Catlin DH, Leder BZ, Ahrens B, Starcevic B, Hatton CK, Green GA, Finkelstein JS. Trace Contamination of Over-the-Counter Androstenedione and Positive Urine Test Results for a Nandrolone Metabolite. JAMA. 2000;284(20):2618-2621. doi:10.1001/jama.284.20.2618
Author Affiliations: Departments of Molecular and Medical Pharmacology (UCLA Olympic Analytical Laboratory) (Drs Catlin and Hatton and Messrs Ahrens and Starcevic), Medicine (Dr Catlin), and Family Medicine (Dr Green), University of California, Los Angeles; and Endocrine Unit, Department of Medicine, Massachusetts General Hospital, Boston (Drs Leder and Finkelstein).
Context Several anabolic steroids are sold over-the-counter (OTC) in the United
States, and their production is not regulated by the US Food and Drug Administration.
Reports have suggested that use of these supplements can cause positive urine
test results for metabolites of the prohibited steroid nandrolone.
Objectives To assess the content and purity of OTC androstenedione and to determine
if androstenedione and 19-norandrostenedione administration causes positive
urine test results for 19-norandrosterone, a nandrolone metabolite.
Design Randomized controlled trial of androstenedione, open-label trial of
19-norandrostenedione, and mass spectrometry of androstenedione preparations,
conducted between October 1998 and April 2000.
Setting Outpatient facility of a university hospital.
Participants A total of 41 healthy men aged 20 to 44 years.
Intervention Participants were randomly assigned to receive oral androstenedione,
100 mg/d (n = 13) or 300 mg/d (n = 11) for 7 days, or no androstenedione (n
= 13); in addition, 4 patients received 10 µg of 19-norandrostenedione.
Main Outcome Measures Content of OTC androstenedione preparations; level of 19-norandrosterone
in urine samples, determined by mass spectrometry, compared among the 3 randomized
groups at day 1 and day 7, and among the participants who received 19-norandrostenedione
from October 1998 to April 2000.
Results All urine samples from participants treated with androstenedione contained
19-norandrosterone, while no samples from the no-androstenedione group did.
Urinary concentrations were averaged for day 1 vs day 7 measurements; mean
(SD) 19-norandrosterone concentrations in the 100-mg/d and 300-mg/d groups
were 3.8 (2.5) ng/mL and 10.2 (6.9) ng/mL, respectively (P = .006). The 19-norandrosterone content exceeded the cutoff for reporting
positive cases (>2.0 ng/mL) in 20 of 24. The androstenedione preparation used
was pure at a sensitivity of 0.1%, but at 0.001% 19-norandrostenedione was
found. For the 4 participants to whom 10 µg of 19-norandrostenedione
was administered, 19-norandrosterone was found in all urine samples. Of 7
brands of androstenedione analyzed at the 1% level, 1 contained no androstenedione,
1 contained 10 mg of testosterone, and 4 more contained 90% or less of the
amount stated on the label.
Conclusion Our study suggests that trace contamination of androstenedione with
19-norandrostenedione is sufficient to cause urine test results positive for
19-norandrosterone, the standard marker for nandrolone use. Oral steroid doses
as small as 10 µg are absorbed and excreted in urine. Some brands of
androstenedione are grossly mislabeled. Careful analysis of androstenedione
preparations is recommended in all studies of its biological effects.
The dietary supplement industry generates annual sales estimated at
US $12 billion,1 and in 1998 up to $800 million
of this was spent on "sports supplements."2
An important consequence of the 1994 Dietary Supplement Health and Education
Act3 is that several androgenic steroids, eg,
androstenedione, dehydroepiandrosterone, 19-norandrostenedione, and 19-norandrostenediol,
became widely available over-the-counter (OTC) in the United States and via
the Internet. This act enables companies to market steroids as nutritional
supplements if they do not claim to diagnose, prevent, or cure disease, in
which case they escape US Food and Drug Administration regulation. Thus, content
and purity of the ingredients are left to the manufacturer.4
Athletes take OTC steroids for their alleged muscle-building properties.
Often the package labels do not warn that they may contain substances that
are banned by sports organizations and may cause a positive urine test result.
Recently, there has been an increase in the number of athletes who have
tested positive for nandrolone, a banned steroid popular in the early years
of athletic doping. Since nandrolone itself is rarely found in urine, these
cases are actually reports of findings of urinary 19-norandrosterone, the
major but inactive metabolite of nandrolone (19-nortestosterone). Administration
of nandrolone fell out of style in the 1980s because athletes know that it
is readily detectable in urine for weeks.5
Our laboratory conducts approximately 15,000 tests per year and reported 10,
23, and 35 cases where 19-norandrosterone was found in urine in 1997 to 1999,
19-Norandrosterone is also found in urine after ingestion of the OTC
steroid 19-norandrostenedione.6 In sport, the
origin of urinary 19-norandrosterone is moot because many sports organizations
prohibit and test for OTC steroids in addition to controlled steroids. 19-Norandrosterone
traces may occur naturally in the urine of untreated males7,8;
therefore, the International Olympic Committee set cutoffs in 1998 at 2 ng/mL
for males and 5 ng/mL for females.
We recently reported that oral androstenedione administration increases
serum testosterone and estradiol levels in young men.9
Because of reports suggesting that dietary supplement use can cause urine
test results positive for 19-norandrosterone, we analyzed these men's urine
samples for 19-norandrosterone. We also assessed the effect of administration
of 10 µg of 19-norandrostenedione on urinary 19-norandrosterone excretion
and examined the purity and content of commercial androstenedione preparations.
The subjects (n = 37, aged 20-40 years) and protocol have been described
previously.9 The men were randomly assigned
to receive no androstenedione (n = 13) or 100 (n = 13) or 300 mg/d (n = 11)
of androstenedione (Sports One, Klein Laboratories, Wallingford, Conn) for
7 days. Urine was collected for 8 hours on days 0 (baseline), 1, and 7. Urinary
19-norandrosterone concentrations were compared on days 1 and 7 and between
the 100- and 300-mg groups by paired t tests. Data
are expressed as the mean (SD), and P<.05 was
considered statistically significant.
Four healthy men, aged 30 to 44 years, who denied use of anabolic steroids
and steroid supplements, participated in the study after giving written informed
consent. Each man collected urine before and after ingestion of 10 µg
of 19-norandrostenedione (>99.0% pure). The UCLA Office for Protection of
Research Subjects approved the study.
Over-the-counter capsules were purchased from stores. Androstenedione,
testosterone, 19-norandrostenedione, and nandrolone were obtained from Sigma
Chemical Company (St Louis, Mo). The capsules' androstenedione content was
determined by high-performance liquid chromatography.
Impurities in excess of 1% or 0.1% of the androstenedione were detected
by LC-MS-MS. We analyzed 18 capsules from brands 1 and 2 for steroid contaminants
at more than 0.1% of androstenedione and selected Sports One (brand 1) for
the androstenedione clinical study.9 We also
analyzed 26 capsules from brands 3 through 9 for androstenedione content and
traces of steroids (>1% of the androstenedione).
We reanalyzed 8 capsules from 4 bottles of brand 1 by LC-MS-MS for 19-norandrostenedione
or nandrolone impurities at more than 0.001% of the androstenedione.
Steroid extraction and derivatization, with minor modifications, have
been previously described.10 The analysis used
a high-resolution mass spectrometer. Full scans of 19-norandrosterone were
obtained on 10 samples with 19-norandrosterone concentrations greater than
We analyzed 18 capsules from brands 1 and 2 to select a brand for the
androstenedione clinical study. This analysis detected impurities at greater
than 0.1% of androstenedione. Both brands were more than 99.9% pure, and brand
1 was selected (Table 1).
We also analyzed 26 capsules from brands 3 through 9 at a less stringent
level (1%) of impurity. Table 1
shows that brand 8 contained no steroids and brand 9 contained 10 mg of
testosterone. The mean androstenedione content of the brands labeled as containing
100 mg varied from 85 to 103 mg. The amounts of androstenedione in brands
1, 3, and 4 were within ±10% of that stated on the label, whereas brands
2 and 5 through 9 contained less than 90%.
19-Norandrosterone was found in the urine samples of all subjects (n
= 24; range, 1.3-23.1 ng/mL) who received androstenedione, in none of the
baseline urine samples (n = 37), and in none of the day 1 (n = 13) or day
7 urine samples (n = 13) from the 0-mg/d group (Table 2). There was no difference in 19-norandrosterone concentrations
between day 1 and day 7 (P = .06 and P = .67 for the 100 mg/d and 300 mg/d groups, respectively); thus,
the mean concentrations of day 1 and day 7 were averaged. The mean 19-norandrosterone
concentrations in the 100- and 300-mg/d groups were 3.8 (2.5) ng/mL and 10.2
(6.9) ng/mL, respectively. These concentrations differed (P = .006), indicating dose dependency
After finding significant amounts of 19-norandrosterone in the urine
samples from men who received brand 1, 2 capsules from each of 4 bottles of
brand 1 androstenedione were reanalyzed using a method 100 times more sensitive
to detect 19-norsteroids if present at a level greater than 0.001% of the
androstenedione. None of the capsules contained nandrolone or 19-norandrostenediol;
however, 7 of 8 contained 0.004% to 0.018% 19-norandrostenedione (mean, 0.0076%
or 7.6 µg per capsule).
The urine samples of all 4 subjects who ingested 10 µg of 19-norandrostenedione
contained both 19-norandrosterone and 19-noretiocholanolone (Table 3), neither of which was found in the baseline urine samples.
The 8-hour recovery of 19-norandrostenedione, measured as equivalents of 19-norandrosterone
and 19-noretiocholanolone, was 48% (range, 32%-66%) of the total 10-µg
We found 19-norandrosterone, a metabolite of the prohibited steroid
nandrolone, in all urine samples obtained after oral administration of androstenedione
and in none of the urine samples collected before androstenedione administration.
The mean concentrations were small. Nevertheless, if these samples were from
athletes, 20 of 24 men would have tested positive for 19-norandrosterone (International
Olympic Committee cutoff, 2 ng/mL for males). Thus, the administration of
androstenedione contaminated with 19-norandrostenedione results in 19-norandrosterone
in the urine, as does the administration of nandrolone. However, the ingestion
of 19-norandrostenedione as a trace impurity in androstenedione is readily
distinguished from that of nandrolone use because the latter results in very
high urine concentrations of 19-norandrosterone (>100,000 ng/mL),6 whereas in this androstenedione study the highest
mean concentration was only 23.1 ng/mL.
Brand 1 was selected for the clinical study because it had the most
consistent content of androstenedione
and less than 0.1% impurities. After the unexpected finding
of 19-norandrosterone in urine, we retested brand 1 for steroid traces if
present at a level greater than 0.001%, which could explain this finding.
Only brand 1 was retested; thus, we have no data on trace contamination of
other androstenedione preparations.
The 19-norandrosterone detected in the androstenedione subjects' urine
samples could result from metabolism of the administered androstenedione to
19-norandrosterone, stimulation of a latent metabolic pathway to produce endogenous
19-norandrosterone, or oral administration of the 19-norandrostenedione impurity
in androstenedione. We cannot exclude metabolism of androstenedione to 19-norandrosterone.
Aromatizable androgens such as androstenedione are readily converted to estrogens
in humans, and this conversion is associated with C-19 demethylation to 19-norsteroids.11
We cannot exclude the possibility of stimulation of a latent pathway
either. There are reports of up to 0.6 ng/mL of 19-norandrosterone in the
urine of untreated males,7,8 of
19-nortestosterone and 19-norandrostenedione in human ovarian follicles,12 and of 19-nortestosterone in the serum of pregnant
women.13 However, none of the 37 subjects who
received androstenedione or the 4 subjects who received 19-norandrostenedione
had 19-norandrosterone levels greater than 0.5 ng/mL in their baseline urine
samples. These 41 cases together with the existing 47 cases7,8,14
add support for the suggestion that the upper limit of endogenous 19-norandrosterone
production does not result in urine concentrations of 19-norandrosterone greater
than 0.6 ng/mL.
The hypothesis that the 19-norandrosterone is derived from the 19-norandrostenedione
contamination is strongly supported by the finding that administration of
19-norandrostenedione produced urinary concentrations of 48.9 ng/mL of 19-norandrosterone
and 16.7 ng/mL of 19-noretiocholanolone in the first 2 hours and a 48% recovery
of the 19-norandrostenedione in urine as the metabolites 19-norandrosterone
and 19-noretiocholanolone. Therefore, the 19-norandrosterone found in this
study is almost certainly due to ingestion of contaminated androstenedione.
The manufacturer of the particular batch of brand 1 androstenedione
that we reanalyzed should not be faulted for a product with a maximum contamination
of 0.018%. This level of contamination is far below the level of 0.1% accepted
by the Food and Drug Administration for pharmaceuticals.15
In contrast, some of the other products contained less drug than was listed
on the label and even misrepresented ingredients. According to the Dietary
Supplement Health and Education Act,3 both
brand 8 (no androstenedione) and brand 9 (fails to identify testosterone)
are misbranded. If we apply the United States Pharmacopoeia16 criterion of ±10% for ethical
pharmaceuticals, then brands 2, 5, 6, and 7 are misbranded (wrong strength).
Thus, 6 of 9 brands tested were misbranded.
Today most steroids are made from plant precursors, eg, diosgenin from
yams and stigmasterol from soya, by semisynthesis17
or bacterial action.18 These publications do
not provide sufficient detail to determine how the contamination of androstenedione
with 19-norandrostenedione might have occurred. It is also possible that the
contamination occurred after synthesis, for example, during capsulation or
some other packaging or formulation step.
Sport organizations that test for anabolic steroids have banned androstenedione
even though it is sold OTC in the United States. The data reported herein
show that contaminated androstenedione administration will, like nandrolone
and 19-norandrostenedione, be detected as 19-norandrosterone in urine and
can produce a positive test result for anabolic steroids in subjects taking
OTC dietary supplements.