source: American Society of Hematology

year: 2017

authors: Nancy S. Green, Rashid Halloway, Adam M. Brickman, Giuseppe Tosto, Clinton B. Wright, Chuanhui Dong, Liyong Wang, Ralph L. Sacco, Mitchell S.V. Elkind, Jose Gutierrez, Sandra Barral

Sickle cell disease is a risk factor for stroke. Findings from previous studies examining an association between sickle trait and stroke have been conflicting (Roach ES, 2005; Caughey MC, et al, 2014).An estimated two million people in the United States (U.S.) carry sickle trait (Key NS, et al, 2015). Carrier prevalence among African American newborns is 7-8% and up to 4% among Caribbean Hispanics, especially among people from the Dominican Republic and certain other nationalities (Lorey FW, et al, 1996; Siddiqui S, et al, 2012). Trait carriers are at increased risk of renal papillary necrosis, chronic kidney disease and thrombosis (Naik RP, et al, Bucknor MD, et al, 2014; Porter B, et, 2014). We hypothesized that cohorts of high-risk elderly populations who carry sickle trait may provide an opportunity to examine the question of cerebrovascular risk within the context of other associated risk factors.

Methods: We examined cross-sectional associations between sickle trait and MRI-defined cerebral infarct and small vessel abnormalities, defined by a quantitative trait of white matter hyperintensities (WMHs), in 1476 participants from two well-phenotyped, high-risk elderly population-based cohorts, the Washington Heights-Inwood Columbia Aging Project (WHICAP) (Brickman AM, et, 2008 and 2010) and the Northern Manhattan Study (NOMAS) (Monteith T, et al, 2014). Participants shared by the cohorts were eliminated from WHICAP data. All 536 African American and 940 Caribbean Hispanic participants in the combined cohorts were genotyped for sickle trait. MRI protocols and WMH phenotyping used in both cohorts were performed as previously described using a 1.5T scanner and common pathologically-informed algorithms with good intra-reader reliability (Brickman AM, 2008; Verhaaren BF, 2015; Gutierrez J, et al, 2015). Univariate and multivariable-adjusted analyses were performed with statistical software packages SAS, SPSS and R. Beta coefficients from the linear regression models (WMH as the outcome variable) were exponentiated for interpretation as odds ratios.

Share

Discuss in Social Wall

Twitter

Share to SCDteam

+myBinder