Patients and Methods

Results

Detectable acute pancreatitis developed in 31 of the 308 patients studied. Twenty three of them were classified as mild with rapid spontaneous recovery. Eight patients (26%) had a severe episode of pancreatitis with prolonged hospital stay or complications.

The median baseline level of serum trypsinogen 2 in the patients with ERCP induced AP was 53 μg/l (range 17–189 μg/l); one hour after ERCP it was 485 μg/l (range 35–6988 μg/l), at six hours 1401 μg/l (range 73–8990 μg/l), and at 24 hours 1008 μg/l (range 90–7311 μg/l) (fig 1). The median concentration at six hours was 26 times the baseline value. The differences between patients with and without ERCP induced pancreatitis were statistically significant at all time points (p<0.001).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

: Box plots showing the concentration of serum trypsinogen 2 (25–75% interquartile range, mean, 95% range, and outliners) in patients without pancreatitis, with mild pancreatitis, and with severe pancreatitis before, and at one, six, and 24 hours after ERCP.

The median baseline level of serum trypsin 2-AAT in patients with postprocedural pancreatitis was 8 μg/l (range 3–114 μg/l) and an 11-fold elevation to 88 μg/l (range 32–944 μg/l) was observed at 24 hours after ERCP (fig 2). This difference was statistically significant (p<0.001).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

: Concentrations of trypsin 2-AAT before, and at one, six, and 24 hours after ERCP.

Six hours after ERCP, the trypsinogen 2 concentrations were higher in severe pancreatitis (median 2440 μg/l, range 93–8990 μg/l) than in mild pancreatitis (median 1090 μg/l, range 278–7400 μg/l) (p=0.19). The ability of trypsinogen 2, trypsin 2-AAT, and amylase to differentiate mild from severe ERCP induced pancreatitis was also evaluated by receiver operating characteristic (ROC) curve analysis. The area under the ROC curve (AUC) was 0.67 for trypsinogen 2 and only 0.60 for trypsin 2-AAT because of the considerable overlapping between the groups. However, amylase had an AUC of only 0.54, which approaches 0.5, and so fails to differentiate between the patients. The median trypsin 2-AAT concentration 24 hours after ERCP in patients with severe pancreatitis was 162 μg/l (range 36–770 μg/l), and in mild postprocedural pancreatitis was 68 μg/l (range 31–940 μg/l) (p=0.38) (fig 2).

A substantial proportion of the patients (29%) had elevated trypsinogen 2 levels (>90 μg/l) and 32% had elevated trypsin 2-AAT levels (>12 μg/l) before ERCP (table 1). The frequency of elevated amylase values before the examination was 21%. We therefore analysed whether an ERCP induced increase above the baseline value would provide better accuracy. With a threefold increase as a cut off value, the sensitivity of trypsinogen 2 one hour after the procedure was 74%, after six hours 93%, and after 24 hours 90%, the specificity being 87%, 91%, and 95%, respectively. For trypsin 2-AAT the sensitivity at 24 hours was 93% and the specificity 90%. Because the diagnosis of pancreatitis was based on amylase, the sensitivity of a threefold increase in amylase activity was, by definition, 100% while the specificity was 81%. The differences in specificity between amylase (53 false positives out of 277) and trypsinogen 2 (15–35 false positives out of 277) as well as trypsin 2-AAT (27 false positives out of 277) were statistically significant (p<0.001). The correlations of trypsinogen 2 and trypsin 2-AAT with amylase activity at different time points were fairly low (correlation coefficients 0.21–0.45) (fig 3). The correlation between trypsinogen 2 at six hours and trypsin 2-AAT concentrations at 24 hours was high (r=0.70) (fig4).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

: Correlations of trypsinogen 2 and amylase before and six hours after ERCP, and trypsin 2-AAT and amylase before and 24 hours after the procedure. Filled circles, ERCP induced pancreatitis; open circles, no pancreatitis.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

: Scattergram of the correlation between trypsinogen 2 at six hours and serum trypsin 2-AAT at 24 hours after the examination (n=308).

Before ERCP, the patients who developed pancreatitis actually had lower trypsinogen 2 values than those who did not. A pre-ERCP trypsinogen 2 concentration higher than 300 μg/l was inversely correlated with the risk of pancreatitis (p=0.02, Fisher’s exact test). An elevated baseline concentration over 32 μg/l for trypsin 2-AAT was also associated with a lower risk, but this was not significant (p=0.27).

The frequency of high baseline values of trypsinogen 2 and trypsin 2-AAT varied between the diagnostic groups. Elevated levels were more common in patients with biliary tract cancer, in whom the median pre-ERCP concentration of trypsinogen 2 was 92 μg/l (range 23–592 μg/l), than in the rest (p=0.07). In patients with pancreatic cancer, the median baseline concentration of trypsinogen 2 was 81 μg/l (range 13–880 μg/l). A smaller rise in trypsin 2-AAT concentrations was observed in these patients. In chronic pancreatitis, the baseline values of the two markers did not differ from those of the rest of the study population.

Table 2 summarises the technical factors of the ERCP procedure and the risk of pancreatitis. The risk of pancreatitis increased with long duration of ERCP, endoscopic papillotomy (EP) to the Wirsungian duct, and stent placement with EP. However, precut papillotomy, sphincterotomy, and the number or volume of contrast media injections into the pancreatic duct had no influence on the risk. No significant differences were seen in liver function tests, infection markers, number of patients with postcholecystectomy state, or medical history. Acinar opacification was not seen in this series. The success rate of the procedure was 92%. One patient died of septic complications giving a 30 day mortality rate of 0.3%. Bleeding requiring endoscopic treatment was seen in three patients and duodenal perforation occurred in two patients.

Discussion

Our results show that trypsinogen 2 and trypsin 2-AAT reflect pancreatic injury after ERCP and support our previous findings on these markers in non-ERCP pancreatitis.18-22 In patients developing pancreatitis raised trypsinogen 2 concentrations were already evident one hour after ERCP and peaked at six hours, whereas the trypsin 2-AAT complex did not show a clear rise until 24 hours, at which time the trend was still increasing. This suggests that the inactive proenzyme is released into the circulation in the initial phase, while activation of trypsinogen and release of active trypsin into the circulation with complex formation are later events. This interesting finding is probably explained by the sequential interactions of trypsin with various protease inhibitors within and outside the pancreas.23 The main intrapancreatic trypsin inhibitor, pancreatic secretory trypsin inhibitor (PSTI), represents a “first-line” defence, whereas formation of the trypsin 2-AAT complex apparently starts later, when the inhibitory capacity of the intrapancreatic inhibitors is exceeded.24 The relatively late formation of trypsin 2-AAT has not been reported before, but is in keeping with the results of earlier studies on the dynamics of trypsin α₁ protease in non-ERCP pancreatitis.23

The patients with severe pancreatitis had the highest concentrations of trypsinogen 2 and trypsin 2-AAT, as was also seen in earlier studies,19-21 but this difference did not reach statistical significance, probably due to the limited number of patients in the severe pancreatitis group. These markers appear to indicate the degree of the primary pancreatic proteolytic insult during the induction of AP. In contrast to our earlier results in non-ERCP pancreatitis, trypsin 2-AAT was not superior to trypsinogen 2 in reflecting the severity of the disease. This is explained by the different time courses of these markers. Trypsinogen 2 peaked during the activation period, whereas trypsin 2-AAT was probably still increasing when the last sample was taken 24 hours after ERCP.

We suggest that a trypsinogen 2 concentration of over 3000 μg/l at six hours after ERCP should prompt the clinician to institute intensive monitoring, aggressive fluid transfusion, and antibiotic therapy.3 ,25 ,26 Whether this assay can be utilised to improve the prognosis of patients with ERCP induced pancreatitis remains to be evaluated in a further trial. A trypsin 2-AAT concentration of over 250 μg/l at 24 hours after ERCP could also be used, but trypsinogen 2 has the advantage of increasing earlier.

In a recent study of Gottlieb et al,27amylase and lipase levels were analysed two hours after the ERCP. The authors reported that the magnitude of the rise which occurred early correlated with the risk of post-ERCP pancreatitis. Since most outpatients undergoing ERCP are discharged within a few hours after the procedure, early detection or firm exclusion of the complication is highly desirable. The use of two hour values of amylase and lipase was recommended for identification of patients at risk and to indicate the need for longer follow up as inpatients. In the present study, the sensitivity of a threefold rise in trypsinogen 2 at one hour was 74% and the specificity 87%; the accuracy increased until six hours. The figures at one hour are comparable with the two hour amylase and lipase levels (sensitivity 82–92% and specificity 52–76%) reported in the study of Gottlieb et al.27 According to the results of this study, the optimal time lapse for the identification of patients at risk with trypsinogen 2 cannot be exactly determined; it is probably between one and six hours.

The high sensitivity and specificity of trypsinogen 2 and trypsin 2-AAT for post-ERCP pancreatitis was not unexpected, because pancreatic trypsinogen activation plays an essential role in the pathophysiology of pancreatitis.17 ,23 It can activate other pancreatic proteins and protease cascades, including the kallikrein-kinin system.28 Hereditary pancreatitis is reported to be caused by a mutation in the trypsinogen gene, which renders cationic trypsin resistant to proteolysis.29 In earlier studies, the sensitivity and specificity of trypsinogen 2 and trypsin 2-AAT concentrations to non-ERCP induced AP approached 100%.19 ,21

In a substantial proportion of patients, the baseline values of all the markers studied were elevated. Trypsinogen 2 and trypsin 2-AAT are not completely specific for acute pancreatitis. In earlier studies on 222 patients with abdominal pain of aetiologies other than acute pancreatitis, elevated levels of trypsinogen 2 over the upper reference limit were detected in about 5%.19-22 Trypsin 2-AAT concentrations were elevated in up to 10%,19 ,22respectively. However, naturally, the specificity of the markers is largely dependent on the cut off level used. Elevated levels have been detected in pancreatic, biliary tract, hepatocellular, and colorectal cancers as well as in chronic pancreatitis, pancreatic pseudocysts, and purulent cholangitis.22 ,30

In 60 patients in our series the indication for ERCP was an abdominal malignancy (table 1). Furthermore, almost all the patients in whom the indication for ERCP was acute biliary pancreatitis had high baseline concentrations of the markers. Thus the more frequent occurrence of high baseline concentrations of trypsinogen 2 and trypsin 2-AAT than of amylase (30% versus 21%) reduces the diagnostic utility of the two former markers for pancreatitis following ERCP. The diagnostic criterion to be used is therefore not the absolute value but the increase in concentration that is induced by ERCP. Consequently, the baseline value must be considered, when the risk of post-ERCP pancreatitis is analysed. With a threefold increase at six hours, the specificity was significantly better than that of amylase. There were seven patients in the post-ERCP pancreatitis group with an elevated baseline level, and all were detected using the threefold elevation as diagnostic method. Even with this method the accuracy of trypsinogen 2 might have been underestimated, since a few patients classified as having mild AP might have had a non-specific rise in amylase and abdominal pain for reasons other than AP.

The negative correlation between a high baseline concentration of trypsinogen 2 and the risk of postprocedural pancreatitis is interesting. It is probably explained by selection of a series with a high rate of underlying pancreatic pathology, especially fibrosis of the gland, which has been reported to reduce the risk of inflammation.31

The 10% incidence of pancreatitis following ERCP in our series is in accordance with previous prospective reports.2 ,10 ,12 The highly variable incidence of post-ERCP pancreatitis in the literature reflects differences in inclusion criteria and definitions of pancreatitis. The methods of data collection are also highly variable—prospective studies with frequent enzyme determinations and clinical follow up or retrospective studies without them.2 ,6 ,32 The relatively high incidence in the present study results from the prospective nature of the trial and the liberal criteria of AP. Furthermore, the proportion of therapeutic ERCPs in our series was somewhat higher (83%) than in most previous reports.10 ,33 Two patients developed late pancreatitis, which was not manifested by increasing trypsinogen concentrations within the short follow up period. Late development of pancreatitis after ERCP has been reported to occur in up to 15% of patients.34

In conclusion, our study indicates that rises in serum trypsinogen 2 and trypsin 2-AAT complex following ERCP accurately reflect the development of pancreatitis, the rise correlating with the severity of the disease. The clinical utility of high serum trypsinogen 2 as an early indicator of severe post-ERCP pancreatitis deserves to be evaluated in a larger prospective controlled study.