Cardiovascular response to short-term fasting in menstrual phases in young women: an observational study

The subjects were recruited from the local university with no incentives. We enrolled seven healthy female students who satisfied the inclusion and exclusion criteria for entry into the study. The inclusion criteria were female gender and age of 20 to 24 years. The exclusion criteria were past and current smoking, alcoholism, high-performance athletes, medication use, oral contraceptive use, irregularity of menstrual cycle, parity, and body mass index ≥ 25 kg/m². All subjects were carefully informed about the purpose and potential risks of this experiment, and all gave written informed consent to participate in this study. The study protocol was approved by the Human Ethics Committee of the Graduate School of Human Development and Environment, Kobe University.

The experiments were conducted from February 2012 to February 2014. The experimental procedure is shown in Fig. 1a and b. All subjects participated in a total of four experimental sessions during two successive phases (follicular and luteal phase in the same menstrual cycle, or luteal phase and follicular phase in the next menstrual cycle) according to a randomized crossover design. In the first menstrual cycle, the first and second sessions were conducted. In the second menstrual cycle, the third and fourth sessions were conducted. In the first and second sessions in the first menstrual cycle, the subjects were randomly assigned to each trial: (1) the meal intake trial or (2) the fasting trial. After a 12-h overnight fast, subjects assigned to the meal intake trial were served a test meal, and those assigned to the fasting trial continued fasting for a further 2 h (total 14 h). If the first session was conducted in the follicular phase, the second session was conducted in the luteal phase in the same menstrual cycle. If the first session was conducted in the luteal phase, the second session was conducted in the follicular phase in the next menstrual cycle. Third and fourth sessions were also conducted in the same way as the first and second sessions. In the first and second sessions, the subjects were randomly assigned to the meal intake trial or the fasting trial. In the third and fourth session, the subjects were assigned to other than the trial assigned in each follicular or luteal phase in the first and second sessions. Subjects were determined as being in the follicular phase or luteal phase based on the occurrence of menstruation and ovulation. Ovulation was recognized by a surge in luteinizing hormone detected using two self-examination kits for luteinizing hormone (L-check FT; Nipro Co., Ltd, Osaka, Japan and P-check; Mizuho Medy Co., Ltd, Saga, Japan).

On the day before testing, the subjects in both the meal intake trial and the fasting trial finished their meals by 21:00 h and went to sleep before 24:00 h. They abstained from drinks containing caffeine, food containing capsaicin, alcohol, and sports activity for 24 h prior to each testing day. After overnight fasting, each subject arrived at the laboratory between 08:00 h and 08:30 h and rested in the supine position for 20 min in an environmentally controlled room (temperature, 23–26 °C; relative humidity, 50–60 %). After resting, baseline physiological and psychological assessments were conducted. Saliva samples were collected at the same time, and stored at −25 °C for assessment of salivary cortisol levels. The physiological assessments included an electrocardiogram, and measurement of systolic blood pressure (SBP) and diastolic blood pressure (DBP). SBP and DBP were measured by the Riva-Rocci-Korotkoff method using a digital manometer (GP-303S, Parama-Tech Co., Ltd, Fukuoka, Japan). A test meal was served for the meal intake trial at 09:50 h, and the subjects finished eating by 10:05 h. Postprandial physiological and psychological data and a saliva sample were obtained at 20, 40, 60, and 80 min after meal intake. For the fasting trial, a test meal was not served, but physiological and psychological data and a saliva sample were obtained 20, 40, 60, and 80 min after the corresponding time point of intake of food in the meal intake trial.

The meal for the experiment consisted of 81 g of bread made from bread flour and white sugar, 26 g of ham, and 200 g of orange juice, which together contained 395 kcal of energy, 15.5 g of protein, 11.3 g of fat, 57.5 g of carbohydrate (energy %: protein/carbohydrate/fat 15/62/23), and 0.8 g of sodium. The overall energy contained in the meal and the energy proportions of carbohydrate, protein, and fat were determined according to the mean intake of food in females in their 20s reported in the National Nutrient Survey in Japan [41]. These meals were served at 36.0 °C. Meal temperature was measured using a non-contact thermometer (Thermo-Hunter, model PT-2LD; Optex Co., Otsu, Japan). All drinks were provided at room temperature. In the fasting trial, each subject continued fasting throughout the experiment for a further 2 h.

The R-R intervals, the time interval between two consecutive R waves in the electrocardiogram, were continuously recorded before and after meals by an ambulatory electrocardiograph monitor (Active Tracer AC-301A; GMS Inc., Tokyo, Japan) with a sampling rate of 1 kHz. Power spectral analysis of HRV was performed using a 16th-order autoregressive model [42] by Kubios HRV analysis software 2.0 (Biomedical Signal and Medical Imaging Analysis Group, Department of Applied Physics, University of Kuopio, Finland) [43].

We analyzed low frequency power (LF: 0.04–0.15 Hz) and high frequency power (HF: 0.15–0.40 Hz) as HRV parameters. LF power reflects sympathetic and parasympathetic modulation of the heart, whereas HF power primarily reflects parasympathetic modulation of the heart [25]. LF and HF power were computed for each minute of the 5-min R-R data [44]. The LF/HF ratio was calculated as an index of sympathovagal balance [25]. The mean of each 1-min HRV was used for statistical analysis.

All frozen samples were thawed, mixed, and centrifuged at 3000 × g for 10 min at room temperature. Samples were analyzed in duplicate with a competitive immunoassay specifically validated for the quantitative measurement of salivary cortisol (Salimetrics LLC, State College, PA, USA). The calibration range was 0.012–3.0 μg/dL.

Data are shown as means ± standard error. The Student’s t test was used to detect differences between the trials at baseline. A two-way repeated-measures analysis of variance (ANOVA) was used to investigate the effects of meal conditions (eating or fasting), the effects of time, and interaction effects between condition and time on physiological parameters. The Bonferroni test was used for post-hoc analysis. Effect sizes are reported as generalized eta squared (η _G ² ), which allows direct comparison of within- and between-subjects designs [45], and interpreted according to Cohen’s recommendation of .02 for a small effect, .13 for a medium effect, and .26 for a large effect [46]. Statistical analysis was performed using SPSS Statistics 19 (IBM Inc., Tokyo, Japan). A P value less than 0.05 was considered statistically significant.

The mean age ± standard deviation of subjects was 22.3 ± 1.0 years. The subjects’ height, body weight, and body mass index were 158.0 ± 3.8 cm, 50.7 ± 4.6 kg, and 20.5 ± 2.1 kg/m², respectively. All subjects completed a meal in approximately 15 min in each session. Table 1 shows baseline data for physiological characteristics. There was no difference in the baseline data.

Changes in heart rate (HR) in the follicular phase are shown in Fig. 2a. Two-way repeated-measures ANOVA showed that time and meal condition had a significant main effect on HR (time, P < 0.001, η _G ²  = 0.060; meal, P < 0.001, η _G ²  = 0.118). There was also a significant interaction effect between time and meal condition for HR (F [4, 24] = 6.34, P < 0.001, η _G ²  = 0.056). Post-hoc testing showed that HR in the eating trial was significantly increased at 20, 40, 60, and 80 min after the meal compared with baseline (20 min, P = 0.026; 40 min, P = 0.004; 60 min, P = 0.017; 80 min, P = 0.045). There was no significant change in HR in the fasting trial. In addition, HR was significantly higher in the eating trial than in the fasting trial at 20, 40, 60, and 80 min after the meal (20 min, P = 0.005; 40 min, P = 0.029; 60 min, P < 0.001; 80 min, P = 0.005). Changes in HR in the luteal phase are shown in Fig. 2b. Two-way repeated-measures ANOVA showed that time and meal condition had a significant main effect on HR (time, P = 0.002, η _G ²  = 0.039; meal, P = 0.012, η _G ²  = 0.120). There was also a significant interaction effect between time and meal condition with HR (F [4, 24] = 5.79, P = 0.002, η _G ²  = 0.030). Post-hoc testing showed that HR in the eating trial was significantly increased at 20, 60, and 80 min after the meal compared with baseline (20 min, P = 0.014; 60 min, P = 0.037; 80 min, P = 0.034). HR in the fasting trial was significantly higher at 80 min after the meal compared with 20 min after the meal (P = 0.029). In addition, HR was significantly higher in the eating trial than in the fasting trial at 20, 40, and 60 min after the meal (20 min, P = 0.001; 40 min, P = 0.009; 60 min, P = 0.025).

Changes in SBP in the follicular and luteal phases are shown in Fig. 3a and b. There was no significant main effect or interaction effect for SBP.

Changes in DBP in the follicular phase are shown in Fig. 4a. Two-way repeated-measures ANOVA showed that time had a significant main effect on DBP (P = 0.03, η _G ²  = 0.019). There was also a significant interaction effect between time and meal condition for DBP (F [4, 24] = 5.48, P = 0.003, η _G ²  = 0.058). Post-hoc testing showed that DBP was significantly higher in the fasting trial than in the eating trial at 40, 60, and 80 min after the meal (40 min, P = 0.014; 60 min, P = 0.049; 80 min, P = 0.03). Changes in DBP in the luteal phase are shown in Fig. 4b. Two-way repeated-measures ANOVA showed that time had a significant main effect on DBP (P = 0.037, η _G ²  = 0.040). There was no significant interaction effect between time and meal condition for DBP. Post-hoc testing showed that DBP was significantly higher in the fasting trial than in the eating trial at 40 min after the meal.

Changes in the LF/HF ratio in the follicular and luteal phases are shown in Fig. 5a and b. Two-way repeated-measures ANOVA showed that meal condition had a significant main effect on the LF/HF ratio in the luteal phase (P = 0.038, η _G ²  = 0.095). Post-hoc testing showed that the LF/HF ratio was significantly higher in the eating trial than in the fasting trial at 60 min after the meal (P = 0.015). There were no other main or interaction effects for the LF/HF ratio.

Changes in HF power in the follicular phase are shown in Fig. 6a. Two-way repeated-measures ANOVA showed that meal condition had a significant main effect on HF power in the follicular phase (P = 0.047, η _G ²  = 0.088). There was also a significant interaction effect between time and meal condition for HF (F [4, 24] = 2.83, P = 0.047, η _G ²  = 0.061). Post-hoc testing showed that HF power was significantly higher in the fasting trial than in the eating trial at 60 and 80 min after the meal (60 min, P = 0.038; 80 min, P = 0.018). Changes in HF power in the luteal phase are shown in Fig. 6b. Two-way repeated-measures ANOVA showed that time had a significant main effect on HF power in the luteal phase (P = 0.042, η _G ²  = 0.060). There was also a significant interaction effect between time and meal condition for HF (F [4, 24] = 3.09, P = 0.035, η _G ²  = 0.071).

Changes in salivary cortisol concentrations in the follicular phase are shown in Fig. 7a. Two-way repeated-measures ANOVA showed that time had a significant main effect on salivary cortisol concentrations in the follicular phase (P = 0.003, η _G ²  = 0.065). There were no other main or interaction effects in salivary cortisol concentrations. Changes in salivary cortisol concentrations in the luteal phase are shown in Fig. 7b. Two-way repeated-measures ANOVA showed that time and meal condition had a significant main effect on salivary cortisol concentrations (time, P = 0.015, η _G ²  = 0.176; meal, P = 0.028, η _G ²  = 0.319). There was also a significant interaction effect between time and meal condition for salivary cortisol concentrations (F [4, 24] = 3.45, P = 0.023, η _G ²  = 0.066). Post-hoc testing showed that salivary cortisol concentrations were significantly higher in the eating trial than in the fasting trial at 20, 40, and 80 min after the meal (20 min, P = 0.011; 40 min, P = 0.011; 80 min, P = 0.037).

There are some limitations to this study. First, the meals used in the present study had a complicated nutrient content, whereas fasting was a simple experiment. Therefore, the underlying mechanism for the present results is still unclear. The autonomic nervous system engages in maintenance of homeostasis by cooperation with other regulatory systems. Therefore, the underlying mechanism in this study may be explained from the viewpoint of whole-body coordination in future studies. Second, we conducted experiments with a controlled setting, which may be different from a non-controlled setting. Third, we did not investigate the long-term effects of fasting. Long-term effects of food intake are ultimately necessary to clarify the influence of daily lifestyles in humans. Therefore, further studies are required to determine this issue. Fourth, the present study does not refer to longer-term stress. Future studies need to test repeated 12-h fasting to verify the effects of fasting on long-term stress. Repeated 12-h fasting may not only affect long-term stress but also address the feasibility and safety of fasting, and the influence of resumed eating on stress levels.