A prospective, single-center, parallel-group, evaluator blinded, randomized trial was performed in the Department of Dermatology at Beijing Friendship Hospital, Capital Medical University in China. The Ethics Committee of our hospital approved the study with the number 2020-P2-248-01. The clinical trial registry number of this study is ChiCTR2100044015 (https://www.chictr.org.cn/showprojen.aspx?proj=29,280). All patients were informed of the risks, benefits, and possible complications of the process before enrollment in the study, and informed consent was obtained from each patient.

We recruited consecutive FPHL patients in our alopecia clinics from October 2020 to August 2021 if they fulfilled the following criteria: (1) aged 18–45 years; (2) female; (3) with normal hormone levels in the blood and regular menstrual cycle (estradiol, progesterone, prolactin, follicle-stimulating hormone, luteinizing hormone, and testosterone were tested during 2–5 days of the menstrual period. Patients with the hormonal level in the reference range of follicular phase could be enrolled); (4) FPHL diagnosis is proved by Sinclair class for II and III. FPHL was diagnosed based on the previous history of hair loss and the clinical and dermatoscopic appearance of the scalp.

Exclusion criteria included the following: (1) presence of other forms of alopecia; (2) presence of relative contraindications of SPT, such as insufficiency of heart, liver, and kidney, low blood pressure, or elevated blood potassium; (3) presence of relative contraindications of MN, such as a history of bleeding disorders or using anticoagulants medications, active infection at the targeted area, and keloidal tendency; (4) pregnancy or lactation; and (5) use of topical or systemic medications that influence hair growth within the past 6 months. We calculated the sample size required for this trial by the PASS 15.0 software (NCSS, LLC, Kaysville, Utah, United States). Referring to a previous study for male androgenetic alopecia with 5% MX solution two times a day (11), the hair density change from week 24 to baseline was 18.8 ± 9.6 /cm² in the MX group and 38.3 ± 11.1/cm² in MX + MN group, respectively. In addition, we made a hypothesis that the hair density change from baseline to week 24 in MX + SPT group was the average of the MX group and MX + MN group because these data in the real world were not available. We set that α = 0.0167, β = 0.10. The sample size needed for the comparison between the MX group and the MX + MN group is 10 for each group. The sample size needed for the comparison between the MX group and the MX + SPT group is 30 for each group. The sample size needed for the comparison between MX + SPT group and the MX + MN group is 35 for each group. Above all, at least 35 samples for each group were required for statistical analyses. Considering the perhaps 10% loss rate of follow-up, the sample size for each group is 39 (35/0.9).

The participants were randomly assigned to three regimens by using a computer-generated randomization table. Since the therapy differed among the three groups, the study was not masked.

All patients were randomly assigned to the following three groups: (1) MX group: The patients received 1 ml of topical 5% MX tincture (Man Di, ZheJiang WanSheng Pharmaceutical Co., Ltd. China) once daily for 24 weeks; (2) MX + SPT group: The patients received oral SPT of 80–100 mg/day and 1 ml of topical 5% MX once daily for 24 weeks; (3) MX + MN group: The patients received MN treatments with the delivery of 5% MX every 2 weeks and 1 ml of topical 5% MX once daily for 24 weeks. All participants used the same household device with the scale (Scalp Solutions Applicator, Bella Medical Technology, China) to ensure the exact volume applied. Patients in MX + MN group received microneedling treatments according to the following protocol: (1) Patients washed their hair before treatment; (2) alcohol cotton balls were used to clean the treatment sites three times; (3) 1 ml of 5% MX was topically applied; (4) an electrodynamic microneedling was inserted at a depth of 0.7–1.0 mm; (5) the treatment endpoints were punctate hemorrhage and redness at the treatment site; (6) patients were told to not wash their hair within 8 h and not to use MX within 24 h. During the entire study period, no other treatment expected to be effective for FPHL was allowed. The enrolled patients were required to maintain the same hair length and color before any treatment and follow-up. All patients were allowed to drop out of the trial if a severe or life-threatening side effect occurred, or if the treatment did not respond as they expected.

Study visits were scheduled for week 12 and week 24. At baseline, patients' information, including age, body mass index (BMI), age of disease onset, disease duration, and family history, were collected. Moreover, dermoscope evaluation, ultrasound biomicroscopy (UBM), physician's global assessment, and patient assessment were also collected at baseline and each visit.

The most severe zone of alopecia (generally is the mid-point of the connection of the left and right external auditory ear canals) from each patient was marked with a red dot tattoo at baseline to ensure consistency. The tattoo was at the center of the target circle zone, and the diameter of this circle was 1.5 cm. Hair was clipped to a length of 1.6 mm in this zone at baseline and follow-ups. The Dermoscope system (Dermoscopy Image Diagnostic Workstation, Dermat Speedy Recovery T&D Co., Ltd. Beijing, China) was used to take pictures of the target zone under 20–folds magnification. Three board-certified dermatologists blinded to the treatment group evaluated the total hair density and hair shaft diameter according to the dermoscopy images. Hair density was calculated as the number of hairs per cm². The area of the photograph was automatically acquired in the dermoscope system. All hairs in the photograph of the target zone were marked manually and then counted by the cell count function in the ImageJ software. For each hair shaft in the photograph, we clicked and chose the diameter of it using the measuring tool of the dermoscope system, and then, the length was automatically calculated. The average diameter of all hairs in every photograph was recorded. UBM in 50 MHz (MD-310 Skin Ultrasound Biomicroscope, Meda Co., Ltd. Tianjin, China) was used to assess the scalp tissue structure. The examination was done using a specially designed cup (26 mm diameter) that forms a water bath environment of 3 cm height. The epidermal thickness, dermis thickness, and average follicle diameter were measured using the segment tool in the UBM screen. Five hair follicles at the same level were chosen to calculate the average hair follicle diameter.

At baseline and at each visit, we took global photographs by a digital camera (M100, Canon, Japan). The posture and position of patients, the angle and position of the camera, and lighting conditions were well–controlled. Three board-certified dermatologists who were blinded to the treatment group completed the photograph evaluation based on 7-point rating scale (−3 = marked worsening, −2 = moderate worsening, −1 = mild worsening, 0 = no change, 1 = mild improvement, 2 = moderate improvement, 3 = marked improvement). Sinclair's stage change was also assessed. Patient assessments consist of the Women's Androgenetic Alopecia Quality of Life Questionnaire (WAA-QoL) (12) and the Sinclair hair shedding score (6-point scale) (13).

Safety assessment included the recording of all adverse events and laboratory testing of serum potassium levels.

Continuous variables were recorded as means ± SDs or median (interquartile range, IQR) depending on whether they were a normal distribution. Ordinal variables were recorded as median (IQR). Categorical variables were recorded as a frequency and percentage. The comparison of continuous variables in normal distribution between baseline and follow-ups or among three groups was analyzed with paired t-test and one-way ANOVA, respectively. The comparison of continuous variables in abnormal distribution and ordinal variables between baseline and follow-ups or among three groups was analyzed with the Wilcoxon rank-sum test and Kruskal-Wallis test, respectively. Quantitative data were analyzed with the chi-squared test. P < 0.05 was considered to be statistically significant. All P-values were 2-sided. Particularly, pairwise comparisons among three groups were post hoc tests, and thus, P < 0.0167 was considered to be statistically significant according to Bonferroni correction. The SPSS 24.0 software (SPSS Inc., Chicago, IL, United States) was used for statistical analysis.

The flowchart of the study is shown in Figure 1. A total of 120 patients were enrolled in the study with 40 patients in each group. Five patients dropped out of the study. There were no significant differences in demographic and clinical characteristics among the three groups at baseline. Details are shown in Table 1.

Table 2 shows the changes in dermoscopic assessment. There were no significant differences in hair density and hair shaft diameter among the three groups at baseline (P = 0.824 for hair density; P = 0.597 for hair shaft diameter).

The hair density significantly increased only in MX + SPT group (P < 0.001) at week 12, while it significantly increased in all three groups at week 24 compared with the baseline (9.95 ± 10.16/cm² in the MX group, 16.76 ± 11.75/cm² in MX + SPT group, and 30.33±15.18/cm² in MX + MN group; P < 0.001 for each group from the baseline). As shown, MX + MN group increased in hair density more than both the MX group and MX + SPT group at week 24 (P < 0.001). Then, MX + SPT group increased in hair density more than MX alone group (P = 0.009).

The hair shaft diameter significantly increased in all groups both at week 12 and week 24 (P < 0.001). However, there were no significant differences among the three groups at week 12 and week 24 (P = 0.249 and 0.905, respectively). The representative dermoscopic images before and after treatments are shown in Figure 2.

Table 3 shows the changes in UBM assessment. There were no significant differences in epidermal thickness, dermis thickness, and average follicle diameter among the three groups at baseline (P = 0.651, 0.905, and 0.106, respectively).

The epidermal thickness significantly increased only in MX + MN group both at week 12 and week 24 (P < 0.001). However, there were no significant changes in the epidermal thickness in the MX group and MX + SPT group during the follow-up (P =0.537 and 0.088 for the MX group at week 12 and week 24; P = 0.343 and 0.562 for MX + SPT group at week 12 and week 24).

The dermis thickness of the three groups all significantly increased both at week 12 and week 24 (P < 0.001). There were no significant differences among the three groups at week 12 and week 24 (P = 0.581 and 0.524, respectively).

The average follicle diameter of the MX + MN group significantly increased both at week 12 and week 24 (P < 0.001). The average follicle diameter had no significant changes in the MX group and MX + SPT group during the follow-up (P =0.238 and 0.436 for the MX group at week 12 and week 24; P = 0.163 and 0.056 for MX + SPT group at week 12 and week 24). The representative UBM images before and after treatments are shown in Figure 3.

Assessed by investigators according to global photographs, 55.27% of patients in the MX group, 86.49% of patients in the MX + SPT group, and 95.00% of patients in the MX + MN group showed improvement at week 24 (Figure 4). Among the three groups, only 5.26% of patients in the MX group got worse. Clinical photographs are shown in Figure 5.

The Sinclair stage of patients among the three groups was improved during the follow-up. The change in MX + MN group was most significant. At week 24, the highest rate of Sinclair I and the minimum rate of Sinclair III were observed in this group. On the contrary, the MX group had the minimum rate of Sinclair I and the highest rate of Sinclair III. Details are shown in Figure 6.

At week 12, the hair shedding scores decreased in MX + SPT group and MX + MN group (P < 0.001), but not in MX group (P = 0.114). At week 24, the hair shedding scores decreased in all groups (P < 0.001). The WAA-QoL scores decreased in all groups at week 12 (P = 0.001, < 0.001 and < 0.001, respectively). At week 24, the WAA-QoL scores also decreased in all groups (P < 0.001). Details are shown in Table 4.

The adverse effects were recorded in all 115 patients. The main adverse effects are listed in Table 5. Scalp pruritus was most common among the three groups and ranked first in MX + MN group (n = 9). The most adverse effects were reported in MX + SPT group (n = 45). Several adverse effects, including menstrual disorder, hyperkalemia, and edema of the limbs, only occurred in this group. The menstrual disorder was the most common adverse effect of the MX + SPT group (n = 15). Infection (n = 1) only occurred in MX + MN group.