INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ALTHOUGH THE MAJORITY of previous research has suggested that substrate utilization during moderate-intensity exercise is different between genders (12, 13, 36-38), some research has demonstrated minimal differences (5, 16, 31). Previous research has also suggested that women may rely more heavily on lipid metabolism while preserving muscle glycogen (36-38). The mechanisms responsible for these gender-dependent differences in fuel selection appear to be correlated with the complex hormonal milieu surrounding ovulation in women.

Some recent investigations have attempted to study a gender-specific response in glucose kinetics to exercise using stable isotope-tracer methodologies (8, 12, 28). Friedlander et al. (12) demonstrated similar rates of glucose appearance (R_a) and glucose disposal (R_d) after training at the same exercise intensity relative to peak O₂ uptake (O_2 peak) in women and men. When data from women were compared with previously collected data from men during an identical training program, women tended to have a lower respiratory exchange ratio (RER), despite similar rates of glucose flux. Friedlander et al. further suggested that this might indicate a higher relative contribution from blood glucose as a percentage of total carbohydrate (CHO) oxidation.

Previous research has attempted to increase the level of experimental control by matching men and women on several training characteristics, including weekly mileage and average training intensity (36-38). These studies have continually demonstrated that women have a lower RER, rely less on muscle glycogen, and oxidize proportionately more lipid than similarly trained men (36-38) during exercise relative to O_2 peak. However, previous research has not determined whether men and women demonstrate different patterns of substrate oxidation and glucose kinetics during exercise relative to the lactate threshold (LT).

Coggan et al. (3) demonstrated that glucose kinetics during exercise were most related to the LT, even when subjects were matched on O_2 peak, fiber type distribution, and capillary density. During exercise at 55% of O_2 peak, glucose R_a and R_d were significantly lower for the high- than for the low-LT subjects. Therefore, normalizing the exercise intensity relative to the LT may reduce the variability in substrate oxidation between the genders. The rationale for selecting an exercise intensity relative to the LT stems from previous research that has demonstrated the concept that exercise performance is more related to the ventilatory threshold (21, 27) and LT (6, 7) than to a percentage of O_2 peak. We hypothesized that, during cycle ergometer exercise at the same percentage of the LT, gender differences in glucose R_a and R_d [per kilogram of fat-free mass (FFM) or total body weight] would be minimized.

The purpose of the present study was to compare glucose kinetics and substrate utilization in men and women during exercise at two intensities: 70 and 90% of O₂ uptake (O₂) at LT (70 and 90% LT, respectively). The response to two exercise intensities was tested on the same day to minimize changes in the hormonal milieu.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects and procedures. Six regularly menstruating women and five men served as subjects for this investigation (Table 1). Before participation, each subject completed a University of Montana Internal Review Board-approved informed consent form and a physical activity history. All subjects were involved in some type of endurance training (4 of the 5 men and 4 of the 6 women were training as recreational triathletes) with an average weekly participation of 5-6 days and a history of regular cycling. All female subjects reported regular menstrual flow for at least the last 6 mo before the study. Female subjects recorded their days of menses and their morning oral temperature for 2 mo before all exercise trials to accurately predict the timing associated with ovulation and the luteal phase. All the female subjects were tested between day 22 and day 27 after the onset of menses [mean estradiol (E₂) = 172 ± 29 nmol/l]. The same women were also tested as part of another study during the follicular phase of the cycle (days 4-6 after the onset of menses, mean E₂ = 55 ± 7.5 nmol/l). We previously demonstrated that glucose R_a and R_d and total CHO oxidation are significantly reduced in the luteal phase compared with the midfollicular phase of the cycle (43). To our knowledge, no published reports have attempted to study gender differences in substrate utilization using female subjects after ovulation, during the luteal phase of the menstrual cycle.

Body composition analyses. Percent body fat and FFM were calculated from hydrostatic weighing corrected for residual lung volume. Subjects performed repeated underwater weight trials until three values within 100 g were obtained. With the subject in a seated position outside the tank, residual volume was measured at least three times using the helium-dilution technique (Collins Modular Lung Analyzer, Greensboro, NC). Body density was converted to percent body fat using the age- and gender-specific formulas suggested by Heyward and Stolarkzyck (20).

Cycling O_2 peak and LT tests. The O_2 peak testing protocol began with three 4-min steady-state stages of increasing power outputs (75, 125, and 175 W for the female protocol and 100, 175, and 250 W for the male protocol) on a Schwinn Velodyne cycle ergometer that was calibrated before each trial. Immediately after the third stage, the power output was increased 25 W/min until volitional exhaustion. The criteria for O_2 peak were a plateau in O₂ from at least two 20-s values, RER > 1.1, and volitional exhaustion.

Two-stage submaximal exercise trial. Subjects reported to the laboratory 10 h after their last meal. An indwelling Teflon catheter was inserted into an antecubital vein in each arm (18-20 gauge, 1.25-in. Angiocath). After a resting blood sample was taken for background isotopic enrichment, a primed (30 µmol/kg) constant (0.42 µmol · kg¹ · min¹) infusion of [6,6-²H]glucose (Cambridge Isotopes Laboratories, Woburn, MA) was initiated into one arm vein. After 90 min of constant infusion, subjects cycled at a power output corresponding to 70% LT for 25 min immediately followed by 25 min of cycling at 90% LT. Expired gases were monitored during the last 5-10 min of each of the two stages for O₂ and RER using the Teem 100 metabolic system with the medium-flow pneumotach, as described above. The metabolic system was calibrated as described above before the collection of expired gas for each workload. Blood samples for glucose and glucose isotopic enrichment were obtained every 5 min during exercise and placed into tubes containing EDTA. Blood samples for all other metabolites and hormones were collected during the last 5 min of each 25-min stage. Subjects refrained from exercising 36 h before the submaximal exercise trials. Also, subjects submitted a 2-day dietary record before the submaximal exercise trials to ensure diet adherence. All subjects consumed 4 g CHO · kg¹ · day¹ before each trial.

Metabolite and hormone assays. Plasma samples were obtained and frozen at 30°C for further analysis of glucose, lactate, glycerol, and insulin. Glucose and lactate concentrations from the submaximal trials were analyzed with an enzymatic assay (Stat 2000 analyzer, Yellow Springs Instruments). Glycerol concentrations were determined with a commercially available spectrophotometric assay (Sigma assay 337A). Insulin and E₂ were measured with commercially available double-antibody radioimmunoassay kits (Diagnostic Products, Los Angeles, CA). All samples were analyzed in duplicate.

Isotopic enrichment and calculation of glucose kinetics. The ratio of [6,6-²H]glucose to unlabeled glucose (isotopic enrichment) was determined by forming the pentaacetate derivative of glucose and using gas chromatography-mass spectroscopy to selectively monitor the peak abundances of mass-to-charge ratio of 200, 201, and 202 (41). Glucose R_a and R_d from the circulation were calculated with the non-steady-state equations of Steele (35) and spline fitting (42), and the average of 20 and 25 min (70% LT) and the average of 45 and 50 min (90% LT) were then determined. The volume of distribution was set at 150 ml/kg for the calculations of R_a and R_d (4).

Substrate oxidation. Total energy expenditure and CHO and fat oxidation patterns were calculated from the equations established by Frayn (11). The relative contributions of plasma glucose and muscle glycogen to total CHO oxidation were established from indirect calorimetry and glucose R_d. It was assumed that 100% of the calculated glucose uptake (R_d) was oxidized in the skeletal muscle (3). However, this assumption may lead to an underestimation of muscle glycogen utilization. Therefore, the calculated relative contribution of muscle glycogen to total CHO oxidation represents a minimal rate of glycogen use. There is no evidence to suggest that there are differences between genders in the percentage of glucose R_d oxidized in the working muscle.

Statistical procedures. Descriptive data were analyzed between men and women using an independent t-test. All other dependent variables were compared using a mixed-design ANOVA (gender × intensity). Specific cell comparisons were done using a series of a priori planned comparisons with the SuperAnova statistical package (Abacus, Berkeley, CA). The number of uncorrected comparisons was limited to the degrees of freedom (df) for the difference among the means. When the number of desired comparisons exceeded the degrees of freedom, the experiment-wise alpha was adjusted [(0.05 · df)/desired number of comparisons]. The level of significance was set at an overall experiment-wise alpha of 0.05. Values are means ± SE.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Descriptive data. Comparative baseline descriptive measurements are summarized in Table 1. There was no statistical difference in the total body weight of the men and women. However, men were taller, with a lower percent body fat and a higher overall FFM. Absolute and relative (ml/kg and ml · kg FFM¹ · min¹) O_2 peak were significantly lower in the women. Blood values for hemoglobin and hematocrit were significantly higher in the men.

Workload intensities. The workload intensities and heart rate responses are summarized in Table 2. There were no differences between men and women in the exercise intensities at either workload expressed relative to the LT. However, the exercise intensity in watts and expressed relative to O_2 peak was significantly higher for the men during the 70 and 90% LT workloads. Similarly, the heart rate response was significantly higher in men during both workloads.

Metabolic data. Substrate oxidation calculated from indirect calorimetry is summarized in Table 3. There were no differences between the men and women for either of the workloads for RER and percent fat and percent CHO oxidized. Comparisons between the two workloads showed that the women did not exhibit a workload-dependent change in RER and percent CHO and fat oxidization. In contrast, the men exhibited an increase in RER and percentage of CHO oxidation with a concomitant decrease in the percentage of fat oxidation across the two intensities. Total energy expenditure (kcal/min) was higher for the men at both workloads. Men and women showed a significant increase in total energy expenditure from 70 to 90% LT. The total CHO and fat oxidation (per kilogram total body mass per minute) was significantly higher for the men at both workloads. The women showed a significant increase in oxidation of both fuels from 70 to 90% LT. However, the men showed a significant increase in CHO oxidation with no change in fat oxidation from 70 to 90% LT. Figure 1 demonstrates the differences between genders for CHO oxidation expressed relative to FFM.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Carbohydrate (CHO) oxidation relative to fat-free mass (µmol · kg FFM1 · min1) during the 2-stage submaximal exercise trials. Values are means ± SE. *P < 0.05 vs. male subjects; P < 0.05 vs. 70% lactate threshold (LT).

Plasma metabolites and insulin. Table 4 summarizes the plasma metabolite and hormone response to the two-stage experimental protocol. Although there were no differences in plasma glucose concentration at rest, the women maintained a higher blood glucose concentration at 70% LT. However, there was no difference in the plasma glucose concentration between the men and women at 90% LT. The men showed a significant difference in the plasma glucose concentration between 70 and 90% LT. However, there were no differences in the plasma glucose concentrations across the workloads in the women. Plasma insulin concentration was significantly higher for the women at rest and during the 70 and 90% LT workloads. The women showed a significant increase in blood lactate concentration from 70 to 90% LT. However, there were no differences between the 70 and 90% LT workloads in the men. There were no differences in plasma lactate concentrations between the men and women at either workload.

Plasma glucose kinetics. Figures 2 and 3 summarize the glucose R_a and R_d at rest and during the 70 and 90% LT workloads relative to FFM (µmol · kg FFM¹ · min¹). There were no differences between men and women in glucose R_a at rest and during the 70 and 90% LT workloads. Although the means demonstrate a trend toward higher values in the men, calculated-effect size using Cohen's d demonstrated a moderate effect (d = 0.506). With the use of Cohen's d, it was calculated that, even with a sample size >15, similar results would be obtained. The same results were observed for glucose R_d. Metabolic clearance of glucose was not significantly different between the men and women for either workload: 5.1 ± 0.7 and 7.9 ± 1.1 ml · kg¹ · min¹ at 70 and 90% LT, respectively, for men and 4.2 ± 0.6 and 6.5 ± 1.3 ml · kg¹ · min¹ at 70 and 90% LT, respectively, for women.

View larger version (12K): [in this window] [in a new window]   Fig. 2.   Rates of glucose appearance (Ra) relative to fat-free mass for the 2-stage submaximal exercise trials. Values are means ± SE.

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Rates of glucose disposal (Rd) relative to fat-free mass for the 2-stage submaximal exercise trials. Values are means ± SE.

1

1

1

1

1

1

1

1

1

1

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Relative contribution of plasma glucose and muscle glycogen to total CHO oxidation for the 2-stage submaximal exercise trials. Values are means ± SE. *P < 0.05 vs. male subjects at 70% LT; P < 0.05 vs. male subjects at 90% LT.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The purpose of this investigation was to determine whether plasma glucose R_a and R_d and CHO oxidation varied between recreationally trained men and women at exercise intensities relative to the LT. Although the present group of men demonstrated a higher glucose R_a and R_d expressed relative to total body weight, when the data were expressed relative to FFM, there were no differences between the groups. The most notable finding in the present investigation was that the relative contribution of blood glucose (to total CHO oxidation) was considerably higher in the women with a concomitant decrease in the contribution of muscle glycogen (Fig. 4).

Previous research has suggested that the reproductive hormone 17-estradiol (E₂) may suppress glycogen utilization (19, 24, 25, 33). Consistently, it was demonstrated that women use less muscle glycogen than similarly trained men during submaximal endurance exercise (19, 24, 25, 33). Kendrick et al. (24, 25) also demonstrated suppression in muscle glycogen use during exercise in rats treated with E₂. Although it is an indirect measure of glycogen utilization, Friedlander et al. (12) suggested that the relative contribution of blood glucose may be higher in women. Tarnopolsky et al. (37) also demonstrated that women have a limited ability to maximize glycogen stores in response to a 4-day CHO-loading protocol and that this limitation may be related to a lower energy intake in women (39). Collectively, the human data suggest less reliance on muscle glycogen by women than by men. However, whether the decrease in glycogen utilization is due to limitations in muscle glycogen availability (or storage) or some other mechanism associated with stimulation of glycogenolysis remains unclear. Although our data do not show direct measures of glycogen depletion, they demonstrate an increased relative contribution of blood glucose to total CHO oxidation in women. The increased reliance on blood glucose in women requires further investigation.

Research by Hansen et al. (18) may in part explain the differences in the relative contribution of CHO sources in women. Ovariectomy in female rats did not change muscle glycogen or GLUT-4 concentration in the skeletal muscle (soleus, epitrochlearis, and flexor digitorum brevis) at rest or in response to insulin-stimulated glucose transport. In contrast, when muscle tissue was subjected to in vitro electrical stimulation, ovariectomy resulted in decreased glucose transport. However, total GLUT-4 protein content was similar between the sham and ovariectomized animals. Although the total GLUT-4 protein content and insulin-stimulated glucose uptake may be unaffected by abnormally low concentrations of ovarian hormones, the translocation of the GLUT-4 proteins to the plasma membrane may be altered. This mechanism requires further investigation. The possibility that E₂ and progesterone may enhance or inhibit the translocation process may provide insight into metabolic regulation during pregnancy.

Campbell and Febbraio (2) demonstrated an E₂-dependent increase in 2-[1-¹⁴C]-deoxy-D-glucose uptake in ovariectomized rats treated with subcutaneous E₂ pellets (timed release of E₂) compared with untreated and progesterone-treated animals. Interestingly, glucose uptake by the muscle was only increased with E₂ administration. Administration of progesterone alone resulted in rates of glucose uptake similar to those in untreated ovariectomized animals. These recent data indicate that circulating levels of E₂ may stimulate GLUT-4 translocation and promote glucose uptake during contraction. Consequently, the effects of progesterone apparently antagonize elevated rates of glucose uptake by the muscle. These results also suggest that the mechanisms for disparity in glucose metabolism at rest between men and women are controlled differently compared with exercise and may further emphasize the importance of blood glucose maintenance in women.

Our glucose kinetic data are in agreement with a recent study by Marliss et al. (28). Regardless of the possible differences in the LT among subjects, glucose R_a (mg · min¹ · kg¹) was similar between men and women during exercise at 88% of O_2 peak. However, glucose R_d (mg · min¹ · kg¹) was lower in the women. When R_a and R_d were expressed relative to FFM (mg · kg FFM¹ · min¹), there were no differences between men and women during the short exercise trial. Our present data show a trend identical to that of Marliss et al. for glucose R_a and R_d expressed relative to FFM (no difference between men and women). Marliss et al. also quantified glucose kinetics during a 120-min recovery period. Although there were no differences during exercise, during the recovery period, women exhibited higher circulating insulin and glucose R_d (relative to FFM) than men. The authors suggest that women may be more efficient at restoring muscle glycogen from endogenous glucose after exercise. The hyperinsulinemic response noted by Marliss et al. during recovery is also of interest given the present data. The women in our study demonstrated a higher circulating insulin concentration at rest and during both exercise intensities than the men. However, circulating insulin levels remained consistently low during exercise. Therefore, it is not likely that the circulating insulin contributed significantly to the dominant contraction-mediated glucose uptake during exercise. Although this did not translate to a higher glucose R_d in the women, it may have contributed to the larger relative contribution of blood glucose to total CHO oxidation.

The majority of past research that has described gender variation in substrate utilization has relied heavily on indirect calorimetry. The present data are in agreement with previous research reporting no difference in RER between men and women (5, 8, 16, 31) but are in disagreement with other reports demonstrating that women have a lower RER during exercise (1, 13, 14, 22, 30, 36-38). One possible explanation for the observed lower RER in women in other studies may be related to differences in LT between the male and female subjects. Coyle et al. (6) and Coggan et al. (3) reported that cyclists with high LT have lower RER values than those with lower LT at a given percentage of O_2 peak. It is possible that the female subjects in some of the studies that reported lower RER had higher LT and thus exercised at a lower percentage of LT than the men when exercised at a given percentage of O_2 peak. Although Phillips et al. (30) showed similar LT between men and women after matching subjects on training and competitive habits (including training intensity, duration, frequency, common competitive distances, and years of training), the possible discrepancies in LT among subjects are not discussed in the previous literature. Interestingly, Davis et al. (8) also compared men and women in the follicular phase at the same percentage of the anaerobic threshold (as calculated from the V-slope method) and observed similar RER values between men and women, which supports the present findings.

Previous literature has suggested that, for an adequate comparison of substrate utilization between men and women, subjects should be carefully selected and equated on detailed training and competitive histories (30, 36-38). However, matching similarly trained men and women without subjecting them to similar training regimens has also been criticized, in that it does not adequately allow for the collection of directly comparable data in men and women (12). Furthermore, a similar O_2 peak does not guarantee a similar metabolic response to exercise above and below a subject's LT (3). Other studies that have compared men and women have noted a significant difference in the O_2 peak values of the subjects. Even when O_2 peak is expressed relative to FFM, women demonstrated values 13.5% lower than men (8). Skinner et al. (34) describe the gender differences across 20 wk of cycle ergometer training in the HERITAGE Family Study. Before training, women (n = 346) demonstrated a 13% lower O_2 peak (ml · kg FFM¹ · min¹) than men (n = 287). The difference in O_2 peak between the men and women in the present study was 19%.

On the basis of the recent work cited above, matching men and women using the criteria established by Phillips et al. (30) has advantages in terms of maximizing internal validity. However, it is unclear how this may suppress external validity and act to minimize the sexual dimorphisms during moderate-intensity exercise. If the data of Skinner et al. (34) represent an average gender difference in the population, an argument could be made that, to maximize external validity, subjects who demonstrate a similar discrepancy in O_2 peak (relative to FFM) should be selected. The rationale for selecting an exercise intensity relative to the LT was taken from previous research that has demonstrated that exercise performance is more related to ventilatory threshold (21, 27) and LT (6, 7) than to O_2 peak. Similarly, the data of Coggan et al. (3) support the concept that glucose R_a and R_d are related to LT. Without controlling for the LT, some subjects may be performing the experimental trial well below LT, whereas others may be forced to work above the LT.

Female subjects in the present study were tested in the luteal phase of the menstrual cycle. CHO oxidation (9, 10, 15, 17, 40, 43) and lactate concentrations (23, 26, 29) are lower during the luteal phase. In a recent study from our laboratory, we noted pronounced differences in fuel utilization and glucose kinetics across the menstrual cycle in fasted female subjects (43). Indirect calorimetry data demonstrated a significantly lower total CHO oxidation for the luteal than for the follicular phase of the cycle during exercise at 90% LT: 82.0 ± 12.3 vs. 93.8 ± 9.7 µmol · kg¹ · min¹. Glucose R_a and R_d were also significantly lower during the luteal than during the follicular phase of the cycle at 90% LT. Although our previous study indicated higher circulating E₂ and, consequently, lower glucose R_a and R_d during the luteal phase, the relative contribution of muscle glycogen was similar for both phases of the menstrual cycle and consistently lower than that of the men. The data of Tarnopolsky et al. (36) demonstrating decreased glycogen depletion in women are in agreement with this observation (although women were tested in the early-midfollicular phase of the menstrual cycle).

In the previous research by Romijn et al. (32), the relative contribution of glucose and glycogen to total CHO oxidation changed in response to the intensity and duration of the exercise. However, Romijn et al. studied glucose kinetics at 25, 65, and 85% of O_2 peak. In the present investigation, the two exercise intensities were different by ~10-12% O_2 peak, which, in part, explains the similar RER and relative contribution of CHO sources at 70 and 90% LT in the female subjects. In contrast, the men demonstrated an increase of ~15% O_2 peak between the two intensities. This is closer to the 20% difference built into the design of Romijn et al. and may explain the intensity-dependent increase in RER noted in the men.

In conclusion, the results of this study indicate that at exercise intensities near the LT, glucose R_a and R_d expressed relative to FFM (µmol · kg FFM¹ · min¹) are similar in men and women. There were also no significant gender differences in total CHO and fat oxidation relative to total energy expenditure. The most notable finding in the present study indicates that the relative contribution of blood glucose to total CHO oxidation is higher in women. The precise mechanism(s) associated with the increased relative contribution of glucose in women may include but may not be limited to E₂- or progesterone-mediated alterations in GLUT-4 translocation or muscle glycogenolysis and decreased muscle and/or hepatic glycogen stores. These results also suggest the need to reevaluate the importance of CHO feedings and the maintenance of blood glucose during exercise in women.