A total of 66 women who had both luteal phase defect and high blood flow impedance of the corpus luteum [corpus luteum-resistance index (CL-RI) ≥ 0.51] were recruited into this study. When serum progesterone concentrations were < 10 ng/ml during the mid-luteal phase, the patient was diagnosed as having a luteal phase defect in this study. CL-RI was measured during the mid-luteal phase, and the cutoff value was determined as described below. The mean age was 32.4 ± 4.3 years (mean ± SD), with a range of 24–41 years. The patients were non-smokers and free from major medical illness including hypertension; they were excluded if they had myoma, adenomyosis, congenital uterine anomaly, or ovarian tumors or if they used estrogens, progesterone, androgens, or had chronic use of any medication, including nonsteroidal anti-inflammatory agents or anticonvulsants.

Blood flow in the corpus luteum was measured as reported previously 16 using a computerized ultrasonography with an integrated pulsed Doppler vaginal scanner [Aloka ProSound SSD-3500SV and Aloka UST-984-5 (5.0 MHz) vaginal transducer, Aloka Co. Ltd, Tokyo, Japan]. The high pass filter was set at 100 Hz, and the pulse repetition frequency was 2–12 kHz, for all Doppler spectral analyses. After the endovaginal probe was gently inserted into the vagina, adnexal regions were thoroughly scanned. The ovary was identified, and color signals were used to detect the area with the highest blood flow within the corpus luteum. Blood flow was identified in the peripheral area of the corpus luteum 16. The pulsed Doppler gate was then placed on that area to obtain flow velocity waveforms. An acceptable angle was less than 60°, and the signal was updated until at least four consecutive flow velocity waveforms of good quality were obtained. Blood flow impedance was estimated by calculating the resistance index (RI), which is defined as the difference between maximal systolic blood flow (S) and minimal diastolic flow (D) divided by the peak systolic flow (S-D/S). Blood flow impedances were examined in the corpus luteum during the mid-luteal phase (6–8 days after ovulation). The day of ovulation was determined by urinary LH, transvaginal ultrasonography and basal body temperature records. Since the interobserver coefficient of variation for Doppler flow measurements in the present study was less than 10%, which is consistent with the reports by Ziegler et al. 32 and Miwa et al. 33, the Doppler flow measurements were judged to be reproducible.

There was a significant negative correlation between CL-RI and serum progesterone concentrations during the mid-luteal phase from the data obtained from 36 women with normal luteal function and 10 women with luteal phase defect (Fig. 1a). Receiver operating characteristic curve (ROC) analysis was performed to determine the cutoff value of the CL-RI providing the best value of the sensitivity and the specificity for determination of normal luteal function and luteal phase defect. A cutoff value of 0.51 provided the best combination with 84.3% sensitivity and 85.6% specificity to discriminate between normal luteal function and luteal phase defect (Fig. 1b).

In order to investigate whether vitamin E or L-arginine treatment has a potential to increase luteal blood flow and to improve luteal function in patients with luteal phase defect, the patients who showed both luteal phase defect and high CL-RI (≥ 0.51) during the mid-luteal phase (6–8 days after ovulation) were given vitamin E (600 mg/day, 3 times per day orally; Eisai Co., Ltd., Tokyo, Japan; n = 18), or L-arginine (6 g/day, 4 times per day orally; Now Foods, IL, USA; n = 14) during the luteal phase of the subsequent menstrual cycle.

Decreased ovarian blood flow is reported to inhibit luteal function via oxidative stress 29. Therefore, to examine a possibility that the decrease in luteal blood flow inhibits luteal function through oxidative stress in women with luteal phase defect, melatonin (3 mg at 22:00 hr orally; KAL, Park City, UT, USA; n = 13) was given as an antioxidant during the luteal phase of the subsequent menstrual cycle. We confirmed that administration of 3 mg of melatonin works as an antioxidant and suppresses oxidative stress in the human ovulatory follicle 34.

Another 10 patients received luteal support with HCG injection (2,000 IU/day, on days 3 and 5 after ovulation; Gonatropin; Asuka Co., Ltd., Tokyo, Japan).

As controls, 11 patients with both luteal phase defect and high CL-RI (≥ 0.51) during the mid-luteal phase received no medication during the subsequent menstrual cycle.

To evaluate the effect of those treatments, serum progesterone concentrations and CL-RI were measured during the mid-luteal phase (6–8 days after ovulation). Ultrasonogrphy was performed before blood sampling for serum progesterone measurement.

Venous blood was taken for the determination of serum progesterone concentrations on the day of the Doppler examination during the mid-luteal phase. Progesterone concentrations were measured by enzyme immunoassay (ST AIA-PACK PROG, Tosoh Co., Ltd., Japan) as reported previously 16. The minimal detectable concentration is estimated to be 0.1 ng/ml. Intra-assay and inter-assay coefficients of variation were 9.9% and 11.3%, respectively.

Single regression analysis, Wilcoxon signed-ranks test, and chi-squared test with Bonferroni correction were carried out using the computer program SPSS for windows 13.0. A value of P < 0.05 was considered significant.

Eighteen patients who had both luteal phase defect and high CL-RI (≥ 0.51) during the mid-luteal phase were given vitamin E during the luteal phase of the subsequent menstrual cycle. Fifteen patients out of 18 (83%) showed improved CL-RI of less than 0.51, and 12 patients (67%) developed serum progesterone concentrations of more than 10 ng/ml (Table 1). In the control group, only one patient out of 11 (9%) showed normal CL-RI and 2 patients (18%) showed normal serum progesterone concentrations (Table 1), suggesting that vitamin E significantly improved CL-RI and serum progesterone concentrations compared with the control (Table 1). Of the 12 patients whose serum progesterone concentrations improved, 11 patients showed improved CL-RI of less than 0.51. Vitamin E significantly decreased CL-RI and increased serum progesterone concentrations between the treatment cycle and the previous cycle (Table 1).

L-arginine treatment improved CL-RI in all the patients (100%), and 10 patients out of 14 (71%) developed serum progesterone concentrations of more than 10 ng/ml (Table 1). Compared with the control, L-arginine significantly improved CL-RI and serum progesterone concentrations (Table 1). L-arginine also significantly decreased CL-RI and increased serum progesterone concentrations between the treatment cycle and the previous cycle (Table 1).

Melatonin treatment improved CL-RI in 4 patients out of 13 (31%) and improved serum progesterone concentrations in 5 patients (38%) (Table 1). These effects were not significant compared with the control (Table 1). Melatonin treatment caused a significant increase in serum progesterone concentrations in the treatment cycle compared with the previous cycle, but the serum progesterone levels were less than 10 ng/ml (Table 1).

HCG treatment improved CL-RI in all the patients (100%), and 9 patients out of 10 (90%) developed serum progesterone concentrations of more than 10 ng/ml (Table 1). Compared with the control, HCG significantly improved CL-RI and serum progesterone concentrations (Table 1). HCG also significantly decreased CL-RI and increased serum progesterone concentrations between the treatment cycle and the previous cycle (Table 1).