Weight gain in menopause: systematic review of adverse events in women treated with black cohosh

Abstract Weight gain is a frequent problem in perimenopausal and postmenopausal women. Cimicifuga racemosa (CR) is a popular treatment option for menopausal symptoms. The aim of this review was to investigate whether there is scientific evidence that CR causes weight gain. We searched our database for medically confirmed, spontaneous adverse events regarding weight gain, literature for case reports and randomized controlled trials. Thirty cases in total were spontaneously reported in 15 years. The causality was not considered certain/likely in any of the cases. A nurse (consumer) assessed the causality as possible. Only one case was published in the literature. However, no change in body fat composition was reported, and the causality seems unlikely. Of the 31 identified studies, 17 were double-blind placebo-controlled, five were double-blind reference-controlled and nine were open reference-controlled. In total, 1839 women were treated with CR for up to 12 months. Two studies reported weight gain as an adverse event; however, no significant differences in weight changes were observed between the groups. One case of weight gain (about 2 kg) was reported, but the authors did not specify in which treatment group. In conclusion, this study provides no scientific evidence that the use of Cimicifuga racemosa causes weight gain in menopausal women.


Rationale
Weight gain is very common among women in perimenopause and postmenopause. Nearly two-thirds of women aged 40-59years are overweight with a body mass index (BMI) of >25 [1]. On average over 5 years, up to 2.1 ± 5.1 kg were gained in this age group [2]. Other studies suggested an increase of 4.0 ± 4.6 kg [3]. The evidence regarding menopause-related changes in body composition (increased fat mass, decreased lean mass) is mixed; some studies support an association and others do not [4]. Cross-sectional comparisons of premenopausal and postmenopausal women and other studies of women throughout the menopause transition have yielded evidence both for and against menopause as the mediator of changes in body composition [5]. The majority of evidence suggests that changes in weight are due to chronological aging whereas changes in body composition and fat distribution are primarily due to the cessation of ovarian function. Weight gain and increased BMI are related to anxiety, depression and low life satisfaction during menopausal transition [6,7]. Furthermore, common disorders in menopausal women can lead to weight gain: sleep disturbances induce adipogenesis, increase appetite and calory consumption and result in overweight and obesity [8]. However, menopause per se does not result in significant weight gain after adjustment for age [9]. Weight gain per se also does not appear to be affected by the hormonal changes after menopause [10].
Hormone therapy is generally accepted as the treatment of choice for the amelioration of menopausal symptoms. However, the Women's Health Initiative (WHI) Study has shown that numerous health concerns are associated with even short-term use of hormone therapy, including stroke, invasive breast cancer and venous thromboembolic disease [11][12][13][14][15]. Although subsequent analysis showed a reduction in cardiovascular diseases in women under the age of 60 years, a growing number of women worldwide use herbal therapies [16]. Cimicifuga racemosa (CR; black cohosh) with or without Hypericum perforatum (St. John's wort) is a successfully used, non-hormonal herbal alternative for the relief of menopausal symptoms with a positive benefit-risk profile [17][18][19].

Objectives
The aim of the current systematic review was to evaluate the scientific evidence regarding any association between the intake of CR and weight gain from the most relevant sources of adverse events reporting.

Eligibility criteria and information sources
For the purpose of this review, we investigated the evidence of weight gain from the most relevant information sources used in the management of adverse events in pharmacovigilance: spontaneous reporting of adverse drug reactions (ADR); case reports in scientific literature; and clinical studies with a high Level of Evidence (LoE) according to the Oxford Center for Evidence-Based Medicine [20] (i.e. randomized controlled trials [RCTs]).

Data collection from spontaneous reporting
Cases of suspected adverse reactions from spontaneous reporting from worldwide reporting sources were included. However, only valid cases concerning weight gain under treatment with CR that were medically confirmed by a doctor or other health-care professional were included. Medical inquiries or reports from consumers as well as cases described on social media platforms were excluded.

Search in scientific literature
PubMed as well as our internal databases for pharmacovigilance and medical affairs were searched for any case reports from the literature of suspected adverse reactions concerning weight gain under treatment with CR. 'CR', 'isopropanolic extract of Cimicifuga (iCR)', 'Remifemin', 'black cohosh', 'Hypericum perforatum', 'St. John's wort' and/or 'weight gain', 'weight increased', 'increase of body weight' and 'BMI increase' were used as key words for the search strategy. No limitations for languages were set. Only valid cases with the four minimal criteria (identifiable source, identifiable patient, identifiable drug and identifiable adverse event) were reviewed.

Selection of clinical studies
Only RCTs using a preparation containing a CR extract to treat menopausal complaints were included in the analysis. CR as a mono preparation as well as combinations with other active ingredients were considered. There were no restrictions for therapy duration, CR dosage or extraction solvent. Clinical studies with a low LoE that did not include a control group were excluded. Due to treatment complexity and possible influence on body weight, studies in cancer patients were not considered.

Synthesis of results
Descriptive statistical methods were used for statistical analysis.
Cases from the three sources (RCTs and case reports from spontaneous reporting and the literature) were evaluated for the suspected drug (CR as a mono preparation or in combination with other active ingredients, such as H. perforatum, and regardless of the extraction solvent). The dose, duration of treatment, source, frequency, seriousness and outcome of ADRs were also considered. The causality was assessed in accordance with the World Health Organization (WHO) criteria [21]. In addition, the case documentation was evaluated regarding the quality of the available data.

Spontaneous reporting
In the period covered by our database (2005-2020) about 17.4 million patients were treated with iCR as mono preparation worldwide. During the same time frame, 30 case reports of weight gain were found in the spontaneous reporting system (Table 1). Eighteen women were treated with a CR mono preparation and 12 with a combination of CR and H. perforatum. Twenty-four cases were reported by pharmacists and only two by physicians. Thirteen patients   were between 40 and 59 years old, and in 16 cases the age was unknown when the adverse event occurred. Treatment duration was between 1 and 24 weeks; in 19 cases, the duration was not reported. For the majority of the cases (n ¼ 27), the daily dose was unknown. In six cases the outcome was reported as resolved. The outcome was not reported or was unknown for the majority of cases (n ¼ 23). In 11 cases the drug was withdrawn, and in 17 cases the consequences of the adverse events on drug intake (e.g. dose decrease) were unknown. For 21 cases, no causality assessment was performed by the reporter. In one case the pharmacist regarded the causality as suspected, and in two further cases the causality was assessed as possible (in one case, the patient was a nurse who consumed the preparation herself).

Case reports in scientific literature
Only one case report was found in the literature [22]. A 44year-old woman presented with a 1-week history of remarkable weight gain of 6.5 kg, headaches and abdominal discomfort. Four weeks before presentation, a gynecologist prescribed a dry extract of CR at a daily dose of 5 mg. The patient took two doses of 1 g acetaminophen because of headaches and reported being a vegetarian and consuming a minimal amount of alcohol.
Because of the difference in leg circumference between the two limbs due to lower leg edema, weight gain and the obvious coagulation activation, further diagnostic procedures were performed. There was remarkable fluid accumulation in the subcutaneous tissue. CR was discontinued. Complete resolution of the symptoms occurred within 1 week. No further case reports about similar reactions during treatment with CR were found in the literature.

Characteristics of the selected studies
Overall, 31 studies with 3827 patients were identified; 1839 patients were in the CR group and 1988 in the control group. In 17 RCTs, patients were randomized to a preparation containing: CR alone (n ¼ 2223), CR in combination with another herbal preparation (n ¼ 1081) or placebo (n ¼ 1142). In the other five double-blind RCTs, 334 patients were treated with CR and 328 with the reference drug. Furthermore, nine reference-controlled studies with 942 patients were found. In these studies, 424 were allocated to CR and 518 to therapy with other active drugs. An overview of the selected studies is presented in Table 2.

Adverse event reporting in clinical studies
Stoll [23] reported two patients in the placebo group who terminated the study because of strong headaches and weight gain. One woman in the estrogen group and three in the CR group experienced minor weight gain. However, therapy was not terminated and the causality with CR was assessed as unlikely. Bai et al. [24] mentioned one case of weight increase in the tibolone group. The causality was assessed as possibly drug related. No cases of weight increase occurred in the CR group.
Another study [25] noted that weight gain was common in both groups; however, no significant group differences were seen. A causality with the study medication was not reported. Osmers et al. [26] found no case of the ADR 'weight increase' in either trial groups, but they reported an increase in the mean/median body weight, which was higher in the placebo group than in the CR group (Table 2). Chung et al. [27] reported a weight gain of 2 kg in one case but did not specify the treatment group (CR or placebo). In all other studies, no adverse events referring to weight gain occurred or were notified.

Synthesis of results
An overview of the causality assessment for all cases of weight gain from all sources (spontaneous reporting, scientific literature and clinical studies) is presented in Table 3.

Discussion
Weight gain is common among women in perimenopause and postmenopause. Nearly two-thirds of women aged 40-59 years are overweight (BMI >25) [1]. Recently, during the COVID-19 pandemic lockdown, a greater percentage of women gained weight compared to men [28]. Among other factors, weight gain increases the risks for many diseases. A recent study found a positive association between BMI and (estrogen receptor-positive) breast cancer risk [15]. One in 50 women aged 50 years and of average weight using a 5-year continuous combined estrogen plus progestin therapy will be diagnosed with breast cancer within 20 years. During years 5-14 of menopausal hormone therapy use, the risk ratios hardly varied for any personal characteristics except BMI. Risk increased with increasing BMI among never users, but not among current users. On the other hand, many disorders (e.g. sleep disturbances, depressive mood) that are common among perimenopausal and postmenopausal women can lead to weight gain [8,29]. Therefore, in most individual cases it remains unclear whether perimenopausal and postmenopausal weight gain is the result of aging or underlying disorders, or possibly medication-induced.
To assess the latter, we evaluated the scientific evidence regarding any association between the intake of CR and weight gain and did not find any causal relationship. A few cases of weight gain were reported spontaneously, but none of them provided sufficient evidence that CR had had an impact on body weight. For the majority of the cases reported by physicians or pharmacists, no scientific assessment of causality was provided by the reporter. In few cases, the outcome was reported as resolved/resolving, and in the majority of reports the outcome was not reported or known. Using the WHO criteria for causality assessment, no evidence was found that CR induces weight gain. In the same period (2005-2020), about 17.4 million patients were treated with the iCR as a mono preparation worldwide. The literature search revealed only one case report about coagulation activation with subsequent weight gain [22], but the causality was regarded as unlikely, and no further cases regarding coagulation activation under CR therapy were found in the literature. RCTs with black cohosh showed no influence on coagulation parameters [30]. Numerous RCTs were conducted with CR (Table 2), but no evidence regarding weight gain was reported.
Our findings concur with the general development of body weight in women over 50 years of age. About 60% of women in menopause are overweight with a BMI >25 [1]. Weight gain was not related to change in menopausal status or to any lifestyle factors measured [2]. Menopause per se does not result in significant weight gain [9]. A recent review concluded that changes in fat mass quantity between premenopausal and postmenopausal women were mainly attributed to age; menopause had no significant additional influence [31]. However, an increase in measures of central fat and a decrease in total leg fat percentage indicated a possible change in fat mass distribution after menopause. The symptoms described in the case report from the literature [22] occurred 3 weeks after starting therapy with CR and resolved 1 week after withdrawal of it. Furthermore, the pathological laboratory values normalized 2-4 weeks after discontinuing treatment. At first glance, a relationship between the intake of CR and the reported side-effects seems to be possible. This assessment, however, has many limitations. The patient had been on a journey abroad shortly before the occurrence of the side-effects. This information was provided in the initial report, but not disclosed in the published case report. Thrombosis after a long flight or driving is a well-known event. Additionally, the patient had been taking pyridostigmine for many years and had been treated with acetaminophen when clinical symptoms occurred. She also sporadically used acetaminophen in a dose of up to 2 g per day; this may explain the slight elevation of transaminases. In summary, a causal relationship between the occurrence of the symptoms and the administration of CR seems to be unlikely. The female patient might have suffered from a thromboembolic event, and the observed weight gain was more likely a symptom of the thromboembolic event and not a change in body composition.
As regards clinical studies, 22 RCTs with a high LoE were included in our analysis. Seventeen studies were with placebo and five studies with other reference drugs. In total, 2885 patients were included. Thereof, 1415 patients were treated with an isopropanolic, an ethanolic or an unspecified CR extract; 1470 patients were allocated to control groups. Only a few single cases of weight gain were reported; however, no significant differences were found between the treatment groups. The causality was regarded as unlikely.

Experimental studies
The results of our analyses of RCTs agree with findings from experimental studies. Sun et al. [32] investigated the longterm effect of iCR on glucose and lipid metabolism in a rat model of menopause. Adult female Sprague-Dawley rats were sham operated (SHAM), ovariectomized (OVX), OVX and treated with estradiol valerate (OVX þ E) or OVX and treated with iCR (OVX þ iCR). Body weight, body composition and blood glucose levels of the animals were monitored. After 3 months of treatment, OVX þ iCR and OVX þ E rats exhibited a significant decrease in body weight gain, body and abdominal fat mass, serum triglyceride levels, hepatic fat accumulation and adipocyte hypertrophy compared with OVX rats.
In a further study [33], CR and some of its constituents were investigated in vitro for their effects on AMP-activated protein kinase (AMPK) compared to metformin in HepaRG cells. CR and its constituents activated AMPK to the same extent as metformin. In mice, CR significantly decreased the average daily and cumulative weight gain [33]. The results of this study demonstrate instead that CR may reduce body weight.
A recent study showed that an active ingredient of CR (23-epi-26-deoxyactein) effectively ameliorated obesity in high-fat, diet-fed mice. It is suggested that this compound may have substantial anti-obesity properties and may also improve menopausal and genetic obesity as well as associated metabolic syndrome [34].
Racho n et al. [35] evaluated the effects of CR consumption on body weight gain, intra-abdominal fat accumulation, plasma leptin, lipids and glucose tolerance and compared them with the effects of 17b-estradiol in vivo. Female Sprague-Dawley rats were ovariectomized and fed soy-free chow with the addition of estradiol-3 benzoate (E2B) or CR. In OVX rats, CR consumption not only decreased enhanced pituitary LH secretion but also attenuated body weight gain and intra-abdominal fat accumulation.
From the clinical perspective, a review of the potential effects of traditional Chinese medicine reported that CR may be beneficial for weight loss [36].
Our results are also in agreement with findings from a retrospective cohort study. In this observational study, women over the age of 40 years were treated with either menopausal hormone therapy (n ¼ 142) or CR (n ¼ 32). In both groups, metabolic serum parameters (lipids, glucose, insulin) and body weight did not change over the follow-up period of 12 months [37].

Limitations
This systematic review has limitations related to the articles from which data were extracted. We included data from different sources and quality of data. Particularly, cases from spontaneous reporting with low quality (LoE 5) can be seen critically. However, such limitations are a general problem and are well known for spontaneous adverse event reporting in pharmacovigilance.
The variety of designs and methodologies used in the studies is another limitation. While it may have been tempting to include only RCTs in the analysis, such an approach would have been critically incomplete from the perspective of good pharmacovigilance practice. Pharmacovigilance is the science and activities relating to the detection, assessment, understanding and prevention of adverse effects, particularly for the signal detection of rare adverse reactions.
Furthermore, details about safety data from clinical study reports were only available for those studies that were conducted with iCR extract. Other information was limited to the available data that had been published.
Another limitation was the variety of doses and extraction procedures used for the products in the studies included in this analysis. For these studies, a second analysis that addresses the doses and extracts most regularly used in clinical practice is needed to minimize the possible effect of the dosage and extract differences. Due to the limited information available in this scientific area, we forwent this analysis. Based on our data, however, we can state that the different doses/extraction procedures for the products analyzed in the present systematic review do not appear to influence the findings.