A genetic polymorphism of the

N-oxidation of trimethylamine in humans

Trimethylamine (TMA) and its N-oxide (TMAO) are normal components of human urine. They are
present in the diet and also derived from the enterobacterial metabolism of precursors such as choline.
Dietary TMA is almost entirely metabolized to and excreted as TMAO. However, the extent to which
TMA undergoes N-oxidation appears to be polymorphic in a British white population study (n = 169).
Two propositi were identified with relative TMA N-oxidation deficiency that was further confirmed by
oral challenge with TMA (600 mg). The study of the families of the two propositi, as well as those of
two identified subjects with trimethylaminuria, under both normal dietary conditions and after oral TMA
challenge strongly indicates that the conditions of impaired N-oxidation is inherited as a recessive trait.
It is proposed that the N-oxidation of TMA in humans is polymorphic and under single gene diallelic
control in which individuals who are homozygous for the variant allele exhibit marked N-oxidation
deficiency and trimethylaminuria. (CLIN PHARmAcoL THER 1987;42:588-94.)

Makram Al-Waiz, M.B., Ch.B., Riad Ayesh, M.R.C.P., Stephen C. Mitchell, Ph.D.,
Jeffrey R. Idle, Ph.D., and R. L. Smith, D.Sc. London, England
Trimethylamine (N,N-dimethylmethanamine) (TMA)
is a volatile tertiary aliphatic amine with a pungent
ammoniacal odor. It is responsible for the characteristic
odor of rotting fish. The amine is found as such, at low
levels, in fresh marine fish but more particularly in the
form of its N-oxide (TMAO), which may reach concentrations as high as 0.5% of body weight.' Both the
free amine and its TMAO are natural components of
human urine; healthy young volunteers excrete about
1 mg TMA and 40 mg TMAO daily, although these
values are markedly influenced by diet particularly if
marine fish are eaten, When the latter are consumed
several hundred mg TMAO may be excreted.
In humans, TMA and TMAO excretion are believed
to arise from two sources. The first is the TMA and
TMAO present as such in marine fish. The second
source is via the enterobacterial metabolism of precursors such as choline present in foodstuffs such as eggs,
liver, and soybean.'
Studies on the metabolic disposition of TMA have
shown that the orally administered amine is absorbed
rapidly and metabolized almost exclusively to TMAO,

From the Department of Pharmacology, St. Mary's Hospital Medical

School.
Supported by the Wellcome Trust and the Arabian Gulf University,
Bahrain.
Received for publication April 17, 1987; accepted July 2, 1987.
Reprint requests: R. L. Smith, D.Sc., Department of Pharmacology,
St. Mary's Hospital Medical School, London W2 1PG, U.K.

588

which is excreted in the urine.' Secondary pathways of

metabolism, such as N-dealkylation, are of little importance for the biotransformation of this compound in

humans.'
However, there are a number of reports in the literature of individuals who excrete large amounts of unoxidized TMA in their urine. Such individuals may acquire a particularly characteristic and unpleasant smell
thought to be caused by the excretion of TMA in the
sweat and exhaled air. This condition was first described
by Humbert et al.' and is now referred to colloquially
as the fish-odor syndrome or, in more specific biochemical terms, as trimethylaminuria. Since 1970 about
16 further cases have been described in the literature
and there have been suggestions and observations that
the condition may well be an inherited one, but this
has never been established. 6,7
Clinical experience has suggested that this phenomenon, albeit in a milder form, may not be as rare as
the literature would indicate. Consequently, it was decided to carry out a random population study to ascertain the distribution of trimethylamine N-oxidation capacity and, having identified trimethylaminuric probands, to carry out appropriate family studies with a
view to establishing the possible genetic basis and mode
of inheritance of the phenomenon.

METHODS
Human volunteers. One hundred sixty-nine healthy
white volunteers (99 men and 70 women) were recruited

VOLUME 42
NUMBER 5

from the staff and students of St. Mary's Hospital Medical School, London. Their ages ranged between 18 and
50 years (mean - 1 SD 21.5 4.4). No subject had
any history of major organic disease and none consumed more than moderate amounts of alcohol although
36 regularly smoked cigarettes. None of the volunteers
had been exposed to any recent drug treatment and none
were taking medication at the time of the study. All of
the volunteers were asked not to eat fish for the 2 days
before the test and during the urine collection itself.
The families of two subjects in the population study
were also examined. In addition, two proband individuals who had been diagnosed previously as having trimethylaminuria (fish odor syndrome)8 were investigated together with all members of their families. All
of these additional volunteers were in good health.
Full informed consent was obtained from all the participants before entering the study and appropriate ethical approval was provided by the local ethics committee.
Experimental procedure. Each volunteer collected a
total 0 to 24-hour urine sample in a container to which
10 ml hydrochloric acid (4 mol/L) had already been
added. This ensured that any TMA excreted in the form
of its free base would be converted to the nonvolatile
hydrochloride salt. The total urine volume was recorded
and an aliquot (20 ml; pH 2.0) was stored at 20 C
until analysis, which was carried out as soon as possible.
Further investigations were undertaken on all family
members. These individuals were given 970 mg trimethylamine hydrochloride (equivalent to 600 mg
TMA free base; Aldrich Chemical Co. Ltd., Gillingham, Dorser, England) in gelatin capsules together with
a glass of water, and the following 0 to 8-hour urine
samples were collected and treated as above.
Certain precautions must be taken during the selection of volunteers because spurious high levels of free
TMA may be recorded during menstruation from the
endometrial tissue9 and in those subjects who have vaginal or urinary tract infections caused by bacterial mediated metabolism.' Such individuals were identified
and excluded from the study.
Analytic procedure. TMA in the urine was analyzed
by a modification of the head-space GC technique as
described previously in the literature."' Urine (5 ml)
and 0.2% v/v aqueous isopropylamine (100 p.A; internal standard) were placed in a 15 ml screw-top septum glass vial. Potassium carbonate (2 to 3 gm) and
potassium hydroxide (five to 10 pellets) were added to
the urine and the vials were sealed with air-tight
polytetrafluoroethylene-lined septum caps. All of these

Trimethylamine oxidation polymorphism

589

procedures were carried out on ice. The vials were then

vortex mixed and heated at 90 C for 20 to 30 minutes
in an aluminum heating block and an aliquot (2 ml) of
the head-space gas generated was injected directly onto
the GLC column with an air-tight plastic disposable
syringe.
TMAO was quantitated as TMA after chemical reduction.' Urine (2 ml) was reduced by titanous sulfate
(0.2 ml 15% [w/v] in 23% tw/v] aqueous H2SO4) in a
septum vial at 30 C for 30 minutes. Samples of the
reduced urine were then diluted with water and analyzed
as described above. All urine samples were analyzed
in duplicate.
The chromatography was performed on a PU 204
model gas chromatograph (Pye Unicam, Cambridge,
U.K.) with a "flame ionization" detector, the injector
port being held at 130 C and the detector at 200 C.
The column (170 cm by 0.4 cm inside diameter) was
packed with 4% (w/w) Carbowax 20 mon/0.8%
(w/w) KOH on Carbopack B graphitized carbon support
(60 to 80 mesh) (Supelco, Inc., Bellefonte, Pa.) with
an oven temperature of 70 C and a nitrogen carrier gas
flow rate of 40 ml/min.
The areas of the relevant peaks were calculated by a
Spectra-Physics 4270 integrator (Pye Unicam) and the
results were expressed in terms of micromoles of TMA
and TMAO excreted in 0 to 24-hour urine samples. In
addition, the percentage of total TMA-related material
excreted in the form of the N-oxide was calculated as
free TMA/total TMA x 100.
follows: total TMA
A metabolic ratio for N-oxidation was also calculated
as TMA (in micromoles)/TMAO (in micromoles) excreted in the 0 to 24-hour urine samples. This ratio was
used to separate more clearly those few individuals who
produced relatively smaller amounts of the TMAO from
the rest of the population.

RESULTS
Analysis. According to the conditions described for
the assay, the retention times obtained for TMA and
the internal standard were 1.9 and 3.2 minutes, respectively, and could be clearly resolved from other
volatile urinary components including ammonia (retention time 0.5 minute), methylamine (retention time 0.9
minute), dimethylamine (retention time 1.5 minutes),
acetone (retention time 4.0 minutes), and ethanol (retention time 4.4 minutes) ( Fig. 1). The limit of detection of TMA was 0.01 p.,g/m1 and calibration curves
for the assay were constructed in the range of 0.2 to
10 ii,g/m1 (Fig. 2). The same urine samples from three
individuals were analyzed on six separate occasions during the course of 1 week and the values obtained for

CLIN PHARMACOL THER

NOVEMBER 1987

590 Al-Waiz et al.

1. 5

4
CONCENTRATION

(#9/m1)

Fig. 2. Calibration curve for the estimation of TMA.

6
MINUTES

Fig. 1. Gas chromatogram of urine from a normal subject

shows peaks for ammonia (A), monomethylamine (B), dimethylamine (C), TMA (D), isopropylamine (E), acetone (F),
and ethanol (G).

TMA and total TMA levels (after N-oxide chemical reduction) were virtually the same (coefficient of variation
3.2 to 6.9). The efficiency of TMAO chemical reduction using the system described was always 98% to
99%.
Repeat analysis of urine samples from 20 subjects
after storage at 20 C for up to 6 months showed
virtually no difference. Comparison of the initial esti-

mates of TMA and TMAO with the repeat values obtained showed no significant difference (Student t test:
t = 1.01, P = 0.324; t = 0.86, P = 0.401, respectively).
Biologic reproducibility. The values obtained for the
percentage of total TMA-related material excreted as
TMAO were found to be reproducible in 24 subjects in
whom complete repeat tests were performed on two
separate occasions at least 6 months apart. When the
first estimate of the percentage of TMAO was compared
with the second estimate, the correlation coefficient
obtained was r = 0.998 and P < 0.001 (Fig. 3). In
addition, estimates were made on two subjects on five
separate occasions over a period of several months and
the values obtained for the percentage of TMAO were
virtually identical (coefficient of variation <1%) (Table I).
Interindividual variation. The results obtained for
the 169 subjects investigated are given in Table II.* The
mean values found for 0 to 24-hour urinary excretion
of TMA and TMAO (mean SD) were 20.3 20.9
Rmol (1.2 1.2 mg) and 505.6 329.7 Kmol
(38.0 24.8 mg), respectively. When the results were
plotted in the form of a histogram relating the percent
*See NAPS document No. 04527 for four pages of supplementary
material. Order from NAPS, c/o Microfiche Publications, PO Box
3513, Grand Central Station, New York, NY 10163-3513. Remit in advance, in US funds only, $7.75 for photocopies or
$4.00 for microfiche. Outside the United States and Canada,
add postage of $4.50 for the first 20 pages and $1.00 for each
of 10 pages of material thereafter, or $1.50 for microfiche
postage.

VOLUME 42
NUMBER 5

Trimethylamine oxidation polymorphism

591

100

80

98

'G.

96

94

92

20

90,
20

40

60

80

100

90

94

92

95

98

100

(FIRST ESTIMATE)
% OF

TOTAL TMA URINARY EXCRETION

AS

TMAO

Fig. 3. Biologic reproducibility of the N-oxidation of TMA; correlation of first and second estimates
of urinary TMAO excretion for 24 volunteers (A); expanded correlation scale for 20 subjects (B).

of total TMA excretion as TMAO to the number of

volunteers, it became apparent that the majority of subjects (167 or 98.8%) lay within one major mode with
values ranging from 89.3% to 99.7% TMAO. However,
two individuals displayed a relative impairment in
TMAO production and lay outside the major mode, with
values of 68.9% and 84.1%, respectively (Fig. 4).
When the data were expressed in terms of a metabolic
ratio that considerably skewed the values, giving relatively more importance to those individuals who may
produce lower amounts of TMAO, these two subjects
were obviously separated from the rest of the population
(Fig. 5). Repeat testing of these two individuals (one
man and one woman) on separate occasions gave the
same results (Table I).
The two subjects previously diagnosed as having trimethylaminuria were shown to be severely impaired in
their ability to produce TMAO and to have values of
10.5% and 22.2% TMAO, respectively. These subjects were added to the distribution histogram for comparison.
Family studies. To investigate the possible role of a
genetic influence as a determinant of the relative impairment of TMAO production, all members of the families of the two propositi identified from the population
study and the entire families of the two previously diagnosed trimethylaminurics were investigated (Fig. 6).
Each subject was investigated with respect to the urinary excretion of TMA and TMAO under normal dietary conditions as described above, whereas for certain
subjects the excretion of these two substances was determined after oral challenge with 970 mg trimethylamine hydrochloride (equivalent to 600 mg TMA base).

Table I. Biologic reproducibility of TMA

N-oxidation among four volunteers
Urinary excretion
(Amol124 hr)

% Total TMA

excretion as

Subject
1

TMA

TMAO

TMAO

10.3
18.5

408
778.2
592.3
458
307.1
536.0
385.4
545.0
236.3
467.0
289.2
247.7
455.2
842.4
341.9

97.5
97.7
98.6
98.3
97.3
94.0
94.6
94.9
95.8
93.6
68.9
68.3
72.3

8.6
7.7
8.5
34.5
22.0
29.0
10.4
31.8

130.6

115.1
174.3
159.1

66.4

84.1

83.7

This test has previously been found to be useful in

characterizing "carrier" individuals with an impaired
N-oxidation capacity that was not evident under the
TMA load derived from dietary origin.8"4 The values
in Fig. 6 show the percentage of the total TMA excretion for each individual in the form of TMAO. The
values in parentheses refer to the same parameter as
determined after oral challenge with TMA.
Family A. Family A consisted of three sisters; two
exhibited severe N-oxidation deficiency and a condition
of trimethylaminuria. Both parents and the third sister

CLIN PHARMACOL THER

NOVEMBER 1987

592 Al-Waiz et al.

60

30

1.

48

24

36

16

24

12

12

60

Cm
0

10

20

30

40

50

60

70

80

11

90

100

% OF TOTAL TMA URINARY EXCRETION AS TMAO

Fig. 4. Population and frequency distribution in the urinary

excretion of TMAO as a proportion of total TMA elimination
for 169 normal healthy volunteers (0). Values for two subjects
with trimethylaminuria
have been added for comparison.

()

are apparently unaffected. However, both parents

showed a significant deficiency in N-oxidation capacity
after oral challenge with TMA. Under normal dietary
conditions, TMAO excretion accounted for 95% and
98% of TMA elimination for the father and mother,
respectively. After TMA challenge, this declined to
78% and 79%, respectively, with a marked increase in
the urinary levels of unoxidized TMA. However, the
unaffected daughter showed virtually no change on
TMA challenge. The clustering together in a single
family of two affected individuals together with an unaffected daughter is suggestive of an inherited trait.
Furthermore, the TMA challenge test indicates that both
parents can be expected to carry a relative deficiency
for N-oxidation and may be considered, on the basis of
a single gene hypothesis, to be "carriers" or heterozygotes whereas the two affected daughters would be
considered to be homozygous affected.
Family B. Family B consisted of three siblings, one
affected and two unaffected, and the two parents. Oral
TMA challenge to both parents significantly reduced
urinary TMAO excretion from 97% under normal dietary conditions to about 75% after challenge, suggesting both to be heterozygous carriers. The affected
sibling displayed major impaired N-oxidation capacity
and marked trimethylaminuria.
Family C. Family C consisted of three siblings, two
daughters and their propositus brother, and the two parents. The father showed no impairment of N-oxidation

0.2

0.33

0.4

0.5

TMA/TMAO RATIO

Fig. 5. Population frequency distribution for the N-oxidation

metabolic ratio (urinary TMA/TMAO excretion in
iimol/24 hr) for 169 normal healthy subjects.

both under normal dietary conditions and after TMA

challenge. By contrast the mother showed evidence of
impairment both under normal dietary conditions and
after TMA challenge. However, the impairment was
not of the same degree as that seen with the three homozygous affected subjects from families A and B and
it was concluded that the mother was heterozygous.
The male propositus of this family showed on TMA
challenge increased impairment of N-oxidation capacity
as did one of the two sisters. Therefore it was concluded
for this family that the mother, the propositus son, and
one of the two daughters were heterozygotes with respect to TMA N-oxidation capacity.
Family D. Family D was comprised of the female
propositus, two brothers, and two parents. The propositus and the mother showed relatively impaired
N-oxidation capacity both under normal dietary conditions and more markedly after oral TMA challenge.
Both the father and one brother showed unimpaired
N-oxidation capacity under both conditions whereas the
second brother showed impairment of N-oxidation only
after oral TMA challenge. It was concluded therefore
that the propositus, one brother, and the mother were
heterozygous for the condition of impaired metabolism
whereas the father and second brother were assigned
the homozygous unaffected genotype.

DISCUSSION
It can be seen from the results presented in this article
that the majority of the population excrete TMA, derived from the diet almost entirely in the form of

VOLUME 42
NUMBER 5

Trimethylamine oxidation polymorphism

Family A

Family

95

98

97

76

(78)

(79)

(96)

(61)

593

97

14

441:2

(91)

95

96

84

(75)

HOMOZYGOUS AFFECTED PROpOSITUS

HETEROZYGOUS AFFECTED CARRIER PROPOSITLIS

Family B
97
(75)

(69)

Family D
98

98

(78)

(96)

76
(62)

94

99

HOMOZYGOUS AFFECTED PROPOSITUS

4
95

69

94

(92)

(52)

(78)

HETEROZYGOUS AFFECTED CARRIER PROPOSITUS

Fig. 6. Family studies on the urinary excretion of TMA and TMAO under normal dietary conditions
and after oral challenge with TMA (600 mg). Values for the latter are shown in parentheses. All
figures refer to percent of the total urinary TMA excretion as TMAO. Affected probands are indicated
by arrowheads.

TMAO. Furthermore, it has been found that the ability

of individuals to N-oxidize TMA under normal, consistent dietary conditions, which excluded fish, is a
reproducible characteristic as shown by the stability of
the urinary TMA/TMAO ratio. However, from the population study two individuals were identified as being
relatively deficient in TMA N-oxidation showing that
there exists interindividual variation with respect to the
occurrence of this reaction. Furthermore, this deficiency is a stable biologic characteristic as shown by
the reproducibility of the ratio of urinary TMA/TMAO.
The family data for these two individuals and the two
subjects with trimethylaminuria investigated provide
the strongest evidence available that the impaired
N-oxidation of TMA and the associated syndrome of
trimethylaminuria is an inborn error of metabolism. It
is proposed that there exists a genetic polymorphism
for the N-oxidation of TMA in humans. The data are
consistent with the hypothesis of a diallelic single gene
mechanism governing the N-oxidation of TMA. One
allele would be responsible for determining extensive
N-oxidation capacity for TMA whereas the other variant
allele would be responsible for determining recessive
impaired N-oxidation. However, final proof of this genetic hypothesis must await the isolation and characterization of the specific gene product, which is pre-

sumably a TMA-oxidase enzyme. Also from this study

it would appear that the frequency of carriers may be
at least 1% in the British white population. This value
may in fact be higher because it is not known what
proportion of heterozygotes may be concealed in the
major mode of the population distribution.
Two further points to emerge from this study
are worthy of comment. First is the principle that
homozygous-affected individuals can be recognized on
the basis of the measurement of the ratio of urinary
TMA/TMAO derived under normal dietary conditions.
In other words, one can use the normal dietary derived
burden of TMA for phenotyping, and the administration
of an exogenous probe drug is not necessary for this
purpose. The second point is that the oral challenge test
with TMA (600 mg) is a very useful diagnostic test for
identifying carriers of the N-oxidation defect.
Previous studies with chickens' have shown that
within one strain there are two groups known as "tainters" and "nontainters" according to the free TMA in
the egg yolk when they were fed a diet containing rape
seed meal, which is a rich source of TMA in the form
of sinapine, choline, and TMAO. This fishy taint was
the result of a gross impairment of the chickens' ability
to convert TMA to TMAO. Further observations' have
demonstrated that the capacity of the fowl for synthe-

594 Al-Waiz et al.

sizing TMA oxidase is inherited and is consistent with
the involvement of a single, autosomal, semidominant
gene. These studies suggest that the variation in TMA
N-oxidation occurs in other animal species and is not
confined only to humans.
The process of TMA N-oxidation takes place mainly
in the liver' and to a minor extent in the kidney:8
Because TMA is a nucleophilic alkylamine, the enzyme
concerned in its N-oxidation is believed to be microsomal flavin adenine dinucleotide-containing monooxygenase:9 Higgins et al." have shown that there is
a major impairment in vitro of TMA N-oxidation by
liver tissue obtained from a patient with trimethylaminuria, thus indicating the presence of a defective
form or frank absence of the enzyme.
Many substances and drugs such as nicotine," nicotinamide, guanethidine, metyrapone," and pinacidil"
are known to undergo N-oxidation, but less extensively
than TMA. Further studies would be suggested to investigate whether the metabolic transformation of these
substances may be deficient in individuals with impaired TMA N-oxidation.

References
Shewan JM. The chemistry and metabolism of the nitrogenous extractives in fish. Biochem Soc Symp
1951;6:28-48.
Tver DF, Russell P. Nutrition and health encyclopedia.
New York: Van Nostrand Reinhold, 1981:100.
Al-Waiz M, Mitchell SC, Idle JR, Smith RL. The metabolism of '4C-labeled trimethylamine and its N-oxide
in man. Xenobiotica 1987;17:551-8.
Al-Waiz M, Mitchell SC, Idle JR, Smith RL. The relative
importance of N-oxidation and N-demethylation in the
metabolism of trimethylamine in man. Toxicology
1987;43:117-21.
Humbert JR, Hammond KB, Hathaway WE, Marcoux
JG, O'Brien D. Trimethylaminuria: the fish-odour syndrome. Lancet 1970;ii:770-1.
Brewster MA, Schedewie H. Trimethylaminuria. Ann
Clin Lab Sci 1983;13:20-4.
Spellacy E, Watts RWE, Goolamali SK. Trimethylaminuria. J Inherited Metab Dis 1979;2:85-8.
Al-Waiz M, Ayesh R, Mitchell SC, Idle JR, Smith RL.
Trimethylaminuria (fish-odour syndrome): an inborn error of oxidative metabolism. Lancet 1987;i:634-5.
Windholz M, ed. The Merck index, ed 10. Rahway, NJ:
Merck and Co, 1983:1387-8.

GUN PHARMACOL THER

NOVEMBER 1987

Brand JM, Galask RP. Trimethylamine: the substance

mainly responsible for the fishy odor often associated
with bacterial vaginosis. Obstet Gynecol 1986;68:682-5.
Brewster M, Schedewie H, MacDonald JA. Rapid identification of trimethylaminuria as a cause of body odor.
Clin Chem 1980;26:1011.
Lowis S, Eastwood MA, Brydon WG. The measurement
of methylamines in biological material using a gas chromatographic head space gas technique. J Chromatogr
1983;278:139-43.
Simenhoff ML, Dunn SR, Asatoor A, Milne MD. Determination of trimethylamine N-oxide in urine. Proc Am
Chem Soc 1977;57.
Al-Waiz M, Mitchell SC, Ayesh R, Idle JR, Smith RL.
Variable N-oxidation of trimethylamine in man. Br J Clin
Pharmacol 1987;23:614-5P.
Pearson AW, Butler EJ, Curtis F, Fenwick GR, HobsonFrohock A, Land DG. Effect of rapeseed meal on hepatic
trimethylamine oxidase activity in the domestic fowl in
relation to egg taint. J Sci Food Agric 1979;30:291-8.
Pearson AW, Butler EJ. Effect of selective breeding and
age on the ability of the domestic fowl (Gallus domesticus) to oxidize trimethylamine. Comp Biochem Physiol
1983;76C:67-74.
Baker JR, Struempler A, Chaykin S. A comparative
study of trimethylamine N-oxide biosynthesis. Biochem
Biophys Acta 1963;71:58-64.
Acara M, Camiolo S, Rennick B. Renal N-oxidation of
trimethylamine in the chicken during tubular excretion.
Drug Metab Dispos 1977;5:82-90.
Ziegler MD, Mitchell HC. Microsomal oxidase IV: properties of a mixed function amine oxidase isolated from
pig liver microsomes. Arch Biochem Biophys 1972;150:
116-25.
Higgins T, Chaykin S, Hammond KB, Humbert JR. Trimethylamine N-oxide synthesis: a human variant. Biochem Med 1972;6:392-6.
Beckett AH, Gorrod JW, Jenner P. The analysis of
nicotine- -N-oxide in urine, in the presence of nicotine
and cotinine, and its application to the study of in vivo
nicotine metabolism in man. J Pharm Pharmacol 1971;
23:555-615.
Oelschlager H, Al Shaik M. Metabolic N-oxidation of
alicyclic amines. In: Gorrod JW, Damani LA, eds. Biological oxidation of nitrogen in organic molecules.
Chichester: Ellis Horwood 1985:66-75.
Eilertsen E, Magnussen MP, Petersen HJ, RastrupAndersen N, Sorensen H, Arrigoni-Martelli E. Metabolism of new antihypertensive agent pinacidil in rat, dog
and man. Xenobiotica 1982;12:187-96.