European Journal of Clinical Nutrition (1997) 51, 514±519
ß 1997 Stockton Press. All rights reserved 0954±3007/97 $12.00
para-aminobenzoic acid used as a marker for completeness of
24 hour urine: assessment of control limits for a speci®c HPLC
method
J Jakobsen1, L Ovesen1, S Fagt1 and AN Pedersen2
1
Institute of Food Chemistry and Nutrition, National Food Agency, Copenhagen, Denmark; and 2Copenhagen County Center of
Preventive Medicine, Department of Internal Medicine C, Glostrup University Hospital, Denmark
Objective and design: The study comprised three protocols. Protocol 1 compared a HPLC method with the
commonly employed colorimetric diazocoupling method. Protocol 2 examined, if the last dosage of paminobenzoic acid (PABA) could be advanced in the old to allow for a delayed age-dependent urinary excretion
of PABA. Protocol 3 established limits for recovery of PABA in 24 h urine applying the HPLC method.
Subjects and setting: A total of 151 healthy volunteers participated in the study of which 140 were accepted. In
protocol 1: 37 subjects aged 20±78 y were included. All subjects took PABA as recommended (80 mg orally at
08.00, 12.00 and 18.00 h). Protocol 2: compared urinary PABA excretion in two groups of 80 y old subjects who
had their last PABA dosage administered at 15.00 h (n ˆ 16) and at 18.00 h (n ˆ 31), respectively. Protocol 3:
comprised 56 subjects aged 20±80 y. In the younger age group (20±59 y; n ˆ 34) PABA was taken as
recommended, whereas in the older age group (60±80 y; n ˆ 22) the last PABA dosage was advanced three
hours.
Results: Protocol 1: HPLC gave signi®cantly lower PABA recovery results compared to colorimetry, the
difference between methods being 23.9 8.5 mg/24 h (P < 0.001). Protocol 2: higher PABA recoveries were
demonstrated with the advanced dosage schedule compared to the recommended schedule (208 14 mg/24 h vs
181 22 mg/24 h; P < 0.001). Protocol 3: PABA recovery with HPLC was 211 12 mg/24 h, and the lower
limit comprising 95% of subjects was 187 mg/24 h. Similar PABA recoveries were demonstrated in the younger
subjects and the older subjects (211 11 mg/24 h vs 211 13 mg/24 h; NS).
Conclusion: An advanced dosage schedule for PABA in the aged is recommended. Because of lower recoveries
with HPLC, the low limit for recovered PABA in a complete 24 h urine differs from the limit based on
colorimetry. This study found a limit of 187 mg/24 h corresponding to the lower 95% con®dence limit for a
single subject.
Sponsorship: Supported by the National Food Agency, the Velux Foundation of 1981, and the Health Insurance
Foundation.
Descriptors: Dietary survey; urine collection; PABA; HPLC
Introduction
Bingham & Cummings (1983) introduced p-aminobenzoic
acid (PABA) as an objective marker to validate the completeness of 24 h urine collection, and showed that the
lowest acceptable recovery for PABA administered as
80 mg tablets 3 times daily with each main meal was
205 mg when using a colorimetric method. However,
some limitations of the use of PABA have been outlined
(Knuiman et al, 1986; Bingham et al, 1988; Roberts et al,
1990; Leclercq et al, 1991; Bingham et al, 1992). Some
data question the analytical method used for PABA determination because the colorimetric method co-determines
aromatic amines, compounds that can originate from the
intake of a number of commonly used drugs, notably
paracetamol and sulphonamides (Berg et al, 1985; Brown
et al, 1974; Hung et al, 1987; Ito et al, 1982; Kastel et al,
1994; Ward et al, 1985). Other data suggest that PABA
recoveries are lower in the elderly compared to the young
Correspondence: J Jakobsen, Institute of Food Chemistry and Nutrition,
National Food Agency, Mùrkhùj Bygade 19, DK-2860 Sùborg, Denmark.
Received 30 December 1996; revised 8 April 1997; accepted 12 April
1997
subjects, probably re¯ecting delayed renal excretion in the
elderly (Leclercq, 1991).
This study describes a speci®c HPLC method for PABA
determination, compares HPLC with the colorimetric diazocoupling method for analysis of PABA in 24 h urine, and
delimits recovery of PABA for completeness of 24 h urine
collection. Further, the study examines if delayed PABA
excretion in the elderly can be overcome by advancing the
dosage schedule.
Materials and methods
Analytical methods
Liquid chromatography was performed by the following
procedure: 1 ml 8 mol/L NaOH was added to 1 mL urine
in a 25 mL conical ¯ask. The ¯ask was covered with tin
foil, heated for 90 min at 121 C, and subsequently cooled.
The volume of the solution was made up to 50 mL with
0.2 mol/L phosphoric acid. This solution was injected to a
reversed phase column C18 (Spherisorb, 5ODS,
25 cm 6 4.6 mm, Phenomenex, Maccles®eld, UK) where
the mobile phase was a mixture of acetonitril and water
(6:1 v/v) containing 0.02 mol/L potassium dihydrogen
phosphate. pH was adjusted to 3.5 with 10% phosphoric
para-aminobenzoic acid used as a marker
J Jakobsen et al
acid. The injection volume was 20 ml and the ¯ow rate
1.25 mL/min. UV-absorption was measured at 290 nm.
PABA eluted after 6±7 min. The amount of PABA was
calculated by use of PABA as external standard. The
HPLC-system (Waters, Milford, MA, USA) was equipped
with a pump (model 6000A), autoinjector (WISP 710B),
and an UV-detector (490E). Waters software Millenium
2010, version 2.1, was used for quanti®cation of the peak
area.
The colorimetric method included alkaline hydrolysis
followed by formation of diazonium salt with a coupling
agent. The procedure was that described by Williams &
Bingham (1986)). A Shimadzu UV 160-A (Kyoto, Japan)
spectrophotometer was used.
All reagents were of analytical grade. Standards of
PABA, p-aminohippuric acid (PAHA) and p-acetamidobenzoic acid (PAAB) were purchased from Sigma (St.
Louis, MO, USA).
Subjects and urine collection
A total of 151 subjects, all non-institutionalised volunteers,
were enrolled in the study. No dietary changes were
imposed, and all followed their normal daily routine
while taking part in the study. For the collection of urine
all subjects were given two 2 l brown bottles containing
10 mL 1 mol/L HCl for preservation, a 500 mL brown
`visiting' bottle, a funnel, a safety pin for attachment to a
suitable place on the underclothing as a reminder, three
80 mg PABAcheck tablets (Laboratories for Applied Biology Ltd., London, UK; batch no BN8659), and a set of
printed instructions. The subjects were asked to record the
time of the start and ®nish of the collection on a form,
together with the time of taking the tablets, any lost specimens and intake of medications during the urine collection
period. Urine was received at the laboratory immediately
after collection, volumes were estimated by weight (speci®c gravity ˆ 1 g/mL), and aliquots of 50 mL were stored
at 720 C (maximum 30 d).
Protocol 1: Comparison between HPLC and colorimetry.
PABA recoveries by HPLC and colorimetry from 37
healthy subjects aged 20±80 y (17 males; 20 females)
participating in a dietary intake study were studied. None
received medications on a regular basis, but two subjects
took paracetoml on the day of urine collection. The subjects
were instructed to take 80 mg PABA three times with main
meals on the day of urine collection, approximately at
08.00, 12.00 and 18.00 h (PABA18).
Protocol 2: PABA recovery in the old. For this part of the
study, 51 subjects were recruited, all 80 y old (27 females;
Statistical analysis
Agreement between colorimetry and HPLC was assessed
according to Bland & Altman (1986). Analysis of data also
included t-test and linear regression. The data were tested
for outliers by Dixon's outlier test (ISO 5725). P-values
(two-sided) < 0.05 were considered statistically signi®cant.
To ®nd a difference of 5% (considered clinically interesting) between each of the age groups 20±59 y and 60±80 y,
and in each of the 80 y old age groups, with an s.d. of 5%
required 26 subjects in each group with a ˆ 0.05 and
b ˆ 0.95. All values are mean s.d., unless otherwise
speci®ed.
Analytical method
Figure 1 shows typical chromatograms of, respectively, a
standard, a 24 h urine without PABA and a urine with
PABA. The concentration of PABA and the corresponding
peak area was linearly related from 0.25 mg/mL to 25 mg/
mL (r2 ˆ 1.00). The ranges of concentrations of PABA in
the collected urines were 1±7 mg/mL in the measuring
solution. Within-run s.d. was 1.0% (n ˆ 10) and day-today s.d. was 2.4% (n ˆ 48), representing, respectively,
repeatability and reproducibility.
Table 1 shows recoveries of PABA and two metabolites
(PAHA and PAAB) by HPLC. Recoveries of PAHAstandards were also tested with alkaline hydrolysation for
Recovery of p-aminobenzoic acid (PABA) and two of its metabolites
Concentration (mg/mL)
Number
Recovery (%) x^ s.d.
range
a
Protocol 3: Limit for PABA recovery. Sixty-three healthy
subjects aged 20±80 y (41 females; 22 males) were
recruited from the staff, and relatives, of the National
Food Agency of Denmark. Subjects 20±59 y old (n ˆ 37)
were instructed to take the PABA in connection with their
main meals, at 08.00, 12.00 and 18.00 h (PABA18). Subjects 60±80 y old (n ˆ 26) were instructed to advance the
last PABA dosage to 15.00 h (PABA15).
A subgroup of 19 subjects was asked to collect 24 h
urine twice, the time interval between collection days at
least seven days apart, to determine within-subject variation for PABA recovery.
All 24 h urines were analysed by HPLC.
Results
Protocols
Table 1
24 males), who participated in a dietary intake study. All
were free of current acute disease, terminal illness or
mental impairment. Twenty-two subjects took a wide
range of drugs regularly, mostly non-steroid antirheumatics
(n ˆ 10), diuretics (n ˆ 6) or antihypertensives (n ˆ 4). Six
subjects took paracetamol. One group (n ˆ 33) took the
PABA with main meals (PABA18), another group (n ˆ 18)
advanced the last PABA dosage to 15.00 h (PABA15).
Recovery of PABA in 24 h urine was analysed by
HPLC.
Spiking with
PABAa
to urine (no PABA)
PABAb
Spiking with
PAHAc
to collected 24h urine (with PABA)
PAABd
200
17
96.8 2.1
94.5±100.5
250
14
100.4 2.5
97.8±107.4
350
4
96.7 0.8
95.7±97.6
225
3
100.8 2.1
98.7±102.9
Urine without PABA to which 200mg PABA/mL was spiked.
The recovery of PABA, p-aminohippuric acid (PAHA) and p-acetamidobenzoic acid (PAAB), respectively, added to 24 h urine (with PABA).
b±d
515
para-aminobenzoic acid used as a marker
J Jakobsen et al
516
Figure 1 Chromatogram of (A) external standard (3 mg PABA/mL); (B) extract of 24 h urine without intake of PABA; and (C) extract of 24h urine with
intake of PABA.
1 h. The recoveries were between 87 and 93%, but
increased to close to 100% if hydrolysation was extended
to 90 min. The analysed content of PABA in PABA tablets
were found to be 97.9 1.8% (n ˆ 13).
The quantitative limit of detection of PABA was
0.06 mg/mL in the measuring solution. The corresponding
limit of detection for the total amount of PABA in the urine
depended on volume. Mean limit of detection being 5 mg
(range 2±15 mg).
PABA spiked to urine (incl. preservative) kept at 25 C
was stable for at least 10 d in a concentration of 200 mg/ml.
During the development of the method meta-hydroxybenzoic acid was used as internal standard. But external
standard calibration was chosen because an unidenti®ed
peak interfered with meta-hydroxybenzoic acid.
None of the subjects collected 24 h urine twice, with and
without intake of PABA. However, based on the chromatograms no interferences appeared with PABA from the
different kind of drugs the subjects had been taking.
Protocols
Table 2 gives a summary of all the data collected, and
analytical results produced by HPLC.
Protocol 1: The urine from two subjects who took paracetamol on the day of urine collection gave much higher
24 h PABA recoveries when analysed by colorimetry
(400 mg and 610 mg, respectively) compared to HPLC
(203 mg and 206 mg, respectively). One subject had for
unknown reasons lower 24 h recovery by colorimetry
(74 mg) than by HPLC (203 mg).
For the remaining 34 urine samples the plot of differences
between methods against their mean, for the calculation of
PABA in 24 h urine, shows reasonable good agreements
(Figure 2). There does not seem to be any obvious relation
between differences and the mean, but colorimetry results
were signi®cantly higher than HPLC, the difference between
the two methods being 23.9 8.5 mg/24 h (P < 0.001).
Table 2 Summary of the results by HPLC from each of the three protocols
Protocol
Subjects (number) original
rejected
Age (y) x s.d.
range
Dosage schedule
Urine volume (ml) x s.d.
range
Duration of collection (h) x s.d.
range
PABA recovery (mg) x s.d.
range
a
b
1
37
2
48 21
20±78
PABA18a
1799 839
590±4278
23.5 0.9
20.4±24.4
201 22
127±232
2
33
2
PABA18a
2134 882
557±4056
23.8 0.7
21.5±25.0
181 22
132±216
80mg p-aminobenzoic acid (PABA) taken at 8.00, 12.00 and 18.00h.
80 mg p-aminobenzoic acid (PABA) taken at 8.00, 12.00 and 15.00h.
3
18
2
80
PABA15b
2019 755
606±3377
24.0 0.7
22.5±25.3
208 14
178±232
37
3
37 10
23±57
PABA18a
1689 570
619±3148
23.7 0.6
22.0±24.5
211 11
190±232
26
4
68 7
60±80
PABA15b
2005 838
946±4830
23.8 0.3
23.0±24.0
211 13
173±227
para-aminobenzoic acid used as a marker
J Jakobsen et al
(n ˆ 31). Likewise, no signi®cant difference was found
between the two subgroups aged 20±59 y from protocol 1
and 3, who had urinary PABA recoveries of 207 16 mg/
24 h and 211 11 mg/24 h, respectively.
Discussion
Figure 2 Comparison between the accepted values of the HPLC method
with the colorimetric (diazocoupling) for urinary p-aminobenzoic acid
(PABA), Blandt-Altman plot.
The colorimetric analysis gave higher PABA recoveries
in 24 h urine than HPLC. The difference is probably
explained by the determination of aromatic amines from
other sources with colorimetry. The basal level of aromatic
amines in urine was not measured speci®cally in our study,
however, in other studies the mean basal excretion has
varied between 13 mg/24 h and 24 mg/24 h (Bingham &
Cummings, 1983; Knuiman et al, 1986).
None of the urine samples analysed in this study
produced peaks in the chromatograms which interfered
with PABA though the subjects had been taken a variety
of drugs on the day of urine collection including paracetamol. The presented HPLC method does not include calculation of glucuronide metabolites which have been detected
PABA recoveries with HPLC in 24 h urine correlated
inversely with age (r ˆ 70.4; P < 0.05).
Protocol 2: Urine from four subjects had to be rejected
(two were outliers, and two because of failure to ingest
PABA dosages). There were no signi®cant differences
between 24 h urine volumes and duration of collection in
the two groups of 80 y old.
PABA recoveries in 24 h urine were signi®cantly lower
in the PABA18 group than in the PABA15 group
(P < 0.001; Table 2).
Protocol 3: 24 h urines from 56 subjects were accepted.
Rejected samples were from one subject who did not taken
PABA, two were not able to carry out the collection,
one was an outlier and three had not collected for full
24 h (13±20 h).
Mean PABA recovery was 211 12 mg/24 h, and the
lower limit comprising 95% of subjects was 187 mg/24 h.
Similar recoveries were demonstrated when the younger
age group (23±59 y) was compared to the older age group
(60±80 y) (NS; Table 2).
In subjects (n ˆ 19) collecting 24 h urine twice, the
recovery of PABA in the ®rst collection period was
208 10 mg/24 h, and in the second collection period
recovery was 212 20 mg/24 h (NS).
In Figure 3A is shown the recoveries of PABA for
subjects in protocol 1 and PABA18 subjects in protocol
2 (dosage schedule: 08.00, 12.00 and 18.00 h; n ˆ 66),
and in Figure 3B the recoveries for subjects in protocol
3 and PABA15 subjects in protocol 2 (dosage schedule
dependent on age: 08.00, 12.00 and 18.00 for ages
< 60 y and 08.00, 12.00 and 15.00 h for ages 60 y;
n ˆ 72). The mean urinary PABA recovery in the group
who had a dosage schedule dependent on age was
210 12 mg/24 h with a lower 95% con®dence limit
of 186 mg/24 h.
The subgroup aged 60±78 y from protocol 1 (PABA18;
n ˆ 14) had urinary PABA recovery of 191 26 mg/24 h,
not signi®cantly different from the recovery of
181 22 mg/24 h in the PABA18 subjects in protocol 2
Figure 3 Recovered p-aminobenzoic acid (PABA) in 24h urine versus
age of the subject. (A) PABA taken at 8.00, 12.00 and 18.00 h (PABA18),
results from protocol 1 and a part of protocol 2. (B) Last dosage advanced
to 15.00 h in subjects 60 y, results from protocol 3 and a part of protocol
2.
517
para-aminobenzoic acid used as a marker
J Jakobsen et al
518
in urine after intake of PABA (Karnes et al, 1985). The
content of these metabolites may be the reason why we
only recovered 88% of the orally administered PABA. In
10 healthy volunteers who took PABA for three days and
who collected urine for three days, the recovery was
reported to be 98.7 3.7% with colorimetry (Roberts et
al, 1990). These results suggest that the alkaline hydrolysation does not convert all PABA metabolites present in
the urine to PABA.
This study showed that PABA excretion was inversely
associated with the age of the subject. PABA excretion
in the subjects aged 80 y who received their third PABA
dosage at 18.00 h (PABA18) was 14% lower than in
subjects aged 20±59 y, a difference similar to that
demonstrated by Leclercq et al (1991) between young
healthy subjects aged 19±39 y and elderly aged 60±89 y.
By advancing the third 80 mg PABA dosage 3 h, from
18.00 to 15.00 h (PABA15) the PABA excretion increased
to the level in subjects aged 20±59 y. The reason for the
lower PABA recovery in the elderly is probably explained
by delayed PABA excretion due to an age-dependent
gradual decline in renal function (Papper, 1973). Incomplete collections of urine have also been suspected to
account for decreased PABA excretion in the elderly
(Leclercq et al, 1991). However, this explanation seems
unlikely in our subjects, who had normal PABA excretions
when the dosage schedule was advanced, and who demonstrated no group differences in 24 h urine volumes and no
reported losses of urine. Renal function was not assessed in
our study.
PABA excretion is almost complete in six hours in
young subjects (Bingham & Cummings, 1983). Excretion
rates after PABA ingestion has not been determined in the
elderly, but most likely some would demonstrate similar
excretion rates as young subjects. Thus, advancing the last
PABA dosage by 3 h might increase the risk of accepting
a urine as complete, when it in fact was not collected for
the full 24 h. The urines from the three subjects in our
study who were rejected due to their information of
inadequate collection durations would have been accepted
as complete.
The between-subject variation in our study was 5.7%
(n ˆ 56), a value smaller than the within-subject variation
calculated to be 7.4% (n ˆ 19). Similar values are reported
for within-subject variation in 8 healthy young subjects
aged 24±47 y (Roberts et al, 1990).
By dividing subjects into two age groups, 20±59 y and
60±80 y, and advancing the last dosage of PABA 3 h for
the group of elderly subjects to 15.00 h instead of 18.00 h it
is possible to assess a limit, for accepting 24 h urine to be
completely independent of age. The recovery of PABA in
24 h urine in 56 subjects (protocol 3) was 211 12 mg/
24 h, and the lower limit comprising 95% of subjects was
calculated to be 187 mg/24 h. Including the group aged
80 y who received their last PABA dosage 3 h advanced
resulted only in a minor and insigni®cant change of mean
PABA recovery (210 12 mg/24 h; n ˆ 72) and the 95%
lower con®dence limit (186 mg/24 h). The corresponding
PABA recovery determined with the colorimetric method
has been found to be 223 9 mg/24 h in 33 healthy
subjects 22±55 y old (Bingham & Cummings, 1983). If
the contribution from other aromatic amines, analysed to be
13 7 mg/24 h, was deducted, the limit would be the same
as determined with HPLC.
Consequently, this study demonstrates that when using
the HPLC method, the lowest acceptable recovery of
PABA for assessing completeness of 24 h urinary collection is 77.9% (187 mg/24 h) in a single observation when
240 mg (three times 80 mg) of PABA is taken. This limit
will be met by 95% of the subjects.
The reliability of PABA used as an objective marker for
the control of the completeness of 24 h urine is still based
on the fact that the subjects do not extend urine collections
beyond 24 h. The presented HPLC method does seem to
make allowance for one commonly encountered source of
error: the use of some widely used drugs, paracetamol and
sulphonamides, on the day of urine collection. No interferences appeared with PABA from the different kind of
drugs, including paracetamol, the subjects had been taking.
Further studies are required to check the extraction procedure for the complete conversion of PABA metabolites to
PABA, and to assess at which age the last PABA dosage
should be advanced.
AcknowledgementsÐWe wish to thank Lene Breum Larsen, Heddi
Worsùe and Kirsten Pinndal for their skilful assistance with the analytical
work.
References
Berg JD, Chesner J & Lawson N (1985): Practical assessment of the NBTPABA pancreatic function test using high performance liquid chromatography determination of p-aminobenzoic acid in urine. Ann. Clin.
Biochem. 22, 586±590.
Bingham S & Cummings JH (1983): The use of 4-aminobenzoic acid as a
marker to validate the completeness of 24 h urine collections in man.
Clin. Sci. 64, 629±635.
Bingham SA, Murphy J, Waller E, Runswick SA, Neale G, Evans D &
Cummings JH (1992): para-Amino benzoic acid in the assessment
of completeness of 24-h urine collections from hospital out-patients
and the effect of impaired renal function. Eur. J. Clin. Nutr. 46, 131±
135.
Bingham SA, Williams R, Cole TJ, Price CP & Cummings JH (1988):
Reference values for analytes of 24-h urine collections known to be
complete. Ann. Clin. Biochem. 25, 610±619.
Blandt JM & Altman DG (1986): Statistical methods for assessing
agreement between two methods of clinical measurement. Lancet i,
307±310.
Brown ND, Lofberg RT & Gibson TP (1974): High performance liquid
chromatographic method for the determination of p-aminobenzoic acid
and some of its metabolites. J. Chrom. 99, 635±641.
Hung CT, Perrier DG, Schicker AM & Zoest AR (1987): Analysis of paminobenzoic acid and its metabolites in nutrient and plasma by ion
pair HPLC in the NBT-PABA pancreatic function test. Biomed. Chrom.
2, 13±16.
Ito S, Maruta K, Imai Y, Kato T, Ito M, Nakajima S, Fujita K & Kurahashi
T (1982): Urinary p-aminobenzoic acid determined in the pancreatic
function test by liquid chromatography, with electrochemical detection.
Clin. Chem. 28, 323±326.
Karnes HT, Riley CM, Curry SH & Schulman SG (1985): Analysis of NBenzoyl-L-Tyrosyl-p-Aminobenzoic acid (Bentiromide) metabolites in
urine by ion-pair high-performance liquid chromatography. J. Chrom.
338, 377±388.
Kastel R, Rosival I & Blahovec J (1994): Simultaneous determination of
p-aminobenzoic acid and its metabolites in urine by high performance
liquid chromatography. Biomed. Chrom. 8, 294±296.
Knuiman JT, Hautvast JGAJ, Van der Heyden L, Geboers J, Joossens
J, Tornqvist H, Jsaksson B, Pietinen P, Tuomilehto J, Poulsen L, Flynn
A, Shortt C, BoÈing H, Yomtov B, Magathues E, Angelico F, Stetanutti
C, Fazio S, Cantina R, Ricci G, Trichopoulou A & Katapoti Z. (1986):
A multi-centre study on completeness of urine collection in 11 European centres. I. Some problems with the use of creatinine and 4aminobenzoic acid as markers of the completeness of collection. Hum.
Nutr. Clin. Nutr. 40C, 229±237.
Leclercq C, Maiani G, Polito A & Ferro-Luzzi A (1991): Use of PABA test
to check completeness of 24-h urine collections in elderly subjects.
Nutrition 7, 350±354.
Papper, S (1973): The effects of age in reducing renal function. Geriatrics
28, 83±87.
para-aminobenzoic acid used as a marker
J Jakobsen et al
Roberts SB, Morrow FD, Evans WJ, Shepard DC, Dallal GE, Meredith CN
& Young VR (1990): Use of p-aminobenzoic acid to monitor compliance with prescribed dietary regimens during metabolic balance studies
in man. Am. J. Clin. Nutr. 51, 485±488.
Ward CD, Rowe DJF, Fraser K, Tanner AR & Smith CL (1985): Problems
of interference in the PABA test for assessment of pancreatic exocrine
function. Clin. Chem. 31, 661.
Williams DRR & Bingham SA (1986): Sodium and potassium intakes in a
representative population sample: estimation from 24 h urine collections known to be complete in a Cambridgeshire village. Br. J. Nutr. 55,
13±22.
519