STRUCTURED ABSTRACT
INTRODUCTION:
Whether history of adverse pregnancy outcomes (APOs) contributes to cognitive decline in women is unclear.
METHODS:
Among parous female Adult Changes in Thought (ACT) participants (aged ≥65 years without dementia at enrollment), we abstracted APO history for those born 1940 or later and enrolled between 2005 – 2020 (n=444). Generalized estimating equations estimated the association between APO history and cognition score, measured using the Cognitive Abilities Screening Instrument (CASI-IRT), and cognitive decline.
RESULTS:
Among all participants, 13% had any history of APO. In adjusted models, women with an APO history had 0.23-point lower CASI-IRT score at any age (95% CI: −0.54, 0.07); this small difference was not statistically significant. We found no evidence of an association between APO history and 4-year cognitive change.
DISCUSSION:
Among women in the ACT study born in 1940 or later, we found no evidence of a significant association between history of APO and lower cognition or cognitive decline in older adulthood.
Keywords: Cognition, Cognitive decline, Pregnancy, Maternal, Hypertensive disorders of pregnancy, Preeclampsia, Preterm birth, Adverse pregnancy outcomes
1. BACKGROUND
Adverse pregnancy outcomes (APOs) such as hypertensive disorders of pregnancy, preterm birth, small-for-gestational-age infant, and stillbirth, affect 1 in 5 pregnancies in the United States1. Histories of some APOs were previously found to be associated with the development of Alzheimer’s disease and related dementias (ADRD) in later life.2–6 A systematic review and meta-analysis found that compared with women with no APO history, women with a history of any APO had moderately higher risk of all-cause dementia and Alzheimer’s disease, and nearly double the risk of vascular dementia.7
One potential pathway by which APOs may contribute to dementia risk, especially vascular dementia, is the long-term higher risk of maternal cardiovascular and cerebrovascular disease in those with APOs8–11. Compelling data suggest shared risk factors for APOs and cognitive impairment, and potential causal pathways between APOs and cognitive impairment.12 Several APOs, including hypertensive disorders of pregnancy, preterm birth, and fetal growth restriction, have a prominent vascular component to their pathophysiology, involving angiogenic imbalance, oxidative stress and systemic endothelial inflammation.13 These same processes have also been associated with both vascular cognitive impairment14,15 and Alzheimer’s disease.16 A history of APOs is strongly linked in epidemiological studies to higher risk and earlier onset of cerebrovascular disease,9,10,17,18 a major contributor to cognitive impairment and dementia.19 Other potential linking pathways between APOs and ADRD may include shared risk factors, including hypertension, obesity, and social influences on health (e.g., financial stress, exposure to systemic racism). In addition, limited data support a possible direct effect of certain APOs, such as preeclampsia, on maternal neurodegeneration and pathological brain aging.12
A major barrier to understanding these associations is the lack of detailed pregnancy outcomes data in prospective cohorts of older adults with longitudinal measures of cognition. Multiple studies have identified higher risk of cognitive impairment and dementia in women with a history of preeclampsia4,6 and other hypertensive disorders of pregnancy.3,5 One study using Swedish registry data found higher risk of vascular dementia in women with a history of APOs including hypertensive disorders of pregnancy, preterm birth, and fetal growth restriction.2 In contrast, the Bogalusa Heart Study, a prospective cohort in a semirural US population of Black and white women, found no association between midlife cognition and self-reported history of APOs.20 Studies using prospectively collected APO information have more consistently found relationships between APOs and cognitive impairment, compared to those using only recalled APOs. However, these studies have focused on hypertensive disorders of pregnancy;3–6,20 the role of other APOs, such as preterm birth, on long-term maternal cognition is unclear. While some studies have shown a relationship between APOs and cognition in older adulthood,5 additional studies are needed to contribute to the understanding of pathways from pregnancy to ADRD risk among those who survive into older adulthood.
In this study, set within the ongoing, prospective Adult Changes in Thought (ACT) cohort study of adults aged 65 years and older, we aimed to evaluate associations between history of APOs and later life cognition function and change in cognitive function over time. We hypothesized that on average, those who experienced APOs would have a lower global cognition score and faster cognitive decline than those with no history of APOs.
2. METHODS
2.1. Study Setting and Design
The ACT study is an ongoing prospective cohort study of adults aged 65 years and older, set within Kaiser Permanente Washington (KPWA) (originally Group Health Cooperative [GHC]), an integrated healthcare delivery system in Washington State. Since 1994, ACT has recruited 5,763 male and female participants, with data available for this analysis until March 2020 (average 7.7 years of follow-up).
ACT study design and methods have been previously described.21 Briefly, at study enrollment, eligible participants were dementia-free adults aged 65 years old or older who had been community-dwelling members of the KPWA integrated healthcare system for at least two years. At baseline and every two years thereafter, ACT participants were invited to complete a detailed survey in the clinic or at home, including the Cognitive Abilities Screening Instrument (CASI), a 40-item instrument that screens for dementia. Participants scoring less than 86 out of 100 were referred to further testing and dementia diagnosis conferred by a multidisciplinary conference of physicians and neuropsychologists; following dementia diagnosis, the CASI was no longer completed.22
2.2. Study Population
From all female ACT participants (n=3,343), we excluded participants born before 1940 (n=2,649) due to limited pregnancy records for this birth cohort. We further excluded participants without any pregnancy information (Supplemental Figure 1), resulting in 444 women eligible for this analysis. This study protocol was reviewed and approved by the institutional review board of KPWA. Participants provided written informed consent.
2.3. Assessment of Reproductive History
A team of three trained abstractors reviewed KPWA paper and electronic health records to abstract detailed reproductive histories for each participant. Information for care received pre-2005 was collected primarily from paper records, and information for care received 2005 and later was collected primarily from electronic health records. A condensed version of the chart abstraction protocol, as well as assessment of inter-rater reliability, is included in the Supplemental Methods, Appendix 1.
2.4. History of Any Adverse Pregnancy Outcomes
We defined a history of any APOs as ever having a pregnancy complicated by any of the following conditions: preterm birth (defined as spontaneous or medically indicated birth prior to 37 weeks gestation), hypertensive disorders of pregnancy (including gestational hypertension, preeclampsia, eclampsia), low birth weight (< 2500 g), small-for-gestational age infant (defined as birth weight <10th percentile for gestational age at delivery, not stratified by sex),23 and stillbirth.
2.5. Assessment of Cognition and Cognitive Decline
Cognition was measured biennially using the CASI (range 0 to 100, with higher scores indicating better cognitive function), a composite score of nine separate subscores: abstraction and judgment, attention, construction praxis, category fluency, language ability, long-term memory, mental concentration, orientation, and short-term memory.24 All participants eligible for our study had at least one complete and valid CASI assessment. Scores calculated from the CASI assessment using item response theory (CASI-IRT)25 were used as the primary outcome. Cognitive subdomain scores for Memory, Executive function, Language, and Visuospatial (MELV) were used as exploratory outcomes (see Supplemental Methods, Appendix 2).
2.6. Covariates and Other Measures
Birth date was used to calculate age at each ACT visit and birth cohort (World War II [WWII] cohort [1940–1945] vs. baby boom cohort [1946–1952]). Additional covariates included self-reported marital status at ACT enrollment and education level at ACT enrollment. Due to clustering of education levels around completion of bachelors and postgraduate degrees in this population, we dichotomized education by completion of a bachelors (less or more).
Other measures of interest included in descriptions of participant characteristics were self-reported at ACT baseline: race and ethnicity; smoking status; history of hypertension, cerebrovascular disease, heart disease, diabetes, and current medication use. We included a binary indicator for whether the respondent had at least one APOE-e4 allele versus none.
2.7. Statistical Analysis
For the primary analysis, we fit linear regression models using generalized estimating equations26,27 with all available global CASI-IRT scores as the outcomes.28,29 All models used an independence working correlation to account for correlation between CASI scores from the same participant; robust (sandwich) standard errors were reported. To estimate adjusted mean differences in CASI-IRT scores at any age 65 and older between individuals with versus without history of APOs, we fit an age-adjusted linear regression model with APO history (any vs. none) and current age and reported the coefficient for APO history. To estimate adjusted mean differences in 4-year rates of cognitive change between those with and without APO histories, we included an interaction between current age and APO history and reported the coefficient for the interaction between age and APO history (scaled to four years) as well as the adjusted mean 4-year change in CASI-IRT score among those with and without a history of APOs. In fully adjusted models, we added birth cohort as a possible confounder, and education level and marital status as possible precision variables, to age-adjusted models. Exploratory analyses were performed for each individual APO; detailed methods are included in the Supplemental Methods, Appendix 3. All analyses were complete-case. R version 4.2.330 was used for all analyses.
3. RESULTS
3.1. Characteristics of the Study Population
Among 444 eligible women, 13.3% had a history of any APOs (Table 1). Women had a median age of 68 at ACT enrollment, and a median follow-up time of 4 years post-enrollment. Women with a history of any APOs were slightly older at first and last birth and more likely to be born during the baby boom era than women without any APOs. Distributions of years of follow-up in this analysis were similar in women with no APOs (median 4; IQR= 2,8) and women with APOs (median = 4; IQR = 2, 6). Participants were age 65 – 79 during study enrollment and follow-up. A total of 11 participants were diagnosed with dementia during follow-up.
Table 1:
Characteristics of Eligible ACT Participants, by APO history.
| Overall | Any APOs | No APOs | |
|---|---|---|---|
| N=59 (13.3%) | N=385 (86.7%) | ||
| Age at first birth, mean (SD)* | 26.1 (5.00) | 27.9 (5.36) | 25.7 (4.87) |
| Age at last birth, mean (SD)* | 30.6 (5.13) | 33.3 (4.22) | 30.0 (5.14) |
| Live births, median (Q1,Q3) | 2 (1, 2) | 2 (1.5, 3) | 2 (1, 2) |
| Age at menopause, median (Q1,Q3)* | 51.0 (47.0, 53.0) | 49.0 (45.0, 52.0) | 51.0 (47.0, 53.0) |
| Years of follow-up, median (Q1,Q3) | 4.0 (2.0, 8.0) | 4.0 (2.0, 6.0) | 4.0 (2.0,8.0) |
| Birth cohort (N, %) | |||
| WWII (1940–1945) | 296 (66.7%) | 34 (57.6%) | 262 (68.1%) |
| Baby boom (1946–1952) | 148 (33.3%) | 25 (42.4%) | 123 (31.9%) |
| Race and ethnicity (N, %) | |||
| Asian | 20 (4.5%) | 17 (4.4%) | |
| Hispanic | 12 (2.7%) | 9 (2.3%) | |
| Non-Hispanic Black | 14 (3.2%) | 9 (2.3%) | |
| Non-Hispanic White | 367 (82.7%) | 44 (74.6%) | 323 (83.9%) |
| Other | 31 (7.0%) | 27 (7.0%) | |
| Highest education level (N, %) | |||
| High School degree or less | 92 (20.7%) | 15 (25.4%) | 77 (20.0%) |
| Associate’s degree or some college | 38 (8.6%) | 34 (8.8%) | |
| Bachelor’s degree | 170 (38.3%) | 20 (33.9%) | 150 (39.0%) |
| Post-secondary degree | 144 (32.4%) | 20 (33.9%) | 124 (32.2%) |
| Dementia diagnosis at end of ACT follow-up (N, %) | 11 (2.5%) | 10 (2.6%) | |
| Characteristics at ACT enrollment | |||
| Age, median (Q1,Q3) | 68.0 (67.0, 70.0) | 68.0 (66.0, 70.0) | 68.0 (67.0, 71.0) |
| Marital status (N, %) | |||
| Married/Partnered | 280 (63.1%) | 40 (67.8%) | 240 (62.3%) |
| All other marital statuses | 164 (36.9%) | 19 (32.2%) | 145 (37.7%) |
| Body mass index, kg/m2, median (Q1, Q3)* | 26.5 (23.4, 30.8) | 26.9 (23.1, 30.6) | 26.5 (23.5, 30.9) |
| Smoking status (N, %) | |||
| Never | 253 (57.0%) | 33 (55.9%) | 220 (57.1%) |
| Past | 171 (38.5%) | 21 (35.6%) | 150 (39.0%) |
| Current | 18 (4.1%) | 14 (3.6%) | |
| Current self-reported medication use (N, %) | |||
| Aspirin | 154 (34.7%) | 21 (35.6%) | 133 (34.5%) |
| Other antithrombotics | 7 (1.6%) | 5 (1.3%) | |
| Antihypertensives* | 165 (37.2%) | 23 (39.0%) | 142 (36.9%) |
| Cholesterol-lowering* | 107 (24.1%) | 14 (23.7%) | 93 (24.2%) |
| Systolic blood pressure, mmHg, mean (SD)* | 136 (18.7) | 136 (20.1) | 136 (18.5) |
| Diastolic blood pressure, mmHg, mean (SD)* | 72.8 (9.77) | 72.1 (8.48) | 72.9 (9.96) |
| Self-reported history of comorbidities (N, %) | |||
| Hypertension* | 167 (38%) | 25 (42%) | 142 (37%) |
| Cerebrovascular disease○* | 14 (3%) | 10 (3%) | |
| Heart disease○* | 22 (5%) | 20 (5%) | |
| Diabetes* | 36 (8%) | 33 (9%) | |
| ≥ 1 APOE-e4 allele (N, %)* | 96 (21.6%) | 13 (22.0%) | 83 (21.6%) |
| Adverse pregnancy outcomes (N, %) * | |||
| Hypertensive disorder of pregnancy**† | 13 (2.9%) | 13 (22.0%) | - |
| Low birthweight (< 2500g)** | 14 (3.2%) | 14 (23.7%) | - |
| Preterm birth** | 15 (3.4%) | 15 (25.4%) | - |
| Small-for-gestational-age infant** | 28 (6.3%) | 28 (47.5%) | - |
| Stillbirth** | 10 (2.3%) | 10 (16.9%) | - |
Adverse pregnancy outcomes were not mutually exclusive; thus, percentages add up to >100%.
% missing of overall sample, by characteristic: age at first birth (12.6%); age at last birth (33.6%); age at menopause (24.1%); body mass index (3.4%); antihypertensives (4.7%); cholesterol lowering medication (4.7%); systolic blood pressure (2.0%); diastolic blood pressure (2.0%); cerebrovascular disease (1%); heart disease (1%); hypertension (1%); diabetes (28%); APOE genotype (34.2%); hypertensive disorder of pregnancy (9.7%); preterm birth (33.8%); low birthweight (57.4%); small-for-gestational age infant (63.3%).
Hypertensive disorders of pregnancy included gestational hypertension, preeclampsia, and eclampsia.
Cerebrovascular disease included stroke, transient ischemic attack, carotid endarterectomy. Heart disease included myocardial infarction, angina, coronary artery bypass graft surgery, and angioplasty.
ACT: Adult Changes in Thought study; APO: adverse pregnancy outcome; APOE: apolipoprotein E; BMI: body mass index; Q: quartile; SD: standard deviation; WWII: World War II. Note: Cells with fewer than five individuals have been suppressed to protect participant privacy, per ACT study policy.
3.2. Associations of cognitive outcomes with any APO history.
At any age ≥ 65 years, women with a history of APOs were estimated to have CASI-IRT scores approximately a quarter of a point lower than those with no APO history who were otherwise similar with respect to birth cohort, education, and marital status (adjusted mean difference: −0.23, 95% CI: −0.54, 0.07); this difference did not reach statistical significance (p > 0.05) (Table 2; Figure 1). Among women with and without an APO history, estimates for the adjusted average 4-year rate of change in cognition were near zero (0.02 point per 4 years [95% CI −0.05, 0.09] and 0.01 point per 4 years [95% CI −0.21, 0.23], respectively; Table 3), with no significant difference in those rates (adjusted average difference: −0.01 point per 4 years [95% CI: −0.25,0.23], p > 0.05; Table 4; Figure 2).
Table 2:
Adjusted mean differences in CASI-IRT score between women with and without adverse pregnancy outcomes
| Age-adjusted modelc | Fully-adjusted modelc | ||
|---|---|---|---|
| Eligible sample (% with APO)a | Adjusted mean differenceb in CASI-IRT score (95% CI) | ||
| Any APO | 444 (13.3%) | −0.25 (−0.54, 0.05) | −0.23 (−0.54, 0.07) |
| Hypertensive disorder of pregnancy | 398 (3.3%) | −0.04 (−0.72, 0.63) | −0.09 (−0.71, 0.54) |
| Low birthweight | 399 (3.5%) | −0.60 (−1.01, −0.18) | −0.48 (−0.98, 0.02) |
| Preterm birth | 400 (3.8%) | −0.57 (−1.16, 0.01) | −0.55 (−1.11, 0.00) |
| Small-for-gestational-age infant | 413 (6.8%) | −0.40 (−0.86, 0.07) | −0.34 (−0.84, 0.16) |
| Stillbirth | 395 (2.5%) | −0.04 (−0.50, 0.43) | 0.05 (−0.50, 0.60) |
Analytic samples included women with either (1) history of the listed APO or (2) no history of APOs.
Relative to women with no APOs.
Age-adjusted model includes only age at cognitive assessment. Fully-adjusted model includes age at cognitive assessment, birth cohort, education level, and marital status.
ACT: Adult Changes in Thought study. APO: adverse pregnancy outcome. CI: confidence interval.
Figure 1:
Adjusted mean differences in CASI-IRT scores (95% CI) between women with a history of the specified APO in any pregnancy compared to women with no APO history of the same age, birth cohort, education level, and marital status.
Table 3.
Adjusted mean 4-year change in CASI-IRT among women with and without APOs.c
| Eligible sample (% with APO)a | Adjusted mean 4-year change in CASI-IRT with no APOs (95% CI)b | Adjusted mean 4-year change in CASI-IRT with APO (95% CI)bb | |
|---|---|---|---|
| Any APO | 444 (13.3%) | 0.02 (−0.05,0.09) | 0.01 (−0.21,0.23) |
| Hypertensive disorder of pregnancy | 401 (3.2%) | −0.01 (−0.08,0.07) | −0.09 (−0.66,0.48) |
| Low birthweight | 189 (7.4%) | 0.07 (−0.04,0.17) | 0.09 (−0.34,0.52) |
| Preterm birth | 294 (5.1%) | 0.03 (−0.06,0.13) | 0.00 (−0.38,0.39) |
| Small for gestational age | 163 (17.18%) | 0.04 (−0.08,0.16) | 0.02 (−0.37,0.41) |
| Stillbirth | 444 (2.3%) | 0.03 (−0.04,0.09) | −0.19 (−0.55,0.16) |
Analytic samples included women with either (1) history of the listed APO or (2) no history of APOs.
Expected change in CASI-IRT calculated for the “average” woman in the study sample with respect to age, birth cohort, education level, and marital status.
Inference on difference in rates in Table 4.
APO: adverse pregnancy outcome. CI: confidence interval
Table 4.
Adjusted average differences in 4-year rate of change in CASI-IRT score between women with and without adverse pregnancy outcomes
| Fully-adjusted modelc | |||
|---|---|---|---|
| Eligible sample (% with APO)a | Adjusted average difference in 4-year rate of change (95% CI)b | ||
| Any APO | 444 (13.3%) | −0.01 (−0.25,0.23) | −0.01 (−0.25,0.23) |
| Hypertensive disorder of pregnancy | 398 (3.3%) | −0.02 (−0.57,0.53) | −0.08 (−0.67,0.50) |
| Low birthweight | 399 (3.5%) | 0.04 (−0.40,0.48) | 0.02 (−0.42,0.47) |
| Preterm birth | 400 (3.8%) | 0.01 (−0.37,0.39) | −0.03 (−0.43,0.37) |
| Small-for-gestational-age infant | 413 (6.8%) | −0.03 (−0.44,0.38) | −0.02 (−0.44,0.39) |
| Stillbirth | 395 (2.5%) | −0.13 (−0.51,0.24) | −0.22 (−0.59,0.15) |
Analytic samples included women with either (1) history of the listed APO or (2) no history of APOs.
Relative to women with no APOs
Age-adjusted model includes only age at cognitive assessment and interaction between age and history of APOs. Fully adjusted model includes age at cognitive assessment, birth cohort, education level, marital status, and interaction between age and history of APOs.
ACT: Adult Changes in Thought study. APO: adverse pregnancy outcome. CI: confidence interval.
Figure 2:
Adjusted mean differences in 4-year change in CASI-IRT scores (95% CI) between women with a history of the specified APO in any pregnancy compared to women with no APO history of the same age, birth cohort, education level, and marital status.
3.3. Associations of cognitive outcomes with individual APO histories.
For all individual APOs other than stillbirth, cognition scores were estimated to be lower among women with a history of the individual APO compared to those without any history of APOs, adjusting for age, birth cohort, marital status, and education level; however, no differences were statistically significant at the pre-defined level (Table 2; Figure 1). Of individual APOs, the largest point estimates were for preterm birth, experienced by 15 women (adjusted mean difference: −0.55 points, [95% CI: −1.11, 0.00]), and low birthweight, experienced by 14 women (expected difference: −0.48 points, [95% CI: −0.98, 0.02]). Adjusted average 4-year rates of change in CASI-IRT scores were similar between women with each individual APO and without a history of any APO (Table 4; Figure 2); the former estimates had wide confidence intervals (Table 3) and no differences in rates of change were statistically significant.
3.4. Exploratory outcomes: cognitive subdomains.
There were no significant differences in adjusted average MELV subdomain scores (Supplemental Table 1) or adjusted average 4-year rates of change in MELV subdomain scores (Supplemental Table 2) between women with a history of any APO or any individual APO and women with no APO history, after consideration of multiple comparisons (using a nominal p-value of 0.0125 for analyses of any APO history and 0.0025 for analyses of individual APO histories).
4. DISCUSSION
In this population-based cohort study of older adult ACT study participants with a history of at least one pregnancy, we evaluated whether a history of APOs was associated with cognition score and cognitive decline. We found no significant association between history of any APO, or individual APOs, and lower adjusted average CASI-IRT scores. We also found no evidence of an association between history of any APO (or any individual APO) and the adjusted average difference in 4-year rates of change in cognition.
4.1. Results in context.
A history of APOs has been inconsistently associated with midlife cognitive outcomes in several studies;3,31 however, few studies have evaluated APO history in relation to cognition among women older than 65 years of age.5 The lack of evidence of an association between history of APO with cognitive score among an older population in our study is consistent with a recent finding regarding hypertensive disorders of pregnancy in the Mayo Clinic Study for Aging (MCSA), which found no evidence that a history of hypertensive disorders of pregnancy, versus normotensive pregnancy, was associated with cognitive performance level in a population of women with a median age of 73.5 However, in the MCSA, a history of hypertensive disorders of pregnancy was associated with greater declines longitudinally in global cognition and attention/executive z scores over a median follow-up time of 5.2 years, compared with history of a normotensive pregnancy.5 In contrast, we found no evidence that a history of APO was associated with a difference in rate of change in cognition. This may reflect heterogeneity of our exposure (any APO), compared with hypertensive disorders of pregnancy specifically. Alternatively, the MCSA analysis, which included over 2000 women, may have been better powered to detect differences.
Additional studies have evaluated a history of APOs in relation to cognition in populations of younger postmenopausal women age <65 years. Nearly all of these studies focused on hypertensive disorders of pregnancy, and results remained inconsistent in this age group.3,4,31,32 In a matched case-control study nested within the population-based Rochester Epidemiologic Project (mean age ~59 years), there was no statistically significant evidence that a history of preeclampsia was associated with cognitive level compared with history of a normotensive pregnancy.31 In the ORACLE study nested within the Rotterdam-based Generation R Study (n=576), a history of a hypertensive disorder of pregnancy was associated with reduced objective cognitive function at 14 years post-pregnancy (mean age = 46 years at cognitive testing).3 In the Coronary Artery Risk Development in Young Adults (CARDIA) cohort, a history of preeclampsia was associated with lower Digit Symbol Substitution Test (DSST) scores and on the Stroop Test of executive function at a mean age of 49 years, but not with working memory.32 In the Family Blood Pressure Project Genetic Epidemiology Network of Arteriopathy (GENOA) study of women with a mean age of ~59 years, a history of hypertensive disorders of pregnancy was associated with lower scores on tests of processing speeds.33 In a cross-sectional analysis of the Queen of Hearts case-control study of premenopausal women in the Netherlands, differences in executive functioning were present in those with a history of preeclampsia (mean age 39 years) compared to those with normotensive pregnancies (mean age 44 years).4 One study, the Bogalusa Heart Study (median age ~48 years), included both hypertensive and non-hypertensive APOs: low birthweight, preterm birth, and hypertensive disorders of pregnancy were not strongly associated with cognitive function; however, miscarriage was associated with better cognitive function.20
4.2. Biological rationale.
Brain imaging studies of women with a history of preeclampsia and other hypertensive disorders of pregnancy reveal evidence of early onset of cerebral small vessel disease,34,35 as well as cortical atrophy.33,36 There is a strong biological premise for investigating the impact of non-hypertensive APOs as well as hypertensive APOs on maternal brain health, given that these disorders share common pathophysiological pathways involving vascular placental dysfunction, inflammation, and oxidative stress.13,37,38 Furthermore, a wide range of APOs, not just hypertensive disorders of pregnancy, are associated with increased cardiovascular and cerebrovascular risk.10,17,18,39,40 However, we did not find evidence of significant differences in cognition in our postmenopausal cohort between those with and without a history of APOs. It is possible that differences in cognition and cognitive decline might be observed at earlier time points in the life course, which then attenuate as those who did not experience APOs “catch up” in the vascular aging process.
4.3. Strengths and limitations.
Our study had key strengths. Our setting within ACT, a prospective population-based study conducted within an integrated healthcare delivery system, allowed us to link the ACT study’s rich cognition data with pregnancy history data newly abstracted from electronic and paper medical records. The ACT study uses a well-validated measure of cognition that captures cognition level and the ability to detect minor changes in cognition in multiple domains.41,42 Furthermore, most studies of APOs in relation to cognition have been limited to hypertensive disorders of pregnancy, while our study evaluated a broader range of APOs in relation to cognitive outcomes.
Our study had limitations. Women eligible for this study enrolled in ACT at a median age of 68 years (IQR: 67.0, 70.0) and had a median follow-up time of approximately 4 years (IQR: 2, 8). Therefore, most women completed CASI assessments at younger ages that may precede periods of more rapid cognitive decline occurring later in life. In fact, women were observed to improve minimally in cognition during follow-up on average. Contributing factors to this improvement could be a “re-testing effect”, in which participants perform better on re-testing compared to the first assessment due to better preparation, or potential measurement error. The former claim is substantiated by select participants whose CASI-IRT score improved consistently over multiple visits, with an overall increase by more than 2 points (equivalent to movement from approximately the 50% percentile to the 97.7% percentile in the population). While the reasons for these improvements are unclear, we observed a small amount of improvement on average in the population. We are unaware of any causes for measurement error affecting women differently by APO status; the only potential reason we are aware of for a differential re-testing effect by APO history is the impact of APO history on cognition itself, i.e. their ability to “learn the test”. Hence, we anticipate that estimated mean differences in cognitive decline and rates of change by APO history remain unbiased.
In addition, we were unable to adjust for pre-pregnancy health factors that may be associated with both APO risk and later life cognition score and cognitive decline, such as pre-pregnancy body mass index and blood pressure. Furthermore, ACT participants may be healthier than the general population43. Women who experienced an APO may be more likely to have had an early death and therefore to not survive to participate in ACT, resulting in survival bias. Therefore, our results generalize to individuals who were previously pregnant and have dementia-free survival through an older age.8 ACT participants have high levels of educational attainment, which may reflect social advantages over the life course that do not correspond to the broader US population.
We did not include participants born before 1940, due to incomplete historic pregnancy and obstetrics data available to us for chart review in this population. We therefore limited our study to more recently enrolled participants, who have shorter follow-up that may not capture eventual cognitive decline. Given the birth cohorts from the 1940s and 1950s in this analysis, the experiences of APOs and related health and social determinants over the life course may not reflect the experiences of more contemporary cohorts. For example, the age standardized rate of hypertensive disorders of pregnancy per 1000 live births has doubled from 38.4 to 77.8 in years 2007 to 2019.1 Among contemporary cohorts, Black women are overwhelmingly more likely to experience APOs than other women.8,44 However, almost all participants in this study identify as non-Hispanic white women, which reflects in part the racial and ethnic composition of Seattle.43
We may have had incomplete capture of pregnancies. While we considered all sources, including notes in the medical records relating to previous pregnancies and complications, we were unable to estimate the number of pregnancies missed because we do not have self-reported number of pregnancy or birth information from ACT participants. However, the average cohort fertility rate of two lifetime pregnancies is comparable to the fertility rates for the period 1971–1987 (when our cohort was aged 25–40) of 2.27 – 1.87 lifetime pregnancies [45].
We also likely had incomplete capture of APOs. Women who had given birth before enrolling in KPWA (then GHC) may have experienced APOs that were not captured retrospectively in later health encounters within KPWA/GHC medical records. Even for women enrolled in GHC for all pregnancies, the clinical recognition and documentation of APOs was likely less accurate historically (50–80 years ago), than in present day. This likely resulted in some degree of non-differential misclassification of our exposure, biasing our results toward the null. We also were unable to capture miscarriages, which may be incompletely documented in medical records.
Clinical definitions of hypertensive disorders of pregnancy have changed over time and across the study period, which may have affected the number of women diagnosed in our cohort. Even for participants with prenatal records in the chart, the majority were handwritten notes and it was difficult to be sure of complete records (for example, blood pressure in the postpartum period). Although we abstracted available diastolic and systolic blood pressure data from the first and final prenatal visits when available, these data were available for only ~25% of our participants’ pregnancies; thus, we were unable to use actual blood pressure recordings to adjudicate the presence of hypertensive disorders of pregnancy according to current definitions.
Due to the small sample size, we only considered the binary indicator of ever versus never experience of APOs. However, the number and severity of APOs may be an important cardiovascular and cerebrovascular risk indicator11,45,46 (e.g., early onset preeclampsia or recurrent preeclampsia, vs. milder forms of APO); this risk was not captured by our study.
In conclusion, in this study of women aged 65 and older, compared to women without a history of APO, we found no evidence that a history of APO was associated with either lower average cognition scores, or a difference in 4-year rates of change in cognition. Larger longitudinal studies are needed, ideally with prospective ascertainment and adjudication of pregnancy outcomes with longer cognitive follow-up, to better understand the relationship between pregnancy complications and cognitive aging in women.
Supplementary Material
ACKNOWLEDGMENTS:
We thank the participants of the Adult Changes in Thought (ACT) study for the data they have provided and the many ACT investigators and staff who steward that data. You can learn more about ACT at: https://actagingstudy.org/. The authors further acknowledge Laura Gibbons for help with CASI-IRT data and interpretation, Paul Crane for help with identifying appropriate cognitive subdomains, and Margie Wilcox, Mary Lyons, and Becky Lederman for their abstraction of data from medical records.
CONFLICTS OF INTEREST/DISCLOSURES:
L.B. Harrington, S.E. Tom, K. Chau, R. Muiruri, C. Krakauer, E. Micks, L.K. McEvoy, A.Z. LaCroix, and E.C. Miller have no disclosures. P.D. Wartko reported funding by grants from Syneos Health to conduct US Food and Drug Administration-mandated research from a consortium of pharmaceutical companies that manufacture long-acting opioids through a contract between Syneos Health and Kaiser Permanente Washington, outside the submitted work. All statements in this report, including its findings and conclusions, are solely those of the authors and do not necessarily represent the views of the National Institute on Aging or the National Institutes of Health.
FUNDING SOURCES:
The present study was supported by the National Institute on Aging (NIA) (R21AG069111 to E. Miller) and the National Institute of Neurological Disorders and Stroke (K23NS107645 to E. Miller). The Adult Changes in Thought study is funded by the National Institute on Aging (U19AG066567). Data collection for this work was additionally supported, in part, by prior funding from the National Institute on Aging (U01AG006781).
NON STANDARD ABBREVIATIONS
- ACT
Adult Changes in Thought
- ADRD
Alzheimer’s Disease and related dementias
- APO
Adverse pregnancy outcome
- CASI
Cognitive Abilities Screening Instrument
- CASI-IRT
Cognitive Abilities Screening Instrument, scored using Item Response Theory
- GHC
Group Health Cooperative
- KPWA
Kaiser Permanente Washington
- MELV
Memory, Executive function, Language, and Visuospatial functioning
Footnotes
CONSENT STATEMENT: This study was approved by the Institutional Review Boards of Kaiser Permanente Washington Research Institute and Columbia University Irving Medical Center. All participants in the ACT study provide written, informed consent at time of enrollment, including consent for medical records review.
REFERENCES
- 1.Freaney PM, Harrington K, Molsberry R, et al. Temporal Trends in Adverse Pregnancy Outcomes in Birthing Individuals Aged 15 to 44 Years in the United States, 2007 to 2019. J Am Heart Assoc 2022;11(11):e025050, doi: 10.1161/JAHA.121.025050 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Andolf E, Bladh M, Moller L, et al. Prior placental bed disorders and later dementia: a retrospective Swedish register-based cohort study. BJOG 2020;127(9):1090–1099, doi: 10.1111/1471-0528.16201 [DOI] [PubMed] [Google Scholar]
- 3.Adank MC, Hussainali RF, Oosterveer LC, et al. Hypertensive Disorders of Pregnancy and Cognitive Impairment: A Prospective Cohort Study. Neurology 2021;96(5):e709–e718, doi: 10.1212/WNL.0000000000011363 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Alers RJ, Ghossein-Doha C, Canjels LPW, et al. Attenuated cognitive functioning decades after preeclampsia. Am J Obstet Gynecol 2023;229(3):294 e1–294 e14, doi: 10.1016/j.ajog.2023.02.020 [DOI] [Google Scholar]
- 5.Mielke MM, Frank RD, Christenson LR, et al. Association of Hypertensive Disorders of Pregnancy With Cognition in Later Life. Neurology 2023;100(19):e2017–e2026, doi: 10.1212/WNL.0000000000207134 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Basit S, Wohlfahrt J, Boyd HA. Pre-eclampsia and risk of dementia later in life: nationwide cohort study. BMJ 2018;363(k4109, doi: 10.1136/bmj.k4109 [DOI] [Google Scholar]
- 7.Miller EC, Conley P, Alirezaei M, et al. Associations between adverse pregnancy outcomes and cognitive impairment and dementia: a systematic review and meta-analysis. Lancet Healthy Longev 2024;100660, doi: 10.1016/j.lanhl.2024.100660 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Parikh NI, Gonzalez JM, Anderson CAM, et al. Adverse Pregnancy Outcomes and Cardiovascular Disease Risk: Unique Opportunities for Cardiovascular Disease Prevention in Women: A Scientific Statement From the American Heart Association. Circulation 2021;143(18):e902–e916, doi: 10.1161/CIR.0000000000000961 [DOI] [PubMed] [Google Scholar]
- 9.Brohan MP, Daly FP, Kelly L, et al. Hypertensive disorders of pregnancy and long-term risk of maternal stroke-a systematic review and meta-analysis. Am J Obstet Gynecol 2023;229(3):248–268, doi: 10.1016/j.ajog.2023.03.034 [DOI] [PubMed] [Google Scholar]
- 10.Wu P, Gulati M, Kwok CS, et al. Preterm Delivery and Future Risk of Maternal Cardiovascular Disease: A Systematic Review and Meta-Analysis. J Am Heart Assoc 2018;7(2), doi: 10.1161/JAHA.117.007809 [DOI] [Google Scholar]
- 11.Miller EC, Kauko A, Tom SE, et al. Risk of Midlife Stroke After Adverse Pregnancy Outcomes: The FinnGen Study. Stroke 2023;54(7):1798–1805, doi: 10.1161/STROKEAHA.123.043052 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Testo AA, McBride C, Bernstein IM, et al. Preeclampsia and its relationship to pathological brain aging. Front Physiol 2022;13(979547, doi: 10.3389/fphys.2022.979547 [DOI] [Google Scholar]
- 13.Brosens I, Pijnenborg R, Vercruysse L, et al. The “Great Obstetrical Syndromes” are associated with disorders of deep placentation. Am J Obstet Gynecol 2011;204(3):193–201, doi: 10.1016/j.ajog.2010.08.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Winder Z, Lee T, Fardo DW, et al. Evaluating vascular cognitive impairment and dementia due to small-vessel disease using plasma levels of PLGF and VEGF-A. Alzheimer’s & Dementia 2020;16(S5):e044576, doi: 10.1002/alz.044576 [DOI] [Google Scholar]
- 15.Hinman JD, Elahi F, Chong D, et al. Placental growth factor as a sensitive biomarker for vascular cognitive impairment. Alzheimers Dement 2023;19(8):3519–3527, doi: 10.1002/alz.12974 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Grammas P Neurovascular dysfunction, inflammation and endothelial activation: implications for the pathogenesis of Alzheimer’s disease. J Neuroinflammation 2011;8(26, doi: 10.1186/1742-2094-8-26 [DOI] [Google Scholar]
- 17.Liang C, Chung HF, Dobson AJ, et al. Infertility, Miscarriage, Stillbirth, and the Risk of Stroke Among Women: A Systematic Review and Meta-Analysis. Stroke 2022;53(2):328–337, doi: 10.1161/STROKEAHA.121.036271 [DOI] [PubMed] [Google Scholar]
- 18.Ananth CV, Patrick HS, Ananth S, et al. Maternal Cardiovascular and Cerebrovascular Health After Placental Abruption: A Systematic Review and Meta-Analysis (CHAP-SR). Am J Epidemiol 2021;190(12):2718–2729, doi: 10.1093/aje/kwab206 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Iadecola C, Duering M, Hachinski V, et al. Vascular Cognitive Impairment and Dementia: JACC Scientific Expert Panel. J Am Coll Cardiol 2019;73(25):3326–3344, doi: 10.1016/j.jacc.2019.04.034 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Harville EW, Guralnik J, Romero M, et al. Reproductive History and Cognitive Aging: The Bogalusa Heart Study. Am J Geriatr Psychiatry 2020;28(2):217–225, doi: 10.1016/j.jagp.2019.07.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Kukull WA, Higdon R, Bowen JD, et al. Dementia and Alzheimer disease incidence: a prospective cohort study. Arch Neurol 2002;59(11):1737–46, doi: 10.1001/archneur.59.11.1737 [DOI] [PubMed] [Google Scholar]
- 22.Crane PK, Trittschuh E, Mukherjee S, et al. Incidence of cognitively defined late-onset Alzheimer’s dementia subgroups from a prospective cohort study. Alzheimers Dement 2017;13(12):1307–1316, doi: 10.1016/j.jalz.2017.04.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Lipsky S, Easterling TR, Holt VL, et al. Detecting small for gestational age infants: the development of a population-based reference for Washington state. Am J Perinatol 2005;22(8):405–12, doi: 10.1055/s-2005-872595 [DOI] [PubMed] [Google Scholar]
- 24.Teng EL, Hasegawa K, Homma A, et al. The Cognitive Abilities Screening Instrument (CASI): a practical test for cross-cultural epidemiological studies of dementia. Int Psychogeriatr 1994;6(1):45–58; discussion 62, doi: 10.1017/s1041610294001602 [DOI] [PubMed] [Google Scholar]
- 25.Crane PK, Narasimhalu K, Gibbons LE, et al. Item response theory facilitated cocalibrating cognitive tests and reduced bias in estimated rates of decline. J Clin Epidemiol 2008;61(10):1018–27 e9, doi: 10.1016/j.jclinepi.2007.11.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.VJ C Generalized Estimation Equation Solver. 2023. Available from: https://cran.r-project.org/web/packages/gee/index.html.
- 27.LIANG K-Y, ZEGER SL Longitudinal data analysis using generalized linear models. Biometrika 1986;73(1):13–22, doi: 10.1093/biomet/73.1.13 [DOI] [Google Scholar]
- 28.Dublin S, Walker RL, Gray SL, et al. Prescription Opioids and Risk of Dementia or Cognitive Decline: A Prospective Cohort Study. J Am Geriatr Soc 2015;63(8):1519–26, doi: 10.1111/jgs.13562 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Gray SL, Dublin S, Yu O, et al. Benzodiazepine use and risk of incident dementia or cognitive decline: prospective population based study. BMJ 2016;352(i90, doi: 10.1136/bmj.i90 [DOI] [Google Scholar]
- 30.Team RC. A Language and Environment for Statistical Computing. R FOundation for Statistical Computing; 2023; [Google Scholar]
- 31.Fields JA, Garovic VD, Mielke MM, et al. Preeclampsia and cognitive impairment later in life. Am J Obstet Gynecol 2017;217(1):74 e1–74 e11, doi: 10.1016/j.ajog.2017.03.008 [DOI] [Google Scholar]
- 32.Dayan N, Kaur A, Elharram M, et al. Impact of Preeclampsia on Long-Term Cognitive Function. Hypertension 2018;72(6):1374–1380, doi: 10.1161/HYPERTENSIONAHA.118.11320 [DOI] [PubMed] [Google Scholar]
- 33.Mielke MM, Milic NM, Weissgerber TL, et al. Impaired Cognition and Brain Atrophy Decades After Hypertensive Pregnancy Disorders. Circ Cardiovasc Qual Outcomes 2016;9(2 Suppl 1):S70–6, doi: 10.1161/CIRCOUTCOMES.115.002461 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Postma IR, Bouma A, de Groot JC, et al. Cerebral white matter lesions, subjective cognitive failures, and objective neurocognitive functioning: A follow-up study in women after hypertensive disorders of pregnancy. J Clin Exp Neuropsychol 2016;38(5):585–98, doi: 10.1080/13803395.2016.1143453 [DOI] [PubMed] [Google Scholar]
- 35.Siepmann T, Boardman H, Bilderbeck A, et al. Long-term cerebral white and gray matter changes after preeclampsia. Neurology 2017;88(13):1256–1264, doi: 10.1212/WNL.0000000000003765 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.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(12):2479–2485, doi: 10.1097/HJH.0000000000001492 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Brosens I, Puttemans P, Benagiano G. Placental bed research: I. The placental bed: from spiral arteries remodeling to the great obstetrical syndromes. Am J Obstet Gynecol 2019;221(5):437–456, doi: 10.1016/j.ajog.2019.05.044 [DOI] [PubMed] [Google Scholar]
- 38.Joo EH, Kim YR, Kim N, et al. Effect of Endogenic and Exogenic Oxidative Stress Triggers on Adverse Pregnancy Outcomes: Preeclampsia, Fetal Growth Restriction, Gestational Diabetes Mellitus and Preterm Birth. Int J Mol Sci 2021;22(18), doi: 10.3390/ijms221810122 [DOI] [Google Scholar]
- 39.Xie W, Wang Y, Xiao S, et al. Association of gestational diabetes mellitus with overall and type specific cardiovascular and cerebrovascular diseases: systematic review and meta-analysis. BMJ 2022;378(e070244, doi: 10.1136/bmj-2022-070244 [DOI] [Google Scholar]
- 40.Crump C, Sundquist J, Sundquist K. Preterm Delivery and Long-Term Risk of Stroke in Women: A National Cohort and Cosibling Study. Circulation 2021;143(21):2032–2044, doi: 10.1161/CIRCULATIONAHA.120.052268 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Li G, Larson EB, Shofer JB, et al. Cognitive Trajectory Changes Over 20 Years Before Dementia Diagnosis: A Large Cohort Study. J Am Geriatr Soc 2017;65(12):2627–2633, doi: 10.1111/jgs.15077 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Suemoto CK, Gibbons LE, Thacker EL, et al. Incident prolonged QT interval in midlife and late-life cognitive performance. PLoS One 2020;15(2):e0229519, doi: 10.1371/journal.pone.0229519 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Davis AC, Voelkel JL, Remmers CL, et al. Comparing Kaiser Permanente Members to the General Population: Implications for Generalizability of Research. Perm J 2023;27(2):87–98, doi: 10.7812/TPP/22.172 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Grobman WA, Parker CB, Willinger M, et al. Racial Disparities in Adverse Pregnancy Outcomes and Psychosocial Stress. Obstet Gynecol 2018;131(2):328–335, doi: 10.1097/AOG.0000000000002441 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Hauge MG, Linde JJ, Kofoed KF, et al. Early-onset vs late-onset preeclampsia and risk of coronary atherosclerosis later in life: a clinical follow-up study. Am J Obstet Gynecol MFM 2024;6(5):101371, doi: 10.1016/j.ajogmf.2024.101371 [DOI] [PubMed] [Google Scholar]
- 46.Auger N, Fraser WD, Schnitzer M, et al. Recurrent pre-eclampsia and subsequent cardiovascular risk. Heart 2017;103(3):235–243, doi: 10.1136/heartjnl-2016-309671 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.