Large Artery Stiffness is Elevated in Women with a Prior Pregnancy Conceived via In Vitro Fertilization

Abstract

In vitro fertilization (IVF) is an effective treatment for infertility; however, IVF may increase future maternal cardiovascular disease (CVD) risk. Vascular dysfunction, characterized by increased large elastic artery stiffness and endothelial dysfunction, predicts future CVD risk. However, if IVF use influences maternal vascular dysfunction is unknown. We studied premenopausal women with a live birth within the past 1–5 years conceived either without medical assistance (n=24) or via IVF (n=15), matched for age, body fat percentage, and seated blood pressure. Large artery stiffness (carotid β-stiffness index) and endothelial function (brachial artery flow-mediated dilation) were assessed after 10 minutes of supine rest following an overnight fast. Large artery stiffness was greater in the IVF group compared with the reference group (6.5±1.9 vs. 5.08±1.2 U, p=0.022); endothelial function did not differ (p=0.619). To explore potential mechanisms, mouse aortas were incubated with serum from a subset of participants (n=11/group) to assess the influence of circulating factors on arterial stiffness (via elastic modulus). The elastic modulus was 28% higher in the group treated with serum from IVF patients relative to the reference group (p=0.078), suggesting circulating factors may contribute to observed differences in large artery stiffness. Follicle-stimulating hormone (FSH) concentrations were greater in the IVF compared with reference group (p=0.026), and FSH concentrations were positively associated with in vivo large artery stiffness (r=0.57, p<0.001). Large artery stiffness is elevated in individuals with a history of IVF-conceived live birth, potentially related to circulating factors, including FSH.

New & Noteworthy.

Infertility is associated with increased risk of cardiovascular disease in women related to vascular dysfunction (i.e., increased large artery stiffness, impaired endothelial function). In vitro fertilization (IVF) is commonly used to treat infertility, but the effects of IVF on maternal vascular function remain unclear. This study demonstrates, for the first time, that large artery stiffness is elevated in women who conceived via IVF within the past 1–5 years, potentially due in part to circulating factors.

Graphical Abstract

INTRODUCTION.

Infertility is associated with an increased risk of future cardiovascular disease (CVD),(1, 2) the leading cause of mortality in women. Infertility is defined as the inability to achieve a successful pregnancy based on a patient’s medical, sexual, and reproductive history, age, physical findings, diagnostic testing, or any combination of factors.(3) In vitro fertilization (IVF) is the most widely used assisted reproductive technology for the treatment of infertility, contributing to more than 10 million births worldwide since 1978.(4) Causes of infertility include female factors such as tubal factors, endometriosis, diminished ovarian reserve and ovulatory dysfunction; male factors such as low sperm count, low motility or morphology or unexplained infertility in the absence of an identifiable medical cause. The IVF process involves controlled ovarian stimulation with exogenous gonadotropins and sex hormones to induce multifollicular development, followed by oocyte retrieval, in vitro fertilization and culture, and subsequent embryo transfer, which may occur in the same cycle (fresh transfer) or after cryopreservation and thawing in a later cycle (frozen transfer). Prior research on the cardiovascular consequences of IVF has primarily focused on offspring, with less attention given to potential maternal risk.(5–7) Moreover, information on the use of assisted reproductive technology, such as IVF, is often not collected or reported in studies evaluating the influence and mechanisms of adverse pregnancy outcomes on CVD risk.(8–11) This represents an important clinical knowledge gap for the growing number of individuals utilizing IVF for family planning.

Vascular dysfunction (i.e., large elastic artery stiffening and endothelial dysfunction) is a key antecedent in the pathogenesis of CVD.(12) Vascular dysfunction is increased with aging and the accumulation of CVD risk factors due in part to the excessive generation of oxidative stress and inflammation.(13) Oxidative stress, characterized as excessive reactive oxygen species (ROS) production relative to antioxidant defense capacity, is a key mechanism underlying the development of vascular dysfunction.(12) Endothelial dysfunction and large elastic artery stiffness both independently predict CVD risk.(14, 15) Chronic exposure to elevated oxidative stress and endothelial dysfunction also induces structural remodeling of the large elastic (e.g., aorta and common carotid) arteries to increase arterial stiffness.(16) Vascular dysfunction is increased in women with infertility,(17–20) however, relatively few studies have evaluated if the process of IVF influences vascular function. Recent data suggest that normal cardiovascular adaptations (e.g., increased cardiac output, decreased peripheral vascular resistance and blood pressure [BP]) in early pregnancy are blunted from conception to the 2^nd trimester of pregnancies conceived by IVF.(21–23) Notably, this effect was specific to pregnancies conceived from a frozen embryo transfer, whereas pregnancies conceived from a fresh embryo transfer did not differ compared with medically unassisted pregnancies.(23) While cardiovascular hemodynamics normalized in the 3^rd trimester and post-partum period in IVF pregnancies in these studies,(21, 24) population studies investigating long-term (>1 year) risk of CVD with prior IVF demonstrate conflicting findings of neutral or increased CVD risk.(25–27) Risk of stroke and new-onset hypertension are elevated 1-5 years after delivery, but not <1 year or >5 years, in pregnancies conceived via IVF(25) suggesting this postpartum period may be important for studying mechanisms of chronic disease risk following IVF. Collectively, these data suggest that IVF may alter early cardiovascular adaptations to pregnancy with a greater effect of frozen embryo transfers. However, whether vascular dysfunction persists in postpartum and influences maternal CVD risk remains unclear.

Therefore, the purpose of this study was to evaluate if vascular function differed in women with a history of live birth 1-5 years earlier conceived via IVF compared with a reference group, and to explore potential underlying mechanisms. We hypothesized that individuals with a history of a live birth conceived from IVF would exhibit elevated large elastic artery stiffness and endothelial dysfunction compared to the reference group matched by age, body composition and blood pressure. Two approaches were used to explore mechanisms underlying potential oxidative stress and inflammation: 1) mouse aortas were incubated in human participant serum obtained from a subset of participants in the study to assess the influence of circulating bioactive factors on vascular function; and 2) correlational analyses were performed between vascular function and blood biomarkers.

METHODS.

Participants.

Female participants aged 18-45 years with a history of a live birth in the past 1-5 years conceived either spontaneously without medical assistance (reference group) or via IVF were recruited from the Denver Colorado Metropolitan area to participate between the years 2021-2023. Participants were recruited via flyer advertisement, mass email, and via registration to recruitment registry through an electronic medical record. Individuals in the IVF group were also recruited via treatment relationship with Shady Grove Fertility Clinic (N.S., C.R., E.C). This study was approved by the Colorado Multiple Institutional Review Board, and procedures were conducted according to The Declaration of Helsinki. All participants provided written informed consent prior to participating. All study visits and measurements were performed at the Colorado Clinical and Translational Sciences Institute (CCTSI) Clinical and Translational Research Center (CTRC).

Participants were apparently healthy and free from cardiovascular, cancer, renal, liver, or respiratory disease as assessed by medical history and standard blood chemistries (e.g., complete blood count, comprehensive metabolic panel, thyroid stimulating hormone). Inclusion criteria included: glucose <126 mg/dL, resting seated BP <140/90 mmHg, and nonsmoking. Individuals using oral contraceptives containing a placebo pill and hormonal and non-hormonal intrauterine devices (IUD) were permitted to participate. Participants were characterized as premenopausal using menstrual cycle characteristics consistent with the Stages of Reproductive Aging Workshop(28) or via serum follicle stimulating hormone concentrations in consultation with a reproductive endocrinologist (N.S.) in individuals with amenorrhea due to a hormonal intrauterine device. Participants were excluded for current or history of major chronic diseases (e.g., CVD, cancer), current use of anti-hypertensive or statin therapy medications, highly physically active defined as self-reported engagement in vigorous intensity exercise >3 days/week or unwillingness to washout vitamin supplements or anti-inflammatory medications for at least 2 weeks prior to testing. Individuals in the control group were excluded for any prior use of assisted reproductive technology. Finally, participants in both groups were matched for age, total body fat and seated blood pressure.

Participant characteristics.

Clinical characteristics were collected as previously described as correlates of cardiovascular health.(29) Seated blood pressure was measured in the morning under fasted conditions with adequate hydration (no caffeine, water encouraged) and abstinence from exercise following ≥ 10 min of seated rest in triplicate in both arms using an oscillometric blood pressure cuff (Carescape V100, GE medical systems). The average of the higher arm is reported in participant characteristics. Dual x-ray absorptiometry (Hologic Horizon) was used to quantify total and trunk body fat and hip and spine bone mineral density. Physical activity was objectively measured for 24 h a day for seven consecutive days using the thigh-worn activPAL3 micro because of prior literature showing that physical activity influences cardiovascular health(30) and may be altered during the postpartum period.(31) Physical activity was analyzed by a single observer (V.N) as previously described.(32) Participants logged daily times they got into bed for the night and got out of bed in the morning and periods of non-wear. Event data were exported with PAL Technologies software (version 8, PAL Technology, Glasgow, Scotland). Waking, non-wear, and sleep periods were demarcated based on participant sleep logs and activPAL event data using a standardized protocol. The start of sleep was identified by sedentary bouts >30 min that occurred around the time of attempted sleep, as reported in the participant log. The end of sleep was determined by the first upright bout that occurred around the time reported wake time. Non-wear and sleep periods were removed from the quantification of waking activity. Sedentary time was assessed as total sedentary time in minutes/day and minutes/day accumulated in bouts of 30 min and 60 minsleep periods were removed from the quantification of waking activity. Sedentary time was assessed as total sedentary time in minutes/day and minutes/day accumulated in bouts of 30 min and 60 min. Physical activity was assessed using steps/day. Moderate-to-vigorous physical activity (PA) was estimated by stepping time with a step cadence ≥100 steps/min. Low intensity PA was estimated as upright or stepping time with a step cadence <100 steps/min. Time in sedentary time, low intensity PA, and moderate-to-vigorous PA was expressed as a percentage of waking wear time to account for variability in sleep duration. (33)

Surveys.

Self-reported reproductive health histories were collected using a standardized reproductive survey administered by trained observers (L.E.D., E.C.) in all participants to collect comprehensive information regarding pregnancy history, adverse outcomes of pregnancy (e.g., pregnancy induced hypertension), and use of assisted reproductive technology. All self-reported reproductive information was reviewed, and infertility diagnosis was confirmed by a board-certified reproductive endocrinologist (E.C., C.R.). Self-reported sleep quality (via Pittsburgh Sleep Quality Index [PSQI]),(34) depressive symptoms ((35)Center of Epidemiological Studies Depression Scale [CES-D]),(36) stress (Perceived Stress Scale)(37) were collected on the day of vascular testing as correlates of cardiovascular health that may be impacted by pregnancy and postpartum.(33, 35, 38)

Vascular testing.

Vascular testing was conducted in the supine position following an overnight fast with adequate hydration as previously described.(29) Participants with menses were scheduled during the early follicular phase (Days 1-6), when possible. Participants with amenorrhea due to hormonal IUD were scheduled at any time. Participants abstained from exercise and alcohol for ≥ 20 h prior to the study visit. Vascular testing was measured following 10 min of quiet rest.

Large elastic artery stiffness.

Carotid systolic and diastolic BP were derived from the carotid waveform and calibrated to the mean and diastolic pressure obtained from a brachial BP cuff and tonometry-acquired brachial BP waveform (Noninvasive hemodynamics [NIHem] workstation; Cardiovascular Engineering, Norwood, MA).(39) Large elastic artery stiffness was measured by common carotid artery stiffness using high-resolution carotid ultrasonography (GE Vivid I) with a 12-MHz linear transducer as described.(29) All images were coded by number, blinded to the study group, and analyzed by a single investigator (B.T.). Briefly, 15-s 2D carotid images were acquired 1–2 cm proximal to the carotid bulb and analyzed using semiautomated edge-detection software (Vascular Analysis Tools v. 5.5; MIA LLC, Coralville, IA) to calculate the maximum (systolic) and minimum (diastolic) diameters and IMT as previously described.(29) Carotid stiffness was measured as carotid β-stiffness index calculated from carotid ultrasonography and central blood pressure derived from applanation tonometry.(39) Carotid β-stiffness index was the primary outcome because it provides an index of arterial stiffness adjusted for distending pressure. Carotid compliance was calculated as described(40) as an alternative measure of arterial stiffness to confirm findings with the primary outcome. Carotid IMT was measured on the far wall during end diastole and averaged across cardiac cycles as a measure of wall thickness.

Macrovascular Endothelial Function.

Brachial artery FMD was measured as described(41) in response to 5 minutes of forearm cuff occlusion (250 mmHg) with duplex ultrasonography (GE Vivid I) using a multi-frequency linear-array transducer according to established guidelines. FMD was measured as a relative change (%). Brachial artery diameters were analyzed in end diastole using a commercially available software package (Vascular Analysis Tools 5.5.1; Medical Imaging Applications, Iowa City, IA) by a single trainer analyzer (B.T.). Data were allometric scaled to account for the influence of baseline diameter.(42)

Blood Sampling.

Blood sampling was performed in the morning under fasted conditions and analyzed by the CCTSI-CTRC Core laboratory as described.(29) Plasma glucose, insulin, total cholesterol (Roche Diagnostic Systems, Indianapolis, IN), high-density lipoprotein (HDL, Diagnostic Chemicals, Ltd. Oxford, CT) were measured using enzymatic/colorimetric methods as correlates of cardiovascular health. Low density cholesterol (LDL) was estimated using the Friedewald equation. Sex hormone concentrations were measured on the day of vascular testing as correlates of vascular function and to account for potential differences in sex hormone concentrations in participants with amenorrhea due to IUD use. Serum concentrations of estradiol, FSH, and progesterone were measured by chemiluminescence (Beckman Coulter). Estrone was measured by radioimmunoassay and total testosterone by a 1-step competitive assay (Beckman Coulter). Oxidized LDL (ELISA, Alpco Diagnostics, Windham, NH) and total antioxidant status (TAS, Randox Laboratories, Oceanside, CA) were measured as markers of oxidative stress and antioxidant levels, respectively. Plasma endothelin-1 and interleukin (IL)-6 (both via enzyme-linked immunoassay) and high-sensitivity C-reactive protein (immunoturbidimetric method) were also measured.

Serum Exposure.

Blood was collected from participants who consented to freezer (−80°F) storage of excess serum for future exploratory analyses. All available samples were subsequently thawed and pooled to match n=11 IVF participants with control participants matched by age, body fat, and seated blood pressure for a total sample of n=22. Aortas were isolated from young adult (4 months of age) female C57BL/6J mice, cleaned of perivascular adipose and connective tissue, and the thoracic region was sectioned into ~1 mm segments. Aorta segments were incubated in DMEM supplemented with 1% gentamycin and 10% human subject serum (total volume 200 μL) in a 96-well plate for 48 h under standard culture conditions, as we have previously described.(43, 44) Thoracic aorta segments from each mouse were incubated with serum from one IVF participant and serum from one reference participant in duplicate using the matching parameters described above. Following incubation, elastic modulus assays were performed using a multi-Myograph (pin myography) System (620M). For mechanical testing, aortic segments were mounted on micropositioner pins in an organ bath containing 5 mL of 1× PBS. Samples underwent a preconditioning phase in which pin displacement was increased by 1 mm and returned to baseline every 30 sec for a total of 3 min. After preconditioning, the Myograph was zeroed, and aortas were stretched until a force of 1 mN was achieved; the corresponding micropositioner displacement was recorded. Pin displacement was then increased stepwise by 0.05 mm, and force (mN) was recorded every 3 min. This was repeated until a >20% decrease in force was observed. Elastic modulus was calculated as the slope of the high-force region of the force-displacement curve using the four data points prior to the tearing of the aorta segment.

Statistical analysis.

Data analysis was performed with R software (v. 4.3.2) and GraphPad prism (v10.6). Data are reported as mean ± SD unless otherwise noted. Data were tested for normality using the Shapiro–Wilk normality test and log transformed when data were not normally distributed. Differences between the IVF and reference group were tested using an Independent T-test for all variables. The primary outcome was large elastic artery stiffness (via carotid β-stiffness index) and a secondary outcome of endothelial function (via brachial artery FMD). The exploratory outcome was aortic elastic modulus measured in mice. Exploratory correlational analyses were conducted using the Pearson product moment correlation coefficient. Biomarkers of oxidative stress were selected a priori to examine their associations with vascular function. Additionally, participant characteristics that showed group differences at p < 0.10 in independent-samples t-tests were identified a priori and included in exploratory correlational analyses to investigate potential underlying mechanisms. Statistical significance was set at p<0.05.

RESULTS.

Participant characteristics.

Participant characteristics for the entire cohort are displayed in Table 1. As expected, the groups did not differ by age, total body fat or seated blood pressure. Participants in the IVF group varied by underlying indication for IVF; however, approximately half experienced unexplained infertility. FSH was higher in the IVF group compared to the reference group (p=0.026), however, FSH was still within expected range for premenopausal women. Fasted cholesterol, triglycerides and blood glucose did not differ between groups (p>0.104). Interleukin-6 (p=0.077) and total antioxidant status (p=0.064) tended to be elevated in the IVF group; however, C-reactive protein and oxidized LDL did not differ between groups (p>0.792). Subjective sleep quality tended to be poorer in the IVF compared with reference group (p=0.097). The groups did not differ by any pregnancy metrics (p>0.418). Steps per day tended greater in reference group (p=0.093), however, no other metrics of physical activity differed between the groups (p>0.193).

Table 1.

Participant characteristics

Variable	Reference Group N=24	IVF N=15	P-value
Age (yrs)	37 ± 3	38 ± 4	0.347
Race n (%)
White	20 (83)	13 (87)	0.829
Black	1 (4)	1 (7)
Asian	1 (4)	1 (7)
American Indian/Alaskan Native	1 (4)	0 (0)
More than one	1 (4)	0 (0)
Ethnicity n (%)			0.688
Non-Hispanic	22 (92)	15 (100)
Hispanic	2 (8)	0 (0)
Infertility type n (%)	-		-
Female		4 (27)
Male		4 (27)
Combination		0 (0)
Unexplained		7 (47)
Frozen embryo transfer n (%)	-	13 (87)	-
Contraception Use n (%)	7 (29)	4 (27)	0.870
Oral Contraceptive	3 (13)	3 (20)
Hormonal IUD	4 (17)	1 (7)
Pregnancy-induced hypertension n (%)a	5 (29)	5 (33)	0.418
Parityb	1 (0.75, 2)	2 (1, 2)	0.518
Time postpartum (years)b	2.7 (1.5, 3.9)	2.3 (1.5, 3.8)	0.729
Body Mass Index (kg/m2)	24.9 (23.1, 27.8)	27.4 (25.6, 31.1)	0.252
Total body fat (%)	36.2 ± 6.4	39.4 ± 6.8	0.154
Trunk body fat (%)	33.5 ± 8.1	37.4 ± 7.4	0.132
Seated systolic BP (mmHg)	109 ± 8	107 ± 10	0.654
Seated diastolic BP (mmHg)	70 ± 6	68 ± 9	0.417
Total cholesterol (mg/dL)	169 ± 29	178 ± 30	0.368
LDL cholesterol (mg/dL)	99 ± 24	111 ± 22	0.145
HDL cholesterol (mg/dL)	50 (47- 59)	49 (45- 53)	0.104
Triglycerides (mg/dL)	74 (58- 98)	82 (66- 93)	0.562
Glucose (mg/dL)	85 ± 7	85 ± 6	0.935
C-Reactive Protein (mg/L)	1.02 (0.38- 2.89)	0.97 (0.53- 3.14)	0.792
Oxidized LDL (U/L)c	47.9 ± 15.3	47.9 ± 12.3	0.999
Interleukin-6 (pg/mL)	0.69 (0.44, 1.32)	1.36 (0.67- 1.79)	0.077
Total Antioxidant Status (mg/dL)	1.20 (1.18, 1.25)	1.27 (1.17, 1.47)	0.064
Estradiol (pg/mL)	66 (40, 139)	105 (49, 158)	0.476
Follicle Stimulating Hormone (mIU/mL)	4.6 (3.3, 6.2)	6.0 (4.5, 7.5)	0.026
Progesterone (ng/mL)	0.20 (0.10, 0.93)	0.40 (0.10, 1.25)	0.566
Total Testosterone (ng/dL)	32 ± 12	28 ± 11	0.320
AMH (ng/mL)c	2.96 (1.27, 3.96)	3.59 (1.31, 5.21)	0.758
PSQI (points)	3 (2, 5)	5 (4, 7)	0.097
Perceived Stress Survey (points)	12.2 ± 4.4	14.2 ± 6.3	0.296
Average steps per dayd	10119 (7527, 11526)	8724 (6071, 10765)	0.093
Average stepping minutesd	130.2 ± 31.8	114.2 ± 30.2	0.193
Average standing minutesd	257.8 ± 67.5	282.7 ± 63.4	0.334
Average sedentary minutesd	503.1 ± 81.0	521.6 ± 58.3	0.501
Total Spine Z-score e	0.00 ± 0.74	0.32 ± 1.05	0.359
Total Hip Z-score e	0.40 ± 0.82	0.38 ± 1.14	0.965

Values are mean ± SD. Non-normally distributed data are expressed as the median (interquartile range). AMH, anti-mullerian hormone; BP, blood pressure; HDL, high density lipoprotein; IUD, intrauterine device; LDL, low density lipoprotein; PSQI, Pittsburg Sleep Quality Index

^aSample size of 33

^bSample size of 31

^cSample size of 36

^dSample size of 27

^eSample size of 26

Cross-sectional differences in vascular function.

Group differences in vascular function for the entire cohort are displayed in Table 2 and Figure 1. Carotid β-stiffness index was greater (p=0.022) in the IVF compared with reference group (Figure 1). Carotid compliance was also lower in the IVF group (p=0.012, Table 2), consistent with the findings with Carotid β-stiffness index. Carotid end-diastolic diameter tended to be greater in the reference compared with IVF group (p=0.051), however, metrics of central blood pressure did not differ between groups (p>0.196). Carotid intima-media thickness was also similar between groups (p=0.592). Endothelial function did not differ between groups when expressed as an absolute or relative change (p>0.619) nor when data were allometrically scaled (p=0.340).

Table 2.

Group differences in vascular function

Variable	Reference Group	IVF	p-value
Central systolic BP (mmHg)	103 ± 8	100 ± 14	0.398
Central diastolic BP (mmHg)	68 (64, 71)	62 (57, 68)	0.196
Central pulse pressure (mmHg)	36 ± 6	37 ± 7	0.932
Brachial systolic BP (mmHg)	107 ± 7	104 ± 12	0.408
Brachial diastolic BP (mmHg)	68 ± 7	64 ± 11	0.188
Carotid compliance (mm/mmHg)	0.129 ± 0.036	0.106 ± 0.038	0.012
Carotid end diastolic diameter (mm)	6.14 ± 0.35	5.89 ± 0.40	0.051
Carotid distension (mm)	0.39 ± 0.21	0.41 ± 0.14	0.689
Carotid intimal-media thickness (mm)	0.55 ± 0.07	0.56 ± 0.04	0.592
Brachial diameter (mm)	3.26 ± 0.43	3.47 ± 0.36	0.129
Absolute flow-mediated dilation (mm)	0.268 ± 0.134	0.264 ± 0.122	0.926
Relative flow-mediated dilation (%)	8.4 ± 4.5	7.7 ± 3.73	0.619

Values are mean ± SD. Non-normally distributed data are expressed as the median (interquartile range). BP, blood pressure.

Figure 1.

Group differences in large artery stiffness measured in vivo via carotid β-stiffness index using carotid ultrasonography in women with a prior pregnancy conceived via in vitro fertilization (IVF, n=14) compared with a reference population (n=23). Groups were matched for age, body composition and blood pressure. Higher values indicate worse vascular function. Group differences were tested using independent T-tests. Large artery stiffness was elevated in the IVF compared with reference group. Figures were created using GraphPad Prism.

Serum-mediated large artery stiffness.

Participant characteristics and vascular data in the subset enrolled in the exploratory study are provided in Supplemental Table 1 and 2. The elastic modulus was 28% higher (p=0.078) in the IVF group relative to exposure to serum from control participants following incubation with human serum (Figure 2). Large artery stiffness measured in vivo tended to be associated with the observed values ex vivo (r=0.41, p=0.056; Supplemental Figure) in the entire cohort.

Figure 2.

Group differences in large artery stiffness measured ex vivo via aortic elastic modulus performed in a subset of participants matched for age, body composition and blood pressure. Aortas from female rodents were incubated in serum from the in vitro fertilization (IVF, n=11) or reference group (n=11). Higher values indicate worse vascular function. Group differences were tested using independent T-tests. Large artery stiffness was elevated in the IVF compared with reference group. Figures were created using GraphPad Prism.

Correlations between large artery stiffness and blood biomarkers were performed with biomarkers of oxidative stress and inflammation as well as using variables with a group difference of p<0.10. Large artery stiffness was not correlated with biomarkers of oxidative stress and inflammation (p>0.765). However, greater FSH concentrations were associated with large artery stiffness (r=0.57, p<0.001) measure in vivo in the entire cohort (Figure 3). The association between FSH concentrations and large artery stiffness remained significant after covarying for chronological age (r=0.50, p=0.002). The association between FSH and large artery stiffness was stronger in the IVF group (stiffness: r=0.70, p=0.005) compared with the reference group (stiffness: r=0.36, p=0.080). FSH was not associated with any indices of central blood pressure or large artery stiffness measured ex vivo (p>0.549). Finally, large artery stiffness was not correlated with variables PSQI or steps per day (p>0.533).

Figure 3.

Scatterplots displaying the relation between follicle stimulating hormone (FSH) and large artery stiffness via carotid β-stiffness index measured in vivo using carotid ultrasonography. Higher carotid β-stiffness index values indicate worse vascular function. Open circles represent the control group with black circles representing the IVF group. Correlations between variables were performed using Pearson Product-Moment Correlations and adjusted for age using Partial Bivariate Correlations. Higher FSH was associated with greater large elastic artery stiffness. Figures were created using GraphPad Prism.

DISCUSSION.

This study evaluated the influence of IVF on vascular function and potential underlying mechanisms. Novel findings include the observation that large elastic artery stiffness was greater in women who conceived via IVF within the prior 1–5 years compared with the reference group, despite no group differences in traditional and non-traditional CVD risk factors. Additionally, large elastic artery stiffness measured ex vivo was greater in female mouse aortas exposed to serum from the IVF group compared with the reference group, consistent with the in vivo findings, and suggesting that factors in the circulation are partially responsible for observed group differences in large elastic artery stiffness. However, contrary to our hypothesis, circulating biomarkers of oxidative stress and inflammation did not explain the observed differences in vascular function. Instead, greater FSH concentrations were correlated with higher large elastic artery stiffness. This observation was independent of how patients conceived, but more pronounced in participants with a prior IVF pregnancy. Collectively, these data provide the first evidence that large artery stiffness may differ in women 1-5 years following pregnancies conceived via IVF.

Influence of IVF on Large Elastic Artery Stiffness and Mechanisms

Large elastic artery stiffness is an independent predictor of CVD event risk,(14) with stiffening of the common carotid arteries predicting stroke risk independent of traditional CVD risk factors and aortic stiffness.(45) Moreover, the common carotid arteries stiffen prior to the aorta with aging,(46) supporting the idea that common carotid artery stiffness may be a more sensitive predictor of the onset of elevated CVD risk. In the present study, large artery stiffness was ~1.4U higher in the IVF compared with the reference group, with prior studies suggesting that 1U difference in large elastic artery stiffness is associated with elevated CVD risk.(45)

To evaluate underlying mechanisms, we performed correlational analyses between factors that differed between the groups and an exploratory serum exposure tissue culture model to explore if differences observed in large elastic artery stiffness between groups were related to circulating factors in the blood or chronic, structural remodeling. Prior research has shown that serum from postmenopausal women increases the stiffness of large central arteries from young adult mice ex vivo.(43) Here, we extend these findings to show that elastic artery stiffness measured ex vivo was 28% greater in aorta rings exposed to IVF serum compared with reference serum, suggesting that group differences in large elastic artery stiffness in vivo are related, in part, to factors in the circulating milieu or bioactive factors in the blood. Notably, the observed difference in serum-evoked large elastic artery stiffness occurred in the absence of differences in traditional blood-based CVD-associated risk factors, further supporting the idea that other factors in the circulating milieu may be responsible for transducing our observed phenotypes.

FSH concentrations were greater in the IVF group relative to the reference group at baseline. Observed FSH concentrations were within normal physiological limits for premenopausal women and mean AMH concentrations, a marker for oocyte quantity, were above the clinical threshold for diminished ovarian reserve.(47) Higher FSH concentrations were related to greater large elastic artery stiffness in the entire cohort; however, the correlation was stronger in the IVF group compared with the reference group. Neither estradiol nor progesterone concentrations were associated with large artery stiffness in the present study (data not shown). The observed findings align with prior literature demonstrating that FSH concentrations are more strongly associated with vascular function compared with other sex hormones in cross-sectional cohorts of healthy women(48–50) and have been implicated in CVD in a preclinical mouse model.(51) The cellular mechanisms underlying this remains unknown, however, prior studies support the idea that FSH influences cardiovascular regulation in females. Higher FSH concentrations were associated with endothelial function and shear hemodynamics in the brachial artery,(52) potentially related to greater sympathetic outflow, in estrogen-deficient postmenopausal women.(53, 54) However, in the current study, no group differences were observed in endothelial function in premenopausal women suggesting that this mechanism does not contribute. Alternatively, other studies suggest that sex hormone concentrations may be subtly altered depending on the menstrual cycle phase in patients with unexplained infertility(55) with early follicular FSH concentrations being elevated in patients with unexplained infertility compared with those without infertility.(55) In the present study, unexplained infertility was the most common indication for IVF. Vascular testing was conducted during the early follicular phase of the menstrual cycle, when possible, however, menstrual cycle phase was not controlled for in individuals with amenorrhea due to the use of hormonal IUDs. Notably, the groups did not differ by the percentage of participants using hormonal IUDs. Standardization of vascular testing to the menstrual cycle is controversial in vascular physiological research.(56, 57) Nevertheless, we cannot determine whether the group differences in FSH concentrations are due to normal fluctuations in sex hormone concentrations across the menstrual cycle, differences in menstrual cycle phase of participants at the time of vascular testing or underlying indication for infertility (e.g., unexplained infertility). Thus, future investigations are warranted to determine whether elevated circulating concentrations of FSH are directly implicated in higher large elastic artery stiffness following IVF pregnancies.

Influence of IVF on Endothelial Function.

Declines in endothelial function are a key antecedent in the pathogenesis of CVD.(58) Endothelial dysfunction contributes to arterial stiffening although this physiological relationship is likely bidirectional.(16) In the present study, endothelial function did not differ between groups, consistent with other studies.(59, 60) Macrovascular endothelial function was not altered following a long agonist IVF protocol to pharmacologically manipulate sex hormone concentrations (via gonadotropin releasing hormone agonist) for ~25 days;(59) however, this study did not adhere to established guidelines for the assessment of brachial artery flow-mediated dilation.(61) In two separate studies, the use of IVF did not influence endothelial function (via EndoPat) 6-12 weeks after delivery nor ~11 months postpartum, although these studies demonstrated divergent findings on group differences in endothelial function during pregnancy.(60, 62) This limited evidence suggests that endothelial function is not altered by the IVF process. The mechanisms underlying this endothelial resilience are unclear but may be related to the vascular protective effects of circulating estradiol(63) and/or due to compensatory changes in circulating factors (e.g., antioxidants) or endothelial cell protein expression. Prior work by our group supports the idea that premenopausal women may be protected from oxidative damage because of the vascular protective effects of circulating estradiol.(64) Alternatively, total antioxidant status and IL-6 tended to be elevated in the IVF group in the present study. However, group differences in these outcomes didn’t meet statistical significance so other mechanisms likely contribute. Collectively, our data and others indicate that endothelial dysfunction is not impaired in pregnancies conceived through IVF and is not the primary mechanism contributing to the elevated large artery stiffness observed in the present study.

Strengths and limitations.

Strengths of this study include the novel evaluation of vascular function in healthy premenopausal women using reference-standard techniques. Additionally, this cohort was well characterized for traditional, non-traditional, and emerging CVD risk factors to help interrogate underlying mechanisms of observed group differences in vascular function. Limitations include, first, the cross-sectional study design prevents inferring causality. Given the cross-sectional study design and focus on the postpartum period in the current study, we cannot evaluate if cardiovascular health differed by group prior to conception. Two prior studies evaluated if IVF altered maternal cardiovascular adaptations to pregnancy compared to pregnancies conceived without medical assistance.(22, 23) Cardiovascular outcomes were studied at multiple time points including pre-pregnancy and peripartum. Cardiovascular outcomes (i.e., cardiac output, left atrial dimension, mitral E’, pulse pressure amplification, augmentation index, global compliance, carotid femoral pulse wave velocity) did not statistically differ between groups prior to conception in these studies. Second, this study was not powered to evaluate group differences due to IVF protocol (fresh vs. frozen) or indication for IVF (e.g., female, male, combination or unexplained infertility, etc.,). Hypertensive disorders of pregnancy are a well-known female-specific CVD risk factor(65) and the risk of hypertensive disorders of pregnancy is elevated in IVF pregnancies.(66) We were not powered to study if vascular function was influenced by the independent or combined effects of hypertensive disorders of pregnancy with/without IVF. However, the two groups were matched for pregnancies complicated by hypertensive disorders of pregnancy, minimizing the potential confounding. Finally, infertility and fertility treatment can negatively impact psychosocial factors including mood, stress and depressive symptoms(67, 68) which may also impact cardiovascular health.(33, 35) While the present study measured self-reported stress and mood around the time of vascular testing, this likely does not reflect the lived experience of a patient experiencing infertility and/or undergoing fertility treatment. Therefore, we cannot evaluate if psychosocial stress associated with infertility or infertility treatment mediates the observed effects of IVF in the present study.

CONCLUSION.

Large elastic artery stiffness, but not endothelial dysfunction, is elevated in people with a history of a live birth conceived via IVF related to factors circulating in the blood, including FSH. Additional research is needed to evaluate the mechanisms underlying the observed group differences, including the regulatory actions of FSH on cardiovascular health in women. Moreover, this study underscores the importance of considering and collecting information on methods of conception (e.g., unassisted, IVF, etc.) in research focused on maternal cardiovascular health.

Supplementary Material

Supplemental Tables and Figure are available here: https://doi.org/10.6084/m9.figshare.32337792

Data availability.

Data will be made available upon reasonable request.