Many tests and techniques have been used to predict or confirm ovulation in women wanting to become pregnant. The direct methods, such as laparoscopy and high-resolution transvaginal ultrasound examination, are too invasive and expensive for repeated or routine use. Various indirect methods have been developed for clinical use based on basal body temperature (BBT), the composition of cervical mucus or changes in estrogen, luteinizing hormone (LH) or progesterone in body fluids. Each of these approaches has advantages and disadvantages, but little is known about their true reliability or test characteristics. Guermandi and colleagues compared commonly used clinical ovulation tests to determine their accuracy.
The authors studied 101 women attending an Italian university clinic for management of infertility. Participants were 18 to 38 years of age (mean age: 32 years) and reported having at least 12 consecutive regular menstrual cycles of 26 to 34 days' duration before the study. The women were of normal body weight (body mass index of 19 to 24 kg per m2) and in good general health. Midluteal serum progesterone levels of at least 8.0 ng per mL (25 nmol per L) were required, and follicle-stimulating hormone (FSH), LH and prolactin levels were determined. Women with elevated FSH and LH concentrations in the early follicular phase were excluded, as were women with elevated prolactin levels in the midluteal phase. Women with signs suggesting polycystic ovary syndrome or any intra-abdominal finding that would interfere with ultrasound examination were also excluded.
Each woman was monitored for a single menstrual cycle by repeated ultrasound scans, measurements of BBT and serum progesterone levels, and urinary tests for LH levels. Transvaginal ultrasound scans were done on day 8 and then every two days until a mean follicular diameter of 14 mm was achieved. Scans were then obtained daily until there was evidence of follicular rupture or disappearance. BBT readings were taken every morning using standardized protocols. Beginning when the mean follicular diameter on ultrasound examination was 14 mm, LH levels were tested every morning and evening with a home testing kit. Testing continuedb until two consecutive positive tests occurred. Serum progesterone levels were obtained six, eight and 10 days after the expected day of ovulation, which was defined as 14 days before expected menses.
Ultrasound examination confirmed follicular development and ovulation in 97 cycles. Urinary LH surge was recorded preceding follicular rupture in all cases. In three cases, no follicular rupture could be demonstrated on ultrasound examination despite positive LH changes. The mean length of the luteal phase was 13 days based on ultrasound examination and 14 to 15 days based on LH surge. Data on BBT were not available in 11 cases. In the remaining 90 cycles, 69 (68 percent) showed a biphasic pattern. Based on BBT data, 65 ovulatory and two anovulatory cycles were confirmed by ultrasound examination. The lowest BBT (nadir) varied considerably. Compared with ultrasonography, BBT detected ovulation with a sensitivity of 0.77, specificity of 0.33 and accuracy of 0.74. A single serum progesterone estimation at the midluteal phase (six days after estimated ovulation) was as effective as repeated measurements. Compared with ultrasonography, a single progesterone level at the midluteal phase had a sensitivity of 0.80, specificity of 0.71 and accuracy of 0.79.
The authors conclude that urinary LH measures provide accurate and convenient tools to predict ovulation. In this study, a surge in urinary LH was consistently detected 72 hours before ovulation. Conversely, women found measurement of BBT to be inconvenient, and it proved to be inaccurate in predicting ovulation. Ovulation could occur from six days before the nadir of BBT until four days afterward. A single midluteal progesterone estimation using a cutoff value of 6.0 ng per mL (19 nmol per L) was also an effective confirmation of ovulation.