| Abstract|| |
The objective is to analyze the experience of using diagnostic and therapeutic ERCP and to attempt identifying the independent factors that predict positive and useful procedures. The medical records of 198 patients seen during the period 1991-1993 were retrospectively reviewed. There were 102 males and 96 females with a mean age of 45.2 SD + 16.7 years. The main indications for performing the procedure were jaundice, abdominal pain, nausea and/or fever. Laboratory tests showed elevated direct bilirubin in 50% of patients, deranged liver enzymes in 43%, and ESR was raised in 51% and fever was documented in 52% of patients. The cannulation of both ducts was successful in 93% of all procedures. The commonest finding was gallbladder and common bile duct stones (CBD). In 54 patients out of 64 with CBD stones, stones were removed endoscopically. Stepwise regressive analyzis has identified age above 40 years, presence of jaundice (total bilirubin > 2.5 mg/ml) and raised ESR (> 25 mm in the 1st hour) as significant factors that independently predict a positive ERCP that revealed findings previously unknown (useful procedure). Diagnostic and therapeutic ERCP is an integral part in management of biliary and pancreatic ducts disorders. The validity of clinical prediction role should be tested prospectively.
|How to cite this article:|
Al Quorain A, Ibrahim EM, Al Sultan AI. Diagnostic and therapeutic value of ERCP and prediction of outcome: A retrospective analysis. Saudi J Gastroenterol 1996;2:138-41
|How to cite this URL:|
Al Quorain A, Ibrahim EM, Al Sultan AI. Diagnostic and therapeutic value of ERCP and prediction of outcome: A retrospective analysis. Saudi J Gastroenterol [serial online] 1996 [cited 2020 Apr 1];2:138-41. Available from: http://www.saudijgastro.com/text.asp?1996/2/3/138/34018
Endoscopic retrograde cholangiopancreatography (ERCP) has become one of the most important diagnostic and therapeutic procedures of the upper gastrointestinal endoscopy. It has been first reported and then introduced in Japan in 1968 and 1970, respectively , . Since then the diagnostic and therapeutic applications of the ERCP have truly revolutionized the management of biliary tract and pancreatic diseases ,,, .
This paper analyzes our experience using ERCP during a three-year period (1991-1993). We also attempted to identify the most significant factors that predict positive studies.
| Patients and Methods|| |
The medical records of all the patients who underwent ERCP at King Fahd University Hospital, Al-Khobar, Saudi Arabia during 1991-1993 were retrieved and retrospectively analyzed.
Two hundred thirteen patients had undergone this procedure using the Olympus-side viewing duodenoscope (Olympus FIT 20) under basal sedation. The study involved only 198 patients, 102 males and 96 females with a mean age of 45.2 SD ± 16.7 (range 10-98) years. The main indications were jaundice, gallbladder disorders and upper abdominal pain. The main symptoms were abdominal pain, nausea/vomiting, anorexia and fever. Physical examinations revealed jaundice in 50%, upper abdominal tenderness in 43.9% and hepatomegaly in 18% of patients.
Comparison of proportions of categorized variables were made using chi-square tests with the Yates correction for continuity where appropriate, and the analysis of variance was used to compare means. Within the derivation set, all variables with a univariate significance level of < 0.01 were entered into a stepwise logistic regression procedure with forward selection to identify a subset of variables that were independent predictors of ERCP. Step selections were based on an asymptotic covariance approximation to maximum likelihood  . Interactions between variables were not tested. The accuracy of the logistic model in discriminating patients with and without a positive ERCP that revealed findings not previously known, in the derivation set was evaluated using the area under a receiver operating characteristic (ROC) curve  , determined by the method of maximum likelihood  . The standard error of each curve was calculated using the methods of Hanley and McNiel , . The discriminatory function of the prediction rule was also analyzed by the cross-tabulation of the potential predictive finding and the outcome by means of a two-by-two contingency table and therefore the sensitivity, specificity, and the total error rate were calculated  . The logistic rule was cross-validated using the "jackknife" technique on a randomly drawn population consisting of one-third of the original derivation set  . The procedure was repeated using different random samples drawn by the computer.
In all analyzes, two-sided P value of less than 0.05 was considered significant. The BMDP Statistical Software programs (PID. P2D, P4F, P7M, and PLR) were used to analyze the data  .
| Results|| |
The laboratory tests showed elevated bilirubin, mainly, direct in 50% followed by deranged liver function test in 43% of patients. Upper abdominal sonography was positive in 50% out of 198 patients. CT scan of the abdomen was requested for 51 patients and showed positive result in 26%. ESR and temperature were elevated in 104 patients (52.5%) for each.
The cannulation of both ducts was successful in 198 out of 213 patients (93%). The results are shown in [Table - 1]. Out of 64 with common bile duct (CBD) stones, 50 patients (78%) have been treated endoscopically with papillotomy and their stones were successfully removed, in some cases the balloon or basket were used. In four (29%) of the remaining 14 patients with CBD stones and minimal biliary stricture, stenting for biliary drainage was successful, later followed by removal of stones using balloon or basket. In the remaining 10 patients with large stones and stricture, the management was surgical after prior placement of stent and drainage. Patients with adenocarcinoma of the ampulla, cholangiocarcinoma and pancreatic tumor were treated palliatively with stenting for biliary drainage. Those with fistulae between CBD and duodenum, choledechal cysts, and leakage from cystic ducts were also managed surgically. Ascaris worms were removed from the CBD after papillotomy in one patient. Two patients with postbulbar tuberculous ulcer were treated with anti-tuberculous drugs. Sixty-nine patients (35%) showed no abnormality with the biliary tree or pancreatic ducts. ERCP procedure was safe with complications in only four patients, three had mild pancreatitis and one developed cholangitis.
The demographic, clinical and radiologic data of 198 adult patients (derivation set) who were subjected to ERCP, were used to construct and to cross-validate a model using stepwise logistic regression procedure. The stepwise model identified age above 40 years, presence of jaundice (total bilirubin > 2.5 mg/ml) and raised ESR (> 25 mm in the 1st hour) y as significant factors that independently predict a positive ERCP that revealed finding(s) not previously known. The area under the receiver operating characteristics was 0.83 (95 % CI, 0.79 to 0.91 %). Thus, the probability that the logistic rule would correctly identify a patient with useful ERCP was 81% and the overall sensitivity and specificity of the rule were 80% and 94%, respectively with a total error rate of 9.1 %.
When the data were re-analyzed, excluding cases with a few missing symptoms, signs, or co-morbidity data, the outcome remained essentially unchanged.
| Discussion|| |
ERCP has outgrown its early limited diagnostic role to encompass an ever increasing number of diagnostic and therapeutic applications. Considerable efforts to enhance the sensitivity of biliary and pancreatic cytology and biopsy in the diagnosis of biliary or pancreatic duct strictures have also been made , . The therapeutic application of interventional endoscopy of the biliary and pancreatic ducts has increased in the recent times , . Papillotomy is the important therapeutic endoscopic biliary procedure.
Retrospectively, analyzes of 198 conservative patients having diagnostic and therapeutic ERCP have been conducted; the results of the diagnostic ERCP [Table - 1] are comparable with other studies locally and internationally , . In our series choledocolithiasis was the most common reason for therapeutic endoscopy which has been the case in other reported studies  . The clearance rate for CBD stones varies between 82-90% in experienced centers , . In our study, 64 patients had CBD stones, of which 54 had their stones removed endoscopically (84%) using the balloon or basket in some cases. The remaining 10 patients had large stones complicated with stricture and thus they were subjected to surgical intervention.
We were able to remove Ascaris worms from the CBD in one patient after papillotomy. Two patients had postbulbar tuberculous ulceration which lead to benign biliary stenosis. These patients were treated with antituberculous drugs, one case was reported from this institution  .
The malignant biliary obstructions in our series have been managed palliatively with stenting and drainage.
Clinical prediction rules have been used to predict the presence of disease, or prognostic outcomes, among groups of patients. They were also intended to help physicians interpret clinical information, make more accurate estimates of disease likelihood, and assess the diagnostic values of various tests and procedures ,, . In our study, three factors were identified as independent predictors for a positive ERCP study that would reveal an abnormality that was not previously known or predicted.
Older age and jaundice have been previously shown to predict the presence of bile duct stones among patients undergoing laparoscopic cholecystectomy  . The authors also concluded that ERCP should be performed preoperatively in patients exhibiting one or more of the strong predictors shown that included age and jaundice among others  .
The independent predictive value of raised ESR was rather interesting. It is conceivable, however, that the predictive value of this indicator only reflected the fact that most of our patients had either gallbladder or CBD stones with associated cholangitis and therefore a raised ESR.
The combination of the three predictors together predicted a useful ERCP in 81 % of patients. However, the derived predictor variables may only reflect the nature of our study population. These variables should be considered when requesting the procedure and balancing its useful yield against its potential morbidity, mortality and cost. Nullification of other predictors can be only explained by their random existence in the population under study.
We conclude that diagnostic and therapeutic ERCP is an integral part of the management of patients with biliary and pancreatic ducts disorders.
The validity of the clinical prediction role, however, should be tested prospectively where the outcome of ERCP could be compared among two groups; one where the procedure is done (or not done) based on the endoscopist's discretion; the decision in the other group will be guided by the model.
| References|| |
|1.||McCurie WS, Shorb Pe, Moscovitz H. Endoscopic cannulation of the ampulla of vater. A preliminary report. Ann Surg 1968;167:752-70. |
|2.||Oi I, Kobayashi S, Kondo T. Endoscopic panreatico-cholangiography. Endoscopy, 1970;2:103-6. |
|3.||Kawi K, Akaska Y, Murakami K, et al. Endoscopic sphinterotomy of the ampula of vater. Gastrointest Endosc 1974;20:148-51. |
|4.||Classen M, Phillip J. Endoscopic retrograde cholangio-panreatography (ERCP) and endoscopic therapy in panreatic disease. Clin Gastroenterol 1974; 13:819-41. |
|5.||Cotton PB. Frontiers of biliary endoscpy. Scand J Gastroenterol 1990,25(supp);175:58-62. |
|6.||Deviere J and Cremer M. Biliary endoscopy. Curr Opin Gastroenterol 1994,10;5:567-72. |
|7.||Afifi AA, Azen SP. Statistical analyzis: A computer-oriented approach. 2nd ed. New York; Academic Press, 1979. |
|8.||Metz CE. Basic principles of ROC analyzis. Semin Nucl Med 1978;8:283-98. [PUBMED] |
|9.||Dorfman DD, Alf E, Jr. Maximum likelihood estimation of parameters of signal-detection theory and determination of confidence intervals; rating method data. J Math Psych 1969;6:487-96. |
|10.||Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic curves derived from the same cases. Radiology 1982;143:29-36. [PUBMED] [FULLTEXT]|
|11.||Hanley JA, McNeil BJ. A method of comparing areas under receiver operating characteristic curves derived from the same cases. Radiology 1983; 148:839-43. [PUBMED] [FULLTEXT]|
|12.||Wasson JH, Sox HC, Neff RK, Goldman L. Clinical prediction rules: application and methodologic standards. N Engl J Med 1985;313:793-9. [PUBMED] |
|13.||Dixon WJ, Brown MB, Engelman L, Jennrich RI. BMDP Statistical Software manual Berkeley: University of California Press, 1993. |
|14.||Shimizu S, Tada M and Kawai K. Diagnostic ERCP. Endoscopy, 1994;26:88-92. |
|15.||Al-Karawi MA and Mohammed AE. Endoscopic retrograde cholangiopancreaticography and sphincterotomy. The RKH Experience. Saudi Med J 1989,10(1):547. |
|16.||Oll DJ, Young GP, Mitchell RB, Chen MYM and Gelfand DW. Therapeutic ERCP: Spectrum of procedures performed in 60 consecutive patients. Abdom Imaging 1994,19:30-3. |
|17.||Al-Mofleh IA. Management of gallstones. Saudi J Gastroenterol 1996;2(l);29-38. |
|18.||Silvis SE, Vennes JA. Endoscopic Retrograde Sphincterotomy. In: Silvis SE, Ed. Therapeutic Gastrointestinal Endoscopy, New York; Igaku-Shoin 1985:198-240. |
|19.||Al-Karawi MA and Mohammed AE. Endoscopic management of bile duct stones at Riyadh Military Hospital: An eight-year experience. Ann Saudi Med, 1991;11(1):62-6. |
|20.||Lambert ME, Belts CD, Hill J, Fragher EB, Martin DF, Tweedle DEF. Endoscopic sphincterotomy: The whole truth. B. J. Surg 1991;78:473-6. |
|21.||Al-Qurain A. Duodenal Tuberculosis, A cause of obstructive jaundice. Hellenic Journal of Gastroenterology 1993;6(1):41-5. |
|22.||Barkun AN, Barkun JS, Fried GM, Ghitulescu G, Steinmetz 0, Pham C, Meakin JL, Goresky CA and the McGill. Gallstone Treatment Group. Useful preditors of bile duct stones in patients undergoing laparoscopic cholecystectomy. Ann Surg 1994;220(l):32-9. |
Abdulaziz Al Quorain
P.O. Box 40001, Al-Khobar 31952
Source of Support: None, Conflict of Interest: None
[Table - 1]