Bacterial lipase and high-fat diets in canine exocrine pancreatic insufficiency: A new therapy of steatorrhea?

GASTROENTEROLOGY 1997;112:2048–2055

Bacterial Lipase and High-Fat Diets in Canine Exocrine Pancreatic Insufficiency: A New Therapy of Steatorrhea? AYAKO SUZUKI, AKIYOSHI MIZUMOTO, MICHAEL G. SARR, and EUGENE P. DIMAGNO Gastroenterology Research Unit, Mayo Clinic and Mayo Foundation, Rochester, Minnesota

Background & Aims: Nutrients and properties of lipases affect survival of lipolytic activity during aboral gastrointestinal transit. Whether different doses and formulations of bacterial lipase and diets affect steatorrhea was tested in pancreatic-insufficient dogs. Methods: A dose of 0–600,000 IU of powdered and 135,000 and 300,000 IU of liquid bacterial lipase was given with a standard meal to 5 dogs with ligated pancreatic ducts. In 4 dogs, 0 or 300,000 IU (normal 6-hour postprandial amount) of powder bacterial lipase was also given with five meals containing 850 kcal with different nutrient caloric densities (mixture design). Coefficients of fat absorption during 72-hour fecal balance studies were used to assess treatments. Results: With the standard meal, powder bacterial lipase reduced steatorrhea in a dose-dependent manner (P Å 0.03), and 135,000 and 300,000 IU of the liquid form decreased steatorrhea more than powder bacterial lipase (P Å 0.017 and 0.057, respectively). Coefficients of fat absorption with 300,000 IU of powder bacterial lipase correlated (r 2 Å 0.79; P õ 0.001) with increasing proportions of fat calories in diets. Conclusions: Liquid bacterial lipase decreases steatorrhea more than powder, and 300,000 IU of powder bacterial lipase ingested with high-fat meals corrects canine pancreatic steatorrhea. The combination of adequate mixing of small amounts (milligrams) of bacterial lipase and high-fat meals abolishes canine steatorrhea and may abolish human pancreatic steatorrhea.

S

teatorrhea caused by chronic pancreatitis rarely is abolished by commercial preparations of porcine lipase because the ingested lipase is inactivated within the lumen by acid1,2 and proteases.3,4 To date, efforts to correct pancreatic steatorrhea have included increasing survival of lipolytic activity by neutralizing acid with antacids,2,5 reducing gastric acid secretion with H2 blockers2,5–8 or omeprazole9 or using enteric-coated microsphere formulations2,10 to protect enzymes from denaturation or lipases that are resistant to acid denaturation.11 Although some strategies correct steatorrhea in a minority of patients, they have not gained popularity because they require multiple medications, which reduces compliance and increases cost. Enteric-coated microsphere formula/ 5e1d$$0018

05-09-97 20:36:49

gasa

tions do not abolish steatorrhea in most patients because pellet size and optimum pH of the coatings for enzyme release preclude adequate mixing of the lipolytic activity with the meal in the upper small intestine. Furthermore, daily ingestion of a large number of microencapsulated preparations may cause colonic strictures in children with cystic fibrosis.12,13 Acid-resistant lipases tested to date, including gastric and fungal lipases, are inactivated by physiological concentrations of bile acids in the lumen of the small intestine.11 Hence, the ideal lipase should resist acid denaturation and proteolytic digestion and maintain lipolytic activity within the milieu of the intestinal tract. We showed that the lipolytic activity of bacterial lipase isolated from Burkholderia plantarii, previously called Pseudomonas glumae,14 survives better in vitro than porcine lipase in human gastric or duodenal juice under luminal conditions present postprandially in patients with exocrine pancreatic insufficiency,15 maintains activity in the presence of physiological concentrations of bile acids, and does not require colipase for activity. Thus, we hypothesize that bacterial lipase may be ideal to alleviate pancreatic steatorrhea. The nutrient composition of meals also may be important in treating pancreatic steatorrhea. Nutrients increase the survival of pancreatic enzymes during storage,16 during incubation in human duodenal juice,17 and during duodenal-ileal transit in normal humans.18 For example, incubating lipase at normal postprandial concentrations (250 IU/mL) or in concentrations found in pancreatic insufficiency (25 IU/mL) with protein and fat, but not carbohydrate, prevents loss of lipolytic activity. These studies suggest that altering the diet might change the response to lipase in pancreatic insufficiency. To test our hypotheses, we developed a canine model, because canine and human pancreatic and gastrointestinal physiology are similar.19 Dogs with chronic pancreatic Abbreviations used in this paper: CCA, coefficient of carbohydrate absorption; CFA, coefficient of fat absorption; CPA, coefficient of protein absorption. q 1997 by the American Gastroenterological Association 0016-5085/97/$3.00

WBS-Gastro

June 1997

LIPASE, DIETS, AND STEATORRHEA 2049

insufficiency from ligation of both pancreatic ducts or after total pancreatectomy survive if carefully managed postoperatively.20 Our strategy and specific aims were to determine the dose response to powder bacterial lipase from B. plantarii while dogs ate standard dog chow and then to use the highest dose that did not abolish steatorrhea to test the effects of diets and another formulation of bacterial lipase (bacterial lipase dissolved in saline and mixed with meals), which might improve the mixing of the lipolytic activity with the ingested nutrients. We also determined if measurements of fecal frequency, consistency, and total or solid stool weights predicted changes in absorption. To mimic amounts of lipolytic activity generally used to treat human pancreatic steatorrhea, we gave a low dose of 30,000 IU of lipolytic activity, which has a mass of 12 mg. This amount of lipolytic activity is contained in eight pancreatin tablets (amounts we have previously tested and used clinically1,2). We also tested 135,000, 300,000, and 600,000 IU of bacterial lipase. We emphasize that giving these higher doses of lipases as pancreatin preparations to humans is impossible because the low specific activity of porcine lipase would require ingesting an exorbitant number (ú35) of tablets. In contrast, administering 300,000–600,000 IU of bacterial lipase requires ingesting only 120–240 mg.

Materials and Methods Preparation of Pancreatic-Insufficient Dogs and Their Maintenance All procedures and experiments were reviewed and approved by the Institutional Animal Care and Use Committee of the Mayo Foundation in accordance with the guidelines of the National Institutes of Health and the Public Health Policy on the Humane Use and Care of Laboratory Animals. We used 5 female dogs weighing between 18 and 21 kg. Dogs were anesthetized with intravenous thiopental sodium (12.5 mg/kg) and maintained with halothane. After a midline celiotomy, all tissue connections between the head of the pancreas and the duodenum were ligated and transected, including both the minor and major pancreatic ducts. The inferior and superior pancreatoduodenal vessels were carefully preserved, as were the mesenteric vascular arcades along the duodenum proximal and distal to the region of the duodenum adjacent to the head of the pancreas. Transection and ligation of all pancreaticoduodenal connections is important. In several dogs, individual ligation and transection of just the major and minor ducts failed to completely abolish intraluminal protease and lipase activity. These dogs did not have malabsorption and were not used for this study. After surgery, the dogs were housed in individual cages. The absence of intraduodenal protease and lipase activity was confirmed in each dog before the study. For maintenance be-

/ 5e1d$$0018

05-09-97 20:36:49

gasa

tween experiments, dogs were fed canned dog food (Hill’s Prescription Diet, Canine i/d; Hill’s Pet Products, Topeka, KS). Each can contained 580 kcal: 48% carbohydrate, 27% fat, and 25% protein as percentage of calories. Each can of dog food contained 15 g of fat as triglyceride, diglyceride, monoglyceride, and fatty acid; 1 g of cholesterol and 1 g of cholesterol ester; and 0.5 g of phospholipid. Dogs were fed two cans in the morning and one can in the afternoon. Ten grams of porcine pancreatin powder (Viokase; AH Robins Co., Richmond, VA) was given with the morning meal, and 7 g was given with the afternoon meal.

Experimental Design At the beginning of each experiment, a carmine stool marker (Chemical Mfg. Corp., Gardena, CA) was given with a meal. When the carmine appeared in the stools, stools were collected for 72 hours. Pancreatic insufficiency experiments began 3 weeks after the operation. For the dose response study, 5 dogs were fed one can of dog food with 100 mL of 0.9% NaCl in the morning and afternoon. Each meal was given without or with powder bacterial lipase (30,000, 135,000, 300,000, or 600,000 IU of lipolytic activity [12, 54, 120, or 240 mg]) sprinkled on the surface of the meal. To determine if different formulations of bacterial lipase affected the response in these 5 dogs, we used 135,000 and 300,000 IU because we found that 600,000 IU of bacterial lipase corrected steatorrhea in 1 and nearly corrected steatorrhea in another dog (96% and 89% coefficients of fat absorption [CFA], respectively; see Results). To formulate the liquid lipase, we dissolved the bacterial lipase (water soluble) in the 100 mL of 0.9% NaCl and mixed it with the dog food immediately before the dog ate the meal. Dogs gulped and swallowed the meal in õ2 minutes. To determine the effect of nutrients on steatorrhea, each of 4 dogs (1 of the 5 dogs used in the dose-response and formulation studies was killed during the study because of refusal to eat the maintenance diet with the large amount of pancreatin) was fed each of the five meals (Figure 1) twice daily for the duration of each balance study. Two balance studies were performed for each of the test meals. With one balance study, no enzymes were given. With the other balance study, 300,000 IU of powder bacterial lipase was given with both daily meals. A maximum of one balance study was performed on each dog per week. Between balance studies (a minimum of 4 days, but usually ú7 days), dogs were fed the maintenance diet with pancreatin, as previously described. Administration of the test meals and bacterial lipase was in random order. The five meals consisted of 850 kcal that had the proportions of nutrients distributed according to a mixture design.21 The nutrient composition of the meals was adjusted by adding to one can of dog food varied quantities of Promod (protein supplement made from whey; Ross Laboratories, Columbus, OH), Polycose (glucose polymers of approximately 1000 daltons in powder form; Ross Laboratories), and corn oil. Meals (numbered in order of percentage of calories as fat in the meals) were low fat, low

WBS-Gastro

2050 SUZUKI ET AL.

GASTROENTEROLOGY Vol. 112, No. 6

Figure 1. (A ) Triangular diagram and (B ) table of the distribution of percentage of calories of fat, protein, and carbohydrate in diets. The horizontal axis is protein from 0% (left) to 100% (right), the right axis is fat from 0% (bottom) to 100% (top), and the left axis is carbohydrate from 0% (top) to 100% (bottom).

protein, and high carbohydrate (meal 1); low fat, high protein, and high carbohydrate (meal 2); relatively even distribution of three nutritional classes (meal 3); high fat, high protein, and low carbohydrate (meal 4); and high fat, low protein, and high carbohydrate (meal 5). The percentage of calories of fat, protein, and carbohydrate in the meals was selected on the basis of an ethical constraint area (daily requirements of each of the nutrients) and by determining what the dogs would eat (palatability) (Figure 1). The mineral content of the five meals was adjusted by adding calcium citrate, sodium citrate, potassium citrate, sodium phosphate, potassium chloride, and magnesium oxide so that the mineral requirement per day was exceeded but nearly equal among the meals (Table 1). The ingredients were similar among meals and consisted of chicken, turkey, beef lung, whole egg, ground corn, liver, animal fat (chicken, pork), rice, cracked

pearled barley, and soybean meal. In particular, the types of fat were similar among the test meals (Kathy L. Gross, Ph.D., personal communication, September 1994) (Table 1). In addition, the five meals had similar energy content because digestibility (proportion of nutrients available for absorption by the dog) of protein, fat, and carbohydrate in canned foods was similar among test meals, i.e., 84.5%–93.5%, 95.5%–97.2%, and 88.6%–91.2%, respectively (Marilyn Colgon, D.V.M., personal communication, August 1994). Each of the meals had a low fiber content (0.9–2.7 g/d). Lastly, because the physical properties among test meals could affect fat absorption, we standardized the total volume and pH among the meals and the osmolality in four of the five meals (Table 1). The bacterial lipase used in these studies (Knoll AG, Ludwigshafen, Germany) has a molecular weight of 30,000, is acid-stable, contains 319 amino acids, and is secreted by B. plantarii during fermentation. In contrast to porcine lipase, this lipase does not require colipase for activity. There are at least two forms of the enzyme containing two or four molecules per asymmetric unit.14 The specific activity of the preparation is 3,000 IU lipolytic activity per milligram.

Analyses The 72-hour stool collections were weighed, and the results are expressed as grams per 24 hours. Fecal fat,22 nitrogen,23 and carbohydrate24 were measured and the results expressed as CFA, coefficient of protein (CPA; nitrogen converted to protein by assuming that protein contained 16% nitrogen) or coefficient of carbohydrate absorption (CCA). Fecal consistency was scored as grade 1 (normal-looking and formed), grade 2 (mushy), or grade 3 (liquid). Fecal frequency, the fecal score (product of consistency and frequency), and total and solid stool weights are reported per 24 hours.

Data and Statistical Analysis Results are expressed as mean values { SE of the coefficients of absorption. To avoid unequal variances of a propor-

Table 1. Meal Composition

Volume (mL ) pH Osmolality (mOsm ) Ca2/ Na/ K/ Cl0 P Mg Fat composition Short chain Median chain Long chain Saturated Unsaturated Saturated/unsaturated

Meal 1

Meal 2

Meal 3

Meal 4

Meal 5

500 5.94 1005 2089 1314 1980 2154 1700 180

510 6.09 757 2090 1250 1980 1869 1700 180

540 5.91 769 2090 1157 1980 1870 1700 180

500 5.99 795 2088 1170 2108 1869 1699 180

520 5.84 753 2090 1136 1980 1911 1700 180

22 0.8 77.2 30.8 69.2 0.45

17.2 0.6 82.2 27.2 72.8 0.37

14.5 1.0 84.5 30.6 69.4 0.44

11.6 0.8 87.6 27.7 72.3 0.38

14.1 1.3 84.6 33.4 66.6 0.5

NOTE. Total volume, pH, and osmolality are values obtained after blending a meal with 100 mL saline. The mineral content (Ca2/, Na/, K/, Cl0, P, and Mg) is expressed as milligrams per day (two meals). Fat composition is the percent distribution of fatty acids in the meal.

/ 5e1d$$0018

05-09-97 20:36:49

gasa

WBS-Gastro

June 1997

LIPASE, DIETS, AND STEATORRHEA 2051

tion or a percentage at their extremes,25 we transformed the proportions to arcsines25 or used fecal fat (grams per 24 hours) to perform statistical analyses. Data were assessed by analysis of variance for repeated measures. In certain analyses, repeated t tests were performed to determine if there were significant differences among treatments; the Bonferroni correction was used whenever appropriate. To analyze differences among meals, we performed analysis of variance for repeated measures with the Tukey correction. Multiple regression analyses were performed to detect relationships among fecal data and nutrients.

Results Figure 3. Fecal fat in response to no treatment (h) or 300,000 IU powder bacterial lipase (j) in 4 dogs with different diets (horizontal axis, ascending order of percentage of fat calories in meals). Values are mean { SE. *P õ 0.0005 vs. meal 1, 2, and 3.

Effects of Exocrine Pancreatic Insufficiency on Body Weight, Appetite, and Blood Glucose Dogs with pancreatic insufficiency lost 7% { 2% (2%–13%) of their body weight but had good appetites and appeared otherwise healthy; postprandial blood glucose concentrations of õ100 mg/dL were not different from normal throughout the experiments (data not shown).

Effect of Liquid Lipase

Dose Response of Powder Bacterial Lipase In dogs with pancreatic insufficiency without treatment, the CFA was less than preoperative values (63% { 6% vs. 96% { 1%; P õ 0.01). Bacterial lipase increased the CFA in a dose-dependent manner (P Å 0.03). Giving 600,000 IU of bacterial lipase decreased steatorrhea ú30,000 IU (P õ 0.05; Figure 2A). With 600,000 IU, 1 dog corrected (96% CFA) and 1 nearly corrected (89% CFA) steatorrhea. By comparison, the

Figure 2. Effects of powder and liquid bacterial lipase in 5 pancreaticinsufficient dogs with a diet of 47% carbohydrate, 30% fat, and 23% protein. KU Å 103 U. (A ) Dose response (mean { SE) to powder bacterial lipase (j) and response to liquid lipase (h). *P õ 0.05. (B ) Response to liquid and powder lipase at 135,000 (L and solid lines) and 300,000 (l and dotted lines) IU. Note the different vertical scales for A and B: (A ) 0%–100%; (B ) 40%–100%. **P Å 0.017 (powder vs. liquid 135,000 IU bacterial lipase) and P Å 0.057 (powder vs. liquid 300,000 IU bacterial lipase).

/ 5e1d$$0018

05-09-97 20:36:49

highest CFA with 300,000 IU was 77%. In pancreaticinsufficient dogs, the CPA was significantly lower than in normal dogs (39% { 3% vs. 75% { 5%; P Å 0.001) and was not altered by bacterial lipase.

gasa

Giving 135,000 and 300,000 IU of liquid lipase increased the CFA more than the same doses of powder lipase (P Å 0.017 and 0.057, respectively; Figure 2B); however, these doses of liquid lipase did not increase CFA ú600,000 IU of powder lipase (Figure 2B). Effect of 300,000 IU of Powder Bacterial Lipase With Different Diets Fecal fat and fat absorption. Without bacterial lipase, fecal fat (grams per 24 hours; Figure 3) and CFA (Figure 4) differed among meals (P õ 0.0001 and P õ

Figure 4. Effects of bacterial lipase on coefficient of fat absorption with different diets in each of 4 dogs. (A ) Response to no treatment (h) or 300,000 IU powder bacterial lipase (j) with different diets (horizontal axis, ascending order of percentage of fat calories in meals). Values are mean { SE. *P õ 0.01. (B ) Each combination of symbol and line is 1 of 4 dogs. r Å 0.835–0.977 for the 4 dogs (P Å 0.002 for slopes vs. 0).

WBS-Gastro

2052 SUZUKI ET AL.

GASTROENTEROLOGY Vol. 112, No. 6

Figure 5. CPA in response to no treatment (h) and 300,000 IU powder bacterial lipase (j) with different diets (horizontal axis arranged in ascending order of percentage of protein calories in meals). Values are mean { SE. *P õ 0.05.

0.006, respectively). Fecal fat was significantly greater with the two high-fat meals (meals 4 and 5; P õ 0.0005). The CFA of meal 3 (relatively even distribution of three nutritional classes) was greater (P õ 0.01) than that of meals 1 (low fat, low protein, high carbohydrate) and 4 (high fat, high protein, low carbohydrate), and the CFA of meal 2 was greater than that of meal 1 (P Å 0.04). In contrast, with powder bacterial lipase, fecal fat was approximately 10 g/24 h among all meals (Figure 3). Bacterial lipase reduced fecal fat with meals 1, 4, and 5 (P õ 0.01), but not with meals 2 or 3 (Figure 3). The amount of absorbed fat correlated with the amount of fat in the meals (r2 Å 0.99; P õ 0.001; data not shown). CFA in response to powder bacterial lipase (Figure 4A) differed among meals (P õ 0.0002) and correlated with increasing percentage of fat calories in the meals (r2 Å 0.79; P õ 0.001). Bacterial lipase increased the CFA of meals 1, 4, and 5 (P õ 0.01), but not of meals 2 or 3 (Figure 4). Within each dog (Figure 4B), CFA directly correlated (r Å 0.835–0.977; slopes vs. 0, P Å 0.002) with the percentage of fat calories in the diet. CPA. CPA differed among meals regardless of treatment without or with bacterial lipase (P õ 0.03 and 0.01, respectively), but bacterial lipase did not affect CPA within each meal. Without bacterial lipase, CPA was lowest with meal 5 (high fat, low protein, high carbohydrate; first column in Figure 5), which was less (P õ 0.05) than meals 2, 3, and 4 (pairs of columns 3, 4, and 5 in Figure 5). CPA directly correlated with percentage of protein calories in meals without or with bacterial lipase (r2 Å 0.46, P õ 0.05 and r2 Å 0.78, P õ 0.001, respectively). The amount of absorbed protein also correlated with the amount of protein in meals with/ 5e1d$$0018

05-09-97 20:36:49

gasa

out or with bacterial lipase (r2 Å 0.72, P õ 0.001 and r2 Å 0.95, P õ 0.001, respectively). CCA. CCA differed among meals without and with bacterial lipase (P õ 0.04; P õ 0.001), and within a meal, CCA was similar regardless of treatment except for meal 4 (high fat, high protein, low carbohydrate). With meal 4, bacterial lipase increased CCA (P Å 0.05; first pair of columns in Figure 6), which was greater than meals 2 and 1 (P õ 0.02; second and fifth columns in Figure 6). Without and with bacterial lipase, there was a trend toward an inverse correlation between CCA and the percentage of carbohydrate calories in meals (r2 Å 0.428, P Å 0.064 and r2 Å 0.419, P Å 0.07, respectively), and a significant direct correlation between the amount of absorbed carbohydrate and the amount of carbohydrate in the meal (r2 Å 0.98, P õ 0.001 and r2 Å 0.95, P õ 0.001, respectively). Fecal frequency, consistency, and score. In pancreatic-insufficient dogs, fecal scores (Table 2) generally paralleled fecal frequencies and consistencies (data not shown) and were significantly greater than preoperative values of the dogs when they were fed the maintenance diet (1.9 { 0.4; P õ 0.02). Fecal scores were also different among meals (Table 2; P õ 0.05), directly correlated with the percentage of protein calories in the meal (r2 Å 0.649, P õ 0.001 and r2 Å 0.597, P õ 0.01, without and with bacterial lipase, respectively), and inversely correlated with the combined percentage of fat and carbohydrate calories in the meal (r2 Å 0.658, P õ 0.01 and r2 Å 0.678, P õ 0.01, without and with bacterial lipase, respectively). Without bacterial lipase, the fecal score of meal 2 (highest protein content) was greater than all other meals (P õ 0.002), and the fecal

Figure 6. CCA in response to no treatment (h) and 300,000 IU powder bacterial lipase (j) in 4 dogs with different diets (horizontal axis, ascending order of percentage of carbohydrate calories in meals). Values are mean { SE. *P õ 0.02.

WBS-Gastro

June 1997

LIPASE, DIETS, AND STEATORRHEA 2053

Table 2. Fecal Score, Total Stool Weight, and Solid Weight in Stool According to Percentage of Fat, Carbohydrate, and Protein as Caloric Value in Five Meals Fecal scorea Meal 1 2 3 4 5

F (%) 18 23 33 43 47

C (%) 66 37 43 21 43

P (%) 16 40 24 36 10

No treatment 5.9 11.7 8.4 8.8 7.4

{ { { { {

d

0.4 0.3e 0.5 0.4 0.4

Total stool weight (g )b

BL 300,000 IU

No treatment

{ { { { {

592 { 37 533 { 89 393 { 54 663 { 41g,h 534 { 49h

6.1 11.6 7.8 6.3 5.1

1.1 1.2 1.3 1.1 1.4f

Solid weight in stool (g )c

BL 300,000 IU 632 579 408 430 377

{ { { { {

85 61 52 18 57

No treatment 153 155 108 195 139

{ { { { {

14 24 17 14g 14h

BL 300,000 IU 151 163 112 131 92

{ { { { {

22 17 10 9 14i

BL; bacterial lipase. a P õ 0.05 among meals with no treatment; P õ 0.02 among meals with BL 300,000 IU. b P õ 0.05 among meals with no treatment. c P õ 0.05 among meals with no treatment and BL 300,000 IU. d P õ 0.01 vs. meals 3 and 4. e P õ 0.002 vs. all other meals. f P õ 0.03 vs. meal 2. g P õ 0.05 vs. meal 3. h P õ 0.03 vs. BL 300,000 IU. i P õ 0.05 vs. meal 2.

score of meal 1 (high carbohydrate, low fat, low protein) was lower than meals 3 and 4 (P õ 0.01). With bacterial lipase, the fecal score of meal 2 was also higher than the other meals and significantly higher than meal 5 (high carbohydrate, high fat, low protein; P õ 0.03), the meal with the lowest fecal score. Weights of the total and solid components of stool. In pancreatic-insufficient dogs in the absence of

bacterial lipase, the total and solid stool weights differed among meals (P õ 0.05 and P õ 0.04, respectively); stool weights of meal 3 (relatively even distribution of three nutritional classes) were the lowest, significantly lower than meal 4 (P õ 0.05; low carbohydrate, high fat, high protein). Bacterial lipase decreased (P õ 0.03) total stool weights of both high fat meals (meals 4 and 5) and the solid fecal weight of meal 5. With bacterial lipase, the solid fecal weight of meal 5 was the lowest, significantly lower than that of meal 2 (P õ 0.05; high carbohydrate, low fat, high protein), which had the highest fecal solid weight with treatment. Without and with bacterial lipase, total (Figure 7) and solid (data not shown) stool weights inversely correlated with CFA. With bacterial lipase, total and solid stool weights also inversely correlated with CCA (r2 Å 0.511, P õ 0.05 and r2 Å 0.475, P õ 0.04, respectively). Without treatment, stool weights did not correlate with meal nutrients; however, with bacterial lipase, total and solid stool weights correlated inversely with the percentage of fat calories (r2 Å 0.696, P õ 0.001 and r2 Å 0.649, P Å 0.002, respectively) and directly correlated with the total percentage of protein and carbohydrate calories (r2 Å 0.697, P õ 0.005 and r2 Å 0.759, P Å 0.001, respectively) in meals. / 5e1d$$0018

05-09-97 20:36:49

gasa

Discussion Treatment of human pancreatic steatorrhea is not optimal and rarely abolishes the condition. In contrast, we show that 300,000 IU (120 mg) of bacterial lipase corrects steatorrhea in canine exocrine insufficiency when ingested with meals containing a high proportion of fat as calories. This amount of lipolytic activity is secreted normally 4–6 hours postprandially (300,000 IU) in humans26 and dogs.27 Thus, human pancreatic steatorrhea might be abolished by giving milligrams of bacterial lipase containing large amounts of lipolytic activity with high-fat meals. This treatment is a drastic departure from current treatment of human exocrine insufficiency and is important to consider in planning clinical trials. In patients with severe exocrine pancreatic insufficiency, 30,000 IU

Figure 7. Relation between total stool weight and CFA with different diets in 4 dogs; (A ) no treatment, (B ) 300,000 IU bacterial lipase (BL). Each combination of symbol and line is 1 of 4 dogs. (A ) r Å 0.38–0.99 for the 4 dogs (P Å 0.036 for slopes vs. 0); (B ) r Å 0.75– 0.95 for the 4 dogs (P Å 0.028 for slopes vs. 0). Note the different horizontal scales for A (20%–80%) and B (60%–95%).

WBS-Gastro

2054 SUZUKI ET AL.

GASTROENTEROLOGY Vol. 112, No. 6

of porcine lipase reduces steatorrhea by ú50% but does not abolish steatorrhea; most patients continue to excrete ¢10–20 g of fat per 24 hours.1,2 In these studies, patients were fed high-fat diets with 100 g of fat in equally divided amounts of 25 g of fat in four meals daily. Under these conditions, abolishing steatorrhea requires gastric acid neutralization, antisecretory agents, or microencapsulation with enteric coatings; however, even these adjuvants do not abolish steatorrhea in all patients. In untreated canine pancreatic insufficiency, fecal fat markedly increased when the caloric proportion of fat in meals was ú40%; CFA was greatest with meal 3 (relatively equal proportions of nutrients as calories). In subjects with untreated pancreatic insufficiency, this meal had a similar CFA as meal 1 (lowest fat, low protein, and the highest carbohydrate) with bacterial lipase. In contrast, with bacterial lipase, fecal fat was approximately 10 g/day among all meals. As we increased the amount of fat in meals, more fat was absorbed, resulting in higher CFA. We interpret these data as indicating that to insure optimal nutrition in subjects with untreated pancreatic insufficiency, meals should have balanced caloric proportions of the nutrients; however, if exogenous lipase is administered, a high-fat diet may improve fat malabsorption. This is an iconoclastic opinion; generally, patients with pancreatic insufficiency restrict intake of fats to relieve symptoms of malabsorption. We have under investigation several possibilities of why fat absorption varies in response to bacterial lipase with meals of different nutrient composition: (1) differences in the survival of lipolytic activity induced by nutrients; (2) alterations of gastrointestinal transit (mixing and coordination between lipase and lipid delivery); (3) differences in duodenal pH and bile acid/micelle concentrations; (4) interaction of fat with other undigested nutrients in the intestinal lumen; and (5) changes in the rate of fatty acid absorption by intestinal cells. However, we conclude that the present study shows that in pancreatic insufficiency, fat absorption is directly dependent on the load of fat in diets when a normal amount of lipase (in this study as bacterial lipase) is given with meals. It is important to abolish steatorrhea. In children with cystic fibrosis, correction of malabsorption may decrease morbidity and increase survival, because correcting malabsorption may restore normal growth and slow the decline of pulmonary function.28 Adults with exocrine pancreatic insufficiency have a significantly shortened life span,29 partly because of an increase in atheromatous large vessel disease and resultant cardiovascular disease.30 Patients with chronic pancreatitis have lower plasma levels of factors that may protect against atherogenesis, including high density lipoprotein-c, apoprotein a1, and / 5e1d$$0018

05-09-97 20:36:49

gasa

lipoprotein a.31 What produces these abnormalities is unknown, but malabsorption, malnutrition, and metabolic abnormalities associated with smoking and drinking may be involved. In general, fecal characteristics (fecal score) correlated directly with the proportion of protein in diets regardless of treatment, but treatment with bacterial lipase and a high proportion of fat in the diet was associated with the lowest fecal score and lowest total and solid fecal weights. We interpret these data to suggest that meals with a high proportion of protein might increase symptoms because of malabsorption, and if fat malabsorption is corrected but protein (and/or carbohydrate) malabsorption is not, symptoms of malabsorption will continue. However, amelioration of steatorrhea by bacterial lipase should reduce symptoms, because total stool weight and solid weight were inversely correlated with CFA and CCA, but not CPA. Furthermore, stool weight is likely a better predictor of the response to treatment of malabsorption (other than measurement of nutrients in stool) than fecal score or frequency. We suggest that in human exocrine insufficiency, treatment with lipase and a high proportion of fat in the diet will reduce fecal frequency and amount of stool, but abolition of symptoms requires abolition of malabsorption of other nutrients, particularly protein. To completely correct malabsorption requires the use of proteolytic enzymes, amylase, and lipase. Further investigations are needed to conclude which diets best alleviate stool symptoms in pancreatic insufficiency. In summary, our findings may have direct clinical significance. A stable lipase that efficiently digests fat intraluminally would greatly simplify treatment of pancreatic steatorrhea by eliminating the need for using acidneutralizing, antisecretory drugs, or enteric coatings. Bacterial lipase may fulfill these requirements. In addition, our study supports ingesting high-caloric, high-fat meals in exocrine pancreatic insufficiency, which would especially benefit children with cystic fibrosis who need better nutritional management to enhance growth and development, and should reduce morbidity and mortality in patients with cystic fibrosis or chronic pancreatitis.

References 1. DiMagno EP, Malagelada J-R, Go VLW, Moertel CG. Fate of orally ingested enzymes in pancreatic insufficiency: comparison of two dosage schedules. N Engl J Med 1977;296:1318–1322. 2. Regan PT, Malagelada J-R, DiMagno EP, Glanzman SL, Go VLW. Comparative effects of antacids, cimetidine, and enteric coating on the therapeutic response to oral enzymes in severe pancreatic insufficiency. N Engl J Med 1977;297:854–858. 3. Thiruvengadam R, DiMagno EP. Inactivation of human lipase by proteases. Am J Physiol 1988;255:G476–G481. 4. Thiruvengadam R, Zinsmeister AR, DiMagno EP. Is human lipase

WBS-Gastro

June 1997

5. 6.

7.

8.

9.

10.

11.

12.

13. 14.

15.

16.

17.

18.

LIPASE, DIETS, AND STEATORRHEA 2055

survival during aboral intestinal transit dependent upon intraluminal chymotrypsin (CT) activity and bile acid concentrations (BAC)? Dig Dis Sci 1987;32:1189. Graham DY. Pancreatic enzyme replacement: the effect of antacids or cimetidine. Dig Dis Sci 1982;27:485–490. Saunders JH, Drummond S, Wormsley KG. Inhibition of gastric secretion in treatment of pancreatic insufficiency. Br Med J 1977; 1:418–419. Regan PT, Malagelada J-R, DiMagno EP, Go VLW. Rational for use of cimetidine in pancreatic insufficiency. Mayo Clin Proc 1978;53:79–83. Zentler-Munro PL, Fine DR, Batten JC, Northfield TC. Effect of cimetidine on enzyme inactivation, bile acid precipitation, and liquid solubilisation, in pancreatic steatorrhea due to cystic fibrosis. Gut 1985;26:892–901. Heijerman HG, Lamers CB, Bakker W. Omeprazole enhances the efficacy of pancreatin (pancrease) in cystic fibrosis. Ann Intern Med 1991;114:200–201. Dutta SK, Rubin J, Harvey J. Comparative evaluation of the therapeutic efficacy of a pH sensitive enteric coated pancreatic enzyme preparation with conventional pancreatic enzyme therapy in the treatment of exocrine pancreatic insufficiency. Gastroenterology 1983;84:476–482. Zentler-Munro PL, Assoufi BA, Balasubramanian K, Cornell S, Benoliel D, Northfield TC, Hodson ME. Therapeutic potential and clinical efficacy of acid-resistant fungal lipase in the treatment of pancreatic steatorrhea due to cystic fibrosis. Pancreas 1992; 7:311–319. Smyth RL, van Velzen D, Smyth AR, Lloyd DA, Heaf DP. Strictures of ascending colon in cystic fibrosis and high-strength pancreatic enzymes. Lancet 1994;343:85–86. Taylor CJ. Colonic strictures in cystic fibrosis. Lancet 1994;343: 615–616. Cleasby A, Garman E, Egmond MR, Batenburg M. Crystallization and preliminary x-ray study of a lipase from Pseudomonas glumae. J Mol Biol 1992;224:281–282. Raimondo M, DiMagno EP. Lipolytic activity of bacterial lipase survives better than that of porcine lipase in human gastric and duodenal content. Gastroenterology 1994;107:231–235. Kelly DG, Sternby B, DiMagno EP. How to protect human pancreatic enzyme activities in frozen duodenal juice. Gastroenterology 1991;100:189–195. Kelly DG, Sandberg RJ, Bentley KJ, Zinsmeister AR, DiMagno EP. Protection of lipolytic activity (L) by nutrients in simulated pancreatic insufficiency (abstr). Pancreas 1988;3:601. Holtmann G, Kelly DG, Sandberg RJ, Bentley KJ, Magocsi L, DiMagno EP. Is the survival of human lipolytic activity (LA) during aboral transit affected by the amount of calories, nutrient, chymo-

/ 5e1d$$0018

05-09-97 20:36:49

gasa

19.

20.

21. 22.

23.

24.

25. 26.

27.

28.

29.

30.

31.

trypsin (CT) or bile acid (BA) entering the duodenum (abstr)? Gastroenterology 1991;100:A276. DiMagno EP, Layer P. Human exocrine pancreatic enzyme secretion. In: Go VLW, DiMagno EP, Gardner JD, Lebenthal E, Reber HA, Scheele GA, eds. The pancreas: biology, pathobiology and diseases. New York: Raven, 1993:275–300. Malfertheiner P, Sarr MG, DiMagno EP. Role of the pancreas in the control of interdigestive gastrointestinal motility. Gastroenterology 1989;96:200–205. Cornell JA. Experiments with mixtures. New York: Wiley, 1981. van de Kamer JH, Huinink HTB, Weyers HA. Rapid method for the determination of fat in feces. J Biol Chem 1949;177:347– 355. Chibnall AC, Rees MW, Williams EF. The total nitrogen content of egg albumin and other proteins. Biochem J 1943;37:354– 359. Layer P, Zinsmeister AR, DiMagno EP. Effects of decreasing intraluminal amylase activity on starch digestion and postprandial gastrointestinal function in humans. Gastroenterology 1986;91: 41–48. Fleiss JL. The design and analysis of clinical experiments. New York: Wiley, 1986:61–62. DiMagno EP, Go VLW, Summerskill WHJ. Impaired CCK-PZ secretion, intraluminal dilution and maldigestion of fat in sprue. Gastroenterology 1972;63:25–32. Koike D, Yamdera K, DiMagno EP. Effect of a wheat amylase inhibitor on canine carbohydrate digestion, gastrointestinal function and pancreatic growth. Gastroenterology 1995;108:1221– 1229. Gaskin K, Gurwitz D, Durie P, Corey M, Levinson H, Forstner G. Improved respiratory prognosis in cystic fibrosis patients with normal fat absorption. J Pediatr 1982;100:857–862. Layer P, Yamamoto H, Kalthoff L, Clain JE, Bakken LJ, DiMagno EP. The different courses of early and late onset idiopathic and alcoholic chronic pancreatitis. Gastroenterology 1994;107: 1481–1487. Gullo L, Stella A, Labriola E, Costa PL, Descovich G, Labo G. Cardiovascular lesions in chronic pancreatitis. A prospective study. Dig Dis Sci 1982;27:716–722. Montalto G, Soresi M, Carroccio A, Scafidi E, Barbagallo CM, Ippolito S, Notarbartolo A. Lipoproteins and chronic pancreatitis. Pancreas 1994;9:137–138.

Received November 18, 1996. Accepted March 3, 1997. Address requests for reprints to: Eugene P. DiMagno, M.D., Gastroenterology Research Unit (AL 2-435), Mayo Clinic, 200 First Street Southwest, Rochester, Minnesota 55905. Fax: (507) 255-6318. Supported in part by a grant from Knoll AG, Ludwigshafen, Germany.

WBS-Gastro