Value
of Random
Urinary
Homovanillic
Acid
and Vanillylmandelic
Acid
Levels
in the
Diagnosis
and Management
of Patients
with
Neuroblastoma:
Comparison
with 24-Hour
Urine
Collections
Mendel
Tuchman,
MD, Christopher
L. Morris,
MD,
Margaret
L. Ramnaraine,
BS, Larry
D. Bowers,
PhD,
and
William
Krivit,
MD, PhD
From the Departments of Pediatrics, and Laboratory Medicine and Pathology, University
of Minnesota, Minneapolis
ABSTRACT. Urinary homovanillic acid (HVA) and van-illylmandelic acid (VMA) levels were determined in ran-dom samples and in 24-hour collections from 13 patients with neuroblastoma and 22 patients without neuroblas-toma. Random sample levels were compared with levels in 24-hour collections and showed a positive correlation of 95% for HVA (N = 59) and 93% for VMA (N = 52).
No false positives or false negatives occurred using
ran-dom samples for diagnosis. Nonneuroblastoma (normal) HVA (N = 126) and VMA (N = 1 19) levels are reported for different age groups. Sequential random HVA and VMA determinations in patients with neuroblastoma during and after therapy are shown. Random urinary
HVA and VMA levels are shown to be adequate for
utilization in the diagnosis of neuroblastoma and sequen-tial determinations of random HVA and VMA are shown to be helpful in the follow-up of those patients. Pediatrics
1985;75:324-328; homovanillic acid, vanillylmandelic acid, neuroblastoma, neural crest tumors, catecholamines.
Catecholamines and their acidic metabolites are
excreted in larger than normal amounts in patients
with tumors originating from the neural crest.’6
The diagnosis is easily established by determination of urinary homovanillic
(4-hydroxy-3-methoxy-phenylacetic) acid and vanillylmandelic
(4-hy-droxy-3-methoxymandelic) acid. During the past 20
years since this test was first used, the
measure-Received for publication Jan 30, 1984; accepted March 15, 1984.
Reprint requests to (MT.) Department of Pediatrics, University
of Minnesota Hospitals, 420 Delaware St, SE, Minneapolis, MN
55455
PEDIATRICS (ISSN 0031 4005). Copyright © 1985 by the
American Academy of Pediatrics.
ments of HVA and VMA have become a routine
test for diagnosis and follow-up of children with
neuroblastoma.7’#{176}
The original methodology assumes that a 24-hour
urine collection is needed, and several authors still
recommend a 24-hour urine collection as the only
way to obtain reliable results.1’’3 The question
whether random urinary HVA and VMA levels are
adequate for diagnosis and follow-up of patients
with neuroblastoma has not been answered.
Pro-longed urine collections especially in young infants
are hard to obtain, time consuming, inconvenient
for patients and staff, and lead to higher cost.’4
We have studied the value of random urinary
HVA
and
VMA
levels
in the
diagnosis
of patients
with neuroblastoma by comparing random urine
samples and 24-hour urine collections. We report
normal values of random urinary HVA and VMA
levels in different age groups, and, finally, we
dem-onstrate the use of sequential random HVA and
VMA
determinations
in follow-up
of patients
with
neuroblastoma.
MATERIALS
AND
METHODS
Comparison
Between
Random
Samples
and
24-Hour Collections
The study included 35 patients ranging in age
from newborn to 18 years; there were 13 patients
with neuroblastoma (as diagnosed by histology) in
various stages of disease, and 22 patients who were
sick at the time of sampling but in whom the
Some patients with neuroblastoma had more than
one urine collection performed. No dietary or drug
restrictions were imposed except in the case of
drugs known to increase HVA or VMA excretion
(eg, adrenalin, L-dopa, dopamine, ACTH, insulin,
caffeine). Twenty-four hour urine collections were
performed on ice, and a paired random sample,
consisting of a single void, was collected arbitrarily
within 24 hours before or after the 24-hour urine
collection. Urine that was not run immediately was
kept at less than 5#{176}C.
HVA
and
VMA
determinations
were
performed
in our laboratory by capillary gas chromatography
as described previously.9 Briefly, the method
in-cludes solvent extraction of urinary acids with
eth-ylacetate and ether, drying the organic phases, and
derivatization for capillary gas chromatographic
analysis. Quantities were then calculated relative
to an internal standard, and concentrations were
expressed relative to urinary creatinine, which
avoids factitious high or low levels due to
differ-ences in urinary levels of concentration.
Determi-nations on 24-hour collections were done in another
laboratory using a packed column gas
chromato-graphic method as described previously.’0 That
method is the standard procedure for urinary HVA
and VMA determination in several laboratories.
Seven samples could not be included in the study
because the other laboratory could not obtain exact
determinations. This problem is relatively common
to packed column chromatography and is due to
interfering urinary substances.’5’7
HVA
and
VMA
determinations
(N
= 59 and 52,respectively) from random samples and 24-hour
collections were statistically analyzed by Student’s t test using a null hypothesis and by linear regres-sion (forced through the origin) for correlation.
Normal
Values
Normal values of HVA and VMA in random
urine samples were obtained from 93 pediatric
pa-tients in whom neural crest tumors have been
ex-cluded and eight healthy, adult laboratory
person-nel. Dietary and drug restrictions were the same as
described above. The ages of patients ranged from
newborn to 32 years. Random urine samples were
collected and run as stated above. HVA and VMA
determinations were arranged into six pediatric age
groups and an adult age group. Mean and standard
deviations were determined for HVA and VMA for
each age group, a total of 126 and 119
determina-tions, respectively, with between eight and 30
de-terminations per group.
Sequential
Determinations
Two patients with stage IV neuroblastoma were
followed after diagnosis for 150 and 200 days,
re-spectively. Sequential random HVA and VMA
levels were done (22 and 19 determinations,
respec-tively) before, during, and after several courses of
chemotherapy. Some sequential determinations
were done within the same day. The results of
sequential determinations were plotted graphically.
RESULTS
Using capillary gas chromatography, we found no
interference in either separating or quantitating
HVA
and
VMA.
The
result
of
t test showed nosignificant differences between paired random
sam-ples and 24-hour urine samples. Mean values are
compared in Table 1 and show small variations.
Because the sample population is not normally
distributed, linear regression was done plotting
ran-dom v 24-hour values for HVA (Fig 1) and VMA
(Fig 2) and showed correlation coefficients (R) of
.951 and .929, respectively. (P < .001). The observed
variabilities (R2) show that 10% of HVA and 14%
of VMA variabilities cannot be accounted for by
the 24-hour urine collections (R2 = .904 for HVA
and .863 for VMA, P < .001). In the time since we
started using random samples, ten cases of
neuro-blastoma have been newly diagnosed; there have
been no false-positive or false-negative results.
Normal values of HVA and VMA are shown in
Table 2, with mean values and standard deviations
for seven age groups. These values closely match
those established for 24-hour urine collections by
the second laboratory. The graph of the mean
val-ues by age group (Fig 3) shows the expected sharp
decline with age and leveling off to adult values. In
patients who do not have neuroblastoma, the ratio
between HVA and VMA levels remains relatively
TABLE 1. Mean Values of Homovanillic Acid (HVA) an
in Random Samples Versus 24-Hour Collections
d Vanillylmandelic Acid (VMA)
Test HVA (zg/mg of
Urinary Creatinine)
VMA (pg/mg of Urinary Creatinine)
Mean ± SE No. of Determinations
Mean ± SE No. of
Determinations
Random samples 95.61 ± 15.45 59 24-h urine collections 90.93 ± 16.23 59
constant. The mean HVA/VMA molar ratio was 1.642 ± 0.242.
Sequential HVA and VMA levels in random urine
samples from two patients with stage IV
neuroblas-toma following the diagnosis are shown in Fig 4.
The levels are slowly declining following multiple
courses of chemotherapy. The sequential
determi-nations done within the same day are all
signifi-cantly above normal values.
DISCUSSION
Increased urinary excretion of catecholamines
and their metabolites in patients with
neuroblas-toma was first reported by Voorhess and
Gard-ner.’8’19 The diagnostic value of VMA in
neuroblas-toma has been emphasized because this acid is the
metabolite of epinephrine and norepinephrine.2#{176}
The
diagnostic
value
of HVA,
the dopamine
me-0 C 0 - -00 #{176}D) 40( 0
tabolite, became apparent when several authors’4’2’
reported that HVA is more consistently elevated in
neuroblastoma
and
that
20%
to 30%
of patients
with neuroblastoma excrete normal amounts of
VMA.2’24
Several authors”3 still feel that HVA and VMA
levels in 24-hour urine collections only, are reliable
enough
in establishing
the
diagnosis
of
neuroblas-toma.
This
concept
persisted
as more
databecame
available regarding the effects of diet on quantities
of catecholamine metabolite excretion as well as
information on the circadian rhythm and seasonal
variations of catecholamine excretion.253#{176} The effects of different diets containing increased
ci%o:#{176} I
0 100 200 300
pg/mg UCr Random Sample
Fig 2.
Vanillylmandelic acid levels in random urinarysamples u 24-hour urine collections. A total of 52
deter-minations were compared. Correlation coefficient was .929 (P < .001). Abbreviation used is: UCr, urinary cre-atinine.
(HVA) and Vanillylmandelic Acid 1
0
100
200
300 400,ug/mg UCr Random Sample
Fig 1. Homovanillic acid levels in random urinary
sam-ples v 24-hour urine collections. A total of 59
determi-nations were compared. Correlation coefficient was 0.951
(P < .001). Abbreviation used is: UCr, urinary creatinine.
TABLE 2. Nonneuroblastoma Levels of Random Urinary Homovanillic Acid
(VMA) in Different Age Groups*
Age HVA (zg/mg of
Urinary Creatinine)
VMA (zg/mg of Urinary Creatinine)
Range Mean ± SD No. of Range Mean ± SD No. of
Determinations Determinations
(N = 126) (N = 119)
0-3 mo 11.3-35.0
(0-46)
22.33 ± 7.46 12 5.9-37.0
(0-56.8)
19.49 ± 10.32 12
3-12 mo 8.4-44.9
(0-46)
27.79 ± 9.57 30 8.4-43.8 (0-22.2)
22.81 ± 10.59 28
1-2 yr 12.2-31.8 (4.7-41.5)
19.65 ± 6.00 16 7.9-23.0
(0-13.6)
14.88 ± 3.85 15
2-5 yr 5-10 yr 3.4-32.0 (3.1-28.7) 6.8-23.7 (0.6-19.4)
15.33 ± 7.69 21
12.80 ± 5.10 25
2.9-23.0 (2.2-11.6) 5.8-18.7 (2-8.3)
11.25 ± 7.61
9.34 ± 3.68
22
21
10-15 yr 3.2-13.6 (0-18.1)
8.12 ± 3.88 13 1.6-10.6
(0-7.6)
5.17 ± 2.51 13
>15 yr 3.2-9.6
(0-5.7)
5.66 ± 1.86 9 2.8-8.3
(0-6.2)
4.52 ± 1.77 8
300
ii
0
a,
E
a,
Fig 3. Mean normal homovanillic acid (HVA) and
va-nillylmandelic acid (VMA) levels in random urinary sam-ples from various age groups. Abbreviation used is: UCr,
urinary creatinine.
0
C)
E
0)0
:D
0)
E
C) 3mo l2mo y 5y lOy 15y
Age
0-0 HVA
o----0 VMA
A Chemotherapy
ii
50 100 150 200
400
amounts of vanillin, vanilla, and phenolic acids on
HVA and VMA excretion were studied by Weetman
et al.’3 They concluded that restricted diet is not
necessary prior to screening children for neuroblas-toma.
Despite the fact that a diurnal variation is pres-ent in the excretion of catecholamines, being higher
in the evening than in the morning,27 Oishi3’ found
that in pheochromocytoma, VMA excretion is
in-variably higher than normal values at any time
during the day. He concluded that random
one-hour urine samples were useful for diagnosis of
pheochromocytoma. Hinterberger and
Barthol-omew’4
have
shown
that
random
specimens
may be
useful in screening for neural crest tumors.
Our
data
show
that
the correlation
of HVA
andVMA levels between random samples and 24-hour
urine collections is high enough to be reliable for
the diagnosis of neuroblastoma. The fact that the
tests were done in two different laboratories
in-creases the reliability of this information. The
ac-curacy of the capillary gas chromatographic method
helps to avoid false positives and false negatives.
The
nonneuroblastoma
“normals”
are suitable
as
control subjects. McKendrick and Edwards” have
found that residence of children in hospitals rather
than at home, sex and presence of tumor other than
neural crest in origin, or postoperative stress have
no effect on VMA excretion. Our study agrees with
their findings. The nonneuroblastoma HVA and
VMA
levels
that
were
obtained
from
sick patients
did not show elevation of excretion in diseases other
than neuroblastoma. The relative constant molar
Days Post Diagnosis
Fig 4. Sequential homovanillic acid (HVA) and
vanil-lylmandelic acid (VMA) determinations in random
un-nary samples of two patients with neuroblastoma: patient
1, a 2#{189}-year-old-girl with stage IV neuroblastoma, and patient 2, a 4-year-old boy with stage IV neuroblastoma. Abbreviation used is: UCr, urinary creatinine.
ratio between
HVA
and
VMA
levels
in our control
subjects probably reflects the activity of the enzyme
dopamine j3-hydroxylase which catalyzes the
hy-droxylation of dopamine to norepinephnine which
is then metabolized to VMA.32 Preliminary results
of a study that we are currently conducting show
that sequential determinations of HVA and VMA
in random urinary specimens are useful in
follow-up of patients with neuroblastoma after surgical
resection or during and after chemotherapy. In view
of the data presented here, we feel that random
urinary specimens are adequate for diagnosis and
follow-up of patients with neuroblastoma.
ACKNOWLEDGMENTS
This work was supported by the following grants:
Hematology training grant NIH-NHLBI-5T32-AM07145,
Oncology training grant NIH-NCI 5T32-CA09445,
and Pediatric Oncology/Marrow Transplantation
Pro-gram Project NIH-NIC 2P01-CA21737.
We thank Dr Les Robison, Sheryl Frankel, and the
personnel of the Steroid Lipid Laboratory of the Univer-sity of Minnesota Hospitals for their help in the
prepa-ration of this work.
REFERENCES
1. Bell M, Steward JK: Neuroblastoma and catecholamine excretion. Lancet 1961;1:1288
2. Voorhess ML, Gardner LI: Studies of catecholamine
excre-tion by children with neural tumors. J Clin Endocrinol
Metab 1962;22:126-133
3. Von Studnitz W: Chemistry and pharmacology of
catechol-amine secreting tumors. Pharrnacol Rev 1966;18:645
4. Bohuon C: Catecholamine metabolism in neuroblastoma. J Pediatr Surg 1968;3:114-118
5. Schweisguth 0: Excretion of catecholamine metabolites in
urine of neuroblastoma patients. J Pediatr Surg
1968;3:118-120
6. Hinterberger H, Bartholomew RI: Catecholamines and their
acidic metabolites in urine and in tumor tissue in neuro-blastoma ganglioneuroma and pheochromocytoma. Clin
Chim Acta 1969;23:169-175
7. Yoshiue M, Yoshioka M, Samirra Z: Determination of
va-nillylmandelic acid and homovanillic acid in urine by gas
chromatography. Chem Pharm Bull 1975;23:3040-3041
8. Soldin SJ, Hill JG: Simultaneous liquid-chromatographic
analysis for 4-hydroxy-3-methoxymandelic acid and 4-hy-droxy-3-methoxyphenylacetic acid in urine. Clin Chem
1980;26:291-294
9. Tuchman M, Crippin PJ, Krivit W: Capillary gas chromato-graphic determination of urinary homovanillic acid and vanillylmandelic acid. Clin Chem 1983;29:828-831
10. Foster-Leiendecker C, Freier .EF: Improved simultaneous
gas-chromatographic analysis for homovanillic acid and
vanillylmandelic acid in urine. Clin Chem 1981;27:2029-2032 1 1. McKendrick T, Edwards WH: The excretion of
4-hydroxy-3-methoxymandelic acid by children. Arch Dis Child
1965;40:418-425
12. Johnsonbaugh RE, Cahill R: Screening procedures for neu-roblastoma: False-negative results. Pediatrics 1975;56:267-270
13. Weetman RM, Rider PS, Oei TO, et al: Effect of diet on
urinary excretion of VMA, HVA, metanephrine and total
free catecholamine in normal preschool children. J Pediatr 1976;88:46-50
14. Hinterberger H, Bartholomew RJ: The chemical diagnosis of ganglioneuroma using random specimens of urine. Aust
Paediatr J 1970;6:222-227
15. Coward RF, Smith P: The gas chromatography of aromatic
acids as their trimethylsilyl derivatives including
applica-tions to urine analysis. J Chromatogr 1969;45:230-243
16. Bjorkman L, McLean C, Steen G: Organic acids in urine
from human newborns. Clin Chem 1976;22:49-52
17. Melchert HV, Hoffmeister H: More on gas liquid chromato-graphic analysis for homovanillic and vanillylmandelic acids
in urine from cases of neuroblastoma. Clin Chem 1978;
24:842-843
18. Voorhess ML, Gardner LI: Catecholamine metabolism in neuroblastoma. Lancet 1960;2:651-652
19. Voorhess ML, Gardner LI: Urinary excretion of
norepineph-rine epinephrine and 3 methoxy-4-hydroxymandelic acid by children with neuroblastoma. J Clin Endocrinol Metab
1961;21:321-335
20. Moyano MB, Bergada C, Becu L: Significance of catechol-amine excretion in the follow-up of sympathoblastomas.
Cancer 1971;27:228-232
21. Brewster MA, Berry DH, Moriarty B, et al: Urinary 3-methoxy-4-hydroxyphenylacetic (homovanillic) and
3-methoxy-4-hydroxymandelic (vanillylmandelic) acids:
Gas-liquid chromatographic methods and experience with 13 cases of neuroblastoma. Clin Chem 1977;23:2247-2249 22. Voorhess ML: Neuroblastoma with normal urinary
cate-cholamine excretion. J Pediatr 1971;78:680-683
23. Helson L, Bethune V, Schwartz MK: Clinical evaluation of the VMA test strip. Pediatrics 1973;51:153
24. Liebner HJ, Rosenthal IM: Serial catecholamines in the radiation management of children with neuroblastoma. Can-cer 1973;32:623-633
25. Waalkes TP, Sjoerdsma A, Creveling CR, et al: Serotonin, norepinephrine and related compounds in bananas. Science
1958;127:648-650
26. Zicha L, Au BS: Uber den einfluss von koffien auf den
tagesrhytmus den vanillinmandelsaure und
homovanillin-saure.Med Welt 1969;20:1674-1678
27. Winkel P, Slob 5: Catecholamine excretion of normal male
adolescents during various periods of the day cycle. Clin
Chim Acta 1973;45:113-119
28. Wisser H, Knoll E: Ergebmisse einer kurzzeitstudie uber die
circadiane rhythmik bei normaler und erhohter
katecholam-inausscheidung in urin. Clin Chim Acta 1975;59:1-7
29. Markianos E, Beckmann H: Diurnal changes in dopamine
3hydroxylase. Homovanillic acid and
3-methoxy-4-hydrox-yphenylglycol in serum of man. J Neural Transm 1976;
39:79-93
30. Yamamoto T, Doi K, Takeuchi Y, et al: Seasonal variation of urinary excretion of total metanephrines. Clin Chim Acta
1976;68:241-246
31. Oishi 5: Random one hour urine catecholamine and 4-hydroxy-3-methoxymandelic acid assays for diagnosis of
pheochromocytoma. Clin Chim Acta 1980;103:335-342
32. Laug WE, Siegel SE, Shaw KMF, et al: Initial urinary