Skip to main content
The Journals of Gerontology Series A: Biological Sciences and Medical Sciences logoLink to The Journals of Gerontology Series A: Biological Sciences and Medical Sciences
. 2019 Dec 6;75(8):1523–1529. doi: 10.1093/gerona/glz217

Sex Differences in the Association Between Pentraxin 3 and Cognitive Decline: The Cardiovascular Health Study

Lindsay M Miller 1,2,✉, Nancy S Jenny 3, Andreea M Rawlings 2,4, Alice M Arnold 5, Annette L Fitzpatrick 6, Oscar L Lopez 7, Michelle C Odden 2,8
Editor: Jay Magaziner
PMCID: PMC7357589  PMID: 31808814

Abstract

Background

The importance of systemic inflammation, measured by C-reactive protein, in cognitive decline has been demonstrated; however, the role of vascular inflammation is less understood. Pentraxin 3 (PTX3) is a novel marker of vascular inflammation.

Methods

We followed adults 65 and older, free of cardiovascular disease (CVD) for up to 9 years (n = 1,547) in the Cardiovascular Health Study. We evaluated the relationship between PTX3 and change in cognitive function, measured using the Modified Mini-Mental State Examination (3MSE), and incident cognitive impairment (3MSE < 80). Mediation by CVD events, and effect modification by sex and apolipoprotein E ɛ4 allele (APOE4) were also examined.

Results

The average decline in 3MSE was 0.77 points per year. The association between PTX3 and change in 3MSE differed between women and men (p = .02). In the adjusted model, each standard deviation higher in PTX3 was associated with a 0.20 greater decline in 3MSE score per year in women over follow-up (95% CI: −0. 37, −0.03; p = .02), compared to no change in men (β = 0.07; 95% CI: −0.08, 0.22). CVD events had a minor effect on the associations. No effect modification by APOE4 was found, although we observed the association of PTX3 and cognitive impairment in women was attenuated and nonsignificant after adjustment for APOE4. There was a paradoxical protective association between PTX3 and reduced cognitive impairment in men, even after adjustment for APOE4.

Conclusions

We found that vascular inflammation was significantly associated with cognitive decline in older women, but not men.

Keywords: Inflammation, Cognitive aging, Biomarkers, Sex differences


Exploring subclinical markers of cognitive decline, and identifying cognitive impairment in its early stages is essential for primary interventions to maintain quality of life in older adults. Inflammation, a commonly identified cause of age-related diseases, is a key factor in the etiology of both cardiovascular diseases (CVD) (1), and cognitive decline with and without dementia (2,3).

The importance of systemic inflammation, commonly measured by C-reactive protein (CRP) in cognitive decline has been well described (2,4–6); however, the role of vascular inflammation with cognitive decline is less understood. Pentraxin 3 (PTX3), a novel vascular inflammatory biomarker in the same family as CRP, is secreted by macrophages and somatic cells when exposed to inflammatory signals (7,8), and has been examined as a marker of CVD and immunity (9,10). Unlike CRP, PTX3 has a distinct vascular pathology, and is produced by cells identified in atherosclerotic lesions including smooth muscle cells and vascular endothelial cells (11). For these reasons, PTX3 may be a more specific risk factor for cognitive decline than CRP.

Only a few studies have evaluated the association between levels of PTX3 and cognitive decline or dementia (12–14), and none have evaluated the mediating effect of CVD events. As evidence suggests that inflammation has a role in predicting CVD events (15), and CVD events have been consistently shown to occur prior to subsequent cognitive decline (16), evaluating the mediating effect can add additional insight into the relationship between vascular inflammation and cognitive function. Although studies have shown sex differences in both the incidence of dementia, the rate of cognitive decline (17,18), as well as differences in pathways and mechanisms, effect modification by sex has not yet been examined in the relationship between PTX3 and cognitive function. However, effect modification by apolipoprotein E ɛ4 allele (APOE4) on the relationship between CRP and cognitive function has been observed previously (19). In the present study, we evaluated the relationship between PTX3 and cognitive function over 9 years of follow-up in the Cardiovascular Health Study, a cohort study of older adults. We examined mediation by CVD events, and effect modification by sex and APOE4.

Methods

Study Population

CHS is a population-based, prospective cohort study designed to examine risk factors for CVD in older adults. The study enrolled 5,201 adults 65 years of age and older in 1989–1990 from four field sites throughout the United States. An additional 687 African American participants were recruited in 1992–1993.

Participant recruitment has been previously described (20). Briefly, participants were recruited from Medicare eligibility lists at four field sites: University of California, Davis in Sacramento County, California, Johns Hopkins University in Washington County, Maryland, Wake Forest University School of Medicine in Forsyth County, North Carolina, and University of Pittsburgh in Pittsburg Pennsylvania. Eligibility included all persons living in the household of the individual sampled, who were noninstitutionalized, who expected to remain in the area for 3 years, and who could give informed consent without the need for a proxy. At baseline and annually during the first 10 years, participants completed clinic visits, physical examinations, health questionnaires, and donation of blood specimens.

The present study was restricted to participants with data on PTX3 (n = 2,118). PTX3 was measured in a subcohort of CHS participants by Jenny and colleagues using a case–control study of CVD (9). The case–control design included participants who had incident stroke, angina, or myocardial infarction (MI) before June 30, 1995, a comparison group with no subclinical CVD and a control group that was randomly selected from the remaining participants. Inverse probability of selection weights was used in this analysis to account for the nonrandom selection of participants for PTX3 assessment.

Pentraxin 3

Serum PTX3 was measured in baseline blood samples using PTX3 (human) Detection Set from Alexis Biochemicals (Axxora, LLC, San Diego, CA); the assay coefficient of variation is 10.2%.

Cognitive Function

Cognitive function and change in cognitive function were measured using two instruments. The primary instrument, the Modified Mini-Mental State Examination (3MSE, range 0-100), assessed global cognitive function and was administered beginning in 1990–1991. Cognitive impairment at the first administration was defined as a 3MSE score of <80. Incident cognitive impairment was defined as a 3MSE score of <80 during two consecutive visit years, or one score of <80 and missing on the following subsequent visit. Executive cognitive ability and processing speed were categorized based on the Digit Symbol Substitution Test (DSST), administered annually through 1999, with a score ranging from 0 to 90, and higher scores representing better cognitive function (21).

To reduce bias from missing data, the Telephone Interview for Cognitive Status (TICS) was used to convert missing values for 3MSE scores during follow-up years 6 through 9 described by Arnold and colleagues (22). After inclusion of TICS, missingness for 3MSE improved from 22% to 19% in follow-up visit 6, from 25% to 19% in follow-up visit 7, from 28% to 20% in follow-up visit 8, and from 30% to 23% in follow-up visit 9.

Potential Confounders and Mediators

All potential confounders were measured at baseline. Demographic variables included age, sex, race (white, black/other), and education (less than high school, high school or GED, some college or degree). Behavioral measures included smoking status (never, former, current), alcohol use defined as none (0 drinks per week), low (1–7 drinks per week for women, or 1–14 drinks for men) or high (>7 drinks per week for women, or >14 drinks per week for men), and body mass index (BMI) was calculated as weight (kg)/height (m)2. Clinical measures included diabetes mellitus (fasting glucose level > 125 mg/dL or the use of glucose-lowering medications), hypertension (seated systolic blood pressure average of ≥160 mmHg, or seated diastolic blood pressure average of ≥95 mmHg, or history of hypertension and use of antihypertensive medication), low-density lipoprotein (LDL) cholesterol (mg/dL), high-density lipoprotein (HDL) cholesterol (mg/dL), and triglycerides (mg/dL). Self-reported medications were transcribed by technicians based on prescription bottles brought in by participants during clinic visits (23). These included 3-hydroxy-3-methylglutaryl coenzyme A inhibitor (statin), aspirin use measured as three or more days in the past 2 weeks (yes or no), nonsteroidal anti-inflammatory (NSAIDS) excluding aspirin, and antihypertensive medication (indication of high blood pressure and use of any of the following: β-blockers, calcium-channel blockers, diuretics, vasodilators, etc.). Kidney disease was measured using serum cystatin C-based estimated glomerular filtration rate (eGFRcys), and was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation (24). Measurement methods for APOE have been published elsewhere (25). APOE4 was dichotomized in this study as having at least one ɛ4 allele (yes, no; APOE4) (25,26). CRP was measured by high sensitivity in-house ELISA; analytic coefficient of variation 8.9% (27), and was transformed using the natural log to avoid skewness. Cardiovascular events included stroke, MI, and the diagnosis of heart failure (HF), and were adjudicated annually by a CHS outcome-assessment committee using standardized criteria (28). Cardiovascular events were included after baseline, and up to 6 months prior to the time-dependent measure of cognitive function to ensure cognitive impairment was not acutely caused by the observed CVD event; this time-window was derived from expert opinion in CHS.

Statistical Analysis

Participants with a history of CVD events at baseline were excluded to ensure primary incident cases upon follow-up (n = 571), which yielded a total analytical sample size of 1,547. PTX3 was transformed using the natural log (ln) as has been previously done in this cohort due to the right skew of the data (9), as were CRP and triglycerides. We summarized baseline characteristics in the subset population. Student’s t tests and pairwise chi-squared (χ 2) tests were used to evaluate differences in baseline characteristics between men and women.

Longitudinal associations were evaluated using an unadjusted and adjusted model. To generate the most parsimonious model, the adjusted model was built using backwards stepwise regression with p ≤ .2 as a criterion for inclusion using variables listed in potential confounders excluding hypothesized mediators. This was done separately in the model for 3MSE and DSST. Longitudinal associations were evaluated using a set of linear mixed models with 3MSE and DSST scores as outcomes and included the cross-sectional ln PTX3 as well as the interaction between ln PTX3 and visit year to evaluate the change in cognitive function over time. Robust variance estimation was used to account for the correlation in cognitive scores across study visits. Additionally, a quadratic visit year term was included in the models to account for nonlinear changes in cognitive function over time (29). The adjusted longitudinal model was build using the same stepwise method and potential confounders. To evaluate the predicted 3MSE scores over follow-up by sex, ln PTX3 was dichotomized using the median value.

To examine the influence of mediating incident CVD events, we compared the difference in the coefficient of the ln PTX3 and visit interaction term before and after adding all 3 CVD events (stroke, MI, HF) to the model. We examined effect modification of the association between ln PTX3 and cognitive function by sex and APOE4 in longitudinal models. We conducted stratified analyses by sex or APOE4 if significant effect modification was observed in the longitudinal analyses. We used Cox proportional hazards regression to evaluate time to incident cognitive impairment. Covariates determined through stepwise regression in the longitudinal model were used to build the adjusted model for Cox regression.

As an exploratory model, we used the Paramed command in Stata to evaluate causal mediation with parametric regression models. This method can calculate controlled and natural direct effects while allowing for an exposure/mediator interaction; in other words, this approach allows for an interaction between ln PTX3 and CVD events. The implementation of this analysis required some changes from the main analysis, which is why we considered this an exploratory approach. First, we needed to dichotomize ln PTX3, so we did so at the median. Additionally, this approach does not easily incorporate repeated measures, so we estimated the cognitive decline for each individual using a multilevel model of 3MSE levels on study year and used this as the outcome of interest. Finally, due to the need to use the average individual slope in 3MSE as the outcome of interest, we used any CVD event during follow-up as the mediator, but this meant that we could not ensure that CVD occurred prior to the changes in 3MSE.

All analyses were performed in Stata (version 14.2, StataCorp, College Station, TX). Significance was defined for all analyses as p < .05.

Results

Compared with men at baseline (n = 567), women (n = 980) were more likely to be never smokers and drinkers. Women were more likely to have hypertension and be on antihypertension medication, statins, and NSAIDs. Furthermore, women had higher baseline levels of LDL and HDL cholesterol, and had higher baseline scores in 3MSE and DSST compared to men (Table 1, p < .05 for all comparisons).

Table 1.

Baseline Characteristics of Participants Stratified by Sex

Total Subset n = 1,547 Women n = 980 Men n = 567
Characteristics Mean (standard error) or Percent
Age (yr) 72.1 (0.17) 72.0 (0.20) 72.3 (0.30)
Black/other 16% 16% 17%
Education
 Less than high school 27% 26% 30%
 High school diploma 28% 30% 25%
 Some college and higher 45% 45% 45%
Smoking***
 Never 49% 59% 28%
 Former 39% 29% 60%
 Current 12% 12% 12%
Alcohol use***
 Never 50% 51% 47%
 Light 40% 40% 39%
 Heavy 10% 9% 14%
Average BMI 26.8 (0.16) 26.9 (0.23) 26.6 (0.19)
Diabetes mellitus 6% 6% 7%
Hypertension* 39% 41% 35%
Any antihypertension medication** 37% 41% 31%
Any statin use*** 2% 2% 0.2%
Any NSAIDS excluding aspirin* 11% 12% 8%
Systolic blood pressure (mm/Hg) 135.8 (0.69) 136.3 (0.90) 134.7 (1.06)
Diastolic blood pressure (mm/Hg)** 71.4 (0.39) 70.7 (0.50) 72.8 (0.60)
Ln C-reactive protein (mg/L)** 0.9 (0.03) 1.0 (0.04) 0.8 (0.05)
Cystatin C (mg/L) 1.0 (0.01) 1.0 (0.01) 1.0 (0.01)
eGFRcys 73.8 (0.57) 73.6 (0.72) 74.4 (0.93)
LDL cholesterol (mg/dL)*** 132.1 (1.23) 136.4 (1.45) 123.9 (1.89)
HDL cholesterol (mg/dL)*** 56.6 (0.52) 60.4 (0.64) 49.2 (0.76)
Ln triglycerides (mg/dL) 4.8 (0.01) 4.8 (0.02) 4.8 (0.02)
Apolipoprotein E ɛ4 allele 26% 26% 27%
Ln Pentraxin 3 (ng/mL)** 0.6 (0.02) 0.6 (0.02) 0.7 (0.03)
DSST score*** 38.3 (0.42) 39.6 (0.53) 35.8 (0.67)
3MSE score** 90.1 (0.29) 90.8 (0.32) 88.8 (0.57)

Note: BMI = Body mass index defined as weight(kg)/height(m)2; DSST = Digit Symbol Substitution Test; Ln = Natural log-transformed; 3MSE = Modified Mini-Mental State Examination; NSAIDS = Nonsteroidal anti-inflammatory medications.

p values based on pairwise χ 2 tests for categorical variables and Student’s t tests for continuous variables.

*p < .05; **p < .01; ***p < .001.

Participants were followed for an average of 4.5 annual visits; the average annual decline in 3MSE was 0.77 points. There was no evidence of effect modification in the longitudinal relationship between PTX3 and cognitive decline by APOE4 for 3MSE or DSST (β: −0.10, 95% CI: −0.41, 0.21, p = .54 and β: 0.08, 95% CI: −0.14, 0.30, p = .49, respectively); however, we found significant effect modification by sex with 3MSE (β: 0.27, 95% CI: 0.04, 0.49, p = .02) and with DSST (β: 0.18, 95% CI: 0.02, 0.34, p = .03). Women with higher baseline ln PTX3 had a steeper rate of decline in predicted 3MSE scores over follow-up compared to men (Figure 1). The interaction between PTX3 and APOE4 was not significant in sex stratified models (β: −0.33, 95% CI: −0.75, 0.08, p = .12 for women; β: 0.27, 95% CI: −0.07, 0.60, p = .13 for men). The PTX3 quadratic term was not significant in either women or men suggesting linear trends (β: −0.09, 95% CI: −0.21, 0.03, p = .16 for women; β: −0.001, 95% CI: −0.12, 0.12, p = .99 for men).

Figure 1.

Figure 1.

Baseline ln Pentraxin 3 (PTX3) dichotomized into high and low levels based on the median value (1.8 ng/mL), and predicted Modified Mini-Mental State Examination (3MSE) score stratified by sex over 9 years of follow-up.

We found a significant longitudinal association between ln PTX3 and 3MSE for the unadjusted and adjusted models in women (Table 2). In the adjusted model, 3MSE scores decreased by an additional 0.20 points per year for each 1-SD higher ln PTX3 at baseline (95% CI: −0.37, −0.03, p = .02). The addition of hypothesized mediators (HF, MI, and stroke) did not attenuate the relationship importantly (β: −0.19, 95% CI: −0.36, −0.02, p = .03). Baseline ln PTX3 was significantly associated with a −0.13 greater decrease in DSST scores in women in the unadjusted model (95% CI: −0.23, −0.03, p = .008), and adjustment for potential confounders did not change the estimated association (β: −0.13, 95% CI: −0.24, −0.03, p = .01). The additional adjustment for CVD events did not attenuate the relationship (β: −0.16, 95% CI: −0.28, −0.05, p = .005; Table 2). Ln PTX3 was not associated with 3MSE or DSST scores in the unadjusted or adjusted models in men.

Table 2.

Multivariable Mixed Effects Analysis Evaluating the Longitudinal Relationship Between Baseline ln Pentraxin 3 (PTX3) and Cognitive Function Over 9 Years of Follow-Up Stratified by Sex

Women n = 980 Men n = 567
Model and Outcome β Coefficient (95% CI) p β Coefficient (95% CI) p
3MSE
 Unadjusted −0.23 (−0.40, −0.06) .009 0.003 (−0.14, 0.14) .97
 Adjusted* −0.20 (−0.37, −0.03) .02 0.07 (−0.08, 0.22) .38
 Mediation† −0.19 (−0.36, −0.02) .03 0.09 (−0.05, 0.23) .22
DSST
 Unadjusted −0.13 (−0.23, −0.03) .008 0.01 (−0.09, 0.12) .81
 Adjusted‡ −0.13 (−0.24, −0.03) .01 0.05 (−0.08, 0.17) .46
 Mediation§ −0.16 (−0.28, −0.05) .005 0.02 (−0.12, 0.15) .82

Note: All coefficients represent the rate of change in cognitive function score per year for each 1 SD higher of baseline ln PTX3 (1.88 ng/mL). CI = Confidence interval; DSST = Digit Symbol Substitution Test; 3MSE = Modified Mini-Mental State Examination.

*Adjusted for study site, age, race, education, alcohol intake, body mass index, diabetes mellitus, aspirin use, LDL cholesterol, cystatin C estimated glomerular filtration rate, apolipoprotein E ɛ4 allele, cystatin C, and C reactive protein.

†Adjusted for covariates in adjusted model,* and potential mediators: incident heart failure, incident myocardial infarction, and incident stroke.

‡Adjusted for study site, age, race, education, alcohol intake, smoking status, diabetes mellitus, nonsteroidal anti-inflammatory medication use, cystatin C estimated glomerular filtration rate, LDL cholesterol, and apolipoprotein E ɛ4 allele.

§Adjusted for covariates in adjusted model,‡ and potential mediators: incident heart failure, incident myocardial infarction, and incident stroke.

In exploratory causal mediation models, the association of high PTX3 and change in 3MSE was robust to adjustment for mediation by CVD events. The natural direct effect of high PTX3 in women was a −0.28 (95% CI: −0.51, −0.05) faster decline in 3MSE per year compared with participants with low PTX3 at baseline (Supplementary Table 1). The controlled direct effects of high PTX3 were stronger in the presence of a mediating CVD event suggesting an interaction between PTX3 and CVD events. The association of high PTX3 and change in 3MSE was in the opposite direction; the natural direct effect of high PTX3 in men was a 0.27 (0.02, 0.52) increase in 3MSE per year compared with men with low PTX3 at baseline. The effect sizes were stronger among those with mediating CVD events, also suggesting an interaction between PTX3 and CVD events.

There were 97 participants with a 3MSE score of <80 at baseline. One hundred and fifty-one (15%) cases of incident cognitive impairment in women, and 86 (15%) cases of incident cognitive impairment in men. In adjusted models, Cox proportional hazards regression showed that baseline ln PTX3 was associated with a 29% greater hazard of incident cognitive impairment in women (95% CI: 0.97, 1.68, p = .09), although this estimate did not reach statistical significance. Paradoxically, the association was in the protective direction for men (HR: 0.78, 95% CI: 0.55, 1.11, p = .17), although it did not reach statistical significance. The addition of APOE to the model moved the estimate toward the null in women (HR: 1.07, 95% CI = 0.75, 1.54, p = .07), and in the protective direction in men (HR: 0.63; 95% CI: 0.43, 0.93, p = .02; Table 3).

Table 3.

Cox Proportional Hazards Regression Evaluating the Relationship Between Baseline PTX3 and Incident Cognitive Impairment Stratified by Sex

Women n = 980 Men n = 567
Model HR (95% CI) p HR (95% CI) p
Unadjusted 1.28 (0.99, 1.65) .06 0.85 (0.62, 1.16) .30
Adjusted* 1.28 (0.97, 1.68) .09 0.78 (0.55, 1.11) .17
Adjusted + APOE4† 1.07 (0.75, 1.54) .70 0.63 (0.43, 0.93) .02
Adjusted + APOE4 + CVD‡ 1.07 (0.75, 1.54) .70 0.59 (0.42, 0.85) .004

Note: Incident cognitive impairment was defined as a Modified Mini-Mental State Examination Score of <80 on two consecutive study visits, or a score of <80 during one visit followed by missing on the subsequent visit. HR = Hazard ratio.

*Adjusted for study site, age, race, education, alcohol intake, body mass index, diabetes mellitus, aspirin use, LDL cholesterol, cystatin C estimated glomerular filtration rate, cystatin C, and C reactive protein.

†Adjusted for covariates in adjusted model* and apolipoprotein E ɛ4 allele.

‡Adjusted for covariates in model† and potential mediators: incident heart failure, incident myocardial infarction, and incident stroke.

Discussion

In this population-based, prospective cohort study of community-dwelling older adults, we found that higher PTX3 was significantly associated with a faster decline in cognitive function scores over follow-up in women, but not in men. Contrary to our expectation, we did not find evidence for a mediating effect of CVD events in this relationship although we found suggestive evidence that mediating CVD events may strengthen the effects of PTX3. These findings highlight the importance of vascular inflammation in cognitive decline. We found evidence of an association with PTX3 with incident cognitive impairment in women, although this association was driven by the presence of the APOE ɛ4 allele. The association of PTX3 and incident cognitive impairment was in the opposite (protective) direction in men, especially after adjustment for the APOE4; however, these unexpected findings should be interpreted with caution. Finally, the association between PTX3 and cognitive decline was independent of CRP suggesting that the relationship between vascular inflammation and cognition function in independent of systemic inflammation.

Our findings are consistent with prior research examining the relationship between inflammatory biomarkers and cognitive function (12–14), and provides further support for a vascular component in the etiology of cognitive decline. Similar to our study, the Ginkgo Evaluation of Memory Study (GEMS) found that APOE4 was a strong confounder for PTX3 and cognitive impairment, which supports the hypothesis that PTX3 and APOE4 share a similar vascular pathway (12).

We found a significant effect of PTX3 on cognitive decline in women but not men, which has been found previously in studies examining the effect of inflammation on cognitive function. For example, the Sydney Memory and Ageing Study found that interleuken-12 was significantly associated with cognitive deficits in women, but not men and Cannon and colleagues in 2011 found that in older adults with metabolic disease, CRP was significantly associated with cognitive decline in women, but not men (30). While it is not clear what role inflammation plays on sex differences in cognitive function, evidence has shown that women have a faster age related decline than men (18), even after controlling for increased lifespan (31). Potential biological explanations for sex differences in cognitive decline include a decrease in sex-hormones(32) and brain glucose metabolism (33), as well as differences in microglial functioning during neurodevelopment (34). Furthermore, men and women have different mechanisms contributing to vascular disease including differences in genes involved in macrophage activation and the expression and dysregulation of inflammatory risk factors (35,36).

Another possible explanation for the observed effect in women and not men is the different timing of vascular risk factors and disease, with men developing CVD risk factors earlier than women (37). It is possible that the association between PTX3 and cognitive decline may be altered when men are younger. Since this study was conducted in older age, we are conditioning on survival. Several studies have found paradoxical beneficial effects of risk factors in older age or among those with manifest disease, and some have attributed this to the depletion of susceptibles or collider stratification bias (38). Since vascular disease often occurs earlier in men, this bias may be more pronounced in men and may contribute to these paradoxical effects. Although it should be noted that this bias is generally believed to have weak effects compared with classic confounding, so it would be unexpected for this to completely explain the sex differences (39,40). The effects of APOE4 are often observed at younger ages, so perhaps the differential timing of vascular disease between men and women also contribute to the differential effect of APOE4 in the analyses of incident cognitive impairment.

While the mechanism by which PTX3 affects cognitive function is not entirely clear, recent evidence has supported a relationship between vascular inflammation and cognitive decline (41). The role of vascular inflammation in cognitive decline may be explained by an increase in inflammatory molecules, such as cytokines and chemokines in the brain, which can increase neuronal damage and reduce phagocytosis activity (42). It may also be attributable to an influx of cerebral infarcts in the brain, which has been shown to exist in patients with both vascular cognitive impairment and dementia, often in the absence of cardiovascular events (43,44). Other evidence suggests that vascular inflammation heightens oxidative stress which can downregulate antioxidant defenses and increase vascular leakage (42).

There were several limitations to this study. First, PTX3 was measured at baseline only, eliminating our ability to evaluate the relationship between changes in PTX3 and changes in cognitive function over time. Additionally, while 3MSE is a common instrument for assessing cognitive function in older adults, evidence has shown a learning effect often witnessed in longitudinal analyses with repeated measures and can underestimate true cognitive decline (45). Moreover, using 3MSE for incident cognitive impairment instead of the more ideal adjudicated dementia or adjudicated mild cognitive impairment in survival analysis may introduce measurement error and potentially bias the hazard ratios, since the cut-point of less than 80 is not a diagnostic parameter. Strengths of this study include the large sample size and longitudinal design. Evaluating the relationship with 9 years of follow-up allows for the identification of a temporal relationship between baseline PTX3 and cognitive function. We did not find a cross-sectional association, but observed a significant effect on the rate of decrease in 3MSE scores in women over time, which may indicate that prospective studies are more appropriate for evaluating the relationship between inflammatory biomarkers and cognitive health. Furthermore, this is the first study to evaluate the mediating effect of adjunct CVD events on the relationship and the effect modification by sex and PTX3 with cognitive decline, providing further insight into the pathology of cognitive impairment.

In summary, we report that in a population-based study of community-dwelling older adults, higher baseline PTX3 was significantly associated with a higher rate of decline in cognitive function scores among women over 9 years of follow-up, and there was minor evidence of mediation by CVD events. We found no evidence to support a relationship between PTX3 and cognitive decline in men, although there is suggested evidence of a paradoxical protective relationship with incident cognitive impairment. Further investigation is warranted to confirm sex differences on the effect of vascular inflammation and cognitive decline, including vascular dementia specifically. By determining appropriate markers of cognitive impairment in its early stages and gaining a better understanding of subgroups affected by higher levels of vascular inflammation, clinicians can make more informed targeted interventions to slow or reduce cognitive decline.

Funding

This work was supported by the National Institute of Health (R01 AG46206), as well as by contracts HHSN268201200036C, HHSN268200800007C, HHSN268201800001C, N01HC55222, N01HC85079, N01HC85080, N01HC85081, N01HC85082, N01HC85083, N01HC85086, and grants U01HL080295 and U01HL130114 from the National Heart, Lung, and Blood Institute (NHLBI), with additional contribution from the National Institute of Neurological Disorders and Stroke (NINDS). Additional support was provided by R01AG023629 from the National Institute on Aging (NIA). A full list of principal CHS investigators and institutions can be found at CHS-NHLBI.org.

This content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Health.

Conflict of Interest

None reported.

Supplementary Material

glz217_suppl_Supplementary_Table_S1

References

  • 1. Shoamanesh A, Preis SR, Beiser AS, et al. Circulating biomarkers and incident ischemic stroke in the Framingham Offspring Study. Neurology. 2016;87:1206–1211. doi: 10.1212/WNL.0000000000003115 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Tegeler C, O’Sullivan JL, Bucholtz N, et al. The inflammatory markers CRP, IL-6, and IL-10 are associated with cognitive function–data from the Berlin Aging Study II. Neurobiol Aging. 2016;38:112–117. doi: 10.1016/j.neurobiolaging.2015.10.039 [DOI] [PubMed] [Google Scholar]
  • 3. Teunissen CE, van Boxtel MPJ, Bosma H, et al. Inflammation markers in relation to cognition in a healthy aging population. J Neuroimmunol. 2003;134:142–150. doi: 10.1016/S0165-5728(02)00398-3 [DOI] [PubMed] [Google Scholar]
  • 4. Hsu PF, Pan WH, Yip BS, Chen RC, Cheng HM, Chuang SY. C-reactive protein predicts incidence of dementia in an elderly Asian Community Cohort. J Am Med Dir Assoc. 2017;18:277.e7–277.e11. doi: 10.1016/j.jamda.2016.12.006 [DOI] [PubMed] [Google Scholar]
  • 5. Choi S-M, Kim BC, Kang KW, et al. Relationship between serum high-sensitivity C-reactive protein levels and cognitive function in patients with Parkinson’s disease. Neurol Asia. 2016;21:349–356. doi:10.3389/fneur.2019.00384 [Google Scholar]
  • 6. Ridker PM, Hennekens CH, Buring JE, Rifai N. C-reactive protein and other markers of inflammation in the prediction of cardiovascular disease in women. N Engl J Med. 2000;342:836–843. doi: 10.1056/NEJM200003233421202 [DOI] [PubMed] [Google Scholar]
  • 7. 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] [PMC free article] [PubMed] [Google Scholar]
  • 8. Mantovani A, Garlanda C, Bottazzi B. Pentraxin 3, a non-redundant soluble pattern recognition receptor involved in innate immunity. Vaccine. 2003;21(Supplement 2):S43–S47. doi:10.1016/S0264-410X(03)00199-3 [DOI] [PubMed] [Google Scholar]
  • 9. Jenny NS, Arnold AM, Kuller LH, Tracy RP, Psaty BM. Associations of pentraxin 3 with cardiovascular disease and all-cause death: the Cardiovascular Health Study. Arterioscler Thromb Vasc Biol. 2009;29:594–599. doi: 10.1161/ATVBAHA.108.178947 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Azzurri A, Sow OY, Amedei A, et al. IFN-gamma-inducible protein 10 and pentraxin 3 plasma levels are tools for monitoring inflammation and disease activity in Mycobacterium tuberculosis infection. Microbes Infect. 2005;7:1–8. doi: 10.1016/j.micinf.2004.09.004 [DOI] [PubMed] [Google Scholar]
  • 11. Mantovani A, Garlanda C, Bottazzi B, et al. The long pentraxin PTX3 in vascular pathology. Vascul Pharmacol. 2006;45:326–330. doi: 10.1016/j.vph.2006.08.011 [DOI] [PubMed] [Google Scholar]
  • 12. Sharma M, Fitzpatrick AL, Arnold AM, et al. Inflammatory biomarkers and cognitive decline: The Ginkgo Evaluation of Memory Study. J Am Geriatr Soc. 2016;64:1171–1177. doi: 10.1111/jgs.14140 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Chi GC, Fitzpatrick AL, Sharma M, Jenny NS, Lopez OL, DeKosky ST. Inflammatory biomarkers predict domain-specific cognitive decline in older adults. J Gerontol A Biol Sci Med Sci. 2017;72:796–803. doi:10.1093/gerona/glw155 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. Fitzpatrick A, Chi G, Hom E, Jenny N, Lopez O, DeKosky S. Pentraxin inflammatory biomarkers and associations with risk of dementia and Alzheimer’s disease in older adults. Alzheimers Dement J Alzheimers Assoc. 2013;9:P235. doi:10.1016/j.jalz.2013.05.447 [Google Scholar]
  • 15. Ridker PM, Buring JE, Shih J, Matias M, Hennekens CH. Prospective study of C-reactive protein and the risk of future cardiovascular events among apparently healthy women. Circulation. 1998;98:731–733. doi: 10.1161/01.CIR.98.8.731 [DOI] [PubMed] [Google Scholar]
  • 16. Pasquier F, Leys D. Why are stroke patients prone to develop dementia? J Neurol. 1997;244:135–142. doi: 10.1007/s004150050064 [DOI] [PubMed] [Google Scholar]
  • 17. De Deyn PP, Goeman J, Vervaet A, Dourcy-Belle-Rose B, Van Dam D, Geerts E. Prevalence and incidence of dementia among 75-80-year-old community-dwelling elderly in different districts of Antwerp, Belgium: the Antwerp Cognition (ANCOG) Study. Clin Neurol Neurosurg. 2011;113:736–745. doi: 10.1016/j.clineuro.2011.07.030 [DOI] [PubMed] [Google Scholar]
  • 18. Proust-Lima C, Amieva H, Letenneur L, Orgogozo JM, Jacqmin-Gadda H, Dartigues JF. Gender and education impact on brain aging: a general cognitive factor approach. Psychol Aging. 2008;23:608–620. doi: 10.1037/a0012838 [DOI] [PubMed] [Google Scholar]
  • 19. Haan MN, Aiello AE, West NA, Jagust WJ. C-reactive protein and rate of dementia in carriers and non carriers of Apolipoprotein APOE4 genotype. Neurobiol Aging. 2008;29:1774–1782. doi: 10.1016/j.neurobiolaging.2007.04.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Tell GS, Fried LP, Hermanson B, Manolio TA, Newman AB, Borhani NO. Recruitment of adults 65 years and older as participants in the Cardiovascular Health Study. Ann Epidemiol. 1993;3:358–366. doi:10.1016/1047-2797(93)90062-9 [DOI] [PubMed] [Google Scholar]
  • 21. Wechsler D. Manual for the Wechsler Adult Intelligence Scale; 1955. http://psycnet.apa.org/psycinfo/1955-07334-000. Accessed October 18, 2016.
  • 22. Arnold AM, Newman AB, Dermond N, Haan M, Fitzpatrick A. Using telephone and informant assessments to estimate missing Modified Mini-Mental State Exam scores and rates of cognitive decline. The cardiovascular health study. Neuroepidemiology. 2009;33:55–65. doi: 10.1159/000215830 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Psaty BM, Lee M, Savage PJ, Rutan GH, German PS, Lyles M. Assessing the use of medications in the elderly: methods and initial experience in the Cardiovascular Health Study. The Cardiovascular Health Study Collaborative Research Group. J Clin Epidemiol. 1992;45:683–692. doi: 10.1016/0895-4356(92)90143-b [DOI] [PubMed] [Google Scholar]
  • 24. Inker LA, Schmid CH, Tighiouart H, et al. Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med. 2012;367:20–29. doi: 10.1056/NEJMoa1114248 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25. Kuller LH, Shemanski L, Manolio T, et al. Relationship between ApoE, MRI findings, and cognitive function in the Cardiovascular Health Study. Stroke. 1998;29:388–398. doi: 10.1161/01.str.29.2.388 [DOI] [PubMed] [Google Scholar]
  • 26. Austin MA, Ordovas JM, Eckfeldt JH, et al. Guidelines of the National Heart, Lung, and Blood Institute Working Group on drawing, processing, and storage for genetic studies; 1996. https://repository.library.georgetown.edu/handle/10822/529655. Accessed October 31, 2016. doi:10.1093/oxfordjournals.aje.a008948
  • 27. Macy EM, Hayes TE, Tracy RP. Variability in the measurement of C-reactive protein in healthy subjects: implications for reference intervals and epidemiological applications. Clin Chem. 1997;43:52–58. [PubMed] [Google Scholar]
  • 28. Ives DG, Fitzpatrick AL, Bild DE, et al. Surveillance and ascertainment of cardiovascular events. Ann Epidemiol. 1995;5:278–285. doi:10.1016/1047-2797(94)00093-9 [DOI] [PubMed] [Google Scholar]
  • 29. Rosso AL, Flatt JD, Carlson MC, et al. Neighborhood socioeconomic status and cognitive function in late life. Am J Epidemiol. 2016;183:1088–1097. doi: 10.1093/aje/kwv337 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Canon ME, Crimmins EM. Sex differences in the association between muscle quality, inflammatory markers, and cognitive decline. J Nutr Health Aging. 2011;15:695–698. doi: 10.1007/s12603-011-0340-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. Ruitenberg A, Ott A, van Swieten JC, Hofman A, Breteler MM. Incidence of dementia: does gender make a difference? Neurobiol Aging. 2001;22:575–580. doi: 10.1016/s0197-4580(01)00231-7 [DOI] [PubMed] [Google Scholar]
  • 32. Long J, He P, Shen Y, Li R. New evidence of mitochondria dysfunction in the female Alzheimer’s disease brain: deficiency of estrogen receptor-β. J Alzheimers Dis. 2012;30:545–558. doi: 10.3233/JAD-2012-120283 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Mosconi L, Mistur R, Switalski R, et al. Declining brain glucose metabolism in normal individuals with a maternal history of Alzheimer disease. Neurology. 2009;72:513–520. doi: 10.1212/01.wnl.0000333247.51383.43 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. Hanamsagar R, Bilbo SD. Sex differences in neurodevelopmental and neurodegenerative disorders: focus on microglial function and neuroinflammation during development. J Steroid Biochem Mol Biol. 2016;160:127–133. doi: 10.1016/j.jsbmb.2015.09.039 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Winham SJ, de Andrade M, Miller VM. Genetics of cardiovascular disease: importance of sex and ethnicity. Atherosclerosis. 2015;241:219–228. doi: 10.1016/j.atherosclerosis.2015.03.021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. Hall JR, Wiechmann AR, Johnson LA, et al. Biomarkers of vascular risk, systemic inflammation, and microvascular pathology and neuropsychiatric symptoms in Alzheimer’s disease. J Alzheimers Dis. 2013;35:363–371. doi: 10.3233/JAD-122359 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. Maas AH, Appelman YE. Gender differences in coronary heart disease. Neth Heart J. 2010;18:598–602. doi: 10.1007/s12471-010-0841-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38. Banack HR, Kaufman JS. The obesity paradox: understanding the effect of obesity on mortality among individuals with cardiovascular disease. Prev Med. 2014;62:96–102. doi: 10.1016/j.ypmed.2014.02.003 [DOI] [PubMed] [Google Scholar]
  • 39. Greenland S. Quantifying biases in causal models: classical confounding vs collider-stratification bias. Epidemiology. 2003;14:300–306. [PubMed] [Google Scholar]
  • 40. Pizzi C, De Stavola B, Merletti F, et al. Sample selection and validity of exposure-disease association estimates in cohort studies. J Epidemiol Community Health. 2011;65:407–411. doi: 10.1136/jech.2009.107185 [DOI] [PubMed] [Google Scholar]
  • 41. Gorelick PB. Role of inflammation in cognitive impairment: results of observational epidemiological studies and clinical trials. Ann N Y Acad Sci. 2010;1207:155–162. doi: 10.1111/j.1749-6632.2010.05726.x [DOI] [PubMed] [Google Scholar]
  • 42. Marchesi C, Paradis P, Schiffrin EL. Role of the renin-angiotensin system in vascular inflammation. Trends Pharmacol Sci. 2008;29:367–374. doi: 10.1016/j.tips.2008.05.003 [DOI] [PubMed] [Google Scholar]
  • 43. Jellinger KA. Morphologic diagnosis of “vascular dementia”—a critical update. J Neurol Sci. 2008;270:1–12. doi: 10.1016/j.jns.2008.03.006 [DOI] [PubMed] [Google Scholar]
  • 44. White L, Small BJ, Petrovitch H, et al. Recent clinical-pathologic research on the causes of dementia in late life: update from the Honolulu-Asia Aging Study. J Geriatr Psychiatry Neurol. 2005;18:224–227. doi: 10.1177/0891988705281872 [DOI] [PubMed] [Google Scholar]
  • 45. Small SA, Stern Y, Tang M, Mayeux R. Selective decline in memory function among healthy elderly. Neurology. 1999;52:1392–1396. doi: 10.1212/wnl.52.7.1392 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

glz217_suppl_Supplementary_Table_S1

Articles from The Journals of Gerontology Series A: Biological Sciences and Medical Sciences are provided here courtesy of Oxford University Press

RESOURCES