cortisol ratio responses to short-term intensive exercise training
Eur J Appl Physiol (2010) 108:1125–1131
Amy R. Lane · Joseph W. Duke · Anthony C. Hackney
Abstract This study examined the eVect of dietary
carbohydrate (CHO) consumption on the free testosterone
to cortisol (fTC) ratio during a short-term intense microcycle
of exercise training. The fTC ratio is a proposed biomarker
for overreaching–overtraining (i.e., training stress
or imbalance) in athletes. The ratio was studied in two
groups, control-CHO (»60% of daily intake, n = 12) and
low-CHO (»30% of daily intake, n =
, of male subjects
who performed three consecutive days of intensive training
(»70–75% maximal oxygen consumption, 60 min per
day) with a dietary intervention (on the day before and
during training). Resting, pre-exercise blood samples were
collected under standardized-controlled conditions before
each day of training (Pre 1, 2, 3) and on a fourth day after
the micro-cycle (Rest). Bloods were analyzed for free testosterone
and cortisol via radioimmunoassay procedures.
Subjects performed no additional physical activity other
than prescribed training. Statistical analysis (ANCOVA)
revealed the fTC ratio decreased signiWcantly (p < 0.01)
from pre-study resting measurement (Pre 1) to the Wnal
post-study resting measurement (Rest) in the low-CHO
group (¡43%), but no change occurred (p > 0.05) in the
control-CHO group (¡3%). Findings suggest if the fTC
ratio is utilized as a marker of training stress or imbalance
it is necessary for a moderately high diet of CHO to be
consumed to maintain validity of any observed changes in
the ratio value.
Keywords Hormones · Endocrine · Diet · Stress ·
Adaptation
Introduction
As the primary fuel source for high-intensity exercise it is
important for athletes to maintain an adequate amount of
carbohydrates (CHO) in their diet. Regrettably, athletes
tend to overestimate their CHO consumption by 10–25%
(Snyder et al. 1995). The body can only store limited
amounts of CHO, needing a continual replenishment
through diet to maintain the high level of training necessary
to remain competitive (Evans and Hughes 1985). Dietary
consumption of CHO has also been shown to aVect the hormonal
response to exercise. SpeciWcally, Galbo et al.
(1979) found that subjects on a low-CHO diet (daily caloric
intake = 11.5% CHO) for several days had elevated levels
of fuel mobilizing hormones compared to subjects on a
high-CHO diet (daily caloric intake = 77% CHO). Thus, a
consequence of ingesting a low-CHO diet is the eVect that
it could have on cortisol and testosterone levels. Without
abundant CHO to replenish both glycogen stores and maintain
suYcient blood glucose levels, cortisol will be secreted
in an eVort to maintain blood glucose through muscle
proteolysis, and amino acid oxidation (Brooks et al. 2005).
Testosterone also displays alterations in bioavailability
with diVering diets. For example, Anderson et al. (1987)
found that testosterone levels decreased in a high-protein
Communicated by Susan Ward.
A. R. Lane · J. W. Duke · A. C. Hackney (&)
Endocrine Section, Applied Physiology Laboratory,
Department of Exercise and Sport Science,
University of North Carolina, CB # 8700,
Fetzer Building, Chapel Hill, NC 27599, USA
e-mail: ach@email.unc.edu
A. C. Hackney
Department of Nutrition, School of Public Health,
University of North Carolina, Chapel Hill, NC, USA
1126 Eur J Appl Physiol (2010) 108:1125–1131
123
diet, compared to a high-CHO diet, while cortisol showed
the opposite response.
The free testosterone to cortisol (fTC) ratio has been suggested
as a potential endocrine biomarker to monitor the
training status of an athlete (i.e., training stress or imbalance)
(Adlercreutz et al. 1986; Hoogeven and Zonderland 1996;
Lehmann et al. 1993; Urhausen et al. 1995). A decrease in
the fTC ratio greater than 30% or an absolute ratio value less
than or equal to 0.35 £ 10¡3 is considered to be indicative of
a negative catabolic state. Lehmann et al. (1993) also proposed
a theoretical model with potential mechanisms for the
development of the Overtraining Syndrome which included
glycogen depletion and hormonal imbalances as potential
causative factors. The current body of research in this area
spans long- and short-term training studies looking primarily
at the response of the fTC ratio to various modes and intensities
of exercise (Vervoorn et al. 1991; Purge et al. 2006;
BanW and Dolci 2006). Only a few studies have taken into
consideration the speciWc eVect of diet on this ratio (Snyder
et al. 1995; Costa et al. 2005; Costill et al. 1988). Even these
studies, however, did not necessarily have strict nutritional
controls or a speciWc intent to determine the eVect of diet per
se on the ratio. Therefore, the purpose of this study was to
determine if, with exercise training held constant, the daily
carbohydrate consumed aVected the fTC ratio and inXuenced
the potential use of this ratio as a biomarker for training during
a short-term intensive micro-cycle.
Methods
Subjects
Twenty endurance-trained males (¸5 days/week for at least
60 min) between the ages of 20 and 42 participated in this
study. All were in good health, and had no history of a current
or chronic medical condition or musculoskeletal injury.
All were informed regarding the protocol and procedures
and signed an informed consent prior to participation as
stipulated by the University of North Carolina institutional
review board.
Procedures
Subjects reported to our laboratory on Wve separate occasions.
The Wrst was an orientation session, followed by
three consecutive days of training sessions (Training 1, 2,
3), and Wnally a resting session (Rest). During the Wrst visit,
subjects signed the informed consent and underwent a medical
and physical screening to conWrm they were healthy
and could participate. The subject’s height (cm), mass (kg),
and body composition were recorded. Body composition
was estimated using skinfold measurement at the chest,
abdomen, suprailiac, and mid-axillary sites. All measurements
were taken in triplicate using Lange® calipers (Cambridge
ScientiWc Industries, Cambridge, MD), and the mean
of the two closest measurement values (§5%) at each site
was used as values in the generalized equation reported by
Golding et al. (1982) to calculate percentage body fat. The
subjects then completed a peak oxygen consumption test
(VO2peak) on a Lode® electronically braked cycle ergometer
(Lode, Groningen, Netherlands) using the protocol of Mac-
Dougall et al. (1991). Respiratory gases were collected
throughout using the Parvo Medics TrueMax® 2400 Metabolic
System (Parvo Medics, Salt Lake City, UT). The
respiratory data collected (oxygen consumption [VO2] carbon
dioxide produced [VCO2], respiratory exchange ratio
[RER]) were used in a regression equation to determine the
workload for their training session visits (see below). Peak
Watt power output, test time duration, ratings of perceived
exertion, and maximal heart rate (HR) were also recorded.
The subjects were each randomly assigned to either the
control- (»60%) or low- (»30%) CHO diet group at their
initial visit. Each group maintained a diet log beginning the
day before the Wrst training session (Training 1) and each
day of training (three consecutive days) as well as the
fourth day (Rest) until the Wnal blood sampling (see below).
Daily diets were free-living (prepared by the subjects), but
guidance as to types of and quantity of food was speciWcally
given to each group member. In addition, to insure the
desired CHO consumption and the maintenance of an isocaloric
status for all subjects during training, supplements
were provided. The control-CHO group subjects were
given a CHO supplement, Polycose® (87.5 g CHO per
day). The low-CHO group was given Boost High Protein®
drink (45 g protein per day). Throughout the study food and
supplement consumption were monitored to insure compliance
with required dietary conditions.
The remaining visits at our laboratory (Training 1, 2, 3)
began at least 5 days after the orientation session. The subject
arrived at the laboratory between 06:00 and 10:00 a.m.,
in an 8-h fasted state, fully rested having done no exercise
in the previous 24 h. The subjects rested in a supine position
for 10 min, and gave an initial 3 ml blood sample (Pre)
using a standard veni-puncture technique. A Polar® (Polar
Electro, Inc., Lake Success, NY) HR monitor was Wxed
around their chest, and they then completed a 10 min
warm-up. After 5 min of seated rest on the cycle ergometer
for resting respiratory and HR measures, they cycled for
60 min at 100% of the individual anaerobic threshold (IAT)
which corresponded to »70–75% of their VO2peak measurement.
The IAT had been determined from each subjects
VO2peak testing results using procedures explained in detail
elsewhere (Daly et al. 2005). Another blood sample (Post)
was taken immediately following the 60 min of cycling.
Every 10 min during the exercise HR (Polar® monitor) and
Eur J Appl Physiol (2010) 108:1125–1131 1127
123
RPE were determined using a Borg’s original 6–20 scale.
At 20-min intervals (minutes 16–20, 36–40, 56–60), the
respiratory gases (VO2, VCO2, RER) were measured. Upon
completing the exercise, the subjects were permitted to
leave the laboratory once their heart rate was below
100 bpm. This exact protocol was repeated for all three
training session days (Training 1, 2, 3).
The Wnal session (fourth session) was 24 h after the
Training 3 session. After arriving at the laboratory the subject
rested in a supine position for 10 min, a Wnal 3-ml
blood sample (Rest) was taken using the same veni-puncture
technique as noted above. The blood sampling protocol
used during the entire study is schematically represented in
Fig. 1.
Biochemical analysis
All blood samples were collected into K2EDTA Vaccutainer
® tubes and put on ice until processing. Whole blood
samples were analyzed for hematocrit (Hct) and hemoglobin
(Hb). The remaining blood samples were centrifuged at
4°C at 3,000 rpm for 10 min and separated. Plasma was aliquotted
and stored at ¡80°C until hormonal analysis. Triplicate
measurements of Hct and Hb were utilized to
determine plasma volume shifts (PV) within, between
training sessions (Dill and Costill 1974). Hormone levels
were measured with speciWc radioimmunoassay (RIA) procedures
for both free testosterone and cortisol (Siemens
Medical Diagnostics, Los Angeles, CA). The sensitivity for
cortisol and free testosterone RIA assays were 0.2 g/dL
and 0.15 pg/mL, respectively. The coeYcient of variation
for the between and within assay replicates within the RIA
analyzes was less than 10%.
Statistical analysis
All analyses were conducted with a statistical software
package (SPSS version 15.0, Chicago, IL) and data are presented
as mean § standard deviation (SD). A one-way
analysis of variance (ANOVA) was used to examine
between-group diVerences in physical characteristics as
well as for the metabolic (VO2, HR, etc.) and performance
(peak Watts, time, etc.) outcomes of the VO2peak testing.
Separate mixed model between-within ANOVAs were used
to examine for between group (control-CHO vs. low-CHO)
diVerences as well as changes over time (Pre 1, Pre 2, Pre 3,
Rest) within the resting cortisol, free testosterone, and fTC
ratio hormonal measures (an analysis of covariance was
also applied to examine the fTC ratio data only; see
“Results”). Separate between-within ANOVAs were used
to look for diVerences in the dietary measurements assessed
(CHO and caloric intake). Additionally, between-within
ANOVAs were completed to examine for diVerences in the
metabolic responses (VO2, HR, RPE) within the training
sessions. Tukey post hoc tests were used if signiWcant
F-ratios were found to determine where signiWcant diVerences
existed. The signiWcance level was set a priori · 0.05.
Results
The physical characteristics and VO2peak testing results are
found in Table 1. There were no signiWcant diVerences
between the groups in any of these variables.
The analysis of the diets indicated there was a highly
signiWcant diVerence (p < 0.001) in the percentage of
Fig. 1 The blood sampling protocol employed in the study. The ‘Pre’
points are resting values (before exercise at Training 1, 2, 3) and the
‘Post’ points are the values after exercise. The ‘Rest’ point is a resting
value for the fourth day of the study. The # symbol denotes when the
blood sample was taken
↓ ↓ ↓ ↓ ↓
↓ ↓
↓ ↓
Table 1 Physical characteristics and VO2peak testing results for both
the ‘control’ and ‘low’ carbohydrate consumption treatment groups
(mean § SD)
No signiWcant between-group diVerences were detected for any of
these measurements
Measurement Control-CHO
(n = 12)
Low-CHO
(n =
Age (years) 27.1 § 5.8 24.0 § 3.4
Height (cm) 179.4 § 6.8 180.0 § 5.6
Mass (kg) 75.0 § 7.3 74.7 § 7.9
Body fat (%) 12.1 § 3.3 9.8 § 3.2
VO2peak (L/min) 4.52 § 0.76 4.58 § 0.41
VO2peak (ml/(kg min)) 60.1 § 6.6 61.9 § 8.6
Max RER 1.19 § 0.07 1.16 § 0.06
Max HR (bpm) 188.4 § 8.8 191.0 § 8.9
Max RPE 19.3 § 0.7 19.3 § 0.7
Test duration (Min) 18.9 § 2.5 17.7 § 1.6
Max workload (Watts) 350.0 § 47.7 315.6 § 37.6
1128 Eur J Appl Physiol (2010) 108:1125–1131
123
carbohydrates consumed on a daily basis between the
groups. The control- and low-CHO groups consumed
58.5 § 4.9% and 31.9 § 2.5% (main eVect) of their daily
dietary intake in CHO, respectively. The total average
(§SD) calories (kcal) per day consumed during the study
period (4 days) by the control group was 3226.7 §
389.8 kcal, and in the low-CHO group was 2804.1 §
604.9 kcal (main eVect; p > 0.05). There was no signiWcant
diVerence found in the caloric intake between the groups, or
over time (p > 0.05) and each group remained iso-caloric in
their intake during the study.
The metabolic responses to the training sessions (Training
1, 2, 3) were remarkably similar between the groups.
The mean VO2 values during the exercise training sessions
were not signiWcantly diVerent between the groups
(p > 0.05). The control-CHO group and low-CHO group
both worked at high-intensity levels, 73.1 and 71.0% (main
eVect), respectively. Heart rate responses during exercise
were also not found to be diVerent between the groups
(p > 0.05), but by 60 min of exercise in Training 1, 2, and 3
the responses were substantially elevated in both groups
(control-CHO range = 156–162 bpm; low-CHO range =
161–163 bpm [overall mean responses during Training 1, 2
and 3]). However, group diVerences were seen for RPE. At
the 60-min measurement point of the Training 1 and Training
3 sessions, the low-CHO group had signiWcantly higher
(p < 0.05) RPE responses (16.0 § 1.7 vs. 17.6 § 1.1, and
15.0 § 1.7 vs. 17.0 § 1.1). SigniWcant plasma volume
reductions (p < 0.01) were observed by 60 min of exercise
for each training session in both groups; however, there
were no signiWcant diVerences between the groups in
the magnitude of the reduction (p > 0.05). Furthermore, the
level of reductions were similar across the three exercise
training sessions in both groups (p > 0.05) (these data are
not reported).
Resting cortisol levels increased signiWcantly (p < 0.04)
in the low-CHO group during the study. At Pre 3
(27.6 § 7.4 g/dL) and at Rest (27.6 § 7.9 g/dL) the levels
were greater than that at Pre 1 (24.1 § 9.1 g/dL) for
this group. Conversely, the control-CHO group showed no
signiWcant (p > 0.05) changes in resting cortisol during the
study (see Table 2).
Resting free testosterone decreased throughout the training
sessions (Pre 1 to Rest) for both groups (see Table 2).
However, these changes were not statistically signiWcant
(p > 0.05), although the magnitude of the decrease
appeared substantially greater in the low-CHO group
(»7.9 pg/ml, Pre 1 to Rest) than in the control-CHO group
(»3.6 pg/ml) and this diVerence nearly reached statistical
signiWcance (p < 0.06; interaction eVect).
The resting fTC ratio levels showed a signiWcant main
eVect for time (p = 0.008) during the study with the values
decreasing progressively from Pre 1 to Rest (see Table 2).
No signiWcant interaction of group by time was detected
(p > 0.05) for the ratio change. However, close examination
of the individual group fTC responses revealed that the
low-CHO group measures decreased more substantially
over time (Pre 1 vs. Rest), while those of the control-CHO
group only decreased somewhat. Furthermore, it was noted
that the initial resting Pre 1 fTC ratio levels between the
groups were somewhat diVerent (»24% diVerence);
although not statistically so (p > 0.05). Nevertheless, due to
these trends toward diVering responses in and between the
groups, an analysis of covariance (ANCOVA) was also
conducted on the fTC ratio in which the initial resting Pre 1
levels of the ratio for each group were used as covariates.
This alternative analysis revealed a signiWcant group by
time interaction eVect for the fTC ratio changes (p < 0.03;
see Fig. 2). The low-CHO group displayed a decrease at
every measurement time, with the Wnal measurement (Rest)
being signiWcantly lower than at the Pre 1 (#43%;
p < 0.0001) low-CHO value. The change over time for the
control-CHO was not signiWcant (#3%, Pre 1 to Rest;
p > 0.05). Furthermore, at the Wnal measurement (Rest) the
low-CHO group value was signiWcantly less than that of the
control-CHO group (p < 0.03).
Plasma volume shifts (PV) were calculated to examine
the inXuence of hemoconcentration and hemodilution on
the resting fTC ratio. That is, from the beginning of the
study (initial resting, Pre 1) in comparison to the remaining
resting measurements (Pre 2, Pre 3, Rest). At the Pre 2,
Pre 3 and Rest time points, the control-CHO group displayed
a somewhat increased PV while the low-CHO
group was somewhat decreased (Pre 2 = + 5.0 § 10.3%
vs. ¡0.7 § 10.3%, Pre 3 = +6.8 § 9.1% vs. ¡3.9 § 9.2%,
Table 2 Cortisol, free testosterone, and fTC ratio data for each session
at the resting (pre-exercise for the Training session 1, 2, and 3 as
well as session 4 [Rest involving no exercise]) measurement
* SigniWcant diVerence (p < 0.05) from Pre 1 measurement (§SD)
Session Control-CHO
(n = 12)
Low-CHO
(n =
Cortisol (g/dL) Pre 1 19.7 § 6.1 24.1 § 9.1
Pre 2 19.6 § 4.0 24.4 § 7.1
Pre 3 19.3 § 5.1 27.6 § 7.4*
Rest 19.4 § 4.4 27.6 § 7.9*
Free testosterone
(pg/ml)
Pre 1 24.7 § 15.2 22.0 § 5.4
Pre 2 22.8 § 10.0 20.3 § 3.8
Pre 3 19.0 § 8.5 19.2 § 5.6
Rest 21.1 § 7.9 14.0 § 4.9
fTC ratio Pre 1 1.90 § 1.98 1.53 § 1.30
Pre 2 1.54 § 0.93 1.21 § 0.67
Pre 3 1.44 § 1.20 1.04 § 0.72
Rest 1.49 § 0.91 0.82 § 0.71
Eur J Appl Physiol (2010) 108:1125–1131 1129
123
Rest = +6.7 § 8.5% vs. ¡6.9 § 8.7%, for the control-CHO
and low-CHO, respectively). In these data there was a main
eVect for group (control-CHO > low-CHO; p < 0.02), but
no time or group by time interaction eVects. The hormonal
concentrations reported (cortisol, free testosterone, fTC
ratio) are not corrected for these PV shifts based upon recommendations
in the literature (McMurray et al. 1995).
Discussion
The intent of this study was to determine if the daily CHO
consumption aVected the fTC ratio in men during a shortterm
intensive micro-cycle. Thus, it was critical to this
study’s research question that the diet and training treatments
applied were of an appropriate nature. The dietary
CHO content of the two groups were signiWcantly diVerent
from one another, which was the dietary treatment desired.
The low-CHO group consumed a diet signiWcantly lower
(»32%) in carbohydrates than the control-CHO group
(»59%). Furthermore, there was no signiWcant diVerence in
the total number of daily calories consumed between the
groups, only the marco-nutrient composition of the calories.
The low-CHO group consumed »23% less CHO during
the study period, than in their normal diets outside of
the study which was signiWcantly diVerent (p < 0.05). The
total daily caloric consumption, however, was not signiWcantly
diVerent within the subjects in each group for the
before study period and during the study period. Therefore,
these Wndings suggest that the diVerences between groups
were of a signiWcantly diVerent dietary composition only,
strongly indicating that the dietary treatment was eVectively
achieved.
The other treatment eVect being pursued in this study
was for the micro-cycle of exercise training to be intensive.
The groups both worked above 70% of their peak oxygen
consumption (approximately 100% of their IAT), speciWcally,
on average the control-CHO and low-CHO groups
exercised at »73 and »71% of VO2peak over the three training
sessions, respectively. This high level of VO2 combined
with the high RPE and HR (»85% of maximum) responses
by the end of each training session for both groups suggests
the desired intensive treatment eVects were also achieved.
The key Wnding in this study was that the resting fTC
ratio signiWcantly and substantially decreased from Pre 1 to
Rest within the low-CHO group, while there was no change
in the control-CHO group ratio throughout the training.
This suggests that a 3-day micro-cycle of high-intensity
exercise can induce a decrease in this ratio if suYcient levels
of CHO (»60% daily intake) consumption in the diet
are not maintained. The control-CHO group demonstrated
that by maintaining adequate levels of CHO (»60%) in
their diet that even with diYcult training on consecutive
days, the fTC ratio can be maintained within a normal
range. Lehmann et al. (1993) suggest maintenance of the
ratio is indicative of a healthy balance between training
load and recovery which is critical to optimal training and
adaptation in athletes. The reduced ratio response of the
low-CHO group matches trends found by other researchers
investigating this ratio in response to exercise training
(Filaire et al. 2002; Vaananen et al. 2004; BanW and Dolci
2006; Costa et al. 2005). These studies include both shortand
long-term training periods. However, these studies did
not control for diet to the same extent as in the current
study. The similarity of ratio response between the current
and other studies provides conWdence that the present
responses were not random chance, but were a result of the
experimental treatments.
It has been suggested that in men, changes in the fTC
ratio are driven by the action of free testosterone, and in
women, cortisol is the more dominant hormone (Urhausen
et al. 1995). The present Wndings appear to follow suit with
this thought as most of the hormonal change was observed
in free testosterone. Although, in the low-CHO group, both
hormones changed from Pre 1 to the Rest measurement 3
days later, free testosterone decreased by 36.1% while cortisol
increased by 14.8%.
Why was resting cortisol elevated over the course of the
study? Upon review of Fig. 3, it appears the increases in
cortisol in response to exercise during the Wrst two training
sessions (Pre to Post comparison) were so great that there
was not adequate time to recover between training sessions
and the resting levels at Pre 3 and Rest were not able to
return to a normal range. Why might such an inadequate
recovery occur? Metabolically, the low-CHO diet in combination
with the exercise would have prevented glycogen
stores from being completely resynthesized between each
training session. This has been shown in the classic work
Fig. 2 Plot of ANCOVA adjusted results for the resting (Pre 1, Pre 2,
Pre 3, Rest) measurements of the fTC ratio (mean § SE) of both
groups (the Pre 1 value for each groups was used as the respective
covariant in the ANCOVA). The asterisk denotes signiWcant diVerence
from Pre 1 within the low-CHO group as well as between both groups
(p < 0.05)
0.6
0.8
1
1.2
1.4
1.6
1.8
2
Pre 1 Pre 2 Pre 3 Rest
fTC Ratio
Control-CHO
Low-CHO
*
1130 Eur J Appl Physiol (2010) 108:1125–1131
123
by Costill in the 1970s (reviewed in Brooks et al. 2005).
This potentially drove a faster and more robust response
from the fuel-mobilizing hormones, including cortisol
(Galbo 1983). Experimentally, Galbo et al. (1979) found
that with several days of a low-CHO diet, during exercise,
the resulting compromised glycemia levels induced a much
more rapid and robust response from cortisol than with a
high-CHO diet. This cortisol eVect has been reported to
persist for some time into the recovery from exercise as
much as 24 h (Hackney 2006).
It is important to note that cortisol is also known to
have a direct suppressive eVect on the testosterone steroidogenesis
process. Therefore, with perpetually elevated
levels of cortisol in response to exercise, and in the recovery
from exercise, there would be a potential for inhibition
on testosterone production via steroid inhibition
(Cumming et al. 1983; Brownlee et al. 2005). Figure 3
reveals that the exercise during the training sessions
caused testosterone to increase (Post) only slightly relative
to the resting (Pre) levels before each exercise session
in the low-CHO group. These eVects became mitigated
over time and by the Wnal training session; the exerciseinduced
elevation barely brought the hormone levels back
to the original pre-study resting value (i.e., Pre 1). This
reduction in free testosterone was inXuential in the ratio
being signiWcantly suppressed. Cortisol, as previously
stated, inhibits testosterone production, and with the
increase witnessed in that hormone (see Table 2; Fig. 3),
it was likely responsible, in part, for the depressed levels
of testosterone in the low-CHO group. This point is
supported by the fact that the change in cortisol levels
during the study (Pre 1 vs. Rest) and the reduction
observed in the fTC ratio (from Pre 1 to Rest) show a
strong negative correlation (r = ¡0.763, p < 0.03), even
though there are only eight subjects in the low-CHO group
(i.e., " cortisol!# fTC ratio).
Additionally, factors may also have inXuenced the testosterone
levels in this study. With the low-CHO diet,
increased levels of the sex hormone binding globulins
(SHBG) may have occurred (Anderson et al. 1987).
Increased SHBG can reduce the bioavailability of free testosterone
by causing more of the hormone to be in the
bound form rather than free (Anderson et al. 1987). Interestingly,
with less circulating free testosterone in the
blood both protein synthesis and glycogen resynthesizing
can become compromised (Hackney 1996). Such eVects
could exacerbate the impact of the low-carbohydrate diet
even more so. That is, with continued exercise in such
conditions there would be accelerated and augmented
endocrine responses in lipolytic regulatory hormones in
an attempt to promote lipolysis and provide energy substrates
(Hackney 1996, 2006). Cortisol is one of the key
lipolytic regulatory hormones and thus a reduction in free
testosterone could perpetuate the cycle of further suppressed
testosterone via elevated cortisol, and add to a
growing disparity with each training session. Testosterone
may also have been decreased as a result of a reduction in
the pituitary gonadotropes (LH, FSH) as a result of the
training sessions, (LH stimulate testosterone production)
(Galbo et al. 1977). These hormones were not measured
in the present study. In men, longer training cycles are
typically needed to cause persistent reductions in these
hormones, which previously published research has
shown (Safarinejad et al. 2009). Finally, testosterone
could also have been suppressed by prolactin, a stressrelated
hormone. In men, prolactin is known, (when at
hyperprolactinemic levels) to have an eVect on the
gonadotropes at the testis, inhibiting testosterone production.
These hyperprolactinemic men have been shown to
display signiWcantly suppressed testosterone levels
(Hackney et al. 2000). In an ancillary project within the
current study, prolactin levels were examined in some of
the subjects (data not reported here). While the levels of
this hormone did increase, they were not to levels indicative
of a hyperprolactinemic state. Therefore, prolactin
may have been responsible for some of the decrease in
testosterone levels, but was likely not a key factor in the
overall suppression.
As noted earlier, based upon literature recommendations
the reported hormonal values are not corrected for hemoconcentration
or hemodilution eVects due to the inXuence
of plasma volume shifting (McMurray et al. 1995). Interestingly,
if we had represented the fTC ratio in this fashion,
the noted eVects for between-group’s diVerences would
have been even greater than reported. That is, the control-
CHO exhibited an overall hemodilution eVect (values are
lower than actual) and the low-CHO exhibited a hemoconcentration
eVect (values are higher than actual). Thus, PV
corrected ratio values would have magniWed the experimental
Wnding of observed ratio decreases in the low-CHO
group.
Fig. 3 Mean cortisol and free testosterone responses at pre-exercise
(resting, Pre) and post exercise (Post) during the training sessions in
the low-CHO group
10
15
20
25
30
35
Pre Post Pre Post Pre Post Rest
Training 1 Training 2 Training 3 Rest
Cortisol & Testosterone
Cortisol
(μg/dL)
Testosterone
(pg/ml)
Eur J Appl Physiol (2010) 108:1125–1131 1131
123
Conclusions
This study has shown that endurance athletes need to be aware
of the eVect one’s diet can have on their ability to maintain
normal endocrine status during periods of high-intensity
endurance training. SpeciWcally, a more suYcient dietary
CHO (»60%) intake allows subjects to maintain a more anabolic
hormonal proWle during the course of performing 3 days
of consecutive high-intensity exercise training.
The fTC ratio has been suggested as a biomarker for
overreaching and overtraining in endurance activities.
Adlercreutz et al. (1986) suggested that a decrease in the
ratio greater than 30% indicates a state of overreaching–
overtraining. In this study, with the low-CHO diet there
was an overall average decrease of 43% in the resting fTC
ratio (see Fig. 2), while the control-CHO group only experienced
a 3% decrease, suggesting the low-CHO subjects
may have been experiencing at least overreaching. Lehmann
et al. (1993), suggest this represents an imbalance
between training load and recovery. It is not possible from
the current data to determine if the low-CHO subjects were
truly overreaching or overtraining as this diagnosis requires
more medical evaluation than this study intended. However,
if the fTC ratio is to be a legitimate biomarker for
such training stress, then the current data suggest that as
little as 3 days of intensive training without adequate
carbohydrate consumption can compromise the measures
validity. The current Wndings strongly support that it is critical
therefore, to control for diet if this ratio is going to be
utilized for diagnostic or informative measures of training
status; otherwise the results could be misleading.
References
Adlercreutz H, Harkonen M, Kuoppasalmi K, Naveri H, Huhtaniemi I,
Tikkanen H, Remes K, Dessypris A, Karvonen J (1986) EVect of
training on the plasma anabolic and catabolic steroid hormones
and their response during physical exercise. Int J Sports Med
7:27–28
Anderson KE, Rosner W, Khan MS, New MI, Pang S, Wissel PS,
Kappas A (1987) Diet–hormone interactions: protein/carbohydrate
ratio alters reciprocally the plasma levels of testosterone and
cortisol and their respective binding globulins in man. Life Sci
40:1761–1768
BanW G, Dolci A (2006) Free testosterone/cortisol ratio in soccer: usefulness
of a categorization of values. J Sports Med Phys Fitness
46(4):611–616
Brooks GA, Fahey TD, Baldwin KM (2005) Exercise physiology:
human bioenergetics and its applications. The McGraw Hill Companies,
Boston
Brownlee KK, Moore AW, Hackney AC (2005) Relationship between
circulating cortisol and testosterone: inXuence of physical exercise.
J Sports Sci Med 4:76–83
Costa RJ, Jones GE, Lamb KL, Coleman R, Williams JH (2005) The
eVects of a high carbohydrate diet on cortisol and salivary
immunoglobulin A (s-IgA) during a period of increase exercise
workload amongst olympic and ironman triathletes. Int J Sports
Med 26(10):880–885
Costill DL, Flynn MG, Kirwan JP, Houmard JA, Mitchell JB, Thomas
R, Han Park S (1988) EVects of repeated days of intensiWed training
on muscle glycogen and swimming performance. Med Sci
Sports Exerc 20(3):249–254
Cumming DC, Quigley ME, Yen SC (1983) Acute suppression of circulating
testosterone levels by cortisol in men. J Clin Endocrinol
Metab 57:671–673
Daly W, Seegers CA, Rubin DA, Dobridge JD, Hackney AC (2005)
Relationship between stress hormones and testosterone with prolonged
endurance exercise. Eur J Appl Physiol 93(4):375–380
Dill DB, Costill DL (1974) Calculation of percentage changes in
volumes of blood, plasma, and red cells in dehydration. J Appl
Physiol 37(2):247–248
Evans WJ, Hughes VA (1985) Dietary carbohydrates and endurance
exercise. Am J Clin Nutr 41:1146–1154
Filaire E, Legrand B, Bret K, Sagnol M, Cottet-Emard JM, Pequignot
JM (2002) Psychobiologic responses to 4 days of increased training
and recovery in cyclists. Int J Sports Med 23(8):588–594
Galbo H (1983) Hormonal and metabolic adaptation to exercise. G
Thieme Stuttgart, New York
Galbo H, Hummer L, Petersen IB, Christensen NJ, Bie N (1977) Thyroid
and testicular hormone responses to graded and prolonged
exercise in man. Eur J Appl Physiol 36:101–106
Galbo H, Holst JJ, Christensen NJ (1979) The eVect of diVerent diets
and of insulin on the hormonal response to prolonged exercise.
Acta Physiol Scand 107(1):19–32
Golding LA, Myers CR, Sinning WE (1982) The Y’s way to physical
Wtness. National Board of YMCA, Chicago
Hackney AC (1996) Testosterone, the hypothalamo-pituitary-testicular
axis and endurance training: a review. Biol Sport 13(2):85–98
Hackney AC (2006) Stress and the neuroendocrine system: the role of
exercise as a stressor and modiWer of stress. Expert Rev Endocrinol
Metab 1(6):783–792
Hackney AC, Dobridge J, Wilson LS (2000) The overtraining
syndrome in athletes: hypothesis as to development and steps to
treatment. Medicina Sportiva 4(1):E15–E28
Hoogeven AR, Zonderland ML (1996) Relationships between testosterone,
cortisol and performance in professional cyclists. Int
J Sports Med 17(6):423–428
Lehmann M, Foster C, Keul J (1993) Overtraining in endurance athletes:
a brief review. Med Sci Sports Exerc 25(7):854–862
MacDougall JD, Wenger HA, Green HJ (1991) Physiological testing
of the high performance athlete. Human Kinetics, Champaign
McMurray RG, Eubanks TK, Hackney AC (1995) Nocturnal hormonal
responses to resistance exercise. Eur J Appl Physiol 72:121–126
Purge P, Jurimae J, Jurimae T (2006) Hormonal and psychological
adaptation in elite male rowers during prolonged training. J Sports
Sci 24(10):1075–1082
Safarinejad MR, Azma K, Kolhi AA (2009) The eVects of intensive,
long-term treadmill running on reproductive hormones, hypothalamus-
pituitary-testis axis, and semen quality: a randomized controlled
study. J Endocrinol 200(3):259–271
Snyder AC, Kuipers H, Cheng B, Servais R, Fransen E (1995) Overtraining
following intensiWed training with normal muscle glycogen.
Med Sci Sports Exerc 27(7):1063–1070
Urhausen A, Gabriel H, Kindermann W (1995) Blood hormones as markers
of training stress and overtraining. Sports Med 20(4):251–276
Vaananen I, Vasankari T, Mantysaari M, Vihko V (2004) Hormonal
responses to 100 km cross-country skiing during 2 days. J Sports
Med Phys Fitness 44(3):309–314
Vervoorn C, Quist AM, Vermulst LJ, Erich WB, de Vries WR,
Thijssen JH (1991) The behavior of the plasma free testosterone/
cortisol ratio during a season of elite rowing training. Int J Sports
Med 12(3):257–263
