Gastrointestinal and pancreatic function in peritoneal dialysis patients: their relationship with malnutrition and peritoneal membrane abnormalities

Gastrointestinal and Pancreatic Function in Peritoneal Dialysis Patients: Their Relationship With Malnutrition and Peritoneal Membrane Abnormalities Abelardo Aguilera, MD, M. Auxiliadora Bajo, MD, PhD, Mauricio Espinoza, MD, PhD, Antonio Olveira, MD, Ana M. Paiva, MD, Rosa Codoceo, MD, PhD, Pilar Garcı´a, MD, Sonia Sa´nchez, RN, Olga Celadilla, RN, Maria Jose Castro, RN, and Rafael Selgas, MD, PhD ● Background: Malnutrition is frequent in peritoneal dialysis (PD) patients, but the contribution of gastrointestinal (GI) dysfunction has not been well established. Methods: We studied GI function in 49 stable PD patients to ascertain its relationship with malnutrition. After an overload fat diet, fecal fat, sugar, starch and nitrogen, intestinal protein permeability (␣1-antitrypsin fecal clearance [C-␣1-AT]), fecal chymotrypsin (CT), GI hormones and gastrin, pepsinogen I and II, cholecystokinin (CCK), gastrin releasing peptide (GRP), and neuropeptide Y (NPY) were measured. Vasoactive intestinal polypeptide (VIP), substance P (SP), and tumor necrosis factor (TNF-␣) and biochemical nutritional markers were evaluated. Results: All patients showed high fecal sugar. Elevated fecal nitrogen was found in 21 patients, 6 with high C-␣1-AT. High fecal starch levels appeared in 21, fat in 20, and low fecal CT in 39 patients. These determinations showed inverse relation with nutritional markers. Increased fecal C-␣1-AT values were associated with lower serum albumin. Fecal CT values showed a negative linear correlation with serum albumin and were inversely associated with retinol-binding protein, normalized protein nitrogen appearance, and serum iron. High plasma levels of pancreatic stimulating hormones were found: gastrin, CCK, and VIP. These levels were higher in patients with a worse pancreatic exocrine function. Higher values of other GI hormones, gastrin, pepsinogen I and II, CCK, GRP, and TNF-␣. Normal concentrations of NPY, VIP, and PS were observed. Conclusion: GI abnormalities (malabsorption, maldigestion, pancreatic dysfunction, and protein losing enteropathy) are present in an important number of PD patients. These features are negatively associated to nutrition. Am J Kidney Dis 42:787-796. © 2003 by the National Kidney Foundation, Inc. INDEX WORDS: Gastrointestinal (GI) malabsorption and maldigestion; protein-losing enteropathy; exocrine pancreatic insufficiency; gastrointestinal (GI) peptides; tumor necrosis factor-␣ (TNF-␣); peritoneal dialysis (PD); malnutrition.

A

VARIETY OF gastrointestinal (GI) tract (GIT) disorders are frequently seen in patients with chronic renal failure (CRF), including stomatitis, gastritis, delayed motility disorders, nodular duodenitis, increased bowel mucosal permeability, microbiological flora alteration, and angiodysplasia.1-6 Alterations in GI peptide plasma levels,7,8 absorption, and mucosal enzymatic activity9-11 have also been described. Many of these uremic GIT abnormalities have been found in postmortem studies or in asymptomatic patients, revealing that some of these phenomenons are clinically inapparent. Changes in pancreatic histology and exocrine function are also commonly reported in CRF.12,13 Several uremic changes such as hyperparathyroidism, chronic acidosis, hypercalcemia, hyperlipemia, elevated plasma levels of GI antagonist hormones (glucagon, somatostatin, pancreatic polypeptide),14,15 or dialysis-related amyloidosis have been implicated.16 Exocrine pancreatic insufficiency has been implicated in the pathogenesis of the wasting syndrome of CRF.17 Increased intestinal permeability has been pre-

viously reported in pre-dialysis subjects.3-5 Potential toxic peptides due to small bowel bacterial overgrowth, acting both in intestinal lumen and plasma, have been suggested to cause proteinlosing enteropathy (PLE).4,5 Inappropriate absorption of glucose, fat, calcium and folate1,9-11 have been described. Local disaccharidase levels were both normal18 and abnormal.9-11 GI hormones, such as cholecystokinin (CCK), gastrin inhibitory peptide, glucagon, neurotensin, vasoactive intestinal polypeptide (VIP), and motilin participate in the neurohormonal control of intestinal transit, hunger-satiety cycle modulaFrom the Servicio de Nefrologı´a, Gastroenterologı´a y Laboratorio de Gastroenterologı´a, Hospitales Universitario de la Princesa y La Paz, Madrid, Spain. Received January 10, 2003; accepted in revised form June 6, 2003. Address reprint requests to Rafael Selgas, MD, PhD, Servicio de Nefrologı´a, Hospital Universitario de la Princesa, Diego de Leo´n 62, 28006 Madrid, Spain. E-mail: [email protected] © 2003 by the National Kidney Foundation, Inc. 0272-6386/03/4204-0024$30.00/0 doi:10.1053/S0272-6386(03)00920-X

American Journal of Kidney Diseases, Vol 42, No 4 (October), 2003: pp 787-796

787

788

AGUILERA ET AL

tion, and meal digestion and absorption. Many of these peptides are catabolized by the kidney and their baseline plasma levels are increased in uremia.1,7,8,19 The role of elevated GI hormone levels in GIT disorders of dialysis patients is uncertain. Elevated plasma levels of cytokines such as tumor necrosis factor-␣ (TNF-␣) and interleukin-1␤ (IL-1␤) have also been implicated in decreasing intestinal absorption and motility.20,21 High levels have been found in dialysis patients,22,23 but their effect on intestinal absorption has not been studied. In addition to GIT changes described in uremia, peritoneal dialysis (PD) may alter the GI function through different mechanisms. These include the mechanical effect of PD fluid that increases intraabdominal pressure, and structural changes of parietal and visceral peritoneum, whose maximal expression is sclerosing-encapsulating peritonitis,24 thus severely decreasing intestinal motility, among other effects. Malnutrition is a frequent finding of PD. However, GIT dysfunctions have not been extensively studied in this population, and their participation in PD malnutrition has not been established. Our aim was to identify GIT dysfunctions and their relationship with malnutrition and peritoneal membrane disorders in PD patients. PATIENTS Our initial intention was to include all active PD patients treated in our unit. As consequence of prior gastroscopic and colonoscopic studies and bacterial overgrowth breath tests guided by anamnesis, 11 patients were considered nonelegible. Local ethics committee and patients were informed about our aim and methods to perform the present study. Finally, we included 49 clinically stable PD patients. No acute disorders were present during the 2 months prior to the study. Patients with known intestinal, pancreatic, or chronic liver diseases; neoplasms; abdominal radiation; infections; or obstructive lung disease were not included. No patient was receiving pancreatic supplements. Causes of chronic renal failure were glomerulonephritis (22.4%), diabetes (18.3%), chronic pyelonephritis (14.8%), polycystic kidney disease (12.2%), nephrosclerosis (12.2%), systemic (8.1%), congenital diseases (6.1%), and unknown (8.2%). The mean period on PD was 30.3 ⫾ 35.7 (1 to 179) months. Nineteen patients did not present GI symptoms and 16 referred mild occasional GI symptoms (dyspepsia, nausea, vomiting, sporadic abdominal pain, intermittent diarrhea, or constipation). Different degrees of anorexia were present in 14 cases. Thirty-nine patients (70.6%) had never been diagnosed of GIT disease. Of the remainder, 8 had been previ-

ously diagnosed of peptic acid disease, 1 of hiatus hernia, and 1 of irritable bowel syndrome and diverticulosis. Thirty-five patients used oral calcium supplements, 32 aluminum-based antiacid, 26 vitamin D3 derivatives, 8 GI prokinetics drugs (domperidone or cisapride), 6 osmotic laxatives (lactulose), 19 ranitidine, 5 oral iron, and 4 omeprazole. Antiacids, laxatives, ranitidine, oral iron, and omeprazole were withdrawn 15 to 10 days before the study. Peritoneal ultrafiltration capacity was considered to be normal in 39 patients (800 ⫾ 200 mL/day, using a combination of 1.36% and 2.27% dextrose), high in 7 patients (using mostly 1.36% dextrose), and low in 3 patients (requiring at least 25% of daily dialysate as 3.86% dextrose).

METHODS We determined: 1. Dialysis adequacy parameters: urea KT/V and normalized protein nitrogen appearance (nPNA) including the residual renal function.25 2. Nutritional markers (fasting conditions): (A) Anthropometric: body mass index (BMI) (weight [kg]/height [m2]). (B) Serum long-term: creatinine, albumin, and cholesterol (colorimetric method; Hitachi 704, Japan). Medium- to short-term: phosphorus, potassium, ferritin, nitrogen (colorimetric method; Hitachi 704), iron (Hitachi 911), prealbumin, retinolbinding protein (RBP), transferrin, fibronectin, ␣1-antitrypsin (immunonephelometric method; Boehring Nephelometer-Terminal S.A., Behringwerke AG, Magnus, Germany), vitamin B12, and folic acid by radio immuno-analysis. 3. Intestinal absorption. All patients received for 3 days a fat overload diet (80 g/day) addressed by a guidediet used in our laboratory. This diet consisted of enriching the food with olive oil. On day number 4 (just 1 day after fat overload), our patients collected the 24-hour stool. Finally, we determined the following parameters in a sample: (A) Stool weight, water content, nitrogen, starch, sugar, and total fat were determined to evaluate malabsorptive syndrome (NIRA method). Fecal nitrogen results mostly from diet protein degradation, but a little amount may be secreted from blood.26 (B) Fecal chymotrypsin (CT) (spectrophotometry) was determined to evaluate pancreatic exocrine function (normal range ⬎23 U/g stool).27 (C) Fecal clearance of ␣1-antitrypsin (fecal C-␣1AT) was determined by the immunonephelometric method to evaluate GIT protein losses (normal ⬍12 mL/24 h). High fecal values of nitrogen with normal C-␣1-AT was determined to define fecal protein maldigestion. If both show high values, it is defined as PLE.26-28 (D) pH, reductors, glucose, and saccharose (Kelly method) were determined to evaluate sugar absorption. 4. GI peptide plasma levels which participate in gastric and intestinal secretion and motility.

GASTROINTESTINAL AND PANCREATIC FUNCTION IN PD PATIENTS

(A) Gastric: plasma gastrin, pepsinogen I and II, CCK, gastrin releasing peptide (GRP), and neuropeptide Y (NPY). (B) Intestinal: VIP, substance P (SP), and TNF-␣. All peptides were determinated by RIA. The CCK, NPY, VIP, GRP, and SP were determined with a kit from Peninsula Labatories (Belmon, CA). Gastrin was determined with a kit from INCSTAR Corporation (Stillwater, MN). Pepsinogen I and pepsinogen II were determined with a kit from SORIN Biomedica Diagnostics S.p.A (Vercelli, Italy). (A) Gastrin: the gammmaDab [125I] gastrin was determined with a precipitating antiserum reagent to separate antibody-bound tracer from unbound tracer. The minimum detectable concentration is 6 pg/mL. Normal range values were from 25 to 115 pg/mL. (B) CCK. The 26 to 33 unsulfated CCK fragment was determined. The sensitivity was calculated from 95% of confidence limit (␣SD) with an interval confidence of 50% (IC50) for 30 replicated samples at the 0 point of the standard curve. The sensitivity is IC50 35 pg/100 ␮L; specificity to CCK 26-33 and CCK 33 has a percentage cross-reactivity of 100%. Values considered normal were 12 to 20 pg/mL. (C) NPY. The sensitivity was IC50 23 pg/tube. The specificity for human, porcine, and rat NPY was 100%. No cross-reactivity was detected for pancreatic polypeptide, VIP, amylin, and human pre-pro-NPY (68-97). Normal values were 20 to 80 pg/mL. (D) Pepsinogen I. Sensitivity is ⬍1 ng/mL at 95% confidence limit. Normal values were 20 to 80 ng/mL. (E) Pepsinogen II. Specificity for this peptide is 100%. Assay sensitivity is 0.3 ng/mL at 95% confidence limit. Normal values were 3 to 20 ng/mL. The ratio of pepsinogen I/pepsinogen II is a noninvasive method to evaluate peptic secretion and functional gastric mucosal status. A high ratio means gastric hypersecretion. The normal range is 2 to 25. (F) VIP. Specificity and sensitivity of 100% (IC50 7 pg/tube). Normal values were 5 to 40 pg/mL. (G) GRP. 14-27 GRP fragment was determinated. Sensitivity (IC50: 25 pg/tube) and specificity of 100% for human and porcine GRP and bombesin. Normal range was between 0.2 and 11 pg/mL. (H) SP. Sensitivity (IC50 4 pg/tube) and specificity of 100% for several SP fragments (1-11, 3-11, 4-11, and 5-11). Normal values were less than 28 pg/mL. (I) TNF-␣. Measured using enzyme amplified sensitive immunoassay performed on microliter plate (Oligoclonal system with monoclonal antibodies; EASIA Medgenix Diagnostics S.A., Belgium, Brussels). Normal values ranged from 3 to 20 pg/mL.

789

A statistical analysis was performed using the Student’s t-test or Wilcoxon for non-paired data and regression analysis. A probability of 95% was considered statistically significant. Results are given as the mean ⫾ standard deviation (SD).

RESULTS

We studied 49 clinically stable PD patients (19 men and 30 women), ranging in age from 22 to 86 years (mean 50 ⫾ 14.2 years). Thirty-seven were on continuous ambulatory PD and 12 were on automatic PD (8 continuous cyclic PD and 4 nocturnal PD). Dialysis Adequacy Parameters The mean weekly urea KT/V was 2.2 ⫾ 0.5, nPNA was 1.04 ⫾ 0.23 g/kg/day, and renal creatinine clearance was 2.17 ⫾ 2.26 mL/min (0.04 ⫾ 0.04 mL/s). General Analytical Data Hemoglobin, 11 ⫾ 1.5 g/dL (110 ⫾ 150 g/L); triglycerides, 160.6 ⫾ 90 mg/dL (1.81 ⫾ 1.02 mmol/L); ferritin, 263.1 ⫾ 213.9 ng/mL (263 ⫾ 214 ␮g/L); blood urea, 146 ⫾ 38 mg/dL (52.1 ⫾ 13.6 mmol/L); creatinine, 9.4 ⫾ 2.2 mg/dL (831 ⫾ 194 ␮mol/L); potassium, 4.4 ⫾ 0.7 mEq/L (4.4 ⫾ 0.7 mmol/L); calcium, 9.5 ⫾ 0.9 mg/dL (2.37 ⫾ 0.22 mmol/L); and phosphorus, 5.5 ⫾ 1.5 mg/dL (1.78 ⫾ 0.48 mmol/L) were measured. Table 1 shows the nutritional markers. Evaluation of GI Protein Digestion and Absorption Twenty-one patients showed elevated fecal nitrogen (Table 2). It is noteworthy that the patients with higher fecal nitrogen had also higher fecal fat and fecal sugar as is demonstrated by a positive linear correlation (r ⫽ 0.71, P ⬍ 0.01; and r ⫽ 0.86, P ⬍ 0.01, respectively). At the same time, the patients with normal fecal nitrogen showed a positive correlation with pepsinogen I/II ratio (r ⫽ 0.51, P ⬍ 0.05). Patients with higher fecal nitrogen showed similar blood urea than the remaining (151.3 ⫾ 36.2 v 147.8 ⫾ 37.5 mg/dL, P ⫽ not significant). Evaluation of Pancreatic Exocrine Function Thirty-nine patients showed lower than normal fecal CT values, including all diabetics. Nondiabetic patients had higher fecal CT levels

790

AGUILERA ET AL Table 1.

Nutritional Markers in PD Patients No. of Patients

Parameter

Mean ⫾ SD

Normal Range

Lower

Normal-Higher

nPNA (g/kg/day) Albumin (g/dL)* Prealbumin (mg/dL)* RBP (mg/dL) Fibronectin (mg/dL) Serum ␣1-AT (mg/dL) Transferrin (mg/dL)* Cholesterol (mg/dL)* Serum Iron (␮g/dL) Vitamin B12 (pg/mL) Folic acid (ng/mL) BMI (kg/m2) Venous pH

1.04 ⫾ 0.23 3.8 ⫾ 0.47 33.4 ⫾ 9.3 12.3 ⫾ 4.6† 40 ⫾ 15.7† 240.5 ⫾ 61.5 255.2 ⫾ 47.6 226.5 ⫾ 46 64.8 ⫾ 23.8 781 ⫾ 338 7.1 ⫾ 3.2 26.5 ⫾ 5.7† 7.35 ⫾ 0.04†

⬎0.9 3.8-5 ⬎30 3-6 25-40 150-350 ⬎200 ⬎100 50-145 150-750 2-10 22-25 7.35-7.45

13 17 19 8 8 5 5 0 11 0 0 6 20

36 32 30 41 41 44 44 49 38 49 49 43 29

NOTE. To convert albumin in g/dL to g/L, multiply by 10; transferrin in mg/dL to g/L, multiply by 0.14; Cholesterol in mg/dL to mmol/L, multiply by 0.02586; iron in ␮g/dL to ␮mol/L, multiply by 0.179; vitamin B12 in pg/mL to pmol/L, multiply by 0.738; folic acid in ng/mL to nmol/L, multiply by 2.266. Abbreviations: RBP, retinol-binding protein; ␣1-AT, ␣1-antitrypsin. *Values accepted as normal in PD patients. †Different from normal range.

than diabetics (14.7 ⫾ 10.7 v 4.8 ⫾ 3.7 U/g, P ⬍ 0.001). Patients using GI prokinetic drugs had lower fecal CT than patients without (5.4 ⫾ 4.1, n ⫽ 8 v 11.3 ⫾ 10.5, P ⬍ 0.05). Fecal CT and serum albumin and SP showed a significant positive linear correlation (r ⫽ 0.33, P ⬍ 0.05; and r ⫽ 0.55, P ⬍ 0.01, respectively). On the contrary, fecal CT and fecal fat showed a negative correlation (r ⫽ ⫺0.33, P ⬍ 0.05); this was strong in patients with abnormal fecal fat (⬎6 g/24 h stool) (r ⫽ ⫺0.5, n ⫽ 20, P ⬍ 0.05). All patients with lower fecal CT showed higher Table 2.

fecal starch values than the remaining (5.7 ⫾ 3.8 v 3.3 ⫾ 1.77 g/24 h stool, P ⬍ 0.01). However, patients with high fecal nitrogen showed lower fecal CT than the remaining (6.5 ⫾ 4.4, n ⫽ 21 v 16.4 ⫾ 8.3 U/g stool, P ⬍ 0.05). Pancreatic stimulating hormones showed a negative correlation with fecal CT: CCK (r ⫽ ⫺0.33, P ⬍ 0.05) and gastrin (r ⫽ 0.32, P ⬍ 0.05). Table 3 shows the negative relationships between fecal CT levels and nutritional markers. No relation was found between fecal CT, venous pH, PTH, calcium, or phosphorus levels.

Intestinal Absorption Parameters No. of Patients

Substance

Mean ⫾ SD

Normal Range

Lower

Normal

Higher

Fecal pH Fecal sugar (g/24 h) Fecal CT (U/g) Fecal fat (g/24 h) Fecal nitrogen (g/24 h) Fecal starch (g/24 h) Fecal water (%) C-␣1-AT (mg/24 h)

6.44 ⫾ 0.71 3.03 ⫾ 1.93* 12.9 ⫾ 10.5* 5.27 ⫾ 3.6 2 ⫾ 1.08* 1.7 ⫾ 3.6* 84.6 ⫾ 5.1 7.28 ⫾ 6.55

6-8 0 ⬎23 0-6 0-2 0 65-85 ⬍12

4 0 39 0 0 0 0 0

45 0 10 29 28 28 25 25

0 49 0 20 21 21 24 6

Abbreviations: CT, chymitrypsim; C-␣1-AT, clearance of ␣1-antitrypsin. *Different from normal range.

GASTROINTESTINAL AND PANCREATIC FUNCTION IN PD PATIENTS Table 3.

Relationship Between Fecal CT and Nutritional Markers

Parameter

Iron (ng/mL)* RBP (mg/dL)* Urea KT/V nPNA (g/kg/day)* Plasma NPY (pg/mL)

Fecal

CT

⬍50 5.5 ⫾ 3.8 ⱖ50 14 ⫾ 10.8 ⬍6 3.3 ⫾ 1.9 ⱖ6 14 ⫾ 10 ⬍2 5.7 ⫾ 2.6 ⱖ2 14.1 ⫾ 8 ⬍0.9 6.8 ⫾ 5.2 ⱖ0.9 15 ⫾ 11 ⬍50 6.4 ⫾ 4.4 ⱖ50 11.8 ⫾ 10.5

n

11 39 8 41 11 38 13 36 18 31

P

⬍0.01 ⬍0.001 ⬍0.01 ⬍0.01 ⬍0.05

*Nutritional marker.

Evaluation of Intestinal Protein Losses Elevated fecal C-␣1-AT (⬎12 mL/24 h stool) was present in 6 of 31 patients. All of them had high fecal nitrogen, defining them as having PLE. This was associated with poor nutritional status: lower serum albumin (3.57 ⫾ 0.57 v 3.98 ⫾ 0.38 g/dL [35.7 ⫾ 5.7 v 39.8 ⫾ 3.8 g/L], P ⬍ 0.05) and transferrin (243 ⫾ 70 v 272 ⫾ 44.3 mg/dL [2.4 ⫾ 0.7 v 2.7 ⫾ 0.4 g/L], P ⬍ 0.05), as well as lower triglycerides (131.3 ⫾ 31.7 v 187 ⫾ 116 mg/dL [1.48 ⫾ 0.36 v 2.11 ⫾ 1.31 mmol/L], P ⬍ 0.05). In 31 patients, a negative linear correlation between C-␣1-AT and nPNA was found (r ⫽ ⫺0.35, n ⫽ 31, P ⬍ 0.05). On the other hand, a positive linear correlation between C-␣1-AT and fecal fat (r ⫽ 0.45, P ⬍ 0.05), fecal starch (r ⫽ 0.44, P ⬍ 0.05), and fecal nitrogen (r ⫽ 0.43, P ⬍ 0.05) was found. The last one was stronger when fecal nitrogen was abnormally high (⬎2 g/24 h stool) (r ⫽ 0.65, n ⫽ 15, P ⬍ 0.01). Evaluation of Sugar Losses (A) Fecal pH (Table 2): Four patients showed fecal pH ⬍ 5. Reductor bodies, fecal glucose, and saccharose were present in only 1 patient. (B) Fecal starch: Twenty-one (9 diabetics) patients showed positive fecal starch. These patients showed lower albumin and prealbumin and higher CCK than the remaining patients (3.6 ⫾ 0.5 v 3.9 ⫾ 0.4 g/dL [36 ⫾ 5 v 39 ⫾ 4 g/L], P ⬍ 0.05; 29.8 ⫾ 9.1 v . 34.9 ⫾ 9.5 mg/dL, P ⬍ 0.05; and 57.3 ⫾ 36.2 v 35 ⫾ 24.1 pg/mL, P ⬍ 0.05, respectively). (C) Fecal sugar: All patients had high fecal

791

sugar levels. This abnormality showed no relationship with biochemical nutritional markers. Strong correlations between fecal sugar and fecal nitrogen (r ⫽ 0.86, P ⬍ 0.001) and fecal fat (r ⫽ 0.48, P ⬍ 0.01) were found. We did not find relation between fecal sugar, GI prokinetics drugs, laxatives, and GI peptide. Evaluation of Fecal Fat Losses Twenty patients had elevated fecal fat losses (⬎6 g/24 h stool). A negative linear correlation between fecal fat and nPNA (r ⫽ ⫺0.32, P ⬍ 0.05) and serum total protein (r ⫽ ⫺0.34, P ⬍ 0.05) was found. Patients with positive fecal fat showed lower prealbumin than the remaining (28.9 ⫾ 9.3 mg/dL v 35.5 ⫾ 8.2, P ⬍ 0.05). Similarly, patients with fibronectin lower than 40 mg/dL had higher fecal fat losses (7 ⫾ 3.8, n ⫽ 20 v 4.33 ⫾ 3.4 mg/24 h stool, n ⫽ 29, P ⬍ 0.05). Fecal fat losses showed a positive correlation with fecal nitrogen (r ⫽ 0.71, P ⬍ 0.01), fecal sugar (r ⫽ 0.48, P ⬍ 0.01), and fecal C-␣1-AT (r ⫽ 0.45, n ⫽ 31, P ⬍ 0.05). We did not find relation between fecal fat losses and prokinetic drugs intake, laxatives, CCK, GRP, VIP, NPY, and TNF-␣ values. Plasma GI Peptides and TNF-␣ Table 4 shows plasmatic levels of gastrin, CCK, GRP, pepsinogen I and II, pepsinogen I/pepsinogen II ratio, and TNF-␣. GI Symptoms and GI Dysfuction We did not find an association between GI symptoms and GI absorption disorders, except for 2 of 3 patients with intermittent diarrhea who showed low fecal CT, suggestive of pancreatic exocrine deficiency. DISCUSSION

Malnutrition in PD patients is clearly associated with a high mortality.29 It is well known that the first requirement for adequate nutritional status is the structural and physiological GIT integrity. Our study demonstrates that in PD patients, severe GIT disorders, both in digestion and absorption, are present. The participation of these GIT disorders in malnutrition has not been previously acknowledged. The predominance of female patients was ab-

792

AGUILERA ET AL Table 4.

Plasma Peptides and Cytokine Implicated in GIT Function in PD Patients No. of Patients

Substance

Mean ⫾ SD

Normal Range

Lower

Normal

Higher

Gastrin (pg/mL) CCK (pg/mL) NPY (pg/mL) GRP (pg/mL) VIP (pg/mL) SP (pg/mL) Pepsinogen I (ng/mL) Pepsinogen II (ng/mL) Pepsinogen I/Pepsinogen II TNF-␣ (pg/mL)

200.9 ⫾ 128.4* 48.4 ⫾ 32* 50.7 ⫾ 27.8 15.2 ⫾ 8.8* 38.5 ⫾ 15.5 8.3 ⫾ 32.4 453.1 ⫾ 298* 31 ⫾ 17.1* 16.2 ⫾ 9.5* 67.8 ⫾ 31.5*

25-115 12-20 20-80 0.2-11 5-40 ⬍27 20-80 3-20 1-25 3-20

1 5 10 0 0 0 0 0 4 0

16 4 33 13 27 48 5 14 36 1

32 40 6 36 22 1 35 26 0 48

*Different from the normal range.

solutely randomized by the cross-sectional methodology applied in the selection of patients. We consider that in the general population or uremic patients there are no data on sex as a factor in GI absorption disorders. Our patients displayed normal range of dialysis adequacy parameters. However, as Table 1 shows, several nutritional parameters are clearly abnormal in this series, as usual in PD patients.29 Although some of our patients showed GIT symptoms, we did not find an association between these symptoms and GIT disorders, except for patients with anorexia who showed high plasma levels of TNF-␣ (a powerful anorexigen) and relatively lower levels of NPY (a powerful orexigen). These findings were previously described.23 On the other hand, 2 of 3 patients with intermittent diarrhea showed low fecal CT, suggestive of pancreatic exocrine deficiency. Evaluation of Protein Digestion and Absorption Isolated protein maldigestion was demonstrated in almost half of our series according to the presence of high fecal nitrogen and normal fecal C-␣1-AT. In uremia, the GIT is exposed to great amounts of creatinine and urea, both being metabolized by intestinal flora. Urea is transformed into ammonia (converted then to urea in the liver by entero-hepatic cycle) and carbon dioxide.9 The absence of a significant relationship between serum urea and fecal nitrogen suggests that fecal nitrogen is mainly derived from dietary protein, ruling out false positive data. The studies on amino acid absorption in uremic patients have shown disorders of brush bor-

der peptidases activity: splitting of oligopeptides and 10-fold reduced glycyl-l-leucine and l-alanyll-proline dipeptidases activity.9-11,30 Moreover, protein malnutrition causes intestinal peptidase deficiency.18 This represents a reciprocal effect of protein-malnutrition and protein maldigestion. The measurement of serum pepsinogens and pepsin-precursors is a noninvasive reliable biochemical method to evaluate peptic secretion and functional gastric mucosal status.31 Hyperpepsinogenemia is a risk factor for peptic acid disease,19,32 whereas hypopepsinogenemia is suggestive of severe atrophic gastritis and intestinal metaplasia.33 In uremia, elevated plasma levels could result from hyperproduction induced by gastrin, CCK, and SP19,34 and/or accumulation due to the lack of renal excretion.1,7,10,19 According to our data, in patients without protein maldigestion or malabsorption (fecal nitrogen ⬍2 g/24 h stool), the positive correlation between pepsinogen I/pepsinogen II ratio and fecal nitrogen could indicate preservation of the normal cycle between protein intake and pepsin secretion, despite the high concentration of intraluminal urea. Other important factors in intestinal absorption are the bile and pancreatic exocrine secretion. The bile is composed of proteolytic enzymes such as trypsin, chymotrypsin, carboxipeptidases, amylase, lipase, water, and bicarbonate. Fecal CT level is used to identify subjects with pancreatic exocrine deficiency with a sensitivity of nearly 80%.35 Surprisingly, in our series, 91% of patients showed lower than normal values, which is

GASTROINTESTINAL AND PANCREATIC FUNCTION IN PD PATIENTS

indicative of pancreatic exocrine deficiency. Another more sensitive test, elastase-1, has been recently proposed to evaluate pancreatic exocrine function. However, until 2001 this test had not been demonstrated to be superior over fecal CT,36 and this is manly applicable for patients with severe pancreatic dysfunction (cystic fibrosis) or taking oral enzyme supplements.37-39 Finally, we cannot classify our patients according to the severity of pancreatic exocrine dysfunction because there are no clear criteria for classification, and the criteria that do exist are based on image techniques.14,40 Several reasons could explain this pancreatic dysfunction: (A) High plasma levels of secretagogue pancreatic peptides such as secretin, gastrin, and CCK (the last 2 being elevated in our increased pancreatic secretion) could be the reason. However, there are different degrees of augmentation in these hormones. Curiously, and according to our data, fecal CT showed a negative correlation between gastrin and CCK, a finding similar to that in patients with chronic pancreatitis,19,40 and might suggest an uremic generalized debilitation of glandular secretion in some patients. It could be assocociated to other factors such as hyperparathyroidism, hypertriglyceridemia, calcium infiltration, amyloidosis, or extracellular volume changes.14-16 There are also hormones with inhibitory effects on pancreatic secretion such as glucagon and human pancreatic polypeptide.1,14,15,19 Unfortunately, as we did not study these substances, we cannot conclude on their final role in pancreatic exocrine function. In particular, among PD patients, there is a stimulation of GI peptides secretion due to the constant glucose peritoneal absorption.23 Moreover, it has also been suggested that dialysate acid pH and hypertonicity may irritate the pancreas head.41 However, we did not find a relationship between the PD glucose concentration administered to our patients and the secretagogue hormones or fecal CT. We wonder about the contribution of this phenomenum in uremic pancreatic function. (B) Abnormalities in composition of pancreatic juice could be the reason. Diminished bicarbonate pancreatic secretion has been reported in uremic patients.1,14,15 We have found indirect signs of abnormal pancreatic juice composition by the presence of slight acid stools, with the

793

mean pH being 6.4 ⫾ 0.7 (4 patients showed fecal pH ⬍5; all of them had starch malabsorption). These patients could have an inadequate fecal pH range for CT activity. A lower intraluminal pH is related to sugar malabsorption,42,43 hypergastrinemia, and hyperpepsinogenemia I.19 Similar to patients with chronic pancreatitis, whose fecal pH is nearly 4,43 our PD patients could suffer a bile acid-like syndrome. On the other hand, decreased amylase and lipase output has been described in uremic subjects.1,14,15,25,26 Our results could support this, due to the fact that we found an inverse relationship between fecal fat and fecal CT in patients with steatorrhea. We have found an association between fecal fat digestive disorder and malnutrition. Our results are in agreement with those published by Sachs et al,17,44 who successfully administered pancreatic supplements to uremic patients, similar to patients with chronic pancreatitis. (C) Protein malnutrition as cause/effect of pancreatic dysfunction. The protein pool is an important factor to maintain an adequate pancreatic exocrine secretion. Pancreas is an affected organ in the first stage of protein deficiency.35,45 Protein malnutrition is the most frequent feature in malnourished PD patients. Supporting this hypothesis, our results show an inverse correlation between fecal CT and serum albumin, and a positive correlation with food intake markers (nPNA and NPY) (Table 3). Both lower nPNA and NPY could act as food intake markers. We recently published that NPY values ⬍50 pg/mL are associated with anorexia and malnutrition in PD patients.23 On the contrary, TNF-␣, a cytokine able to induce cachexia, was elevated according to lost renal function.20,23 However, serum TNF-␣, an excellent inflammatory marker, did not show significant positive correlation with nutritional markers, suggesting the partial independence of serum albumin from a hidden inflammatory process. As a consequence, it is possible to admit the role of serum albumin as a nutritional marker. (D) Structural pancreatic changes in uremia have been found in 44% of cases. The main histologic features are dilatation of acini, flattening of acinar cells with loss of zymogengranules, and inspissated secretions with dilatation of tiny ductules.1,13,46 Long-term dialysis and drugs could

794

AGUILERA ET AL

Fig 1. Summary of GIT abnormalities found in our patients. All of the patients studied showed some GIT abnormality (maldigestion, malabsorption, or PLE).

modify this uremic pancreatic tract changes.46 Accelerated atherosclerosis, uremic neuropathy, and long survival in dialysis could be involved in the pathogenesis of these conditions.14,47-49 Unfortunately, we did not study pancreatic structure. PLE has been reported in uremic patients.3-5 Elevated fecal C-␣1-AT, the definitive diagnostic criteria for PLE, was found in 6 of 31 patients. Protein malnutrition is a generally accepted consequence of PLE caused by Crohn’s or other diseases.50 When we analyzed the causes of PLE in our patients, the influence of CCK could be discarded, because it is known that CCK promotes intestinal translocation of immunoglobulins, albumin, and electrolytes.51 Also, we found no correlation between CCK and fecal C-␣1-AT. In a previous work, we found an association between PLE and lower ultrafitration, high urea and creatinine MTCs, long-term PD, and clinical peritonitis history. We hypothesized that changes in peritoneal membrane chronically induced by PD may produce peritoneal lymphatic and venular dilatation, resulting in local protein translocation into the intestinal lumen.52 In regards to fecal sugar loss, curiously, all of our patients had sugar malabsortion, but lower fecal pH only was found in 4 patients. Three reasons could explain this finding. In uremic status the column acquires some renal function as water and sodium excretion. It is possible that in PD patients, the glucose absorbed crosses the peritoneum to the blood. Finally, the glucose could be eliminated through the column.53 On the other hand, reduced activity of mucosal sacar-

ase, lactase, and maltase with the subsequent sugar malabsorption have been reported in CRF.10,11,51 This could be the second explanation to our finding. However, intestinal dipeptidases deficiency was not found by others.18 Finally, the inverse relationship between fecal sugar and gastrin levels could be related to acid hypersecretion, which decreases fecal pH and causes sugar malabsorption. However, in our patients, this possibility could be because real acid stool was found only in 4 cases. In regards to fecal fat loss, to digest a fat molecule the GIT indemnity, lymphatic vessels, and normal composition of bile are necessary.35 This explains the close relationship between fecal fat and the remaining of fecal parameters (fecal nitrogen, sugar, and C-␣1-AT).35 We found fat malabsorption in 20 patients pointing to a GI malabsorptive disorder that potentially may worsen the malnutrition of PD patients. The inverse relation between fecal fat, nutritional parameters, and fecal CT (r ⫽ ⫺0.3) could indicate a uremic bile composition abnormality as responsible at least in part of fat malabsorption. Figure 1 summarizes the GIT abnormalities found in our patients. On the other hand, we explored the association between inflammation, GI peptides, and GI function. Recently, it has been published that TNF-␣ is able to induce intestinal malabsorption and decrease in the intestinal motility.20,21 We did not find this relationship; however, an improvement in the nutritional status after treating the gastroparesia of dialysis patients has been described.54

GASTROINTESTINAL AND PANCREATIC FUNCTION IN PD PATIENTS

Finally, the early diagnosis of these GIT uremic disorders, and the eventual treatment with pancreatic enzymes,44 antagonist H2 drugs, antibiotics in cases of intestinal bacterial overgrowth, and prokinetics drugs,54 could improve and prevent the nutritional status in PD patients. Moreover, in accord with our results in PD patients, GI malabsorption and maldigestion are silent processes (nonassociated to GI symptoms). CONCLUSION

Fecal sugar, fat and starch malabsorption, protein maldigestion, PLE, and exocrine pancreatic insufficiency are frequently found in PD patients. These severe GI abnormalities are associated with malnutrition, and especially when exocrine pancreatic insufficiency exists, where the protein deficiency could cause low fecal chymotrypsin conditioning a vicious cycle. GI peptides, peritoneal membrane transport disorders, and high plasma concentration of TNF-␣ could play an indirect role in the pathogenesis of intestinal maldigestive and malabsorptive in PD patients. REFERENCES 1. Kang JY: The gastrointestinal tract in uremia. Dig Dis Sci 38:257-268, 1993 2. Zukerman GR, Mills BA, Koehler RE, Siegel A, Harter HR, DeSchryver-Kecskemeti K: Nodular duodenitis pathologic and clinical characteristics in patients with endstate renal disease. Dig Dis Sci 28:1018-1024, 1983 3. Johansson SV, Odar-Cederlo¨ f I, Plantin LO, Strandberg PO: Albumin metabolism and gastrointestinal loss of protein in chronic renal failure. Acta Med Scand 201:353358, 1977 4. Magnusson M, Magnusson KE, Sundqvist T, Denneberg T: Increased intestinal permeability to differently sized poluthylene glycols in uremic rat: Effects of low- and high- protein diets. Nephron 56:306-311, 1990 5. Magnusson M, Magnusson K-E, Sundqvist T, Denneberg T: Impaired intestinal barrier fuction measured by differently sized polyethylene glycols in patients with chronic renal failure. Gut 32:754-759, 1991 6. Vaziri ND, Dure-Smith B, Miller R, Mirahmadi MK: Pathology of gatrointestinal tract in hemodialysis patients: An autopsy study of 78 cases. Am J Gastroenterol 80:608611, 1985 7. Sirinek KR, O’Dorisio T, Gaskill HV, Levine BA: Chronic renal failure: Effect of hemodialysis on gastrointestinal hormones. Am J Surg 148:732-735, 1984 8. Hegbrant J, Thysell H, Ekman R: Plasma levels of gastrointestinal regulatory peptides in patients receiving mantenance hemodialysis. Scand J Gastroenterol 26:599604, 1991 9. Wizemann V, Ludwimg R, Kuhl R, Burgmann R:

795

Digestive-absorptive function of the intestinal brush in uremia. Am J Clin Nutr 31:1642-1646, 1978 10. Denneberg T, Lindberg T, Berg NO, Dahlqvist A: Morphology, dispeptidases and disaccharidases of small intestinal mucosa in chronic renal failure. Acta Med Scand 195:465-470, 1974 11. McNair A, Olsen J: Disaccharidase activity in chronic renal failure. Acta Med Scand 195:93-96, 1974 12. Araki T, Ueda M, Ogawa K, Takao T: Histological pancreatitis in end-state renal disease. Int J Pancreatol 12:263269, 1992 13. Vaziri ND, Dure-Smith B, Mirahmadi M: Pancreatic pathology in chronic dialysis patients—An autopsy study of 78 cases. Nephron 46:347-349, 1987 14. Abu-Alfa A, Ivanovich P, Mujaus SK: Uremic exocrine pancreopathy. Nephron 48:94-100, 1988 15. Esker AH, Mujais SK: Uremic exocrine pancreophaty. Int J Artif Organs 10:135-138, 1987 16. Campistol JM, Cases A, Torras A, et al: Viceral involvement of dialysis amyloidosis. Am J Nephrol 7:390393, 1987 17. Sachs EF, Hurwitz FJ, Bloch HM, Milne FJ: Pancreatic exocrine hypofunction in the wasting syndrome of end-state renal disease. Am J Gastroenterol 78:170-173, 1983 18. El-Lakany S, Eagon PK, Gavaler JS, Schade RR, Whiteside T, Van Thiel D: Gastrointestinal fuction, morphology, and immune status in uremia. Nutrition 6:461-468, 1990 19. Walsh JH: Gastrointestinal peptide hormones, in Sleisinger F (ed): Gastrointestinal Disease Pathophysiology, Diagnosis and Management (ed 4). Philadelphia, PA, Saunders, 1989, pp 78-106 20. Cerami A: Tumor necrosis factor as a mediator of shock, cachexia and inflamation. Blood Purif 11:108-117, 1993 21. Bueno L, Fioramonti J: Neurohormonal control of intestinal transit. Reprod Nutr Dev 34:513-525, 1994 22. Pereira BJG, Shapiro L, King AJ, Falagas ME, Strom JA, Dinarello CA: Plasma levels of IL-1␤, TNF-␣, and their specific inhibitors in undialyzed chronic renal failure, CAPD and hemodialysis patients. Kidney Int 45:890-896, 1994 23. Aguilera A, Codoceo R, Selgas R, et al: Orexigen (neuropetide Y) and anorexigen (tumoral necrosis factor alpha and cholecystokinin) plasma levels in peritoneal dialysis (PD) patients. Their relationship with nutritional parameters. Nephrol Dial Transplant 13:1476-1483, 1998 24. Honda K, Nitta K, Horita S, Yumura W, Nihei H: Morphological changes in the peritoneal vasculature of patients on CAPD with ultrafiltration failure. Nephron 72: 171-176, 1996 25. Selgas R, Bajo MA, Fernandez-Reyes MJ, et al: An analysis of adequacy in a selected population on CAPD for over 3 years: The influence of urea and creatinine kinetics. Nephrol Dial Transplant 8:1244-1253, 1993 26. Codoceo R, Mun˜ oz-Codoceo R, Lama R, Garcia P, Rebollo F: Screening of malabsorption by mean infrared reflectance analysis (NIRA). J Pediatr Gastroenterol Nutr 24:470, 1997 27. De Pedro C, Codoceo R, Vazquez P, Hernanz A: Fecal

796

chymotrypin levels in children with pancreatic insufficiency. Clin Biochem 19:338-340, 1986 28. Codoceo R, Lama R, Hernanz A: Alteraciones del vaciamiento ga´ strico en pacientes con reflujo-gastro-esofa´ gico: Papel de la colecistoquinina. Quim Clin 13:362, 1994 29. Yung G, Kopple J, Lindholm B, et al: Nutritional assessment of continuous ambulatory peritoneal dialysis patients: An international study. Am J Kidney Dis 17:462471, 1991 30. Sterner G: Small intestine function in experimental uremia. Contr Nephrol 60:188-191, 1988 31. Plebani M: Pepsinogens in health and disease. Crit Rev Clin Lab Sci 30:273-328, 1993 32. Paimela H, Ha¨ rko¨ nen M, Karonen S-L, Tallgren LG, Stenman S, Ahonen J: Relation between serum group II pepsinogen contration and the degree of brunner’s gland hyperplasia in patients with chronic renal failure. Gut 26:198202, 1985 33. Muto S, Asana Y, Hosoda S, Miyata M: Hypochlorhydria and hypergastrinemia and their association with gastrointestinal bleeding in undialyzed and hemodialyzed patients. Nephron 50:10-13, 1988 34. Beglinger C, Hildebrand P, Meier R, et al: A physiological role for cholecystokinin as a regulator of gastrin secretion. Gastroenterology 103:490-495, 1992 35. Riley SA, Turnberg LA: Maldigestion and malabsorption, in Sleisinger F (ed): Gastrointestinal Disease Pathophysiology, Diagnosis and Management (ed 4). Philadelphia, PA, Saunders, 1989, pp 1009-1026 36. Luth S, Teyssen S, Forssmann K, Kolbel C, Krummenauer F, Singer MO: Fecal elastase-1 determination: “Gold standard” of indirect pancreatic function test? Scand J Gastroenterol 36:1092-1099, 2001 37. Walkowiak J, Nousia-Arvanitakis S, Agguridaki C, et al: Longitudinal follow-up of exocrine pancreatic function in pancreatic sufficient cystic fibrosis patients using the fecal elastase-1 test. J Pediatri Gastroenterol Nutr 36:474-478, 2003 38. Dominic R, Franzini C: Fecal elastase-1 as a test for pancreatic function: A review. Clin Chem Med 40:325-332, 2002 39. Lankisch PG, Schmidt I, Konig H, et al: Fecal elastase-1: Not helpful in diagnosing chronic pancreatic associated with mild to moderate exocrine pancreatic insufficiency. Gut 42:551-554, 1998 40. Go´ mez-Cerezo J, Garces MC, Codoceo R, et al: Postpandrial glucose-dependent insulinotropic polypeptide and insulin responses in patients with chronic pancreatitis wit and without secondary diabetes. Regul Pept 67:201-205, 1996

AGUILERA ET AL

41. Caruana RJ, Wolfman NT, Karstaedt N, Wilson DJ: Pancreatitis. An important cause of abdominal symptoms in patients on patients on peritoneal dialysis. Am J Kidney Dis 7:135-140, 1986 42. Moses R, Panebianco P, Dinoso VP, et al: Decreased pancreatic bicarbonate (HCO3) responsiveness to graded doses of secretin (s) in end stage renal disease. Gastroenterology 86:1190, 1984 43. Nakamura T, Kikuchi H, Takebe K, et al: Correlation between bile acid malabsorption and pancreatic exocrine dysfunction in patiens with chronic pancreatitis. Pancreas 9:580-584, 1994 44. Sachs EF, Bloch HM, Milne FL: Pancreatic supplementation in end-state renal disease. Nephron 37:120-122, 1984 45. Korsten MA, Lieber CS: Nutrition in pancreatic and liver disorders, in Shils ME, Olson JA, Shike M (eds): Modern Nutrition in Health and Disease (ed 8). Philadelphia, PA, Lea and Febiger, 1994, pp 1066-1080 46. Chachati A, Godon JP: Effect of haemodialysis on gastrointestinal tract pathology in patients with chronic renal failure. Nephrol Dial Transplant 1:233-237, 1987 47. Heidbreded E, Schafferhans K, Heidland A: Disturbances of peripheral and autonomic nervous system in chronic renal failure. Effects of hemodialysis and transplant. Clin Nephrol 23:222-228, 1985 48. Dooley CP, Valenzuela JE: Duodenal volumen and osmoreceptors in the stimulation of human pancreatic secretion. Gastroenterology 86:23-27, 1984 49. Henderson JR, Daniel PM, Fraser PA: The pancreas as a sigle organ: The influence of the endocrine upon the exocrine part of the gland. Gut 22:158-167, 1981 50. Graham JH: Protein-losing gastroenteropathy, in Sleisinger F (ed): Gastrointestinal Disease Pathophysiology, Diagnosis and Management (ed 4). Philadelphia, PA, Saunders, 1989, pp 283-290 51. Freier S, Eran M, Faber J: Effect of cholecystokinin and of its antagonist, of atropine, and of food on the release of immunoglobulin A and immunoglobulin G specific antibidies in the rat intestine. Gastroenterology 93:1242-1246, 1987 52. Aguilera A, Bajo MA, Selgas R, et al: Protein losing enterophaty is associated with peritoneal functional abnormalities in peritoneal dialysis patients. Perit Dial Int 20:284289, 2000 53. Friedman EA: Bowel as a kidney substitute in renal failure. Am J Kidney Dis 28:943-950, 1996 54. Ross EA, Koo LC: Improved nutrition after the detection and treatment of occult gastroparesis in nondiabetic dialysis patients. Am J Kidney Dis 31:62-66, 1998