- Email: [email protected]
Simultaneous Assessment of Fat Maldigestion and Fat Malabsorption by a Double-Isotope Method Using Fecal Radioactivity
GASTROENTEROLOGY 1985;88:47-54
Simultaneous Assessment of Fat Maldigestion and Fat Malabsorption by a Double-Isotope Method Using Fecal Radioactivity NIELS THORSGAARD PEDERSEN and HANNE HALGREEN Medical Department F and Department of Clinical Chemistry. Glostrup Hospital. Copenhagen. Denmark
F4C]Triolein and [3H]01eic acid are useful tracers of dietary triglycerides and free fatty acids. The combustion/liquid scintillation counting technique was found a practicable method of estimating the fecal activity of these tracers. When ingested with the nonabsorbable marker 51CrCP, the fecal excretion of 14C and 3H could be estimated accurately from samples of two stools. This technique was investigated as a test of lipid assimilation in a prospective, blind study of 84 consecutive patients suspected of malassimilation. The fecal excretion of 14C was a useful alternative to fecal fat: the patients with "normal lipid assimilation" excreted :::::10.4% of the dose ingested, whereas 25 of the 26 patients with steatorrhea ("unequivocal malassimilation") excreted >10.4%. Fecal fat was normal in 18 patients who exhibited other signs of malassimilation ("equivocal malassimilation"); in 5 of these patients fecal excretion of 14C was >10.4%. Lastly, there was a significant correlation between fecal 14 C and fecal fat (r = 0.82, P < 0.001). The fecal 14CPH ratio was found useful as an index of pancreatic digestive function, being <1.3 in patients with "normal" lipid digestion and ~1.3 in 18 of the 19 patients with pancreatic steatorrhea and in 19 of the 27 patients with severely reduced exocrine pancreatic function. The test reReceived December 2. 1983. Accepted July 11. 1984. Address requests for reprints to: Niels Thorsgaard Pedersen. M.D .. Medical Department 2. Kommunehospitalet. Oester Farimagsgade 5. DK-1399 Copenhagen K. Denmark. This work was supported by grants from the Danish Medical Research Council (J.No. 512-21015 and J.No. 12-1753) and from the Danish Hospital Foundation for Medical Research. Region of Copenhagen. the Faroe Islands. and Greenland (J.No. 46/80). The authors thank Mrs. Grethe Hansen and Mrs. Ingrid Emanuel of the Department of Clinical Physiology and Mrs. Lene Appelt and Mrs. Mariann Hansen of Medical Department F for their technical assistance. They also thank Mrs. Marianne Iversen for dietetic assistance. © 1985 by the American Gastroenterological Association 0016-5085/85/$3.30
suIt was found widely independent either of the amount of carrier fat or the quality of the fecal collections and was therefore useful as an outpatient test. As it provided information about the pathogenesis of malassimilation and was at least as sensitive and specific as a fat "balance study," the test may be a useful alternative to fecal fat measurement.
When estimating lipid assimilation from measurements of the fecal fat excretion. the perpetually taken dietary lipids serve as test substance. The assimilation of a definite quantity of lipids can therefore not be established. Instead, a 3-day fat "balance study" is done. Feces cannot be collected for examination of the test diet residues until the elapse of an ingestion period of at least the gastrointestinal transit time, and then fluctuations in dietary intake and colonic function and incomplete fecal collections may cause uncontrollable error (1). Nonabsorbable gastrointestinal markers may be used to show the excretion of the residues of the test diet and to some extent to monitor the completeness of the fecal collections (1,2), but do not permit correction for the above errors (2,3). At least one-fifth of the fat "balance studies" will therefore be unsuccessful (1). These problems could be avoided by estimating the assimilation of a synthetic lipid. [14C]Triolein has been reported to be useful as a tracer of dietary triglycerides (4,5). When ingested with the no nabsorbable marker 51CrCb, the [14 Cjtriolein excretion can be estimated from fecal samples (4). The excretion of [3H]oleic acid. a tracer of free fatty acids, can be estimated in the same way (5). Triglycerides make up the major part of the dietary lipids. The fecal excretion of [14C]triolein may therefore be a useful alternative to fecal fat (4). The assimilation of [14C]triolein includes digestion, mainly catalyzed by pancreatic lipase, and absorp-
48
GASTROENTEROLOGY Vol. 88, No.1, Part 1
THORSGAARD PEDERSEN AND HALGREEN
tion, whereas [3HJoleic acid is absorbed directly. The ratio between the excretion of 14C and 3H may therefore reflect pancreatic digestive function (5). Thus estimation of the excretion of simultaneously ingested [14CJtriolein and [3H]0Ieic acid appears useful not only as a test of lipid assimilation, but also as a differential test of mal digestion and malabsorption.
Methods Patients Eighty-four patients admitted consecutively over a 2-yr period with symptoms suggesting malassimilation entered in the study. Informed, written consent was obtained and the study was approved by the Medical Ethical Committee of Copenhagen County.
Clinical Procedures The patients were examined in the daily routine work without knowledge of the results of the present investigation. Whenever possible, the diagnosis was established on morphologic criteria, e.g., x-ray, ultrasonography, endoscopy, endoscopic retrograde cholangiopancreatography, and biopsy. The function of the diseased organs was then evaluated by means of the following specific function tests: Exocrine pancreatic function was estimated from the Lundh test by measuring the meal-stimulated duodenal lipase concentration (6). the normal value being >240 kU/L. The 25-g D-xylose test was used to estimate jejunal absorptive surface area; a 1-h plasma xylose concentration >2.2 mmoliL was considered normal (7). Micellar solubilization was estimated from the mealstimulated duodenal bile acid concentration and a bile acid concentration >5.4 mmol/L was considered normal (8).
Finally, the Schilling II test was used to estimate distal ileal function and thereby indirectly the enterohepatic circulation and the micellar solubilization. A 24-h urinary excretion >10% of ingested [58 Co]B 12 -intrinsic factor was considered normal (9). Fecal fat was estimated as the average excretion from the pooled 3-day collection (10). <7 g/day being normal (1).
Classification of Patients Chronic pancreatitis was diagnosed when the Lundh test was abnormal in a patient with at least one of the following findings: previous episodes of acute pancreatitis, pancreatic calcifications, abnormal ultrasonography, and typical endoscopic retrograde cholangiopancreatography findings. Lipolysis was considered reduced when the Lundh test demonstrated a meal-stimulated duodenal lipase concentration <50 kU/L (Le., <20% of normal) (11,12). When, additionally, the fecal fat was >7 g/day, the patient was classified as having unequivocal
mal digestion (Table 1). Those patients who despite a very low duodenal lipase concentration excreted <7 g fat/day were classified as having equivocal mal digestion (Table 1). One patient with a severely reduced duodenal lipase concentration and steatorrhea, due to pancreatic cancer, was classified as having unequivocal mal digestion. Two patients suffering from diarrhea and loss of weight after gastric resection were classified as having equivocal maldigestion (Table 1); they did not fulfill the above-mentioned criteria for chronic pancreatitis, but showed severely reduced meal-stimulated intestinal lipase activity and excreted 6-7 g fat/day. Eight patients with chronic pancreatitis as just defined exhibited only a moderately reduced exocrine pancreatic function, i.e., a duodenal lipase concentration between 240 kU/L and 50 kU/L. These patients excreted <7 g fat/day and were consequently regarded as having normal lipid assimilation (Table 1). Seven patients with various well-established diseases of the small intestine (Crohn's disease of the ileum and distal jejunum, intestinal resection > 1 00 cm, celiac disease, pseudoobstruction, and bacterial overgrowth because of intestinal involvement of systemic sclerosis) and 1 patient with a biliocolonic fistula were classified as having unequivocal malabsorption because of one or more abnormal specific function tests, i.e., xylose test, Schilling test, and duodenal bile acid concentration test, and steatorrhea (Table 1). Another 8 patients had manifest small intestinal disease (Crohn's disease of the ileum, intestinal diverticulosis with bacterial overgrowth, and celiac disease) and abnormal specific function tests, but none had steatorrhea at examination and were therefore classified as having Table 1. Clinical Diagnoses and Clinical Findings Pancreatic disease Chronic pancreatitis Pancreatic cancer Intestinal disease Crohn's disease Intestinal resection Stagnant loop syndrome Lactase deficiency Celiac disease Systemic sclerosis Intestinal ischemia Gastric disease Sequelae after gastric resection Gastric ulcer Gastric cancer Miscellaneous VVatery diarrhea Irritable bowel syndrome Biliary disease Chronic alchoholism Infectious gastroenteritis Totals
No.
nla
33 32
8
ema
uma
emd
umd
6
18
1 18 5 3
1 1
4 2 2
3 4
2
1
1 1 10
1
6
3
3
1
3 1
23 9
9
1
1 84
40
10
2 5
1 1
1
5 1 5 1
6
1 3
2
1
7
8
19
ema, equivocal malabsorption; emd, equivocal maldigestion; nla, "normal" lipid assimilation; uma, unequivocal malabsorption; umd, unequivocal maldigestion.
January 1985
equivocal malabsorption (Table 1). Two other patients were classified ai'! having equivocal malabsorption: 1 patient with diarrhea after gastric resection and a fat excretion of 11 g/day, but with normal specific function tests and intestinal cultures, and 1 patient with clinically classified irritable bowel syndrome, whose fecal excretion of 8 g fat/day was the only abnormal finding. The specific function tests and the fecal fat measurement showed no evidence of malassimilation in 40 patients, including the 8 patients with moderate chronic pancreatitis. They were thus classified as having normal lipid assimilation (Table 1).
FECAL ["4C]TRIOLEIN/[3H]OLEIC ACID TEST
49
out problems of crossover counting and with constant quenching and counting efficiencies (5). According to the findings in the previous studies (4,5), the fecal excretion of 14C and 3H was estimated by extrapolating the activities of 14C, 3H, and 51Cr in the two stools to 100% recovery of 51Cr.
Reproducibility In 16 consecutive patients, the 14C and 3H excretion was estimated twice under the conditions described above within an interval of 14 days. The first of these measurements was used as the test result in the study.
Investigational Procedures Estimation of the fecal excretion of 14C and 3H was done together with measurement of fecal fat and without knowledge of the clinical findings. Details and specifications of the techniques used have been given previously (4,5). Briefly, the patients were served a fat- and caloriefixed diet of 100 g fat and 2100 kcal (8786 kJ) per day from the morning of the day before the ingestion of the isotopes until lunch after the test meal. The breakfast supplied 42 g fat and 900 kcal (3765 kJ). After 1 day on this diet the patient ingested with the breakfast 10 p.Ci of glycerol-tri[1- 14C]01eate ([14C]triolein), 20 p.Ci of [9,10(n)-3H]01eate ([3H]01eic acid), and 20 p.CJ. of 51CrCh, [14C]Triolein and [3H]01eic acid were stored and handled in 100 mg of "cold" triolein on lump sugar; 51CrC13 was dissolved in 100 ml of water. All the isotopes were supplied by the Radiochemical Centre, Amersham, U.K. The fecal collections were started immediately after breakfast. Each stool was collected separately in a sealed plastic bag. After 3 days, the stools were checked for 51Cr by means of a portable y-detector and if 51Cr was present the fecal collections were stopped (78 patients), otherwise the collections continued with daily checks until 51Cr was passed. The two stools with the highest activity of 51Cr were identified and used in the following procedures: the activity of 51Cr was estimated by counting the stools in a whole-body ,},-scintillation counter against weighed standards of the batch ingested. After homogenization (Colworth Stomacher 400, UAC House, London, U.K.), duplicate aliquots of 300 mg were taken from each stool for estimation of 14C and 3H. The aliquots were oxidized at 700°C in oxygen-enriched air in a Tri-Carb Oxidizer (Packard Instrument Co., Inc., United Technologies, Downers Grove, Ill.). In this apparatus the 14C-labeled materials were oxidized into 14COZ and automatically trapped into a scintillation liquid mixture, specifically absorbing COz (Carbosorb and Permaflour V, Packard), in one counting vial with an efficiency of 95% and a coefficient of variation of 3.3% (n = 17) (4). The 3H-labeled materials were oxidized into 3HzO and trapped into a scintillation liquid, specifically absorbing HzO (Monophase 40, Packard), in another counting vial with an efficiency of 98% and a coefficient of variation of 15% (n = 17) (5). All materials other than COz and HzO, including 51Cr, were discharged into a special reservoir. In this way, 14C, 3H, and 51Cr were physically separated with an efficiency >99.5% (5), which enabled liquid scintillation counting of 14C and 3H with-
Influence of Carrier Fat For 14 days, 17 consecutive patients were randomly served either the diet described above or an isocaloric low-fat diet (30 g fat/day and 8786 kJ). The breakfast of this low-fat diet supplied 6 g fat and 3765 kJ. The isotopes were ingested at breakfast on the second day. The test diets were continued until lunch. Fecal collections and measurements were done as described above. The findings on the high-fat diet were used as the test result.
Radiation Dose The radiation dose from [14C]triolein was <20 mrem/p.Ci (13) and that from [3H]01eic acid was <2 mremlp.Ci (13); the whole-body burden from 51CrC13 was <0.2 mreml p.Ci, giving a total whole-body radiation dose of <244 mrem per test.
Statistics The log values of fecal fat and fecal 14C were normally distributed. The level of demarcation between "normal" and abnormal fecal 14C was calculated as the mean + 2 SD of the log values of fecal 14C in the patients with "normal lipid assimilation," whereas the highest fecal 14Cf3H ratio in the patients with normal lipid digestion (Le., normal lipid assimilation and equivocal or unequivocal malabsorption) was taken as the level of demarcation between a normal and an abnormal fecal 14Cf3H ratio. The linear regression analysis of the log values was used to estimate the correlation between fecal 14C and fecal fat and the reproducibility of fecal 14C. In all other analyses nonparametric statistics were used (Mann-Whitney test, Wilcoxon's test of paired differences). Five percent significance limits were used in all tests.
Results Fecal
14
C
The mean fecal 14C excretion in the patients with normal lipid assimilation was 2.5% of the dose
50
THORSGAARD PEDERSEN AND HALGREEN
% of dose
GASTROENTEROLOGY Vol. 88. No.1. Part 1
••
Fecal1"c
100
o 50
• • ••
8 o
o
10
0
2 SO
o
o
o
o o
• •
o
••
If
o
0
1
5
2.5
Figure 1. The fecal HC excretion in the patient groups. Mean ± 2 SD of fecal 14C in the patients with "normal lipid assimilation" is indicated on the ordinate. O. patients with fecal 14C/3H ratio <1.3; •. patients in whom the ratio was 21.3.
8
o o
~
ct
X
o o
~
00 0
ego 0 0
2S0
0.5
o
NORMAL LIPID ASSIMILATION n='O
EQUIVOCAL UNEQUIVOCAL MALABSORPTION n=10 n=7
EQUIVOCAL UNEQUIVOCAL MAlOIGESTION n=8 n=19
ingested, with 10.4% as the upper limit of normal fecal 14C excretion (Figure 1). The fecal 14C excretion correlated significantly with fecal fat (r = 0.82, P < 0.001) (Figure 2) and was significantly higher in both the patients with unequivocal malabsorption/mal digestion and those with with equivocal malabsorption or equivocal maldigestion than in the patients with normal lipid assimilation (p < 0.001 and p = 0.05, respectively) (Figure 1). g/day
Fecal 3 H The patients with normal lipid assimilation excreted up to 15.7% of the ingested dose of [3H]0Ieic acid (Table 2). The patients with unequivocal malabsorption and those with unequivocal maldigestion excreted significantly more 3H (p < 0.01 for both). Thus, 6 of the 7 patients with unequivocal malabsorption and 4 of the 19 with unequivocal mal digestion excreted >15.7% of 3H, the excretion
Fecal fat
100
• •
50
20
10
7
2
1
•
1
2
5
• • •• • •
,
•
•
• •
\
• •
••
Figure 2 . The correlation between the fecal 14C excretion and fecal fat. The marked parts of the axes indicate the normal v
•
•
• ••
.,.....
5
.
•
• • • • • • •• • • •• •••• • • • • •• .a· • • ••
•
•
• •
•
Fecal 14C
7
10
20
50
100 "10 of dose
January 1985
FECAL [14 C]TRIOLEIN/[3H]OLEIC ACID TEST
Table 2. Fecal 3H Excretion and the Fecal 14Cf3H Ratio
Normal lipid assimilation Equivocal malabsorption Unequivocal malabsorption Equivocal maldigestion Unequivocal mal digestion
No.
Fecal 3H median (range)
40
5.0% (1.3 %-15.7%)
0.6 (0.3-1.3)
10
6.0% (3 .0%-27.6% )
0.9 (0.2-1.2)
7
35.4% (12 .0%-46.7% )
1.2 (0.5-1.3)
8
5 .0% (3.0%-10.9°,{,)
0.6 (0.3-1.4)
19
11.8% (4.0%-41.2%)
3.2 (0.9-15.5)
Fecal 14Cf3 H median (range)
being significantly higher in the patients with unequivocal malabsorption than in those with unequivocal maldigestion (p < 0.01) (Table 2). No difference in fecal 3H was found between the patients with normal lipid assimilation and the equivocal groups (Table 2). Fecal 14CPH Ratio
The fecal 14C/3 H ratio was <1.3 in all of the patients with normal lipid assimilation and equivocalor unequivocal malabsorption (Table 2), and was therefore considered to indicate normal lipid digestion. The fecal 14CPH ratio was >1.3 in 18 of the 19 patients with unequivocal mal digestion, this difference being significant (p < 0.001), whereas there was no difference between the patients with equivocal
Fecal 14C PH
,
maldigestion and those with normal lipid assimilation. Figure 3 depicts the relation between the duodenal lipase activity and the fecal 14CPH ratio, and shows that a fecal 14CPH ratio > 1.3 only appeared in patients with severely reduced exocrine pancreatic function. Simultaneous Estimation and Lipid Absorption
of Lipid
Digestion
The results of the fecal estimations of 14C and 14CPH were then combined. Fecal 14C ::5 10.4% of the ingested dose was considered as the test result normal lipid assimilation; fecal 14C > 10.4% of the dose and fecal 14CPH < 1.3 were considered to indicate malabsorption; and fecal 14C > 10.4% of the dose and fecal 14CPH ::::: 1.3 were considered as the test result mal digestion. As there are three possible test results, the sensitivity and specificity of each of these were calculated, both for all of the patients in the study and for the patients in whom lipid assimilation could be established definitely [Le., the patients with normal lipid assimilation and those with steatorrhea (the unequivocal groups)] (Table 3).
Reproducibility The results of the duplicate estimations of fecal 14C correlated significantly, the correlation
1
115,S)
10 •
8
Figure 3. The relation between the fecal 14C/3H ratio and the meal-stimulated duodenal lipase concentration in the 76 patients in whom the Lundh test was done. Normal lipase conc entration was >240 kU/L.
51
•
6
4
3 2 1.3 . . . . _
•
.
• • .... :c• ....• '.
• • •• • .::~ •• •
240
500
50
••
1
••
• KU/I
• •
DUODENAL LIPASE CONCENTRATON
52
THORSGAARD PEDERSEN AND HALGREEN
GASTROENTEROLOGY VoL 88. No. 1 , Part 1
Table 3 . Sensitivity and Specificity of the Fecal [14C]Triolein![3HJOl ei c Acid Test Test result Normal lipid assimilation Malabsorption Maldigestion
Sensitivity 30/31 (25 /26 65/72 (58/58 57/65 (47/48
= 0 ,98 = 0.96)
= 0 .90 = 1.00) = 0 ,88 = 0 .98)
Specificity 39/53 (39/40 10/12 (7/8 19/19 (18/18
= = = = = =
0.74 0.98) 0.83 0 ,88) 1.00 1 ,00)
The "diagnostic sensitivity" of a test result was calculated as th e number of patients in whom the test correctly excluded a specific diagnosis divided by the number of patients in whom the test excluded that specific diagnosis , The diagnostic specificity was calculated as the number of patients in whom the test correctl y indicated a specific diagnosis divided by the number of patients in whom the test indicated that specific diagnosis, The results from the patients with "normal li pid assimilation" and the unequivocal groups are shown in the parentheses,
coefficient being 0.81 and the slope of the regression line being 0.99. Qualitatively, 15 of the 16 duplicate estimations of fecal 14C and fecal 14Cf3H gave the same result; in 1 patient the test result changed from mal digestion to malabsorption. .
Influence of Carrier Fat The fecal excretion of 14C was significantly higher on the high-fat diet than on the low-fat diet (median 5.1%, range 0.6%-132.7% vs. median 4.9%, range 0.3%-91.7%) (p < 0.05), whereas there was no difference between the fecal 14CPH ratio of the highfat and the low-fat diet. Qualitatively, all of the duplicate measurements gave the same test result.
Discussion Estimation of the assimilation of acutely ingested isotope-labeled lipids is a logical alternative to measurement of fecal fat and several radiolipid assimilation studies have been published (4,5,1424). The easily measurable y-emitting 131I-labeled triglycerides were used initially (14,15), but later abandoned because of instability (15,16). In recent years, the stable 14C_ and 3H-labeled fatty acids have been used as tracers, mostly in breath tests (17-21) , but also in tests based on measurement of postprandial serum radioactivities (12,23). The most direct way of estimating the assimilation of one dose of [14C]triolein and [3H]oleic acid is by calculating the difference between the dose ingested and the dose excreted. Hitherto, this has been made difficult by the technical problems of estimating the activity of the l3-emitting 14C and 3H in feces: feces contain lipophilic dyes (e.g. , bilirubin), which make
the scintillation liquid cloudy (quenched) even after lipid extraction, thus resulting in poor counting efficiencies and crossover counting (4,5,24). Further, not more than two isotopes may be present in the sample. The combustion/liquid scintillation counting technique, previously used in studies of bile acid absorption (9), solved these problems and was also found practicable (4,5). The excretion of 14C and 3H is easily estimated when the two are ingested together with a nonabsorbable marker (4,5 ,14,24). No ideal lipid phase marker exists (3,25). Investigations have been done with insoluble chromic salts as a repeatedly ingested marker of dietary lipids (26) and as an acutely ingested marker of isotope-labeled bile acids (9). These studies found that the transit of the lipids was faster than that of the marker, which made the marker-corrected data useless. However, our previous studies did not disclose any systematic difference in the transit of 51CrCl3 and the radiolipids (4 ,5) . Like other investigators (14) , we concluded that 51CrC13, though not ideal, was useful as a marker of gastrointestinal transit of [14C]triolein -and [3f1]oleic acid (4,5): the coefficient of variation of the ratio between 14C and 51Cr and between 3H and 51Cr was 31% in all stools passed (23 patients), but when two stools were used to estimate the fecal 14C and 3H excretions, these estimates correlated highly with the total excretions measured (r = 0.99), were reproducible (coefficient of variation = 6%), and were not further improved by taking more stools (4). But even then the estimated excretion may deviate from the true one (e.g., up to 133% of the ingested dose), although this was mostly seen in patients with massive steatorrhea in whom less than one-thousandth of the stool was oxidized. In this study, we used the two stools with the highest activities of 51Cr to estimate the fecal excretion of 14C and 3H, but any of the stools could have been used for the estimation, so long as they contained sufficient radioactivity for accurate measurement, i.e., >5% of the dose ingested. Thus, the coefficient of variation of estimating fecal 14C from any two stools was 16%, and that of 3H was 18% (17 patients). This means that the result of the technique was less dependent on the quality of the fecal collections than is the fecal fat measurement, and explains the success of the test in all the patients studied. In addition, it was only necessary to collect feces until some of the radioactivity was passed, and for this a y-detector or a visible marker could be used. In this study the needed collection period was 18-124 h (median 38 h). Finally, the fecal excretion of 14C and 3H, like the percentage of dietary fat excreted (27), was only slightly affected by the amount of fat ingested. The
January 1985
result of the test was thus only moderately dependent on patient compliance and the test could therefore be used in outpatients. In fact, the reproducibility was estimated in outpatients and was still found as good as the reproducibility of the most carefully conducted "fat balance" study (1). Because triglycerides constitute the major part of the dietary lipids, we expected the fecal excretion of the triglyceride tracer, [14C]triolein, to reflect fecal fat. This expectation was fulfilled, the correlation between fecal fat and fecal 14C being highly significant (Figure 2). The excretion of [3H]0Ieic acid, which is a tracer of the digestive product of triolein, was expected to be equal to the excretion of [14C]triolein in patients with normal lipid assimilation and malabsorption, i.e., the fecal 14CJ3H ratio was expected to be ~1. In the patients with maldigestion, the fecal excretion of 3H was expected to be normal, but this was not the case in 4 of the patients. One reason for this may be reduced micellar solubilization of the lipolytic products-and [3H]0Ieic acid-due to precipitation of the bile acids because of a more acidic milieu in patients with pancreatic insufficiency (28). Another reason may be that jejunal absorption was reduced by the high concentration of lipids in the jejunum (29). The fecal excretion of 14C, however, exceeded fecal 3H, i.e., the fecal 14CJ3H ratio was >1.3, reflecting digestive function (Figure 3) . In fact the relationship between the fecal 14CJ3H ratio and the meal-stimulated duodenal lipase concentration (Figure 3) exactly mimicked that between fecal fat and duodenal lipase activity (11,12). An exact estimation of lipid assimilation is not always possible in a group of consecutive patients. In our study, the examinations and investigational procedures were done once within 1 mo, after which treatment was often begun, and this made reassessment of the assimilation at the time of investigation impossible. The clinical findings conflicted in 18 of the 84 patients. These patients were therefore classified as having equivocal malabsorption or equivocal maldigestion and constitute a subgroup of special interest when investigating a new test of lipid assimilation. All except 1 patient in the equivocal groups suffered manifest disease in the upper gastrointestinal tract compatible with malassimilation, but fecal fat was normal in 16 of the 18 patients and only insignificantly increased in the last 2 patients. However, taken as a group, the equivocal patients seemed less diseased than the unequivocal patients. Chronic pancreatitis was advanced both in the patients with equivocal and unequivocal maldigestion. The exocrine pancreatic function, as estimated from the Lundh test was, however, more reduced in the patients with unequivocal mal digestion (median 7.8 kUlL, range 0.2-20 kU/L) than in the patients with
FECAL ['4C]TRIOLEIN/[3H]OLEIC ACID TEST
53
equivocal maldigestion (median 20 kUlL, range 2.042 kUlL) (p < 0.001). Differences in colipase secretion (not measured) may explain why some patients had steatorrhea, whereas others, with equally low lipase concentrations, seemingly had normal fecal fat excretion (12). Another reason may be that the measurement of fecal fat was distorted by incomplete fecal collections and constipation in some patients in the equivocal groups. In 5 of these 18 patients, the fecal [14C]triolein/[3H]0Ieic acid test indicated malassimilation in agreement with the specific function tests (Figure 1). In this investigation, only patients with gastrointestinal symptoms were studied, no healthy controls were included. No conclusion can thus be made with regard to normal 14C and 3H excretion. From a clinical point of view it is, however, more interesting to see whether the test can differentiate between patients with diarrhea caused by malassimilation and those with diarrhea from other causes. When looking at the patients in whom lipid assimilation could be established, i.e., normal lipid assimilation and the unequivocal groups, only one false-positive and one false-negative test result were found, and the test discriminated correctly between malabsorption and mal digestion in all except 1 patient in the unequivocal groups (Figure 1). The fecal [14C]triolein/[3H]0Ieic acid test thus seemed specific and more sensitive than the traditional fecal fat measurement. Great efforts have been made to elaborate a nonfecal test of lipid assimilation, e.g., the breath test (1721) and measurement of the postprandial serum radioactivity of 14C and 3H (22,23). These tests are more simple than the one described here, but the results are sometimes distorted by kinetics, unrelated to the assimilation (17,21). The breath test has mostly been studied in patients with obvious steatorrhea (corresponding to our unequivocal groups) and normal subjects, and the results are conflicting (1722). None of these studies showed the breath test to be more sensitive or specific than our test (17-21). Furthermore, the breath test provides no quantitative information (17,18,21)' and it necessitates two tests on alternate days to discriminate between malabsorption and maldigestion (20). Measurement of the postprandial serum radioactivity of 14C and 3H is not as sensitive as our test (22,23). Until more substantial investigations of the alternative methods are published, we believe that fecal collections and measurements are unavoidable. The fecal [14C]triolein/[3H]0Ieic acid test often needs shorter fecal collection time and less fecal handling than the measurement of fecal fat and is therefore less unpleasant. In all the radiolipid assimilation tests the isotopes
54
THORSGAARD PEDERSEN AND HALGREEN
were served at breakfast after a fast of several hours. Whether or not the sensitivity of the test would improve if the isotopes were served later in the day after some fatty meals has not been clarified. The radiation dose of 14C and 3H is modest, but nevertheless their long half-lives raise some concern in using these isotopes in humans, especially in children, in whom estimation of lipid assimilation from samples of feces would be particularly useful. However, 14C and 3H could possibly be replaced by the nonradioactive isotopes 13C and 2H, which are measurable by mass spectrometry after oxidation of feces (30). Expensive equipment was used in this investigation. A portable y-counter could replace the wholebody y-scintillation counter (data not shown), but the oxidizer is essential (price ~$25,000). [14C]Triolein and [3H]oleic acid are stable, both before ingestion and in feces (4,5), and can be transported for measurement at specialized laboratories. The costs of the isotopes, counting liquids, and salaries amount to ~$30 per test. On the other hand, the test can be done on outpatients and is at least as reliable as the inpatient "fat balance" study. Because it provides information about the pathogenesis of malassimilation, the fecal [14 C]triolein![3H]oleic acid test seems profitable both from a clinical and an economical point of view.
References 1. Thorsgaard Pedersen N, Halgreen H. Faecal fat and faecal weight. Reproducibility and diagnostic efficiency of various regimens. Scand J GastroenteroI1984;19:350-4. 2. Davignon I. Simmonds WI. Ahrens EH. Usefulness of chromic oxide as an internal standard for balance studies in formulafed patients and for assessment of colonic function. J Clin Invest 1968;47:127-38. 3. Fordtran JS. Marker perfusion techniques for measuring intestinal absorption in man. Gastroenterology 1966;51:1089-93. 4. Thorsgaard Pedersen N. Estimation of lipid assimilation from faecal samples using I4C-triolein as tracer and 5ICrCI3 as nonabsorbable marker. Scand J Clin Lab Invest 1983;43:3238.
5. Thorsgaard Pedersen N. Estimation of assimilation of simultaneously ingested I4C-triolein and 3H-oleic acid as a test of pancreatic digestive function. Scand J Gastroenterol 1984; 19:161-9. 6. Worning H, Miillertz S. pH and enzymes in the human duodenum during digestion of a standard meal. Scand J GastroenteroI1966;1:268-83. 7. Beck IT, Rona S, McKenna RD, Kahn DS. Evaluation of 200 D( + )-xylose blood level time curves as index of intestinal function. Am J Dig Dis 1962;7:936-48. 8. Westergaard H. Duodenal bile acid concentration in fat malabsorption syndromes. Scand J Gastroenterol 1977;12:11522. 9. Fromm H, Thomas pJ, Hofmann AF. Sensitivity and specificity in tests of distal ileal function: prospective comparison of bile acid and vitamin B12 absorption in ileal resection patients. Gastroenterology 1973;64:1077-90.
GASTROENTEROLOGY Vol. 88, No.1, Part 1
10. Van de Kamer JH, Ten Bokkel Huinink H, Weijers HA. Rapid method for determination of fat in faeces. J BioI Chern 1949;177:347-55. 11. DiMagno EP, Go VLW, Summers kill WHJ. Relations between pancreatic enzyme outputs and malabsorption in severe pancreatic insufficiency. N Engl J Med 1973;288:813-5. 12. Gaskin KJ, Durie PR, Lee L, Hill R, Forstner GG. Colipase and lipase secretion in childhood-onset pancreatic insufficiency. Gastroenterology 1984;86:1-7. 13. Thorsgaard Pedersen N, Marqversen J. Metabolism of ingested I4C-triolein. Estimation of radiation dose in tests of lipid assimilation using I4C_ and 3H-Iabelled fatty acids. Eur J Nue! Med 1981;6:327-9. 14. Ditchburn RK, Smith AH, Hayter CJ. The assessment of fat absorption in man utilizing single stools or incomplete faecal collections after oral administration of radioactive triolein with an unabsorbed radioactive marker. Int J Appl Radiat Isotop 1974;25:167-76. 15. Wagner A, zum Winkel K. Signifikanz der Fettresorptionspriifungen mit I3II-markierten Lipiden. Med Klin (Miinchen) 1971;66:431-9. 16. van Handel E, Zilversmit DB. Limitation of radioiodine as label for fat. J Lab Clin Med 1958;52:83-9. 17. Newcomer AD, Hofmann AF, DiMagno EP, Thomas pJ, Carlson GL. The triolein breath test. Gastroenterology 1979;76:613. 18. Strange RC, Reid I. Holton D, Jewell NP, Percy-Robb IW. The glyceryl I4C-tripalmitate breath test: a reassessment. Clin Chim Acta 1980;103:317-23. 19. West PS, Levin GE, Griffin GE, Maxwell JD. Comparison of simple screening tests for fat malabsorption. Br Med J 1981;282:1501-4. 20. Goff JS. Two-stage triolein breath test differentiates pancreatic insufficiency from other causes of malabsorption. Gastroenterology 1982;83:44-6. 21. Thorsgaard Pedersen N, Andersen BN, Marqversen J. Estimation of 14C-triolein assimilation as a test of lipid assimilation. Breath test or measurement of serum radioactivity? Scand J Gastroenterol 1982;17:309-16. 22. Thorsgaard Pedersen N. Estimation of 14C-triolein assimilation from postprandial serum radioactivity of I4C. Scand J Clin Lab Invest 1983;43:415-20. 23. Adlung I. Grazikowske H. Diagnosis of fat absorption with I4C-tripalmitate/3H-palmitic acid. Scand J Gastroenterol 1979;14:587-92. 24. Nelson LM, MacKenzie JF, Russel RI. Measurement of fat absorption using ('H) glycerol triether and (14C)glycerol trioleate in man. Clin Chim Acta 1980;103:325-34. 25. Saunders DR, O'Brien TK, Smith K. Disappointment with triethers as markers for measuring triglyceride absorption in man. Gut 1972;13:867-70. 26. Hofmann AF, Poley JR. Role of bile acid malabsorption in pathogenesis of diarrhea and steatorrhea in patients with ileal resection. Gastroenterology 1972;62:918-34. 27. Woolf GM, Miller C, Kurian R, Jeejeebhoy KN. Diet for patients with a short bowel: high fat or high carbohydrates? Gastroenterology 1983;84:823-8. 28. Regan PT, Malagelada J-R, DiMagno EP, Go VLW. Reduced intraluminal bile acid concentrations and fat mal digestion in pancreatic insufficiency: correction by treatment. Gastroenterology 1979;77:285-9. 29. Ammon HV, Thomas PI. Phillips SF. Effects of long chain fatty acids on solute absorption: perfusion studies in the human jejunum. Gut 1977;8:805-13. 30. Schoeller DA, Klein PD, MacLean WC, Watkins JB. Fecal 13 C analysis for the detection and quantitation of intestinal malabsorption. J Lab Clin Med 1981;97:439-48.
Simultaneous Assessment of Fat Maldigestion and Fat Malabsorption by a Double-Isotope Method Using Fecal Radioactivity NIELS THORSGAARD PEDERSEN and HANNE HALGREEN Medical Department F and Department of Clinical Chemistry. Glostrup Hospital. Copenhagen. Denmark
F4C]Triolein and [3H]01eic acid are useful tracers of dietary triglycerides and free fatty acids. The combustion/liquid scintillation counting technique was found a practicable method of estimating the fecal activity of these tracers. When ingested with the nonabsorbable marker 51CrCP, the fecal excretion of 14C and 3H could be estimated accurately from samples of two stools. This technique was investigated as a test of lipid assimilation in a prospective, blind study of 84 consecutive patients suspected of malassimilation. The fecal excretion of 14C was a useful alternative to fecal fat: the patients with "normal lipid assimilation" excreted :::::10.4% of the dose ingested, whereas 25 of the 26 patients with steatorrhea ("unequivocal malassimilation") excreted >10.4%. Fecal fat was normal in 18 patients who exhibited other signs of malassimilation ("equivocal malassimilation"); in 5 of these patients fecal excretion of 14C was >10.4%. Lastly, there was a significant correlation between fecal 14 C and fecal fat (r = 0.82, P < 0.001). The fecal 14CPH ratio was found useful as an index of pancreatic digestive function, being <1.3 in patients with "normal" lipid digestion and ~1.3 in 18 of the 19 patients with pancreatic steatorrhea and in 19 of the 27 patients with severely reduced exocrine pancreatic function. The test reReceived December 2. 1983. Accepted July 11. 1984. Address requests for reprints to: Niels Thorsgaard Pedersen. M.D .. Medical Department 2. Kommunehospitalet. Oester Farimagsgade 5. DK-1399 Copenhagen K. Denmark. This work was supported by grants from the Danish Medical Research Council (J.No. 512-21015 and J.No. 12-1753) and from the Danish Hospital Foundation for Medical Research. Region of Copenhagen. the Faroe Islands. and Greenland (J.No. 46/80). The authors thank Mrs. Grethe Hansen and Mrs. Ingrid Emanuel of the Department of Clinical Physiology and Mrs. Lene Appelt and Mrs. Mariann Hansen of Medical Department F for their technical assistance. They also thank Mrs. Marianne Iversen for dietetic assistance. © 1985 by the American Gastroenterological Association 0016-5085/85/$3.30
suIt was found widely independent either of the amount of carrier fat or the quality of the fecal collections and was therefore useful as an outpatient test. As it provided information about the pathogenesis of malassimilation and was at least as sensitive and specific as a fat "balance study," the test may be a useful alternative to fecal fat measurement.
When estimating lipid assimilation from measurements of the fecal fat excretion. the perpetually taken dietary lipids serve as test substance. The assimilation of a definite quantity of lipids can therefore not be established. Instead, a 3-day fat "balance study" is done. Feces cannot be collected for examination of the test diet residues until the elapse of an ingestion period of at least the gastrointestinal transit time, and then fluctuations in dietary intake and colonic function and incomplete fecal collections may cause uncontrollable error (1). Nonabsorbable gastrointestinal markers may be used to show the excretion of the residues of the test diet and to some extent to monitor the completeness of the fecal collections (1,2), but do not permit correction for the above errors (2,3). At least one-fifth of the fat "balance studies" will therefore be unsuccessful (1). These problems could be avoided by estimating the assimilation of a synthetic lipid. [14C]Triolein has been reported to be useful as a tracer of dietary triglycerides (4,5). When ingested with the no nabsorbable marker 51CrCb, the [14 Cjtriolein excretion can be estimated from fecal samples (4). The excretion of [3H]oleic acid. a tracer of free fatty acids, can be estimated in the same way (5). Triglycerides make up the major part of the dietary lipids. The fecal excretion of [14C]triolein may therefore be a useful alternative to fecal fat (4). The assimilation of [14C]triolein includes digestion, mainly catalyzed by pancreatic lipase, and absorp-
48
GASTROENTEROLOGY Vol. 88, No.1, Part 1
THORSGAARD PEDERSEN AND HALGREEN
tion, whereas [3HJoleic acid is absorbed directly. The ratio between the excretion of 14C and 3H may therefore reflect pancreatic digestive function (5). Thus estimation of the excretion of simultaneously ingested [14CJtriolein and [3H]0Ieic acid appears useful not only as a test of lipid assimilation, but also as a differential test of mal digestion and malabsorption.
Methods Patients Eighty-four patients admitted consecutively over a 2-yr period with symptoms suggesting malassimilation entered in the study. Informed, written consent was obtained and the study was approved by the Medical Ethical Committee of Copenhagen County.
Clinical Procedures The patients were examined in the daily routine work without knowledge of the results of the present investigation. Whenever possible, the diagnosis was established on morphologic criteria, e.g., x-ray, ultrasonography, endoscopy, endoscopic retrograde cholangiopancreatography, and biopsy. The function of the diseased organs was then evaluated by means of the following specific function tests: Exocrine pancreatic function was estimated from the Lundh test by measuring the meal-stimulated duodenal lipase concentration (6). the normal value being >240 kU/L. The 25-g D-xylose test was used to estimate jejunal absorptive surface area; a 1-h plasma xylose concentration >2.2 mmoliL was considered normal (7). Micellar solubilization was estimated from the mealstimulated duodenal bile acid concentration and a bile acid concentration >5.4 mmol/L was considered normal (8).
Finally, the Schilling II test was used to estimate distal ileal function and thereby indirectly the enterohepatic circulation and the micellar solubilization. A 24-h urinary excretion >10% of ingested [58 Co]B 12 -intrinsic factor was considered normal (9). Fecal fat was estimated as the average excretion from the pooled 3-day collection (10). <7 g/day being normal (1).
Classification of Patients Chronic pancreatitis was diagnosed when the Lundh test was abnormal in a patient with at least one of the following findings: previous episodes of acute pancreatitis, pancreatic calcifications, abnormal ultrasonography, and typical endoscopic retrograde cholangiopancreatography findings. Lipolysis was considered reduced when the Lundh test demonstrated a meal-stimulated duodenal lipase concentration <50 kU/L (Le., <20% of normal) (11,12). When, additionally, the fecal fat was >7 g/day, the patient was classified as having unequivocal
mal digestion (Table 1). Those patients who despite a very low duodenal lipase concentration excreted <7 g fat/day were classified as having equivocal mal digestion (Table 1). One patient with a severely reduced duodenal lipase concentration and steatorrhea, due to pancreatic cancer, was classified as having unequivocal mal digestion. Two patients suffering from diarrhea and loss of weight after gastric resection were classified as having equivocal maldigestion (Table 1); they did not fulfill the above-mentioned criteria for chronic pancreatitis, but showed severely reduced meal-stimulated intestinal lipase activity and excreted 6-7 g fat/day. Eight patients with chronic pancreatitis as just defined exhibited only a moderately reduced exocrine pancreatic function, i.e., a duodenal lipase concentration between 240 kU/L and 50 kU/L. These patients excreted <7 g fat/day and were consequently regarded as having normal lipid assimilation (Table 1). Seven patients with various well-established diseases of the small intestine (Crohn's disease of the ileum and distal jejunum, intestinal resection > 1 00 cm, celiac disease, pseudoobstruction, and bacterial overgrowth because of intestinal involvement of systemic sclerosis) and 1 patient with a biliocolonic fistula were classified as having unequivocal malabsorption because of one or more abnormal specific function tests, i.e., xylose test, Schilling test, and duodenal bile acid concentration test, and steatorrhea (Table 1). Another 8 patients had manifest small intestinal disease (Crohn's disease of the ileum, intestinal diverticulosis with bacterial overgrowth, and celiac disease) and abnormal specific function tests, but none had steatorrhea at examination and were therefore classified as having Table 1. Clinical Diagnoses and Clinical Findings Pancreatic disease Chronic pancreatitis Pancreatic cancer Intestinal disease Crohn's disease Intestinal resection Stagnant loop syndrome Lactase deficiency Celiac disease Systemic sclerosis Intestinal ischemia Gastric disease Sequelae after gastric resection Gastric ulcer Gastric cancer Miscellaneous VVatery diarrhea Irritable bowel syndrome Biliary disease Chronic alchoholism Infectious gastroenteritis Totals
No.
nla
33 32
8
ema
uma
emd
umd
6
18
1 18 5 3
1 1
4 2 2
3 4
2
1
1 1 10
1
6
3
3
1
3 1
23 9
9
1
1 84
40
10
2 5
1 1
1
5 1 5 1
6
1 3
2
1
7
8
19
ema, equivocal malabsorption; emd, equivocal maldigestion; nla, "normal" lipid assimilation; uma, unequivocal malabsorption; umd, unequivocal maldigestion.
January 1985
equivocal malabsorption (Table 1). Two other patients were classified ai'! having equivocal malabsorption: 1 patient with diarrhea after gastric resection and a fat excretion of 11 g/day, but with normal specific function tests and intestinal cultures, and 1 patient with clinically classified irritable bowel syndrome, whose fecal excretion of 8 g fat/day was the only abnormal finding. The specific function tests and the fecal fat measurement showed no evidence of malassimilation in 40 patients, including the 8 patients with moderate chronic pancreatitis. They were thus classified as having normal lipid assimilation (Table 1).
FECAL ["4C]TRIOLEIN/[3H]OLEIC ACID TEST
49
out problems of crossover counting and with constant quenching and counting efficiencies (5). According to the findings in the previous studies (4,5), the fecal excretion of 14C and 3H was estimated by extrapolating the activities of 14C, 3H, and 51Cr in the two stools to 100% recovery of 51Cr.
Reproducibility In 16 consecutive patients, the 14C and 3H excretion was estimated twice under the conditions described above within an interval of 14 days. The first of these measurements was used as the test result in the study.
Investigational Procedures Estimation of the fecal excretion of 14C and 3H was done together with measurement of fecal fat and without knowledge of the clinical findings. Details and specifications of the techniques used have been given previously (4,5). Briefly, the patients were served a fat- and caloriefixed diet of 100 g fat and 2100 kcal (8786 kJ) per day from the morning of the day before the ingestion of the isotopes until lunch after the test meal. The breakfast supplied 42 g fat and 900 kcal (3765 kJ). After 1 day on this diet the patient ingested with the breakfast 10 p.Ci of glycerol-tri[1- 14C]01eate ([14C]triolein), 20 p.Ci of [9,10(n)-3H]01eate ([3H]01eic acid), and 20 p.CJ. of 51CrCh, [14C]Triolein and [3H]01eic acid were stored and handled in 100 mg of "cold" triolein on lump sugar; 51CrC13 was dissolved in 100 ml of water. All the isotopes were supplied by the Radiochemical Centre, Amersham, U.K. The fecal collections were started immediately after breakfast. Each stool was collected separately in a sealed plastic bag. After 3 days, the stools were checked for 51Cr by means of a portable y-detector and if 51Cr was present the fecal collections were stopped (78 patients), otherwise the collections continued with daily checks until 51Cr was passed. The two stools with the highest activity of 51Cr were identified and used in the following procedures: the activity of 51Cr was estimated by counting the stools in a whole-body ,},-scintillation counter against weighed standards of the batch ingested. After homogenization (Colworth Stomacher 400, UAC House, London, U.K.), duplicate aliquots of 300 mg were taken from each stool for estimation of 14C and 3H. The aliquots were oxidized at 700°C in oxygen-enriched air in a Tri-Carb Oxidizer (Packard Instrument Co., Inc., United Technologies, Downers Grove, Ill.). In this apparatus the 14C-labeled materials were oxidized into 14COZ and automatically trapped into a scintillation liquid mixture, specifically absorbing COz (Carbosorb and Permaflour V, Packard), in one counting vial with an efficiency of 95% and a coefficient of variation of 3.3% (n = 17) (4). The 3H-labeled materials were oxidized into 3HzO and trapped into a scintillation liquid, specifically absorbing HzO (Monophase 40, Packard), in another counting vial with an efficiency of 98% and a coefficient of variation of 15% (n = 17) (5). All materials other than COz and HzO, including 51Cr, were discharged into a special reservoir. In this way, 14C, 3H, and 51Cr were physically separated with an efficiency >99.5% (5), which enabled liquid scintillation counting of 14C and 3H with-
Influence of Carrier Fat For 14 days, 17 consecutive patients were randomly served either the diet described above or an isocaloric low-fat diet (30 g fat/day and 8786 kJ). The breakfast of this low-fat diet supplied 6 g fat and 3765 kJ. The isotopes were ingested at breakfast on the second day. The test diets were continued until lunch. Fecal collections and measurements were done as described above. The findings on the high-fat diet were used as the test result.
Radiation Dose The radiation dose from [14C]triolein was <20 mrem/p.Ci (13) and that from [3H]01eic acid was <2 mremlp.Ci (13); the whole-body burden from 51CrC13 was <0.2 mreml p.Ci, giving a total whole-body radiation dose of <244 mrem per test.
Statistics The log values of fecal fat and fecal 14C were normally distributed. The level of demarcation between "normal" and abnormal fecal 14C was calculated as the mean + 2 SD of the log values of fecal 14C in the patients with "normal lipid assimilation," whereas the highest fecal 14Cf3H ratio in the patients with normal lipid digestion (Le., normal lipid assimilation and equivocal or unequivocal malabsorption) was taken as the level of demarcation between a normal and an abnormal fecal 14Cf3H ratio. The linear regression analysis of the log values was used to estimate the correlation between fecal 14C and fecal fat and the reproducibility of fecal 14C. In all other analyses nonparametric statistics were used (Mann-Whitney test, Wilcoxon's test of paired differences). Five percent significance limits were used in all tests.
Results Fecal
14
C
The mean fecal 14C excretion in the patients with normal lipid assimilation was 2.5% of the dose
50
THORSGAARD PEDERSEN AND HALGREEN
% of dose
GASTROENTEROLOGY Vol. 88. No.1. Part 1
••
Fecal1"c
100
o 50
• • ••
8 o
o
10
0
2 SO
o
o
o
o o
• •
o
••
If
o
0
1
5
2.5
Figure 1. The fecal HC excretion in the patient groups. Mean ± 2 SD of fecal 14C in the patients with "normal lipid assimilation" is indicated on the ordinate. O. patients with fecal 14C/3H ratio <1.3; •. patients in whom the ratio was 21.3.
8
o o
~
ct
X
o o
~
00 0
ego 0 0
2S0
0.5
o
NORMAL LIPID ASSIMILATION n='O
EQUIVOCAL UNEQUIVOCAL MALABSORPTION n=10 n=7
EQUIVOCAL UNEQUIVOCAL MAlOIGESTION n=8 n=19
ingested, with 10.4% as the upper limit of normal fecal 14C excretion (Figure 1). The fecal 14C excretion correlated significantly with fecal fat (r = 0.82, P < 0.001) (Figure 2) and was significantly higher in both the patients with unequivocal malabsorption/mal digestion and those with with equivocal malabsorption or equivocal maldigestion than in the patients with normal lipid assimilation (p < 0.001 and p = 0.05, respectively) (Figure 1). g/day
Fecal 3 H The patients with normal lipid assimilation excreted up to 15.7% of the ingested dose of [3H]0Ieic acid (Table 2). The patients with unequivocal malabsorption and those with unequivocal maldigestion excreted significantly more 3H (p < 0.01 for both). Thus, 6 of the 7 patients with unequivocal malabsorption and 4 of the 19 with unequivocal mal digestion excreted >15.7% of 3H, the excretion
Fecal fat
100
• •
50
20
10
7
2
1
•
1
2
5
• • •• • •
,
•
•
• •
\
• •
••
Figure 2 . The correlation between the fecal 14C excretion and fecal fat. The marked parts of the axes indicate the normal v
•
•
• ••
.,.....
5
.
•
• • • • • • •• • • •• •••• • • • • •• .a· • • ••
•
•
• •
•
Fecal 14C
7
10
20
50
100 "10 of dose
January 1985
FECAL [14 C]TRIOLEIN/[3H]OLEIC ACID TEST
Table 2. Fecal 3H Excretion and the Fecal 14Cf3H Ratio
Normal lipid assimilation Equivocal malabsorption Unequivocal malabsorption Equivocal maldigestion Unequivocal mal digestion
No.
Fecal 3H median (range)
40
5.0% (1.3 %-15.7%)
0.6 (0.3-1.3)
10
6.0% (3 .0%-27.6% )
0.9 (0.2-1.2)
7
35.4% (12 .0%-46.7% )
1.2 (0.5-1.3)
8
5 .0% (3.0%-10.9°,{,)
0.6 (0.3-1.4)
19
11.8% (4.0%-41.2%)
3.2 (0.9-15.5)
Fecal 14Cf3 H median (range)
being significantly higher in the patients with unequivocal malabsorption than in those with unequivocal maldigestion (p < 0.01) (Table 2). No difference in fecal 3H was found between the patients with normal lipid assimilation and the equivocal groups (Table 2). Fecal 14CPH Ratio
The fecal 14C/3 H ratio was <1.3 in all of the patients with normal lipid assimilation and equivocalor unequivocal malabsorption (Table 2), and was therefore considered to indicate normal lipid digestion. The fecal 14CPH ratio was >1.3 in 18 of the 19 patients with unequivocal mal digestion, this difference being significant (p < 0.001), whereas there was no difference between the patients with equivocal
Fecal 14C PH
,
maldigestion and those with normal lipid assimilation. Figure 3 depicts the relation between the duodenal lipase activity and the fecal 14CPH ratio, and shows that a fecal 14CPH ratio > 1.3 only appeared in patients with severely reduced exocrine pancreatic function. Simultaneous Estimation and Lipid Absorption
of Lipid
Digestion
The results of the fecal estimations of 14C and 14CPH were then combined. Fecal 14C ::5 10.4% of the ingested dose was considered as the test result normal lipid assimilation; fecal 14C > 10.4% of the dose and fecal 14CPH < 1.3 were considered to indicate malabsorption; and fecal 14C > 10.4% of the dose and fecal 14CPH ::::: 1.3 were considered as the test result mal digestion. As there are three possible test results, the sensitivity and specificity of each of these were calculated, both for all of the patients in the study and for the patients in whom lipid assimilation could be established definitely [Le., the patients with normal lipid assimilation and those with steatorrhea (the unequivocal groups)] (Table 3).
Reproducibility The results of the duplicate estimations of fecal 14C correlated significantly, the correlation
1
115,S)
10 •
8
Figure 3. The relation between the fecal 14C/3H ratio and the meal-stimulated duodenal lipase concentration in the 76 patients in whom the Lundh test was done. Normal lipase conc entration was >240 kU/L.
51
•
6
4
3 2 1.3 . . . . _
•
.
• • .... :c• ....• '.
• • •• • .::~ •• •
240
500
50
••
1
••
• KU/I
• •
DUODENAL LIPASE CONCENTRATON
52
THORSGAARD PEDERSEN AND HALGREEN
GASTROENTEROLOGY VoL 88. No. 1 , Part 1
Table 3 . Sensitivity and Specificity of the Fecal [14C]Triolein![3HJOl ei c Acid Test Test result Normal lipid assimilation Malabsorption Maldigestion
Sensitivity 30/31 (25 /26 65/72 (58/58 57/65 (47/48
= 0 ,98 = 0.96)
= 0 .90 = 1.00) = 0 ,88 = 0 .98)
Specificity 39/53 (39/40 10/12 (7/8 19/19 (18/18
= = = = = =
0.74 0.98) 0.83 0 ,88) 1.00 1 ,00)
The "diagnostic sensitivity" of a test result was calculated as th e number of patients in whom the test correctly excluded a specific diagnosis divided by the number of patients in whom the test excluded that specific diagnosis , The diagnostic specificity was calculated as the number of patients in whom the test correctl y indicated a specific diagnosis divided by the number of patients in whom the test indicated that specific diagnosis, The results from the patients with "normal li pid assimilation" and the unequivocal groups are shown in the parentheses,
coefficient being 0.81 and the slope of the regression line being 0.99. Qualitatively, 15 of the 16 duplicate estimations of fecal 14C and fecal 14Cf3H gave the same result; in 1 patient the test result changed from mal digestion to malabsorption. .
Influence of Carrier Fat The fecal excretion of 14C was significantly higher on the high-fat diet than on the low-fat diet (median 5.1%, range 0.6%-132.7% vs. median 4.9%, range 0.3%-91.7%) (p < 0.05), whereas there was no difference between the fecal 14CPH ratio of the highfat and the low-fat diet. Qualitatively, all of the duplicate measurements gave the same test result.
Discussion Estimation of the assimilation of acutely ingested isotope-labeled lipids is a logical alternative to measurement of fecal fat and several radiolipid assimilation studies have been published (4,5,1424). The easily measurable y-emitting 131I-labeled triglycerides were used initially (14,15), but later abandoned because of instability (15,16). In recent years, the stable 14C_ and 3H-labeled fatty acids have been used as tracers, mostly in breath tests (17-21) , but also in tests based on measurement of postprandial serum radioactivities (12,23). The most direct way of estimating the assimilation of one dose of [14C]triolein and [3H]oleic acid is by calculating the difference between the dose ingested and the dose excreted. Hitherto, this has been made difficult by the technical problems of estimating the activity of the l3-emitting 14C and 3H in feces: feces contain lipophilic dyes (e.g. , bilirubin), which make
the scintillation liquid cloudy (quenched) even after lipid extraction, thus resulting in poor counting efficiencies and crossover counting (4,5,24). Further, not more than two isotopes may be present in the sample. The combustion/liquid scintillation counting technique, previously used in studies of bile acid absorption (9), solved these problems and was also found practicable (4,5). The excretion of 14C and 3H is easily estimated when the two are ingested together with a nonabsorbable marker (4,5 ,14,24). No ideal lipid phase marker exists (3,25). Investigations have been done with insoluble chromic salts as a repeatedly ingested marker of dietary lipids (26) and as an acutely ingested marker of isotope-labeled bile acids (9). These studies found that the transit of the lipids was faster than that of the marker, which made the marker-corrected data useless. However, our previous studies did not disclose any systematic difference in the transit of 51CrCl3 and the radiolipids (4 ,5) . Like other investigators (14) , we concluded that 51CrC13, though not ideal, was useful as a marker of gastrointestinal transit of [14C]triolein -and [3f1]oleic acid (4,5): the coefficient of variation of the ratio between 14C and 51Cr and between 3H and 51Cr was 31% in all stools passed (23 patients), but when two stools were used to estimate the fecal 14C and 3H excretions, these estimates correlated highly with the total excretions measured (r = 0.99), were reproducible (coefficient of variation = 6%), and were not further improved by taking more stools (4). But even then the estimated excretion may deviate from the true one (e.g., up to 133% of the ingested dose), although this was mostly seen in patients with massive steatorrhea in whom less than one-thousandth of the stool was oxidized. In this study, we used the two stools with the highest activities of 51Cr to estimate the fecal excretion of 14C and 3H, but any of the stools could have been used for the estimation, so long as they contained sufficient radioactivity for accurate measurement, i.e., >5% of the dose ingested. Thus, the coefficient of variation of estimating fecal 14C from any two stools was 16%, and that of 3H was 18% (17 patients). This means that the result of the technique was less dependent on the quality of the fecal collections than is the fecal fat measurement, and explains the success of the test in all the patients studied. In addition, it was only necessary to collect feces until some of the radioactivity was passed, and for this a y-detector or a visible marker could be used. In this study the needed collection period was 18-124 h (median 38 h). Finally, the fecal excretion of 14C and 3H, like the percentage of dietary fat excreted (27), was only slightly affected by the amount of fat ingested. The
January 1985
result of the test was thus only moderately dependent on patient compliance and the test could therefore be used in outpatients. In fact, the reproducibility was estimated in outpatients and was still found as good as the reproducibility of the most carefully conducted "fat balance" study (1). Because triglycerides constitute the major part of the dietary lipids, we expected the fecal excretion of the triglyceride tracer, [14C]triolein, to reflect fecal fat. This expectation was fulfilled, the correlation between fecal fat and fecal 14C being highly significant (Figure 2). The excretion of [3H]0Ieic acid, which is a tracer of the digestive product of triolein, was expected to be equal to the excretion of [14C]triolein in patients with normal lipid assimilation and malabsorption, i.e., the fecal 14CJ3H ratio was expected to be ~1. In the patients with maldigestion, the fecal excretion of 3H was expected to be normal, but this was not the case in 4 of the patients. One reason for this may be reduced micellar solubilization of the lipolytic products-and [3H]0Ieic acid-due to precipitation of the bile acids because of a more acidic milieu in patients with pancreatic insufficiency (28). Another reason may be that jejunal absorption was reduced by the high concentration of lipids in the jejunum (29). The fecal excretion of 14C, however, exceeded fecal 3H, i.e., the fecal 14CJ3H ratio was >1.3, reflecting digestive function (Figure 3) . In fact the relationship between the fecal 14CJ3H ratio and the meal-stimulated duodenal lipase concentration (Figure 3) exactly mimicked that between fecal fat and duodenal lipase activity (11,12). An exact estimation of lipid assimilation is not always possible in a group of consecutive patients. In our study, the examinations and investigational procedures were done once within 1 mo, after which treatment was often begun, and this made reassessment of the assimilation at the time of investigation impossible. The clinical findings conflicted in 18 of the 84 patients. These patients were therefore classified as having equivocal malabsorption or equivocal maldigestion and constitute a subgroup of special interest when investigating a new test of lipid assimilation. All except 1 patient in the equivocal groups suffered manifest disease in the upper gastrointestinal tract compatible with malassimilation, but fecal fat was normal in 16 of the 18 patients and only insignificantly increased in the last 2 patients. However, taken as a group, the equivocal patients seemed less diseased than the unequivocal patients. Chronic pancreatitis was advanced both in the patients with equivocal and unequivocal maldigestion. The exocrine pancreatic function, as estimated from the Lundh test was, however, more reduced in the patients with unequivocal mal digestion (median 7.8 kUlL, range 0.2-20 kU/L) than in the patients with
FECAL ['4C]TRIOLEIN/[3H]OLEIC ACID TEST
53
equivocal maldigestion (median 20 kUlL, range 2.042 kUlL) (p < 0.001). Differences in colipase secretion (not measured) may explain why some patients had steatorrhea, whereas others, with equally low lipase concentrations, seemingly had normal fecal fat excretion (12). Another reason may be that the measurement of fecal fat was distorted by incomplete fecal collections and constipation in some patients in the equivocal groups. In 5 of these 18 patients, the fecal [14C]triolein/[3H]0Ieic acid test indicated malassimilation in agreement with the specific function tests (Figure 1). In this investigation, only patients with gastrointestinal symptoms were studied, no healthy controls were included. No conclusion can thus be made with regard to normal 14C and 3H excretion. From a clinical point of view it is, however, more interesting to see whether the test can differentiate between patients with diarrhea caused by malassimilation and those with diarrhea from other causes. When looking at the patients in whom lipid assimilation could be established, i.e., normal lipid assimilation and the unequivocal groups, only one false-positive and one false-negative test result were found, and the test discriminated correctly between malabsorption and mal digestion in all except 1 patient in the unequivocal groups (Figure 1). The fecal [14C]triolein/[3H]0Ieic acid test thus seemed specific and more sensitive than the traditional fecal fat measurement. Great efforts have been made to elaborate a nonfecal test of lipid assimilation, e.g., the breath test (1721) and measurement of the postprandial serum radioactivity of 14C and 3H (22,23). These tests are more simple than the one described here, but the results are sometimes distorted by kinetics, unrelated to the assimilation (17,21). The breath test has mostly been studied in patients with obvious steatorrhea (corresponding to our unequivocal groups) and normal subjects, and the results are conflicting (1722). None of these studies showed the breath test to be more sensitive or specific than our test (17-21). Furthermore, the breath test provides no quantitative information (17,18,21)' and it necessitates two tests on alternate days to discriminate between malabsorption and maldigestion (20). Measurement of the postprandial serum radioactivity of 14C and 3H is not as sensitive as our test (22,23). Until more substantial investigations of the alternative methods are published, we believe that fecal collections and measurements are unavoidable. The fecal [14C]triolein/[3H]0Ieic acid test often needs shorter fecal collection time and less fecal handling than the measurement of fecal fat and is therefore less unpleasant. In all the radiolipid assimilation tests the isotopes
54
THORSGAARD PEDERSEN AND HALGREEN
were served at breakfast after a fast of several hours. Whether or not the sensitivity of the test would improve if the isotopes were served later in the day after some fatty meals has not been clarified. The radiation dose of 14C and 3H is modest, but nevertheless their long half-lives raise some concern in using these isotopes in humans, especially in children, in whom estimation of lipid assimilation from samples of feces would be particularly useful. However, 14C and 3H could possibly be replaced by the nonradioactive isotopes 13C and 2H, which are measurable by mass spectrometry after oxidation of feces (30). Expensive equipment was used in this investigation. A portable y-counter could replace the wholebody y-scintillation counter (data not shown), but the oxidizer is essential (price ~$25,000). [14C]Triolein and [3H]oleic acid are stable, both before ingestion and in feces (4,5), and can be transported for measurement at specialized laboratories. The costs of the isotopes, counting liquids, and salaries amount to ~$30 per test. On the other hand, the test can be done on outpatients and is at least as reliable as the inpatient "fat balance" study. Because it provides information about the pathogenesis of malassimilation, the fecal [14 C]triolein![3H]oleic acid test seems profitable both from a clinical and an economical point of view.
References 1. Thorsgaard Pedersen N, Halgreen H. Faecal fat and faecal weight. Reproducibility and diagnostic efficiency of various regimens. Scand J GastroenteroI1984;19:350-4. 2. Davignon I. Simmonds WI. Ahrens EH. Usefulness of chromic oxide as an internal standard for balance studies in formulafed patients and for assessment of colonic function. J Clin Invest 1968;47:127-38. 3. Fordtran JS. Marker perfusion techniques for measuring intestinal absorption in man. Gastroenterology 1966;51:1089-93. 4. Thorsgaard Pedersen N. Estimation of lipid assimilation from faecal samples using I4C-triolein as tracer and 5ICrCI3 as nonabsorbable marker. Scand J Clin Lab Invest 1983;43:3238.
5. Thorsgaard Pedersen N. Estimation of assimilation of simultaneously ingested I4C-triolein and 3H-oleic acid as a test of pancreatic digestive function. Scand J Gastroenterol 1984; 19:161-9. 6. Worning H, Miillertz S. pH and enzymes in the human duodenum during digestion of a standard meal. Scand J GastroenteroI1966;1:268-83. 7. Beck IT, Rona S, McKenna RD, Kahn DS. Evaluation of 200 D( + )-xylose blood level time curves as index of intestinal function. Am J Dig Dis 1962;7:936-48. 8. Westergaard H. Duodenal bile acid concentration in fat malabsorption syndromes. Scand J Gastroenterol 1977;12:11522. 9. Fromm H, Thomas pJ, Hofmann AF. Sensitivity and specificity in tests of distal ileal function: prospective comparison of bile acid and vitamin B12 absorption in ileal resection patients. Gastroenterology 1973;64:1077-90.
GASTROENTEROLOGY Vol. 88, No.1, Part 1
10. Van de Kamer JH, Ten Bokkel Huinink H, Weijers HA. Rapid method for determination of fat in faeces. J BioI Chern 1949;177:347-55. 11. DiMagno EP, Go VLW, Summers kill WHJ. Relations between pancreatic enzyme outputs and malabsorption in severe pancreatic insufficiency. N Engl J Med 1973;288:813-5. 12. Gaskin KJ, Durie PR, Lee L, Hill R, Forstner GG. Colipase and lipase secretion in childhood-onset pancreatic insufficiency. Gastroenterology 1984;86:1-7. 13. Thorsgaard Pedersen N, Marqversen J. Metabolism of ingested I4C-triolein. Estimation of radiation dose in tests of lipid assimilation using I4C_ and 3H-Iabelled fatty acids. Eur J Nue! Med 1981;6:327-9. 14. Ditchburn RK, Smith AH, Hayter CJ. The assessment of fat absorption in man utilizing single stools or incomplete faecal collections after oral administration of radioactive triolein with an unabsorbed radioactive marker. Int J Appl Radiat Isotop 1974;25:167-76. 15. Wagner A, zum Winkel K. Signifikanz der Fettresorptionspriifungen mit I3II-markierten Lipiden. Med Klin (Miinchen) 1971;66:431-9. 16. van Handel E, Zilversmit DB. Limitation of radioiodine as label for fat. J Lab Clin Med 1958;52:83-9. 17. Newcomer AD, Hofmann AF, DiMagno EP, Thomas pJ, Carlson GL. The triolein breath test. Gastroenterology 1979;76:613. 18. Strange RC, Reid I. Holton D, Jewell NP, Percy-Robb IW. The glyceryl I4C-tripalmitate breath test: a reassessment. Clin Chim Acta 1980;103:317-23. 19. West PS, Levin GE, Griffin GE, Maxwell JD. Comparison of simple screening tests for fat malabsorption. Br Med J 1981;282:1501-4. 20. Goff JS. Two-stage triolein breath test differentiates pancreatic insufficiency from other causes of malabsorption. Gastroenterology 1982;83:44-6. 21. Thorsgaard Pedersen N, Andersen BN, Marqversen J. Estimation of 14C-triolein assimilation as a test of lipid assimilation. Breath test or measurement of serum radioactivity? Scand J Gastroenterol 1982;17:309-16. 22. Thorsgaard Pedersen N. Estimation of 14C-triolein assimilation from postprandial serum radioactivity of I4C. Scand J Clin Lab Invest 1983;43:415-20. 23. Adlung I. Grazikowske H. Diagnosis of fat absorption with I4C-tripalmitate/3H-palmitic acid. Scand J Gastroenterol 1979;14:587-92. 24. Nelson LM, MacKenzie JF, Russel RI. Measurement of fat absorption using ('H) glycerol triether and (14C)glycerol trioleate in man. Clin Chim Acta 1980;103:325-34. 25. Saunders DR, O'Brien TK, Smith K. Disappointment with triethers as markers for measuring triglyceride absorption in man. Gut 1972;13:867-70. 26. Hofmann AF, Poley JR. Role of bile acid malabsorption in pathogenesis of diarrhea and steatorrhea in patients with ileal resection. Gastroenterology 1972;62:918-34. 27. Woolf GM, Miller C, Kurian R, Jeejeebhoy KN. Diet for patients with a short bowel: high fat or high carbohydrates? Gastroenterology 1983;84:823-8. 28. Regan PT, Malagelada J-R, DiMagno EP, Go VLW. Reduced intraluminal bile acid concentrations and fat mal digestion in pancreatic insufficiency: correction by treatment. Gastroenterology 1979;77:285-9. 29. Ammon HV, Thomas PI. Phillips SF. Effects of long chain fatty acids on solute absorption: perfusion studies in the human jejunum. Gut 1977;8:805-13. 30. Schoeller DA, Klein PD, MacLean WC, Watkins JB. Fecal 13 C analysis for the detection and quantitation of intestinal malabsorption. J Lab Clin Med 1981;97:439-48.