Association Between HDL Cholesterol and Incident Cognitive Impairment in Black and White Adults in The Reasons For Geographic and Racial Differences in Stroke Study

Robert Rosenson (  robert.rosenson@mssm.edu ) Mount Sinai Hospital Mount Sinai Heart https://orcid.org/0000-0002-5599-0633 Mary Cushman University of Vermont Emily C. McKinley The University of Alabama at Birmingham School of Medicine Paul Muntner University of Alabama at Birmingham Zhixin Wang University of Alabama at Birmingham Tomas Vaisar University of Washington Jay Heinecke University of Washington Christy Tangney Rush University Suzanne Judd University of Alabama Lisandro D. Colantonio University of Alabama at Birmingham

≥ 45 years of age between 2003 and 2007. Overall, 42% of enrolled participants were Black and 55% were women (9). The study was approved by the institutional review boards at all participating institutions, and all participants provided written informed consent.
Serum HDL-C was measured using blood collected at baseline by colorimetric re ectance spectrophotometry. For the current analysis, we included 26,977 study participants without CI, de ned by a score ≥5, at their rst Six Item Screener (SIS) assessment. During follow-up, cognitive status was assessed using the SIS beginning in December 2003, and the three battery tests in the Consortium to Establish a Registry for Alzheimer's Disease (CERAD): Word List Learning, Word List Delayed Recall and Semantic Animal Fluency tests (10). CERAD and SIS tests were administered by telephone at 18-24 month and 12-month intervals, respectively. We excluded participants who did not complete ≥1 assessment for each of the 3 CERAD battery tests during follow-up (n=6,119), did not have valid data on HDL-C (n=867), or who had a history of stroke at baseline or a stroke event during follow-up (n=1,613). The nal cohort included 18,378 participants. ICI was de ned as a score ≥1.5 standard deviations (SD) below the age, sex, and education-based stratum-speci c means on 2 or 3 of the CERAD battery tests at any single determination during follow up.

Statistical analysis
We described participant characteristics and used Poisson regression models with robust variance estimators to calculate the risk ratio (RR) for ICI associated with HDL-C levels in a series of progressively adjusted models. We adjusted for income and education, smoking, second-hand smoke, body mass index, physical activity, alcohol consumption, use of HRT, and menopausal status among women. We further adjusted for systolic blood pressure and diagnosis of hypertension, and C-reactive protein (CRP). We adjusted for triglycerides due to inverse association with HDL-C. All analyses were conducted among White women, Black women, White men, and Black men, separately.

Results:
The characteristics of study participants and the distribution of HDL-C strati ed by sex and race are presented in Table 1. Overall, 5.2% of White women, 11.6% of Black women, 5.8% of White men and 9.2% of Black men developed ICI over a median follow up of 9.6 years. After adjustment for age and region of residence (Model 1), HDL-C <40 mg/dL versus ≥60 mg/dL was associated with an increased RR (95%  b Based on the number of drinks that participants self-reported having per week during their baseline computer-assisted telephone interview and categorized as follows: • No alcohol consumption: 0 drinks per week.
• Moderate alcohol consumption: >0 to 7 drinks per week for women and >0 to 14 drinks per week for men.
• Heavy alcohol consumption: >7 drinks per week for women and >14 drinks per week for men. Discussion: In this large prospective, population-based study of White and Black women and men, low HDL-C (<40 mg/dL) was associated with an increased risk for ICI in White and Black women, but not in men. The higher risk of ICI was present before and after adjustment for factors known to contribute to HDL-C. However, adjustment for triglycerides attenuated these associations in both White women and Black women. Examination of race-sex differences is important because of well-known differences in levels of HDL-C between women and men and triglyceride levels between White and Black people (11).
Multiple prospective population-based studies, (1, 3, 4) and a pooled analysis of 11 cohort studies including 5,658 participants, reported that lower levels of HDL-C were associated with cognitive decline (10). These studies included participants with a history of stroke. In contrast, there was no association between HDL-C and cognitive function in an analysis of 5,033 stroke-free women and men who participated in the Tromso study (5). The results from the current study are consistent with the ndings from the Tromso study.
Few studies have analyzed sex-speci c differences in HDL-C and ICI. The China Health and Retirement Longitudinal Study (CHARLS) investigated ICI in 6,792 Chinese stroke-free adults aged 45 years and older, and paradoxically reported a higher HDL-C level was associated with a higher risk of cognitive decline in women, but a lower risk of cognitive decline in men (12). Among women <60 years of age, the adjusted odds ratio for cognitive decline associated with HDL-C levels of 40-59 mg/dL and ≥60 mg/dL versus<40 Fasting triglyceride levels were not associated with cognitive decline in participants in CHARLS. In the REGARDS study, an increased risk of ICI in Black and White women with low HDL-C concentrations was present after multivariate adjustment with the exception of triglycerides. Elevated triglycerides were associated with an increased risk of non-Alzheimer's dementia and stroke in a prospective study of

125,727 individuals participating in the Copenhagen General Population Study and Copenhagen City
Heart Study (13). Triglyceride-rich lipoproteins increase atherosclerosis contributing to increased risk of stroke and non-Alzheimer's dementia in Blacks and Whites, but not apparently in Chinese.
HDL-C is a common metric for HDL; however, the cholesterol content of HDL is an inadequate metric for the multifarious functions of HDL (14). Emerging evidence from genetic, clinical and experimental studies supports the involvement of surface protein components of HDL in cognitive function (15). In the current study, we were unable to determine the association of HDL-C with the plethora of anti-oxidant and antiin ammatory properties of speci c HDL subspecies that may have important physiological involvement in mitigating the risk of cognitive decline (14). However, the association between a low level of HDL-C in

Conclusion:
In conclusion, low HDL-C was associated with ICI in White and Black women. Adjustment for lifestyle factors partially attenuated the association between low HDL-C and ICI. Triglyceride-rich lipoproteins share many of the same determinants as HDL-C, but have distinct properties that may contribute to ICI.

Declarations
Ethics approval and consent to participate Study methods were approved by the institutional review boards at participating institutions and all participants provided written informed consent for the storage of their clinical data for research purposes as part of the Reasons for Geographic and Racial Differences in Stroke (REGARDS) project.

Consent for publication:
Not applicable

Availability of data and materials
Quali ed researchers trained in human subject con dentiality protocols may request access to the data that support the ndings of this study by contacting the REGARDS Study at REGARDS@uab.edu Competing interests: The following authors received support for this work: ECM reports Amgen funding to institution.
PM reports NIH funding to institution, Amgen funding to institution, and Amgen consulting fees.
ZW reports no disclosures.