Serum complex of trypsin 2 and (alpha)(sub 1) antitrypsin as diagnostic and prognostic marker of acute pancreatitis: clinical study in consecutive patientsBMJ 1996; 313 doi: https://doi.org/10.1136/bmj.313.7053.333 (Published 10 August 1996) Cite this as: BMJ 1996;313:333
- Johan Hedstrom, research fellowa,
- Vesa Sainio, consultantb,
- Esko Kemppainen, lecturerb,
- Reijo Haapiainen, associate professorb,
- Eero Kivilaakso, professorb,
- Tom Schroder, consultantb,
- Jari Leinonen, research fellowa,
- Ulf-Hakan Stenman, associate professora
- a Department of Clinical Chemistry and Second Department of Surgery, Helsinki University Central Hospital, FIN 00290 Helsinki, Finland
- b Department of Surgery, Helsinki University Central Hospital
- Correspondence to: Dr U-H Stenman, Department of Clinical Chemistry.
- Accepted 16 May 1996
Objective: To estimate the usefulness of serum concentrations of the complex of trypsin 2 and (alpha)1 antitrypsin in diagnosing and assessing the severity of acute pancreatitis in comparison with serum C reactive protein, amylase, and trypsinogen 2 concentrations (reference markers).
Design: Markers were measured in consecutive patients admitted with acute abdominal pain that was either due to pancreatitis or to other disease unrelated to the pancreas (controls).
Setting: Department of surgery of a teaching hospital in Helsinki.
Subjects: 110 patients with acute pancreatitis and 66 with acute abdominal diseases of extrapancreatic origin. On the basis of the clinical course, acute pancreatitis was classified as mild (82 patients) or severe (28 patients).
Main outcome measures: Clinical diagnosis of acute pancreatitis and severity of the disease.
Results: At admission all patients with acute pancreatitis had clearly raised concentrations of trypsin 2-(alpha)1 antitrypsin complex (32 μg/l), whereas only three of the controls had such values. Of the markers studied, trypsin 2-(alpha)1 antitrypsin complex had the largest area under the receiver operating curve, both in differentiating acute pancreatitis from extrapancreatic disease and in differentiating mild from severe disease.
Conclusions: Of the markers studied, trypsin 2-(alpha)1 antitrypsin complex was the most accurate in differentiating between acute pancreatitis and extrapancreatic disease and in predicting a severe course for acute pancreatitis.
This complex can be accurately measured in a sensitive immunoassay
In this study the diagnostic and prognostic accuracy of serum concentrations of trypsin 2-(alpha)1 antitrypsin complex was determined in acute pancreatitis
The complex was more accurate than trypsinogen 2, C reactive protein, and amylase in differentiating between acute pancreatitis and extrapancreatic disease and in predicting a severe course for the disease
If the immunoassay could be automated determination of concentrations of trypsin 2-(alpha)1 antitrypsin complex could greatly improve the diagnosis of this common and potentially lethal disease
Typical clinical features with raised serum amylase concentrations are still the main diagnostic criteria in acute pancreatitis, but they are helpful in only 80% of cases.1 In the remainder, the diagnosis is possible only on the basis of the medical history, clinical course, and findings on radiology (after ultrasonography and computed tomography) and laparotomy.2 3 4 5 6 Concentrations of amylase may not be high initially, and high concentrations may occur in abdominal disorders that can be confused with acute pancreatitis.7 8
Acute pancreatitis has an unpredictable outcome, with serious complications in 20-30% of cases9 10 and a mortality of 10-20%.10 11 Rapid assessment of the severity of disease facilitates prompt institution of supportive treatment and improves prognosis.1
Proteolytic enzymes may have a major role in the pathophysiology of pancreatitis,12 13 14 and serum trypsinogen concentration seems to reflect pancreatic damage.15 The ratio of the two major trypsinogen isoenzymes, trypsinogen 1 and trypsinogen 2, in normal plasma is 3.8,16 but in acute pancreatitis the serum concentrations of trypsinogen 2 are greatly increased.15
Pathological intrapancreatic activation of trypsinogen to trypsin occurs in acute pancreatitis.17 In blood, active trypsin is inactivated by (alpha)2 macroglobulin and (alpha)1 antitrypsin.14 If the inhibitory capacity is exceeded activation of other pancreatic enzymes may be triggered, leading to severe complications.17
The complex of trypsin and (alpha)1 antitrypsin can be specifically measured with an assay based on a capture antibody specific for trypsin 2 and a detector antibody recognising (alpha)1 antitrypsin. Concentrations of trypsinogen 2 reflect leakage of enzymes from the pancreas and those of trypsin 2-(alpha)1 antitrypsin complex the release of active trypsin 2.18 Thus concentrations of the complex may reflect the severity of the disease. Activation of trypsinogen can be estimated by measurement of trypsinogen activation peptides in urine, and this method has been successfully used to assess the severity of acute pancreatitis.19
We investigated the accuracy of serum concentrations of trypsin 2-(alpha)1 antitrypsin complex in diagnosing and assessing the severity of acute pancreatitis in comparison with those of the reference markers trypsinogen 2, C reactive protein, and amylase.
Patients and methods
We studied a consecutive series of 110 patients with acute pancreatitis and 66 with acute gastrointestinal extrapancreatic disease admitted to this department of surgery from March 1992 to November 1993. The final diagnosis of pancreatitis was based on findings of upper abdominal pain accompanied by the typical appearance of acute pancreatitis on ultrasonography or computed tomography. If severe pancreatitis was suspected, contrast enhanced computed tomography was used to detect necrosis of the pancreas. Ultrasonography or computed tomography was performed in 103 (94%) patients with acute pancreatitis. In seven patients the diagnosis was based on clinical data alone (table 1).
The median age for all patients with acute pancreatitis was 43 years (range 21-91); 88 were men. The cause of the disease was regarded as alcohol misuse in 82 patients (table 1). A biliary origin was deduced when gall stones were seen on ultrasonography, endoscopic retrograde pancreatography, or laparotomy or at necropsy. The patients were classed according to clinical outcome as having mild (82 patients) or severe disease (28 patients).
All patients with mild acute pancreatitis improved spontaneously with conservative treatment, and none developed local pancreatic or systemic complications. Their median stay in hospital was six days (table 1).
Patients with severe acute pancreatitis had one or more major local or systemic complications of the disease—that is, respiratory failure (17 patients), renal failure (eight patients), septicaemia (seven patients), abscess (two patients), and necrosis of pancreas (17 patients). Four patients died of haemorrhagic pancreatitis, which was verified at necropsy. Twenty six patients required intensive care, and seven patients underwent surgery, either necrosectomy or distal pancreatic resection. Necrosis of the pancreas was confirmed by contrast enhanced computed tomography, by laparotomy, or at necropsy. The median stay in hospital was 23 days (table 1).
Patients with acute abdominal disease of extrapancreatic origin were categorised into the control group on the basis of clinical, radiographic, endoscopic, or surgical findings. The diagnoses were acute appendicitis (11 patients), acute biliary disease (10 patients), intestinal obstruction (six patients), peptic ulcer or oesophagogastritis (seven patients), diverticular disease (five patients), gastroenteritis (five patients), genitourinary disease (three patients), and other gastrointestinal disease (19 patients). Their median age was 45 years (range 20-90). Thirty eight patients were women.
First samples were collected within 12 hours of admission in all control patients and in 35 patients with acute pancreatitis. In 64 patients the initial sample was collected 12-24 hours after admission and in 11, 24-48 hours after admission. Later blood samples were collected daily between 0700 and 0800. Consecutive samples were obtained from patients with acute pancreatitis during the first three days after admission. Serum samples were stored at −20°C until analysed.
Amylase was measured by an enzymatic method with a Hitachi 705E analyser using Amylase EPS reagents (Boehringer Mannheim, Mannheim, Germany). The reference range was 70-300 U/l (median 180 U/l).
Trypsinogen 2 and trypsin 2-(alpha)1 antitrypsin complex were determined by time resolved immunofluorometric assays by a technician who was unware of the clinical diagnosis. The reference range for trypsinogen 2 was 18-90 μg/l and that for trypsin 2-(alpha)1 antitrypsin complex 2.3-12 μg/l.16
Serum C reactive protein was assayed by an immunoturbidometric method with antiserum and standards from Orion Diagnostica (Espoo, Finland). The upper reference limit was 10 mg/l.20
The ability of various tests to differentiate between mild and severe acute pancreatitis and non-pancreatic disease was estimated on the basis of sensitivity and specificity at clinically relevant cut off points. The validity of the tests was further evaluated by analysis of receiver operating characteristic curves. The area under the curve of the receiver operating curve describes the accuracy of the test, 1 indicating 100% sensitivity and specificity and 0.5 no discriminatory power.21 Specificity and odds ratios were calculated by Fisher's test with cut off points that gave 90% and 95% sensitivity. The specificities for different markers were compared using McNemar's test. High sensitivity values were selected because of the importance of not missing severe pancreatitis.
There were no significant differences in concentrations of trypsin 2-(alpha)1 antitrypsin complex or trypsinogen 2 between men and women in either group. Therefore men and women were considered together. In acute pancreatitis concentrations of trypsinogen 2 and trypsin 2-(alpha)1 antitrypsin complex increased much more than did those of C reactive protein and amylase. There was also less overlap in concentrations of the complex than in those of the other markers between patients and controls. In all patients with acute pancreatitis the concentrations of the complex during the first two days were above the upper reference limit (12 μg/l) (fig 1). All patients with acute pancreatitis, but only three of the controls, had a trypsin 2-(alpha)1 antitrypsin value >/=32 μg/l. When the patients were compared with the controls, trypsin 2-(alpha)1 antitrypsin had the largest area under the curve of the markers studied (table 2, fig 2). When patients with severe pancreatitis were compared with those with mild disease in the 35 samples obtained within 12 hours of admission, trypsin 2-(alpha)1 antitrypsin had a larger area under the curve than did C reactive protein (table 2). When all samples taken within 24 hours (n = 99) and 48 hours (n = 110) of admission were considered, the accuracy of trypsin 2-(alpha)1 antitrypsin remained high, while that of C reactive protein increased and approached the accuracy of the complex (table 2). The median values and range of values for the various markers are shown in table 3.
Specificity and odds ratios were calculated at cut off values giving 95% and 90% sensitivity. At high sensitivity trypsin 2-(alpha)1 antitrypsin complex was more specific than C reactive protein in differentiating between severe and mild acute pancreatitis in samples taken 12-48 hours after admission (P<0.01 with McNemar's test) (table 4). Thus, trypsin 2-(alpha)1 antitrypsin complex was the best in differentiating acute pancreatitis from extrapancreatic disease and mild from severe acute pancreatitis. Amylase was poor in differentiating severe from mild acute pancreatitis, whereas C reactive protein, as expected, did not differentiate acute pancreatitis from extrapancreatic disease (figs 1 and 2).
A larger proportion of patients with their first attack had severe disease than those with a recurrent attack (P<0.0001).
TIME COURSE PROFILE OF THE MARKERS
Time course profiles of trypsin 2-(alpha)1 antitrypsin complex (10 patients) and trypsinogen 2 (10 patients) showed similar patterns (fig 3). In severe acute pancreatitis concentrations of trypsin 2-(alpha)1 antitrypsin complex and trypsinogen 2 mostly peaked in the initial sample and slowly decreased but remained high during the following days. In two out of three patients with mild acute pancreatitis the peak was observed on the second day. One patient with increasing concentrations of trypsin 2-(alpha)1 antitrypsin complex and trypsinogen 2 died of necrotising acute pancreatitis confirmed at necropsy. Eight patients out of 10 showed decreasing concentrations of trypsin 2-(alpha)1 antitrypsin complex and trypsinogen 2 when the initial sample was compared with that on day 3.
C reactive protein concentrations (18 patients) were often low initially and tended to increase in mild acute pancreatitis. C reactive protein increased rapidly in eight out of nine patients, the peak being observed on day 3 in seven of the nine patients (fig 3). Five patients with severe pancreatitis had had a very high concentration of C reactive protein (>200 mg/l) at presentation.
Peak amylase (n = 13) values were observed in the initial sample in all cases of severe acute pancreatitis and in five out of eight cases of mild acute pancreatitis. Eleven of the 13 patients had decreasing concentrations of amylase. Interestingly, some patients with severe disease had low amylase concentrations (fig 3).
Clearly raised serum concentrations of trypsin 2-(alpha)1 antitrypsin complex were found at presentation in all cases of acute pancreatitis, and a normal serum concentration of this complex excluded the disease. The lowest concentration of trypsin 2-(alpha)1 antitrypsin complex in acute pancreatitis (32 μg/l) was about three times the upper reference limit in healthy controls. Among the other markers, trypsinogen 2 and C reactive protein concentrations were normal in three cases each, and amylase values were normal in six. Thus trypsin 2-(alpha)1 antitrypsin complex differentiated better between pancreatic and extrapancreatic disease than did trypsinogen 2, C reactive protein, and amylase concentrations. We assumed that trypsin 2-(alpha)1 antitrypsin complex would reflect pancreatitis induced release of active trypsin 2 into the circulation. Our results suggest that this is the case. All patients with severe acute pancreatitis requiring intensive care had a trypsin 2-(alpha)1 antitrypsin complex concentration above 200 μg/l at presentation. A single clearly false increase in this concentration was observed (fig 1). Interestingly, this patient had ovarian cancer, a tumour which may produce trypsin.22
PREDICTION OF DISEASE SEVERITY
At presentation trypsin 2-(alpha)1 antitrypsin complex concentration predicted the severity of acute pancreatitis much better than trypsinogen 2 and amylase concentrations. The serum concentration of C reactive protein increased with time, and when all the samples taken within 48 hours of admission were considered C reactive protein and trypsin 2-(alpha)1 antitrypsin complex had similar accuracy. However, during the first 12 hours after admission trypsin 2-(alpha)1 antitrypsin was more accurate than C reactive protein (area under the curve 0.82 v 0.73). Amylase concentration did not reflect the severity of acute pancreatitis, and in many patients the values were lower in severe than in mild disease. The area under the curve for amylase for differentiation between mild and severe acute pancreatitis was 0.57—that is, close to 0.5, which indicates no discriminative power.
Because early identification of patients with severe acute pancreatitis is desirable, we have emphasised sensitivity in our analysis. Especially at cut off values giving high sensitivity trypsin 2-(alpha)1 antitrypsin complex was more accurate than C reactive protein. Time course profiles showed that the concentrations of trypsin 2-(alpha)1 antitrypsin complex and trypsinogen 2 were already high at presentation and mostly decreased during the first three days of admission, although they were still greatly raised on the third day. The concentrations of C reactive protein were often low at presentation but increased after one to two days. Therefore the accuracy of C reactive protein was low during the first 12-24 hours after admission. The concentrations tended to be low, with many values within the normal range at admission. Follow up samples were not available for all patients. Therefore the time course studies should be interpreted with caution.
There was a higher proportion of men among patients with acute pancreatitis whereas there was an excess of women in the control group. However, there were no significant differences in the marker concentrations between the sexes. In fact trypsinogen values in female controls tended to be higher than those in the male controls (P = 0.15), whereas the opposite was true in the patients with acute pancreatitis (P = 0.45-0.50). Therefore the difference in sex distribution did not improve the accuracy of trypsin 2-(alpha)1 antitrypsin complex as a marker.
We conclude that of the markers studied, trypsin 2-(alpha)1 antitrypsin complex is the most accurate in differentiating between acute pancreatitis and extrapancreatic disease and in predicting a severe course of the disease at presentation. Trypsin 2-(alpha)1 antitrypsin complex and trypsinogen 2 were determined by double antibody sandwich assay using time resolved fluorescence for detection. However, because the typical concentrations are 100 to 1000 times the detection limit, the method can easily be adopted to other assay formats. The assay time is routinely two hours, but with automation the assay could easily be performed in minutes. If available on an automatic analyser on a 24 hour basis, the assay for trypsin 2-(alpha)1 antitrypsin complex could greatly improve the diagnosis of this common and potentially lethal disease.
Funding Grants from the Finska Lakaresallskapet, the Sigrid Juselius Foundation, and the Finnish Academy of Sciences.
Conflict of interest None.