Abstract
Objective: This study examined relationships, by pregnancy histories, between bone mineral density (BMD) and coronary artery calcification (CAC) in postmenopausal women.
Methods: Forty women identified from their medical record as having pre-eclampsia (PE) were age/parity-matched with 40 women having a normotensive pregnancy (NP). Vertebral (T4–9) BMD and CAC were assessed by quantitative computed tomography in 73 (37 with PE and 36 with NP) of the 80 women. Analyses included linear regression using generalized estimating equations.
Results: Women averaged 59 years of age and 35 years from the index pregnancy. There were no significant differences in cortical, trabecular or central BMD between groups. CAC was significantly greater in the PE group (p = 0.026). In multivariable analysis, CAC was positively associated with cortical BMD (p = 0.001) and negatively associated with central BMD (p = 0.036). There was a borderline difference in the association between CAC and central BMD by pregnancy history (interaction, p = 0.057).
Conclusions: Although CAC was greater in women with a history of PE, vertebral BMD did not differ between groups. However, both cortical and central BMD were associated with CAC. The central BMD association was marginally different by pregnancy history, suggesting perhaps differences in underlying mechanisms of soft tissue calcification.
Introduction
Pre-eclamptic pregnancy is a risk factor for future cardiovascular disease including chronic hypertensionCitation1, coronary artery calcification (CAC)Citation2, adverse cardiac eventsCitation3, and strokeCitation4. In previous studies of postmenopausal women, including those using menopausal hormone treatmentsCitation5, CAC negatively correlated with both areal bone mineral density (aBMD) as measured by dual-energy X-ray absorptiometry (DXA) at the femur and lumbar spineCitation6–8 and with lumbar spinal volumetric bone mineral density (vBMD) in the trabecular region, as measured by quantitative computer tomography (QCT)Citation9. Markers of bone turnover such as osteocalcin and the degradation products of C-terminal telopeptides of type I collagen increased during pre-eclamptic pregnanciesCitation10,Citation11; however, these markers did not correlate with significant decreased aBMD during the pregnancy term, as measured using bone ultrasound at the heel and DXA at the femur and lumbar spineCitation10–12.
Our group has previously reported that CAC is greater in women with a history of pre-eclamptic pregnancy compared to age- and parity-matched women with a history of normotensive pregnancyCitation2. However, long-term follow-up studies of BMD in postmenopausal women with histories of pre-eclampsia have yet to be conducted. Therefore, the aim of this study was to examine the impact of pregnancy history, specifically a history of pre-eclampsia, on the relationship of BMD with CAC in postmenopausal women.
Methods
The study was approved by the Mayo Clinic Institutional Review Board and all women gave written informed consent. Participants were recruited from the Rochester Epidemiology Project as previously detailedCitation2. In brief, 40 women who had experienced pre-eclamptic pregnancies (PE) and 40 age- and parity-matched women who had experienced normotensive pregnancies (NP) between the years 1976 and 1982 were recruited. All participants were white Caucasian. Clinical exclusion criteria included previous diagnosis of myocardial infarction, congestive heart failure, stroke, dementia, cancer (excluding non-melanoma skin cancer), autoimmune disease, and neurological conditions. Demographic and clinical data were obtained through review of the medical record, participant interview, and physician examination. Thoracic QCT was acquired in all but one participant (Siemens Sensation 64, Siemens Medical Solutions, Forchheim, Germany), allowing for assessment of CAC (using the Agatston scoring methodCitation13) and thoracic vertebral vBMD as previously described and validatedCitation14,Citation15. Both CAC and thoracic vertebral vBMD were obtained from the same scans. Both measurements were ancillary studies to a main protocol, the primary outcome of which was detection of cognitive impairment in a sample of 40 women with and 40 women without a history of hypertensive pregnancyCitation16. Based on a two-sided two-sample t test, assuming equal group variances with a type I error of 0.05, the study had 80% power to detect an effect size of 0.63 for the comparison of mean cognitive domain scores. This size effect was based on data derived from the Kronos Early Estrogen Prevention Cognitive and Affective StudyCitation17.
Acquisition of vBMD by QCT at the site of the thoracic spine allowed for differential analysis of three bone layers: cortical (dense exterior), trabecular (spongy interior), and central (subset of trabecular bone, designed to be similar to trabecular values as reported from DXA scans)Citation14,Citation15. Briefly, the Spine Cancer Assessment program (Biomedical Imaging Resources, Mayo Clinic) was utilized manually to segment all thoracic vertebrae in the field of the image, at both the superior and inferior endpoints. An elliptical mask was then manually applied in the transverse view to exclude the spinal processes. The software compared the voxel composition of the central slice of each vertebra to a calibration phantom (Mindways Software, Austin, TX, USA) that was placed in each scan, producing accurate vBMD values of the central, cortical and trabecular regions.
An exploratory study using a customized protein arrayCitation18 assessed levels of osteopontin (OPN) and osteoprotegerin (OPG) in serum from a subset of age-matched participants with and without calcification, stratified on pregnancy group (total n = 39, 14 from the NP group and 25 from the PE group). Serum levels of both proteins are presented as intensity (average pixel value).
Baseline characteristics are shown as number (percent) or median (25th, 75th percentiles), as appropriate, and were compared between the PE and NP groups using Pearson χ2 tests for categorical variables and Wilcoxon rank sum tests for continuous variables. For the purposes of describing central, cortical, and trabecular vBMD endpoints, repeated measurements for each outcome were collapsed into a single measure per participant by averaging over vertebral values. For group comparisons, however, normal linear regression models were fitted for each endpoint using all available values, with generalized estimating equations to account for correlation of repeated measurements on the same participant. In models adjusting for age and body mass index (BMI), differences in vBMD between pregnancy groups were estimated with least squares (LS) means and 95% confidence intervals.
In addition, the independent association of vBMD measures with CAC score, after adjustment for pregnancy hypertension, was examined using multivariable regression. Because CAC scores were highly skewed with excessive ties at zero, we used a semi-parametric rank-based modeling approach described in detail previouslyCitation2. Specifically, we fitted a proportional odds ordinal logistic model with the 'log-odds of greater CAC' regressed on the BMD measures and pregnancy history group, with and without their interactions. Terms for trabecular BMD were ultimately removed from the model to avoid multicollinearity. Evidence of differential associations between BMD measures and CAC according to pregnancy history was based on the tests for interaction embedded in this model.
For primary analyses, CT scans of five participants included misplaced calibration phantoms that were not recognized by the Spine Cancer Assessment program. Although previous studies have reported high concordance between measures of vBMD acquired from calibrated and uncalibrated scansCitation15, we repeated these analyses after exclusion of uncalibrated scans; differential results, when observed, were reported. As a separate exploratory analysis of the two ancillary protein measures, we fitted linear models to compare log-transformed values by pregnancy history and adjusted for presence of calcification due to the stratified sampling design. All analyses were conducted using the SAS statistical software package (version 9.4, SAS Institute Inc., Cary, NC, USA).
Results
The final number of participants included in this analysis were 37 in the PE group and 36 in the NP group. Reasons for exclusion included failure to obtain QCT imaging (n = 1) and distorted/corrupted imaging files (n = 6). Clinical variables that were significantly greater in the PE group at the time of the assessments included: BMI, blood insulin concentration, HOMA insulin resistance (HOMA-IR), diagnosis of hypertension, and use of antihypertensive medication (). None of the women self-reported taking calcium supplements with or without a vitamin D supplement and only two women reported using a calcium channel antagonist for hypertension.
Table 1. Demographic and clinical variables of women with histories of normotensive and pre-eclamptic pregnancies.
Data on a total of 338 thoracic vertebrae, ranging from T4 to T9, were captured from all 73 participants within the field of view of the thoracic QCT scan, although this entire range was not captured for each participant. The median number of thoracic vertebrae scanned per participant was five (ranging from three to six); all scanned vertebrae were included in the comparisons between the two pregnancy groups. All QCT scans (n = 73) included T7, T8 and T9 measurements, while 71 (97%) included T6, 43 (59%) included T5, and five (7%) included T4.
There were no significant differences in vBMD for any of the regions studied (cortical, central, or trabecular) between women relative to their pregnancy history (, and ). Differences remained insignificant both after exclusion of uncalibrated scans and after adjustment for age and BMI (data not shown).
Figure 1. Comparison of volumetric bone mineral density (vBMD) by pregnancy history (normotensive and pre-eclamptic) across three cortical (left panel), central (middle panel), and trabecular (right panel) bone regions. Data points represent the average vBMD derived from all vertebrae scanned for a single study participant. Bar represents median, box represents the 25th and 75th percentiles, and the vertical lines represent the range. No significant differences were observed between groups for any bone region.
Figure 2. Estimated differences in bone mineral density (BMD) between normotensive pregnancy and pre-eclamptic pregnancy groups (age- and body mass index-adjusted).
Table 2. Volumetric bone mineral densities and bone area differences for women with histories of normotensive or pre-eclamptic pregnancies. Data are given as median (25th, 75th percentiles). Descriptive statistics are based on a single measure per individual from averaging over vertebra-specific measurements, whereas the analysis for a group difference was conducted using all available measurements in linear models after adjustment for age and body mass index.
Coronary calcification was greater in women with a history of PE () with CAC scores ranging from 0 to 625 AU in this group (). The relationship between BMD and CAC appeared to vary in direction and magnitude between the groups (NP, PE), with seemingly higher positive correlations observed in the PE group for each of the vBMD measures (cortical, trabecular, and central; , ). In the additive multivariable model relating BMD measures with CAC by pregnancy history, cortical BMD was positively associated with CAC (p < 0.001) while central BMD was negatively associated with CAC (p = 0.021). The tests for interaction in the non-additive model indicated no evidence of a differential association for cortical BMD and CAC (p = 0.309). However, there was a trend for central BMD and CAC by pregnancy history (p = 0.057). Based on contrasts derived from the interaction effect, there was a significant negative association between central vBMD and CAC in the NP group, whereas no significant association was observed in the PE group (). In sensitivity analyses, similar results were obtained when uncalibrated scans were excluded, and when additionally controlling for BMI and HOMA-IR in the model.
Figure 3. Regression analysis comparing volumetric bone mineral density (vBMD) from cortical (left panel), trabecular (middle panel) and central (right panel) regions of the spine with coronary artery calcification (coronary artery calcification (CAC) scores in log of Agatston units) in menopausal women with histories of pre-eclamptic and normotensive pregnancies. Both calibrated and uncalibrated scans were included. Each data point represents the average vBMD of all captured vertebrae of a single study participant, along with their CAC score.
Table 3. Relationships between coronary artery calcification (CAC) and bone mineral density (BMD) in postmenopausal women with histories of normotensive or pre-eclamptic pregnancies.
In the subset analysis performed on 39 women, median (IQR) osteoprotegerin levels were 0.27 (0.12, 0.60) in the NP group compared to 0.58 (0.36, 0.89) in the PE group (, left panel), whereas median (IQR) osteopontin levels were 0.90 (0.66, 2.41) in the NP group and 0.71 (0.27, 0.95) in the PE group (, right panel). After adjusting for presence of calcification (yes vs. no) to account for the stratified sampling design, women with a history of PE had significantly elevated osteoprotegerin levels (p = 0.005) and significantly decreased osteopontin levels (p = 0.044) compared to women with a history of NP.
Figure 4. In the subset of 39 women in whom data were available, median with 25th and 75th quartile range of intensity (average pixel values) for osteoprotegerin (OPG, left panel) and osteopontin (OPN, right panel) measured by customized protein array in serum from women with histories of normotensive or pre-eclamptic pregnancies. After adjustment for calcification (positive vs. negative), there were significant differences in OPG (p = 0.005) and OPN (p = 0.044) levels between groups.
Discussion
There are three important findings from this study. First, the vertebral total vBMD did not vary by pregnancy history in women 35 years after the index pregnancy. This finding is consistent with those of shorter-term prospective, observational studies that found no differences in aBMD, as measured using bone ultrasound at the heel at 14–20 and 36–38 weeks during the pregnancy among women with normotensive, gestational hypertension or pre-eclamptic pregnancies, and DXA at the femur and lumbar spine within 2 days after deliveryCitation12,Citation19. Thus, the current study extends these observations to 35 years after the index pregnancy.
The second important finding of this study is that there was a significant association of cortical and central vBMD with CAC. In particular, results of our multivariable analysis revealed a negative partial association between central vBMD and CAC, an observation that is consistent with previous studies noting a negative correlation between CAC and vBMD in the general population of postmenopausal womenCitation6–9. The additional finding of a positive partial association for cortical vBMD with CAC suggests that contrast between cortical and central BMD may be more critical than either region alone.
Third, there was borderline evidence of a differential relationship of central BMD with CAC between the two groups, which may be related to differences in calcium metabolism and differentiation of osteoclasts in women with and without histories of PE. Higher serum levels of osteoprotegerin and lower osteopontin levels in women with a history of PE are consistent with decreases in osteoclast differentiation and bone resorptionCitation20. Although both of these proteins have been implicated in BMD and in vascular calcification, it is not clear that the calcification processes in these distinct tissues are the sameCitation21, as emerging data from transcriptional profiling of calcifying vascular smooth muscle cells and osteoblasts identified distinct patterns of gene regulation between the two types of cellsCitation22. Although there is a relationship between these proteins, type 2 diabetes and BMDCitation23,Citation24, studies have reported such associations for men and not womenCitation24.
It is unlikely that the difference in these protein levels and the presence of CAC was due to differences in calcium-related medications or vitamins since the use of these products was minimal. Calcium homeostasis changes during pregnancy and differs significantly between pregnancies with pre-eclampsia and those without. Urinary calcium excretion as early as the first trimester is lower in pregnancies that subsequently develop pre-eclampsia compared to normal pregnancyCitation25, and calcium supplementation lowers the rate of pre-eclampsia in pregnant women and is now recommended during pregnancyCitation26. Whether differences in calcium metabolism persist in women after pregnancy in the two groups is not clear. Lower calcium intake during pregnancy appears to be associated with increased risk of subsequent development of hypertension, especially in women with a history of hypertension in pregnancyCitation27, and calcium supplementation in women with a history of pre-eclampsia was reported to lower blood pressureCitation28. The relationship of calcium to blood pressure may be through decreased production of the vasodilator nitric oxide, as the enzymatic production is calcium-dependentCitation29. Taken together, these data are consistent with the possibility of long-lasting effects or perhaps genetic factors that influence calcium metabolism in women who have experienced pre-eclamptic pregnancy. As CAC is an important predictive indicator for cardiovascular and non-cardiovascular diseaseCitation30,Citation31, results from the present study warrant investigation into calcium and bone regulatory pathways in women with histories of pre-eclampsia that might predispose these women to CAC.
A limitation of the present study is the small number of women that were prospectively assessed. Clearly, the present observations need to be confirmed in larger numbers of women in whom pregnancy history is defined. This limitation is balanced by the strength of the present study in that pregnancy histories were carefully identified by review of the medical records and not by self-report. In addition, the phenotypic characterization of these asymptomatic women is consistentCitation2,Citation16,Citation32–35 with the epidemiological evidence supporting that having a history of pre-eclampsia puts a woman at risk for future cardiovascular diseaseCitation36–39.
Conclusions
Although the risk of coronary calcification increases with a history of PE, vBMD does not differ in women with a history of PE from age- and parity-matched women with a history of NP. The negative association of central vBMD with CAC in the NP group contrasts with the lack of association (negative or positive) of central vBMD with CAC in the PE group. These associations of vBMD with CAC, in addition to differences in proteins associated with osteoclast differentiation between the groups, may reflect differences in regulatory processes associated with soft tissue calcification that might put women at risk for pre-eclampsia and future cardiovascular disease.
Conflict of interest
The authors report no conflict of interests. The contents are solely the responsibility of the authors and do not necessarily represent the official view of the National Institutes of Health.
Additional information
Funding
References
- Magnussen EB, Vatten LJ, Smith GD, Romundstad PR. Hypertensive disorders in pregnancy and subsequently measured cardiovascular risk factors. Obstet Gynecol 2009;114:961–70
- White WM, Mielke MM, Araoz PA, et al. A history of preeclampsia is associated with a risk for coronary artery calcification 3 decades later. Am J Obstet Gynecol 2016;214:519 e1–8
- Mosca L, Benjamin EJ, Berra K, et al. Effectiveness-based guidelines for the prevention of cardiovascular disease in women–2011 update: a guideline from the American heart association. Circulation 2011;123:1243–62
- Bushnell C, McCullough LD, Awad IA, et al. Guidelines for the prevention of stroke in women: a statement for healthcare professionals from the American heart association/American stroke association. Stroke 2014;45:1545–88
- Bakhireva LN, Barrett-Connor EL, Laughlin GA, Kritz-Silverstein D. Differences in association of bone mineral density with coronary artery calcification in men and women: the rancho bernardo study. Menopause 2005;12:691–8
- Barengolts EI, Berman M, Kukreja SC, Kouznetsova T, Lin C, Chomka EV. Osteoporosis and coronary atherosclerosis in asymptomatic postmenopausal women. Calcif Tissue Int 1998;62:209–13
- Celik C, Altunkan S, Yildirim MO, Akyuz M. Relationship between decreased bone mineral density and subclinical atherosclerosis in postmenopausal women. Climacteric 2010;13:254–8
- Choi SH, An JH, Lim S, et al. Lower bone mineral density is associated with higher coronary calcification and coronary plaque burdens by multidetector row coronary computed tomography in pre- and postmenopausal women. Clin Endocrinol 2009;71:644–51
- Banks LM, Lees B, MacSweeney JE, Stevenson JC. Effect of degenerative spinal and aortic calcification on bone density measurements in post-menopausal women: links between osteoporosis and cardiovascular disease?. Eur J Clin Invest 1994;24:813–17
- Dorota DK, Bogdan KG, Mieczyslaw G, Bozena LG, Jan O. The concentrations of markers of bone turnover in normal pregnancy and preeclampsia. Hypertens Pregnancy 2012;31:166–76
- Gorzelak M, Darmochwal-Kolarz D, Jablonski M, et al. The concentrations of osteocalcin and degradation products of type I collagen in pregnant women with pre-eclampsia. Eur J Obstet Gynecol Reprod Biol 2001;98:23–7
- Sowers M, Scholl T, Grewal J, Chen X, Jannausch M. IGF-I, osteocalcin, and bone change in pregnant normotensive and pre-eclamptic women. J Clin Endocrinol Metab 2001;86:5898–903
- Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M Jr., Detrano R. Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol 1990;15:827–32
- Miyabara Y, Camp J, Holmes D, et al. Coronary arterial calcification and thoracic spine mineral density in early menopause. Climacteric 2011;14:438–44
- Miyabara Y, Holmes D, Camp J, Miller VM, Kearns AE. Comparison of calibrated and uncalibrated bone mineral density by CT to DEXA in menopausal women. Climacteric 2012;15:374–81
- Fields JA, Garovic VD, Mielke MM, et al. Preeclampsia and cognitive impairment later in life. Am J Obstet Gynecol 2017;217:74 e1–11
- Gleason CE, Dowling NM, Wharton W, et al. Effects of hormone therapy on cognition and mood in recently postmenopausal women: findings from the randomized, controlled KEEPS-Cognitive and Affective Study. PLoS Med 2015;12:e1001833
- Miller VM, Lahr BD, Bailey KR, Heit JA, Harman SM, Jayachandran M. Longitudinal effects of menopausal hormone treatments on platelet characteristics and cell-derived microvesicles. Platelets 2016;27:32–42
- To WW, Wong MW. Bone mineral density changes in pregnancies with gestational hypertension: a longitudinal study using quantitative ultrasound measurements. Arch Gynecol Obstet 2011;284:39–44
- Shapses SA, Cifuentes M, Spevak L, et al. Osteopontin facilitates bone resorption, decreasing bone mineral crystallinity and content during calcium deficiency. Calcif Tiss Int 2003;73:86–92
- Tintut Y, Demer LL. Recent advances in multifactorial regulation of vascular calcification. Curr Opin Lipidol 2001;12:555–60
- Alves RD, Eijken M, van de Peppel J, van Leeuwen JP. Calcifying vascular smooth muscle cells and osteoblasts: independent cell types exhibiting extracellular matrix and biomineralization-related mimicries. BMC Genomics 2014;15:965
- Vattikuti R, Towler DA. Osteogenic regulation of vascular calcification: an early perspective. Am J Physiol Endocrinol Metab 2004;286:E686–96
- Berezin AE, Kremzer AA. Circulating osteopontin as a marker of early coronary vascular calcification in type two diabetes mellitus patients with known asymptomatic coronary artery disease. Atherosclerosis 2013;229:475–81
- McMaster KM, Kaunitz AM, Burbano de Lara P, Sanchez-Ramos L. A systematic review and meta-analysis of hypocalciuria in pre-eclampsia. Int J Gynaecol Obstet 2017;138:3–11
- Hofmeyr GJ, Lawrie TA, Atallah AN, Duley L, Torloni MR. Calcium supplementation during pregnancy for preventing hypertensive disorders and related problems. Cochrane Database Syst Rev 2014;(6):CD001059
- Egeland GM, Skurtveit S, Sakshaug S, Daltveit AK, Vikse BE, Haugen M. Low calcium intake in midpregnancy is associated with hypertension development within 10 years after pregnancy: the Norwegian mother and child cohort study. J Nutr 2017;147:1757–63
- Hofmeyr GJ, Seuc AH, Betran AP, et al. The effect of calcium supplementation on blood pressure in non-pregnant women with previous pre-eclampsia: an exploratory, randomized placebo controlled study. Pregnancy Hypertens 2015;5:273–9
- Boeldt DS, Bird IM. Vascular adaptation in pregnancy and endothelial dysfunction in preeclampsia. J Endocrinol 2017;232:R27–44
- Handy CE, Desai CS, Dardari ZA, et al. The association of coronary artery calcium with noncardiovascular disease: the multi-ethnic study of atherosclerosis. JACC Cardiovasc Imaging 2016;9:568–76
- Kianoush S, Al Rifai M, Cainzos-Achirica M, et al. An update on the utility of coronary artery calcium scoring for coronary heart disease and cardiovascular disease risk prediction. Curr Atheroscler Rep 2016;18:13
- Miller VM, Garovic VD, Bailey KR, et al. Pregnancy history and blood-borne microvesicles in middle aged women with and without coronary artery calcification. Atherosclerosis 2016;253:150–5
- Garovic VD, Milic NM, Weissgerber TL, et al. Carotid artery intima-media thickness and subclinical atherosclerosis in women with remote histories of preeclampsia: results from a Rochester Epidemiology Project-based study and meta-analysis. Mayo Clin Proc 2017;92:1328–40
- Jayachandran M, Garovic VD, Mielke MM, Bailey KR, Lahr BD, Miller VM. Characterization of intravascular cellular activation in relationship to subclinical atherosclerosis in postmenopausal women. PLoS One 2017;12:e0183159
- Raman MR, Tosakulwong N, Zuk SM, et al. Influence of preeclampsia and late-life hypertension on MRI measures of cortical atrophy. J Hypertens 2017;35:2479–85
- Weissgerber TL, Turner ST, Bailey KR, et al. Hypertension in pregnancy is a risk factor for peripheral arterial disease decades after pregnancy. Atherosclerosis 2013;229:212–16
- Weissgerber TL, Milic NM, Milin-Lazovic JS, Garovic VD. Impaired flow-mediated dilation before, during, and after preeclampsia: a systematic review and meta-analysis. Hypertension 2016;67:415–23
- Garovic VD, August P. Preeclampsia and the future risk of hypertension: the pregnant evidence. Curr Hypertens Rep 2013;15:114–21
- Garovic V, Bailey K, Boerwinkle E, et al. Hypertension in pregnancy as a risk factor for cardiovascular disease later in life. J Hypertens 2010;28:826–33