O'Connor, KW 1989 A/S/G/E survey of endoscopic sedation and monitoring practices. Gastrointest Endosc 1990; 36: S13–18. PubMed

55 Lazzaroni, M & Porro, GB. Preparation, Premedication, Surveillance Endoscopy 2001; 33 (2): 103–8.

56 McCloy, R. (1997) Safety, sedation and emergencies in endoscopy. In: Practical Endoscopy (eds. M Shephard and J Mason). p. 139. Chapman & Hall, London.

57 American Society for Gastrointestinal Endoscopy. Sedation and monitoring of patients undergoing gastrointestinal endoscopic procedures. Gastrointest Endosc 1995; 42 (6): 626–9. PubMed

58 Sieg, A, Hachmoeller-Eisenbach, U & Heisenbach, T. How safe is premedication in ambulatory endoscopy in Germany? A prospective study in gastroenterology specialty practices. Dtsch Med Wochenschr 2000; 125 (43): 1288–93. PubMed

59 The American Society for Gastrointestinal Endoscopy. Preparation of patients for gastrointestinal endoscopy. Gastrointest Endosc 1998; 48: 691–4. PubMed

60 Wong, RC. The menu of endoscopic sedation: all-you-can-eat, combination set, a la carte, alternative cuisine, or go hungry. Gastrointest Endosc 2001; 54 (1): 122–6. PubMed

61 Zuccaro, G Jr Sedation and sedationless endoscopy. Gastrointest Endosc Clin N Am 2000; 10 (1): 1–20. PubMed

62 Banks, MR, Kumar, PJ & Mulcahy, HE. Pulse oximetry saturation levels during routine unsedated diagnostic upper gastrointestinal endoscopy. Scand J Gastroenterol 2001; 36: 105–9. PubMed

63 Bell, GD. Premedication, preparation and surveillance. Endoscopy 2000; 32: 92–100. PubMed

64 Bell, GD & Jones, JG. Routine use of pulse oximetry and supplemental oxygen during endoscopic procedures under conscious sedation: British beef or common sense? Endoscopy 1996; 28 (8): 718–21. PubMed

65 Charlton, JE. Monitoring and supplemental oxygen during endoscopy. Br Med J 1995; 310: 886–7.

66 Sandlin, D. Capnography for nonintubated patients: the wave of the future for routine monitoring of procedural sedation patients. J Perianesth Nurs 2002; 17 (4): 277–81. PubMed

67 Roduit, J, Jornod, P, Dorta, N, Blum, AL & Dorta, G. Antibiotic prophylaxis of infective endocarditis during digestive endoscopy: over- and underuse in Switzerland despite professed adherence to guidelines. Endoscopy 2002; 34 (4): 322–4. PubMed

68 American Society for Gastrointestinal Endoscopy. Antibiotic prophylaxis for gastrointestinal endoscopy. Gastrointest Endosc 1995; 42 (6): 630–4. PubMed

69 Antibiotic prophylaxis in gastrointestinal endoscopy.http://www.bsg.org.uk

70 Dajani, AS, Taubert, KA, Wilson, W, Bolger, AF, Bayer, A & Ferrieri, P et al. Prevention of bacterial endocarditis. Recommendations by the American Heart Association. Circulation 1997; 96: 358–66. PubMed

71 Antibiotic prophylaxis for gastrointestinal endoscopy.http://www.asge.org

72 Nelson, DB. Infection control during gastrointestinal endoscopy. J Laboratory Clin Med 2003; 141: 159–67.

73 Ponchon, T. Transmission of hepatitis C and prion diseases through digestive endoscopy: evaluation of risk and recommended practices. Endoscopy 1997; 29: 199–202. PubMed

74 Ramakrishna, B. Safety of technology: infection control standards in endoscopy. J Gastroenterol Hepatol 2002; 17: 361–8. PubMed

75 Working Party for E/S/G/E. ESGE guidelines for the prevention of endoscopic transmission of type C hepatitis and update in Creutzfeldt–Jakob disease. Endoscopy 1997; 29: 203–4. PubMed

76 Draganov, P & Cotton, PB. Iodinated contrast sensitivity in ERCP. Am J Gastroenterol 2000; 95 (6): 1398–401. PubMed

77 Neale, G. Reducing risks in gastroenterological practice. Gut 1998; 42(1): 139–42. PubMed

Copyright © Blackwell Publishing, 2004