173,. 507-512 (1948). 6. Buttery, J. E., Kua, S.L., and DeWitt, G. F., The ortho-toluidine ... ed., Iowa State University Press, Ames, Iowa, 1967. reaction of D-xylose ...
CLIN. CHEM. 25/8, 1440-1443 (1979)
A Simplified, Colorimetric Micromethod for Xylose in Serum or Urine, with Phioroglucinol Thomas J. Eberts,’
Richard H. B. Sample, Melvin R. Glick,2 and Gregory H. Ellis
We have developed a simplified xylose assay procedure requires only 10 mm and requires 50 rL of serum or
that
5 tL of urine. The reaction
with phloroglucmnol
is more
sensitive than the classic p-bromaniline color reaction, and requires
only 4 mm of heating for color development. A is mixed with the specimen directly, without prior protein precipitation. Analytical recovery of xylose added to serum was quantitative; precision studies resulted in a between-day coefficient of variation of 5.2%. Glucose, single reagent
which has significant potential for interference in most other xylose procedures, reacts under the test conditions only to the extent of 70 mol of apparent xylose per liter
for a 5.5 mmol/L solution of glucose. The new procedure has been valuable in the assessment of malabsorption, especially in children and infants, where serum xylose is the preferred measurement.
Additional Keyphrases:
pediatric
chemistry
malabsorp-
tion
The xylose absorption test has been widely used as an index of small intestinal function, in both adult and pediatric populations. It remains a valuable technique to be recommended in screening for patients with suspected celiac disease or sprue prior to small bowel biopsy, and may obviate unnecessary biopsies with attendant morbidity and discomfort to patients (1). Introduced in 1937 (2), the method involved an oral dose of D-xylose with determination of urinary xylose after a 5-h collection period. Problems associated with assessment of urinary xylose clearance include factors affecting renal blood flow and glomerular function, and have been recently detailed (3). These problems usually lead to a decreased xylose excretion rate and a false assumption of malabsorption if overlooked. More practical disadvantages of the test include the difficulty of obtaining complete urine collections, especially in the pediatric population, and the requirement for fasting during the urine collection period. To eliminate many of these drawbacks, serum xylose has been assessed in the pediatric (4) and, more recently, adult (3) groups. These studies have shown maximum discrimination between absorption in normal subjects and the lower absorption rate in subjects with untreated malabsorption syndromes 1 h after oral doses of Department of Clinical Pathology, Indiana University Medicine, 1100W. Michigan Street, Indianapolis,
School of
IN 46223
‘Present address:South Bend Medical Foundation, South Bend, IN 46601. 2To whom correspondence and reprint requests should be directet Presented as an abstract at the Spring 1979 meeting of the American Society of Clinical Pathologists, New Orleans, LA, April 1979. Received May 4, 1979; accepted June 7, 1979. 1440
CLINICAL CHEMISTRY,
Vol. 25, No. 8, 1979
xylose as small as 5 g, regardless of age. Hawkins (4) showed that urinary excretion of xylose was more variable than blood concentration. All studies showedgood correlation with results of biopsy of the small intestine, when the lower value for normal serum xylose was 0.3 mmol/L. Many clinical laboratories now measure serum and urinary xylose by the method for free pentoses of Roe and Rice (5) in which pentose, heated in strong acid, forms furfural. Condensation with p-bromaniline produces a pink color that is measured photometrically. The method has proven reliable, but has several drawbacks, including a low sensitivity to xylose. This is particularly significant for the relatively low concentrations of xylose found in serum, especially when the 0.1 g/kg body weight dose is used. The assay takes several hours for completion and requires a protein precipitation step and an incubation period of 70 mm. Glucose interference can be significant, especially at higher blood glucose concentrations.
Other xylose methods, both manual and automated, have been based on the o-toluidine reaction (6, 7), with the choice of wavelength for absorbance determination favoring the pentose. Some procedures have included an additional incubation step with glucose oxidase to remove this sugar before analysis (3, 7, 8). Haeney et al. (3) utilized an automated continuous-flow system based on reduction of alkaline ferncyanide, after treatment of the serum with glucose oxidase to remove glucose. An enzymatic assay (9) utilizing D-xylose isomerase coupled to D-glucitol dehydrogenase is specific and sensitive, but the reagents are not generally available. Excretion of labeled CO2 in breath and xylose in urine and serum has been assessed by a liquid scintillation technique (10) after oral administration of [1-’4C]xylose, but this is of limited use in most circumstances. Trinder (11) has described a urine and plasma xylose assay that combines the advantages of sensitivity and selectivity with stable color and commonly available reagents. Tninder’s procedure, based on the Tollens’ test (12) for pentoses, relies on reaction of the furfural with phloroglucinol to produce a colored compound with high molar absorptivity. He applied the phloroglucinol reaction after preliminary protein precipitation. In our study we have shortened and simplified the Trinder modification of Tollens’ test by minimizing the incubation and color-development time and by eliminating the preliminary protein precipitation step. In addition, we report the use of a single reagent, to simplify the method further.
Materials and Methods Reagents and equipment. Phloroglucinol (1,3,5-trihydroxybenzene) and D-xylose were obtained from Sigma Chemical Co., St. Louis, MO 63178. The color reagent consists of 0.5 g of phloroglucinol, 100 mL of glacial acetic acid, and 10
07
0.S
0.5
C
0.4
a
0
0
a
a
4
4
0.3
0.2
-.----.c
C
-.--
0.1
400
----a--
2
0
W.w.I.n5th
4
S TI,.,i
Fig. 1. Absorbance spectra for xylose or glucose, traced after heating phloroglucinol and specimen for 4 mm. A, serum plus xylose. 0.13 mmol/L; B, xylose standardsolution,0.13 mmol/L; C, glucose standard solutIon, 1.66 mmol/L; 0. senin plus xylose. 0.13 mmol/L, read against serum blank (all other ctrves were vs.reagentblank). E, xylose standardsolution, 0.26 mmol/L
mL of concentrated hydrochloric acid. This reagent is stable four days at room temperature, if protected from light Xylose
centrations.
standard solutions are prepared by dissolving D-xylose in saturated benzoic acid to make 0.7, 1.3, and 2.6 mmol/L conA heat block (Lab-Line
Instruments,
Inc. Meirose
Park IL 60559) set at 105 #{176}C and a spectrophotometer set at 554 nm, with flow-through sampling device (Gilford Instruments, Oberlin, OH 44074) were used. Adding a small amount of water to the holes in the heat block assists in rapid heat transfer at the beginning of the 4-mm heating step. Xylose assay. Place 50 iL of serum, plasma, or xylose solutions, or 5 tL of urine in 16X 100mm disposable test tubes and add 5 mL of phloroglucinol color reagent. Heat all tubes exactly 4 mm at 100 #{176}C, then cool to room temperature in water. After mixing, read the absorbances at 554 nm. Adjust the spectrophotometer to zero absorbance with a reagent blank containing water (50 giL) and phloroglucinol reagent (5.0 mL) before reading the standard solutions, and with a serum blank (50 .tL of xylose-free serum plus 5.0 mL of phioroglucinol reagent) before reading the absorbances of the tubes containing serum or plasma. These “blank” solutions are heated and cooled along with the other solutions before
Table 1. Apparent Xylose Concentration from Added Compounds, When Proposed Procedure Is Used Compound
added,
g/L
Glucose, 1.0
Apparent xylOse, mmol/L
S
10
M’n.,t#{149})
Fig. 2. Absorbance at 554 nm with different incubation times at 100 #{176}C, with use of phloroglucinol reagent as described in proposed method Spectrophotometer was set to zero with water. A. serum plus xyiose,0.13 ,rsnol/L; B, xylosestandardsolution,0.13 mmol/L; C. glucose standard solution, 1.66 mmol/L
reading absorbances. A single serum blank (we use a “normal” control serum) is used for the entire analytical run. Statistical methods. Statistical calculations were performed by calculating the paired Student’s t -test and correlation coefficients by standard methods (13).
Results Absorbance spectra of phloroglucinol
heated with aqueous
xylose standard solution, normal serum, and glucose standard solutions are presented in Figure 1. Absorbance readings at 554 nm minimize the contribution of glucose to the xylose measured. Linearity of the resulting xylose standard curve, when read at 554 nm, extends to at least 13.2 mmol/L. Figure
2 shows the absorbance obtained with variations in time of incubation in the heat block. The absorbance of the xylosephloroglucinol
complex after 4 mm of incubation
is more than
90% of maximum, while the glucose-phloroglucinol absorbance is still less than 40% of the final absorbance achieved with longer heating. Glucose reacts under these conditions only to the extent of 70 mol of apparent xylose per liter for a 5.5 mmol/L solution of glucose. Color due to other (non-xylose) serum and urine constituents is also kept to a minimum by the short incubation period; the use of the serum-phloroglucinol blank for patients’ specimens further decreases the effect of these interferences. Other conditions and compounds that might be encountered in the laboratory analysis of serum or urine xylose were evaluated for their effect on the xylose results by this method (Table 1). Serum separators of the silicone gel barrier type and commonly encountered anticoagulants, including ethylenediaminetetraacetate (EDTA), heparm, citrate, and fluoride-oxalate formulations, did not affect results when used according to manufacturer’s directions to
Fructose, 0.1
0.08 0.53 0
Galactose,
0
obtain serum or plasma. Moderate hemolysis and turbidity related to the presence of lipids also have no effect on this
0
xylose procedure.
Ribose, 0.1
Maltose,
0.1
0.1
Lactose, 0.1
0
Arabinose, 0.1 Methylxanthines, 0.05 Acetylsalicylic acid, 0.1 Acetaminophen, 0.1
0.47 0.13 0 0
Ascorbic acid, 0.1
0
Phenobarbital, 0.2 BilIrubin, 0.204
0 0.06
Table 2. Precision of the Proposed Xylose Method WIthIn-run Mean ± SD,
Serum A (n = 20) Serum B (n = 20) Serum C (n = 24)
mmol/L 0.24 ± 0.009
0.58 ± 0.0 17
-
CV, %
Day-to-day Mean ± SD, mmol/L
CV,
0.13 ± .007
5.2
%
3.8 3.0
CLINICAL CHEMISTRY, Vol. 25, No. 8, 1979
1441
Table 3. Statistical Analysis of Results Obtained by the Present Method (y) and Other Methods (x) Trinder Regression Coefficient
line
of determination (r2) Standard error of estimate No. of determinations
Roe and RIce
y = 0.733x + 0.758
1.05x+0.184 0.988
y mmol/L
0.892
6.521 34
2.136 20
Liquid chromatograply
y
=
0.987x - 0.9 13 0.978 2.809 22
#{149} Slope and intercept are significantly different from 1.0 and 0.0, respectively (p < 0.05).
In addition to the interference studies reported in Table 1, we assessed the specificity of the proposed micromethod for xylose by comparing the results with a “high-performance” liquid-chromatographic separation technique (14). Results from this comparison suggest that the direct, single-reagent phioroglucinol technique described here is more selective for xylose than is the longer method of Roe and Rice. Analytical recovery of xylose added to serum (34 different sera) averaged 100.1%. The range of recoveries was 97.2100.3% during this study, over the xylose concentration range of 0.08-9.6 mmol/L. These recovery values, obtained after addition of xylose to sera from apparently normal individuals, represent the amount of xylose recovered under actual conditions of analysis, because no special allowance was made for the glucose contained in each serum. To measure the precision of the new method, we analyzed serum to which aqueous xylose standard solution was added to obtain concentrations of xylose of about 0.25 and 0.60 mmol/L. These were analyzed 20 times each in a single analytical run. Table 2 shows the means, standard deviations, and coefficients of variation obtained during this study, as well as the day-to-day precision when a single, frozen serum was analyzed each time the assay was done. A statistical analysis of the results obtained by our method, in comparison with those from the method of Trinder (11), of Roe and Rice (5), and of the liquid-chromatographic method is shown in Table 3. Good agreement and correlation were demonstrated between our method and the Trinder and liquid-chromatographic methods, but there was a significant difference between results by our method and the method of Roe and Rice. Figure 3 demonstrates the correlation between our method and the p -bromaniline procedure; results by the latter tended to be slightly higher. The normal range of xyboseconcentration in serum was not evaluated formally during the eourse of this study. Because
of the variation in dosage and time intervals for blood xylose analysis, as well as the difficulty of obtaining adequate urine collections, evaluation of the critical cutoff limits for xylose depend greatly on the circumstances of each test. Further studies of the cutoff limits are in progress. DISCusSIOn Our method is a simplification of the method of Trinder for xylose determinations. It is suitable for serum analysis directly, without need for removing glucose when normal concentrations of it are present in serum, and does not require protein precipitation. It has been in use in our laboratory for over a year with satisfactory results, including in children who have received the 0.1 g/kg body weight xylose dose. The sensitivity of the present method is the same as that described by Trinder. Our rapid, simple technique has been valuable in assessment of malabsorption, especially in infants. By observing the incidence of normal vs. malabsorption results in patients, we obtained close correlation with the p-bromaniline results so that, in each case, the same clinical decision would have been reached regardless of method. The proposed modification of the classical Tollens’ test for pentoses provides a simple, rapid, and clinically useful method for serum or urine xylose measurements. By eliminating the protein precipitation and dilution steps, we have developed a procedure that can be completed easily within 10 mm. Our modification of the Tollen’s reaction results in a significant simplification and savings in time compared with other acceptable methods, making it popular with our technical staff. Its excellent reproducibility, linearity, and accuracy are comparable to or exceed those of the p-bromaniline procedure. Specifically, interference from glucose with our method is less than from that of Roe and Rice. Our method is sufficiently sensitive for serum determinations on small sample volumes, making it especially useful for children, where urine collections are difficult and serum analyses are preferred.
0.8
We thank Dr. Richard M. Thompson for his analysis of serums containing xylose, by use of the liquid-chromatographic method he developed.
0.7
0 #{163}
‘
0
References
E #{176}
. 0
a
1. Rolles, C. J., Kendall, M. J., Nutter, S., and Anderson, C. M., One-hour blood-xylose screening-test for coeliac disease. Lancet ii, 1043-1045 (1973). 2. Helmer, 0. M., and Fouts, P. J., Gastro-intestinal studies VIII. The excretion of xylose in pernicious anemia J. Clin. Invest. 16,343-349
0.5
0
a a.
0.4 .
0
0.3
(1937).
3. Haeney, M. R., Culank, L. S., Montgomery, R. D., and Sanunons, H. G., Evaluation of xyloae absorption as measured in blood and urine: A one-hour blood xylose screening test in malabsorption. Gastroenterology 75, 393-400 (1978). 4. Hawkins, K. I., Pediatric xylose absorption test: Measurements in blood preferableto measurements in urine. Clin. Chem. 16,749-752
E 0
0.2
0
xa. 0.1 (
.;#{248}lP:’‘0
0.1 0.2 0.3 XyIos.. mMoI/Utsr.
,
I
0.4 0.5 0.6 0.7 by So. - Rc. m.thod
0.8
Fig. 3. Correlation between serum xylose resufts by the method of Roe and Rice and tte. present mettioci - - -, line of Identity 1442 CLINICAL CI#{128}MISTRY,Vol. 25, No. 8, 1979
(1970).
5. Roe, J. H., and Rice, E W., A photometric method for the determination of free pentoses in animal tissues. J. Biol. Chem. 173, 507-512 (1948). 6. Buttery, J. E., Kua, S. L., and DeWitt, G. F., The ortho-toluidine
method for blood and urine xylose. Clin. Chim. Acta 64, 325-328 (1975).
7. Braidwood,
J. L., and Smith, M. B., A sensitive automated method of serum and urine xylose. Clin. Biochem. 9,
for the determination
19-21 (1976). 8. Smith, M. B., and Braidwood, J. L., A simple ultramicro method for the determination of serum xylose. Clin. Biochem. 4, 118-122 (1971). 9. Kersters-Hilderson, H., Van Doorslaer, E., De Bruyne, C. K., and Yamanaka, K., Quantitative determination of D-xylose by a coupled reaction of D-xylose isomerase with n-glucitol dehydrogenase. Anal. Biochem. 80,41-50 (1977). 10. Roberts, R. K., Campbell, C. B., Bryant, S. J., and Adames, L.,
Xylose-1-14C absorption test: The use of urine, serum and breath analysis, and comparison with a coloriinetric assay. Aus. N. Z. J. Med. 6,532536 (1976). 11. Trinder, P., Micro-determination of xylose in plasma. Analyst 100, 12-15 (1975). 12. Oshitna, K., and Tollens, B., Ueber Spectral-reactionen des Methylfurfurols. Ber. Dtsch. Chem. Ges. 34, 1425 (1901). 13. Snedecor, G. W., and Cochran, W. G., Statistical
Methods, 6th
ed., Iowa State University Press, Ames, Iowa, 1967. 14. Thompson, R. M., Analysis of mono- and disaccharides by high-performance liquid chromatography of the henzyloxime-perbenzoyl derivatives. J. Chromatogr. 166, 201-212 (1978).
CLINICAL CHEMISTRY, Vol. 25, No. 8, 1979
1443
