Unraveling the associations of age and menopause with cardiovascular risk factors in a large population-based study

This study presents a unique view of reproductive aging, independently of biological aging. We observed an association of CVD risk factors with menopausal status within several clusters of annual age strata, indicating that this relationship cannot be explained by the effects of chronological aging alone. The magnitude of differences in CVD risk factors between menopausal status groups did vary with age, highlighting the added role of chronological aging. Based on these results, it seems likely that both chronological aging and menopausal status contribute to the CVD risk profile of aging women.

Naturally postmenopausal women had lower adjusted SBP levels across a large age range than pre-, peri or surgically postmenopausal women. Prior reports found a later reproductive stage to be associated with increased blood pressure [9, 16, 26], while others reported a lack of any association after adjustment for age [6, 7, 13, 27, 28]. A longitudinal study in 193 women was the first to detect a decreased SBP level in post- compared to premenopausal women [29], hypothesizing that a diminishing ovarian reserve exhibits a protective effect on increasing SBP levels. By design, we cannot confirm this hypothesis, but our results do contest previous reports of an adverse blood pressure milieu in a peri- and postmenopausal state [9, 16, 26].

Where lipid levels are concerned, previous findings are less ambiguous and correspond well to our results. LDL-c and TC levels are widely thought to be influenced by the menopausal transition [6] or associated with menopausal status [4, 5, 7, 10, 30–33]. In fact, the approximate difference in LDL-c levels of 11 mg/dL (0.28 mmol/L) observed by Matthews et al. [6] between the year preceding and following the final menstrual period fits well within the range of our observations. The decrease of estradiol throughout the menopausal transition may not play a role in this regard, as TC and LDL-c levels did not correlate with total or free estradiol in 99 postmenopausal women [34]. On the other hand, post-menopausal hormone therapy was associated with a better lipid profile compared to placebo in a meta-analysis of 28 trials [35]. Another explanation is the reduced activity of LDL-c receptors or lipoprotein lipase in a postmenopausal state [36, 37].

In our population, differences in LDL-c and TC levels between menopausal status groups only became evident after the age of 45, after which LDL-c and TC levels more sharply increased in the peri- and postmenopausal groups. While a rapid increase in lipid levels was previously linked to the menopausal transition [4, 6], our results do suggest that chronological aging is equally involved. Indeed, the adjusted difference in TC and LDL-c values in the interval of 45–50 years was equal to the maximum observed differences between the menopausal status groups. It may be possible that, with increasing age, the availability of compensatory mechanisms to neutralize hyperlipidemia diminishes.

Surgically postmenopausal women, having undergone a bilateral oophorectomy, had consistently higher BMI and TG levels than the remaining women in the same age stratum, the latter even after adjusting for BMI. Others observed similar results [13, 38–42], with the odds of becoming obese specifically increasing after bilateral oophorectomy [41]. Interestingly, the adjusted BMI of pre-, peri- and naturally postmenopausal women hardly differed throughout the study population, which is in line with previous findings [38], but at odds with the observation that the menopausal transition influences fat distribution [15, 32].

For the past two decades, the relationship of menopause with CVD risk factors has been studied extensively through a myriad of ways. As most previous research was based on smaller study populations, often with significantly differing ages between pre- and postmenopausal groups, we hope to provide a substantial contribution to this age-old question with our study. Its strengths are the use of a large study population, with the ability to compare menopausal status groups and CVD risk factors within yearly age strata, over a wide age range. The protocolled assessment of study parameters and relative lack of missing information limit the chance of bias. Unfortunately, this was not quite the case for the classification of menopausal status. It is likely that some postmenopausal women using hormonal contraception or HT were classified as premenopausal due to the report of a regular cycle, and that some premenopausal women with an irregular cycle were wrongly classified as peri- or postmenopausal [43]. The exclusion of women using exogenous hormones did not have an obvious impact on the overall results, with the exception that the lipid profile of young postmenopausal women appeared notably more unfavorable than the other groups in this analysis. It is possible that this difference is due to the putative benefits of hormone supplementation in young women in particular [44], or incorrect classification of premenopausal women using hormonal contraception as postmenopausal. In order to be considered postmenopausal, women had to report in the questionnaire that they had entered menopause in addition to reporting an amenorrhea of at least a year, which makes large-scale misclassification in this category less likely. Moreover, the finding of very young women with non-iatrogenic menopause corresponds to our observations in clinical practice and other Dutch cohort studies and could therefore well be a realistic representation. Finally, due to the small number of women with surgical menopause, there is insufficient power to separately compare this group in all yearly strata. However, as this group of women represents a different clinical entity than natural menopause, we chose to maintain this classification.

As this was a cross-sectional study, our observations are limited to associations without drawing conclusions on causality. A previous study was able to longitudinally follow CVD risk factors [6], providing important information on the changes surrounding the menopausal transition. It is by definition impossible to distinguish these changes from general aging throughout the menopausal transition in the same participant, however, which is why our current study provides an important contribution from a different perspective. Although we are able to confirm previous reports of differences in lipid parameters based on menopausal status, the clinical implications of the observed differences may be limited. A reduction of LDL-c of 1.0 mmol/L was associated with a 22% decreased rate of major vascular events in an extensive meta-analysis of individual patient data [45], but this is difference is 2.5 to 10 times larger than menopause-related differences in this study or the study by Matthews et al. [6], and in fact more approximate to the differences found with 20 years of chronological age. It may be that the increased risk of cardiovascular events observed in post-menopausal women, the causality of which is a matter of debate in itself [11, 12, 46], is mediated through other pathways such as oxidative stress and inflammation [47]. A previous proposal of lipid screening of women entering the menopausal transition [6] may therefore not prove beneficial. That being said, vigilance of changing lipid parameters in high-risk women as they pass both biological and reproductive aging thresholds may be worthwhile.