EFFECT OF MENSTRUAL CYCLE PHASE ON THE GLUCOSE-INSULIN RELATIONSHIP AT REST AND DURING EXERCISE

Abstract

The purpose of this study was to investigate the effect of menstrual cycle phase on the insulin/glucose relationship at rest and response to sub-maximal exercise. Eight eumenorrheic women completed exercise sessions (60 min, 70% V̇O₂max) in the follicular (day 7±2) and luteal (day 20±2) phases of their cycles. At 45 min before each exercise session the subject consumend an oral glucose load (OGL) of 1 gkg⁻¹ body weight in a 400 ml solution. Blood samples were obtained before the OGL (−45 min) immediately before exercise, and at 15 min intervals until the end of exercise (60 min). Results indicated that serum glucose and insulin increased (P<0.01) due to the OGL and decreased (P<0.01) due to exercise. No significant phase differences were observed for the glucose responses: however, insulin levels were greater immediately before exercise in the luteal than follicular phase (47.2±5.8 vs. 39.3±5.1 U·ml⁻¹, respectively). The insulin-to-glucose ratio (I/G) was calculated and also showed significant changes. The I/G ratio was greater in the luteal than in the follicular phase at 0 min (6.5±0.4 vs. 5.3±0.6 mU·mmol⁻¹, respectively). These data would suggest that in eumenorrheic women the menstrual cycle impacts the insulin/glucose relationship at rest, but has no effect during exercise. Whether the mechanism of the resting effect is due to changes in pancreatic and/or target tissue functions in unclear.

Introduction

Large cyclical fluctuations in reproductive hormones occur as part of the normal women’s menstrual cycle. Of particular metabolic interest within these hormonal fluctuations are the changes found in sex steroid hormones, the oestrogens and progestogens. Substantial animal research indicates that oestradiol and progesterone can affect both lipid and carbohydrate metabolism [10,12,13,19,20]. Relative to lipids, these hormones facilitate an increased synthesis of triglycerides and enhanced lipolytic actions in muscle/adipose tissues [8,11,14,19]. Conversely, elevated levels of sex steroids and inhibit hepatic gluconeogenesis and glycogenenolysis, and decrease glycogen utilization in muscle [10,12,13,17]. These metabolic effects are thought to be due, in part, to direct (lipoprotein lipase activation [19]) and indirect (hormonal, augmented growth hormone and insulin release [3,8]) actions.

Thus far, the metabolic effects of the sex steroid hormones have principally been investigated in animals, but limited human research points to analogous findings. Nicklas et al. [18] have shown that muscle glycogen resynthesis, following diet-exercise induced glycogen depletion, is enhanced in women during their luteal phase. Similarly, it has been reported that resting circulating levels of lipids are higher, as are fat oxidation rates, in the luteal than in the follicular phase of the cycle [3,5]. In humans, the mechanism of the altered metabolism across the menstrual cycle is unclear. One hypothesis is that insulin-dependent physiological processes are increased due to the augmented circulating insulin levels found in the luteal phase when oestrogen levels are elevated [3,16]. This hypothesis, however, apparently has nor been thoroughly evaluated in humans.

Therefore, the purpose of this study was to investigate the insulin-glucose relationship in the mid-follicular phase of the menstrual cycle when sex steroid hormones are low, and in the mid-luteal phase of the menstrual cycle, when sex steroid hormones are elevated. These responses were evaluated under two conditions; a resting stste, and in response to a submaximal exercise test. This latter condition was examined because factors affecting carbohydrate/lipid metabolism significantly impact on the overall substrate metabolism in exercise [1,3,4,5]

Materials and Methods

Subjects.

A total of eight subjects were recruited to participate in this study. Their physical and menstrual characteristics are as follows (means ±SE): age 25.0±2.5 yrs., body mass 60.7±2.3 kg, and menstrual cycle lenght 29±1 days. The subjects were all classified as eumenorrheic, based upon reviewing menstrual histories and basal body temperature records. Furthermore, these subjects had participated in several previous menstrual cycle-related studies from our laboratory. This had allowed us to periodically monitor their menstrual status for approximately 6 – 12 months prior to this study, by evaluating urinary or blood hormonal levels [5,18]. All subjects had been competitive athletes at one time in the past, but now were only recreationally active. Currently, they were regularly (approx. 4 – 7 days a week) performing aerobic-type activities for 30 – 60 min daily.

Exercise Sessions.

The subjects completed a maximal oxygen uptake (V̇O₂max) test on a cycle ergometer. The protocol for this maximal test involved the subjects pedaling at 60 rpm for 4 min periods at an initial work load of 60 W, with increasing comparable work loads until volitional fatigue was reached. The menstrual cycle phase during which the V̇O₂max occurred was not controlled as it has been shown that cycle phase does not appear to impact the outcome of this measure [6]. Respiratory gases were collected continuously and monitored for V̇O₂ and carbon dioxide production (V̇CO₂). These measures were assessed with an S3A Applied Electrochemistry O₂ analyzer and a Beckman LB 2 CO₂ analyzer interfaced with an Apple IIe microcomputer (Rayfield Systems, Inc. VT, USA).

For the submaximal exercise sessions the subjects reported to the laboratory at a standardized time (12 h post-prandial) for the insertion of a catheter into the antecubial vein, then rested for 20 min. A blood sample was taken, followed by the subject consuming an oral glucose load (OGL) of 1 g glucose per kg body weight in 400 ml of chilled water (3°C) and a 45 minute rest. The subject next begun exercising on a cycle ergometer for 60 min at 70% of their V̇O₂max (as determined from the maximal exercise test). At 15 min-intervals during the exercise session respiratory gases were analyzed to determine V̇O₂, V̇CO₂ and the respiratory exchange ratio (V̇CO₂V̇O₂) [15]. After the pre-OGL blood sample (−45 min), subsequent blood samples were collected at 15 min-intervals until completion of the exercise session (i.e. −45, −30, −15, 0, 15, 30, 45, 60 min, respectively). Procedural aspects of the study led to the actual blood sampling time to vary somewhat from desired times. However, the deviations were never greater than 3 min from the scheduled collection time.

Exercise sessions in the follicular phase took place on day 7±2 (mean SE) of the cycle (first day of menses was counted as day 1), and the luteal exercise session took place on day 20±2 of the cycle. The order of the follicular vs. luteal exercise sessions was randomized and performed within 2 consecutive menstrual cycles. All testing took place during the winter/early spring seasons of the year. Subjects were asked to replicate their diet and physical activity, and to avoid sexual activity for 72 h before each of the exercise sessions.

Analytical Procedures.

Blood samples were collected in sterile tubes which were placed on ice until centrifuged (20 min, 3000xg, 4°C). The separated serum was aliquoted and stored at −80°C until later analysis. The blood samples were analyzed for insulin, glucose, and triglycerides. The insulin assay was conducted by commercial double antibody radioimmunoassay kits (INCSTAR, Stillwater, MIN, USA). The mean within- and between-assay errors for the insulin assays were 7.2 and 9.3%, respectively. Glucose was determined in whole blood immediately deproteinized in 8% percholoric acid and assayed colorimetrically (Sigma Chemicals, St.Louis, MO, USA). Triglycerides were also assessed using a colorimetric technique (Sigma Chemicals, St.Louis, MO, USA). All determinations were performed in duplicates.

Statistical Analysis.

ANOVA (for replicate measurements) was applied to data evaluation. When significant (P≤0.05) F-ratios were found, the Fisher post-hoc test was used to determine significantly different means. All data are reported as means ±SE.

Results

Exercise Responses.

The subjects had V̇O₂max values of 2.72±0.12 l· min⁻¹ or 44.8±2.0 ml· kg⁻¹· min⁻¹. These data and the physical training backgrounds of the subjects led us to consider these women to be of a moderately trained nature [15]. Table 1 shows the V̇O₂ requirements for the sub-maximal exercise sessions. No significant differences were noted between the trials. The overall exercise intensities were 70.1±1.2 and 71.3%±2.4 for the luteal and follicular test, respectively.

Table 1.

The triglyceride responses, oxygen uptake requirements (V̇O2) and respiratory exchange ratio (RER) responses during the experimental sessions in the follicular (FP) and luteal (LP) phases of the menstrual cycle (means SE)

Phase	Time (min)
	−45	0	15	30	45	60
Triglycerides (mg·dl−1)
FP	65.6 ± 12.0	69.1 ± 14.0	72.1 ± 13.3	64.1 ± 10.5	65.1 ± 8.2	69.7 ± 8.3
LP	63.4 ± 10.2	63.9 ± 11.1	67.3 ± 10.9	64.7 ± 9.2	66.8 ± 7.6	71.4 ± 7.4
V̇O2 (l·min−1)
FP	—	—	1.95 ± 0.10	2.01 ± 0.11	1.94 ± 0.13	1.97 ± 0.13
LP	—	—	1.98 ± 0.12	1.94 ± 0.12	1.94 ± 0.13	1.91 ± 0.11
RER
FP	—	—	0.984 ± 0.008	0.974 ± 0.014	0.955 ± 0.016	0.956 ± 0.016
LP	—	—	0.964 ± 0.012	0.953 ± 0.011	0.923 ± 0.011	0.911 ± 0.014

Table 1 also shows the respiratory exchange ratio responses during the submaximal exercise sessions. The ratio was found to decline in both the luteal and follicular phase trials. In the follicular trial this decline nearly reached statistical significance (P<0.01) while decline was significant in the luteal phase (P<0.06). Furthermore, a significant phase difference was obserwed at the 60 min point with the luteal value being lower than the follicular (P<0.04).

Blood Responses.

Figure 1 shows the glucose changes in the blood. The OGL produced a significant (P<0.01) elevation from −45 min in the glucose levels at −30, −15, and 0 min. Exercise subsequently introduced a significant decline from the pre-exercise (−30, −15, 0 min) glucose levels (P<0.01). The glucose levels during exercise (15 – 60 min) were not significantly less than the pre-OGL (−45min). There were no significant differences in any of the glucose responses between the two menstrual phases.

Fig. 1.

The glucose responses to the oral glucose load (OGL) and the exercise test in the follicular (FOL) and luteal (LUT) phases of the menstrual cycle

*-significantly higher from the pre-OGL and 15 – 60 exercise in both phases of the menstrual cycle

The insulin changes in the blood and are depicted in Figure 2. A significant (P<0.01) rise in resting insulin level occurred due to the introduction of the OGL, subsequently exercise produced a significant reduction although not different from pre-OGL (−45min), in insulin levels (P<0.01). A significant phase difference was observed in the peak insulin response. At 0 min (pre-exercise) the insulin was greater (P<0.02) in the luteal than follicular phase (47.2±5.8 vs 39.3±5.1 U·ml⁻¹, respecively).

Fig. 2.

The insulin responses to the oral glucose load (OGL) and the exercise test in the follicular (FOL) and luteal (LUT) phases of the menstrual cycle

*-significantly different from the respective follicular phase value

The insulin to glucose ratio (I/G) was calculated and the results are depicted in Figure 3. A significant (P<0.01) rise in the resting I/G ratio due to the introduction of the OGL was obserwed at 0 min. The exercise resulted in a significant decline (P<0.01) in the ratio (0 min compared to 60 min). Furthermore, there was a significant phase difference observed at 0 min (pre-exercise), with the I/G ratio being greater (P<0.05) in the luteal than follicular phase (6.5±0.4 vs 5.3±0.6 U·mol⁻¹, respectively).

Fig. 3.

The insulin:glucose ratio responses to the oral glucose load (OGL) and the exercise test in the follicular (FOL) and luteal (LUT) phases of the menstrual cycle

*-significantly different from the respective follicular phase value

Triglyceride responses are shown in Table 1 and were unaffected by the OGL, exercise, and menstrual cycle phase (P>0.05).

Discussion

The major finding from this study is that blood insulin responses to a glucose challenge varied across the menstrual cycle phases. However, the blood glucose response to neither the glucose challenge nor exercise was different in the luteal or follicular phases.

The rise in glucose and insulin in response to the OGL, and subsequent decline due to exercise, were expected findings and agree with other studies in the literature [4,8]. The greater luteal insulin response to the OGL at 0 min is congruent with previously published reports for animals [16]. However, this phase dependent insulin response has apparently not been demonstrated in trained eumenorrheic women before, although it is analogous with reports for women using oral-contraceptives [11]. We speculate the greater luteal insulin response is attributed primarily to the elevated oestrogens levels of the luteal phase effecting the beta-cell of the pancreas and/or the target tissue receptors of insulin [13,16]. The exact physiological mechanism causing this effect can not be determined from the present data. It should be noted that this finding of an augment luteal response is in direct contradiction to the findings of Bonen et al. [2]. In that study no phase difference in insulin response was observed after an OGL of 1.5g per kg body weight. We are uncertain as to the reason for the variance in our findings, however the subject training level, ages, and the OGL protocol does differ considerable between our two studies.

The lower respiratory exchange ratio in the luteal phase at 60 min suggest that a greater amount of lipid was used for energy production during that point in exercise for the luteal than follicular phase. In fact, if standard calorimetry calculations are applied to the ratios at 60 min [15], the lipid oxidation rates are 0.31±0.05 vs. 0.16±0.06g· min⁻¹ for luteal vs. follicular, respectively (P<0.05). This phenomenon has been shown before in exercising women by other investigators [5]. The reason for this event has been hypothesized to be related to the direct and indirect actions sex steroid hormones (primarily oestrogens) have on lipolytic processes. In the rat it has been shown that lipoprotein lipase activity and circulating lipids are greater in the presence of elevated oestradiol [7,14]. Additionally, growth hormone (which increases lipolysis) release in animals and humans also seems to be facilitated by increases sex steroid levels in the blood [8,18]. It is interesting that the calculated lipid oxidation rate difference at 60 min took place even in the presence of elevated glucose levels at the onset of exercise. A greater luteal dependence on lipid as an energy substrate could result in a carbohydrate sparing effect which could be aid to endurance based physical performance [1,4]. However, whether greater levels of lipid were actually mobilized and oxidized during the luteal phase can not be definitely confirmed from the respiratory exchange ratio measure alone. Additional measures, such as blood free fatty acid levels, would allow us to be more conclusive on this issue. We were able to assess blood triglyceride levels as an indicator of lipid responses, but as noted, this measure did not change significantly. Nonetheless, the present findings on lipid oxidation are some interesting preliminary information.

In conclusion, the present findings do indicate that the resting insulin-glucose relationship, in response to a carbohydrate challenge, varies between the menstrual cycle phases. The present findings also suggest that the typical insulin-glucose response during exercise in unaffected by cycle phase. However, substrate response to exercise seem to vary across the menstrual cycle phases. The findings from this study heve direct methodological implications for investigators conducting metabolic related research involving eumenorrheic women in a resting as well as exercise state. The results of this study would suggest that the issue of how the sex steroid hormones affect the metabolic physiology of women would warrant further investigation.