Baseline characteristics of patients
A total of 268 ERCP procedures were performed in 179 patients (77.09% males) aged ≥65 years (Group A). The mean age of Group A was 69.12 ± 4.65 years. In Group B, a total of 612 ERCP procedures were performed in 358 patients aged < 65 years (mean age 42.68 ± 12.97 years). There were statistically significant differences between the groups in terms of age of onset of CP, incidence of pancreatic stones, and incidence of chronic concomitant diseases (Table 1). Compared with the control group, geriatric patients had an older age of onset of CP (63.51 ± 10.29 years vs 37.63 ± 13.71 years; P<0.001). The incidence of pancreatic stones was also higher in Group A (94.97% vs 77.93%; P<0.001). Chronic concomitant diseases, including cardiovascular diseases (P<0.001), neurologic diseases (P=0.005), pulmonary diseases (P=0.005), liver disease (P=0.002), diabetes mellitus (P=0.009), and hypertension (P<0.001), were more prevalence in Group A than in Group B. According to the M-ANNHEIM classification, the proportion of CP patients at stage 0, I, II, III, IV was 5.59% vs 7.82%, 54.75% vs 58.10%, 31.84% vs 25.42%, 6.70% vs 7.54%, and 1.12% vs 1.12% for Group A and B, respectively (all P>0.05).
ERCP findings and interventions
Endoscopic findings on ERCP included dilation of the main pancreatic duct (MPD) in 218 (81.34%) ERCPs, periampullary diverticula in 19 (7.09%) ERCPs, and common bile duct (CBD) stricture in 11 (4.10%) ERCPs (Table 2). There were no significant differences between the two groups (all P>0.05).
Among those patients with radiopaque pancreatic stones ≥5 mm, 142 (52.99%) and 337 (55.07%) ESWL procedures were performed before ERCP in Group A and B, respectively (P=0.569). There were 13 (4.85%) and 34 (5.56%) dorsal pancreatogram procedures in Group A and B, respectively (P=0.669). Of the six patients with pancreas divisum in Group A, five underwent dorsal pancreatogram; one did not due to failed cannulation. In both groups, major papilla sphincterotomy (50.37% vs 52.45%; P=0.570) and minor papilla sphincterotomy (3.36% vs 3.92%; P=0.686) were performed before further manipulation. In both groups, dilation of PD stricture involved bougie dilation (18.28% vs 19.93%; P=0.569) or balloon dilation (5.97% vs 9.48%; P=0.084). In the process of stone removal, an extraction balloon (64.55% vs 68.14%; P= 0.298) or an extraction basket (3.73% vs 3.92%; P=0.893) were used. PD stents (5F-10F diameter) were placed in 179 (66.79%) ERCPs in Group A and 421 (67.48%) ERCPs in Group B (P=0.840) (Table 2).
Success rate of ERCP
The ERCP success rate was 92.16% (247/268) in Group A and 92.32% (565/612) in Group B; there was no statistically significant difference between the two groups (P=0.936) (Table 3). There were also no differences in complete (85.07% vs 84.97%; P=0.967) or partial success (7.09% vs 7.35%; P=0.890) between the two groups. In Group A, there was a total of 21 ERCP failures, eight of which were caused by abnormal duodenal papilla (four due to swelling of the papilla, one due to a small opening of the papilla, and three due to no identification of the papilla). Another main reason for ERCP failure was abnormal MPD, including MPD distortion (2/21), stricture (1/21), or stone blockage (2/21). Other causes of failure included duodenal stenosis (1/21), pancreas divisum (2/21), and unknown causes (5/21).
Complications of ERCP
The overall incidence of post-ERCP complications was similar in the two groups (7.09% vs 5.72%; P=0.436) (Figure 1). There were no statistical differences in the incidence of PEP (4.85% vs 4.58%), bleeding (1.12% vs 0.33%), infection (1.12% vs 0.65%), or basket impaction (0 vs 0.16%) between the two groups (all P>0.05). Although the incidence and type of complications were similar, the risk of moderate to severe complications in elderly patients was significantly increased compared to younger patients; almost all complications in Group B were mild (2.61% vs 0.16%; P=0.002). All complications were resolved with conservative medical treatment, except for two cases of moderate bleeding in Group A.
PEP is the main post-ERCP complication. In this study, PEP was documented in 13 cases in Group A, including eight mild cases, three moderate cases, and two severe cases. In comparison, there were 28 PEPs in Group B, of which almost all were mild (27/28) and only one case of PEP was moderate. The risk of moderate to severe PEP in elderly patients was significantly higher compared to younger patients (1.87% vs 0.16%; P=0.017). In Group A, no PEP occurred among CP patients at stage 0, III, and IV, and the incidence of PEP among CP patients at stage I and II was 9.18% (9/98) and 7.02% (4/57), respectively. In Group B, with increasing stage, the incidence of PEP was 7.14% (2/28), 7.69% (16/208), 4.40% (4/91), 3.70% (1/27), and 0% (0/4), respectively.
Three of 268 (1.12%) ERCP procedures in Group A developed post-ERCP bleeding. No patients with bleeding had a history of anti-platelets/anti-coagulants prior to procedure. There was one mild case of bleeding caused by pancreatic pseudocyst. The patient was treated by placing a pancreatic plastic stent for drainage of the pseudocyst in the tail of the pancreas. On the second day after ERCP, the patient presented abdominal pain. An emergency upper abdominal CT scan showed bleeding from the pseudocyst in the tail of the pancreas. The bleeding was successfully treated by intravenous injection of hemocoagulase agkistrodon. There were two moderate cases of bleeding caused by cardiac mucosal laceration. Two patients with MPD stones were treated by extraction of the stone fragments at ERCP. When the endoscopy was withdrawn, the cardiac mucosa laceration and bleeding was observed. The patients' hemoglobin levels declined significantly requiring blood transfusion treatment. For these two cases, bleeding was successfully stopped utilizing intraoperative and postoperative placement of hemostatic clips under the endoscope, hot probe hemostasis, and local injection of hemostatic drugs. There was no ERCP-related perforation or death and anaesthesia related adverse events.
Risk factors for PEP
Given the high incidence of PEP in our geriatric cohort, we performed a risk factor analysis for PEP (Table 4). Univariate analysis showed that three patient-related factors and two intervention-related factors were significantly associated with PEP, including female gender (P=0.047), pancreas divisum (P=0.031), pancreatic stones (P=0.099), dorsal pancreatogram (P=0.007), and ESWL (P=0.015). Factors with P values less than 0.10 were used to establish a multivariate model. After multivariate analysis, four factors were determined to be independently related to PEP. Female gender (OR, 3.40; 95% CI, 1.02-11.31; P=0.046), pancreas divisum (OR, 7.15; 95% CI, 1.01-50.62; P=0.049), and dorsal pancreatogram (OR, 7.40; 95% CI, 1.63-33.64; P=0.010) significantly increased the risk of PEP and ESWL prior to ERCP significantly reduced the risk of PEP (OR, 0.15; 95% CI, 0.03-0.70; P=0.016).