Is the daily FSH, LH and PdG relation an accurate predictor of luteal-follicular transition? A machine learning approach for hormonal self-testing
Methods, such as calendar calculations, examination of cervical mucus and a change in basal body temperature, as well as ovarian morphology, have been historically used to monitor and predict ovulation . However, all these methods rely on physiological signs which are in turn governed and triggered by hormonal changes. Therefore, a more direct and primary observation of the menstrual cycle and ovarian activity is the individual hormonal fluctuations . LH, FSH, Estrogen (E), and PdG are four key urinary metabolites , largely studied as markers of the cycle because of their abundance, and ease of self-detection at the point-of-care . The relationship between urinary and ovarian excretion rates of such metabolites is well known , with strong correlations between urinary and serum concentrations .
The variability of the menstrual cycle, and the continuum of ovarian function throughout the reproductive life , makes fertility prediction, based solely on information from previous cycles, overly inaccurate [8,9]. Today, point-of-care urinary hormone assays are non-invasive, provide timely results without the need for expensive equipment, and are easier than serum tests . Also, self-testing promotes awareness and responsibility in patients at home .
The timing of ovulation following the hormonal variations along the cycle: a rise or surge, a peak, or decrease; has been widely studied and predicted with high accuracy based on this evidence . LH has a sharp rise about 24 to 48 hours (av. ~33h) before the ovulation date, this is also referred to as an LH surge or rise . The LH peak occurs around 6 to 28 hours (av. ~17h) before ovulation . Similarly, the FSH level rises 2 to 19 hours (av. ~12h) before ovulation and shortly after LH . FSH has another particular characteristic: it experiences a dip approximately 6 days before ovulation during the variable phase of the cycle.
The Fertile Window usually identifies 3 to 6 days when fertility is highest each cycle, although it can span up to 10 days . After ovulation, the egg survives for up to 24 hours , during this time it can be fertilized by sperm, however, sperm can survive from to 2 up to 6 days after intercourse , depending on sperm quality and interaction with cervical fluid. This gives the characteristic distribution of the chances of getting pregnant from the ovulation date, this is the most important characteristic of the Fertile Window, its correlation to the ovulation day.
The Pearl Fertility app runs calculations directly on the data captured in the phone without the need of an internet connection. This calculations of the statistical distribution of the predicted ovulation result in an ovule displayed as a circle on the date with the highest probability of ovulation. From this calculation, a Fertile Window is calculated from an proprietary statistical model built with results of over 20 clinical studies. The ovulation calculation and fertile days estimation with the less variability is the one shown to the user.
1) Pearl builds a hormone profile overtime and predicts ovulation with every measurement.
2) The best prediction (less variability) is chosen to display the Fertile Window
3) The Fertile Window is calculated and shown as a Flower Graph which opening and closing is a function of the probability of getting pregnant if intercourse happens on the given date.