Race, Ethnicity, Psychosocial Factors, and Telomere Length in a Multicenter Setting

Abstract

BackgroundLeukocyte telomere length(LTL) has been associated with age, self-reported race/ethnicity, gender, education, and psychosocial factors, including perceived stress, and depression. However, inconsistencies in associations of LTL with disease and other phenotypes exist across studies. Population characteristics, including race/ethnicity, laboratory methods, and statistical approaches in LTL have not been comprehensively studied and could explain inconsistent LTL associations.MethodsLTL was measured using Southern Blot in 1510 participants from a multi-ethnic, multi-center study combining data from 3 centers with different population characteristics and laboratory processing methods. Main associations between LTL and psychosocial factors and LTL and race/ethnicity were evaluated and then compared across generalized estimating equations(GEE) and linear regression models. Statistical models were adjusted for factors typically associated with LTL(age, gender, cancer status) and also accounted for factors related to center differences, including laboratory methods(i.e., DNA extraction). Associations between LTL and psychosocial factors were also evaluated within race/ethnicity subgroups (Non-hispanic Whites, African Americans, and Hispanics).ResultsBeyond adjustment for age, gender, and cancer status, additional adjustments for DNA extraction and clustering by center were needed given their effects on LTL measurements. In adjusted GEE models, longer LTL was associated with African American race (Beta(β)(standard error(SE)) = 0.09(0.04), p-value = 0.04) and Hispanic ethnicity (β(SE) = 0.06(0.01), p-value = 0.02) compared to Non-Hispanic Whites. Longer LTL was also associated with less than a high school education compared to having greater than a high school education (β(SE) = 0.06(0.02), p-value = 0.04). LTL was inversely related to perceived stress (β(SE) = -0.02(0.003), p<0.001). In subgroup analyses, there was a negative association with LTL in African Americans with a high school education versus those with greater than a high school education(β(SE) = -0.11(0.03), p-value<0.001).ConclusionsLaboratory methods and population characteristics that differ by center can influence telomere length associations in multicenter settings, but these effects could be addressed through statistical adjustments. Proper evaluation of potential sources of bias can allow for combined multicenter analyses and may resolve some inconsistencies in reporting of LTL associations. Further, biologic effects on LTL may differ under certain psychosocial and racial/ethnic circumstances and could impact future health disparity studies.

Serum polychlorinated biphenyls and leukocyte telomere length in a highly-exposed population: The Anniston Community Health Survey

“…Racial differences in LTL are important to consider in epidemiologic studies. African Americans tend to have longer telomeres (Hunt et al, 2008; Lynch et al, 2016; Rewak et al, 2014) and a faster rate of telomere attrition than whites (Diez Roux et al, 2009), although this is not observed in all studies (Diez Roux et al, 2009; Okuda et al, 2002), including ours. Additionally, a statistical interaction by race has been reported in studies of cardiovascular disease and LTL (Carty et al, 2015).…”

Prediagnosis Leukocyte Telomere Length and Risk of Ovarian Cancer

“…Therefore, we cannot rule out the possibility that these associations in our study are due to chance or may reflect another exposure that can alter telomere length. For example, depression, which has been associated with reduced telomere length (53, 54), was recently associated with an increased risk of ovarian cancer (55). Strengths and limitations of this study are worth careful consideration.…”

Psychological Profiles in the Prediction of Leukocyte Telomere Length in Healthy Individuals

“…On the other hand, the inclusion of relatively healthy individuals allows the evaluation of relations between variables of interest that are unconfounded by the effects of disease processes on these same variables. Moreover, the sample consisted of mostly French speaking Caucasians which may limit the generalizability of our study to other ethnicities [116, 117]. …”