|Home | About | Journals | Submit | Contact Us | Français|
Sphincter of Oddi dysfunction is a complex and poorly understood syndrome that usually manifests as pain of apparently biliary or pancreatic origin in the absence of an organic cause after conventional investigations. Transduodenal sphincteroplasty with transampullary septectomy (TDS/TAS) is a therapeutic option.1 Although biliary manometry is the gold standard diagnostic test, the morphine–prostigmine provocation (Nardi) test is sometimes used to screen patients. A fourfold increase in either serum amylase or lipase and reproduction of pain after intramuscular injection of 10 mg morphine (to induce sphincteric spasm) and 1 mg prostigmine (to stimulate pancreatic exocrine secretions) is considered a positive test.2,3 However, the enzymatic changes may also occur in healthy subjects and in patients with irritable bowel syndrome,3,4 bringing the clinical value of the test into question. Conversely, when the Nardi test was carried out on 70 patients with chronic abdominal pain, enzyme elevation was detected in only 23 patients (33%), 16 (70%) of whom had ampullary stenosis.5
We undertook these studies to characterise the response to the Nardi test in 20 healthy young adult subjects and 24 patients with sphincter of Oddi dysfunction before and after TDS/TAS. Our diagnosis of sphincter of Oddi dysfunction was based on clinical history, HIDA scanning, cross sectional imaging, endoscopic retrograde cholangiopancreatography, and exclusion of luminal pathology. All patients had significant debilitating upper abdominal pain requiring regular strong analgesia and had a poor quality of life. Patients were Milwaukee class I or II and had microscopic stenosis/fibrosis, but biliary manometry was not done routinely. We excluded from the healthy subject study those with recurrent abdominal pain, gallstones, or previous abdominal surgery, and those on regular medicines or with allergies.
Subjects abstained from alcohol for at least 24 hours. After sampling blood for baseline concentrations of amylase and lipase, subjects were given 1 ml of 0.9% saline intramuscularly as a sham test. Blood was resampled one hour after this injection. Subjects were then given 10 mg morphine and 1 mg prostigmine intramuscularly simultaneously and blood samples were collected hourly for five hours. Abdominal pain was quantified using visual analogue pain scores. Reference ranges were 23 to 300 U/l for serum lipase and 30 to 110 U/l for serum amylase. The 24 patients subsequently underwent TDS/TAS and the Nardi test was repeated 12 months postoperatively. The gastrointestinal quality of life index (GIQLI)6 was measured preoperatively and 12 months postoperatively. The two studies were approved by the ethics committees of the University of Nottingham Medical School and Nottingham University Hospitals. Informed written consent was obtained.
The median (range) age of the 14 male and six female volunteers was 21 (18 to 23) years. Serum lipase and amylase concentrations increased more than fourfold over baseline in 15 (75%) and 13 (65%) healthy subjects, respectively, after morphine–prostigmine provocation. The median (interquartile range) baseline and maximum post‐provocation concentrations of serum lipase were 120 (74 to 138) and 2930 (344 to 7824) U/l, respectively; corresponding values for serum amylase were 61 (44 to 81) and 220 (91 to 545) U/l (p<0.001, Wilcoxon signed ranks test). Figure 11 shows the maximum increase in enzymes over baseline for each subject, none of whom experienced pain during the study.
The median (range) age of the three male and 21 female patients was 42 (25 to 64) years. The post‐provocation enzymatic increases were significantly greater preoperatively than postoperatively (fig 22).). Preoperatively, 21 of the 24 patients (87.5%) had a greater than fourfold increase in amylase or lipase or both after morphine–prostigmine provocation, compared with 12 (50%) postoperatively (p=0.01, χ2 test with Yates' correction). While pain was reproduced after pharmacological provocation in 20 patients (83%) preoperatively, this occurred in only four (17%) postoperatively. The mean (95% confidence interval) GIQLI score improved from 61 (55 to 66) preoperatively to 92 (82 to 102) postoperatively (p<0.0001). Twelve months postoperatively, eight patients (33%) were symptom‐free and 12 (50%) had reduced analgesia requirement. Four patients (17%) with no improvement also had a positive postoperative Nardi test.
In healthy volunteers increases in serum lipase and amylase after the Nardi test lack specificity. In patients with sphincter of Oddi dysfunction, pain is a universal accompaniment of enzyme elevation after the Nardi test. TDS/TAS significantly obtunds the enzymatic and nociceptive responses to the Nardi test in only 50% of patients with sphincter of Oddi dysfunction. The lack of specificity of the Nardi test makes it unsuitable as an objective diagnostic test for this condition.
Conflict of interest: None declared.