CLINICAL CHEMISTRY
Original Article
Lipase and Pancreatic Amylase Activities
in Tissues and in Patients
with Hyperamylasemia
FRED APPLE, PH.D, PETER BENSON, M.D., LYNNE PREESE, MT, M.B.A.,
STEVEN EASTEP, M.D., LAURA BILODEAU, M.D.,
AND GREG HEILER, M.D.
admitted with possible acute pancreatitis. The serum lipase level
remained higher than normal longer than either the total amylase
and pancreatic amylase levels. In patients with hyperamylasemia
of pancreatic origin, a poor correlation was observed at admission
between serum pancreatic amylase and serum lipase. Not all
patients with elevated lipase had an elevated pancreatic amylase
level and vice versa. However, in every patient pancreatic disease
would have been detected by the elevation of either lipase or
pancreatic amylase levels. Diagnostic efficiency for pancreatic
disease using serum pancreatic amylase, lipase, and total amylase
tests was 94.1%, 76.5%, and 64.7%, respectively. These data
suggest that lipase and pancreatic amylase tests are specific for
the pancreas and might be considered replacements for total
amylase as the stat or routine laboratory test for the diagnosis
of pancreatic tissue injury. (Key words: Pancreatic injury; Lipase;
Pancreatic amylase; Serum enzyme assays) Am J Clin Pathol
1991; 96:610-614
Lipase, pancreatic amylase, and total amylase activities were
measured in nondiseased and diseased human pancreatic tissues
and in six different locations of the human digestive system. In
addition, it was determined whether serum lipase and pancreatic
amylase tests could replace the total amylase test to improved
diagnostic efficiency in the evaluation of acute pancreatitis in
hyperamylasemic patients. Nondiseased pancreatic tissue contained 4.5 times more lipase activity than total amylase activity.
Diseased pancreatic tissue contained less activity for both lipase
and total amylase compared to normal tissue. The total amylase
activity of the pancreas was comprised solely of pancreatic amylase. Tissue obtained from six different anatomic locations in
the digestive system contained 35 to 45 times less lipase and
total amylase activity compared to the pancreas. Total amylase
activity of the digestive system tissues were comprised of 25%
pancreatic and 75% salivary isoamylases. Lipase, pancreatic
amylase, and total amylase levels also were determined in serial
serum samples from 17 consecutive hyperamylasemia patients
Several organs contribute to amylase activities in serum,
and a variety of extrapancreatic disorders contribute to
an increase in total serum amylase activity.3 In normal
individuals, total serum amylase activity results from a
mixture of pancreatic isoamylase (found almost exclusively in the pancreas) and salivary isoamylase (found in
various organs). Overall pancreatic isoamylase constitutes
about 40% of the total amylase activity.4 The acinar cells
of the pancreas also are the primary source of serum lipase.
Thus, serum lipase activity should increase the specificity
for the diagnosis of acute pancreatitis.
The diagnosis of acute pancreatitis is not uncommonly
made
In a recent
From the Department of Laboratory Medicine and Pathology, Hen- in the presence of normoamylasemia.
nepin County Medical Center, Minneapolis, Minnesota.
study by Clavien and associates,5 68% of 65 normoamylasemic patients with acute pancreatitis were found to have
Received November 5, 1990; received revised manuscript and accepted
elevated lipase levels. Until recently, the clinical efficacy
for publication May 17, 1991.
of pancreatic amylase and lipase assays have been limited
Supported in part by Boehringer Mannheim Diagnostics and Robert
Ganz, M.D., for the collection of tissue specimens.
by the lack of specific and sensitive automated assays.2,6
Address reprint requests to Dr. Apple: Clinical Laboratories #812,
Hennepin County Medical Center, 701 Park Avenue South, Minneapolis, Improvements in technology for both the lipase and pancreatic isoamylase assays now afford laboratory workers
Minnesota 55415.
The diagnosis of acute pancreatitis has always been a
challenge to clinicians. During the past 10 years, techniques such as ultrasonography, laparotomy, endoscopic
retrograde cholangiopancreatography, and computed axial
tomography, as well as clinical findings have aided in the
evaluation of pancreatitis.' However, the diagnosis often
has relied on laboratory measurement of serum total amylase activity.2 With the increasing incidence of pancreatitis during the past several years, efforts have been concentrated on improving the specificity and simplicity of
enzyme assays for the diagnosis of acute pancreatitis.
610
APPLE ET AL.
Lipase and Pancreatic Amylase Activity in Hyperamylasemia
the ability to provide clinicians with improved diagnostic
enzymology testing for the evaluation of patients with
acute or chronic abdominal distress with possible pancreatitis, replacing the nonspecific serum total amylase
7-9
assay.
The present study had two primary goals. First, we
quantitated the activity of lipase, pancreatic amylase, and
total amylase in nondiseased and diseased human pancreases and in human digestive system tissues to determine
tissue specificity for each enzyme. Second, we determined
the clinical diagnostic efficiency of pancreatic amylase and
lipase assays in serum from patients admitted through the
emergency department with the possible diagnosis of acute
pancreatitis.
611
tomic locations of the digestive system, as well as from
six different pancreases.
Tissues were trimmed to remove excess connective tissue and immediately frozen at — 70°C and thawed overnight at 4°C before the analyses. All tissues were homogenized in 1- to 5-mL ice-cold phosphate buffer (0.2 mol/
L potassium phosphate, pH 7.0) and centrifuged to remove cellular debris. The supernatants were analyzed for
total amylase, pancreatic amylase, and lipase activities,
as described below (in the Patients section). Results are
expressed as U/g wet weight tissue or as U/mg total protein. Agarose gel electrophoresis was used to determine
qualitatively the distribution of tissue isoamylases.10
Patients
METHODS
Tissues
Pancreases were harvested within 9 hours after death
from two male subjects (A and B) who died as a result of
head trauma in motor vehicle accidents. Neither had clinical or gross indication of pancreatitis. Pancreases also
were harvested within 7 hours after death from a female
subject (C) and a male subject (D) who died as a result of
trauma in a motorcycle accident and from a drug overdose, respectively. Both of the latter subjects had known
histories of alcoholism and chronic pancreatitis. Each of
the latter intact pancreases at harvest had gross indications
of patchy fat necrosis and hemorrhage. Pancreas fragments
(100-150 mg) were obtained from six anatomic locations
from the head (location 6) to tail (location 1) of the pancreas (Table 1). Visible areas of necrosis were avoided
during tissue sampling. In addition, tissue specimens (5
mg) also were obtained at biopsy from six different ana-
TABLE 1. LIPASE AND AMYLASE ACTIVITIES IN
NORMAL AND DISEASED PANCREATIC TISSUE
FROM FOUR SUBJECTS
Lipase, U/g
Nondiseased
Anatomic
Location
1
2
3
4
5
6
Mean
SD
A
1624
1772
1492
2124
1904
1900
1802
224
B
1700
1730
1555
2010
2040
1985
1836
201
Amylase, U/g
Diseased
Nondiseased
Diseased
C
D
A
B
C
D
540
2320
1120
1535
1938
824
1380
678
950
1850
820
1295
1685
1108
1301
438
320
388
304
524
428
444
401
82
380
382
320
595
480
425
434
97
180
432
344
132
112
90
215
140
265
295
422
206
226
252
277
77
Location I denotes the tail and location 6 denotes the head of the pancreas; results are expressed
as U/g wet weight of tissue; amylase activity is 100% pancreatic amylase.
SD • standard deviation.
Serum was obtained at admission as well as approximately 12, 24, and 48 hours after admission in 17 consecutive hyperamylasemic patients admitted through the
emergency department with the possible clinical diagnosis
of acute pancreatitis. Serum was frozen at - 20°C until
time of assay. For enzyme clearances, mean enzyme values were calculated at each time point (Fig. 2). The diagnosis of pancreatitis and admission to the hospital initially was based on a history of abdominal pain that was
consistent with acute pancreatitis. The enzyme values obtained in the emergency department were not used to
establish the diagnosis but were used for clinical confidence. The etiology of the pancreatitis in most cases (9
of 14) was alcohol abuse. The etiologies of the remaining
five cases were unknown. The final diagnoses in 35% of
cases were confirmed by laparotomy, computed tomography, ultrasonography, or at autopsy. Serum total amylase and lipase (with colipase cofactor) activity were determined on the automated Kodak Ektachem 400 analyzer (Eastman Kodak, Rochester, NY) according to the
manufacturer's protocols. Reference ranges were 37-117
U/L and 12-198 U/L, respectively. Pancreatic isoamylase
was determined on a Cobas-Bio (Roche Diagnostic Systems, Montclair, NJ) using the Boehringer Mannheim
(Indianapolis, IN) enzymatic colorimetric test, pancreatic
alpha—amylase p-nitrophenylmalthoheptaoside.9 In this
assay, the activity of human salivary isoamylase is inhibited by a monoclonal antibody that does not affect pancreatic isoamylase. The activity of pancreatic isoamylase
is then measured at 37°C using a p-nitrophenyl-D-maltopeptaoside substrate. The reference range was 0-115
U/L. (It should be noted that under the current assay
conditions, total amylase and pancreatic amylase levels
are measured by different assays; therefore, 1 U/L of total
amylase does not equal 1 U/L of pancreatic isoamylase.)
Diagnostic efficiency, sensitivity and predictive values of
a positive test were calculated.
Vol. 96 • No. 5
612
CLINICAL CHEMISTRY
Original Article
2200 -
A, NORMAL
G
LIPASE, U/G
D
a
2000 -
• a
1800-
a
•
a
•
1600 -
D
•
3
1400-
1
300
i
1
400
600
500
AMYLASE, U/G
3000
3
B, DISEASED
2000
•
D
a
LU
<
Q.
•
1000
—i
0
1
100
•
1
•
200
•
1
1
300
1
1
400
1
500
AMYLASE, U/G
FIG. 1. Relationship between amylase and lipase activity (U/g) in tissues
obtained from nondiseased (normal) and diseased pancreases.
RESULTS
Table 1 shows the lipase and amylase activities from
six distinct anatomic locations in nondiseased and diseased pancreatic tissue from four subjects. From tail to
head in nondiseased tissue (subjects A and B), lipase activity (1819 U/g) was 4.5 times greater than amylase activity (418 U/g) per gram of wet weight tissue. In tissue
obtained from subjects (C and D) with a history of chronic
pancreatitis, there was substantially less activity for both
lipase (1340 U/g) and amylase (246 U/g), representing
26% and 41% depletions. Fourfold differences in amylase
and lipase activities existed within each diseased pancreas.
Figure 1 shows that although there was an excellent correlation (r = 0.90) between tissue lipase and amylase in
nondiseased pancreases, there was a very poor correlation
(r = 0.17) in enzyme activity in the diseased tissue. Table
2 shows the total amylase, pancreatic amylase, and lipase
activities is six different locations in the digestive system
compared to the pancreas. The pancreas contained 35 to
45 times more activity per milligram of total protein for
total amylase, lipase, and pancreatic amylase than any of
the six different digestive system locations. The total amylase activity of the pancreas consisted solely of pancreatic
amylase, and the digestive system total amylase activity
was comprised of less than 50% pancreatic amylase at all
six sites.
The serum enzyme clearance profiles for total amylase,
pancreatic isoamylase, and lipase from the 17 consecutive
emergency department admissions with a possible diagnosis of acute pancreatitis are shown in Figure 2. Although
there was a good correlation of total amylase with pancreatic isoamylase for all samples (r = 0.81), the correlation improved (r = 0.95) when samples with normal
lipase were excluded. A poor correlation existed in admission specimens between serum pancreatic isoamylase
and lipase activities, as shown in Figure 3. Table 3 shows
the sensitivity, diagnostic efficiency, and predictive value
of a positive test for all three assays based on the admission
serum specimen. The diagnostic sensitivity and efficiency
with this small group of patients for acute pancreatitis
was best using serum pancreatic amylase, followed by lipase and total amylase.
DISCUSSION
The diagnosis of pancreatitis often is difficult to make
because this condition must be differentiated from other
abdominal disorders with similar clinical features.' Until
TABLE 2. TOTAL AMYLASE, PANCREATIC AMYLASE, AND LIPASE ACTrVITY IN THE PANCREAS AND
IN SIX DIFFERENT ANATOMIC LOCATIONS OF THE HUMAN DIGESTIVE SYSTEM
Location
Total Amylase
Pancreatic Amylase
Lipase
Pancreas (n = 6)
Duodenal bulb (n = 4)
Pyloric antrum (n = 11)
Angular notch (n = 1)
Corpus (n = 15)
Gastroesophagel junction (n = 1)
Esophagus(n = 3)
100% pancreatic
0.60 (0.73)
0.07(0.13)
0.010
0.08(0.15)
0.001
0.94(1.5)
9.0 (7.0)
0.28 (0.45)
0.02 (0.03)
0.001
0.01 (0.02)
0.001
0.02 (0.3)
7.8 (3.2)
0.13(0.22)
0.03 (0.04)
0.010
0.21 (0.36)
0.001
0.01 (0.01)
' Results are expressed as mean (standard deviation) of activities, U/mg total protein.
AJ.C.P. •November 1991
APPLE ET AL.
Lipase and Pancreatic Amylase Activity in Hyperamylasemia
available, improved serum lipase assay (with colipase cofactor) and pancreatic amylase assay (which uses inhibiting
monoclonal antibodies against salivary amylase) have
been shown clinically to be interchangeable tests for the
diagnosis of pancreatitis. 5 ' 7 ' 9121415 Although it is not a
thoroughly representative population for studying diagnostic efficiency, the current study demonstrated that
serum pancreatic amylase and serum lipase were the most
clinically sensitive and diagnostically efficient assays in
patients treated in the emergency department for possible
acute pancreatitis. This correlated with several other larger
population studies. Lott and co-workers7 showed that the
clinical sensitivity and specificity of lipase assay (100%
and 62%, respectively) was much better than total amylase
assay (96% and 34%, respectively) in 175 patients admitted
for acute pancreatitis. Clavien and co-workers5 showed
that acute pancreatitis and normamylasemia were not an
uncommon combination; however, the serum lipase level
was elevated in 68% of 65 normamylasemic cases. For
the emergency diagnosis of acute pancreatitis in more than
250 cases of varied clinical findings, lipase and pancreatic
amylase were considerably more sensitive and specific,
with few to no false-positive results when compared to
total amylase.14 Finally, Van Lente and associates15 have
shown that lipase and pancreatic amylase are interchangeable serum markers, with high clinical sensitivity
and specificity for acute pancreatitis. These findings demonstrate that the measurement of total amylase is not asefficient as the new and improved serum lipase- and pancreatic-specific amylase assays.
Day After Admission to Hospital
o
o
0
200
400
600
LIPASE U/L
800
613
1000
1200
FlG. 2. Enzyme clearances after the onset of abdominal pain in patients
admitted for pancreatitis for serum lipase, pancreatic amylase, and total
amylase activities; ULN = upper limit of normal reference range. Each
time point represents the mean value of each enzyme for all 17 patients.
FlG. 3. Relationship between admission serum pancreatic amylase and
lipase activities in all patients admitted with possible diagnosis of acute
pancreatitis (r = 0.16).
recently, the serum total amylase test has been the diagnostic laboratory test of choice in this setting. However,
several studies (as well as the tissue and serum results
described in the current study) show that serum lipase
and serum pancreatic (specific) amylase assays are more
sensitive and specific for the diagnosis of pancreatic in2,4,5,7,11-15
The rapid, simple-to-perform, and readily
jury
Multiple factors contribute to the absence of hyperamylasemia in the presence of hyperlipasemia. These include an earlier return to normal for serum amylase during
hospitalization (Fig. 2) and the inability of diseased pancreases to release amylase (Table 1). In the present investigation, we describe two additional mechanisms. First,
lipase activity was four times greater than amylase activity
in the pancreas (Table 1). Second, pancreatic tissue obtained from chronic pancreatitis subjects demonstrated a
substantial decline in both amylase and lipase activity,
with amylase activity showing a greater decrease compared
to lipase (41 % versus 26%; Table 1, Fig. 1). These findings
TABLE 3. SENSITIVITY, EFFICIENCY, AND PREDICATIVE
VALUES OF SERUM PANCREATIC ENZYME ASSAYS
Sensitivity (%)
Efficiency (%)
Positive predictive
value (%)
Vol. 96 • No. 5
Total
Amylase
Lipase
Pancreatic
Amylase
64.3
64.7
87.5
76.5
92.9
94.1
90.0
85.7
100
614
CLINICAL CHEMISTRY
Article
would explain why acute pancreatitis, normoamylasemia,
and hyperiipasemia are not an uncommon finding in
clinical practice. Regarding the specificity of elevated
serum lipase and pancreatic amylase, ourfindingsshowed
that only the pancreas contained substantial activities of
either enzyme when compared to several other anatomic
locations in the digestive system (Table 2). To our knowledge, this is the first report to measure tissue lipase activities in the pancreas and digestive system. Our finding of
a small quantity and a mixture of pancreatic and salivary
isoamylases in nonpancreatic tissue correlated with the
findings of Whitten and associates.3
From chart review in the current study, it appears that
elevations of either pancreatic amylase or lipase represent
direct pancreatic injury or secondary release from the
pancreas due to other pathologic conditions. Combined
measurement of pancreatic amylase with lipase or total
amylase with lipase did not increase the sensitivity or diagnostic efficiency (89%, 85%, and 86%, respectively).
Specificity calculations were not performed in the current
study because of the high prevalence of disease (pancreatitis, 14 of 17 patients) in the small patient population
studied.
From ourfindingswe conclude that (1) lipase and pancreatic amylase levels are specific for the pancreas; (2) for
the diagnosis of acute pancreatitis, clinical sensitivity and
specificity are improved by measuring lipase and pancreatic amylase levels; and (3) ourfindings,together with
those of others, provide sufficient information to consider
replacement of the measurement of total serum amylase
activity as an indicator of acute pancreatitis. However,
the elimination of the measurement of serum total amylase in clinical laboratories will depend on the general
availability of instrumentation for the measurement of
serum lipase (colipase) and serum pancreatic amylase levels. When serum lipase and pancreatic amylase activities
are found to be elevated, it should suggest some form of
pancreatic tissue injury.
REFERENCES
1. Moossa AR. Diagnostic tests and procedures in acute pancreatitis.
N Engl J Med 1984; 311:639-643.
2. Kolars JC, Ellis CJ, Levitt MD. Comparison of serum amylase pancreatic isoamylase and lipase in patients with hyperamylasemia.
DigDisSci 1984;29:289-293.
3. Whitten RO, Chandler WL, Thomas MG, et al. Survey of amylase
activity and isoamylases in autopsy tissue. Clin Chem 1988; 34:
1552-1555.
4. Panteghini M, Pagani F. Diagnostic value of measuring pancreatic
lipase and the P3 isoform of the pancreatic amylase isoenzyme
in serum of hospitalized hyperamylasemic patients. Clin Chem
1989;35:417-421.
5. Clavien PA, Robert J, Meyer P, et al. Acute pancreatitis and normoamylasemia. Ann Surg 1989; 210:614-620.
6. Koehler DF, Eckfeldt JH, Levitt MD. Diagnostic value of routine
isoamylase assay of hyperamylasemic serum. Gastroenterology
1982; 82:887-890.
7. Lott JA, Patel ST, Sawhney AK, et al. Assays of serum lipase: Analytical and clinical considerations. Clin Chem 1986; 32:12901302.
8. Rosenblum JL. Direct, rapid assay of pancreatic isoamylase activity
by use of monoclonal antibodies with low affinity for macroamylasemic complexes. Clin Chem 1988; 34:2463-2468.
9. Tietz NW, Burling A, Gerhardt W, et al. Multicenter evaluation of
a specific pancreatic isoamylase assay based on a double monoclonal-antibody technique. Clin Chem 1988; 34:2096-2101.
10. Gillard BK. Quantitative gel-electrophoretic determination of serum
amylase isoenzyme distributions. Clin Chem 1979; 25:1919-1923.
11. Werner M, Steinberg WM, Pauley C. Strategic use of individual and
combined enzyme indicators for acute pancreatitis analyzed by
receiver-operator characteristics. Clin Chem 1989; 35:967-971.
12. Kazmierczak SC, VanLente F. Incidence and source of hyperamylasemia after cardiac surgery. Clin Chem 1988; 34:916-919.
13. Hafkenscheid JC, Hessels M, Wetzels JF. Clearance of pancreatic
and salivary amylase in normal subjects. Clin Chem 1985; 31:
162-163.
14. d'Eril GM, Bosoni T, Lesi C. Pancreatic amylase in serum for differential diagnosis of acute pancreatitis and acute abdominal diseases. Clin Chem 1989; 35:2142-2143.
15. VanLente F, Kazmierczak SC. Immunologically-derived pancreatic
amylase, pancreatic lipase, and total amylase compared as predictors of pancreatic inflammation. Clin Chem 1989; 35:1542.