Survey of current practices among members of CAG in the follow-up of patients diagnosed with gastric ulcer

Can J Gastroenterol Vol 13 No 6 July/August 1999 489 Department of Community Health Sciences, University of Calgary, Calgary, Alberta Correspondence: Dr L Sutherland, Professor and Head, Department of Community Health Sciences, University of Calgary, 3330 Hospital Drive NW, Calgary, Alberta T2N 4N1. Telephone 403-220-7181, fax 403-270-7307, e-mail lsutherl@ucalgary.ca 11th INTERNATIONAL COURSE ON THERAPEUTIC ENDOSCOPY

I t has been recommended practice for a number of years that patients found to have a gastric ulcer at endoscopy should have multiple biopsies taken from the ulcer margin or base (1).It is further recommended that, following a six-to eight-week course of ulcer healing therapy, the endoscopy should be repeated and further biopsies be taken from the affected area to exclude occult malignancy (2)(3)(4).Furthermore, repeat endoscopy should be continued until healing occurs.It is suggested that such a practice leads to the detection of early gastric cancers at a treatable stage.The merits of this practice have been questioned by a number of investigators (5)(6)(7)(8)(9).However, international guidelines still recommend repeat endoscopy in the follow-up of all cases of gastric ulcer (10,11).
There are many reasons why endoscopists may elect not to follow-up gastric ulcers endoscopically.The lesion may appear obviously benign.Ulcers in the body of the stomach and fundus tend to arouse more suspicion (4).There may be associated nonsteroidal anti-inflammatory drug (NSAID) use or helicobacter infection, or the patient's age or medical condition may dissuade the endoscopist from performing further invasive tests.In some cases the patient may relocate or default from follow-up.
We examined data from the Alberta Endoscopy Project to identify patients with gastric ulcer and follow-up endoscopy with a view to validating or refuting this report (12).During the three-year study period, gastric ulcer was noted in 1597 patients.However, only 684 of these patients underwent a second endoscopy (42.8%) (95% CI 40.4% to 45.3%).In those who did have a second procedure, the apparent reason for repeat endoscopy was frequently to obtain biopsies that could not be taken initially because of recent hemorrhage.This suggests a change in practice that has important implications with regard to existing guidelines and cost issues.
The primary aim of our study was to assess, via a survey, the proportion of members of the Canadian Association of Gastroenterology (CAG) performing repeat endoscopy in patients with gastric ulcer.We also wished to ascertain the factors influencing their practice in relation to the endoscopic follow-up of such patients.Data were also recorded on Helicobacter pylori evaluation and treatment in this setting.

MATERIALS AND METHODS
A cross-sectional survey of physician members of CAG was performed.A study package containing a covering letter, self-administered questionnaire and return-addressed, stamped envelope was mailed to each individual inviting their participation in the study.Questionnaire: The questionnaire was developed by the authors (Appendix A) and was pilot tested among local Calgary gastroenterologists to assess comprehension and content.The physician was specifically asked to report his or her actual practices rather than what he or she perceived practice to be elsewhere.
The questionnaire recorded age, years since primary qualification as a physician and the type of hospital in which they practised (teaching versus nonteaching).Those who did not perform any upper gastrointestinal endoscopies were requested to return the form without further completion.Otherwise, the respondent was asked to indicate the proportion of patients with gastric ulcer in whom they repeated the endoscopy.Four possible answers were provided: 95% to 100% (termed for the purpose of analysis as 'almost all'), 75% to 94%, 50% to 74% and below 50% (collectively termed for the purpose of analysis as 'less than 95%').Following this, they were asked to choose from a list of factors that might influence them to perform follow-up endoscopy in this population.This was followed by a list of factors that might influence them not to repeat the procedure.They were then asked to indicate whether they had changed their practice recently and what factors had influenced any change.Last, a series of questions regarding the diagnosis and treatment of H pylori and gastric ulcer disease was posed.Design and statistical considerations: Members of CAG formed the study population.The survey was a crosssectional questionnaire-based descriptive study.Results were expressed as the proportion of respondents (and confidence interval) expressing a particular view.Where appropriate, a two-tailed t test, Student's t test and c 2 tables were used for comparisons between groups.A Bonferroni correction was applied to all P values less than 0.05 to avoid introducing a type I error due to multiple comparisons.

RESULTS
Three hundred and ninety-seven questionnaires were mailed to eligible subjects.Responses were received from 241 individuals constituting a 60.7% response rate (95% CI 55.7 to 65.5%).Fourteen individuals did not perform upper gastrointestinal endoscopy.Thus, the study population comprised 227 individuals.The mean age of respondents was 48.6 years (95% CI 47.4 to 47.9 years), and the mean number of years since primary qualification as a physician was 24.5 years (95% CI 23.2 to 25.7).Two hundred were practising as gastroenterologists or as internists with an interest in gastroenterology.One respondent did not indicate his specialty.The remainder were pediatric gastroenterologists (13 respondents), hepatologists (three respondents), surgeons (nine respondents) or diagnostic radiologist (one respondent).The small numbers in these groups precluded meaningful comparisons for interspecialty differences in practice.
Of the respondents, 57.3%, 22.5%, 9.2% and 11.0% indicated that they performed follow-up endoscopy in 95% to 100%, 75% to 94%, 50% to 74% and fewer than 50% of patients with gastric ulcer, respectively.There was no significant difference in practices between physicians working in teaching and nonteaching hospitals (P=0.88).
The subsequent analyses compare practices between those performing repeat endoscopy in the 95% to 100% group ('almost all patients') to the remainder (fewer than 95%).The mean ages of physicians in the two groups were very similar (48.5 versus 48.7 years, respectively; P=0.86).There was no significant difference in the length of time in medical practice between the two groups (24.3 versus 24.7 years; P=0.75).In the group performing repeat endoscopy in almost all patients the most common reasons influencing this choice were to ensure healing (86.9%), to confirm the benign nature of the lesion (80.8%), a history of complicated ulcer (39.3%) and a family history of cancer (40%).In those performing repeat endoscopy in fewer than 95% of patients with gastric ulcer, the most common reasons for carrying out a follow-up procedure were the confirmation of healing (68%), the confirmation of the benign nature of the lesion (66%), a history of complicated ulcer (55.7%) and the ulcer site (38%) (Figure 1).
In those performing a repeat procedure in 'almost all patients' with gastric ulcer, patient ill health (50%) and a history of NSAID use (36.9%) were the most common reasons for not repeating the procedure.In those performing followup in fewer than 95% the most common factors discouraging the physician from repeating the endoscopy were NSAID use (81.4%), patient ill health (60%) and young patients (54.6%) (Figure 2).
Of the respondents, 21.5% indicated that they had changed their practice recently.Of these, 95.6% attributed this to the role of H pylori.In excess of 97% of respondents indicated that they routinely check for H pylori infection in patients with gastric ulcer, with 99.5% indicating that they would prescribe eradication therapy in this setting.However, only 35.4% would routinely confirm eradication.In this subgroup, 64.8% indicated that they would use endoscopic-based diagnostic methods for this purpose, and 57.9% indicated that they would use the urea breath test.

DISCUSSION
This cross-sectional survey indicates that approximately 57% of the physician membership of CAG report that they invariably perform follow-up endoscopy in patients found to have a benign gastric ulcer at an initial procedure.This practice is not related to the age of the physician, the setting in which he or she practises or the length of time that the physician is in practice.Confirmation of healing was the most commonly cited reason for repeating endoscopy among all those surveyed.
There are many possible reasons underlying this divergence from practice guidelines.The importance of repeating endoscopy in patients with gastric ulcer to rule out occult malignancy is a divisive issue.A number of studies and case reports have highlighted the risk of occult gastric cancers occurring in apparently benign ulcers (4,(13)(14)(15)(16)(17)(18).However, several studies suggest that the risk of gastric cancer in apparently benign gastric ulcer is very low (5)(6)(7)9).In a large study of 1812 Japanese patients with gastric ulcers, 38 patients developed gastric cancer a mean of 6.1 years after the initial diagnosis of gastric ulcer (8).Only nine of the 31 cases where the cancer site was known developed in the site of the original ulcer.This suggests that the incidence of gastric cancer occurring in the actual site of a benign ulcer is very low even in a population with a high risk of gastric cancer.
It appears that an experienced endoscopist's macroscopic impression combined with adequate biopsies and/or brush cytology has a high sensitivity for detecting malignant transformation in gastric ulcers (6).Some authors report sensitivities approaching 100% when seven to 10 ulcer biopsies are taken (19,20).However, Bytzer et al ( 21) reported a 1.1% prevalence of gastric cancer among 773 patients with an apparently initial benign ulcer who were followed until healing.In six of 10 of these the malignant transformation did not become manifest until a range of two to 10 years after the initial diagnosis (21).The authors report that 250 endoscopies would have to be performed to detect one curable cancer.In the Canadian health care system this would cost $140,000 dollars per cancer diagnosed.
Gastric malignancies can ulcerate.However, whether benign ulcers become malignant is controversial.H pylori is regarded as a class 1 gastric carcinogen (22).Whether eradication of H pylori reduces the risk of malignant transformation in gastric ulcers is unknown.Successful eradication of the bacterium, in the setting of H pylori-associated gastric ulcer, reduces recurrence rates at one year to 0% to 8%.This suggests that for these patients confirmation of eradication by urea breath testing may be sufficient follow-up (23).In our survey, approximately 20% of respondents indicated that they had recently changed their practice in the follow-up of gastric ulcers as a result of the role played by H pylori.
Patients developing gastric ulcers while taking NSAIDs appear to have a lower risk of developing gastric ulcer-associated malignancy (21).Among those surveyed in the present study, 56.1% indicated that a history of NSAID use would influence them not to repeat endoscopy.
The performance of repeat endoscopies for gastric ulcer healing constitutes 10% of the workload in some hospitals (21).This is an area where workload could be reduced.In an era in which the cost of health care is important, prospective studies on the natural history of the subtypes of gastric ulcer are warranted, taking into account the role of NSAID use and helicobacter infection.In the meantime, Canadian gastroenterologists vary widely in their individual practices.
The possibility of response bias affecting the outcome of our study cannot be discounted.Another dilemma of any survey on physician practice is whether the responses reflect actual clinical practice.Each physician in our study was requested to complete the questionnaire according to his or her actual practices rather than how he or she perceived a practice should be.There is evidence from other studies that self-reported behaviours by physicians are a valid measure of actual behaviour (24).We sought to confirm the accuracy of their self reporting by comparing the responses for a subset of individuals for whom we had in-practice evidence of their follow-up of gastric ulcers in the Alberta Endoscopy project.All of the physicians overestimated the proportion of endoscopies they repeated in this setting.However, the database may have underestimated this proportion because any gastric ulcers diagnosed during the initial phase of data entry would have been regarded as first time diagnoses, whereas a proportion of them would have been follow-up procedures.In addition, approximately half of the ulcers occurred in the setting of NSAID use; 82% of physicians who responded to the present survey and who took part in the Alberta Endo-scopy Project indicated that this would influence against repeat endoscopy.

CONCLUSIONS
Fifty seven per cent of CAG members indicate that they repeat endoscopy in virtually all patients with a benign gastric ulcer.The primary reasons for doing so include ruling out occult malignancy and confirmation of healing.In those who repeat endoscopy in fewer than 95% of cases the most common reasons for not performing endoscopy are a history of NSAID use and patient ill health.Comparison of the results of our study with in-practice data from the Alberta Endoscopy Project suggests a discrepancy between actual and reported practice.This discrepancy may be best resolved by a prospective study on physician practice in this area.It appears that many of the malignancies occurring in patients with gastric ulcers occur after a long period of nonhealing.The eradication of H pylori may be of benefit in this regard.

Figure 1 )Figure 2 )
Figure 1) Factors influencing physicians in favour of repeat endoscopy in benign gastric ulcer.Benign?To confirm benign nature; bleed History of bleeding; family ca Family history of cancer; Hp Helicobacter pylori infection; personal exp Personal experience