APMIS 113: 513–6, 2005 Copyright C APMIS 2005

Printed in Denmark . All rights reserved

ISSN 0903-4641

Trimethylamine and trimethylamine oxide levels in

normal women and women with bacterial vaginosis reflect a
local metabolism in vaginal secretion as compared to urine
H. WOLRATH,1 B. STÅHLBOM,2 A. HALLÉN3 and U. FORSUM1
1
Division of Clinical Microbiology, 2Division of Occupational and Environmental Medicine, Department of
Molecular and Clinical Medicine, Linköping University and 3Department of Dermatology and
Venereology, University Hospital, Uppsala, Sweden

Wolrat H, Ståhlbom B, Hallén A, Forsum U. Trimethylamine and trimethylamine oxide levels in

normal women and women with bacterial vaginosis reflect a local metabolism in vaginal secretion as
compared to urine. APMIS 2005;113:513–6.
The smell of rotten fish is one of the characteristics of bacterial vaginosis (BV), and is due to trimethyl-
amine (TMA). Trimethylamine can be found in human urine, although most of it occurs as the
nonvolatile oxide (TMAO) form. The fraction TMA/TMAO can be expected to be the same in differ-
ent body fluids if no local production of TMA occurs. In women with BV, TMAO in the vaginal fluid
is expected to be chemically reduced by the local bacterial flora to the much more odorous TMA. We
have therefore studied the local vaginal production of TMA in vaginal secretion compared to the
general TMA-TMAO metabolism that was measured in urine using gas chromatography. Both vaginal
fluid and random urine samples were collected from women, with and without BV, attending a Swedish
clinic for sexually transmitted diseases, and these samples were analyzed for TMA and TMAO. The
results show that a local production of TMA occurs in the vagina that is not part of the general
metabolism of TMA-TMAO.
Key words: TMA; TMAO; bacterial vaginosis; urine.
U. Forsum, Division of Clinical Microbiology, Universitetssjukhuset, S-581 85 Linköping, Sweden.
e-mail: urban.forsum/imk.liu.se

Trimethylamine (TMA) is a pungent volatile amounts of non-oxidized TMA, is excreted

amine that is responsible for the characteristic
odor of rotten fish. Fresh marine fish contain
low levels of the amine but higher levels of its
N-oxide form, trimethylamine oxide (TMAO).
It is well known that TMA is normally formed
in the human body, mainly from dietary chol-
ine, where eggs, liver and soybeans are the main
sources. Intestinal bacteria also form TMA
from TMAO, which is present in seafood. TMA
is absorbed from the gut and re-oxidized in the
liver to TMAO, which, together with small
Received 2 November 2004.
Accepted 14 February 2005.
through the urine.
Healthy young adults excrete about 1 mg of
TMA and 40 mg of TMAO daily, although
these levels are markedly influenced by diet, par-
ticularly when containing marine fish (1). When
marine fish is a dietary component, several hun-
dred mg of TMAO may be excreted (2). How-
ever, in some individuals the N-oxidation pro-
cess is impaired, resulting in the excretion of
higher levels of TMA through urine, breath,
sweat and vaginal secretion (2, 3). These find-
ings were first described by Humbert et al. (4),
and are now referred to as the ‘‘fish odor syn-
drome’’ or trimethylaminuria.
513
 WOLRATH et al.
TABLE. TMA and TMAO content of vaginal fluid and urine in normal women and women with bacterial vaginosis
Diagnosis TMA (mg/g vaginal fluid)1 TMA (mg/g urine)2
Range Median Mean (SD) Range
BV (nΩ20) 0–62.5 22.4 24.5 (18) 0.06–1.48 0.18
Normal (nΩ17) 0–14.8 1.8 4.0 (4.5) 0–0.65
TMA is also responsible for the fishy odor
characteristic of bacterial vaginosis (BV). Crud-
en et al. (5) have shown that a few isolates of
Mobiluncus spp. can reduce TMAO to TMA,
indicating that vaginal TMA is produced by an-
aerobic bacteria occurring in elevated numbers
in BV. The results of the so-called sniff test, an
often used diagnostic procedure for BV, could
thus be influenced by TMA produced in the gut
and liver metabolism of TMA-TMAO. In this
study we examined the relationship between lo-
cal vaginal production of TMA from TMAO
and the general TMA-TMAO metabolism that
is reflected in the urinary secretion of these
amines. Our results indicate that in women at-
tending a sexually transmitted disease (STD)
clinic, the vaginal TMA level mainly reflects loc-
ally produced amine.
MATERIALS AND METHODS
Women of childbearing age (nΩ37, age 20–40 years,
mean 27) who attended an STD clinic in Uppsala,
Sweden, during a 2-month period volunteered to par-
ticipate in this study. The clinic attracts women with
any kind of lower genital tract infection, not only
STDs. Twenty-two women (60%) were taking oral
contraceptives, and four had an IUD.
A glass vial together with a plastic loop was weigh-
ed using a 0.0001 g scale. Vaginal samples from the
lateral fornix were taken with the plastic loop, which
was then placed in the vial and the new weight was
noted. Two ml of 1 M HCl was added to the sample
before storage at ª20 æC until analysis. An additional
sample of vaginal fluid was taken from the lateral
fornix and smeared onto a glass slide, air-dried, and
saved for later batch-wise Gram staining. The Gram-
stained smears were scored according to Nugent’s cri-
teria (6), in a fully blinded manner by another mem-
ber of the research group, after the TMA determi-
nations. At the STD clinic a preliminary diagnosis
was made according to Hallén et al. (7), based on
pH, the sniff test and clue cells in a wet mount of
vaginal fluid. The patients were also screened for
Chlamydia trachomatis and Candida.
Random urine samples from all 37 women were
collected in glass vials and acidified with concen-
514
TMAO (mg/g vaginal fluid)2
Median Mean (SD) Range Median Mean (SD)
0.53 (1.5) 0–209 5.6 25.5 (49)
0.12 0.2 (0.17) 0–38.3 4.5 6.75 (8.9)
trated HCl (37%, 2 ml per 100 ml), and stored at
4 æC until analyzed.
Analysis of TMA and TMAO in vaginal samples
and urine samples was performed by gas chromato-
graphy (8). TMA fractions are presented as percent
of the total amount of TMA and TMAO. If no
TMAO was present, TMA was set at 100%.
Statistical methods
Student’s t-test and the Wilcoxon rank-sum test
were used.
RESULTS
Eight of the thirty-seven participating women
had a diagnosis of Chlamydia or Candida infec-
tion. Nugent scoring revealed BV in 20 women
(Nugent scores of 7–10) and normal flora in 17
(Nugent scores of 0–4). The Hallén method id-
entified 18 BV cases out of the 20 that were di-
agnosed by Nugent scoring. Two cases identified
as BV by the Hallén method were not verified
as BV by Nugent scoring.
The concentrations of TMA and TMAO in
vaginal fluid and urine in women diagnosed
with BV and in those diagnosed as normal ac-
cording to Nugent scoring are presented in the
table. There was a significantly higher concen-
tration of TMA in the vagina of women with
BV than in normals, and a significantly higher
fraction of TMA out of total TMA-TMAO in
the vagina as compared to urine. The patient
groups had similar levels of TMAO in the urine
and in the vagina. Thus, in women without BV,
we found TMA in the range 0–100% (median
25%) of the total TMA-TMAO concentration
in vaginal fluid, and in the range 0.23–6.7%
(median 1.0%) in urine. In women diagnosed
with BV, we found TMA in the range 0–100%
(median 65%) of the total TMA-TMAO con-
centration in vaginal fluid and in the range 0–
5% (median 0.75%) in urine. Two women with-
out BV had detectable levels of TMA in urine,
1.6% and 1.2%, respectively, but not in vaginal
fluid (0%). In three women without BV, all
 VAGINAL TMAO IN BACTERIAL VAGINOSIS
TABLE continued
2
TMAO (mg/g urine) TMA/TMAπTMAO (% vaginal)2
Range Median Mean (SD) Range Median
3.83–251 20.7 33.6 (54) 0–100 65
4.95–61.43 23.5 26.2 (16.7) 0–100 25
1
(p⬍0.01), 2BV vs normal and vaginal vs urine N.S.
TMA-TMAO was present as TMA in vaginal
fluid. In two women with BV, no TMA was
present in the vaginal fluid. These women had
1.22% and 0.33%, respectively, of the total
TMA -TMAO concentration in the form of
TMA in urine.
DISCUSSION
The so-called sniff test, when KOH is added to
vaginal secretions to release the odor of TMA,
has proven to be an important and expedient
criterion for diagnosing BV. However, it is un-
known whether the content of TMA in vaginal
secretion can also be influenced by diet, as is the
case for urine. Furthermore, it is reasonable to
assume that there is an individual variation in
the TMA N-oxidation capacity that is revealed
in the TMA-TMAO relationship in the vaginal
secretion. We have previously shown that TMA
can be present in the vaginal fluid of women
who do not have BV, but in amounts below 5
mg/g of vaginal fluid (9). The TMA content in
vaginal fluid can therefore be due to individual
variations in diet, the BV-related reduction of
TMAO to TMA, and, possibly, the individual
TMA N-oxidation capacity.
Sardas et al. (3) showed that women with a
fishy-smelling discharge, independent of diag-
nosis, had a significantly higher concentration
of TMA in the urine compared to the control
group. However, the concentration of TMA in
the vaginal secretion of these women was not
investigated and there are no published reports
of the relation between TMA-TMAO in vaginal
secretion compared to TMA-TMAO in urine in
women with BV.
In the present study, we performed Nugent
scoring and compared it to the diagnostic test
at the STD clinic, and investigated the concen-
trations of TMA and TMAO in both vaginal
fluid and urine. In women with and without BV,
TMA/TMAπTMAO (% in urine)2
Mean (SD) Range Median Mean (SD)
61 (1.1) 0–4.95 0.75 0.9 (1.0)
43 (1.8) 0.23–6.7 1.0 0.23 (1.4)
the percentage of TMA of the total TMA-
TMAO was significantly higher in vaginal fluid
than in urine. As expected and also shown by
us in earlier studies, the percentage of TMA in
vaginal fluid was significantly higher in women
with BV than in normals. However, in the pres-
ent study we detected TMA concentrations as
high as 15 mg/g vaginal fluid in women without
BV, in contrast to our previous study where the
concentrations were less than 5 mg/g of vaginal
fluid. This might reflect the sample of women
studied. Caution should be exercised in inter-
preting these differences before more data on
dietary and metabolic status of included women
have been collected.
Interestingly, two women diagnosed with BV
had no TMA in the vaginal fluid. We also found
three women without BV who had only TMA
in the vaginal fluid. It is possible that the Nug-
ent score may not have reflected the true BV
state of these women. The main outcome of our
study is that in women with BV there is a local
production of TMA in the vagina. However, it
is possible that TMA levels in vaginal fluid are
also influenced by liver and gut flora met-
abolism, which in turn might affect the diag-
nosis of bacterial vaginosis when the whiff test
is used as criterion. It thus seems important to
study the TMAO content in vaginal fluid and
the probable bacterial source of local produc-
tion of TMA, and to confirm the findings of
this study in well-characterized populations that
reflect the various dietary sources and metabolic
determinants of TMA/TMAO.
We would like to thank the staff of the STD depart-
ment at the University Hospital in Uppsala for all
their help with the patient samples. The research was
approved by the institutional review board, Uppsala.
We also thank Bodil Carlsson for Gram staining all
smears and for scoring of the smears. This research
was supported by an ALF grant from the University
Hospital in Linköping.
515
 WOLRATH et al.
REFERENCES
1. Tver DF, Russell P. Nutrition and Health Encyc-
lopedia. New York: Van Nostrand Reinhold,
1981:100.
2. Al-Waiz M, Ayesh R, Mitchell SH, Idle JR,
Smith RL. A genetic polymorphism of the N-
oxygenation of trimethylamine in humans. Clin
Pharmacol Ther 1987;42:588–94.
3. Sardas S, Akyol D, Green RL, Mellon T,
Gökmen O, Cholerton S: Trimethylamine N-oxi-
dation in Turkish women with bacterial vag-
inosis. Pharmacogenetics 1996;6:459–63.
4. Humbert JR, Hammond KB, Hathaway WE,
Marcoux JG, O’Brien D. Trimethylaminuria: the
fish-odour syndrome. Lancet 1970;ii:770–1.
5. Cruden DL, Galask RP. Reduction of trimethyla-
mine oxide to trimethylamine by Mobiluncus
516
strains isolated from patients with bacterial vag-
inosis. Microb Ecol Health Dis 1988;1:95–100.
6. Nugent RP, Krohn MA, Hillier SL. Reliability of
diagnosing bacterial vaginosis is improved by a
standardized method of Gram stain interpreta-
tion. J Clin Microbiol 1991;29:297–301.
7. Hallen A, Påhlson C, Forsum, U. Bacterial vag-
inosis in women attending STD clinic: diagnostic
criteria and prevalence of Mobiluncus spp. Geni-
tourin Med 1987;63:386–9.
8. Lundh T, Ståhlbom B, Åkesson B. Methods for
the determination of dimethylethylamine and di-
methylethylamine-N-oxide in air, plasma and
urine samples. Intern J Environ Anal Chem
1991:48;203–7.
9. Wolrath H, Borén H, Hallén A, Forsum U. Tri-
methylamine content in vaginal secretion and its
relation to bacterial vaginosis. APMIS 2002;110:
819–24.