INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

IT HAS BEEN ESTABLISHED that hormonal changes induced by menopause accelerate the reduction of bone mineral density, elevating the risk of osteoporosis for affected women (4). Several treatments are recommended for osteoporosis, including calcium supplementation, exercise, and estrogen replacement therapy (9). Greater bone mineral density (or attenuation of its loss) from these treatments may be caused by mechanical loading of bone, nutrient status, or endocrine changes (23). Estrogen may have an effect on bone mineral density through calcium metabolism or by increasing levels of growth hormone (GH) (10, 23). The process of biological aging also includes a reduced capability for cellular protein synthesis, leading to a decline in muscle mass. It is known that aging is associated with decline in GH pulsatility, pulse amplitude, and insulin-like-growth factor-I (IGF-I) concentrations, and it has been suggested that the decline in bone mineral density and muscle mass is associated with reduced GH release (29).

GH has been shown to increase in response to running exercise in eumenorrheic women (17). Low-volume resistive exercise also elicits a GH increase in eumenorrheic women during the luteal phase, but not during the follicular phase, of the menstrual cycle (19). Certain high-volume resistive exercise protocols elicit a GH response as well (21). There are only sparse data concerning GH responses to exercise in postmenopausal women. Exercise may stimulate a greater GH response for women receiving hormone replacement therapy (HRT) than for those not taking estrogen (NHRT). An increased GH response could have a beneficial effect on biological aging. The exercise-induced response may be especially evident for women who take oral estrogen rather than transdermal estrogen. The latter produces a tonic level of estradiol (E₂) and does not pass through the liver in high concentrations but distributes E₂ directly into the peripheral circulation (1). It has been shown that postmenopausal women receiving oral estrogen have significantly higher 24-h GH concentrations and pulse amplitudes than postmenopausal women taking transdermal estrogen (8, 27); however, treatment with oral and transdermal E₂ that results in similar concentrations of circulating E₂ increases GH concentrations to the same extent (8).

Prolactin (PRL) and GH share amino acid homologies, and the genes for PRL and GH have similar structure and organization; PRL receptors in humans are also stimulated by GH (28). Both PRL and GH levels have been shown to respond to exercise (6, 14, 17). The lactotrophic cells in the anterior pituitary are increased in size and number by estrogen (28); thus HRT may enhance PRL response to exercise in postmenopausal women. This is the first study to determine whether HRT affects GH and PRL responses to aerobic exercise in postmenopausal women.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Subjects. Seventeen women were recruited through newspaper advertisements. The women provided written consent for participation in the study. Eight untrained women were on oral HRT [5 naturally and 3 surgically postmenopausal; mean age 49.38 ± 3.10 (± SE) yr]. Nine untrained women were NHRT subjects (7 naturally and 2 surgically postmenopausal, age 53.0 ± 2.98 yr). The HRT subjects were taking Ogen, Premarin, or Estratest daily. Criteria for participation in the study included 1) being in postmenopausal status, either natural or surgical (removal of both ovaries and uterus); 2) being able to complete 30 min of moderate treadmill exercise; 3) not having chronic illnesses, such as diabetes mellitus, liver or gallbladder disease, coronary heart disease, malignancy, anemia, or renal failure; and 4) not taking any medication that could alter hormone concentrations, e.g., -blockers, glucocorticoids, diuretics, or other hormonal or hormone-mimetic medications. All subjects were deemed healthy through medical history screening and a graded-exercise test with a 12-lead electrocardiogram. Women were judged to be postmenopausal by 1) surgical removal of both ovaries and uterus, 2) absence of menses for at least 1 yr, and 3) baseline gonadotropin levels. The study was approved by the Southeastern Louisiana University Committee for the Use of Humans and Animals as Research Subjects and was completed in accordance with the Declaration of Helsinki.

Nutritional status. To verify that diet was similar for the HRT and NHRT groups, each subject completed a 3-day food record during the week before an experimental and a control trial. By using a computer program (Nutritionist IV-N², version 3; Salem, OR), food records were analyzed for 1) total kilocalorie intake, 2) percent of diet that was carbohydrate, fat, and protein, and 3) saturated and unsaturated fat intake. Subjects were asked to refrain from exercise or alcohol ingestion for 48 h before testing.

Session 1. Subjects completed a preexperimental session for determination of fitness level and to become familiarized with the treadmill. Skinfold thickness was assessed at four sites: 1) triceps, 2) abdomen, 3) suprailiac, and 4) thigh (13). Maximal O₂ uptake (O_{2 max}) was assessed on a treadmill with an automated system to measure O₂ consumption (O₂). Every 30 s, an expiratory air volume was assessed with a heated pneumotach (series 3813; Hans Rudolph, Kansas City, MO) and pressure transducer (VRCD/HC-1; Consentius Technologies, Sandy, UT), and expired O₂ and CO₂ were analyzed (S-3A/1 and CD-4 analyzers; Ametek, Pittsburgh, PA). Equipment was interfaced (OUS/MC; Consentius Technologies) to a personal computer, and values were recorded every 30 s. Before each O_{2 max} determination, the O₂ and CO₂ gas analyzers were calibrated with gases of known composition. The treadmill protocol began at 2 miles/h (mph) and 2% grade, and every 2 min it was increased by 0.5 mph and 2% grade until 3.5 mph was reached; thereafter, treadmill grade only was increased by 3% every 2 min until exhaustion. All subjects were determined to have reached O_{2 max} when either the primary criterion of a plateau in O₂ with an increase in workload was met or two of three secondary criteria were noted: 1) reaching predicted maximal heart rate, 2) respiratory exchange ratio >1.0, or 3) a rating of perceived exertion (10-point Borg Scale) of 9 or 10 (11). Descriptive data are shown in Table 1. Age, height, weight, body mass index (BMI), sum of skinfolds, and O_{2 max} of the subjects in the HRT and NHRT groups were not statistically different.

Sessions 2 and 3. Subjects reported to the exercise physiology laboratory at 0745 after an overnight fast. A registered nurse inserted an intravenous catheter (20 gauge, 32 mm; Jelco, Tampa, FL) and attached a normal saline lock. With the subject in a sitting position, resting blood samples were collected from the catheter at 0830 (40 min before exercise; 40) and at 0910 h (10 min before exercise; 10). For each blood draw, the first 3 ml of blood (with saline from the catheter lock) was withdrawn into a discard tube preceding a 25-ml draw. The catheter was then flushed with physiological saline solution (3 ml) to maintain patency. Subjects completed 30 min of treadmill exercise at 79.16 ± 1.2% O_{2 max} (HRT group) and 80.18 ± 0.91% O_{2 max} (NHRT group). Percent O_{2 max} values were not significantly different in NRT and NHRT groups. Exercise intensity was maintained by adjusting speed and grade of the treadmill. Blood was collected during [after 15 min of exercise (+15)] and immediately after treadmill exercise [0-min recovery (R0)]. Additional blood draws were taken in a sitting position at 10, 20, 35, 50, 65, and 80 min postexercise (R10, R20, and so on). Each subject completed a second trial that served as a control, and blood was sampled at the same diurnal times as during the exercise trial but with the subject resting in a sitting position.

Blood analyses. For each blood draw, samples were collected in two 10-ml whole blood tubes for endocrine determinations, a 5-ml EDTA tube for hematocrit and hemoglobin assays, and a 3-ml sodium fluoride/potassium oxalate tube for a colorimetric plasma lactate analysis (Sigma Chemical, St. Louis, MO). Whole blood was centrifuged, and serum was divided into aliquots and frozen (20°C) for subsequent determination of GH, PRL, luteinizing hormone (LH), follicle-stimulating hormone (FSH), and E₂. Hematocrit (microcapillary method) and hemoglobin values (Sigma Chemical) were used to determine plasma volume change (7) and to correct hormone levels for hemoconcentration shifts that could inherently elevate hormone levels (18). Lactate levels were documented in the present study to demonstrate the effects of the exercise on elevated metabolic activity. Baseline concentrations of LH, FSH, and E₂ were determined to verify reproductive hormone status of the women and to provide a complete endocrine profile of each subject.

Statistics. The data were analyzed by using two different statistical approaches. First, to examine the total response of GH and PRL to exercise and estrogen replacement, we computed an integrated area under the curves (AUC) for GH and PRL response by using a trapezoidal method after subtracting averaged baseline hormone concentration for each subject. AUC is a summary statistic often used in endocrine research (e.g., Ref. 19) that allows the cumulative effect of a stimulus on an endocrine parameter to be examined. Independent t-tests were used to determine whether hormone concentrations differed between the HRT and NHRT groups. These were one-tailed analyses, as we expected GH and PRL to be higher in the HRT group than in the NHRT group. To ascertain whether AUC for GH and PRL differed between experimental (exercise) and control trials, dependent t-tests were conducted. These were also one-tailed analyses, because we expected GH and PRL to be higher during the exercise trial than during the control trial.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Time period 40 and 10 hematocrit and hemoglobin concentrations were within the normal range for resting values. Plasma volume shifts did not exceed 7.4% between any two time points during the exercise trial. Age, height, weight, BMI, O_{2 max}, and sum of skinfolds were not significantly different between HRT and NHRT groups. Lactate peaked at 5.79 ± 0.80 mM (+15) for HRT subjects and at 4.60 ± 0.65 mM for NHRT subjects (R0). Resting and peak lactate values were not significantly different between the HRT and NHRT groups during exercise.

GH. GH levels peaked at R0 at 10.8 ± 1.60 and 4.90 ± 1.18 ng/ml for for HRT and NHRT groups, respectively (Fig. 1). The results of t-tests comparing AUC for GH levels in HRT and NHRT groups revealed that both exercise trials were significantly higher than control trials, and AUC for GH levels in the HRT group was significantly higher than that for NHRT in the exercise trial (Table 2).

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Data represent means ± SE of growth hormone concentrations for experimental () and control () trials before (40 and 10 min), during (+15 min), and after exercise (R0 to R80 min) for hormone replacement therapy (HRT) group (n = 8) and for experimental () and control () trials before, during, and after exercise for non-HRT (NHRT) group (n = 9). Time, in min. * Significantly different values, HRT compared with NHRT; P <0.05. Values were significantly higher for both exercise groups than for controls.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Data represent means ± SE prolactin concentrations for experimental () and control () trials before, during, and after exercise for HRT group (n = 7) and for experimental () and control () trials before, during, and after exercise for NHRT group (n = 9). Time, in min, as in Fig. 1. * Significantly different values for HRT group (solid symbols) compared with controls (open symbols).

PRL. PRL peaked at 12.67 ± 2.58 ng/ml (R0) for HRT subjects and at 9.04 ± 2.17 ng/ml (R10) for the NHRT group (Fig. 2). The t-test results indicated AUC values for PRL in the HRT and NHRT groups during exercise were significantly higher than those in control trials, but not significantly different from each other (Table 2).

Potential confounding variables. Mean FSH, LH, and E₂ concentrations were within the normal range, as established by our laboratory, for postmenopausal women on HRT and NHRT (Table 3). As expected, E₂ was significantly higher in the HRT group, and FSH was significantly higher in the NHRT group. One potential variable influencing GH responses to exercise is fitness level of the subjects. The results of an ANCOVA, using O_{2 max} as the covariate and AUC for GH as the dependent measure, revealed a significantly higher AUC for GH for the HRT group [F(1, 15) = 6.16; P < 0.05]. Adjusted means for GH are shown in Table 2. Moreover, we evaluated the relationship between O_{2 max} and the magnitude of the exercise-induced change in GH concentrations (AUC) by using a Pearson correlation coefficient, which revealed a small, nonsignificant correlation (r = 0.27, P > 0.10).

View this table: [in this window] [in a new window]   Table 3.   Baseline (10) growth hormone, prolactin, gonadotropin, and estradiol concentrations

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

We have demonstrated that GH responses to treadmill exercise are higher in postmenopausal women receiving HRT than in those not on HRT. However, in response to treadmill exercise, PRL increases in postmenopausal women regardless of their estrogen status. Our research design controlled for GH pulsatility by using a control trial and short intervals between blood sampling. This is the first study to document effects of HRT in postmenopausal women on their GH and PRL responses to a well-quantified exercise bout.

Circulating GH and IGF-I concentrations have been shown to decline with age in women of reproductive age and postmenopausal women (27). The age-related decline in GH pulses may be explained partially by an increase in somatostatin tone, because the administration of agents that inhibit somatostatin release, i.e., cholinergic agonists and arginine, have been demonstrated to increase GH release in older persons (23, 29). There is evidence that GH and IGF-I facilitate anabolic functions of bone formation and protein synthesis in muscle (2). The data from our investigation underscore the importance of incorporating exercise into the lifestyle of postmenopausal women, whether on or off estrogen replacement, to reduce catabolic effects associated with aging.

Evidence suggests that oral estrogen administration to postmenopausal women elicits an increase in GH-pulse height, individual GH-pulse amplitude, and incremental GH-pulse amplitude (8, 26). Although both oral and transdermal HRT have been shown to increase blood levels of GH, neither has been shown to completely reverse age-related reductions (2). Mechanisms suggested for the E₂ effect on increased GH include the expression of E₂ and somatostatin receptors in the hypothalamus by GH-releasing hormone neurons (24) and reduced IGF-I negative feedback effect on GH secretion (1, 8). We have previously shown that exercise-induced increases in GH do not result in short-term increases in IGF-I concentrations (20); therefore, we did not measure IGF-I values in the present study.

Dawson-Hughes et al. (5) reported a slightly higher GH response to light exercise in postmenopausal women treated for 2 wk with ethynyl E₂. The exercise intensity was not well quantified (walking at ~1 mph as determined by a pedometer) and appeared to be considerably lower than that used in the present study. It has been shown that the mode and intensity of exercise affect the GH response (17-19, 21). We have previously documented elevated GH concentrations in young women in response to treadmill running (17) and to resistive exercise during the luteal phase of the menstrual cycle (19). One mechanism explaining enhanced exercise-induced increases in GH increases via E₂ includes increased pituitary sensitization leading to increased release of GH-releasing hormone (19), which could explain the observations from the present study. Another mechanism, hemoconcentration, was ruled out as an explanation for increased GH and PRL levels in the present study, because peak percent increases in GH and PRL were well above the greatest percent PV reduction that would have caused increased hormonal concentrations.

It has been shown that PRL secretion is increased by E₂ through 1) increasing baseline secretion, 2) increasing release of thyroid-releasing hormone, and 3) reducing the effects of dopamine agonists on inhibition of PRL secretion (25). PRL responses to exercise have been observed in eumenorrheic runners, but the PRL responses were absent in amenorrheic runners (22), and those findings were suggested to be caused by extreme ovarian suppression. Our data, however, suggest that treadmill exercise at ~80% of O_{2 max} increases PRL concentrations in postmenopausal women, regardless of their HRT status.

Keizer et al. (16) showed that PRL rises in trained and untrained eumenorrheic women when they are exercised to exhaustion (above lactate threshold). Because lactate levels were increased during exercise in the present study, it would appear that lactate threshold was exceeded and produced a PRL response even in the NHRT postmenopausal women.

For a number of reasons, we do not think there was a difference in physical exertion between the HRT and NHRT groups. First, the percentages of O_{2 max} maintained by the NHRT and HRT groups were similar. Second, lactate concentrations in the HRT and NHRT groups were not significantly different at rest or peak exercise. Third, the rating of perceived exertion reported for the NHRT group was similar to that of the HRT group during exercise (5.10 for HRT group vs. 5.30 for NHRT group on the 10-point Borg scale). Fourth, the percent of maximum heart rate maintained by both groups during exercise was not significantly different (90.23 vs. 89.49% for HRT vs. NHRT group, respectively; P > 0.05).

It has previously been shown that higher fitness levels are associated with greater GH response to exercise (3). Not only was O_{2 max} similar between groups (P > 0.05), but even with the variance accounted for by O_{2 max} removed in an ANCOVA analysis, the GH concentrations were still significantly and substantially higher in the HRT group compared with the NHRT group. Moreover, no significant relationship between fitness level and GH response (AUC) was found.

Diet may also affect GH concentrations (12); however, it is unlikely that diet affected the results of the present study. Food records indicated that the diets of subjects in the HRT and NHRT groups were not significantly different in total caloric intake as well as in percent carbohydrate, fat, and protein; moreover, the nutrient composition and caloric values were close to recommended daily allowance values.

The exercise response of other hormones may also be affected by HRT. For example, we have recently documented an enhanced exercise response of the adrenal hormones dehydroepiandrosterone and cortisol in the same subjects who participated in the present study (15). Future investigations are needed to examine exercise responses and adaptations of hormones affected by E₂.

In summary, this is the first study to demonstrate an augmented response of GH to treadmill exercise by postmenopausal women on HRT and similar PRL responses in postmenopausal women regardless of HRT or NHRT status. The effect of HRT and exercise may serve to enhance the obvious health benefits produced by HRT or by exercise alone.