Effect of CIMicifuga racemosa on metaBOLIC parameters in women with menopausal symptoms: a retrospective observational study (CIMBOLIC)

CIMBOLIC is an investigator-initiated monocentric retrospective observational study. Coded health-related data of all patients above age 40 presenting for first-time consultation at the Menopause Centre of the Department of Obstetrics and Gynecology, Inselspital Bern, between 01.04.2009 and 31.12.2016 were abstracted from medical records. Routine assessment comprised personal and family history including reproductive age [15], medication, lifestyle factors (smoking, alcohol), questionnaire Menopause Rating Scale (MRS)-II [16], anthropological parameters (body weight, height, body mass index, blood pressure, and waist circumference), laboratory blood tests, and risk evaluation of certain non-communicable diseases (cardiovascular disease, breast cancer, osteoporosis), respectively (supplementary file 1). The selection of the treatment option (CR or MHT) was based on the above-described routine assessment by the physician, a benefit–risk evaluation, and the patients wish at the time of consultation. For the final analysis, all patients treated with either Ze 450 (Cimifemin^® forte = 13 mg dry extract, Cimifemin^® uno = 6.5 mg dry extract or Climavita^® forte = 13 mg dry extract) or any MHT (but not both) during the follow-up period and having at least one follow-up visit including blood tests were eligible. The study was approved by the cantonal ethics committee (Kantonale Ethik-kommission Bern (KEK) Switzerland No 2016-01179).

Data from medical records were transferred to a REDCap (Research Electronic Data Capture) database. Generation, transmission, storage, and analysis of health-related personal data within this project were according to the current Swiss legal requirements for data protection and were performed according to the Ordinance HRO Art. 5. This observational study is reported according to the STROBE statement (see www.​strobe-statement.​org).

Patients were grouped according to whether they were treated with Ze 450 or MHT during the follow-up period. Continuous and categorical variables are presented by group using median with lower and upper quartiles (lq, uq), and number and percentage of patients, respectively. Baseline characteristics between groups were compared using Wilcoxon rank-sum tests and Fisher’s exact tests for continuous and categorical variables, respectively. Linear mixed-effects regression models were used for the analysis of a change over time within and between groups. The endpoint values at baseline and each follow-up were used as dependent variable and the treatment group (i.e., MHT or Ze 450), the time, and the interaction of time and group as fixed explanatory variables. A random intercept was included for each patient to account for intra-patient correlations. The models were fitted with restricted maximum likelihood, the degrees of freedom were calculated using the Satterthwaite approximation and 95% confidence intervals (95% CI) based on the t distribution. Marginal effects for each group were calculated using Stata’s margin command and expressed as mean change per year with 95% CI and a p value. The difference between groups (i.e., the interaction term of group and time) was expressed as mean difference per year (Ze 450 minus MHT) with 95% CI and a p value. For the main analysis, all patients in the full analysis set were included, regardless whether a baseline and a follow-up assessment were available for a specific endpoint. For a sensitivity analysis, only patients with a baseline and at least one follow-up measurement were included (referred to as complete case analysis). We used a multi-variable linear mixed-effects regression to adjust for confounding. All baseline variables with a p value < 0.2 in the baseline comparison and less than 30% missingness in both groups were included as covariates. An overall control of the type I error rate was not done because of the explorative nature of the study. The reported p values were not corrected for multiplicity and have to be interpreted accordingly. All analyses were done with Stata [StataCorp. Stata Statistical Software: Release 14. StataCorp, College Station, TX: StataCorp LP., 2015.] or R [R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria, 2016. URL http://​www.​R-project.​org. ISBN 3-900051-07-0.].

Figure 1 displays the patient flow. Of 769 patients screened, 174 were eligible for the final analysis. Of those, 32 were treated with Ze 450 and 142 with MHT, respectively. Current systemic MHT regimens comprised mainly estrogen–progestogen combinations (n = 89; 63%), while the remainder received either estrogens only, progestogens only, estrogen–androgen combinations, or tibolone, respectively. Only few patients in the Ze 450 group (n = 3; 9%) but most in the MHT group (n = 103; 73%) had a history of MHT use. However, at first consultation few patients were using Ze 450 (n = 8; 25%) or MHT (n = 49; 35%), respectively. All Black cohosh user had a Ze 450 product. Of those, the majority had a daily dosage of Ze 450 at 13 mg dry extract (n = 25; 78%; recommended dose 1 tablet daily).

Table 1 presents subjects’ anthropological, reproductive, and lifestyle parameters at baseline. There was no significant intergroup difference but for reproductive stage (p = 0.038). The majority of women treated with Ze 450 were postmenopausal (n = 24; 83%), while only one in two women treated with MHT fell into this category (n = 66; 55%). About one-third of MHT-treated women were either premenopausal (n = 21; 17%) or in the early menopausal transition (n = 23; 19%). In both groups, mean follow-up was 12 months (Ze 450 lq, uq 12.0, 24.0; MHT lq, uq 12.0, 36.0; p = 0.91). There was a significant difference in the number of follow-up visits (p = 0.011). 88% (Ze 450) and 61% (MHT) of subjects had only one follow-up visit. A few women treated with Ze 450 had two (n = 2; 6%), three (n = 1; 3%), or four (n = 1; 3%) follow-up visits, compared to 41 (29%) with two, 12 (8%) with three, and 3 (2%) with four follow-up visits in the MHT group.

Table 2 presents the type and intensity of menopausal symptoms at baseline, assessed by the questionnaire MRS-II. In both groups, median intensity of overall menopausal symptoms (MRS-II total score) at baseline was moderate (defined as score 9–16) to severe (defined as score ≥ 17) (p = 0.46). When differentiating between menopausal symptom subdomains, both groups suffered from moderate vegetative symptoms (defined as score 5–8) (p = 0.10) and moderate (defined as score 2–3)-to-severe (defined as score ≥ 4) urogenital symptoms (p = 0.89), respectively. Psychological menopausal symptoms were moderate (defined as score 4–6) at baseline in both groups with women choosing MHT later on being slightly but significantly more affected than women choosing Ze 450 (p = 0.011).

We did not find any evidence for a within-group change in body weight, carbohydrate metabolism (glucose, insulin, and HOMA-IR), and serum lipids (total cholesterol, LDL-cholesterol, HDL-cholesterol, and triglycerides), neither in women treated with Ze 450 nor with MHT (Table 3). For MHT, we found a significant improvement over time of the MRS-II total score (− 0.99 [95% CI − 1.42, − 0.55] per year; p < 0.0001), MRS-II vegetative subscore (− 0.24 [95% CI − 0.45, − 0.03] per year; p = 0.023), MRS-II psychological subscore (− 0.48 [95% CI − 0.71, − 0.25]; p < 0.0001), and MRS-II urogenital subscore (− 0.28 [95% CI − 0.45, − 0.11]; p = 0.001). Similarly, treatment with Ze 450 significantly improved the MRS-II vegetative subscore (− 0.81 [95% CI − 1.57, − 0.04]; p = 0.039) and urogenital subscore (− 0.64 [95% CI − 1.26, − 0.01]; p = 0.045) (Table 3). MRS-II total score had the same trend (− 1.43 [95% CI − 3.16, 0.30]; p = 0.11). Between-group comparisons did not reveal significant differences for any endpoint (Table 3).

To adjust for potential confounding, we included baseline variables with a p value < 0.2 in the baseline comparisons and less than 30% missingness (diastolic blood pressure, age at menarche, reproductive stage, smoking status, MRS-II vegetative subscore, and MRS-II psychological subscore) in multi-variable models (Table 4). Results were similar as in the univariable models. For MHT, we found evidence for a change of the MRS-II total score (− 0.87 [95% CI − 1.26, − 0.48] per year; p < 0.0001), the psychological subscore (− 0.42 [95% CI − 0.60, − 0.24] per year; p < 0.0001), and the urogenital subscore (− 0.24 [95% CI − 0.42 to − 0.07] per year; p = 0.005). For Ze 450, we found some evidence for a change of the total score (− 1.82 [95% CI − 3.65, 0.02]; p = 0.05) and the urogenital subscore (− 0.79 [95% CI − 1.60, 0.02]; p = 0.05). We did not find any evidence for a between-group difference. Furthermore, we did not find within or between-group changes for body weight and serum metabolic parameter (total cholesterol, LDL-cholesterol, HDL-cholesterol, triglycerides, fasting glucose, insulin, and HOMA-IR) (Table 4).