In evaluating malabsorption syndromes, the 72-h

Size: px

Start display at page:

Download "In evaluating malabsorption syndromes, the 72-h"

Elmer Simon
6 years ago
Views:

1 GASTROENTEROLOGY 1989:96:421-7 Sudan Stain of Fecal Fat: New Insight Into an Old Test M. R. KHOURI, G. HUANG, and Y. F. SHIAU Gastrointestinal Section, Department of Medicine, Veterans Administration Medical Center and The University of Pennsylvania, Philadelphia, Pennsylvania The 72-h fecal fat determination is used as the gold standard to document the presence of steatorrhea. Although the Sudan stain for fecal fat is advocated as a sensitive screening test, a quantitative correlation between the 72-h fecal fat quantitation and the fecal Sudan stain is lacking. This study was designed to examine the staining properties of different classes of purified lipids in an experimentally defined artificial matrix, and to elucidate the reasons for the lack of quantitative correlation between these two tests. Our results indicate that the "neutral fat" stain without acidification or heating identifies triglyceride; and at an appropriate ph, the "neutral stain" also identifies fatty acid. The "split fat" stain with acidification and heating identifies both triglyceride and fatty acid. After acidification, fatty acid soaps are converted to the nonionized fatty acid. Thus, fatty acid soaps can be identified indirectly as fat droplets that are stained by the split fat stain. Although cholesterol is stained with Sudan stain after heating, upon cooling, cholesterol forms crystals of anhydrous cholesterol, making its staining pattern distinct. Neither the neutral fat nor the split fat stain can detect phospholipid or cholesteryl ester. The 72-h fecal fat determination is a measure of the total fatty acid content after a specimen is saponified. The resulting fatty acids are derived from a variety of endogenous and exogenous sources, including free fatty acids, soaps of fatty acids, triglycerides, cholesterol esters, and phospholipids. Therefore, the 72-h fecal fat quantitation does not differentiate between the primary sources of the measured fatty acid. It is concluded that the 72-h fecal fat determination is not specific for documenting triglyceride (fat) malabsorption. Until new methods are developed that specifically measure fecal triglyceride and fatty acid, the Sudan stain of fecal fat appears to be a more specific method for detecting the presence of triglyceride and fatty acid in a matrix. In evaluating malabsorption syndromes, the 72-h fecal fat quantitation has been used as the gold standard test to document steatorrhea (1,2). The simple Sudan stain on a routine fecal specimen has been advocated as a screening procedure to document the presence of fecal fat (3). Although the accuracy and reproducibility of the Sudan stain is reported to be high (4,5), the correlation between the amount of fecal fat detected by Sudan stain and the 72-h fecal fat measurement is poor. Drummey et al. (6) defined Sudan stain positivity based on the size and number of fat droplets per high-power microscopic field. In Drummey's series, 41 of 41 patients with fecal fat content of >15 g, had a strongly positive Sudan stain. However, 16 of 17 patients with g of fecal fat and 12 of 16 patients with 6-10 g of fecal fat also had a strongly positive Sudan stain. A strongly positive Sudan stain can be seen when the fecal lipid content varies from 6 to >15 g. These data suggest that a quantitative correlation between the Sudan stain and the 72-h fecal fat measurement is lacking. As feces contain different kinds of lipids (7,8), the poor correlation between the 72-h fecal fat determination and the fecal Sudan stain might be due to the fact that these tests measure different forms of fecal lipids. To advance our understanding of the interpretation of the Sudan stain, the staining properties of purified lipids were examined in an artificial matrix, Metamucil (Proctor & Gamble, Cincinnati, Ohio). The effect of saponification on the lipids quantitated by the 72-h fecal fat determination was also examined. Our study reveals that each class of lipids has distinct staining properties. We conclude that the differences in the staining properties of lipids may account for the poor quantitative correlation between the 72-h fecal fat measurement and fecal fat detected by the Sudan stain by the American Gastroenterological Association /89/$3.50

2 422 KHOURI ET AL. GASTROENTEROLOGY Vol. 96, No.2, Part 1 Materials and Methods Methods The purpose of this study was to identify the staining properties of the individual classes of lipids. An artificial, lipid-free matrix was used instead of feces. The matrix was made by adding 20 ml of distilled water to 1 g of Metamucil to form a matrix with a pasty consistency. A known amount of purified lipid (triolein, oleic acid, cholesterol, cholesteryl ester, or phosphotidyl choline) was added to the matrix and mixed thoroughly with a rotary mixer. As normal feces contains a maximum of 5% fat (2), the concentration of lipid in the matrix was varied from 0.5% to 5%. After vigorous mixing, a small sample of the lipid-containing matrix was placed on a slide, stained with Sudan III, and examined by light microscopy. Sudan Stain for Fat Reagents. The following reagents were used. (a) Sudan stain solution: an excess of Sudan III dye was dissolved in 95% ethanol and stirred for 30 min. The solution was filtered through a Whiteman No. 42 filter to obtain a precipitant-free Sudan stain solution. (b) Acetic acid (36%). (c) Ethanol (95%). Procedure. The following procedure was followed. (a) "Neutral fat" stain: a small sample was transferred to a glass slide and two drops of 95% ethanol and two drops of Sudan stain solution were added and mixed thoroughly. After a glass cover slip was placed on top of the sample, the sample was examined by light microscopy. (b) "Split fat" stain: a small sample was transferred to a glass slide and two drops of 36% acetic acid and two drops of Sudan stain solution were added and mixed thoroughly. After a cover slip was placed on top of the sample, the sample was heated on a Bunsen burner until bubbling. The slide was then examined by light microscopy. Interpretation of the Sudan stain. For quantitating the amount of fat detected microscopically, the criteria established by Drummey et al. (6) were used. They are as follows: normal (+): up to 100 fat droplets per high-power field, the diameter of fat droplets being <4 JLm; increased (2+): up to 100 fat droplets per high-power field, the diameter of fat droplets varying between 4 and 8 JLm; markedly increased (3+): >100 fat droplets per highpower field, the diameter of the fat droplets varying from 6 to 75 JLm. Lipid Extraction, Separation, and Identification To examine the effect of the chemical treatment process on the lipid measured by the 72-h fecal fat determination, the conventional method for the 72-h stool quantitation was used (1). Van de Kamer's method. The conventional method for the 72-h fecal fat determination involves the saponification of a homogenized 5-g sample for 20 min with a 33% potassium hydroxide and ethanol-isoamyl alcohol mixture. The sample is then acidified and extracted into petroleum ether. The solvent is evaporated and the residue is dissolved in ethanol and titrated with a standard solution of sodium hydroxide. The weight of the lipid in the specimen is calculated and expressed in grams per 24 h. In our study, the extracted lipids were subjected to lipid identification. Lipid identification. After lipid extraction, the solvent is blown dry and then reconstituted in 1 ml of heptane; 25 JLl of the solution is spotted on a silica G plate for thin-layer chromatographic separation. The plate is developed in a petroleum ether/ether/acetic acid (70:20:4, vol/vol/vol) solvent system. Once the separation is complete, the plate is placed in an iodine chamber for staining. Each lipid class is identified by comparing the separated lipids with standards plated concomitantly. Materials The following materials were used: oleic acid: 99% pure, from Sigma Chemical Co., S1. Louis, Mo., used without further publication; triolein: obtained from Sigma; lecithin: soy bean lecithin obtained from Sigma; cholesterol: obtained from Sigma; Metamucil: obtained from Proctor & Gamble, Cincinnati, Ohio. Results Triglyceride, fatty acid, cholesterol, and phospholipid are the major fecal lipids. Thus, these purified lipids were added to the Metamucil matrix and stained with Sudan red III using both the neutral and split fat staining techniques (Table 1). Neutral Fat Stain of Lipids in a Metamucil Matrix As normal stool contains up to 5% fat, the concentration of triglyceride added to the Metamucil matrix was varied from 0.5% to 5% in separate experiments. When the triglyceride samples were examined with the neutral fat stain, triglyceride was stained readily and appeared as round, red-orange stained, refractile droplets. Although both the 1% Table 1. Staining Properties of Different Classes of Lipids Triglyceride Fatty acid Phospholipid Cholesterol Cholesteryl oleate " See Table 2. Neutral fat stain Stained ph-dependent" Stained upon heating, forms crystals of anhydrous cholesterol Split fat stain acidification Stained Stained Same as neutral fat stain

3 February 1989 SUDAN STAIN OF FECAL FAT: A NEW INSIGHT 423 Figure 1. The staining of triglyceride by Sudan red III illustrating that there are significant differences in the size of the fat droplets when the amount of triglyceride in the matrix is varied from 1% to 5% (x430). A. The neutral stain of a 1 % triglyceride matrix. B. The neutral stain of a 5% triglyceride matrix. and the 5% triglyceride samples were strongly positive in accordance with the criteria of Drummey et al. (6), there were significant differences in the size of the fat droplets when the 1 % (Figure la) and the 5% (Figure lb) samples were compared. When the 1 % and the 5% oleic acid matrices were stained with the neutral fat stain, the fatty acid particles picked-up the stain poorly. Although the majority of the oleic acid droplets did not stain at all, a few stained weakly as light orange droplets (Figure 2A). As some fatty acid droplets stained poorly and others did not stain, the staining variability could not be attributed to inadequate mixing of the fat droplets with the Sudan dye. Fatty acid molecules exist as protonated or ionized species, and in the form of soaps (7). Thus, the variable staining of fatty acid might reflect the chemical state of fatty acid in the matrix. To investigate this possibility, the ph of the matrix was measured (Table 2). In the presence of 5% fatty acid, the ph of the matrix was 5.6, which is higher than the pka of the fatty acid (9). Consequently, most of the fatty acid molecules were in the ionized form. In this state, fatty acid cannot be stained with Sudan red III. When the sample was acidified and then stained with Sudan red III, without the application of heat, the staining of the fatty acid droplets was enhanced (Figure 2B), suggesting that nonionized fatty acid can be readily stained. The split fat stain can identify triglyceride as well. However, this is not the result of the hydrolysis of triglyceride by acidification. Conversely, when sodium hydroxide was added to the sample, fatty acid lost its ability to form fat droplets and could not be stained, suggesting that the sodium soaps of fatty acid cannot be stained with Sudan red III. When a 1 % phospholipid matrix was examined with the neutral fat stain, the Sudan stain was negative. That is, no fat droplets were identified. At room temperature, cholesterol forms crystals and does not interact with Sudan stain. When the cholesterol-containing samples were heated, cholesterol stained avidly with the Sudan dye. However, upon cooling the samples to room temperature, the

424 KHOURI ET AL. GASTROENTEROLOGY Vol. 96, No.2, Part 1 Figure 2. The Sudan staining properties of 5% fatty acid in a matrix (x430). A. At ph 5.6, fatty acid picks up the neutral stain poorly. B.

crystalline pattern of anhydrous cholesterol formed (Figure 3), demonstrating that the Sudan staining characteristics of cholesterol are clearly different from fatty acid and triglyceride.

4 424 KHOURI ET AL. GASTROENTEROLOGY Vol. 96, No.2, Part 1 Figure 2. The Sudan staining properties of 5% fatty acid in a matrix (x430). A. At ph 5.6, fatty acid picks up the neutral stain poorly. B. After acidification with two drops of acetic acid, the Sudan staining of the fatty acid droplets is enhanced. crystalline pattern of anhydrous cholesterol formed (Figure 3), demonstrating that the Sudan staining characteristics of cholesterol are clearly different from fatty acid and triglyceride. When cholesteryl ester was examined, the neutral fat stain failed to stain cholesteryl ester. edly heated, a few droplets stained weakly. We concluded that for practical purposes, phospholipid and cholesteryl ester cannot be stained with the split fat staining technique. Split Fat Stain of Lipids in Metamucil Among the major stool lipids, triglyceride and fatty acid can be stained by Sudan red III whereas phospholipid cannot. The 72-h fecal fat determination may differ from the Sudan stain by including phospholipids in the measurement. To examine this possibility, a matrix containing 1% phosphatidyl choline was subjected to the conventional method for quantitative fecal fat determination as described by van de Kamer et al. (1). After saponification, acidification, and extraction, the extract was reconstituted with Metamucil, and stained with Sudan red III. Microscopic examination revealed that the lipid extract had the staining properties of fatty acid. That is, it did not stain with the neutral fat stain, but stained positively for split fat. Moreover, when the extract was subjected to thin layer chromatographic separation, fatty acid was the predominant lipid identified. When each of the lipid samples was first acidified, heated, and then stained for split fat, both triglyceride and fatty acid were readily stained. Phospholipid and cholesteryl ester, on the other hand, did not stain. When the phospholipid and the cholesteryl ester samples were acidified and repeattable 2. Effect of ph on the Sudan Staining Properties of Fatty Acid Acidification Control (ph 5.6) Alkalinization b Bright orange droplets Variable Most = not stained Few = stained weakly No droplet formation Acidification: two drops of 36% acetic acid. two drops of 1 N sodium hydroxide. b Alkalinization: Identification of Lipids After Saponification

February 1989 SUDAN STAIN OF FECAL FAT: A NEW INSIGHT 425 Figure 3. The distinct crystalline pattern of anhydrous cholesterol stained with Sudan dye (x430).

0.5%, 1%, and 5% triolein and 0.5%, 1%, and 5% oleic acid, were stained with the neutral and split fat staining techniques.

5 February 1989 SUDAN STAIN OF FECAL FAT: A NEW INSIGHT 425 Figure 3. The distinct crystalline pattern of anhydrous cholesterol stained with Sudan dye (x430). Quantitative Sudan Stain of Triglyceride and Fatty Acid To determine the feasibility of using the Sudan stain for a quantitative assessment of stool fat, individual samples of the matrix, containing 0.5%, 1%, and 5% triolein and 0.5%, 1%, and 5% oleic acid, were stained with the neutral and split fat staining techniques. The samples were examined by light microscopy (magnification X400), and the number and size of fat droplets were assessed. It was found that in all the fat-containing samples the number of fat droplets per high-power field was >100. Yet when the size of the droplets was compared in the 1% and the 5% samples, a considerable overlap in the range of droplet size was apparent. However, by comparing a group of the largest droplets in the 1% sample with the largest in the 5% sample, a significant difference could be appreciated. To objectively analyze the differences in droplet size, four high-power fields were selected and photographed using 35-mm color slide film (x 430). The photographic slides were subjected to computer-assisted image analysis to quantify the size of the fat droplets. The 15 largest fat droplets were selected by the observer, and the diameter and the surface area of the fat droplets were measured. The largest 10 of the 15 selected fat droplets from each slide were used for the computer-assisted analysis, which revealed a good correlation between the diameter of the fat droplets and the amount of fat added to the matrix (Figure 4) :. : ;-.:..::.. :. 0.5'1. 1.0'1. 5.0'1. 0.5'1. 1.0'1. 5.0'1. FA FA FA TG TG TG i- I J: Figure 4. The relationship between the diameter of fat droplets and fat concentration when analyzed by computerassisted image analysis. A good correlation between the amount of fat in the matrix and the diameter of fat droplets is demonstrated.

6 426 KHOURI ET AL. GASTROENTEROLOGY Vol. 96, No.2, Part 1 Discussion The fat content of feces from normal individuals is estimated to be no more than 7 g/day when the diet is standard, containing g of fat per day. As the fecal output is estimated to be about 120 g/day (2,10-12), fecal fat c;m account for a maximum of 5% of the normal fecal weight. In the series of Drummey et a1. (6), the Sudan stain was negative in 85% of the normal stools examined. In the remaining 15%, the neutral and split fat stains were weakly positive. A strongly positive Sudan stain, by Drummey's criteria, has more than 100 fat droplets per high-power field, with the droplet diameter varying from 6 to 75 /Lm. The presence of this degree of fat on fecal Sudan stain is indicative of steatorrhea. In the present study; we have shown that an artificial matrix containing 1% triglyceride yields a strongly positive neutral fat stain when the samples are examined by light microscopy with the aid of a micrometer. Similarly, a matrix containing 1% fatty acid gives a strongly positive split fat stain. The degree of positivity demonstrated in the samples containing 1% triglyceride or fatty acid was abnormally high when evaluated by Drummey's criteria. These findings suggest that the feces of normal individuals must have a triglyceride and fatty acid content of <1 %. Therefore, the coefficient of triglyceride and fatty acid absorption must be >99%. Because fecal fat, as measured by van de Kamer's method, is ~ 5 of % stool weight, it must be derived from sources other than triglyceride and fatty acid. Biliary phospholipids and cholesterol esters, and lipids derived from the turnover of intestinal epithelial cells and gut bacteria, must be the principal sources of lipid found in the feces of normal individuals. Thus, the 72-h fat quantitation measures more than just dietary triglyceride and its lipolytic by-products. Phospholipids and cholesterol esters both contain fatty acid and are measured as part of the 72-h fecal fat determination. However, these lipids cannot be stained with Sudan red III. Consequently, the lack of quantitative correlation between the Sudan stain and the amount of fat measured by the van de Kamer method for the 72-h determination can be explained by the fact that these tests measure different forms of fecal lipids. The amount of triglyceride and fatty acid in the matrix correlates well with the size of the fat droplets assessed microscopically. Although the diameter of the fat droplets, estimated by micrometer, was >6 /Lm in the 1 % triglyceride and fatty acid matrices, the computer-assisted image analysis revealed that the diameter of the fat droplets in the 1 % triglyceride samples was slightly <6 /Lm, indicating that the micrometer estimations are less accurate for assess- ing droplet size. In addition, the computer-assisted image analysis revealed that both diameter (Figure 4) and surface area of the largest fat droplets are excellent discriminating parameters for assessing the amount of fat in samples. In summary, based on the examination of purified lipids in a predefined matrix, we found that the Sudan stain is a specific test for detecting triglyceride and fatty acid. The neutral fat stain identifies triglyceride. Moreover, at an appropriate ph, the neutral fat stain also identifies fatty acid. The split fat stain detects both triglyceride and fatty acid. Cholesterol can be stained with Sudan dye after gentle heating. Upon cooling to room temperature, cholesterol forms distinct crystals of anhydrous cholesterol that are readily identified and easily distinguished from droplets of triglyceride and fatty acid. Although phospholipid and cholesteryl ester cannot be stained with Sudan dye, they are measured as part of the 72-h stool fat determination. The 72-h fecal fat determination quantitates the total amount of fatty acid in a sample, but does not differentiate between the primary sources of the fatty acid. We conclude that the Sudan stain is more specific than the 72-h fecal fat quantitation for detecting the presence of triglyceride and fatty acid in a matrix. The Sudan stain, if used properly, may provide quantitative information about the fecal fat content. However, to validate the clinical application of the Sudan stain for fecal fat quantitation, a more specific method for quantitating fecal triglyceride and fatty acid content is needed. References 1. Van de Kamer JH, Huinink HTB, Weyers HA. Rapid method for determination of fat in feces. J Bioi Chern 1949;177: Raffensperger EC, D'Agostino F, Manfredo H, Ramirez M, Brooks FP, O'Neill F. Fecal fat excretion: an analysis of four years' experience. Arch Intern Med 1967;119: Phil HD, Helper OE. Stains for fat in feces. Am J Clin Pathol 1953;23: Ghosh SK, Littlewood JM, Goddard D, Steel AE. Stool microscopy in screening for steatorrhoea. J Clin PathoI1977;30: Luk GD. Screening for steatorrhea (lett). Gastroenterology 1979;77: Drummey GD, Benson JA, Jones CM. Microscopical examination of the stool for steatorrhea. N Engl J Med 1961;264: Shiau YF, Popper DA, Reed M, Umstetter C, Capuzzi D, Levine GM. Intestinal triglycerides are derived from both endogenous and exogenous sources. Am J Physiol 1985;248: G Carey MC. Small DM. The characteristics of mixed micellar solution with particular reference to bile. Am J Med 1970;49: White A, Handler P, Smith EL, Hill RL. The lipids. In: Principles of biochemistry. New York: McGraw-Hill, 1978: Pimparkar BD, Tulsky EG, Kalsar MH, Bockus HL. Correlation

7 February 1989 SUDAN STAIN OF FECAL FAT: A NEW INSIGHT 427 of radioactive and chemical fecal fat determinations in the malabsorption syndrome. Am J Med 1961;30: Skala I, Krondall A. Vulterinova M, et al. Composition of feces in steatorrhea of different etiology: mutual relationship between volume of feces, water, dry matter, nitrogen and fat content. Am J Dig Dis 1968;13: Wollaeger EZ, Comfort MW, Weir JWS, Osterb AE. The total solids, fat and nitrogen in the feces. Gastroenterology 1946;6: Received February 1, Accepted September 12, Address requests for reprints to: Yih-Fu Shiau, M.D., Ph.D., Gastrointestinal Section (l11gi), Veterans Administration Medical Center, University and Woodland Avenues, Philadelphia, Pennsylvania This study was supported by funds from the Medical Research of the Veterans Administration. Dr. Khouri is supported by Training Grant AM from the National Institutes of Health.

A New Method of Quantitative Fecal Fat Microscopy and Its Correlation With Chemically Measured Fecal Fat Output

Clinical Chemistry / QUANTITATIVE FECAL FAT MICROSCOPY A New Method of Quantitative Fecal Fat Microscopy and Its Correlation With Chemically Measured Fecal Fat Output Kenneth D. Fine, MD, and Frederick