CLIN. CHEM. 40/8, 1571-1574 (1994)
#{149} Other
Areas of Clinical
Chemistry
HPLC Determination of Polyethylene Glycol 400 in Urine: Oligomeric Profile in Healthy and Celiac Disease Subjects Alexis
Oliva,’
Honorlo
Armas,2
and Jos#{233} B. Farina”3
The absorption of orally administered polyethylene glycol (PEG) has been used to assess intestinal permeability. We describe a simple HPLC technique to determine the
oligomeric profile of PEG excreted in urine. We measured the total (%) PEG excreted in 6 h and the ratio of the four smallest oligomers to the three largest oligomers (expressed as mean percentages). The proposed method differentiates distinct groups of subjects with varying degrees of intestinal permeability detected by intestinal biopsy. The percent of PEG excreted and the oligomer ratio values for heatthy subjects were, respectively, 30.1 ± 3.87 and 0.35 ± 0.03; for celiac patients on a gluten-free diet, 24.5 ± 6.65 and 0.45 ± 0.16; and for celiac patients, 15.0 ± 5.93 and 1.12 ± 0.55. Indexing Terms: intestinal permeability/chromatography,
liquid
The permeability of the intestinal mucosa to molecules varies with molecular size; hence, substances of different molecular sizes such as xylose (1), lactulose (2, 3), inulin (4), urea (5), creatinine (6), mannitol (2, 6), EDTA (7), and polyethylene glycol (PEG) (8) have been used in tests of mucosal integrity under various physiological and pathological conditions (9-11 )4 Cellar disease is especially interesting because it is common (0.03% of the white population) (12), and a diagnostic laboratory test could obviate the need for biopsy. We used PEG because it has a low oral toxicity [although topical toxicity is well documented (13)], is not degraded by intestinal flora, is easily quantifiable in urine, and >90% of the dose administered is excreted in urine within 6 h after administration. Moreover, the distribution of the various homologs in urine may differentiate healthy subjects and patients. We chose a low-molecular-mass PEG, PEG 400, for the present studies because the distribution of the various oligomers and their chromatographic resolution combine to make its analysis notably easier than that of heavier PEGs. In this study we focused on (a) the validation of HPLC determination of PEG 400 in urine, (b) the study of urinary excretion of PEG 400 in healthy subjects to determine the importance of the intersubject variability ‘Departamento de Ingenierla QuImica y Tecnologla Farmac#{233}utics, Facultad de Farmacia, Universidad de La Laguna, 38200 La Laguna, Tenerife, Spain. 2Servicio de Pediatrf a, Hospital Universitario de Canarias, Facultad de Medicina, Universidad de La Laguna. 3Author for correspondence. Fax 3422 630095; E-mail
[email protected]. 4Nonstandard abbreviations: PEG, polyethylene glycol; X,,, total of excreted PEG at 6 h from aiministration with respect to the dose administered; C, quotient between smaller and larger molecules; and ANOVA, analysis of variance. Received December 6, 1993; accepted April 1, 1994.
as well as the influence of gastric emptying, and (c) the estimation of variables affecting the interpretation of the oligomeric proffle of PEG in urine, to establish the capacity of the PEG 400 test to differentiate subjects according to their intestinal permeability.
Materials and Methods Sample Treatment To validate the analytical method, we prepared six calibrators at concentrations of 0.75, 1.5, 2.25, 3.0, 3.75, and 4.5 g/L. Every concentration was replicated four times for urine samples and six times for the mobile phase. To quantify PEG 400 (Serva, Heidelberg, NY) in the urine samples, we used the method proposed by Tagesson and Sjodahl (14). Extrelut 20 columns (Merck, Darmstadt, Germany) were used as the extraction solid phase, and 5 mL of each urine sample was diluted with 10 mL of water and 5 mL of methanol. The whole volume (20 mL) was then added to the column and eluted with chloroform; the eluate was evaporated at 30#{176}C in a nitrogen atmosphere and the dry residues of the calibrators were dissolved at 37#{176}C in 5 mL of the mobile phase. The dry residues of the unknown samples were appropriately diluted to obtain concentrations within the calibration range. The samples to be analyzed were filtered with 0.45-pm (pore size) filters (Miffipore Waters Chromatography, Milford, MA).
Analytical Method Filtered samples (50 L) were injected into a Waters chromatograph (Millipore Waters) equipped with an autosampler (Wisp Model 710B) and a pump (Model 510). The effluent was analyzed with a differential refractometer (Model 410) and the voltage output was integrated with a Model 730 integrator. The mobile phase was a mixture of methanol and water (30:70) at a flow rate of 1 mL/min, and an ODS column (-Bondapak C18, Millipore Waters) was the stationary phase. Deionized water prepared with a MilliQ apparatus (Millipore Waters) was used throughout; all other chemicals and reagents were from Merck. Urinary Excretion of PEG 400 Three groups of subjects were used. The procedures followed were in accordance with the ethical standards of our institution’s responsible committee and with the Helsinki Declaration of 1975, as revised in 1983. No subjects were used who consumed drugs or alcohol before or during the study. Group I: healthy individuals. This group consisted of six healthy volunteers (three men and three women, 23-28 years old, 52-75 kg) with no history of renal or CLINICAL CHEMISTRY, Vol. 40, No. 8, 1994
1571
intestinal permeability disorders. The volunteers received a medical checkup, and gave written consent to their participation in the study. They were given a single dose of 10 gof PEG 400 in 100 mL of water after a 12-h fast. A cross-assay (2 x 2 Latin square) was used, whereby the six subjects were divided into two groups of three each. The first group fasted before and during the assay, whereas the second took PEG 400 with a continental breakfast. Liquid was not restricted in either of the groups. After 1 week the assay was repeated with the conditions interchanged. In both cases the urine samples were collected 1, 2, 4, and 6 h after administration. Group II: celiac patients on a gluten-free diet. This group included five children with celiac disease (four girls and one boy, 2-14 years old, 9-62 kg) who were in the second diagnostic phase of their illness and who had been on a gluten-free diet for 2 years. Permission for their participation in the study was granted by their parents. To obtain the samples of intestinal mucosa we used a pediatric Kilby capsule for intestinal biopsy, introduced via mouth or nose. From the biopsy data, celiac subjects on a gluten-free diet showed normal intestinal viii except for two subjects who exhibited slight partial atrophy. All the children received PEG 400 (0.175 g/kg body weight) diluted in 100 mL of water after a 12-h fast. Urine was collected during the 6 h after administration. Group III: celiac patients. This group consisted of 16 children (eight girls and eight boys, 1-17 years of age, 9-61 kg) who were in the first stage of celiac disease and hence were on a diet in which gluten was still present. The permeabifity test was performed exactly as above. The celiac patients showed subtotal intestinal atrophy, except for two who showed moderate partial atrophy. Results and DIscussIon As a preliminary step to the study of urinary PEG excretion, the chosen analytical method was validated both in the mobile phase, with no impurities that might affect determination, and in urine. The separation of PEG 400 by HPLC resolved several peaks corresponding to the various oligomers Of Mr 282-634. To quantify PEG 400 we used the total area of all peaks. The percentage of the total area for individual peaks allowed estimation of the distribution of the oligomers. Up to eight PEG oligomers could be resolved (Fig. 1), although we selected the seven best quantifiable in urine. The minimum quantifiable concentration of PEG 400 in the mobile phase was 0.75 gIL. The mean extraction recovery for urine was 83% (n = 24), with CV 6.74%, a value close to that obtained by Delahunty and Hollander (15) with chloroform extraction. As with the mobile phase, the analytical technique detected 0.75 g/L PEG 400 concentrations in urine, the proposed linear model being acceptable with a CV of 5.9%. The regression lines and their respective equations are shown in Fig. 2. To clarify and complement the information available (16, 17), we studied PEG 400 urinary excretion in six
1572 CUNICALCHEMISTRY,Vol.40, No.8, 1994
LI 15
5 Fig. 1. Chromotograms
15
25mm
25mm
of PEG 400 inmobile phase (A) and urine (B).
healthy adults; PEG 400 urinary excretion is virtually the same in adults as in children (17), and the use of healthy children as controls would pose obvious practical difficulties.
The experimental design used to carry out the assay was a 2 x 2 Latin square. The statistical model took the following form: (1) where uk was the observation made of the ith individual during the jth period and in the kth test; j.t, overall mean; a, effect due to the ith individual; , effect due to the jth period; Tk, effect due to the kth assay; and e, random error normally distributed with zero mean and cr2 variance. We measured the total PEG excreted (X) in 6 h and the ratio C of the four smallest oligomers to the three largest oligomers (expressed as mean percentages). The experimental design enabled us to carry out a three-way analysis of variance (ANOVA) for the selected variables. Table 1 shows the findings for each subject. Independent of the response used, the ANOVA results, taking as response X,, and the ratio C, showed that the intersubject variability was not significant (a = 0.05, P = 0.257 and 0.168, respectively), although the influence of the ingestion of food (assay factor) differed according to which response was used in the ANOVA. Although not significant when X, is used (P = 0.799),
so
a r
= =
36 0.999
s3a,x=SOS
60
20 y =
(15346±105)#{149}X+(199±307)
0 0
0
0.75
2.25
1.5
3
Concentration
70
3.75
4.5
(g/L)
n24 r = 0.993 1940
60
U
326
370
414
458
502
546
590
634
Molecular mass (Da) Fig. 3. Distribution of various oligomers of PEG 400 for various intervals of urine collectionin group I subjects during the fasting
20 y
10 0
0.75
1.5
assay (mean
=(12312±309).X+(642±903)
2.25
3
3.75
4.5
Concentration (g/L) Fig. 2. Regression lines for total area vs concentration of PEG 400 in mobile phase (upper) and urine (lower).
Table 1. Results of PEG 400 assay Assay 1 Subject 1 2 3
4 5 6
on group I subjects. Assay 2
Small Large Small Large mol., % mol., % C X, mol.. % mol., % 26.8 73.2 0.37 27.3 24.8 75.2 28.4 71.6 0.39 24.7 25.4 74.6 29.4 70.6 0.37 30.3 26.5 73.5 26.8 73.2 0.37 38.8 24.8 75.2 32.6 67.4 0.48 20.2 26.1 73.9 26.5 73.5 0.36 35.9 22.5 77.5
C X,, 0.33 33.8 0.34 31.7 0.36 33.1 0.33 31.0 0.35 30.0 0.29 37.3
emptying
primarily influences the oligomeric profile of PEG 400 in urine when C is used (P = 0.0175), so we confined ourselves to using the data of the fasting control subjects for later contrast with the other groups. Figure 3 shows the oligomeric profile for the selected control subjects (fasting) at each interval of urine collection. The proffle is completely unaltered throughout the collection period, so the test may be simplified as far as collection times are concerned; in principle, a urine collection period of 1 h would enable us to obtain the same results when interpreting the oligomeric profile. The results of the test carried out on the other groups are given in Table 2. In agreement with the results obtained by Lifschitz et al. (18), the celiac patients show an important reduction in X.
SD).
Table 2. Results of PEG 400 assay on group II and group Ill subjects. Group
Small mat., %
1
28.4
Large mol., % 71.5
C 0.40
2 3
22.4
77.5
0.29
29.8
70.2
0.42
27.5
72.5
43.0
57.0 42.1 50.8 48.9 42.5 27.8 44.4
0.38 0.75 1.37 0.96 1.04 1.35 2.59 1.25 1.35 1.01 0.85 0.68
Subject
4 5 6 7 8 9 10
11
12 13 14
15 gastric
±
16 17 18 19 20
21
57.9 49.2 51.1 57.5 72.2
55.6 56.8 50.2 46.2 40.8 45.3 35.7 41.6 47.7
29.5 68.1
43.2 49.8 53.8 59.2 54.6 6.43 58.3 52.3 70.5 31.9
0.83 0.55 0.71 0.91 0.41 2.13
29.3 24.9 14.4 20.6
33.5 10.8 10.7 6.44 10.7
8.60 19.7 16.9 19.6 22.9 13.5 12.2 27.1 12.0 10.0 24.7 14.6
The differences among the three groups are evident from the mean oligomeric profiles in urine for each group (Fig. 4). There is a clear inversion between the percentages of smaller and larger molecules in the celiar patients in group III. To determine the statistical significance of these differences between the three groups, a one-way ANOVA CLINICAL CHEMISTRY, Vol. 40, No. 8, 1994
1573
difference between groups I and II is not significant; it is our opinion that the control group used for contrast with celiac patients could be composed exclusively of subjects consuming no gluten, with these subjects serving as their own controls in future comparisons, when gluten might be reintroduced into their diet. This would require that the number of subjects in this group be increased, which means that the minimum significant difference would be reduced, allowing a clearer distinction to be made
between
References 1. Menzies IS. Intestinal
326 370
414
458
502
546
590 634
Molecular mass (Da) FIg. 4. Mean (± SD) oligomericprofiles in urine. was carried out, with the following as the statistical model: Y
=
equation
being used
(2)
z + a4j +
where Y is the jth observation of the ith subject, is the overall mean, a.t, is the effect of the ith treatment, and eij is. the random error normally distributed with zero mean and o variance. The ANOVA results for X, and quotient C showed that there are significant differences (a = 0.05, P = 0.0022 and 0.0002, respectively) among the three groups. This enabled us to apply Scheff#{233}’s test to contrast pairs of means (Table 3) and consequently to determine which groups of subjects differed from each other. From this comparison, the use of X and quotient C equally allows differentiation between normal subjects (group I) and celiac patients (group Ill), and between celiac patients and those with the gluten eliminated from their diet for 2 years (group II). The results of the test also allow us to infer that the
Table 3. S#{243}heff test for contrast between pairs of means. X,, 6.26
15.8
uSDa
10.0 7.92
C
MSD
0.049 0.727 0.678
0.717
9.47 8.40 a Minimum significant differencefor F = 3.4 with 2 and 24 df(a b.C,d Means of X and C for groups I, II, and Ill, respectively.
1574 CLINICALCHEMISTRY,Vol.40, No.8, 1994
0.565 0.606 = 0.05).
means.
permeability in coeliac disease. Gut 1972; 13:847-9. 2. Menzies IS, Pounder R, Sukha H, Laker MF, Bull J, Wheeler PG. Creamer B. Abnormal intestinal permeability to sugars in villous atrophy. Lancet 1979;ii:1107-9. 3. Pearson AD, Eastham EF, Laker MF, Craft AW, Nelson R. Intestinal permeability in children with Crohn’s disease and coeliac disease. Br Med J 1982;285:20-1. 4. Loehry CA, Axon ATR, Hilton PJ, Hider RC, Creamer B. Permeability of the small intestine to substances of different molecular weight. Gut 1970;11:466-70. 5. Loery CA, Kingham J, Baker J. Small intestinal permeability in animals and man. Gut 1973;14:683-8. 6. Fordtran JS, Rector FC, Ewton MF, Soter N, Kinney J. Permeability characteristics of the human small intestine. J Cliii Invest 1965;12:1935-44. 7. Bjarnaseon J, Peters TJ. In vitro determination of small intestinal permeability: demonstration of a persistant defect in patients with coeliac disease. Gut 1984;25:145-50. 8. Chadwick VS, Phillips SF, Hofmann AF. Measurements of intestinal permeability using low molecular weight polyethylene glycol (PEG) 400.1. Chemical analysis and biological properties of PEG 400. Gastroenterology 1977;13:241-6. 9. Jackson PG, Lessof MH, Baker R, Ferrett J, Macdonald DM. Intestinal permeability in patients with eczema and food allergy. Lancet 1981;i:1285-6. 10. Johansen K, Stintzing G, Magnusson KE. Intestinal permeability assessedwith polyethylene glycols in children with charrhea due to rotavirus and common bacterial pathogens in a developing community. J Pediatr Gastroenterol Nutr 1989;9:30713. 11. Tagesson C, Bengtsson A. Intestinal permeability to different sized polyethylene glycol in patients with rheumatoid arthritis. Scand J Rheumatol 1983;12:124-8. 12. Casamitjana N. Enfermedad celiaca. Offarm 1990;9:49-51. 13. Herold DA, Rodeheaver GT, Bellamy WT, Fitton LA, Bruns DE, Eldlich RF. Toxicity of topical polyethyleneglycol.Toxicol Appl Pharmacol 1982;65:329-35. 14. Tagesson C, SjodahlR. Passage ofmolecules through the wall of the gastrointestinal tract. Urinary recovery of different-sized polyethyleneglycols after intravenousand intestinal deposition. Scand J Gastroenterol 1984;19:315-20. 15. Delahunty T, Hollander D. New liquid-chromatographic method for measuring polyethyleneglycolin urine. Cliii Chem 1986;32:351-3. 16. IrvingCS, Lifschitz CH, Marks LM, NicholsBL, Klein PD. Polyethyleneglycolpolymersoflow molecularweight as probes of intestinal permeabilityI.Innovationsin analysis and quantitation.J Lab Clin Med 1986;107:290-8. 17. MacClung HJ, Powers PA, Sloan HR, Kerzner B. A new micromethod forthe quantification oflow molecular weight oligomere ofpolyethyleneglycol.Clin Chim Acta 1983;134:245-54. 18. Lifschitz CH, Irving CS, Marks LM, Klein PD, FinegoldMJ, Nichols BL. Polyethyleneglycolpolymers oflow molecular weight as probes ofintestinal permeability. II. Application to infants and children with intestinal disease. J Lab Clin Med 1986;108:37-43.
