The menopause transition, also known as perimenopause, is the beginning of menstrual irregularities when symptoms of female sex hormone deficiency begin. Menopausal change is associated with higher prevalence of metabolic syndrome and cardiovascular risk factors, as well as alterations in mood, sleep, diet and other lifestyle factors.
However, whether the reported changes associated with menopause are due to hormonal alterations, psychological changes associated with the transition, natural ageing, social and behavioural factors of midlife or genetic vulnerability is less clear and warrants further exploration.
Subsequently, unfavourable fasting blood measures
and a shift to an atherogenic lipid profile (increases in total cholesterol, low-density lipoprotein cholesterol (LDL-C), and apolipoprotein B) occur independently of age, due to menopause.
However, less is known regarding the impact of menopause on the integrated postprandial metabolic response.
Given that humans spend the majority of their day in the postprandial (1–8 h post eating) phase (∼18 h/d) and postprandial lipaemia and glycaemia are independent risk factors for cardiovascular diseases (due to their downstream effects on inflammation, oxidative stress, haemostatic function and lipoprotein remodelling), studies exploring multi-factorial postprandial responses with respect to menopausal status are needed. Furthermore, the gut microbiome is increasingly recognised as an important regulator of metabolism and is associated with multiple cardiometabolic risk factors.
Whilst alterations in gut microbiome composition have been shown during the menopausal transition, its role in increased metabolic risk faced by menopausal women remains unclear.
In light of the well-recognised changes that occur in lifestyle and body composition upon menopause, research furthering our understanding of the key metabolic and microbial changes occurring in concert may help provide tailored lifestyle and dietary advice for women during their menopausal transition and post-menopause. This study leveraged the densely phenotyped ZOE PREDICT cohort to, firstly, characterise lifestyle, diet and health measures in pre-, peri- and post-menopausal women and, secondly, explore the physiological changes of menopause with a focus on postprandial metabolism and the gut microbiome. We report; 1) differences in body composition, fasting blood measures, postprandial metabolites, lifestyle, diet, microbiome and mood across sex, age and menopausal status, 2) an independent association of menopause with postprandial glucose responses in an age-matched subgroup, 3) a protective association between menopausal hormone therapy (MHT) use and visceral fat and fasting and postprandial measures, and 4) a mediation effect of diet and bacterial species on visceral fat and inflammation, by menopause status.
The ZOE PREDICT 1 cohort
Compared to the average UK population, PREDICT 1 participants were older (mean age 46 vs 41 years respectively), had a lower BMI (26 vs 28 kg/m2 respectively), were less likely to smoke and the proportion of males was also lower (27% vs 49 % respectively).
Cohort characteristics and relationships with sex, age and menopausal status
Differences in the association of age with metabolic traits between males and females
Associations with menopausal status
Table 2Characteristics of the PREDICT 1 cohort across menopausal status.
*Scoring ranges from lowest “all of the time” to highest “none of the time”. ANCOVA (continuous) and logistic regression (categorical) adjusted for age and BMI. Anthropometric traits were adjusted for age only. Differences between menopausal groups also adjusted for MHT use and smoking status. All p-values adjusted for multiple testing (FDR<0.05); * p<0.05, ** p<0.01, ***p<0.001.+ PREDICT 1 females who self-reported menopausal status.
Menopausal status associations with postprandial responses and glycaemic variability
Therefore, as humans spend the majority of their day in the postprandial metabolic state, we next examined differences in postprandial responses between pre- and post-menopausal females (adjusting for age, BMI, MHT use and smoking status). Clinic postprandial metabolic responses differed between groups (Figure 2 f-i), with significantly higher glucose2hiauc and insulin2hiauc (pvs. pre-menopausal females.
Given the differences in postprandial glycaemia measured in the tightly controlled clinic setting, we then explored glycaemia in the remote phase of the study using CGM data. We examined different features of glycaemic responses, including glycaemic variability (measured by coefficient of variation), time spent in range, mean day long glucose concentration, as well as glucose2hiauc and peak0-2h following meals of varying macronutrient composition. Mean day-long glucose concentrations and glycaemic variability (examined using 2-4 free-living days of the PREDICT 1 remote phase) were higher in post-menopausal females (5.1±0.53 mmol/L and 17.6±4.3 %) compared to pre-menopausal females (4.9±0.54 mmol/L and 15.6±4.00 %), p<0.002 (ANCOVA) (Supplementary Table 6). Pre-menopausal females also elicited a more favourable TIR (3.9–5.6 mmol/L) (70.8%±16.9) compared with post-menopausal females (68.8%±15.6), p<0.05 (ANCOVA).
Menopausal status also corresponded to a state of greater inter-individual variability (coefficient of variation) in post- vs. pre-menopausal females for postprandial insulin (30 min, pre-menopause 89% vs. post-menopause 200%) and HOMA-IR (pre-menopause 69.4% vs. post-menopause 82.6%) (Supplementary Figure 1 and Supplementary Table 8).
Menopausal status associated with microbiome composition
Therefore, we examined gut microbiome composition, within-sample richness (number of species) and within-sample diversity (Shannon) in pre- and post-menopausal females (n=564). The relative abundances of microbiome species differed with menopausal status, with eight species significantly differentially abundant after correction for multiple testing (pFigure 2k). Of these, four species had significantly higher abundances in pre- vs. post-menopausal females, whereas four species had higher abundances post-menopause (pBacteroides ovatus has been associated with younger age in a large meta-analysis,
but none of the other species were associated with age or BMI. Microbiome richness and diversity were not significantly different (Supplementary Table 9).
Is the association of menopause with metabolic traits independent of age?
When investigating differences in microbiome composition, in the age matched cohort, the abundances of six species were significantly different between pre- and post-menopausal females but were not significant after FDR correction (Supplementary Table 9). From the eight species previously identified in the total cohort, four species showed the same directional trend but were not significant after FDR correction (Eubacterium hallii, Bifidobacterium adolescentis, Faecalibacterium prausnitzii, Oscillibacter sp PC13). The remaining four species, (Harryflintia acetispora, Bacteroides ovatus, Lawsonibacter asaccharolyticus and Clostridium disporicum) were not abundant in the same directional trend in the age-matched subgroup.
Due to the deficiency in female sexual hormones observed post-menopause we also examined measures in age-matched 1) males, 2) pre-menopausal females and 3) post-menopausal females (age range 47–56y, Supplementary Table 10). Females, both pre- and post-menopausal, had significantly lower SBP and ASCVD 10y risk compared to males. However, pre-menopause was associated with more favourable fasting blood glucose and GlycA concentrations, and mean day long glucose concentrations (CGM) compared to males (p<0.05) (ANCOVA), while male levels were more similar to post-menopausal females. Interestingly, glycaemic variability, which worsened post-menopause (compared with age-matched pre-menopausal females), was also significantly higher than males (p=0.007) (ANCOVA), suggesting unfavourable blood glucose variability independent of age and gender.
The association between post-menopausal hormone therapy use and metabolic health
we examined the link between MHT use and metabolic health in post-menopausal females (Table 3; adjusted for age, BMI and smoking status). Measures that were significantly different with menopausal status in the total cohort were examined along with a selection of measures with previously known associations with MHT (visceral fat and lipids (LDL-C and triglyceride6hiauc)).
Post-menopausal females using MHT (n=35) had lower visceral fat mass, favourable fasting blood concentrations (for glucose, insulin, cholesterol (total and LDL)), and lower postprandial lipaemia (triglyceride6hiauc) compared to non-MHT users (n=172), pTable 3). Furthermore, to disentangle the effect of genetics, we examined our predominantly twin cohort and identified six post-menopausal MZ twin pairs discordant for MHT use. HbA1c showed significant differences between discordant twin pairs (p=0.004) (Paired t-test) (Supplementary Table 12). We observed a more favourable body composition (lower BMI, weight, visceral fat) and blood biomarker concentrations (lower glucose, insulin, TG and GlycA) in twins using MHT vs. those non-MHT users (although not significant).
Table 3Menopausal hormone therapy use in post-menopausal females.
*Post-menopausal females who self-reported currently taking MHT.
Mediating effects of sleep, physical activity, diet and microbiome on the link between menopausal status and key metabolic health indicators
Four microbiome species (Collinsella intestinalis, Eggerthella lenta, Flavonifractor plautii and Ruminococcus gnavus) acted as partial mediators in the association between menopause status and GlycA (proportion mediated; 5–10%) (Figure 4b and Supplementary Table 13). The species Flavonifractor plautii and Eggerthella lenta acted as a partial mediators in the association between menopause status and visceral fat (proportion mediated; 5%) and glucose (peak0-2h) (proportion mediated; 8%, pOscillibacter sp 57 20 acted as partial mediators in the association between menopause status and glucose (peak0-2h) (proportion mediated; −6%) (Figure 4c) (p
In the current study, we demonstrate that post-menopause status is associated with unfavourable changes in body composition, fasting and postprandial blood profiles (including inflammation and postprandial glycaemia), diet, sleep and gut microbiome. We further explored the independent role of menopause from age and observed poorer sleep and diet, as well as higher postprandial glycaemic measures post-menopause, alongside a protective association of MHT use with visceral fat, fasting (glucose and insulin) and postprandial (triglyceride6hiauc) blood measures. Finally, we investigated the association between modifiable risk factors on metabolic changes in menopause, finding a mediating effect of diet and a gut bacterial species and visceral fat, glycaemia and inflammation, by menopause status. Changes in key metabolic health indicators observed in menopause may therefore be attenuated by targeting the gut microbiome and diet.
Here, we show that in addition to 2h iAUC for glucose and insulin, post-menopausal females had a more unfavourable glycaemic variability, TIR and day-long glucose, measured by CGM in free-living days. Continuous glycaemic features capture day-to-day glycaemic excursions, including peak concentration, nadirs ‘below baseline’, glycaemic variability and TIR, each associated with downstream metabolic effects, including oxidative stress, inflammation, and increased cardiovascular and diabetes disease risk.
In this cohort, we did not see differences in postprandial TG independent of age. To the best of our knowledge, one previous study has compared postprandial TG response between pre- and post-menopausal females using a sequential meal postprandial investigation.
Females were subdivided into both younger and older pre- and post-menopause subgroups but no differences in postprandial TG were observed due to menopausal status.
Differences were observed between young and old pre-menopausal groups, suggestive that major increases observed in postprandial TG may occur during the later pre-menopause years.
This study highlights that inherent biological differences exist between males and females while also demonstrating the protective effects of oestrogen through the controlled comparison of age-matched males and pre- and post-menopausal females. Multiple aspects of health and glucose homeostasis are regulated differently between sexes, and our findings highlight the added complexity introduced by the female menopausal transition.
However, MHT is available in multiple forms and doses, and the impact of different therapies is beyond the scope of this research. For example, transdermal oestradiol and micronised progesterone are not associated with a risk of venous thromboembolism compared to oral oestrogen with a synthetic progestogen.
Given the appropriate dose and type, MHT can reduce biological ageing and provide effective treatment to alleviate menopausal symptoms and confer protective cardiometabolic effects in appropriate candidates.
which has been associated with multiple diseases outside of this organ. The gut microbiota metabolizes oestrogen-like compounds consumed in plant foods (phytoestrogens), including lignans (derived from a variety of plant foods) and isoflavonoids (found in soy foods).
Administration of isoflavone-rich soy foods to post-menopausal females can lead to elevated levels of gut microbial derived oestrogen-like metabolites
and changes in some bacteria including increases in Bifidobacterium and decreases of Clostridiaceae, which play roles in inflammatory diseases.
Further, changes in oestrogen receptor (ER-β) expression have been shown to affect microbiota composition
and the gut microbiomes of pregnant females were profoundly altered during the third trimester, when oestrogen is at its peak.
Thus a reciprocal relationship may exist between oestrogen and the microbiome which may modulate the health of menopausal females. Our findings show differences in abundances of species post-menopause, including pro-inflammatory and obesogenic bacteria. Of interest, four species in part modulated the relationship between menopause and GlycA, a marker of inflammation. Our previous research associated these species with unfavourable cardiometabolic health, diet and inflammatory outcomes,
in line with previous research.
For example, Ruminococcus gnavus, a prevalent gut microbe, produces a potent, inflammatory polysaccharide recognized by innate immune cells through the toll-like receptor 4 (TLR4) and is associated with multiple inflammatory diseases.
GlycA is strongly associated with cardiovascular and diabetes risk
; thus, our data suggests that inflammation may be reduced through intervention aimed at improving the gut microbiome post-menopause.
which are known to impact metabolic health outcomes, these are potential targets to alleviate some of the downstream unfavourable health effects associated with menopause. For example, in our cohort, post-menopausal females consumed higher intakes of dietary sugars and reported poorer sleep. These are both associated with more pronounced postprandial glycaemia
and increased risk of type-2 diabetes and cardiovascular disease.
Further, a decrease in physical activity energy expenditure and a shift to a more sedentary lifestyle associated with menopause
was recently demonstrated to be a direct effect of declining oestrogen
which may have been captured with more quantitative measures of physical activity in this study. These observed changes in diet and physical activity may increase the risk for positive energy balance and weight gain over time. Healthy dietary patterns such as the Mediterranean diet have been associated with improved weight and vasomotor symptoms in post-menopausal females
and certain foods have also been linked to later onset of menopause i.e. intakes of green and yellow vegetables as well as fresh legumes.
Positive health effects associated with diet quality, may be due to higher contents of dietary fiber, complex carbohydrates, vitamins, minerals, polyunsaturated fatty acids, and phytochemicals. Diets high in plant-based foods may be naturally rich in isoflavones which may play a role in the protective effects some diet patterns such as Mediterranean diet have on menopause. Research shows that 60% of women consulted healthcare providers during their menopausal transition seeking support for menopausal symptoms and treatment,
highlighting an opportune window to target modifiable factors including diet and lifestyle as well as considering MHT.
Limitations of this cross-sectional analysis include; 1) potential inaccuracy in the self-reported time since menopause and self-identification of menopausal status due to ambiguity in determining amenorrhea; 2) cross-sectional data preventing identification of causal relationships; 3) inability to fully account for age-related effects despite the creation of an age-matched subgroup and assessment of MHT discordance; 4) inability to determine menstrual cycle regularity, current contraception use, MHT type (transdermal vs. oral) and other current medication use. However, the data presented links changes in postprandial metabolism, metabolic syndrome factors, mood, sleep, diet and the gut microbiome in a single deeply phenotyped cohort. This may help inform specific hypotheses to design dietary intervention studies examining the impact of menopause status on postprandial metabolism and microbiome composition.
In summary, the physiological effects of menopause are numerous and the menopause transition is a time of great change and unfavourable metabolic effects. Whilst this transition is inevitable, this analysis demonstrates that approaches can be taken to attenuate the adverse cardiometabolic sequelae, including a focus on modifiable factors, such as diet, microbiome and use of MHT in appropriate candidates.
Study design and developed concept: S.E.B, A.M.V, J.W, G.H, R.D, N.S, P.W.F, T.D.S. Data collection: S.E.B, I.L, J.W, G.H, T.D.S. Data analysis: K.M.B, F.A. Study coordination: S.E.B, J.W, I.L, G.H, T.D.S. Writing the manuscript: K.M.B, S.E.B, A.M.V, W.L.H, I.L, K.K, J.W, J.E.M, L.R.N, L.D, J.M.O, A.T.C, T.D.S. Obtained funding: J.W, G.H, T.D.S. Accessed and verified the data: K.M.B, F.A, S.E.B, A.M.V. Decided to submit the manuscript: S.E.B and K.M.B. All authors read and approved the final manuscript.
Declaration of interests
TDS, JW and GH are co-founders of ZOE Ltd (ZOE). AMV, FA, LMD, NS, PWF, SEB, TDS receive payments as consultants to ZOE. GH, IL, JW, KK, RD are employed by ZOE. AMV, GH, IL, JW, KK, LMD, NS, PWF, RD, SEB, TDS also receive options in ZOE. Other authors have no conflict of interest to declare.