| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Studies |
Division of Endocrinology and Metabolism, Department of Medicine (J.A.K., A.W., M.L.H.); Division of Biostatistics and Epidemiology, Department of Health Evaluation Sciences (K.S.P.); Department of Human Services (A.W.); General Clinical Research Center (J.Y.W., A.W.); and the National Science Foundation Center for Biological Timing (M.L.H.), University of Virginia, Charlottesville, Virginia 22908
Address all correspondence and requests for reprints to: Dr. Mark L. Hartman, Eli Lilly and Co., Lilly Corporate Center, Drop Code 5015, Indianapolis, Indiana 46285.
Abstract
Exercise of appropriate intensity is a potent stimulus for GH and
cortisol secretion. Circadian and diurnal rhythms may modulate the GH
and cortisol responses to exercise, but nutrition, sleep, prior
exercise patterns, and body composition are potentially confounding
factors. To determine the influence of the time of day on the GH and
cortisol response to acute exercise, we studied 10 moderately trained
young men (24.1 ± 1.1 yr old; maximal oxygen consumption,
47.9 ± 1.4 mL/kg·min; percent body fat, 13.2 ± 0.6%).
After a supervised night of sleep and a standard meal 12 h before
exercise, subjects exercised at a constant velocity (to elicit an
initial blood lactate concentration of 
2.5 mmol/L) on a treadmill
for 30 min on 3 separate occasions, starting at 0700, 1900, and
2400 h. Blood samples were obtained at 5-min intervals for 1
h before and 5 h after the start of exercise; subjects were not
allowed to sleep during this period. Subjects were also studied on 3
control days under identical conditions without exercise. There were no
significant differences with time of day in the mean blood lactate and
submaximal oxygen consumption values during exercise. The differences
over time in serum GH and cortisol concentrations between the exercise
day and the control day were determined with 95% confidence limits for
each time of day. Exercise stimulated a significant increase in serum
GH concentrations over control day values for approximately 105–145
min (P < 0.05) with no significant difference in
the magnitude of this response by time of day. The increase in serum GH
concentrations with exercise was followed by a transient suppression of
GH release (for 55–90 min; P < 0.05) after
exercise at 0700 and 1900 h, but not at 2400 h. Although the
duration of the increase in serum cortisol concentrations after
exercise was similar (150–155 min; P < 0.05)
at 0700, 1900, and 2400 h, the magnitude of this increase over
control day levels was greatest at 2400 h. This difference was
significant for approximately 130 min and approximately 40 min compared
to exercise at 1900 and 0700 h, respectively
(P < 0.05). The cortisol response to exercise at
0700 h was significantly greater than that at 1900 h for
about 55 min (P < 0.05). A rebound suppression of
cortisol release for about 50 min (P < 0.05) was
observed after exercise at 2400 h, but not 0700 or 1900 h.
Both baseline (before exercise) and peak cortisol concentrations were
significantly higher at 0700 h than at 1900 or 2400 h
(P < 0.01). We conclude that time of day does not
alter the GH response to exercise; however, the exercise-induced
cortisol response is modulated by time of day.
EXERCISE OF appropriate intensity is a potent stimulus for GH and cortisol secretion. The GH and cortisol responses to exercise are both dependent on the relative exercise workload, but other factors modulate these hormonal responses, including the mode and duration of exercise, anaerobic vs. aerobic exercise, prior meal ingestion, and fitness level of the subject (1, 2, 3, 4, 5, 6, 7, 8, 9, 10). Underlying pulsatile and circadian/diurnal rhythms of these hormones potentially may modulate the exercise response observed (11, 12, 13, 14, 15, 16).
At rest, GH secretion is pulsatile and is influenced by age, gender, nutrition, sleep, body composition, fitness, and sex steroid hormones (17). During a 24-h period, spontaneous GH secretion is maximal at night in close association with slow wave sleep (11, 12, 17, 18). The GH response to an iv bolus injection of GHRH is maximal at night, particularly during slow wave sleep (13, 14). Circadian and/or diurnal rhythms may influence GH secretion independently of sleep, as GH secretion is still maximal at night when subjects are kept awake (12). Food ingestion may account in part for the lower levels of GH secretion during the day; ingestion of high fat meals is known to blunt the GH response to exercise (4, 17).
Cortisol secretion is also pulsatile. The amplitude and frequency of cortisol secretory pulses are modulated by a circadian rhythm (15, 16). Circulating cortisol concentrations are maximal in the early morning hours just before awakening as a result of increased cortisol secretory pulse amplitude and frequency. The amplitude of cortisol secretory pulses progressively decreases throughout the day until cortisol concentrations are quite low in the evening (16). This circadian rhythm is largely independent of sleep, although an inhibitory effect of nocturnal, but not diurnal, sleep on cortisol concentrations has been documented in humans (18). Meals stimulate cortisol release in humans. Exercise performed immediately after food ingestion results in a blunted cortisol response to the exercise stimulus (1, 2).
Few studies have assessed the impact of time of day on the GH and cortisol responses to exercise. Two studies reported no effect of time of day on the GH response to exercise. One study was limited by the omission of an appropriate control (nonexercise) day (19). The second study employed a low exercise intensity and infused iv glucose during the test (20). Four studies have reported that the cortisol response to exercise is similar in the morning and evening (2, 19, 21, 22). The statistical methods used in these studies did not allow inspection of possible differences over time in the cortisol response to exercise compared with resting conditions at different times of day. A fifth study employed a low exercise intensity that elicited a cortisol response in the afternoon, but no response in the morning or evening (20).
The present study examined the influence of time of day on the GH and cortisol responses to exercise. Hormonal responses to exercise in the early morning and early evening and at midnight were compared with those on identical control (nonexercise) days. This study controlled for the confounding effects of meals and sleep and employed a statistical approach that enabled the difference in hormone concentrations to be compared across time with 95% confidence limits.
Subjects and Methods
Subjects
Ten healthy, moderately trained, male subjects completed this study. All subjects underwent a detailed medical history and physical examination and provided informed, written consent in accordance with the guidelines established by the University of Virginia human investigation committee. The subjects were nonsmokers, were not taking any medication known to affect GH or cortisol secretion, had not undergone transmeridian travel in the past 8 weeks, were not night shift workers and regularly went to bed between 2200–2400 h. Screening laboratory data revealed normal hepatic, renal, metabolic, hematological, thyroid, and gonadal function.
Experimental design
Body composition and aerobic exercise capacity [maximal oxygen consumption (VO2 max)] were determined approximately 1 week before the first admission. Subjects were admitted to the General Clinical Research Center (GCRC) on six separate occasions for hormonal measurements. Three of the admissions were exercise days, and three admissions were control days. The study times for the admissions were 0600–1200, 1800–2400, and 2300–0500 h. The exercise times in each of these study periods were 0700, 1900, and 2400 h, respectively. These times were chosen because they are three popular times for recreational exercise (before and after day shift work and after evening shift work). These times also encompass the diurnal rhythms of both cortisol and GH secretion. Although the daily peak of cortisol secretion usually occurs earlier than 0700 h, it was not feasible to conduct the exercise studies at 0400 or 0500 h. Controls were studied at the same time of day as the exercise group, but no exercise occurred. The order of all six admissions was randomized.
VO2 max
Maximal oxygen consumption was determined using a continuous treadmill protocol. The treadmill velocity was initially set at 100 m/min at 0% grade, and every 3 min the velocity was increased by 10 m/min. Subjects were verbally encouraged throughout the test, and the test was terminated when the subject reached volitional fatigue. Metabolic measures were collected using standard open circuit spirometric techniques (2900Z metabolic cart, SensorMedics, Yorba Linda, CA). VO2 max was chosen as the highest VO2 attained during the test. From an indwelling venous catheter located in a forearm vein, blood samples were taken at rest and at the end of each 3-min stage while the subjects continued running. A saline solution (0.9% NaCl) was infused after each blood sample to prevent clotting. Samples were analyzed immediately for lactate concentrations with an automated lactate analyzer (model 2700, YSI, Inc., Yellow Springs, OH).
Body composition
Percent body fat was assessed by hydrostatic weighing as previously described (10). Briefly, each subject was weighed in air on an Accu-Weigh beam scale (Metro Equipment Company, Sunnyvale, CA) accurate to 0.1 kg and subsequently weighed underwater with a 9-kg Chatillon autopsy scale (Chatillon, New York, NY) accurate to 10 g. Residual volume was measured on land using an O2 dilution technique (23). Relative body fat was converted from body density using the equation of Brozek et al. (24).
Study day protocol
Subjects were admitted to the GCRC the evening before the study and slept in the GCRC to standardize sleeping conditions (lights out at 2200 h) before the study day. All subjects were given a standard meal that contained 16 Cal/kg BW (50% carbohydrate, 30% fat, and 20% protein) 12 h before the start of blood sampling. After the standard meal, subjects fasted until the study was completed 18 h later. With the exception of the exercise tests, subjects refrained from vigorous activity, but were allowed to walk around the GCRC unit. An indwelling cannula was inserted in a forearm vein 1 h before the start of blood sampling. Serial blood sampling (2.5 mL/sample) was initiated 1 h after iv placement, and samples were taken at 5-min intervals over 6 h. On three of the visits, subjects exercised at a constant velocity on the treadmill for 30 min (starting at 0700, 1900, and 2400 h). The initial velocity chosen was the velocity associated with a blood lactate concentration of 2.5 mmol/L on the incremental VO2 max treadmill test. This velocity was chosen so that the lactate response to 30 min of exercise would not exceed 4.0 mmol/L (25). During the submaximal exercise test, metabolic measures were collected using standard open circuit spirometry as described above. Subjects were not allowed to sleep during the blood-sampling period. As the subjects were significantly sleep deprived during the 2300–0500 h study periods, electroencephalographic monitoring was performed to record any possible episodes of sleep. No episodes of sleep exceeding 9 min occurred.
Assays
Serum GH was measured in duplicate with an immunoradiometric assay (IRMA) using standards diluted in human serum (Nichols Institute Diagnostics, San Juan Capistrano, CA). The assay sensitivity was 0.2 µg/L, and the mean inter- and intraassay coefficients of variation (CVs) were 8.6% and 4.9%, respectively. This assay was chosen (instead of a chemiluminescence assay) because the primary end point of interest was the peak GH response to exercise rather than the low concentrations of GH between pulses. At the time the study was conducted (1993–1994), the GH chemiluminescence assay had not been fully validated for the concentration range of interest in exercise studies. Cortisol was measured with a chemiluminescence assay (Nichols Institute Diagnostics). The assay sensitivity was 38.6 nmol/L, and the mean inter- and intraassay coefficients of variation (CVs) were 7.0% and 8.4%, respectively. The samples from all study days for each subject were run in the same assay. Serum hormone concentrations were determined by a procedure previously described (26). Briefly, standard curves were evaluated by weighted nonlinear least squares analysis using three different response functions. Uncertainties (SD) associated with each hormone concentration were estimated empirically, considering the variance associated with both the assay response and standard curve evaluations. The standard curve parameters and response function to the variably weighted response data were optimized. Confidence limits for the standard curve parameters were then calculated (26). The function yielding the lowest absolute sum of squared residuals was chosen for analyzing the samples.
Deconvolution analysis of hormone concentrations
Integrated serum GH and cortisol concentrations (area under the curve) for the 6-h time periods were calculated as previously described (27). A multiple parameter deconvolution method was employed to derive quantitative estimates of GH secretory events from the serum GH concentrations (28). This was done to allow simultaneous estimates of the subject-specific monoexponential half-life of endogenous GH. It was assumed that a Gaussian distribution of secretory rates approximates each pulse of GH secretion (29). The sensitivity of the GH IRMA did not allow for estimation of basal GH secretion, so GH secretory rates were assumed to decay to zero (11). GH secretory pulses were considered significant if the fitted amplitude could be distinguished from zero (e.g. pure noise) with 95% statistical certainty. The GH secretory pulse half-duration, GH half-life of elimination, and GH distribution volume were assumed to be constant throughout the study period for each individual (11). The mass of GH secreted per pulse was estimated as the area of the calculated secretory pulse (28). The endogenous GH production rate for each 6-h period was estimated as the product of the number of secretory pulses and the mean GH mass secreted per pulse. Deconvolution analysis was not performed on the serum cortisol concentrations due to the long half-life of cortisol and the short duration (6 h) of blood sampling.
Statistical analysis
Results are expressed as the mean ± SE, unless
otherwise stated. A one-way ANOVA with repeated measures was used to
examine blood lactate and VO2 responses to
submaximal exercise. A 2 x 3 ANOVA (condition x time of
day) with repeated measures was employed to determine mean differences
in integrated hormone concentrations, peak hormone concentrations, and
attributes of GH secretion and clearance. Mean comparisons were
examined when mean differences were observed. An 
level of
P < 0.05 was chosen a priori.
As GH and cortisol were measured at 5-min intervals for 6 h,
paired comparisons at each time point were not appropriate. The
spontaneous changes in hormone concentrations over the course of the
day make statistical comparison of the areas under the hormone
concentration curves for the control and exercise days potentially
misleading. To assess objectively the stimulation of GH and cortisol
release by exercise against a background of spontaneous pulsatile
secretion modulated by diurnal and circadian rhythms, we normalized
each subject’s exercise GH and cortisol concentration time trends to
the subject’s control GH and cortisol concentration time trends,
respectively. Both absolute and percent changes have been used to
analyze such data (30). We had previously determined that
absolute differences were more independent of control conditions than
ratios or percent change for GH data and therefore chose differences as
the method to analyze the present data (31). For the three
times of day, we subtracted each subject’s serum hormone concentration
at each time point during the control day from their hormone
concentration at the same time point during the exercise day (exercise
day minus control day). A regression curve (flexible regression spline)
was derived for the change in serum hormone concentrations (
serum
hormone, exercise day minus control day) over time with simultaneous
(taking into account multiple time points) 95% confidence bands. These
curves are smoothed versions of the raw data, with the amount of
smoothing carefully chosen to be able to estimate the mean profiles
without overfitting. The 95% confidence regions were derived using a
variation of the bootstrap technique (see Appendix in Ref.
31 for further details). The use of this 95% confidence
region avoids the problem of multiple pointwise comparisons. A
significant hormone response to exercise was defined as occurring when
the lower 95% confidence limit for the regression curve was more than
zero. Significant suppression of hormone release was defined as
occurring when the upper 95% confidence limit for the regression curve
was less than zero. To compare the magnitude of the hormone response to
exercise at the different times of day, we determined the difference
between the regression curves (with simultaneous 95% confidence
regions) at each time point for 0700 and 1900 h (0700 minus
1900 h), 0700 and 2400 h (0700 minus 2400 h), and 1900
and 2400 h (1900 minus 2400 h), as previously described
(31).
Results
Subject characteristics
The subjects studied had a mean age of 24.1 ± 1.1 yr, weight of 69.8 ± 2.3 kg, and height of 174 ± 1.6 cm. The subjects had a VO2 max of 47.9 ± 1.4 mL/kg·min and a percent body fat of 13.2 ± 0.6%.
Submaximal VO2 during exercise
Each exercise bout was initiated at the same velocity to elicit
initially a similar blood lactate level and VO2.
Throughout the 30 min of exercise, the mean blood lactate level was
3.5 ± 0.6, 3.9 ± 0.6 and 3.9 ± 0.9 mmol at 0700,
1900, and 2400 h on the exercise day, respectively. This
corresponded to a mean submaximal VO2 of
40.1 ± 0.9, 41.4 ± 0.4, and 41.8 ± 1.5 mL/kg·min,
respectively (85% VO2 max). There were no
significant differences in blood lactate or submaximal
VO2 by time of day.
Mean GH and cortisol responses to exercise
The mean GH and cortisol concentrations over the 6-h period are
shown in Figs. 1
and 2. A rise in response to exercise was
seen in both GH and cortisol concentrations. The effect of underlying
diurnal rhythms of GH and cortisol release can be seen by inspection of
the data from the control days. Spontaneous increases in mean serum GH
concentrations occurred in the evening and night hours despite the fact
that subjects did not sleep (Fig. 1). Serum cortisol concentrations
rose during the early morning hours, so that baseline levels before
exercise were higher before exercise at 0700 than at 1900 or 2400
h (Fig. 2).
|
|
Integrated GH concentrations and attributes of GH secretory pulses
and clearance rates during the 6-h periods are shown in Table 1. No differences in the 6-h integrated
GH concentrations, 6-h GH production rates, or attributes of GH
secretory pulses were found among each of the three times of day in
either the control or exercise conditions. The immediate GH response to
exercise was also assessed by calculating the 2-h integrated GH
concentrations (30-min exercise + 90-min recovery), and there were no
significant differences by time of day. Analysis of the data by
condition (exercise vs. control) revealed that integrated GH
concentrations were about 160% higher on the exercise day than on the
control day as the result of a significant increase in GH production
rates (P < 0.01) with no change in the half-life of GH
disappearance. This enhanced secretion was related to a significant
increase in the mass of GH secreted per pulse despite a decrease in
secretory pulse half-duration (P < 0.05). The exercise
stimulus did not affect the number of GH secretory pulses.
|
depicts the differences over
time in serum GH concentrations between the exercise and control days
for each of the three times of day. In each graph, the solid
line represents the regression curve for the difference, and the
dotted lines represent the 95% confidence intervals. A
significant GH response or suppression occurred when both 95%
confidence limits for the regression curve were greater or less than
zero, respectively. Inspection of these graphs reveals that at all
exercise times (0700, 1900, and 2400 h), serum GH concentrations
were significantly increased by the exercise stimulus compared with the
control day (P < 0.05). The exercise-induced
stimulation of GH release lasted approximately 145 min at 0700 h,
105 min at 1900 h, and 125 min at 2400 h. A significant
suppression of GH release occurred between 205 and 260 min after the
onset of exercise at 0700 h and 150 and 240 min after the onset of
exercise at 1900 h (P < 0.05). There was no
significant suppression of GH release after exercise at 2400 h. To
compare the magnitude of the GH response to exercise at the different
times of day, we determined the difference between the regression
curves at each time point for 0700 and 1900 h (0700 minus
1900 h), 0700 and 2400 h (0700 minus 2400 h), and 1900
and 2400 h (1900 minus 2400 h). Regression curves with
simultaneous 95% confidence bands for these differences were
calculated. These graphs, which are shown in Fig. 4, demonstrate that there were no
significant differences in the GH response to exercise by time of
day.
|
|
Integrated cortisol concentrations on the control days were 127,643 ± 6,460, 44,006 ± 6,630, and 74,770 ± 14,298 min·nmol/L for the 0700, 1900, and 2400 h study days, respectively. On the exercise days, the integrated cortisol concentrations were 135,919 ± 12,232, 65,123 ± 7,325, and 103,942 ± 13,363 min·nmol/L for the 0700, 1900, and 2400 h study days, respectively. ANOVA of these integrated concentrations revealed a significant effect of time of day across conditions (P < 0.01), demonstrating the dominant effect of the circadian rhythm in cortisol release. This effect is further illustrated by the mean baseline cortisol concentrations (1 h before exercise), which were greater (P < 0.01) at 0700 h (411.1 ± 49.7 nmol/L) than at 1900 and 2400 h (168.3 ± 16.6 and 162.8 ± 46.9 nmol/L, respectively). The effect of exercise on integrated cortisol concentrations did not reach statistical significance (P = 0.071), and the interaction between time of day and condition also did not reach statistical significance. Peak cortisol concentrations in response to exercise were significantly greater at 0700 h (729.2 ± 85.2 nmol/L) than at 1900 h (454.0 ± 66.3 nmol/L) or 2400 h (634.6 ± 76.9 nmol/L; P < 0.01); the peak concentration at 2400 h was significantly greater than at 1900 h (P < 0.05). These peak concentrations on the exercise days were significantly higher (P < 0.01) than the highest concentrations observed on the control days (644.8 ± 54.7, 189.5 ± 33.3, and 313.2 ± 28.5 nmol/L for 0700, 1900, and 2400 h study days, respectively). However, a significantly greater percent increase from mean baseline cortisol concentrations (1 h before exercise) to peak cortisol values occurred after exercise at 2400 h (600%; P < 0.01) compared with 1900 h (200%) and 0700 h (150%).
The differences over time in serum cortisol concentrations between the
exercise and control days for each of the three times of day are shown
in Fig. 5 with 95% simultaneous
confidence limits. Exercise at each of the selected times of day
significantly stimulated cortisol release for approximately 150–155
min (P < 0.05). There was a brief period of rebound
suppression of cortisol release occurring between 240 and 290 min after
exercise at 2400 h (P < 0.05), but this was not
observed after exercise at 0700 and 1900 h.
|
demonstrates
that for a brief period of time (55 min) the cortisol response to
exercise was greater at 0700 h than at 1900 h; this
difference was significant between 25 min before and 30 min after
exercise began (P < 0.05). Similarly, cortisol
concentrations were transiently higher (40 min) after exercise at
2400 h compared with exercise at 0700 h; this difference was
significant between 130 and 170 min after exercise began
(P < 0.05). The cortisol response to exercise was
significantly greater at 2400 h than at 1900 h; this
difference was significant for approximately 130 min, between 10 and
140 min after exercise began (P < 0.05).
|
The primary purpose of this study was to investigate whether
circadian and diurnal rhythms underlying the secretion of GH and
cortisol modulate the responses of these hormones to aerobic exercise.
Consistent with earlier reports (3, 6, 8, 9, 19, 20, 32),
exercise stimulated a dramatic increase in serum GH concentrations for
approximately 105–145 min. The magnitude of the GH response was
independent of time of day. The exercise-induced increase in serum GH
levels was followed by a transient suppression of GH release (55–90
min) after exercise at 0700 and 1900 h, but not at midnight. In
contrast, the cortisol response to exercise was modulated by time of
day. Although the duration of the increase in serum cortisol
concentrations after exercise was similar (150–155 min) at 0700,
1900, and 2400 h, the increase over control day levels was
greatest at 2400 h and smallest at 1900 h, and an
intermediate response occurred at 0700 h. These differences in
serum cortisol concentrations between exercise and control days were
transient (40–130 min) and were most apparent when the effects over
time were studied, as the 6-h integrated cortisol concentrations did
not differ significantly at each time of day.
Both spontaneous and GHRH-stimulated GH secretion are maximal at night (11, 12, 13, 14, 17, 18). In contrast, the GH secretory response to exercise at 0700, 1900, and 2400 h did not differ significantly in the present study. Two other studies with different experimental designs have reported similar findings. Galliven et al. observed that the GH response to high intensity exercise (90% VO2 max) was similar in the morning and evening, but no comparison was made to control (nonexercise) conditions (19). Scheen et al. observed no difference in the GH secretory response to 3 h of moderate intensity (40–60% VO2 max) exercise initiated at 0500, 1430, or 2330 h compared with resting conditions (20). The present data suggest that aerobic exercise of sufficient intensity is able to override any diurnal or circadian rhythms underlying GH release when the potentially confounding variables of meals and sleep are controlled. This latter point is important, because subjects ate a standard meal 12 h before each exercise bout and were not allowed to sleep during blood sampling. These conditions differ from normal daily life, where meal ingestion may blunt the GH response to exercise (4). In addition, exercising in the evening may decrease GH secretion during the first few hours of sleep (33).
The increase in serum GH concentrations with exercise was related to an
approximately 2-fold increase in GH secretion rates with no significant
change in the half-life of GH disappearance, as estimated by
deconvolution analysis. Exercise increased (2-fold) the mass of GH
secreted per pulse and decreased (20%) the secretory pulse
half-duration, but did not change the number of GH secretory pulses
compared with control conditions. These effects of exercise on GH
secretory pulse attributes were similar to our previous observations
(32). In the present study there were no significant
effects of time of day on GH secretory pulse attributes in response to
exercise.
After the increase in serum GH concentrations with exercise, a significant suppression of GH release occurred between 2.5 and 4 h after the onset of exercise at 0700 and 1900 h. This may reflect autonegative feedback of GH on its own secretion, as suggested by studies in rats (34). It is unlikely that pituitary stores of GH were depleted, because repeated bouts of exercise (70% VO2 max) elicit similar GH secretory responses even if spaced by only 1 h of rest (32). The absence of a similar rebound suppression of GH release after exercise at 2400 h may reflect the diurnal rhythm of enhanced GH secretion at night even when subjects are awake (12). As the sensitivity of the GH IRMA used in this study was insufficient to measure GH concentrations below 0.2 µg/L, it is possible that the magnitude and timing of the suppression of GH release after exercise may not have been precisely determined.
The circadian rhythm had a significant effect on the cortisol response
to aerobic exercise. The increase in cortisol concentrations over
control day levels was greatest at 2400 h and smallest at
1900 h, and an intermediate response occurred at 0700 h.
Previous studies have reported that that the cortisol response to
exercise is similar in both the morning and evening (2, 19, 21, 22), or that exercise elicits a cortisol response in the
afternoon, but no response in the morning or evening (20).
Three features of our experimental design may account for the
differences between the present and past findings. First, an exercise
intensity was employed (85% VO2 max) that was
adequate to reliably stimulate cortisol release; a threshold of 60%
VO2 max has been proposed by Few
(7). Two studies reporting disparate results from the
present findings used exercise intensities between 40–60%
VO2 max (2, 20). Second, the study
design controlled for the possible effects of meals, sleep, and prior
exercise. The cortisol response to exercise is blunted by prior meal
ingestion, and the postprandial increase in serum cortisol
concentrations is attenuated by prior exercise (1, 2).
Third, the pattern of cortisol release over time under exercise
conditions was compared with that observed at rest at the identical
time of day for each subject. Early studies did not include a control
day and simply compared cortisol concentrations during exercise to the
preexercise levels on the same day (2, 5, 7, 19). This
approach does not take into account the changing cortisol
concentrations over the course of the day due to the circadian rhythm.
Recent studies have compared the area under the curve for serum
cortisol concentrations on exercise and control (nonexercise) days
(20, 21, 22). Using this approach, the cortisol response to
exercise (70% VO2 max for 40 min) was reported
to be similar in the morning (0800 h) and the evening (2000 h)
(21). However, using area under the curve analysis or
assessing peak hormone concentrations alone does not enable changes
over time to be detected. The statistical method employed in the
present study enabled differences in serum cortisol levels between
exercise and control conditions to be assessed at each time point with
95% confidence limits. The effect of time of day on these differences
could then be assessed over time with 95% confidence limits. This
approach revealed that time of day does indeed modulate the cortisol
response to exercise, albeit for short periods of time (40–130
min).
The effect of time of day on the cortisol response to exercise cannot be completely accounted for by the preexercise baseline cortisol concentrations. The highest peak cortisol concentrations after exercise occurred at 0700 h, with the second highest vales at 2400 h; the peak response at 1900 h was the lowest. In contrast, the largest percent increase from baseline to peak values was 600% at 2400 h compared with 200% and 150% at 1900 and 0700 h, respectively. As baseline cortisol levels were similar at 1900 and 2400 h, a greater increase over time occurred in response to exercise at 2400 h than at 1900 h. Furthermore, although baseline cortisol levels were significantly higher at 0700 than at 1900 h, the increases in cortisol levels over time on the exercise compared with the control days were similar at 0700 and 1900 h, with a transiently greater response at 0700 h. Thus, assessing the cortisol response to exercise in relation to the baseline cortisol concentration alone is inadequate (21).
Time of day apparently influences the incremental response, but not the peak response, to pharmacological tests of cortisol secretory reserve. The peak cortisol response 30 min after an iv bolus of ACTH is similar in the morning (0800 h) and afternoon (1600 h), but cortisol concentrations rise more rapidly after ACTH administration in the afternoon than in the morning (35). The incremental cortisol response observed in response to insulin-induced hypoglycemia and ovine CRH is also greater in the afternoon than in the morning, although peak cortisol concentrations do not differ by time of day (36, 37). For clinical purposes, the peak response to ACTH is more valuable than the incremental cortisol response when assessing adrenal cortical reserve, with a peak response exceeding 550 nmol/L (20 µg/dL) considered a normal response (35, 38). In the present study mean peak cortisol responses exceeded this threshold at 0700 and 2400 h, but not at 1900 h. This suggests that the cortisol response to exercise is regulated in complex ways that differ from that observed in response to hypoglycemia, CRH, or ACTH. However, it should be pointed out that the times of day that testing was performed in the present study differed from those in previous studies of pharmacological tests, and thus a direct comparison of results is not possible.
No differences by time of day were observed for mean submaximal VO2 or blood lactate levels during exercise in the present study. Previous studies with different experimental designs have observed higher peak lactate accumulation in the afternoon during exhaustive constant power exercise (39) and greater work capacity in the afternoon and evening compared with the morning (40). In the present study the treadmill velocity was chosen to produce similar blood lactate concentrations and VO2 values. Velocity was not adjusted for changes in blood lactate levels achieved, and subjects did not exercise to exhaustion. Anecdotally, when asked which exercise bout seemed the most difficult, subjects could usually choose one time of day; however, the ratings of perceived exertion (Borg scale) obtained during each exercise bout did not reflect any time of day differences. Thus, the effect of time of day on the cortisol responses to exercise cannot be explained by differences in the physiological or psychological response to exercise.
In conclusion, when the confounding factors of meals, prior exercise, and sleep are controlled, time of day has no effect on the magnitude of the GH response to exercise. In contrast, the cortisol response to exercise is modulated by time of day. Peak cortisol concentrations in response to exercise were highest at 0700 h, followed by 2400 and 1900 h in that order. In contrast, maximal increases in cortisol concentrations over time in comparison to control day conditions occurred at 2400 h, followed by 0700 and 1900 h, in that order. These data suggest that the circadian rhythm of cortisol secretion influences the cortisol response to exercise.
Acknowledgments
We acknowledge the following individuals for their assistance with the present project: Ms. Sandra Ware-Jackson and the staff of the General Clinical Research Center at University of Virginia for their help in performing this study; Ginger Bauler and Catherine Kern for performing the immunoassays; Pete Hellmann and Dr. Paul Suratt for assisting with the sleep studies; Drs. Michael Johnson and Johannes D. Veldhuis for providing the multiple parameter deconvolution program; Dr. Frank Harrell for advice about the statistical analysis; and David Boyd for providing data management assistance.
Footnotes
1 This work was supported in part by grants from the NIH (AG-10997 to
M.L.H. and RR-00847 to the General Clinical Research Center and CDMAS
Laboratory at the University of Virginia) and the National Science
Foundation Center for Biological Timing (Grant DIR 89–20162). 
2 Present address: Department of Exercise Science, Syracuse
University, 820 Comstock Avenue, Room 201, Syracuse, New York 13244.
3 Present address: 32 Ivy Trails, Weaverville, North Carolina
28787.
Received February 15, 2000.
Revised November 15, 2000.
Revised February 21, 2001.
Accepted February 23, 2001.
References
This article has been cited by other articles:
 |
|
 |
|
  J. Gibney, M.-L. Healy, and P. H. Sonksen The Growth Hormone/Insulin-Like Growth Factor-I Axis in Exercise and Sport Endocr. Rev., October 1, 2007; 28(6): 603 - 624. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. R. Collier, E. Collins, and J. A. Kanaley Oral arginine attenuates the growth hormone response to resistance exercise J Appl Physiol, September 1, 2006; 101(3): 848 - 852. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Weltman, J. Y. Weltman, C. P. Roy, L. Wideman, J. Patrie, W. S. Evans, and J. D. Veldhuis Growth hormone response to graded exercise intensities is attenuated and the gender difference abolished in older adults J Appl Physiol, May 1, 2006; 100(5): 1623 - 1629. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. M. Hershberger, M. R. McCammon, J. P. Garry, M. T. Mahar, and R. C. Hickner Responses of Lipolysis and Salivary Cortisol to Food Intake and Physical Activity in Lean and Obese Children J. Clin. Endocrinol. Metab., September 1, 2004; 89(9): 4701 - 4707. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. KOSKA, L. KSINANTOVA, R. KVETNANSKY, D. HAMAR, M. MARTINKOVIC, and M. VIGAS Effects of Endurance Training on Endocrine Response to Physical Exercise after 5 Days of Bed Rest in Healthy Male Subjects Ann. N.Y. Acad. Sci., June 1, 2004; 1018(1): 569 - 575. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Kanaley, R. Dall, N. Moller, S. C. Nielsen, J. S. Christiansen, M. D. Jensen, and J. O. L. Jorgensen Acute exposure to GH during exercise stimulates the turnover of free fatty acids in GH-deficient men J Appl Physiol, February 1, 2004; 96(2): 747 - 753. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. C. Woodson, D. Macintosh, M. Fleshner, and D. M. Diamond Emotion-Induced Amnesia in Rats: Working Memory-Specific Impairment, Corticosterone-Memory Correlation, and Fear Versus Arousal Effects on Memory Learn. Mem., September 1, 2003; 10(5): 326 - 336. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
