HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Schuette, S. A. Articles by DiMagno, E. P. Search for Related Content PubMed PubMed Citation Articles by Schuette, S. A. Articles by DiMagno, E. P.

Effect of Triglyceride Structure on Fecal Excretion of ¹³C-Labeled Triglycerides

BioChemAnalysis Corp (S.A.S., M.J.), Chicago, Illinois
Center for Stable Isotope Research Inc (M.J.), Chicago, Illinois
Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio (M.B.C., S.K., T.S)
Metabolic Solutions Inc., Nashua, New Hampshire (D.A.W.)
Mayo Clinic and Foundation, Rochester, Minnesota (E.P.D.)

Address reprint requests to: M Janghorbani, BioChemAnalysis Corp., 2201 W Campbell Park Dr, Chicago, IL 60612-3501. E-mail: mjanghor{at}hotmail.com

	ABSTRACT

TOP FOOTNOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS CONCLUSIONS ACKNOWLEDGMENTS REFERENCES

 
Objective: The aim of this work was to determine the effects of specific changes in the structure of ¹³C-labeled triglyceride (TG*) on its fecal excretion relative to total stool fat excretion determined simultaneously in patients with reduced exocrine pancreatic function.

Methods: A series of 47 studies were conducted in 26 young cystic fibrosis (CF) patients and 11 adult patients with chronic pancreatitis over a five year period. Each test consisted of ingesting a single high fat test meal containing both ¹³C-labeled triglyceride (TG*) and dysprosium chloride (DyCl₃) a nonabsorbable marker of intestinal transit; in most studies the food colorant brilliant blue (FD&C blue #1) was administered along with the DyCl₃. The TG*s tested were: P*P*P* = TRIPALMITIN-1,1,1-¹³C₃; SO*S = 2-OCTANOYL-1,3-DISTEARIN-2-octanoyl-1,2-¹³C₂; and P*LP* = 2-LAURYL-1,3-DIPALMITIN-dipalmitoyl-1,1,2,2-¹³C₄. Ingestion of the test meal was followed by collection of individual stools for at least 72 hours. Stools were analyzed for ¹³C-Excess (¹³C*), total fat, and Dy.

Results: Excretion of P*LP* showed a high degree of linear correlation with stool fat (r² = 0.924) over a wide-range of fecal fat values. Excretion of SO*S was also significantly correlated with stool fat, but its excretion was less than 10% at all levels of steatorrhea and the slope of the regression line relating TG* excretion to stool fat was some four to five times smaller than observed for P*LP*. Fecal excretion of P*P*P* was highly correlated with stool fat (r² = 0.941) in patients with moderate steatorrhea (<25 g fat/24 hours) and the slope of the regression line (3.20) was considerably greater than for P*LP*. Only results from those studies in which stool collections were complete (Dy excretion >90%) were utilized in the statistical comparisons (36 of 47 studies).

Conclusions: The observed highly significant linear correlation between P*LP* and stool fat over the entire range of steatorrhea suggests that P*LP* excretion may be a suitable surrogate for fecal fat in patients with reduced exocrine pancreatic function. Because fecal excretion of TG* administered as described can be accurately determined by sampling only two visually marked stools, development of a noninvasive test to replace the current 72-hour stool fat test using this approach is possible. Use of other engineered TG*s and/or labeled fatty acids, may provide a method for non-invasive in vivo assessment of the specific defect(s) leading to steatorrhea in other patient groups.

Key words: ¹³C-label, malabsorption, triglyceride structure, cystic fibrosis, chronic pancreatitis

Abbreviations: Dy = dysprosium • TG* = ¹³C-labeled triglyceride • cystic fibrosis = CF • ¹³C* = ¹³C-Excess • P*P*P* = TRIPALMITIN-1,1,1-¹³C₃ • SO*S = 2-OCTANOYL-1,3-DISTEARIN-2-octanoyl-1,2-¹³C₂ • P*LP* = 2-LAURYL-1,3-DIPALMITIN-dipalmitoyl-1,1,2,2-¹³C₄

	INTRODUCTION

TOP FOOTNOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS CONCLUSIONS ACKNOWLEDGMENTS REFERENCES

 
Steatorrhea (excess fecal fat) occurs commonly in patients suffering from a wide range of disorders, and its quantitative assessment is important in their diagnosis and management [1]. The current "Gold Standard" for this purpose is the 72-hour fecal fat test [1,2]. This test, however, suffers from important limitations. The method relies on the quantitative relationship between dietary intake of fat and its fecal excretion [1,2]. For this method to be accurate, patients must consume an adequate and known intake of fat for at least 72 hours and collect all stools corresponding to this period of intake. These requirements have rendered the approach impractical and underutilized [3]. In addition, interpretation of the results from this test is always suspect, because of questions regarding patient compliance and quantitative collection of stools.

Another important limitation of the current 72-hour fecal fat test is that finding steatorrhea does not indicate the mechanism(s) causing steatorrhea: (a) abnormal or reduced lipolysis caused by decreased exocrine pancreatic secretion of lipase (e.g., cystic fibrosis, chronic pancreatitis, or pancreatic cancer), (b) decreased intraluminal concentrations of bile acids leading to both maldigestion of dietary triglycerides and malabsorption of the products of lipolysis (e.g., decreased bile acid secretion or intraluminal precipitation of bile acids caused by a wide range of hepatobiliary/pancreatic disorders), and (c) malabsorption due to diseases of the small intestinal mucosa causing deceased absorption of the products of lipolysis (e.g., celiac disease).

And finally, the 72-hour fecal fat test becomes abnormal only when disturbances in the above-mentioned mechanisms result in extreme functional losses. For instance, loss of 90% to 95% of exocrine pancreatic capacity is necessary before fecal fat excretion becomes abnormal [4,5]. This is an important limitation since such a test is of no value in cases of mild derangements in lipolytic/absorptive functions and is of special concern in assessment of those with progressive disease, e.g., infants and children with CF who are likely to become pancreatic insufficient as the disease progresses. In the particular case of steatorrhea resulting from exocrine pancreatic insufficiency, the lack of sensitivity is related to the relative ease with which most dietary triglycerides (TGs) are hydrolyzed by the pancreatic lipase-colipase system. The majority of dietary TGs consist of "mixed TGs", i.e. TGs containing a combination of long-chain saturated and unsaturated as well as medium chain fatty acids [6,7]. Such "mixed TGs" present little steric hindrance to lipolytic enzymes and are readily hydrolyzed [6,7]. Previously published reports have indicated that increasing this steric hindrance by manipulating the TG structure leads to less efficient hydrolysis and increased fecal excretion [8]. Thus, it seems likely that fecal excretion of an engineered TG could be enhanced in relation to decreasing exocrine pancreatic function by exploiting this phenomenon, potentially leading to a test with increased sensitivity where appropriate.

We have previously shown that concurrent ingestion of a known amount of ¹³C-labeled triglyceride (TG*) and the nonabsorbable gastrointestinal marker DyCl₃ eliminates the need for control of dietary fat and complete collection of multiple stools [9], two of the important limitations of the 72-hour fecal fat test [1–3]. Here, we report results from a series of studies focusing on the effect of changes in TG structure on its subsequent fecal excretion by patients with steatorrhea related to pancreatic insufficiency. Specifically, we explored the relationship between fecal excretion of several engineered TG*s and fecal fat in CF patients and patients with chronic pancreatitis as affected by alterations in the chain length of the sn-2 fatty acid moiety.

	MATERIALS AND METHODS

TOP FOOTNOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS CONCLUSIONS ACKNOWLEDGMENTS REFERENCES

 
Preparation of Labeled Triglycerides
A series of ¹³C-labeled triglycerides were purchased from Cambridge Isotope Laboratories (Andover, MA) either from the company’s general inventory or prepared specifically for these projects. In each case, isotopic purity (>90%) and purity from contaminants were assessed both by the manufacturer and by an independent laboratory. The triglycerides utilized were: P*P*P* = TRIPALMITIN-1,1,1-¹³C₃; SO*S = 2-OCTANOYL-1,3-DISTEARIN-2-octanoyl-1,2-¹³C₂; and P*LP* = 2-LAURYL-1,3-DIPALMITIN-dipalmitoyl-1,1,2,2-¹³C₄.

Patients
Two groups of patients with expected reductions in exocrine pancreatic function participated in a series of studies examining the effect of alterations in TG* structure on the relationship between its fecal excretion and steatorrhea as assessed by the 72-hour fecal fat test. One group consisted of 26 children and young adult patients with CF (9 male and 17 female). All were currently receiving treatment through the Cystic Fibrosis Center at Cincinnati Children’s Hospital Medical Center (Cincinnati, OH). This group of CF patients had all been classified as pancreatic insufficient, and were currently receiving pancreatic enzyme replacement therapy. (One CF patient refused enzyme therapy as part of her normal care and was, thus, studied in the absence of enzymes. As a result, her stool fat was determined to be 126 g/day.)

At the time of study all were in a stable period of their disease relative to digestive function. Most studies were conducted on an outpatient basis with the entire stool collection carried out at home and while the patient consumed a self-selected diet. A few patients began their absorption study at the end of a hospital stay initiated for pulmonary exacerbation of their disease and the stool collections were then completed at home. The age range of the CF patients studied was 3 to 32 years at the time of study [mean = 15.3 ± 6.3 (SD)]. These 26 patients completed a total of 36 tests over a period of five years; 15 tests were carried out with P*P*P*, 10 tests with SO*S, and 11 tests with P*LP*.

A second group of volunteers consisted of 11 adult patients diagnosed as having chronic pancreatitis (5 male and 6 female). The diagnosis of chronic pancreatitis was made after a thorough evaluation to exclude other causes of chronic abdominal pain or steatorrhea. A diagnosis of chronic pancreatitis was established if a score of 4 or more was achieved using the following scoring system: 4, pancreatic calcification; 4, typical histological changes; 3, characteristic findings on endoscopic retrograde cholangiopancreatography; 2, pancreatic exocrine insufficiency (quantitative fecal fat >7 g/day or abnormal cholecystokinin test); 2, acute attacks of pancreatitis and/or chronic upper abdominal pain; and 1, diabetes mellitus. This scoring system has been described and validated previously [10,11]. All chronic pancreatitis patients were receiving treatment through the Pancreas Clinic at Mayo Clinic and Foundation (Rochester, MN). The presence of steatorrhea had not previously been established in all patients recruited. The pancreatitis patients ranged in age from 24 to 73 years with a mean age of 51.8. Studies carried out by chronic pancreatitis patients utilized P*LP* as the test TG* and were conducted at home after each received detailed oral and written instructions from the research nurse/coordinator. Patients received all necessary study materials from the research coordinator, and were in daily phone contact with the coordinator during the time of study. Stool samples were returned to the Pancreas Clinic by commercial courier.

All study protocols were approved by the Institutional Review Board of Cincinnati Children’s Hospital Medical Center or Mayo Clinic and Foundation as appropriate. All patients and/or legal guardians gave written consent and children over the age of 11 gave assent.

Clinical Protocols
All tests consisted of ingesting a single test meal containing the ¹³C-labeled triglyceride (TG*) of interest along with the nonabsorbable gastrointestinal marker DyCl₃ followed by complete stool collections. An initial study was performed in nine CF patients to investigate the relative gastrointestinal transit kinetics of TG* in comparison to DyCl₃. In this study, after an overnight fast or four hours after the previous meal, each patient consumed 0.70 g of P*P*P* mixed into 20 g peanut butter. The P*P*P* dose was fed as part of a test meal along with an oral dose of DyCl₃ solution (1.00 mg Dy). The composition of the test meal was varied according to patient’s preferences, but provided approximately 30% of the patient’s daily caloric needs and approximately 40% of calories as fat. Because many of the CF patients were children, we felt it was unethical to study their fat absorption in the absence of enzyme therapy for several days. Thus, the CF patients consumed their normal complement of therapeutic enzymes during the test meal and throughout the entire stool collection period. Quantitative collection of stools was initiated immediately after consumption of the test meal and continued for 120 hours.

All other studies were carried out in a manner similar to that described above with the following changes: 1) Chronic pancreatitis patients consumed the test meal and TG* in the absence of supplemental pancreatic enzymes and abstained from enzyme therapy throughout the stool collection period; 2) TG* doses were mixed manually with 10–20 g of reduced calorie butter substitute (I Can’t Believe Its Not Butter! Light, Lipton, Englewood Cliffs, NJ) instead of peanut butter; 3) Dose of TG* used was increased to 1.00 to 2.00 g when TG*s other than P*P*P* were ingested to insure sufficient ¹³C-enrichment of stool samples for good analytical precision. Increasing the TG* dose was necessary either because the TG* used contained fewer ¹³C-atoms/molecule of TG* (i.e., SO*S), or because the digestibility of the TG* was expected to be greater due to a reduction in the chain length of the fatty acid in the sn-2 position relative to P*P*P* (i.e., SO*S and P*LP*); 4) DyCl₃ dose was incorporated into a gelatin capsule which also contained 100–300 mg D-glucose and 50–60 mg brilliant blue; the Dy capsule was consumed midway through the test meal; 5) Seventy-two hours after consumption of the test meal, a second gelatin capsule was consumed containing only brilliant blue and D-glucose in order to mark the end of the 72-hour intake period. For all of these protocols, complete stool collections began immediately after consumption of the test meal and continued until fecal excretion of the second dose of brilliant blue was noted by the patient. Excretion of the brilliant blue visual marker was readily observed as it imparted a green color to at least a portion of one or more stools.

For all studies, stools were collected individually and frozen until prepared for analysis. For all but our initial study, designed to examine the relative gastrointestinal transit kinetics of P*P*P* and Dy, 72-hour stool composites were prepared from stools bracketed by the two doses of brilliant blue. An accurately weighed portion of each composite was dried at 100° C for 48 hours and then re-weighed. Each dried stool sample was then pulverized to form a fine powder; stools containing a large amount of unabsorbed dietary fat retained a pasty consistency after drying. Triplicate portions of each powdered stool composite were analyzed for total fat, ¹³C-Excess (¹³C*; ¹³C present in sample of interest in excess of natural abundance) and Dy.

Stool samples from our initial study were first dried and prepared for analysis as individual stools because of the design of the experiment; these individual stools were analyzed for ¹³C* and Dy only. At a later point, we created "72-hour composites" starting with the first stool collected after TG* dose ingestion by combining accurately weighed aliquots of dried stool powder from the individual stools; each aliquot represented a constant fraction of the initial wet stool weight. These individual aliquots were well mixed before sampling for composite analyses.

Analytical Methods and Calculations
Triplicate aliquots of dried stool composite (250 mg) were re-hydrated with 2 mL deionized water and four drops of concentrated HCl. Fat was then extracted according to the method of Jeejeebhoy et al. [12] and total fat determined gravimetrically [13]. This method was adopted so that total fat and ¹³C* could be determined from the same aliquot of dried stool. In addition, stool composites from the chronic pancreatitis patients were also sampled prior to drying and analyzed for total fat according to the method of van de Kamer [14] for comparison with the gravimetric method; the latter fat measurements were carried out by the Department of Laboratory Medicine at Mayo Clinic.

After weighing (for gravimetric determination of total fat), each sample of fat residue was re-dissolved in chloroform, quantitatively transferred to a screw-topped vial, and the chloroform evaporated to allow for storage and transport. Each so-prepared fat extract was re-dissolved in chloroform and analyzed for total carbon, and the ratio ¹³C/¹²C (atom% ¹³C) in triplicate using a Europa Scientific 20/20 isotope ratio mass spectrometer equipped with Automated Nitrogen Carbon Analyzer; these analyses were carried out by Metabolic Solutions Inc. (Nashua, NH) as described previously [9]. A set of fat extract standards was prepared for each TG* studied. These standards, whose isotopic contents covered the expected range, were prepared from unlabeled stool extract and spiked with known increments of TG*. An appropriate set of standards was analyzed with each batch of sample extracts. The content of TG* equivalents (mg) in each stool sample was then estimated using the regression equation for ¹³C* vs. TG*. The term TG* equivalents has been used to facilitate comparisons between the dose of TG* ingested and the excretion of ¹³C-label in all forms (i.e., undigested TG*, labeled free fatty acid, sn-1,2-diglyceride, etc.) as well as between TG*s containing differing amounts of ¹³C-label. Finally, excretion of ¹³C-label for each ingested TG* has been expressed as % of ingested oral dose; the dose ingested was individually corrected for any TG*/margarine mixture not removed from its storage container and, thus, not consumed.

The Dy content of dry powdered stool composite was determined by neutron activation analysis according to our previously published method [15]; neutron activation analyses were carried out by Dr. Steven Morris at the University of Missouri (Columbia, MO). Analytical performance data for both ¹³C* and Dy measurements in stool samples have been previously reported in detail [9,15]. We utilized the approach of administering a single dose of TG* simultaneously with the nonabsorbable fecal marker DyCl₃ as part of efforts to develop a practical approach to measurement of steatorrhea based on analysis of a single or small number of stools; these data have been reported in a separate manuscript [9]. Herein, Dy recovery data have only been used to check for the completeness of stool collection—at least during the time when Dy is excreted. A Dy recovery of >90% was considered to indicate a complete stool collection [16,17].

Statistical Methods
SPSS Version 10 was utilized for statistical analysis (SPSS, Inc., Chicago, IL). Regression analysis was used to examine the relationship between ¹³C* excretion and total stool fat. Regression parameters between treatment groups were compared using Analysis of Covariance; p-values of less than 0.05 were considered to indicate statistically significant differences. Welch’s t test was used to compare the age distribution of CF vs. chronic pancreatitis patients because the variances of the two groups were heterogeneous.

	RESULTS

TOP FOOTNOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS CONCLUSIONS ACKNOWLEDGMENTS REFERENCES

 
Stool Fat Analyses
Stool fat values determined for the same stool composites by direct gravimetric measurement of extracted lipids [13] and by the commonly used titrimetric method of van de Kamer [14] were compared using linear regression analysis. Stool fat values determined by the two methods were highly correlated (r² = 0.987, p < 0.01). The slope of the regression line comparing the two methods was 1.254, significantly greater than 1.0 (p < 0.05), and the y-intercept was 0.017 (not different from 0). A slope greater than 1.0 was expected because the gravimetric method measures all lipid components, whereas the titrimetric approach measures only lipids present as free fatty acids after saponification [18]. The observed magnitude of the difference between the two methods compares favorably with that reported in the literature (25% vs. 22%, respectively [18]).

We chose to utilize the gravimetric approach to the measurement of stool fat for several reasons: 1) It allowed us to measure stool fat and ¹³C-isotope enrichment in the same sample of stool extract; 2) It measures excretion of all lipid components and, thus, may correlate better with reduced pancreatic function in those patients with more severe steatorrhea [18]; and 3) The titrimetric procedure introduces additional unquantifiable error due to the loss of volatile fatty acids during saponification, and it requires use of an assumed mean molecular weight for stool fatty acids that is likely to vary among individuals [12,18].

TG* and Fat Excretion
The relationship between ¹³C-label and fecal fat excretion is shown in Fig. 1 for P*P*P*, P*LP*, and SO*S in those studies in which stool collections were complete (>90% excretion of Dy; 36 of 47 studies). These data illustrate a number of striking differences in this relationship as a function of the structure of the TG* fed. Fecal excretion of P*P*P* varied linearly with fecal fat (r² = 0.941, n = 8) over the range of fecal fat corresponding to mild steatorrhea (<25 g fecal fat/24 hours). However, P*P*P* was much less well digested/absorbed than dietary fat as evidenced by the slope of the regression line relating excretion of the two (slope = 3.20, significantly > 1.0; p < 0.05). Excretion of P*P*P* by patients with more severe steatorrhea (>25 g fecal fat/24 hours, n = 4) plateaued at a value of 68 ± 8.7% of dose [mean ± SD] and was, thus, no longer correlated with fecal fat (data not shown). In stark contrast to the behavior of P*P*P*, SO*S excretion was always a small percentage of ingested dose. Even patients excreting more than 50 g of fat/day excreted <10% of the ingested SO*S. While excretion of SO*S was relatively well correlated with fecal fat (r² = 0.843, n = 8), the slope of the regression line was only 0.120, some thirty times smaller than for P*P*P*. The results for P*LP* are intermediate between P*P*P* and SO*S. Excretion of P*LP* was highly correlated with fecal fat (r² = 0.924, n = 16) over a wide range of fecal fat values (from normal to 126 g fat/24 hours) and the slope of the regression line was 0.592.

View larger version (18K): [in this window] [in a new window]   Fig. 1. Correlation between P*P*P* (•), P*LP* (), and SO*S () excretion and stool fat for studies in which Dy excretion for the entire stool composite was >90%. All filled symbols represent tests carried out in CF patients whereas the open symbol () represents tests performed in patients with chronic pancreatitis. Excretion of all TG*s is expressed as percent of oral dose and all stool fat values as g/24 hours. Linear regression analysis was performed using SPSS Version 10 and the least squares fit line is shown for each TG*. The calculated regression parameters for each TG* are as follows: P*P*P* r2 = 0.941, slope = 3.20 and y-intercept = 9.04; P*LP* r2 = 0.924, slope 0.592 and y-intercept = 4.02; SO*S r2 = 0.845, slope = 0.120 and y-intercept = 0.900.

	DISCUSSION

Comparative interpretation of the excretion behavior of the labeled TGs used in these studies is somewhat complicated because of differences in the sn position of the ¹³C-labeled fatty acid on the glycerol molecule. Both P*P*P* and P*LP* contained ¹³C-label in the sn-1,3 positions, whereas SO*S was labeled only in the sn-2 position. Labeled palmitic acids released from the sn-1,3 positions may have been poorly absorbed, although the literature data in this regard are contradictory [19–22]. On the other hand, any labeled octanoyl monoglyceride or octanoic acid released during small intestinal lipolysis can be expected to have been well-absorbed [22,23]. In addition, recent analytical work indicates that little bacterial metabolism of undigested SO*S occurs in the colon of CF children [24]. Thus, significant bacterial metabolism of ¹³C-labeled octanoic acid is unlikely to be responsible for the observed differences in ¹³C-label excretion from SO*S vs. P*LP* and P*P*P*. Taken together, our data suggest that the efficiency of in vivo lipolysis can be markedly influenced by the judicious engineering of the TG* marker used.

We investigated the relationship between P*LP* excretion and fecal fat both in CF patients and patients with chronic pancreatitis. As a group, the CF patients were significantly younger than the pancreatitis patients (mean age 16.2 vs. 51.8 years, respectively; p < 0.05) because of the genetic, developmental nature of their disease [25]. While the underlying disease process that leads to exocrine pancreatic insufficiency is clearly different for the two patient populations, the primary resultant gastrointestinal defects that lead to malabsorption of dietary fat—lack of luminal pancreatic lipase/co-lipase activity, low small intestine luminal pH, and a reduced bile acid pool—are common to both disease states [3,25]. Thus, we chose to analyze the relationship between P*LP* excretion and stool fat for both patient populations taken together. As can be seen from our data, the degree of fat malabsorption varied considerably among the pancreatitis patients, who were all studied in the absence of exogenous enzyme therapy, based on both P*LP* excretion and 72-hour stool measurements. In fact, there was a high degree of linear correlation between P*LP* excretion and fecal fat for this group of patients alone (r² = 0.904, p < 0.01). In contrast, 10 of 11 CF patients who received P*LP* were studied while consuming their full complement of therapeutic enzymes. As a result, this portion of the data set covers only a narrow range of fecal fat values and by themselves reveal little about the correlation between P*LP* excretion and 72-hour stool fat. Importantly, however, considering the entire P*LP* data set (all CF and pancreatitis patients with complete stool collections), P*LP* and stool fat are highly linearly correlated (r² = 0.924, n = 16).

The studies reported here suffer from several important experimental constraints imposed in order to simulate conditions under which a practical TG* excretion test could be routinely carried out, and to encourage patient participation: (1) We did not require constant fat intake from meal to meal or day to day and permitted ad libitum food consumption; (2) We could not experimentally determine the completeness of stool collections for the entire 72-hour period. The latter affected exclusively fecal fat measurements as Dy could have been used to correct TG* excretion values. In the current context, Dy excretion was only used to eliminate incomplete collections (Dy excretion <90%) because the stool fat values could not be corrected. Using Dy to correct only the TG* value would not have made the data pair any more valid. (3) Absorption of a single bolus of TG* was compared to three-days worth of fat intake. Total fat intake likely varied both in amount and TG composition from meal to meal. In the CF patients, the enzyme dose per g fat also may have varied somewhat from meal to meal. These experimental limitations resulted in an undetermined amount of variability in our estimates of fecal fat excretion and introduced undetermined scatter in the correlation plots. For these reasons, we believe that the "true" correlation coefficients for excretion of TG* and fecal fat are better than those observed (Fig. 1).

In spite of these limitations, the observed linear correlation between P*LP* and fecal fat is strong (r² = 0.924, p < 0.01) and the range of linearity covers a wide range of fecal fat values. Thus, we believe that P*LP* may be a suitable surrogate for fecal fat for the purpose of assessing steatorrhea in patients with reduced pancreatic function. Obviously, before such a test could be used clinically, further development is needed. The sensitivity and reproducibility of P*LP* excretion to detect steatorrhea when fat excretion is only slightly elevated needs to be established. Importantly, the upper limits of normal P*LP* excretion in healthy control subjects needs to be determined. This would be useful both for determining when P*LP* excretion is truly elevated and represents steatorrhea as well as for potentially setting a "goal" for normalizing P*LP* excretion during enzyme therapy. Because TG* excretion, following its simultaneous ingestion with Dy and brilliant blue, can be accurately quantitated from the analysis of a sample of only two readily identifiable stools [9], development of a test to replace the 72-hour stool fat test using this approach is clearly possible. The clinical usefulness and acceptance of such a test remains to be determined.

Application of specifically engineered ¹³C-labeled TGs or labeled fatty acids, used in the manner described here, opens the possibility for in vivo evaluation of mechanisms of steatorrhea. For instance, use of a suitably labeled fatty acid may permit evaluation of intestinal fat malabsorption in celiac disease. Similarly, labeled fatty acids could be appropriately combined with labeled TGs in order to evaluate the significance of derangements in pancreatic lipolysis, bile acid disturbance, and intestinal malabsorption in patients for whom the defect leading to steatorrhea is unknown or multi-factorial.

Our work builds on that of earlier investigators who administered radio-labeled triolein (either ¹³¹I- or ¹⁴C-labeled) as a tracer for dietary lipid [16,17]. Both Pedersen [16] and Ditchburn et al. [17] found fecal excretion of radio-label to be significantly correlated with fecal fat in patients with steatorhea (r = 0.82 and 0.94, respectively), and to readily identify abnormal fat excretion in most patients. In addition, Pedersen and Halgreen [26] showed that by administering both ¹⁴C-labeled triolein and ³H-labeled oleic acid it was possible to differentiate malabsorption from maldigestion in most cases. Widespread application of their approach has been limited, in part, by the unacceptability of using radio-isotope labeled tracers in many circumstances. More recently, Murphy et al. [27] examined the absorption of ¹³C-labeled tripalmitin in both CF patients receiving enzyme therapy and healthy controls. These investigators found that the CF patients excreted significantly more ¹³C-label and total fat than controls, but found only a weak association between excretion of ¹³C-label and total lipid in the stool (r = 0.53; p > 0.09). The weak correlation observed by Murphy et al. [27] vs. our robust observation (r = 0.970) could be due the many differences in experimental methodologies employed, such as: dose of ¹³C-label, method of label incorporation into dietary fat, approach to measurement of ¹³C-excess, and method used to quantitate total stool fat. The fact that we were able to quantitate total fecal lipid and ¹³C-Excess in the same sample of stool extract may have reduced our analytical variability. In addition, we only examined the relationship between ¹³C-label excretion and stool fat in CF patients and did not include control subjects in our correlation.

	CONCLUSIONS

TOP FOOTNOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS CONCLUSIONS ACKNOWLEDGMENTS REFERENCES

 
The studies reported here and elsewhere [9] lead to a number of conclusions that may be of value in the clinical assessment of steatorrhea: (a) a noninvasive method utilizing (non-radioactive) stable isotopes can be developed to take advantage of all the important attributes of the 72-hour fecal fat test in the assessment of steatorrhea while eliminating the major experimental limitations related to the need for control of dietary fat and complete fecal collections for several days, (b) the nature of the labeled TG permits potential development of noninvasive methods with greater sensitivity to exocrine pancreatic insufficiency than is currently possible, and (c) this approach may permit assessment of the contributions of various mechanisms in patients with steatorrhea of unknown etiology or that is multi-factorial.

	ACKNOWLEDGMENTS

TOP FOOTNOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS CONCLUSIONS ACKNOWLEDGMENTS REFERENCES

 
The work reported here was supported, in part, by NIH/SBIR-STTR grants R44 DK48190 and R42 DK48537 and by M01-RR08084 from the NCRR-NIH. The authors would like to thank Mr. Rodney Sandburg and Ms. Tammy Dahl, R.N., for their important technical and clinical expertise in carrying-out the studies in chronic pancreatitis patients. In addition, the authors would like to thank Dr. Minu Patel, Assistant Professor of Biometrics at the University of Illinois (Chicago, IL), for his expert statistical advice.

	FOOTNOTES

TOP FOOTNOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS CONCLUSIONS ACKNOWLEDGMENTS REFERENCES

 
The procedure described herein is covered by U.S. Patent No. 6,006,754 held by BioChemAnalysis, Corp.

Received February 21, 2003. Accepted June 6, 2003.

	REFERENCES

TOP FOOTNOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS CONCLUSIONS ACKNOWLEDGMENTS REFERENCES

 

1. Westergaard H: Differential approach to steatorrhea. In Dietschy (ed): "Disorders of the Gastrointestinal Tract, Disorders of the Liver, Nutritional Disorders." New York: Grune & Stratton, pp76 –80,1976 .
2. Couper R: Pancreatic function tests. In Walker WA (ed): "Pediatric Gastrointestinal Disease." Ontario: BC Decker, pp1515 –1528,2000 .
3. Owyang C, Levvitt MD: Chronic pancreatitis. In Yamada T (ed) "Textbook of Gastroenterology." Philadelphia: JB Lipponcott Co, pp.2091 –2112,1995 .
4. DiMagno EP, Go VL, Summerskill WH: Relations between pancreatic enzyme outputs and malabsorption in severe pancreatic insufficiency. N Engl J Med288 :813 –815,1978 .
5. Regan PT, Malagelada JR, DiMagno EP, Glanzman SL, Go VL: Comparative effects of antacids, cimetidine and enteric coating on the therapeutic response to oral enzymes in severe pancreatic insufficiency. N Engl J Med297 :854 –858,1977 .[Abstract]
6. Small DM: The effects of glyceride structure on absorption and metabolism. Ann Rev Nutr11 :413 –434,1991 .[Medline]
7. Bracco U: Effect of triglyceride structure on fat absorption. Am J Clin Nutr60(Suppl) :1102S –1109S,1994 .
8. Hashim SA, Babayan VK: Studies in man of partially absorbed dietary fats. Am J Clin Nutr31(Suppl) :273S –276S,1978 .
9. Schuette SA, Janghorbani M, Cohen MB, Krug S, Schindler T, Wagner DA, Morris SJ: Dysprosium chloride as a nonabsorbable gastrointestinal marker for studies of stable isotope-labeled triglyceride excretion in man. J Am Coll Nutr22 :379 –387,2003 .[Abstract/Free Full Text]
10. 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. Gastroenterology107 :1481 –1487,1994 .[Medline]
11. Maringhini A, Nelson DK, Jones JD, DiMagno EP: Is the plasma amino acid consumption test an accurate test of exocrine pancreatic insufficiency? Gastroenterol106 :488 –493,1994 .[Medline]
12. Jeejeebhoy KN, Ahmad S, Kozak G: Determination of fecal fats containing both medium and long chain triglycerides and fatty acids. Clin Biochem3 :157 –163,1970 .[Medline]
13. Tietz NW: "Textbook of Clinical Chemistry." Philadelphia: WB Saunders,1986 .
14. van de Kamer JH, Huinick HTB, Weyers HA: Rapid method for the determination of fat in feces. J Biol Chem177 :347 –355,1949 .[Free Full Text]
15. Schuette SA, Janghorbani M, Young VR, Weaver CM: Dysprosium as a nonabsorbable marker for studies of mineral absorption with stable isotope tracers in human subjects. J Am Coll Nutr12 :307 –315,1993 .[Abstract]
16. Pedersen NT: Estimation of lipid assimilation from faecal samples using ¹⁴C-triolein as tracer and ⁵¹CrCl₃ as non-absorbable marker. Scan J Clin Lab Med43 :323 –328,1983 .
17. Ditchburn RK, Smith AH, Hayter CJ: Use of unabsorbed radioactive marker substances in a reassessment of the radioactive triolein test of fat absorption. J Clin Path24 :506 –514,1971 .[Abstract/Free Full Text]
18. Henry RJ, Cannon DC, Winkelman JW: "Clinical Chemistry; Principles and Techniques." New York: Harper & Row, pp.1475 –1483,1974 .
19. Kalivianakis M, Minich DM, Bijleveld CMA, van Aalderen WMC, Stellaard F, Laseur M, Vonk R, Verkade HJ: Fat malabsorption in cystic fibrosis patients receiving enzyme replacement therapy is due to impaired intestinal uptake of long-chain fatty acids. Am J Clin Nutr69 :127 –134,1999 .[Abstract/Free Full Text]
20. Watkins JB, Klein PD, Schoeller DA, Kirschner BS, Park R, Perman JA: Diagnosis and differentiation of fat malabsorption in children using ¹³C-labeled lipids: trioctanoin, triolein, and palmitic acid breath tests. Gastroenterology82 :911 –917,1982 .[Medline]
21. Murphy JL, Jones AE, Stolinski M, Wootton SA: Gastrointestinal handling of [1-¹³C] palmitic acid in healthy controls and patients with cystic fibrosis. Arch Dis Child76 :425 –427,1997 .[Abstract/Free Full Text]
22. Greenberger NJ, Skillman TG: Medium-chain triglycerides. N Engl J Med280 :1045 –1058,1969 .
23. Durie PR, Newth CJ, Forstner GG, Gall DG: Malabsorption of medium-chain triglycerides in infants with cystic fibrosis: Correction with pancreatic enzyme supplements. J Pediatr96 :862 –864,1980 .[Medline]
24. Slater C, Ling SC, Preston T, Weaver LT: Analysis of ¹³C-mixed triacylglycerol in stool by bulk (EA-IRMS) and compound specific (GC/MS) methods. Isotopes Environ Health Stud38 :79 –86,2002 .[Medline]
25. Gaskin K: Exocrine pancreatic dysfunction 1. Cystic fibrosis. In Walker WA (ed): "Pediatric Gastrointestinal Disease," Ontario: BC Decker, pp.1353 –1370,2000 .
26. Pedersen NT, Halgreen H: Simultaneous assessment of fat maldigestion and fat malabsorption by a double-isotope method using fecal radioactivity. Gastroenterology88 :47 –54,1985 .[Medline]
27. Murphy JL, Laicho KM, Jones AE, Wootton SA: Metabolic handling of ¹³C labelled tripalmitin in healthy controls and patients with cystic fibrosis. Arch Dis Child79 :44 –47,1998 .[Abstract/Free Full Text]

This Article Abstract Figures Only Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Schuette, S. A. Articles by DiMagno, E. P. Search for Related Content PubMed PubMed Citation Articles by Schuette, S. A. Articles by DiMagno, E. P.