Randomized, Double-blind, Placebo-controlled Clinical Trial for Evaluating the Efficacy of Fractional CO2 Laser Compared With Topical Estriol in the Treatment of Vaginal Atrophy in Postmenopausal Women

Vera L. Cruz, MD; Marcelo L. Steiner, MD, PhD; Luciano M. Pompei, MD, PhD; Rodolfo Strufaldi, MD, PhD; Fernando L. Afonso Fonseca, PhD; Lucila H. Simardi Santiago, MD, PhD; Tali Wajsfeld, MD; Cesar E. Fernandes, MD, PhD


Menopause. 2018;25(1):21-28. 

In This Article


A total of 45 participants from a total of 50 preselected women were randomized into three treatment groups: 2 participants from the L group and 1 from the E group were lost to follow-up (Figure 1). As a result, primary outcomes were analyzed per protocol and by ITT analysis. No adverse effects of the fractional CO2 laser treatment or pain during laser application were observed during the study.

Figure 1.

Study flowchart.

No significant differences were found at baseline among groups. Participants characteristics, VHI, and FSFI were similar with the exception of burning, significantly milder in the E group (Table 1).

ITT analysis showed no differences in VHI average score among groups at baseline (P = 0.8) and week 8 (P = 0.5), although a significant difference among groups at week 20 was observed (P < 0.01). Per protocol analysis showed that VHI average score was significantly higher at week 8 (P < 0.05) and week 20 (P < 0.01) in comparison to baseline in all study arms. LE group also showed incremental improvement from week 8 to week 20 in the VHI score (P = 0.01) and the L group had a lower VHI score at week 20 compared with other study arms (P < 0.05) (Figure 2). Mean difference between groups L and E, L and LE, and E and LE at week 20 were -2.87 (95% CI: -5.99 to 0.26), 4.73 (95% CI: 2.42–7.07), and 1.87 (95% CI: -0.59 to 4.31), respectively.

Figure 2.

VHI score of different treatment arms at multiple time-points. a Wilcoxon test P<0.05, all groups; aa Wilcoxon test P<0.001, all groups; aaa Wilcoxon test, LE week 8 vs week 20, P = 0.01; b Kruskal-Wallis test, L vs E and LE, P<0.05; Friedman test for multiple timepoints, P<0.001, all groups.

L and LE groups showed a significant improvement of dyspareunia, burning, and dryness, and the E arm presented improvement of reported dryness (P < 0.001) in ITT analysis (Table 2). VAS symptoms were comparatively milder in the E group at baseline, and burning was shown to be significantly lower compared with the same symptom in the other treatment arms. Therefore, burning assessment at week 20 in the E group was directly compromised by baseline findings (P = 0.014; R 2 = 0.119). As ITT, per-protocol analysis, analysis also showed a significant difference at baseline among groups for burning (P = 0.02) and no significant improvement of burning symptoms in the E group (P = 0.5). Mean difference between groups L and E, L and LE, and E and LE at week 20 were 0.05 (95% CI: -1.27 to 1.36), 0.35 (95% CI: -0.83 to 1.53), and 0.40 (95% CI: -0.98 to 1.78) for dyspareunia; 0.13 (95% CI: -1.49 to 1.75), -0.93 (95% CI: -1.96 to 0.09), and -0.80 (95% CI: -1.49 to 1.75) for dryness; and 0.13 (95% CI: -0.73 to 1.0), -0.07 (95% CI: -0.96 to 0.82), and 0.07 (95% CI: -0.71 to 0.84) for burning, respectively.

Study arm sizes were considered adequate for detecting differences in VHI scores, and the LE, L, and E groups presented large effect sizes of 0.85, 0.72, and 0.81, respectively. Effect size estimates were also calculated for VVA symptoms, and the LE, L, and E groups presented effect sizes of 0.63, 0.45, and 0.33 for burning, 0.87, 0.82, and 0.75 for dryness, and 0.68, 0.60, and 0.52 for dyspareunia, respectively.

Table 3 summarizes FSFI full-scale scores and individual domain analysis. The LE group presented significant improvement of the total FSFI score (P = 0.02) and in individual domains of pain (P = 0.02), desire (P = 0.005), and lubrication (P = 0.02) compared with baseline. A significant difference in pain at baseline among all arms was found (P = 0.04), particularly between the LE and E groups (P < 0.05), but these findings did not persist at subsequent assessments where an overall symptomatic improvement was observed. A difference among groups (P = 0.006) in pain domain score was also observed at week 20.

No difference was observed in vaginal smear samples among groups (Table 4). The L and E groups had an improvement of MV of Meisels at week 8, but only the L group presented a significant MV increase also at week 20 (P = 0.01). In the LE group a nonsignificant increase of MV value was found, although MV increased from 48.4 ± 25.3 to 60.4 ± 8.6 at week 20. A decrease in the percentage of P cells per higher power field occurred in weeks 8 and 20 compared with baseline in the L and E groups, but only the L group also presented an incremental decrease of P cells for both week 8 (P = 0.03) and week 20 (P = 0.02). Cytological evaluation of vaginal smears was compromised due to inadequate sampling and preparation, resulting in different number of viable samples for analysis between and within groups.