This three-arm blind randomized clinical trial with a negative control group was conducted between August 2019 and August 2020. Participants were postmenopausal women who were referred to comprehensive health service centers in Qazvin Province to receive postmenopausal care. The trial protocol has been peer-reviewed and published [23].

To identify eligible participants, a member of the research team approached comprehensive health service centers in Qazvin, including Buin Zahra, Takestan, and Alborz counties. A total of 105 eligible postmenopausal women were recruited from 35 medical centers using convenience sampling. The inclusion criteria were as follows: (i) being menopausal for at least 1 year, (ii) being married, (iii) being sexually active, and (iv) having sexual desire. The exclusion criteria were: (i) use of intravaginal medications; (ii) moderate to severe utero/genital infection (e.g., vaginitis, cervicitis); (iii) receiving hormone therapy; (iv) having experienced a stressful event during the previous quarter (e.g., death of a loved one); (v) having undergone a hysterectomy or other gynecological surgery; (vi) presence of any other endocrine disease such as Cushing’s disease, diabetes, hyperprolactinemia, or hyperparathyroidism; (vii) having experienced premature ovarian failure; (viii) having any organ abnormalities, sarcoidosis, kidney disease, or histoplasmosis; and (ix) spousal disorder that prevents sexual activity (e.g., diabetes).

According to the results of a previous study [24], using a type I error of 0.05, test power of 80%, and average effect size of 0.6, the minimum number of participants in each group was 25. Considering a 40% attrition rate, 35 individuals from each group were invited to participate, resulting in a total sample size of 105 individuals.

The eligible participants were randomly assigned to three groups for 8 weeks of treatment: intervention (vaginal suppository containing 1,000 units of vitamin D3), placebo (vaginal suppository placebo), or control (no treatment). The allocation sequence was prepared using a random allocation software. Intervention type was then written inside sealed opaque envelopes based on the allocation sequence and coded on the envelope. Questionnaires were coded using the same codes as those used for the intervention. Therefore, participants who received Intervention completed a questionnaire with the same code.

The suppositories were prepared and packaged in coded forms by a pharmaceutical company. Therefore, the participants, presenters, evaluators, and statisticians did not know whether the suppository contained vitamin D. After the statistical analysis, suppository codes were obtained from the pharmaceutical company.

The primary outcome of the present study was a change in sexual functioning assessed using the Female Sexual Function Scale (FSFI) before, immediately after, 1 month after, and 2 months after the intervention. The FSFI, a 19-item self-reported questionnaire, was designed by Rosen et al. [25] in 2000 to assess female sexual functioning. This scale assesses six aspects of female sexual functioning: desire, arousal, lubrication, orgasm, satisfaction, and pain [25]. Psychometric evaluation of the Iranian version was assessed and confirmed to have good reliability and validity [26].

Another questionnaire was used to record the participants’ demographic characteristics, including age, education, menopausal age, spouse’s age, spouse’s educational level, body mass index, and number of pregnancies and deliveries. All participants were interviewed by a trained research team that was blinded to their study group allocations. The interviewer obtained informed consent from all participants. The questionnaires were completed once they agreed to participate. The interviewer helped those who were illiterate answer the questions by recording their answers.

Prior to the intervention, sexual performance was assessed using the FSFI. The trial was implemented based on the assigned interventions: vitamin D suppository (group A), placebo (group B), and control (group C).

The methods of preparation and storage of the vitamin D and placebo suppositories were previously outlined in detail [23]. The vitamin D and placebo suppositories consisted of a common Suppocire AM-15 base and had the same shape, color, and weight. Only one identification code (A or B) was placed on the packaging to match it to the envelope. During the first meeting, envelopes were provided to the participants in groups A and B according to the allocation sequence. They were instructed to insert a suppository 3 cm into their vagina before going to bed. Before inserting the suppository into the vagina, the participants were told to: (i) wash their hands with soap and water, (ii) insert the suppository into the vagina at a depth of two fingertips; and (iii) use only one suppository each time. The treatment schedule was every night in the first 2 weeks and every other day for the following 6 weeks. Follow-up visits were also scheduled. All procedures were conducted under the supervision of a gynocologist who was also a member of the research team.

The risk of vitamin D toxicity occurs with the long-term use of high doses (>10,000 [27,28] to 15,000 [29] units/daily). The total dose of vitamin D in the present study was 32,000 units over 8 weeks. Therefore, the risk of toxicity was considered low and within the safe range considering that it was much lower than the potentially toxic dose based on current evidence [27–30]. The participants were asked not to take any oral vitamin D supplements during the study period. In addition, all symptoms of vitamin D toxicity were explained to the participants for self-monitoring of adverse responses.

Sexual functioning was assessed immediately and 1 and 2 months after the completion of the 8-week intervention. Moreover, symptoms of potential vitamin D toxicity and side effects were evaluated at each visit.

This study was approved by the research committee of the University of Medical Sciences (approval number: IR.QUMS.REC.1397.117). The study protocol was prospectively registered with the Iranian Clinical Trial Registration System (IRCT20180704040346N1) on October 10, 2018. After obtaining the necessary permits, patients were invited to participate in the study. Written informed consent was obtained from all participants. All of the research processes were monitored and verified by an agency selected from the Research Committee of the University of Medical Sciences.

SPSS version 24 (IBM SPSS Statistics, Armonk, NY, USA) was used for the data analysis. The normal distributions of the variables (total sexual functioning scores and subscale scores) were examined using the Shapiro-Wilk test for the indicators of central tendency, dispersion, and histogram diagram. Intergroup comparisons were performed to evaluate the distribution of variables based on the proposed Imbens and Rubin [31] method considering the standardized mean difference (MD) criteria of <0.25 for continuous quantitative variables and <10% risk difference index for categorical variables. Owing to the normal distribution of the data, one-way repeated-measures analysis of variance (ANOVA) was used to compare the changes in mean FSFI scores between the study groups at different time points. In each case, the prerequisites for performing repeated-measures ANOVA (including the sphericity test and homogeneity of variance) were checked. Owing to the lack of sphericity default, the results were reported with the Greenhouse-Geisser correction. In cases in which the ANOVA test was significant for repeated measures, the Sidak post hoc test was used to determine the differences. Moreover, an analysis of covariance was performed to compare groups according to the corrected sexual functioning score to control for the mean baseline score. Bonferroni correction was used for multiple comparisons. The effect sizes of the MD and the standardized MD (SMD) were calculated based on Cohen’s d. A Cohen’s d effect size of 0.2–0.5 was considered a small effect size; 0.8–0.5 was considered a medium effect size; and >0.8 was considered a large effect size [32].

The MD was examined to assess the minimal clinically important differences (MCID). The MCID for outcomes can be defined via three methods based on evidence from previous studies and calculated using anchor- and distribution-based methods [33]. Two approaches were used to define the MCID for FSFI scores. First, based on a literature review, Krychman et al. [34] reported an MCID of 4.2 for the total FSFI score. As the MCID of patient-reported measures might differ among patients and clinical contexts [35], the second approach of calculating the MCID based on the distribution-based method was used. According to Jacobson et al. [36], the MCID was calculated using the following for $\text{MCID}=1.96\times S{D}_{base}\times \sqrt{2\times (1-\text{ICC})}$.

Considering an standard deviation_base of 4.82 for the intervention group and an intraclass correlation coefficient of 0.908 from the first and last measurements (for a more conservative approach), the calculated MCID of 4.05 was used to assess the clinical significance of the intervention and placebo. If the MD was higher than the calculated MCID at any follow-up point, it was considered clinically significant. The significance level for all tests was P<0.05.

As shown in Fig. 2 and Tables 2, 3, in the intervention group using the vitamin D suppository, significant improvement in sexual functioning was observed immediately after and 1 month after the intervention (P<0.001 vs. baseline). Examining the MD between these two time points (adjusted mean, 26.36 and 23.69 immediately after and 1 month after the intervention, respectively) with the baseline (mean, 19.16) value also confirming that a minimal clinically important difference was achieved (7.17% and 4.53%, respectively, compared to the MCID of 4.05). However, at 2 months after the intervention, no difference in sexual functioning scores was noted compared to baseline (P=0.65; adjusted mean, 20.54; MD, 1.38).

In the placebo group, only at the first follow-up (immediately after the intervention; adjusted mean, 23.10) was sexual functioning significantly better than at baseline (mean, 17.93; MD, 5.17; P=0.001; Fig. 2). At 1 and 2 months after the intervention, the sexual functioning score was not significantly or clinically different from that baseline (adjusted mean at 1 month after the intervention, 21.22; P=0.35; MD, 3.29; adjusted mean at 2 months after the intervention, 19.30; P=1.00; MD, 1.37; Fig. 2). In the control group, a decrease in sexual functioning score was observed over time, which is expected in such individuals. This decline was statistically significant at 2 months (adjusted mean, 16.16) compared with baseline (mean, 20.03). However, it had not yet reached an MCID of 4.05 (P<0.001; MD, −3.87) (Table 2).

The intergroup differences are presented as pairwise comparisons between the study groups at each time point in Tables 2, 3 and Fig. 2. The FSFI subscales are compared among groups in Table 4. Comparing the intervention and placebo groups based on the adjusted value using the baseline score of the FSFI and the spouse’s age immediately and 1 month after the intervention, the sexual functioning score in the intervention group was significantly higher than that in the placebo group based on Cohen’s d (immediately after: MD, 3.25; Cohen’s d=0.85; 1 month after: MD, 2.48; Cohen’s d=0.51). Over time, the effect size of the intervention decreased; therefore, the effect size of the intervention group was not significantly different from that of the placebo group at the third follow-up (2 months post-intervention).

The sexual functioning score was higher in the intervention versus control group at all three follow-up points (MD between the two groups at the three time points post-intervention: 7.19, 6.59, and 4.38, respectively, greater than the MCID of 4.05). The vitamin D intervention had a significant effect on sexual functioning versus control (Cohen’s d=1.87, 1.37, and 0.77, respectively, for the three follow-up points).

The sexual functioning score of the placebo group was significantly higher than that of the control group at all three follow-up points, but the difference was not clinically important (MD between the two groups at all three time points post-intervention was less than MCID of 4.05).

No side effects were reported during the intervention or follow-up periods.

The strength of the present study is its use of a randomized clinical trial design with three groups (intervention, placebo, and negative control). However, this study has some limitations. First, the participants were healthy postmenopausal women, which may limit the generalizability of our results to postmenopausal women with underlying diseases. A self-reporting method was used to assess the key outcome variables. Moreover, the spouse’s sexual functioning was not assessed, which may have affected the outcomes. Some contributing factors, including psychological and hormonal factors, were not assessed. In addition, only 2 months of follow-up was performed after the intervention; therefore, further studies are required to examine the longer-term effects. The study did not examine vaginal cytology or measure serum vitamin D levels due to funding constraints, which could also be collected in future studies. Because the standard treatment for vaginal atrophy in menopausal women is low-dose topical estrogen, future studies should compare vitamin D interventions with various standard therapies.

Clinical improvement in sexual functioning (based on clinically important differences) was observed only in the intervention group administered vitamin D (versus control). Although the scores increased in the placebo versus control group, they were not clinically important. The present study showed that vaginal vitamin D suppositories improved the sexual functioning of postmenopausal women in the short term with no side effects. These short-term effects may limit the use of this protocol. Future studies are required to assess the effects of a daily administration regimen that may yield long-lasting improvements in sexual functioning. Moreover, psychological and hormonal factors as well as serum levels of vitamin D should be assessed both during and after treatment.