Int. J. Pancreatol. 9 Copyright 1991 by The Humana Press Inc. All rights of any nature whatsoever reserved. 0169-4197/91/8(3): 253-262/$2.00

Exocrine Pancreatic Function and Chronic Unexplained Dyspepsia A Case-Control Study R o s s C. S m i t h , *,~ N i c h o l a s J. T a l l e y , ~ O w e n F. Michael Jones, ~ and SheUa L o Waller ~

Dent,:

lDepartments of Surgery and Medicine, University of Sydney and Clinical Computing Division, Royal North Shore Hospital, Sydney, Australia," and 2National Centre for Epidemiology and Population Health, Australian National University

Received June 4, 1990; Revised August 6, 1990; Accepted August 14, 1990

Summary The aim of this study was to determine the prevalence of diminished mean tryptic activity (MTA) in duodenal juice of patients with nonulcer dyspepsia following injection of a Lundh test meal. Two separate studies were undertaken. The first examined a consecutive group of 100 patients with suspected pancreatic disease referred over a period of 2 yr. In these patients, receiver-operating-characteristic analysis was used to determine the point of best discrimination, which occurred at 7/xEq/mL/min when nonpancreatic disease patients were compared with those with chronic pancreatitis or pancreatic cancer. The second study involved 22 patients with endoscopically confirmed nonulcer dyspepsia. The MTA for the patients with unexplained dyspepsia did not differ from 17 healthy controls, but a further 16 pancreatic disease control patients had significantly decreased values ( p < 0 . 0 1 ) . Six patients with unexplained dyspepsia (27~ had an MTA of 7 ~ E q / m i n / m L or less, indicating impaired pancreatic function, which was more frequent than in healthy controls (p =0.02). These six patients had significantly different symptoms, being more likely to have pain radiating through to the back (p = 0.017) and pain waking them from sleep (p = 0,002), and less likely to have postprandial pain (p = 0.045). It was of interest that the alcohol intake was not greater in these six patients. It is concluded that pancreatic disease may explain the symptoms of some patients with nonulcer dyspepsia.

Key Words: Essential dyspepsia; Lundh test; pancreatitis. *Author to whom all correspondence and reprint requests should be addressed.

International Journal of Pancreatotogy

253

Volume 8, 1991

254

Smith et al.

INTRODUCTION Chronic pancreatic disease may be an underdiagnosed cause of nonulcer dyspepsia (NUD), since a relationship between NUD and reduced exocrine pancreatic function has been reported in two studies (1,2). Nonulcer dyspepsia, which can be defined as chronic upper abdominal pain or nausea where endoscopy reveals no evidence of peptic ulcer, oesophagitis, or cancer, is a heterogeneous syndrome that can be related to the irritable bowel syndrome, gastro-esophageal reflux, biliary tract disease, or other disorders (3). Once these recognizable diseases and disorders are excluded, the remaining NUD patients have chronic unexplained dyspepsia, provisionally termed essential dyspepsia (3,4). There have been no studies of pancreatic function in this subgroup of NUD patients. It was hypothesized that, if chronic pancreatic disease is an important cause of NUD symptoms, pancreatic function test abnormalities would be found in patients with essential dyspepsia. Measurement of the mean tryptic activity (MTA) of duodenal juice after a standard test meal, the Lundh test, is a recognized reliable and inexpensive test of pancreatic exocrine function in patients with severe pancreatic disease (5). One group has shown it to more readily diagnose minimal pancreatic disease than the secretin pancreozymin test (6), whereas others have shown it to be less diagnostic (7) in such patients. The Lundh test results during the routine investigation of pancreatic disease have been used to determine the value that allows the best discrimination between normal and severe pancreatic disease. In a further group of patients with nonulcer dyspepsia who have undergone a Lundh test meal, this point has been used to determine whether such patients have disturbance in pancreatic function. METHODS

Patient Population Study 1: Clinically Suspected Pancreatic Disease Patients Patients in whom pancreatic disease was suspected because of pain, diarrhea, jaundice, or unexplained loss of weight were referred for a Lundh test. On final diagnosis, patients were grouped into those without pancreatic disease (patient controls) and those with proven pancreatic disease, the diagnosis being independent of either the Lundh test or ultrasound results~ The criterion for proven chronic pancreatitis was pancreatic duct abnormality demonstrated by endoscopic retrograde pancreatogram or calcific changes on a plain abdominal radiogram, and the criterion for pancreatic cancer was either an operative biopsy or fine needle aspiration cytology. Patients in whom pancreatic disease could not be proven or disproved formed the suspected pancreatic disease group. Acute pancreatitis was characterized by an acute illness associated with either ultrasound evidence of pancreatic swelling or an acute rise in serum amylase > 1100 IU/L. It was likely that some of the patients categorized as acute pancreatitis had relapsing chronic pancreatitis, but evidence for this was not available. All but one patient with acute pancreatitis had a Lundh test six or more weeks after the acute episode.

International Journal o f Pancreatology

Volume 8, 1991

Pancreatic Function and Dyspepsia

255

Study 2: Nonulcer Dyspepsia Patients Informed consent was obtained from all patients and controls. The patient population was a consecutive group with endoscopically diagnosed NUD. The exclusion criteria have been previously defined (4). Patients with clinical evidence of the irritable bowel syndrome (discomfort associated with altered bowel function or relieved by defecation), gastro-esophageal reflux, or radiological or ultrasound evidence of cholelithiasis (all patients underwent oral cholecystography or ultrasonography) were excluded. In addition, three patients with a diagnosis of pancreatic disease made prior to index endoscopy were excluded. Twenty-two patients with essential dyspepsia entered this study. The mean age of the patients was 58.5 yr, and females outnumbered males 1.8:1. Five patients (23%) had endoscopic evidence of nonerosive chronic gastritis, and three patients (14070) had a past history of peptic ulcer that was diagnosed more than 6 mo before the study. At the time of entry into the study, endoscopy was carried out on these patients, yielding no evidence of ulcer. Data on symptoms and alcohol intake in the 6 mo before diagnosis were collected using structured questionnaires on the patients with dyspepsia (4,8). The patients were then followed up by telephone interview every 2 mo for 2 yr, when a questionnaire regarding symptoms and investigations was completed (9). This methodology is reliable (I0,11).

Control Populations Healthy Controls Seventeen healthy volunteers who were staff or friends of staff were studied. They had no history of dyspepsia, peptic ulceration, or alcoholism. Their mean age was 44 yr with a female to male ratio of 1.1:1.

Pancreatic Disease Patients Sixteen patients with chronic pancreatic disease entered into the study. Of these, 15 patients had chronic pancreatitis, and one had pancreatic carcinoma. The diagnoses were made at laparotomy (for four patients) or at endoscopic retrograde cholangiopancreatography (for four patients), whereas the remaining eight patients had clinical and radiological evidence of chronic calcific pancreatitis related to alcohol intake. The mean age was 47.8 yr, and females outnumbered males 1.3:1.

Pancreatic Function Testing L u n d h test was performed with two minor modifications of the method described previously. Calcium caseinate replaced casilan in the meal, and two consecutive 1-h collections of duodenal juice were made. The tryptic activity of each was measured by the pH method described by Wiggins (12) using BAEE (N-benzoyl-L-arginine) as the substrate. After an overnight fast, a flexible 109-cm 12 French duodenal tube was placed, under radiological control, into the duodenum. A standard Lundh test meal was then given containing 15 g calcium caseinate, 15 g corn oil, and

International Journal o f Pancreatology

Volume 8, 1991

256

Smith et al.

40 g glucose made up to 300 mL with water (13). The duodenal aspirate was collected for 2 h. At the end of the study, the tube position was checked radiologically. The pH of the aspirate was tested in study 2. If the aspirate had an acidic pH, it was discarded (six subjects). The remaining 55 subjects had the mean tryptic activity measured by titration using the substrate N-benzoyl-arginine ethyl ester (BAEE) (14). The results were expressed in ~Eq/min/mL.

Statistical Analysis Receiver-operating-characteristic analysis (ROC) (15) was used to determine the point of optimum discrimination between patients without pancreatic disease and those with chronic pancreatitis and carcinoma of the pancreas~ In study 2, the data were approximately normally distributed, and analysis of variance (ANOVA) was used to compare the means of the patient and control groups (16). Since the difference was significant, multiple comparison testing was performed using the studentized range. Then, taking 7/zEq/min/mL as the optimal cutoff value differentiating normal from abnormal pancreatic function, patients with essential dyspepsia and healthy controls were compared using Fisher's exact test (16). The symptoms and alcohol intake in patients with and without reduced pancreatic function were also compared with Fisher's exact test or Pearson's chi-squared test, as appropriate (16). Comparisons of age (using ANOVA) and sex (using the Pearsons chi-squared test) between the three groups were also performed (16). All probability values calculated were two-tailed. The significance level was set at 0.05. It was calculated that, to detect a difference of 10/zEq/min/mL between the patients with essential dyspepsia and the healthy controls, nine subjects per group would give a power of 90~ at the 0.05 level (17). The Medical Ethics Review Committee and the Radiation Safety Committee of the Royal North Shore Hospital approved the study.

Results Study 1 Over a 3-yr period, 100 patients who were referred for a Lundh test underwent a successful study. In a further 18, the test was unsuccessful, either because the patient vomited or because the tube failed to enter the duodenum. Of the 100 patients, 52 had a proven pancreatic disorder, including 14 with acute and 24 with chronic pancreatitis and 14 with carcinoma of the pancreas. Of the latter group, 11 had a primary and three a secondary tumor. A further five patients had no evidence of pancreatic disease, but had had previous gastric surgery that might have interfered with pancreatic function. Thirtythree did not have a pancreatic disorder. In the remaining ten patients, pancreatic disease could not be excluded. Six of these had alcoholic liver disease. The MTA for patients in each group appears in Fig. 1. There was clear separation of nonpancreatic disease from those with chronic pancreatitis and carcinoma of the pancreas. Using these three groups for ROC analysis, with

International Journal of Pancreatology

Volume& 1991

Pancreatic

Function

and Dyspepsia

257

40

c

30

i Ib

~" 2 0

W

eo

9

~

:~

~

.

!

e

r

:E 10

............~

i ............. .,........................... i ...............

.........................

............ 1 " 1 . . . . . . . . . . . . . . . . . . . . . . . . .

I ............ J

. . . . . . . . . . . " . . . . . . . . . . . . * ...............

;I;

9

|

"3

"

o

o t

1

CONTROL PG

~

t

I

i

AP

CP

CaP

NO FINAL DIAGNOSIS

Fig. 1. Mean:tryptic activity (MTA) in patients presenting with symptoms suggestive of resulting from pancreatic pathology in their different clinical groups. Control--pain not owing to pancreatic pathology; PG--previous partial gastrectomy; AP--acute pancreatitis; CP--chronic pancreatitis; CaP--carcinoma of the pancreas; No Final Diagnosis--pancreatic disease not fully excluded. The dotted lines denote the values of uncertain significance. 10o

uEq/ml/m

>-60

>I--1-

940 la/ m

20

0

"

0

20

40

60

80

1 O0

FALSE POSITIVE

Fig. 2. ROC curve for patients presenting with either pancreatic cancer or chronic pancreatitis and patients without pancreatic disease. the m a x i m u m discrimination at < 7 ~ E q / m i n / m L and with 81% sensitivity, no false positives and 93% correct diagnoses result (Fig. 2). Four of the 13 patients with acute pancreatitis, in w h o m a Lundh test was performed 6 wk after the acute episode, had an MTA below 7 ~ E q / m i n / m L , as did one o f the six patients with alcoholic liver disease, but this did not occur in those who had previous gastric surgery.

International Journal of Pancreatology

Volume 8, 1991

258

Smith et al. 40

.E 30 E

| o

9 ,J

{ 20 9

10

o | 9

9 9

o 9

-. Qo o * e9 o

~* o* 9

ESSENTIAL HEALTHY PANOREATIC DYSPEPSIA OONTROLS OONTROLS

Fig. 3. Mean tryptic activity (MTA) in the three patient groups. Study 2 The results of the pancreatic function testing (Fig. 3 and Table 1) showed no significant differences between the dyspepsia patients and control groups regarding sex, but the healthy controls were significantly younger than the patients with dyspepsia (p