Breast skin temperature rhythms in relation to ovulation.

Time series analysis was used to characterize circadian breast skin temperature rhythms and to assess their potential in natural family planning. Study subjects were premenopausal women volunteers who were drawn mainly from research staff at the Tenovus Institute and from a group of women in Cardiff practicing natural family planning. These women had no palpable signs of any breast abnormality and 2 women from the Institute were taking oral contraceptives (OCs). These chronophysiological studies of the breast also indluded data from a clinically healthy 9-year-old premenarcheal girl. All measurements of breast skin temperature were generally taken at 30 minute intervals from 0700 to 2300 during wake span for each day of the study. The extremely good linear response of all series to temperature changes irrespective of the instrument meant that when calibration data have been applied to observations made during these studies it is unlikely that the absolute accuracy falls outside 0.10 degrees Centigrade. Prior to the statistical analysis of temperature changes during the menstrual cycle cosinor analysis was used to investigate the existence of circadian rhythms in women with clinically health breasts for different age groups. Circadian rhythms were demonstrated in all series investigated and the overall 3-year mean and standard deviation of the average mesor amplitude and acrophase for 4 sensors located on each quadrant of the left breast were calculated using group cosinor analyses and found to be 35.10 (0.32) degrees Centigrade 0.43 (0.10) degrees Centigrade and 16 hours and 27 minutes. Comparable values for similarly located sensors obtained from 4 subjects exhibiting normal ovulatory cycles and studied for 1 complete menstrual cycle were 34.13 (0.17) degrees Centigrade 0.64 (0.05) degrees Centigrade and 21 hours and 6 minutes. Those for 2 subjects on OCs studied for 1 cycle were 33.78 (0.24) degrees Centigrade and 22 hours and 9 minutes. Computer simulation and practical experiments based on changes in the residual sums of squares from target values obtained for days in the cycle prior to ovulation indicated that this periovulatory increase in temperature is possibly detectable within 24 hours. The use of the chronobra and associated statistics may be of value in signalling the onset of the infertile phase of the menstrual cycle.