ORIGINAL ARTICLE: Clinical Endoscopy
Nationwide, population-based data from 11,074 ERCP procedures from the Swedish Registry for Gallstone Surgery and ERCP Lars Enochsson, MD, PhD, Fredrik Swahn, MD, Urban Arnelo, MD, PhD, Magnus Nilsson, MD, PhD, Matthias Löhr, MD, PhD, Gunnar Persson, MD, PhD Stockholm, Jönköping, Sweden
Background: The Swedish Registry for Gallstone Surgery and ERCP (GallRiks) is the first nationwide Web-based quality registry for gallstone surgery and ERCP in the world. In this article we report data from 11,074 ERCPs performed in 2007 and 2008. Objective: The aim of this study is to present outcomes, safety data, and success rates of ERCPs performed in Sweden. Design: Data gathering from a medical record database. Patients: This study reviewed 11,074 ERCPs performed in 2007 and 2008. Methods: In GallRiks, data concerning surgery performed for gallstone disease as well as all ERCPs are recorded. The registry is approved by the Swedish Surgical Society and is based on an Internet platform with online data registration. The online program includes 30-day follow-up information as well as the opportunity to retrieve electronic reports on demand. The present data represent 76% of all ERCPs performed in Sweden in 2007 and 95% of those performed in 2008. The database also has been validated, indicating a complete match between the medical records and the database in 97.3% of ERCP cases. Main Outcome Measurements: Cannulation success and perioperative and postoperative complications. Results: A successful bile duct cannulation was achieved in 92% of the ERCPs performed. The presence of common bile duct stones was the predominant finding and was seen in 36.8% of examinations. Perioperative and postoperative complication rates were 2.5% and 9.8%, respectively. The rate of ERCP-induced pancreatitis was 2.7%, and the total 30-day mortality rate in the database was 5.9% but varied significantly among the different diagnostic groups. The indications for ERCP differed between high-volume and low-volume centers, indicating an adequate referral pattern of complex cases in Sweden. Limitations: GallRiks registration is voluntary and thus not 100%. This makes selection bias a possibility. Conclusion: ERCP is widely used at Swedish hospitals, with acceptable cannulation success rates and perioperative and postoperative complication rates similar to established standards. GallRiks is a population-based nationwide registry with good data validity and high inclusion rates regarding ERCPs. (Gastrointest Endosc 2010;72:1175-84.)
In Sweden, about 11,000 cholecystectomies (open and laparoscopic) are performed each year for benign indications.1 To assess the outcomes of surgical as well as endoscopic interventions in gallstone disease, the Swedish
Registry for Gallstone Surgery and ERCP (GallRiks, www. ucr.uu.se/gallriks) was established on May 1, 2005. The registry is supported by the Swedish Surgical Society and is estimated to include data from ⬎85% of all interven-
Abbreviations: ASA, American Society of Anesthesiologists; CBDS, common bile duct stone; GallRiks, The Swedish Registry for Gallstone Surgery and ERCP.
Current affiliations: Departments of Surgical Gastroenterology and Clinical Science, Intervention, and Technology (L.E., F.S., U.A., M.N., M.L.), Karolinska University Hospital, Karolinska Institutet, Stockholm; Department of Surgery (G.P.), Ryhov Hospital, Jönköping, Sweden.
DISCLOSURE: Drs Enochsson, Nilsson, and Löhr were supported by unconditional research funds at the Karolinska Institutet. No other financial relationships relevant to this publication were disclosed. Copyright © 2010 by the American Society for Gastrointestinal Endoscopy 0016-5107/$36.00 doi:10.1016/j.gie.2010.07.047
Reprint requests: Lars Enochsson, MD, PhD, Department of Surgery, Clintec, GastroCentrum, K53, Karolinska University Hospital, SE-141 86 Stockholm, Sweden.
Received March 9, 2010. Accepted July 28, 2010.
www.giejournal.org
Volume 72, No. 6 : 2010 GASTROINTESTINAL ENDOSCOPY
1175
Data from Swedish Registry for Gallstone Surgery and ERCP
tional procedures performed for gallstone disease in Sweden in the year 2008. Because bile duct injury, a rare but serious complication of gallstone surgery, is often treated by endoscopic stenting, data from all ERCPs, and not only those with a common bile duct stone (CBDS) indication, are also included in the registry. ERCP is an advanced interventional endoscopic procedure with a significantly higher complication rate than that of cholecystectomy.2-4 A 30-day morbidity rate as high as 15.9% has been reported in the literature, with pancreatitis being the single most common complication.2 There are, however, no large population-based studies documenting quality outcomes and complication rates of ERCP. This article presents data from, to our knowledge, the first nationwide population-based database of ERCPs in the world.
METHODS GallRiks The registry has from the start been financially supported by the Swedish National Board of Health and Welfare. The registry includes data for both open and laparoscopic surgery of the gallbladder (gallstone indication) as well as all ERCPs, with the aim of obtaining a complete registration of interventional gallstone treatment and its complication rates. The registry is approved by the Swedish Surgical Society, which also appoints the members of the GallRiks board. GallRiks uses an Internet platform (www.ucr.uu.se/gallriks) with online data registration of the procedures and 30-day follow-up information, as well as the opportunity to retrieve electronic reports on demand. The primary aim of GallRiks is to provide participating hospitals and physicians with continuously updated results on indications for surgery and surgical and endoscopic methods used as well as complications and patient satisfaction with the care delivered. The local hospitals’ outcome data can be analyzed online over any given time period as specified by the local user from 1 month upward. These local hospital data can be compared with the national data and thus used to evaluate and thereby support efforts for quality improvement at each hospital. Examples of the data that can be included and compared in the online success report for ERCP are given in Figure 1. Furthermore, each participating hospital has an appointed coordinator who has the right to download the local hospital data from GallRiks and to compile them and present them to the local hospital. Another aim for the registry is to use the national database for scientific analysis of the epidemiology and treatment of gallstone disease in Sweden. All the parameters that are possible to include in a report from the registry are presented in the Appendix (available online at www.giejournal.org). A program for validation of data in the registry also has been instituted. 1176 GASTROINTESTINAL ENDOSCOPY
Volume 72, No. 6 : 2010
Enochsson et al
Take-home Message ●
This large population-based analysis of 2 years’ yield of ERCPs in GallRiks shows that ERCP is widely used at Swedish hospitals and that the quality is good, including a low complication rate. This study may serve as a baseline description of the ERCP practice and outcomes in Sweden today. Future studies may show whether the introduction of a nation-based quality registry can serve its purpose to aid hospitals to further improve the quality of ERCP practice.
Four experienced doctors make hospital visits to ensure that adequate resources have been assigned for registration and follow-up. They compare a random sample of 25 medical records with the recordings made in the database. Corresponding to the relationship between gallstone surgery and ERCP, an appropriate distribution of cases among the 25 medical records is chosen. This number was considered to be appropriate for the reviewers to manage during a working day. The audit is funded by the registry. The current goal is that each hospital will be audited in this manner at least once every 3 years, given the financial resources of the registry. The results from the first 25 hospitals audited indicate a complete match between the medical records and the database in 97.3% of ERCP cases.5 In addition to this validation process by GallRiks, there is also the possibility of comparing the GallRiks data with those of several other Swedish registries, such as the Swedish Hospital Discharge Registry and the Swedish National Population Registry. Since its inception, an increasing number of Swedish hospitals have joined the registry, and the data presented here represent 76% of ERCPs performed in Sweden in 2007 and 95% of those performed in 2008.5 The coverage was calculated by comparing the number of procedures registered in GallRiks with data from the Swedish National Board of Health and Welfare.
Data collection The endoscopist registers the ERCP in the database immediately after the procedure is completed. At the 30day follow-up, the online registration is usually performed by the appointed coordinators at each participating hospital, usually 1 doctor and 1 nurse, and it is in most cases the nurse who registers the 30-day follow-up information. Cases in which it is unclear whether there is a complication or not are discussed between the coordinators. There is also a key coordinator at GallRiks with whom these matters can be discussed. The registry is voluntary, but because the results of the audits and the registration frequency of all hospitals are published in annual reports, which are public, there is considerable unofficial pressure on the hospitals to have an up-to-date and reliable registration. www.giejournal.org
Enochsson et al
Data from Swedish Registry for Gallstone Surgery and ERCP
Figure 1. Example of parameters included in the online success report for ERCP.
Data analyzed In 2007 and 2008, a total of 11,623 ERCPs were registered. The 30-day follow-up information was incomplete for 549 patients, and therefore 11,074 ERCPs performed on 8088 patients in 51 hospitals were available for evaluation.
Statistical methods and data analysis The Kruskal-Wallis nonparametric 1-way analysis of variance test was used when results from ⬎2 groups were simultaneously compared. The Mann-Whitney test was used to test statistical differences between 2 groups, and in the case of a normal distribution the t test was used for uncorrelated means. The chi-square test was used to evaluate hypotheses of variables in contingency tables, and the Fisher exact test was used in the case of small expected frequencies. In addition, descriptive statistics and graphical methods were used to characterize the data. The study uses multiple hypothesis testing, in which each hypothesis was analyzed separately, and the existence of patterns in the results and the consistency of the results were considered in the analysis. All analyses were carried out with JMP www.giejournal.org
version 8.0.1 (SAS Institute, Cary, NC) for Mac OS X version 10.5.7 (Apple, Cupertino, Calif), and the 5% level of significance was considered.
RESULTS Characterization of the population under study The majority of patients (7256, 66%) had an emergency/ urgent indication for ERCP, whereas 3816 (34%) of the examinations were scheduled elective procedures (in 2 cases, data were missing). Of all ERCPs, 1477 (13%) were performed on an outpatient basis. There was a predominance of women (55%) over men. The overall mean age was 67.6 years. There was a wide variation between the number of ERCPs per patient (1-11) during the study period (Table 1). The total number of ERCPs registered in 2008 was 6466, which represents approximately 70 ERCPs per 100,000 inhabitants in Sweden. Given that the degree of national coverage of the registry in 2008 is estimated to be 95%, the Volume 72, No. 6 : 2010 GASTROINTESTINAL ENDOSCOPY
1177
Data from Swedish Registry for Gallstone Surgery and ERCP
TABLE 1. No. of ERCPs per patient during study period No. of ERCPs
No. of patients
%
11
1
0.1
10
2
0.2
9
2
0.2
8
5
0.4
7
9
0.6
6
28
1.5
5
49
2.2
4
159
5.7
3
366
9.9
2
1308
23.6
1
6159
55.6
Total
8088
100
Enochsson et al
However, it is important to emphasize that the indication for the ERCP did not always correspond with the outcome. For example, in 8.4% of examinations with the indication of CBDS, no stones were detected.
Successful bile duct cannulation A successful bile duct cannulation was achieved in 92% of cases. Successful bile duct cannulation is defined in the registry as contrast material injection into the bile duct, whereas failure is defined as failure by the endoscopist to cannulate the papilla after it was visualized. The bile duct cannulation rate, however, differed depending on the indication. In patients with suspected CBDS, the cannulation success rate was 95.4%, whereas in the group with known malignancy it was only 83.4% (P ⬍ .05; Fig. 3). The technique of cannulating with a guidewire seems to have been generally adopted in Sweden, because this method was used in 83.7% of the examinations. This technique is used more frequently in high-volume hospitals (91.2%) than in intermediate- and low-volume hospitals (82.6% and 79.2%, respectively; P ⬍ .05).
Pathologic findings annual rate of ERCPs performed in Sweden is approximately 74 per 100,000 inhabitants.
Hospital and endoscopist ERCP volumes There was great variation in the annual frequency of ERCPs between different hospitals as well as between different endoscopists, as shown in Figure 2. Two highvolume hospitals (performing ⬎1000 procedures during the study period) did 22% of the ERCPs. Intermediatevolume hospitals (200-1000 ERCPs during the study period) did 56% of the procedures, whereas low-volume hospitals (⬍200 ERCPs) performed 22% of the procedures. Endoscopists performing ⬎200 ERCPs during the study period did 24% of the procedures. Of the 177 endoscopists, 55 did only ⱕ10 ERCPs each during the 2-year period.
Indications The 2 major indications for ERCP in the database (Table 2) were CBDS (2926 procedures, 26.4%) and unspecified obstructive jaundice (2808 procedures, 25.4%). A previously diagnosed malignancy was the primary indication for ERCP in 1140 (10.3%) cases, but additional cancers might be present in the group of unspecified obstructive jaundice. Suspected bile duct leakage was the indication in 291 (2.6%) cases. In low-volume centers, CBDS was the indication for ERCP in 844 (33.9%) cases, whereas in highvolume centers CBDS was the indication in only 517 (21.3%) cases, thus indicating a different case mix by volume. There were also differences in the case mix among high-, intermediate-, and low-volume centers regarding interventional endoscopic diagnosis and treatment of sclerosing cholangitis and chronic pancreatitis. 1178 GASTROINTESTINAL ENDOSCOPY
Volume 72, No. 6 : 2010
CBDS was the predominant finding at ERCP and was seen in 4074 (36.8%) examinations (Table 3). In lowvolume hospitals CBDS was found in 1011 (40.6%) cases, in intermediate-volume centers CBDS was found in 2359 (38.3%) of the ERCPs, and high-volume hospitals had a slightly different profile with only 704 (29.0%) cases of CBDS (P ⬍ .0001). A solitary bile duct stone was found in 1683 (41.3%) patients with CBDS, 2 to 5 stones in 1891 (46.4%) patients, and ⬎5 stones in 500 (12.3%) patients. A stenosis of the bile ducts was found in 3312 (29.9%) examinations. The cause of the stenosis was estimated by the examiner to be malignant in 1555 (47.0%) of these cases, benign in 496 (15.0%), and unclear in 323 (9.8%). In 28.3% of examinations in which stenosis was found, the examiner was not able to perioperatively classify the type of stenosis. The percentage of stenoses that could be classified seemed to be somewhat higher in high-volume centers (35.3%) compared with that of intermediate-volume (29.6%) and low-volume (25.5%) centers (P ⬍ .0001). Malignant stenosis, on the other hand, seemed to be more frequent in low-volume hospitals at 348 cases (54.7%) compared with 403 cases (47.0%) in high-volume hospitals. In 13% of examinations, no pathologic findings were noted.
Interventional procedures ERCP is largely used as an interventional modality (Table 4). This is illustrated by the fact that in 6880 (75%) of the 9117 examinations in which a previous sphincterotomy had not been performed, some interventional procedure such as a sphincterotomy, precut sphincterotomy, or dilatation of the sphincter was made. Primary dilatation of the native papilla as the sole measure is, however, rarely www.giejournal.org
Enochsson et al
Data from Swedish Registry for Gallstone Surgery and ERCP
Figure 2. The number of ERCPs performed at each hospital and by each endoscopist during the study period. The borders between high-volume, intermediate-volume, and low-volume centers are indicated by the vertical dotted lines.
performed (0.1%). Complete CBDS extraction was successfully achieved in 77.1% of examinations. High-volume centers had a slightly higher success rate (81.1%) than low-volume centers (72.5%; P ⫽ .0008). In 38% of examinations, ⱖ1 endostents were placed. Plastic stents were most frequently used (79%), and the choice of metallic stents was evenly distributed between covered and uncovered.
Sedation Therapeutic procedures require good anesthesia, and the majority of ERCPs were performed with the patient under unconscious sedation with either propofol or general anesthesia (54%), whereas conscious sedation with midazolam was used in 46%. In the propofol group, 40% of procedures were performed with monitored anesthesia care.
American Society of Anesthesiologists categories There was a fairly even distribution of patients between the different hospital categories regarding the American Society of Anesthesiologists (ASA) physical status classification system (Table 5). The overall majority of patients were in the ASA 2 category. High-volume centers, though, tended to treat patients with higher degrees of comorbidity. www.giejournal.org
Complications Perioperative. The overall complication rate was 2.5% (Table 5). Bleeding and contrast material extravasation accounted for 0.6% each. Low-volume centers had a low, but still increased, perioperative complication rate compared with both high- and intermediate-volume centers. Postoperative. The overall postoperative complication rate in the whole ERCP dataset was 9.8%. After ERCP, pancreatitis was registered in 2.7% of cases. Pancreatitis was defined in the registry as serum amylase levels of ⬎3 times the normal value together with persistent abdominal pain for ⬎24 hours. The pancreatitis rate in high-volume hospitals was significantly higher (3.7%) than in low- and intermediate-volume hospitals (both 2.4%). We found a significantly increased frequency of post-ERCP pancreatitis in the group in which the pancreatic duct was cannulated (5.3%) compared with the group in which only the bile duct was cannulated (1.8%; P ⬍ .05). Postoperative bleeding that required additional intervention accounted for 0.9% of patients. A perforation of the bowel or bile ducts after the procedure occurred in 0.3% of patients.
Mortality Mortality rates of the patients in the registry were obtained from the Swedish National Population Registry. The Volume 72, No. 6 : 2010 GASTROINTESTINAL ENDOSCOPY
1179
Data from Swedish Registry for Gallstone Surgery and ERCP
Enochsson et al
TABLE 2. Indications for ERCP High-volume center (>1000)
Intermediatevolume center (200-1000)
n
%
n
%
n
%
High-tointermediate P value
Common bile duct stone
2926
26.4
517
21.3
1565
25.4
Obstructive jaundice
2808
25.4
575
23.7
1646
Malignancy
1140
10.3
282
11.6
Cholangitis
1008
9.1
121
Other
623
5.6
Acute pancreatitis
598
Stent dysfunction
Total
Low-volume center (<200) n
%
High-to-low P value
⬍.0001
844
33.9
⬍.0001
26.7
.0034
587
23.6
NS
584
9.5
.0037
274
11.0
NS
5.0
675
11.0
⬍.0001
212
8.5
⬍.0001
216
8.9
275
4.5
⬍.0001
132
5.3
⬍.0001
5.4
146
6.0
379
6.2
NS
73
2.9
⬍.0001
574
5.2
53
2.2
401
6.5
⬍.0001
120
4.8
⬍.0001
Scheduled reintervention
568
5.1
210
8.6
261
4.2
⬍.0001
97
3.9
⬍.0001
Suspected bile leakage
291
2.6
72
3.0
155
2.5
NS
64
2.6
NS
Scheduled ERCP control
204
1.8
75
3.1
80
1.3
⬍.0001
49
2.0
.0138
Chronic pancreatitis
169
1.5
97
4.0
61
1.0
⬍.0001
11
0.4
⬍.0001
Primary sclerosing cholangitis
133
1.2
60
2.5
54
0.9
⬍.0001
19
0.8
⬍.0001
Prophylaxis vs gallstone pancreatitis
19
0.2
2
0.1
10
0.2
NS
7
0.3
NS
Scheduled stone extraction after ESWL
13
0.1
3
0.1
9
0.1
NS
1
0.0
NS
11,074
100
2429
100
6155
100
2490
100
Indication
Total
NS, Not significant; ESWL, extracorporeal shock-wave lithotripsy.
Figure 3. Bile duct cannulation success rates in patients for whom the indication was common bile duct stone or malignancy. Superficial cannulation was registered in this study when the endoscopist was able to fill the common bile duct with contrast material but was unable to introduce the catheter. CBDS, Common bile duct stone.
overall 30-day mortality rate was 5.9% (Table 5). The 30-day mortality rate in patients with CBDS was 1.5%, whereas it was significantly higher in those with malignant stenosis, as high as 14.7%. In high-volume hospitals, 301180 GASTROINTESTINAL ENDOSCOPY
Volume 72, No. 6 : 2010
day mortality in patients with malignant stenosis was 17.4%, whereas it was 14.2% and 12.9% in intermediateand low-volume hospitals, respectively. Furthermore, mortality rates were higher in patients with higher ASA www.giejournal.org
Enochsson et al
Data from Swedish Registry for Gallstone Surgery and ERCP
TABLE 3. Pathologic findings High-volume center (>1000)
Intermediatevolume center (200-1000)
n
%
n
%
n
%
High-tointermediate P value
CBDS
4074
36.8
704
29.0
2359
38.3
⬍.0001
No CBDS
7000
63.2
1725
71.0
3796
11,074
100
2429
100
1
1683
41.3
280
2-5
1891
46.4
6-10
292
⬎10
Total
Low-volume center (<200) n
%
High-to-low P value
1011
40.6
⬍.0001
61.7
1479
59.4
6155
100
2490
100
39.8
994
42.1
NS
409
40.5
NS
314
44.6
1080
45.8
NS
497
49.2
NS
7.2
57
8.1
164
7.0
NS
71
7.0
NS
208
5.1
53
7.5
121
5.1
.0200
34
3.3
.0002
Total
4074
100
704
100
2359
100
1011
100
Stenosis
3312
29.9
857
35.3
1819
29.6
636
25.5
No stenosis
7762
70.1
1572
64.7
4336
70.4
1854
74.5
11,074
100
2429
100
6155
100
2490
100
Malignant
1555
47.0
403
47.0
804
44.2
348
54.7
Benign
496
15.0
204
23.8
211
11.6
81
12.7
Unclear
323
9.8
100
11.7
170
9.3
53
8.3
Not stated
938
28.3
150
17.5
634
34.9
154
24.3
3312
100
857
100
1819
100
636
100
Finding
Total No. of CBDSs specified
Total
⬍.0001
⬍.0001
Stenosis specified
Total
NS
.0034
CBDS, Common bile duct stone; NS, not significant.
classification scores both in patients with CBDS and in those with malignant stenosis.
DISCUSSION In the present study, we evaluated data from 11,074 ERCPs from GallRiks. The aim of the study was to present a descriptive analysis of hospital caseloads, indications, pathologic findings, cannulation success, sedation techniques, complications, and mortality of ERCPs performed in Sweden. By integrating gallbladder surgery and ERCP in the same database, with a common 30-day follow-up, GallRiks provides a unique opportunity to study the clinical course of gallstone treatment, regardless of whether transabdominal surgery, endoscopy, or both were used. The data presented here demonstrate a great variation regarding the case mix and number of ERCPs performed at different hospitals and by different endoscopists in Sweden. Success as well as complication rates, however, seem to be evenly distributed and of good standards. The strength of the study is the prospective documentation of data in a registry with national coverage reprewww.giejournal.org
senting up to 85% of all ERCPs performed in Sweden during the studied time period. Furthermore, a program for onsite validation of data is in progress, and preliminary data from more than one-third of the participating hospitals indicates good data validity. Moreover, the large sample size serves to minimize the risk of random error, thus ensuring good precision even in subanalyses. In Sweden as a whole, the success rate for cannulation of the bile ducts is 92%, which is lower than in some high-volume centers6,7 but probably reflects the true figure of successful cannulation in ERCPs throughout Sweden today. In Swedish high-volume centers, the ratio of CBDS (704) to stenosis (857) was 0.82, whereas it was 1.30 (2359/1819) and 1.59 (1011/636) in intermediate- and lowvolume centers, respectively, supporting a notion of a different case mix between hospital groups. Furthermore, in the database there was an extremely wide difference in the volume of ERCPs registered at individual centers during the study period (ranging from 1 to 1262). There were, however, no statistically significant differences regarding bile duct cannulation success between high-volume and low-volume groups. Volume 72, No. 6 : 2010 GASTROINTESTINAL ENDOSCOPY
1181
Data from Swedish Registry for Gallstone Surgery and ERCP
Enochsson et al
TABLE 4. Interventional procedures High-volume center (>1000)
Intermediatevolume center (200-1000)
n
%
n
%
n
%
High-tointermediate P value
Sphincterotomy
6010
54.3
1122
46.2
3498
56.8
Precut sphincterotomy
421
3.8
90
3.7
241
Precut sphincterotomy and sphincterotomy
437
3.9
204
8.4
Dilatation
12
0.1
2
No intervention
1627
14.7
Previous sphincterotomy
1957
Not stated
Total
Low-volume center (<200) n
%
High-to-low P value
⬍.0001
1390
55.8
⬍.0001
3.9
NS
90
3.6
NS
163
2.6
⬍.0001
70
2.8
⬍.0001
0.1
8
0.1
NS
2
0.1
NS
264
10.9
906
14.7
⬍.0001
457
18.4
⬍.0001
17.7
635
26.1
959
15.6
363
14.6
610
5.5
112
4.6
380
6.2
118
4.7
11,074
100
2429
100
6155
100
2490
100
Successful
3140
77.1
571
81.1
1836
77.8
733
72.5
Unsuccessful
725
17.8
114
16.2
386
16.4
225
22.3
Not stated
209
5.1
19
2.7
137
5.8
53
5.2
4074
100
704
100
2359
100
1011
100
Endostent
4246
38.3
1113
45.8
2270
36.9
863
34.7
No stent
6828
61.7
1316
54.2
3885
63.1
1627
65.3
11,074
100
2429
100
6155
100.0
2490
100
Plastic stent
3367
79.3
901
81.0
1716
75.6
.0005
750
86.9
.0004
Covered metallic stent
466
11.0
131
11.8
303
13.3
NS
32
3.7
⬍.0001
Uncovered metallic stent
413
9.7
81
7.3
251
11.1
.0004
81
9.4
NS
Total
4246
100
1113
100
2270
100
863
100
1
3684
86.8
875
78.6
2025
89.2
784
90.8
2
477
11.2
192
17.3
211
9.3
74
8.6
3
69
1.6
38
3.4
26
1.1
5
0.6
4
16
0.4
8
0.7
8
0.4
0
0.0
Total
4246
100
1113
100
2270
100
863
100
Intervention Sphincterotomy
Total Complete CBDS extraction
Total
NS
.0008
Endostent
Total
⬍.0001
⬍.0001
No. of stents
NS, Not significant; CBDS, common bile duct stone.
The accuracy of ERCP as a diagnostic procedure is dependent on the quality of the imaging equipment as well as the skills of the endoscopic team. The indication for performing ERCP in the present data was unspecified obstructive jaundice in 25.4% of cases and suspected bile leakage in 2.6% of cases. Thus, ERCP is still in many cases 1182 GASTROINTESTINAL ENDOSCOPY
Volume 72, No. 6 : 2010
both a diagnostic as well as a therapeutic procedure. It is, however, important that unindicated procedures should not be undertaken, because they impose the attendant procedural risks without the potential for therapeutic benefit. The high positive yield (89%) from ERCPs performed for the indication of suspected postoperative bile leakage www.giejournal.org
Enochsson et al
Data from Swedish Registry for Gallstone Surgery and ERCP
TABLE 5. ASA categories, complications, and mortality rates High-volume center (>1000)
Intermediatevolume center (200-1000)
n
%
n
%
n
%
High-tointermediate P value
1
2716
24.5
626
25.8
1473
23.9
2
5134
46.4
983
40.5
2974
3
2809
25.4
703
28.9
4
395
3.6
115
5
17
0.2
Not registered
3
Total
Low-volume center (<200) n
%
High-to-low P value
NS
617
24.8
NS
48.3
⬍.0001
1177
47.3
⬍.0001
1516
24.6
⬍.0001
590
23.7
⬍.0001
4.7
180
2.9
⬍.0001
100
4.0
NS
2
0.1
11
0.2
NS
4
0.2
NS
0.0
0
0.0
1
0.0
NS
2
0.1
NS
11,074
100
2429
100
6155
100
2490
100
Perioperative complications
279
2.5
51
2.1
150
2.4
NS
78
3.1
.0255
Bleeding
66
0.6
11
0.5
44
0.7
NS
11
0.4
NS
Contrast material effusion
67
0.6
13
0.5
36
0.6
NS
18
0.7
NS
Postoperative complications
1090
9.8
271
11.2
579
9.4
.0161
240
9.6
NS
Pancreatitis
299
2.7
89
3.7
150
2.4
.0027
60
2.4
.0123
Bleeding
95
0.9
26
1.1
47
0.8
NS
22
0.9
NS
Perforation
31
0.3
8
0.3
15
0.2
NS
8
0.3
NS
Total
654
5.9
148
6.1
373
6.1
NS
133
5.3
NS
CBDS
60
1.5
7
1.0
38
1.6
15
1.5
Malignant stenosis
229
14.7
70
17.4
114
14.2
45
12.9
30-day mortality by ASA category
ASA 1
ASA 2
ASA 3
ASA 4
CBDS
0.2%
0.7%
3.8%
13.2%
Malignant stenosis
4.6%
8.1%
21.4%
34.3%
ASA category
Total Perioperative complications
Postoperative complications
30-day mortality
ASA, American Society of Anesthesiologists; NS, not significant; CBDS, common bile duct stone.
underscores the importance of having this procedure available in hospitals performing gallstone surgery, because many of the minor bile duct injuries can be successfully treated with ERCP and endoscopic stenting. The post-ERCP pancreatitis rate of 2.7% is in accordance with the literature, in which equivalent rates of between 1.3% and 6.7% have been reported in large-scale prospective studies.2,8-11 In our data, we noted a higher frequency of pancreatitis in high-volume centers compared with both intermediate- and low-volume centers (Table 5). One could speculate that the variable case mix, www.giejournal.org
with a greater number of examinations involving exploration and manipulations in the pancreatic duct (as for indications like chronic pancreatitis) could be an explanation for this occurrence. Our finding of a significantly increased postERCP pancreatitis frequency of 5.3% in examinations in which the pancreatic duct was cannulated compared with 1.8% when only the bile duct was cannulated also supports this hypothesis. The frequency of GI hemorrhage and perforation are in the low range of previously reported figures in the Volume 72, No. 6 : 2010 GASTROINTESTINAL ENDOSCOPY
1183
Data from Swedish Registry for Gallstone Surgery and ERCP
literature.2,9-11 One can only speculate upon the reason for this, but many of the reported large-scale prospective studies that constitute the foundation for the guidelines on the management of CBDSs were published almost 10 years ago.12 Since then, several technical improvements have increased the safety and success rates of ERCP. One might also speculate whether improved diathermy equipment contributed to a decrease in the risk of bleeding at sphincterotomy. The technique for cannulating the bile ducts with a guidewire and changing the various ERCP catheters and sphincterotomes over the existing guidewire is frequently used in Sweden. This technique reduces the risk of cannulating and injecting contrast material into the pancreatic duct by mistake and may thereby theoretically reduce the risk of ERCP-induced pancreatitis. It should, however, be noted that the registry does not differentiate between techniques by which the guidewire was introduced but just notes the fact that a guidewire was used during the procedure to facilitate the cannulation. Freeman et al,8 in a prospective multicenter study, demonstrated the importance of both of these technical skills, as well as avoidance of several pancreatic duct cannulations and a careful selection of patients, to avoid post-ERCP pancreatitis. An overall mortality rate of 1.5% for patients with a primary indication of CBDS may be high, but as shown in Table 5, the mortality rates are dependent on the ASA classification. It is also important to emphasize that 30-day mortality rates are obtained automatically from the Swedish National Population Registry and thus do not differentiate between mortality caused by the procedure, the general condition of the patient, or the nature of the disease itself. This large population-based analysis of 2 years’ yield of ERCPs in GallRiks shows that ERCP is a widely used procedure at Swedish hospitals. The outcomes, quality, and complication rates are similar to international standards. This study may serve as a baseline description of ERCP practice and outcome in Sweden today. Future stud-
1184 GASTROINTESTINAL ENDOSCOPY
Volume 72, No. 6 : 2010
Enochsson et al
ies may show whether the introduction of a nation-based quality registry can serve its purpose to aid hospitals to further improve the quality of ERCP practice. ACKNOWLEDGMENT The authors acknowledge Melroy A. d’Souza for his valuable input and assistance in preparing the manuscript. REFERENCES 1. Rosenmuller M, Haapamaki MM, Nordin P, et al. Cholecystectomy in Sweden 2000-2003: a nationwide study on procedures, patient characteristics, and mortality. BMC Gastroenterol 2007;7:35. 2. Christensen M, Matzen P, Schulze S, et al. Complications of ERCP: a prospective study. Gastrointest Endosc 2004;60:721-31. 3. Colton JB, Curran CC. Quality indicators, including complications, of ERCP in a community setting: a prospective study. Gastrointest Endosc 2009;70:457-67. 4. Kapral C, Duller C, Wewalka F, et al. Case volume and outcome of endoscopic retrograde cholangiopancreatography: results of a nationwide Austrian benchmarking project. Endoscopy 2008;40:625-30. 5. Persson G, Enochsson L, Sandblom G. Årsrapport för Svenskt kvalitets register för gallstenskirurgi [Swedish]. In: Årsrapport 2007, 4th ed. Jönköping; 2008. p. 1-35. 6. Chen CM, Tay KH, Hoe MN, et al. Endoscopic retrograde cholangiopancreatography management of common bile duct stones in a surgical unit. Aust N Z J Surg 2005;75:1070-2. 7. Enochsson L, Lindberg B, Swahn F, et al. Intraoperative endoscopic retrograde cholangiopancreatography (ERCP) to remove common bile duct stones during routine laparoscopic cholecystectomy does not prolong hospitalization: a 2-year experience. Surg Endosc 2004;18:367-71. 8. Freeman ML, DiSario JA, Nelson DB, et al. Risk factors for post-ERCP pancreatitis: a prospective, multicenter study. Gastrointest Endosc 2001;54:425-34. 9. Freeman ML, Nelson DB, Sherman S, et al. Complications of endoscopic biliary sphincterotomy. N Engl J Med 1996;335:909-18. 10. Loperfido S, Angelini G, Benedetti G, et al. Major early complications from diagnostic and therapeutic ERCP: a prospective multicenter study. Gastrointest Endosc 1998;48:1-10. 11. Masci E, Toti G, Mariani A, et al. Complications of diagnostic and therapeutic ERCP: a prospective multicenter study. Am J Gastroenterol 2001; 96:417-23. 12. Williams EJ, Green J, Beckingham I, et al. Guidelines on the management of common bile duct stones (CBDS). Gut 2008;57:1004-21.
www.giejournal.org
Enochsson et al
Data from Swedish Registry for Gallstone Surgery and ERCP
Appendix Figure 1. The complete list of included optional parameters in GallRiks.
www.giejournal.org
Volume 72, No. 6 : 2010 GASTROINTESTINAL ENDOSCOPY 1184.e1
Data from Swedish Registry for Gallstone Surgery and ERCP
Enochsson et al
Appendix Figure 1. continued.
1184.e2 GASTROINTESTINAL ENDOSCOPY
Volume 72, No. 6 : 2010
www.giejournal.org
Enochsson et al
Data from Swedish Registry for Gallstone Surgery and ERCP
Appendix Figure 1. continued.
www.giejournal.org
Volume 72, No. 6 : 2010 GASTROINTESTINAL ENDOSCOPY 1184.e3