We surveyed 49 female university students using a questionnaire and interview. Inclusion criteria were as follows: 1) no history of varus and valgus sprains in the past 6 months; 2) no history of surgery on the lower leg; 3) no oral contraceptive or other hormone-stimulating medication usage in the preceding 6 months ; and 4) physically active less than three times per week. Among the students who were screened, 14 women (mean age, 21.1 ± 0.3 years; mean height, 159.0 ± 4.5 cm; mean weight, 53.0 ± 6.1 kg; mean cycle days, 30.1 ± 2.8 days) with regular menstrual cycles and biphasic BBTs (indicative of ovulatory cycles) were enrolled. This study was approved by the University Ethics Review Committee (Approval Number 17946). In addition, this study complied with the Declaration of Helsinki, and was conducted only after written consent was obtained from the study participants, who had been fully informed (in both verbal and written form) of the nature of the experiment.
To assess the inter-rater reliability of the measurements, 10 adult men were subjected to the same procedures on different days. The study content was fully explained to the subjects, and written, informed consent was obtained from all subjects.
Evaluation of the menstrual cycle
Based on the completed questionnaires and interviews conducted in the 49 female subjects, we asked 26 of them who had regular menstrual cycles and agreed to participate in this study to measure and record their BBT every morning for 1 to 2 months preceding the start of the experiment. Subjects were provided with basal thermometers (Citizen Electronic Thermometer CTEB503L, Citizen Systems Co., Ltd., Tokyo, Japan) for this purpose. To estimate the ovulation date, subjects were provided with ovulation kits (Doctor’s Choice One Step Ovulation Test Clear; Beauty and Health Research, Inc., CA, USA) to be used from the day after the end of menstruation. Since luteinizing hormone (LH) in urine and serum have been shown to correlate with each other , the ovulation date was estimated using the ovulation kit results as a substitute for blood sampling. A recording sheet for creation of a BBT table was prepared, and daily BBT, menstrual period, and ovulation kit results were recorded. Based on these data, the first day of menstruation was considered day 1, and the mean BBT up to day 6 was calculated. When the BBT for three consecutive days after ovulation (as determined by the ovulation kit) was at least 0.2 °C higher than this mean value, it was judged that the subject exhibited a biphasic cycle of low and high temperatures . Women with biphasic cycles were classified as having a normal ovulatory pattern, while those with monophasic cycles were considered to have an anovulatory pattern [19, 20]. Of the 26 women whose menstrual cycles were monitored, two were excluded because their BBT was monophasic; ATFL and GJL were measured in the remaining 24 subjects. The final enrolled study population consisted of 14 women who had a cycle length of 25 to 38 days  and biphasic BBTs during the menstrual cycle, and in whom ATFL length and GJL measurements were performed. Ten of the 24 subjects were excluded for the reasons indicated in Fig. 1.
Timing of measurement
ATFL length and GJL measurements were taken once in each of the four phases of the menstrual cycle; these phases consisted of the early follicular phase, late follicular phase, ovulatory phase, and luteal phase.
ATFL and GJL were measured in the early follicular phase from 3 to 4 days after the start of menstruation, in the late follicular phase from 3 to 4 days after the end of menstruation, in the ovulation phase from 2 to 4 days after the day when the ovulation kit indicated a positive result, and in the luteal phase from 5 to 10 days after the start of the high temperature phase. In consideration of possible diurnal variations, all measurements in all subjects were performed between 8:00 a.m. and 12:00 p.m..
Ultrasound imaging was performed using ultrasonography (Aplio 500, Toshiba Medical Systems, Tochigi, Japan) with a 10-MHz linear probe. The test positions used were identical to those in a previous study [23, 24] and were performed in the following order: (1) neutral ankle position with about 30° of plantar flexion, with the subject lying on their side and the lower extremity positioned on the bed; and (2) anterior drawer stress to the ankle, performed about 3 cm proximal to the lateral malleolus (Fig. 2). Ankle stress conditions were applied with a Telos Stress Device (Telos SE, Aimedic MMT, Japan). Anterior drawer stress was applied to a magnitude of 120 N for all subjects. The measurement was performed thrice, once each by three examiners, two examiners performing the test using ultrasonography, and one examiner performing the test using the Telos Stress Device. With the subject’s ankle in approximately a neutral position, the examiner palpated the anterolateral aspect of the lateral malleolus and talus. Next, the examiner applied ultrasound conducting gel over the lateral aspect of the ankle and positioned the ultrasound probe. The examiner then oriented the probe to view the cross-sectional representation of the lateral malleolus, kept on the right side of the screen, while the lateral talar articular surface cartilage and the neck of the talus, where the ATFL attaches, were identified (Fig. 3). After optimizing the image and centering these bony landmarks within the field of view, the examiner saved the three images and removed the probe. Next, the stress device was applied to the ankle and three images of the ATFL were obtained while performing the anterior drawer stress by application of a posteriorly directed force of 120 N through the tibia (Fig. 2).
Ultrasonographic image analysis was performed using an ultrasonic diagnostic imaging system. The ATFL length was measured as the linear distance (mm) between the landmarks. The anterolateral aspect of the lateral malleolus was identified as the ATFL origin, and the peak of the talus was used as the insertion point. The average of the values measured from the three images was adopted. ATFL length data from each subject were
used to calculate each subject’s normalized length change with application of anterior drawer stress (AD%) using the formula [(L stress – L neutral) /L neutral] × 100, where L is the ATFL length in millimeters .
GJL was measured using the University of Tokyo joint laxity test  (Fig. 4). Mobility was measured at the spine, and bilaterally at the hip, knee, ankle, shoulder, elbow and wrist. Each item was assigned a value of 1 point, and a total of seven positions were measured; for the six major bilateral joints (i.e., aside from the spine), the left and right positions were assigned a value of 0.5 points each. For items with joint angle as the criterion, the joint angle was measured using a goniometer. Joint angle measurements were performed by one operator and recorded by one operator.
We investigated the reproducibility of the ankle anterior drawer stress test measurements in 10 adult males (mean age, 21.0 years; mean height, 176 ± 6.5 cm; mean weight, 68.9 ± 6.3 kg) without orthopedic diseases or pain in the lower limbs. Measurement was performed using the above-described ATFL length measurement method; again, the measurement was performed three times, and the average of the three measurements was used. The measurement was repeated on two or more separate days within 1 week, and the intraclass correlation coefficient (ICC) (1, 3) was calculated. The resulting ICC (1, 3) for the ATFL measurements was 0.92–0.93. According to the criteria of Landis et al.,  reproducibility is considered to be almost perfect when the ICC is 0.81 or more. Therefore, the reproducibility of ATFL length measurement in this study was considered to be high.
Statistical analyses were performed using SPSS (version 24.0, SPSS Japan Inc., Tokyo, Japan). A one-way repeated measures analysis of variance was used to compare AD% and GJL in each phase of the menstrual cycle. Pearson’s correlation test was used to assess the relationship between AD% and GJL in each phase. Pearson’s chi-squared test was used to compare differences in assessments at the spine, and bilaterally at the hip, knee, ankle, shoulder, elbow and wrist in the ovulation phase. The level of significance was set at 5%.