Impact of estrogens on resting energy expenditure: A systematic review

The fear of weight gain is one of the main reasons for women not to initiate or to early discontinue hormonal contraception or menopausal hormone therapy. Resting energy expenditure is by far the largest component and the most important determinant of total energy expenditure. Given that low resting energy expenditure is a confirmed predictive factor for weight gain and consecutively for the development of obesity, research into the influence of sex steroids on resting energy expenditure is a particularly exciting area. The objective of this systematic review was to evaluate the effects of medication with natural and synthetic estrogens on resting energy expenditure in healthy normal weight and overweight women. Through complex systematic literature searches, a total of 10 studies were identified that investigated the effects of medication with estrogens on resting energy expenditure. Our results demonstrate that estrogen administration increases resting energy expenditure by up to +208 kcal per day in the context of contraception and by up to +222 kcal per day in the context of menopausal hormone therapy, suggesting a preventive effect of circulating estrogen levels and estrogen administration on weight gain and obesity development.

(HC) or menopausal hormone therapy (MHT). [7][8][9][10][11] According to current data, however, this concern seems to be unfounded. Neither a sex hormone-related significant weight-increasing nor weightreducing effect has been clearly confirmed or refuted. Cochrane analyses on the topic of the influence of combination contraceptives, 12 progestin-only contraceptives, 13 and MHT 14 on body weight showed no association between sex hormone use and weight changes. However, the duration of hormone administration in the majority of the studies included in these Cochrane analyses was probably too short to detect significant effects of sex hormone medication on body weight. Overall, the majority of studies lack clear evidence to definitively rule out the possibility of sex hormone-induced weight change.
Obesity is the result of a chronic disturbance in energy homeostasis caused by an imbalance between energy intake (EI) and total energy expenditure (TEE). The term "energy intake" is almost selfexplanatory and refers to the intake of energy through macronutrients (carbohydrates, proteins, fats) and alcohol, whereas the term "total energy expenditure" is a more complex one and consists of three main components: first, resting energy expenditure (REE); second, dietary energy expenditure (DEE, also called "thermic effect of food" or "dietinduced thermogenesis"); and third, physical activity energy expenditure (PAEE). [15][16][17] REE represents the largest component of TEE (60%-80% [18][19][20] and is defined as the metabolic rate required to maintain vital physiological functions from the cellular to the systemic level of an organism that is awake, in mental and physical rest, fasting, and in a thermoneutral environment. [15][16][17] DEE describes the energy required for the digestion, absorption, and storage of ingested nutrients. For individuals consuming an average mixed diet, DEE amounts to approximately 10% of TEE. 15,16,18,21 PAEE is the most variable component of TEE and can be significantly increased by intentional physical exercise. Given the typically sedentary lifestyle of Western societies, it averages 25%-35% 22 of TEE ranging from 15% in very inactive individuals to up to 50% in highly active individuals. 18 Both a reduction in EI and an increase in PAEE are necessary for sustainable weight loss but are difficult to implement in the long term.
For this reason, research has recently focused increasingly on the possibility of achieving a negative energy balance by increasing REE.

| Resting energy expenditure and basal proton leak
As mentioned above, REE is defined as the energy required by an organism to maintain homeostasis. [15][16][17] The majority of the energy required for REE is generated through the breakdown of organic substrates in the course of cellular respiration in the mitochondria. The goal of cellular respiration is to produce adenosine triphosphate (ATP), the universal energy source for cellular functions. The main part of ATP production takes place in the respiratory chain, the final step of cellular respiration. However, the coupling of oxidative processes to ATP production has limited efficiency. Part of the energy of the proton motive force is "lost" through a basal proton leak at the inner mitochondrial membrane and is thus uncoupled from ATP production and directly converted into heat. Hence, REE is the sum of the energy required for ATP production (about 70%-80% of REE) on the one hand, and the energy converted into heat through the basal proton leak (20%-30% of REE), on the other. [22][23][24][25][26][27][28] The latter enables the organism to maintain body temperature in a thermoneutral situation without the need to actively initiate other homeostatic thermoregulatory mechanisms. 29,30 The basal proton leak accounts for a significant proportion of this so-called obligatory thermogenesis. 27,31,32 Still, the exact mechanism behind this process remains controversial. 33 However, reinforcing the process of "wasting" energy through proton leaks represents a possibility to modulate the energy balance toward the negative and is therefore considered a potential and promising target for the treatment of obesity. Whereas the basal mitochondrial proton leak cannot be intentionally modulated, there is an inducible/adaptive/facultative proton leak exclusively in brown adipose tissue (BAT) mitochondria, which, in contrast to the basal proton leak, can directly be regulated due to its dependence on the activity of the tissue-specific so-called uncoupling protein 1 (UCP1). 29,30 1.3 | Brown adipose tissue and adaptive proton leak Mammals have two types of adipose tissue that play opposite roles in the body's energy metabolism: white lipid-storing adipose tissue (WAT) and brown lipid-burning adipose tissue. WAT cells contain a single large lipid vacuole and a few mitochondria, whereas BAT cells are characterized by multilocular small lipid droplets and a large number of well-developed mitochondria. WAT not only has the ability to store excess energy in the form of triglycerides and release this energy as needed in the form of free fatty acids but, in recent years, has also gained an expanded role as a metabolically and endocrinologically highly active organ. Through the release of various biochemically active substances, such as adipokines, it is involved in communication between cells, tissues, and organs. 34 In contrast to energy storage, the main function of BAT is to generate heat on demand through the oxidation of free fatty acids, thus consuming energy. This process is part of the so-called adaptive thermogenesis (synonyms: facultative or nonshivering thermogenesis). 29,30,35,36 It is the mass of mitochondria in the BAT cells that enables this form of energy release and gives this particular fat tissue its characteristic brown color. In contrast to the regular mitochondria, the BAT mitochondria are characterized by the ability to actively uncouple cellular respiration from the synthesis of ATP and consequently to convert the energy released during fatty acid oxidation directly into heat, circumventing ATP production. This biochemical short circuit is made possible by the unique expression of UCP1, an inducible ion channel at the inner mitochondrial membrane that allows the re-entry of protons, from the intermembrane space into the mitochondrial matrix, bypassing the membrane-bound ATP synthase and thus dissipating the energy of the proton gradient directly into heat, as described previously (Figure 1). 25,29,35 The concept of additional stimulation of this energetically wasteful tissue with the resulting increase in energy expenditure and its potential influence to counteract the development of obesity and related diseases has kept scientists busy for more than 40 years. 37 In fact, in animals (rodents), the enhancement of BAT thermogenesis to maintain healthy body weight is now well established. [38][39][40][41][42][43] In humans, however, it has long been assumed that only infants possess a physiologically relevant amount of BAT. 29,44 Following the recent rediscovery of metabolically active BAT also in adult humans, [45][46][47][48][49] its experimental use has once again attracted increased interest in clinical and non-clinical studies. 38,[50][51][52][53][54] BAT thermogenesis is controlled by the sympathetic nervous system (SNS). The ventromedial nucleus of the hypothalamus (also referred to as the ventromedial hypothalamus, VMH) is the central key center responsible for the regulation of BAT thermogenesis and thus is superior to the SNS. 29,[55][56][57][58] The physiological stimulus to activate the VMH-SNS-BAT axis is cold. 45,46,49,[59][60][61][62] Subsequent release of norepinephrine from the SNS leads to activation of BAT via binding to its β3 adrenergic receptors (β3 AR). This is followed by an increase in the activity of intracellular lipolytic enzymes and the breakdown of cytosolic triglycerides to free fatty acids, which are directed into the mitochondria, where they both allosterically disinhibit UCP1 present in the inner mitochondrial membrane and themselves serve as fuel for fatty acid oxidation. Simultaneously, glucose and fatty acid uptake from the periphery into BAT cells is increased to replenish intracellular lipid stores. 29,30,35,49 Hence, the BAT is the only tissue that can literally "burn fat." However, stimulation of the BAT by the SNS leads not only to an increase in its metabolic activity but also to an expansion of its volume as well as to the so-called "browning," which refers to the recruitment of brown adipocytes by differentiation from brown adipocyte precursors/preadipocytes in WAT or by transdifferentiation of existing white adipocytes. These stimulation-induced thermogenically competent cells in WAT are referred to as brite (brown-in-white) or beige adipocytes. As the name implies, they exhibit phenotypic and functional features common to white and brown adipocytes. In summary, BAT is a flexible tissue that can be recruited by specific triggers and atrophies in their absence. 35,38,53,56,58,63,64 There is growing evidence that BAT volume and activity in humans are associated with increased energy expenditure, 46,51,[65][66][67] decreased BMI [45][46][47][68][69][70][71] as wells as improved glucose, [72][73][74][75] and lipid metabolism. 73,75 Thus, people with a higher volume of metabolically active BAT also appear to be more likely to be protected from developing obesity and its related diseases.
Because more or less intensive, intermittent as well as continuous cold stimuli are not tolerable for humans in the long run to provide continuous activation of the BAT, other substances have been F I G U R E 1 Obligatory versus adaptive thermogenesis. The goal of cellular respiration is the production of adenosine triphosphate (ATP), the universal energy source for cellular functions. Only 70%-80% of the energy of the proton motive force is used for ATP production. 20%-30% of this energy is "lost" through a basal proton leak at the inner mitochondrial membrane and is converted directly to heat, thus uncoupling it from ATP production. These two processes together are referred to as obligatory thermogenesis. Although the basal mitochondrial proton leak cannot be intentionally modulated, there is also an adaptive proton leak that, unlike the basal proton leak, can be directly regulated because of its dependence on the activity of the so-called uncoupling protein 1 (UCP1) at the inner mitochondrial membrane. Thus, cellular respiration can be actively uncoupled from ATP synthesis when needed, and the energy released during cellular respiration can be directly converted to heat bypassing ATP production. This process is called adaptive thermogenesis.

| Estrogens and energy metabolism
The function of estrogens as important regulators of energy balance is well established. [86][87][88] Using the example of ovarian insufficiency, for example, at menopause, as a consequence of ovariectomy or ovarian function suppression, it has been sufficiently demonstrated that low estrogen levels are associated with increased EI 87  if physiological variations of estrogens affect REE, it is reasonable to assume that exogenous administration of these hormones also has an impact on REE. However, data on this are scarce.

| Estradiol and its central and peripheral effects on brown adipose tissue thermogenesis
Estrogen receptor α (ERα), expressed by both the central nervous system (CNS) and adipose tissue, is considered the key mediator of estrogen action on energy homeostasis. 86,87,114 With respect to energy expenditure, estrogens primarily exert both central and peripheral stimulatory effects on BAT thermogenesis. [84][85][86]115 Consistent with this is the observation of a sexual dimorphism in the BAT. 116 As far as is currently known, women exhibit greater BAT mass and are more responsive to BAT-activating stimuli than men. 47,[117][118][119][120] Moreover, these sex differences disappear after menopause. 121,122 In the CNS, estrogens exert their multiple functions in energy balance by binding to ERα, which is expressed at a variety of sites, predominantly in the hypothalamus. [123][124][125] The most important actions of estrogens on energy expenditure, in particular on BAT thermogenesis, take place nucleus-specifically in the VMH. Any electrical, pharmacological, and hormonal stimulation of this nucleus increases sympathetic outflow to the BAT via intermediate interconnections in brainstem areas such as the nucleus raphe pallidus and the inferior olive, thereby increasing BAT thermogenesis. 55,58,83,[126][127][128][129] Among the major molecular mechanisms mediating the action of estrogens in the hypothalamus is the inhibition of hypothalamic AMPactivated protein kinase (AMPK). 83 AMPK is an enzyme that plays a key role in regulating cellular processes in situations of cellular energy deficiency. Central activation of AMPK leads to an increase in energy production, that is, synthesis of ATP, and a decrease in energy wasting, ensuring the maintenance of cellular energy homeostasis. [130][131][132][133][134] Thus, centrally, estrogens lead to increased sympathetic tone via central inhibition of AMPK, with subsequent activation of BAT thermogenesis and energy "wasting." 83 Other substances known for their effects on metabolic homeostasis and whose common pathway is to increase BAT-specific energy expenditure via AMPK inhibition in the VMH include leptin, thyroid hormones, nicotine, bone morphogenetic protein 8b, and glucagon-like peptide 1 and its analogs. 78,126,[135][136][137] Finally, recent data have shown that not only BAT activation but also the browning of WAT is regulated via AMPK inhibition in the VMH. 115 Another key mechanism by which estrogens lead to an increase in sympathetic tone and thus activation of BAT and energy wasting is through modulation of the lipotoxic effects of reactive lipid species in the VMH. The deposition of lipophilic metabolites such as ceramides in the VMH associated with overeating, also known as hypothalamic lipotoxicity, leads to the impairment of various cellular functions such as the triggering of stress in the endoplasmic reticulum (ER). The result is reduced sympathetic tone, which leads to reduced BAT thermogenesis and food-independent weight gain. 138 It has been demonstrated that centrally administered estradiol (E2) can counteract the central dysregulation of the energy metabolism triggered by the accumulation of lipid metabolites in the hypothalamus. The binding of E2 to ERα results in a selective reduction of ceramide-induced lipotoxicity and consequently of ER stress in neuronal cells of the VMH, which in turn leads to a markedly increased BAT activity, weight loss, and improvement of glucose metabolism. 139,140 Peripherally estrogen modulation of BAT occurs by a direct action on ERα in adipocytes. As in the CNS, the effect of estrogens in adipocytes is mediated by AMPK. In contrast to central regulation, in the periphery, AMPK is activated by the binding of estrogen to its receptor. As in BAT triggered by the binding of norepinephrine to the β3-AR, activation of ERα in WAT leads to intracellular lipolysis with release of free fatty acids followed by disinhibition of UCP1. Furthermore, estrogen activity in white adipocytes leads to the expression of BAT-specific genes (e.g., of β3 AR and other markers for mitochondrial biogenesis) and consequently to browning ( Figure 2). 141 The mechanisms underlying estrogen-induced BAT thermogenesis described herein suggest promising clinical effects on REE, body weight development, and glucose and lipid metabolism. To date, results on this topic have almost exclusively come from animal studies, whereas findings in humans are still sparse and inconclusive.
Thus, the aim of this systematic review was to evaluate the effects of medication with natural and synthetic estrogens on REE in healthy normal weight and overweight (BMI 18.5-29.9 kg/m 2 ) women.

| Study selection process
The search strategy yielded a total of 973 articles published between 1929 and 2021. After the removal of duplicates by HJ 429, unique articles were identified. Titles, abstracts, and full texts were doublescreened independently by the authors SW and KW and tested against the inclusion criteria. Disagreements between individual judgments were resolved in a bilateral discussion. All English-, German-, or French-language clinical trials that investigated the effects of medication with natural and synthetic estrogens on REE in healthy normalweight and overweight (BMI 18.5-29.9 kg/m 2 ) women were included ( Figure 3).

| Data extraction
Relevant information from the included studies was collected for data extraction according to a predesigned protocol established by the two reviewers (SW and KW). For all studies, predefined study data (see Tables 1 and 2) were collected independently by the two reviewers (SW and KW), tabulated according to the PICO elements, F I G U R E 2 Estradiol and its central and peripheral effects on brown adipose tissue thermogenesis. Binding of estradiol (E2) to estrogen receptor α (ERα), expressed by both the central nervous system (CNS) and the adipose tissue, exerts stimulatory effects on brown adipose tissue (BAT) thermogenesis. Centrally, estradiol leads to increased sympathetic nervous system (SNS) activity by binding to its receptor in the ventromedial hypothalamus (VMH) and via intermediate connections in brainstem areas such as the nucleus raphe pallidus (RP) and the inferior olive (IO). Ultimately, triggered by the binding of noradrenaline to the β3 adrenergic receptor (β3 AR), intracellular lipolysis occurs in the BAT and white adipose tissue (WAT) with release of free fatty acids (FF) from triglycerides (TG), followed by disinhibition of UCP1. However, stimulation of the BAT by the SNS not only leads to an increase in its metabolic activity but also to an expansion of its volume as well as to the so-called "browning" of WAT. In the periphery, estradiol leads directly to metabolic activation and browning of WAT via binding to the ERα. among others, and then cross-checked against each other to ensure correct data collection. The main result of interest, the change in REE as a result of estrogen medication, was expressed variously in the different studies, either in absolute numbers, in p-values to indicate the significance of differences, or in purely literal textual terms.

| Data analysis and risk of bias assessment
The risk of bias was assessed for all included randomized controlled trials (RCTs) using the Cochrane Risk of Bias 2 (RoB 2) tool 144 and for all other included non-RCTs using the Thomas H. Quality assessment tool for quantitative studies, 145 which is independent of study design.
All included studies were reviewed by two of the review authors (SW and KW) to independently assess the risk of bias (assessment of methodological quality). Disagreements were resolved by bilateral consensus. With the exception of the study by Eck et al. 146 in which the methodological quality was rated as strong, corresponding to a low risk of bias, the methodological quality of all other studies [147][148][149][150][151][152][153][154] was only rated as moderate (Tables 3 and 4).

| Characteristics of included articles
Our systematic review identified a total of 10 studies investigating the effects of medication with natural and synthetic estrogens on REE and encompassing 461 healthy normal-weight and overweight women (Tables 1 and 2). [146][147][148][149][150][151][152][153][154][155] The 10 included studies consisted of six studies in which estrogens were used in the context of HC 146,[149][150][151][152]154 and four studies in which estrogens were used in the context of MHT. 147,148,153,155 Study designs were either RCTs, 150,151,153 prospective cohort studies, 147 or cross-sectional studies. 146,148,149,152,154,155 The sample size of each cohort with the population of interest ranged from 6 155 to 99 subjects. 152 The type of F I G U R E 3 PRISMA 2020 flow diagram.
T A B L E 1 Overview over the basic characteristics of the included studies.
Author  147,153 to no further specification. 155 CEE dose and mode of application in MHT varied from 0.625 mg/d orally 148 to no further information. 155 EE dose and form of application in contraception varied from 20 to 50 mcg/d orally 146,[149][150][151]154 to a generalized statement of inclusion of all HC methods, and thus also of progestin-only methods. 152 The type of progestin used in MHT varied between medroxyprogesterone acetate (MPA), 148,153 progesterone, 147 and no further details regarding the used progestin. 155 Two of the four MHT studies did not specifically distinguish between sequential and continuous-combined use of the progestin, 148,155 whereas the other two studies reported sequential use. 147,153 The progestins in the context of HC were levonorgestrel, 149,154 desogestrel, 149,154 and norethisterone, 146,149,154 or there was no more specific information about this detail. [150][151][152] The mean age of the subjects enrolled in the study groups of interest in the six contraceptive trials ranged between 19 152 and 26 149

| Impact of estrogens on resting energy expenditure
Regarding the six studies on contraceptives, 146,[149][150][151][152]154 the picture that clearly emerges is that the various HC methods exert a neutral influence on REE or may even slightly increase it. Thus, in the crosssectional study by Piers et al., the measured REE of women on combined oral contraceptives (OCs) adjusted for body weight was significantly higher than that of women not on combined OCs by 2.4% and +32 kcal/d (p = 0.014), respectively. REE adjusted for fat mass (FM) and fat-free mass (FFM), two major confounders of REE, was also significantly higher in women taking combined OCs as compared with women not taking combined OCs, by 2.8% corresponding to +37 kcal/d (p = 0.024). 154 In the study by Diffey et al., the REEenhancing effect of combined OCs was even more pronounced.
Although the raw (unadjusted) results showed a 3.7% increase in REE, equivalent to +208 kcal/d in combined OC users compared with combined OC never-users, the difference was not significant. However, after applying an analysis of covariance with either body weight or a combination of FM and FFM as covariates, the differences between T A B L E 2 (Continued) Author, year of publication, One of the greatest weaknesses of our systematic review, which limits the significance of its results, is not at least the insufficient availability of high-quality studies on our specific research question.
One of the strengths of our systematic review is the comprehensive literature search, covering not only estrogens but all sex steroids, to identify all relevant studies on this topic and to ensure that the available evidence is presented as fully as possible. We also restricted inclusion to normal-weight and overweight women to exclude confounding factors such as metabolic adjustments and hormonal alterations that typically occur when women are obese or underweight. As is typical for systematic reviews in general, our article is characterized by an objective methodology that ensures transparency and reproduc- Studies have shown that hormonal fluctuations during the menstrual cycle influence REE. [102][103][104][105][106][107][108] According to our data, suppression of these fluctuations by HC seems to have neutral or even positive effects on body weight regulation in terms of an increase in REE.
Regarding the influence of MHT on weight development in perimenopause and postmenopause, it should be pointed out that the weight gain associated with perimenopause and postmenopause is rather a consequence of the age-related tendency to reduced physical activity, the associated loss of muscle mass, and the overall reduced energy requirement. 157 Contrary to mainstream belief, MHT actually counteracts this age-related tendency to gain weight, with one of the mechanisms being an increase in REE.
The estrogenic effect on BAT activity may be the underlying reason for this observation. Indeed, considering the magnitude of the estrogen-associated increase in REE (ranging from no effect 146,147,[150][151][152]155 to +32 kcal/d, 154 +37 kcal/d, 154 +60 kcal/ d, 153 +208 kcal/d, 149 and +222 kcal/d 148 ), this corresponds to the additional energy expenditure associated with activation of BAT thermogenesis. It is estimated that BAT-specific energy expenditure can account for up to 2.5%-8% of REE. 36,48,158 The limited data available on this issue quantify the actual contribution of BAT thermogenesis to TEE at 7 kcal/d in thermoneutral environments to approximately 200 kcal/d in mild cold exposure. [158][159][160][161] This rather small contribution of BAT to TEE is mainly due to its low overall tissue mass of about 50-500 g 162 rather than its metabolic activity, which is still several times higher compared with other tissues of the human body. 161 As mentioned at the very beginning, the development of obesity is the consequence of a chronic disturbance of energy homeostasis with a persistent positive energy balance. The average weight gain underlying obesity, which usually commences in young adulthood, is 0.5-0.7 kg per year. This corresponds to a daily positive energy balance of just 12-17 kcal/day at an average energy density of 8840 kcal per kg body weight and lies thus within the range of the BAT-specific contribution to TEE 161,163 and estrogen-mediated activation of BAT thermogenesis, respectively. Hence, the contribution of BAT to REE may be small but not negligible. It seems that low but sustained activity of BAT thermogenesis may significantly counteract the development of obesity. However, current estimates of BAT thermogenesis capacity are at the low end of what would potentially be clinically possible through its chronic stimulation by various endogenous or exogenous factors. 161 It has been shown that the metabolic activity of BAT can be increased many-fold as well as BAT mass can be newly formed and existing depots can be significantly enlarged even in adulthood by long-term cold exposure or cold acclimatization 159 Finally, it should be emphasized that estrogens have an anorexigenic effect and thus prevent increased EI from compensating for increased REE. 168 Together with this fact, our results suggest that the energy balance changes in a negative direction under the influence of estrogens. Thus, the widespread fear of weight gain solely as a result of taking HC or MHT is therefore unsubstantiated. This article is intended to help clinicians educate women accordingly and alleviate their fears in this regard.