Reports of fibrosing colonopathy have heightened concern about pancreatic enzyme dosing in patients with cystic fibrosis (CF) (1). Because the risk for colon injury from pancreatic enzymes is dose related, preventing excessive enzyme intake has become important in patient management (2). Conversely, because malnutrition negatively affects prognosis, maximizing nutrient absorption and the patient's nutritional status are important therapeutic goals. Pancreatic enzyme dosing usually is based on the patient's body weight, age, and usual fat intake (3). Lipase, which is the component of the pancreatic enzyme preparation responsible for improving fat absorption, is inactivated irreversibly in an acidic environment. To decrease enzyme inactivation by gastric acid in the small intestine, drugs that modify intestinal pH have been prescribed as adjuvant therapy, especially for patients receiving high enzyme dosages. The objective of this study was to measure the effect of gastric acid suppressant therapy with either ranitidine or omeprazole on fat absorption in patients with CF receiving a pH-sensitive, enteric-coated microtablet. Although this issue has been examined previously, small numbers of patients and the inability to control for potential differences in dietary intake throughout various treatment modalities have limited these studies. To evaluate these issues, we collected stool from inpatients who received rigorously controlled diets and varying adjuvant therapies.
METHODS
Ten adults (aged 18-36 years) and 12 children (aged 6-17 years) (15 males) participated in a double-blind, placebo-controlled crossover study. All fat balance studies were performed in the General Clinical Research Center (CRC). All subjects had CF, confirmed by sweat test, and pancreatic insufficiency. Subjects with clinically mild malabsorption underwent positive fat stain or 3-day fat-balance study to confirm fat malabsorption before entry into the study. No patient was receiving agents that modify intestinal pH, other than the study medications. Pregnant patients or subjects with cholestasis (direct bilirubin concentration >1.5 mg/dl) or hepatosplenomegaly were excluded. Patients received a controlled diet based on an analysis conducted during 3 days of eating their usual home diets. The diet for each fat-balance study period was kept constant for fat content and the number of meals and snacks per day. All subjects were changed to a dosage of Pancrease MT10 or MT16 equaling their usual home dosage for labeled lipase. All subjects previously had taken enteric-coated microtablet or microsphere product.
Subjects were initially studied at baseline while receiving their usual home enzyme dosage. Nine subjects with fat absorption greater than 88% at baseline had enzyme dosages decreased by 25% before testing with adjuvant therapy. All subjects received study enzymes from the same lot. Adjuvant therapy was started 3 days before each admission. Children weighing less than 40 kg were given ranitidine 5 mg/kg or 10 mg/kg daily, divided into two equal doses 30 minutes before breakfast and dinner; children weighing more than 40 kg and adults received 150 mg or 300 mg twice daily. Adults were also tested while receiving omeprazole 20 mg daily, 30 minutes before breakfast. Each group was also tested while receiving placebo. The order of treatment was randomly assigned. Nurses witnessed the ingestion of all study medications.
Carmine red, 1,000 mg, was administered at the time the controlled diet was started, and a second dose of carmine red was administered 72 hours later. Stool collection was started after the first red stool was passed and continued until the second red marker was passed. A stool log was kept, and each stool was bagged and labeled with the date, time, and subject name to allow verification of a complete collection. Stools were kept frozen until analysis. The CRC metabolic kitchen prepared all food and drink, and any uneaten portion was tabulated. The CRC dietary staff used Diet Planner software (version 2.05, 1993; University of California, San Francisco, CA, U.S.A.) to calculate calorie and fat intake. Quantitative fat analysis was performed according to the method of Van de Kamer et al. (4). Fat absorption was calculated as 72-hour dietary fat intake (g) - total fecal fat excretion (g), divided by the 72-hour dietary fat intake (g) × 100. The CRC review board and the University of Florida Institutional Review Board approved the study. Informed consent and assent was obtained from all subjects and parents of children.
RESULTS
One adult dropped out of the study after completing two of the four treatment arms. For 12 of 96 (12.5%) admissions, recognizable errors were made in the protocol (stool missed, extra stool saved, or extra meal given). Analyses were performed including and excluding these admissions. Table 1 lists subject enzyme dosing and fat absorption results. Paired t tests comparing each adjuvant treatment with placebo were performed to examine the mean difference in fat absorption. Figure 1 displays each subject's change in fat absorption (treatment minus placebo). Marked intrasubject variability in fat absorption was noted between the baseline admission and placebo admission in subjects on the same enzyme dosage and dietary fat intake. The mean difference in fat absorption between admissions was 2.63% with a standard deviation (SD) of 8.29%. The greatest negative change was 19.95%, and the greatest positive change was 14.37%.
Despite the above-noted variability, each treatment was associated with minimal improvement in mean fat absorption, 2.90% for low-dose ranitidine, 1.63% for high-dose ranitidine, and 1.50% for omeprazole. None of these changes were statistically different from absorption measured for patients receiving placebo, P = 0.11, P = 0.44, and P = 0.51, respectively. Subgroup analysis of the adults showed a 4.97% mean improvement in fat absorption for the group receiving low-dose ranitidine, P = 0.003. Subgroup analysis for the other adjuvant therapies in adults and children showed no difference in mean fat absorption. Additional paired t test analysis, excluding the admissions with known errors, also showed no significant change in fat absorption with low-dose ranitidine, high-dose ranitidine, and omeprazole, 3.25% (P = 0.07), 1.41% (P = 0.56), and 0.16% (P = 0.95), respectively.
To account for the crossover design used in this study (drug, period, and age category effect), a linear model F test analysis comparing ranitidine (5mg/kg or 10 mg/kg daily) with placebo was performed. No overall drug effect in the three-drug model (adults and children) was observed (P = 0.32). A second linear model F test analysis in adult subjects, comparing all four drug treatments (placebo, low-and high-dose ranitidine, and omeprazole), also showed no difference in mean fat absorption (P = 0.15). A retrospective power analysis showed the approximate power to be 70% for both the three-and four-drug model F tests. No order effect or crossover effect was noted.
As noted in the methods section, nine subjects with fat absorption greater than 88% on baseline collection had enzyme dosages decreased by 25%. Mean absorption while receiving the usual enzyme dosage was 91.4%, whereas mean absorption while receiving the lower enzyme dosage and placebo for adjuvant therapy was 89.4%; mean difference, -2.77%; SD, 8.11%;P = 0.34. Two subjects had baseline absorptions greater than 88% but did not have enzyme dosage decreased because they were receiving only one MT10 capsule per meal or snack. In the adjuvant placebo portion of the study, 11 subjects (50%) had absorptions less than 88%. Additional analysis of these subjects also failed to show a difference in mean fat absorption, P = 0.37 in the three-drug model and P = 0.33 in the four-drug model.
DISCUSSION
Complete correction of fat malabsorption with the standard therapy of enteric-coated pancreatic enzymes in patients with pancreatic insufficiency secondary to CF is not possible in all subjects. Many factors may contribute to incomplete correction, including suboptimal adherence with therapy, timing of enzyme administration, gastric emptying, bacterial overgrowth, bile acid abnormalities, small bowel disease, and acid destruction of enzymes. Lipase is irreversibly inactivated at a pH less than 4.5. Pancrease MT, the product used in this study, has an enteric coating that dissolves at pH 5.5 to protect the enzymes from inactivation in the acid environment of the stomach (5). Subjects with CF have excessive gastric acid production and diminished bicarbonate output from the pancreas (5). Both of these factors contribute to a decreased intraluminal pH in the proximal intestine. The postulated benefit of acid-modifying drugs is to increase the pH and therefore allow release of enzymes in the proximal intestine and prevent enzyme destruction after release. However, the pH effect (>4) in the duodenum must last at least 90 minutes to allow time for enzyme action and to prevent destruction of released enzyme (6).
Previous studies of adjuvant therapy with enteric-coated enzymes in subjects with CF have yielded mixed results. Three previous studies showed improvement in fat absorption in CF subjects receiving enteric-coated enzyme and misoprostol, a synthetic prostaglandin analog that inhibits gastric acid secretion and stimulates duodenal bicarbonate secretion (7-9). However, this improvement was found only in subjects who had persistent fat malabsorption while receiving enteric-coated enzymes. Some subjects with fat absorption greater than or equal to 90% did not show improvement or showed worsened fat absorption with misoprostol therapy. Thus, subject responses were quite variable to misoprostol therapy, and the improvement in fat absorption in these studies may have been related to the action of misoprostol on bicarbonate secretion in conjunction with suppression of gastric acid output.
Gow et al. (10) showed enteric-coated enzymes to be superior to conventional enzymes, but the addition of antacid or cimetidine therapy failed to show a further overall improvement in mean fat absorption. However, three subjects with persistent fat malabsorption while receiving the enteric-coated enzymes alone did correct to normal fat absorption with combination therapy. Heijerman et al. (11) showed that omeprazole therapy was effective when used with a high dosage of enteric-coated pancreatin supplement in patients with fat absorption of less than 90% while receiving enteric-coated enzymes alone. In this study, absorption with low-dose pancrease plus omeprazole therapy did not differ from that of high-dose pancrease therapy alone. In our study, we down-titrated our enzyme dosage, making our patients similar to those in Heijerman's low-dose enzymes/omeprazole group. In a subsequent study, Heijerman (12) evaluated the effect of omeprazole therapy on several tests of pancreatic function, including urinary PABA secretion, secretion of pancreatic polypeptide, and fecal fat excretion. In this study, actually larger than the first, fecal fat results were more variable and absorption did not significantly improve with omeprazole therapy. This demonstrates that fecal fat absorption can be highly variable despite similar experimental protocols, a finding that our study duplicated. Another study showed significant improvement in fat absorption and weight gain with famotidine plus enzyme therapy when compared with placebo plus enzymes (13). However, in this study, individual fecal fat absorption values were not listed and examination of the figure depicting the data shows that 8 of 10 subjects had little improvement in fecal fat absorption (2-6%), whereas 2 subjects had dramatic improvement in fecal fat absorption with famotidine therapy (37% and 38%). The different and often contradictory results noted in the literature may relate to a number of factors, including differing actions of medications (misoprostol vs. omeprazole and famotidine), level of enzyme dosing (high dosage vs. low dosage), lack of response to adjuvant therapy (cimetidine), and exaggerated response in some patients (famotidine study) (7-13). These studies were all performed in the outpatient setting. Therefore, dietary intakes could not be controlled rigorously. Assumed compliance with experimental medications and assumed completeness and proper handling of stool collections must be interpreted with caution. The largest of these studies included, as did ours, 22 patients, whereas the average patient number was 13.
In contrast to the aforementioned studies, the current investigation was performed under rigorously controlled conditions in the CRC. Dietary intakes were standardized and constant between therapeutic arms, and medication intake and proper handling of stool specimens was assured. Despite this, our study did not show significant improvement in fat absorption when adjuvant therapy was used. Although the 4.97% difference in mean fat absorption noted in adults given low-dose ranitidine therapy as compared with placebo was statistically significant, it is difficult to reconcile with the lack of response noted when patients were given high-dose ranitidine or omeprazole therapy. We do not have a good explanation for this finding. Patient numbers and the unexpected large variability in fat absorption limited the power of this study. A power analysis performed before the study assumed less variability in fecal fat and suggested adequate power to detect a 10% difference. We found marked variability in fat absorption when the groups were considered together with subjects receiving the same enzyme dosage and identical 3-day dietary fat intakes under the tightly defined protocol used in the CRC (baseline vs. placebo: mean difference in fat absorption was 2.63%, greatest negative change was 19.95%, greatest positive change was 14.37%;P = 0.49; SD, 8.29). However, when the groups are subdivided according to those with positive and those with negative change from baseline, the variability is less marked. The standard deviation indicates that fat absorption measured by a 3-day fat-balance study in a CRC can vary, plus or minus, almost 10% in the same subject on the same enzyme dosage and diet. This variability limited our study's ability to detect 5% to 10% changes in fat absorption. More importantly, these data suggest that even under optimally controlled conditions, quantitative fecal fat determinations vary widely from day to day. This finding casts doubt on the utility of these studies in the day-to-day management of patients with CF, and suggests that interpretation of therapeutic studies using this parameter as the therapeutic endpoint may be problematic.
A potential criticism of this study is the fact that during the baseline portion of the study, 8 of 22 patients had fat absorption greater than or equal to 90%, despite a prior decrease in enzyme dosage. With additional testing, 7 of 22 patients had fat absorption greater than or equal to 90% while receiving placebo. Only four patients had fat absorption greater than or equal to 90% in both tests. This further demonstrates the inherent variability in fat absorption in these subjects. Even when the group of patients in whom fat absorption never exceeded 90% was analyzed separately, we noted no therapeutic effect of adjuvant therapies.
In conclusion, adjuvant therapy with ranitidine and omeprazole did not result in significantly improved fat absorption under the rigorously controlled conditions used in this study. Finally, and perhaps most importantly, these data highlight the variability of the 3-day fat-balance study, even when performed in a controlled environment. Therefore, a more precise measure of fat absorption would be of great clinical use.
Acknowledgements:
The authors thank Michael J. Dallas, MStat, and Dan Bowling, Mstat, for help with statistical methods. Glaxo-Wellcome, Merck, and Ortho-McNeil provided the drugs administered during the study. The Clinical Research Center provided inpatient facilities, nursing care, statistical expertise, and nutritional support for the study.
REFERENCES
1. Smyth RL, van Velzen D, Smyuth AR, et al. Strictures of ascending colon in cystic fibrosis and high strength pancreatic enzymes. Lancet 1994; 343:85-6.
2. Mac Sweeney EJ, Oades PJ, Buchdahl R, et al. Relation of thickening of colon wall to pancreatic-enzyme treatment in cystic fibrosis. Lancet 1995; 345:752-6.
3. Borowitz DS, Grand RJ, Durie PR. Use of pancreatic enzyme supplements for patients with cystic fibrosis in the context of fibrosing colonopathy. J Pediatr 1995; 127:681-4.
4. Van de Kamer JH, Ten Bokkel Huinink H, Weyers HA. Rapid method for the determination of fat in feces. J Biol Chem 1949; 177:347-55.
5. Youngberg CA, Berardi RR, Howatt WF, et al. Comparison of gastrointestinal pH in cystic fibrosis and health subjects. Dig Dis Sci 1987; 32:472-80.
6. Graham DY. Enzyme replacement therapy of exocrine pancreatic insufficiency in man. N Engl J Med 1977; 296:1314-7.
7. Robinson P, Sly PD. Placebo-controlled trial of misoprostol in cystic fibrosis. J Pediatr Gastroenterol Nutr 1990; 11:37-40.
8. Robinson PH, Sly PD, Smith AL. Effect of misoprostol on fat malabsorption in cystic fibrosis. Arch Dis Child 1988; 63:1082-1091.
9. Cleghorn GJ, Shepherd RW, Holt TL. The use of a synthetic prostaglandin E1 analogue (misoprostol) as an adjunct to pancreatic enzyme replacement in cystic fibrosis. Scand J Gastroenterol 1988; 23(suppl 143):142-7.
10. Gow R, Francis P, Bradbear R, et al. Comparative study of varying regimens to improve steatorrhea and creatorrhea in cystic fibrosis: effectiveness of an enteric-coated preparation with and without antacids and cimetidine. Lancet 1981; 2:1071-4.
11. Heijerman HG, Lamers CB, Bakker W. Omeprazole enhances the efficacy of pancreatin (pancrease) in cystic fibrosis. Ann Intern Med 1991; 114:200-1.
12. Heijerman HG, Lamers CB, Bakker W, et al. Improvement of fecal fat excretion after addition of omeprazole to pancrease in cystic fibrosis is related to residual exocrine function of the pancreas. Dig Dis Sci 1993; 38:1-6.
13. Carroccio A, Pardo F, Montalto G, et al. Use of famotidine in severe exocrine pancreatic insufficiency with persistent maldigestion on enzymatic replacement therapy. A long-term study in cystic fibrosis. Dig Dis Sci 1992; 37:1441-6.
© 2002 Lippincott Williams & Wilkins, Inc.