Telomere Length and Magnetic Resonance Imaging Findings of Vascular Brain Injury and Central Brain Atrophy: The Strong Heart Study

Telomere Length and Magnetic Resonance Imaging Findings of Vascular Brain Injury and Central... Abstract Telomeres are repeating regions of DNA that cap chromosomes. They shorten over the mammalian life span, especially in the presence of oxidative stress and inflammation. Telomeres may play a direct role in cell senescence, serving as markers of premature vascular aging. Leukocyte telomere length (LTL) may be associated with premature vascular brain injury and cerebral atrophy. However, reports have been inconsistent, especially among minority populations with a heavy burden of illness related to vascular aging. We examined associations between LTL and magnetic resonance imaging in 363 American Indians aged 64–93 years from the Strong Heart Study (1989–1991) and its ancillary study, Cerebrovascular Disease and Its Consequences in American Indians (2010–2013). Our results showed significant associations of LTL with ventricular enlargement and the presence of white matter hyperintensities. Secondary models indicated that renal function may mediate these associations, although small case numbers limited inference. Hypertension and diabetes showed little evidence of effect modification. Results were most extreme among participants who evinced the largest decline in LTL. Although this study was limited to cross-sectional comparisons, it represents (to our knowledge) the first consideration of associations between telomere length and brain aging in American Indians. Findings suggest a relationship between vascular aging by cell senescence and severity of brain disease. aging, brain, central brain atrophy, magnetic resonance imaging, telomere length, vascular aging, vascular brain injury Telomeres are DNA regions comprising thousands of TTAGGG repeats that protect the ends of mammalian chromosomes from shortening during mitotic cell replication (1). Telomere shortening reflects the inability of replication enzymes to copy lagging ends of DNA strands, as well as the presence of arbitrary DNA damage from oxidative stress and inflammation (2). Such attrition eventually triggers cell senescence, a state in which the cell can no longer replicate and ultimately dies (1). Therefore, measures of leukocyte telomere length (LTL) reflect the accumulated cellular impact of stress and inflammation. Shorter LTL and more rapid LTL loss independently predict cardiovascular disease morbidity and mortality, as well as incident diabetes (3–7). Data from the Cardiovascular Health Study suggest that a difference of approximately 1,000 base pairs in LTL is associated with a 3-fold increase in risk of myocardial infarction, even after accounting for conventional risk factors (8). Attrition of LTL may also be associated with increased risk of vascular brain injury (VBI). Studies examining LTL in relation to overt VBI, such as stroke, have produced inconsistent results (2, 5, 8–12). One study with 100 Europeans identified a potential association between shorter LTL and measures of cortical atrophy and white matter hyperintensities (WMH), although statistical power and generalizability were limited (13). A larger study with white, black, and Hispanic participants demonstrated that LTL was associated with brain and hippocampal volumes, including white matter volume, with shorter telomeres corresponding to smaller volumes (14). Other studies have established associations between telomere attrition and dementia (15), Alzheimer disease (16, 17), Parkinson disease (18), Huntington disease (15), adult glioma (19), diabetes (20), arterial stiffness (21), and cardiovascular disease (6, 7, 22). However, no prior study has examined the association of LTL attrition with subclinical VBI in a minority population with a high incidence of cerebrovascular disease; nor has any research with a health disparities population assessed a wide range of findings from cranial magnetic resonance imaging (MRI)—including WMH, white matter grade, infarcts, hemorrhages, sulcal widening, ventricle enlargement, overall brain volume, and hippocampal volume. American Indians experience a particularly high incidence and prevalence of diseases related to vascular aging, such as coronary heart disease, stroke, hypertension, and diabetes (23–25). Examining LTL in relation to a range of measures of covert, imaging-defined VBI in this population might therefore offer more sensitive etiological insights than existing studies. In addition, given the possibility that many cerebral morbidities related to telomere attrition might share a similar biological mechanism, studies of the factors associated with degenerative and vascular brain abnormalities could elucidate important mechanistic pathologies. The results of such work could inform future preventive or therapeutic research among American Indians and other health disparities populations with similar exposure patterns. Therefore, we conducted a cross-sectional study to investigate whether VBI and cerebral atrophy, as measured by cranial MRI, were associated with shorter LTL in a community sample of American Indians. METHODS Setting The Strong Heart Study assembled the largest population-based cohort of elderly American Indians ever recruited, with 4,549 study participants enrolled between 1989 and 1991 (26). Participants were 45–74 years of age at enrollment and represented 13 tribal communities across 3 major geographic regions, including the Northern Plains, Southern Plains, and Southwest. In follow-up efforts, several waves of longitudinal data were collected over the next 2 decades. The Cerebrovascular Disease and Its Consequences in American Indians (CDCAI) Study, also known as the Strong Heart Stroke Study, was a follow-up examination of survivors from the original cohort conducted between 2010 and 2013. It involved the administration of cranial MRI and assessments of cognitive function (27). Details on methods used in the Strong Heart Study (26) and the CDCAI Study (27), including informed consent procedures and MRI examinations (28), have been previously published. Selection Eligibility criteria for the present study included 1) availability of LTL data from the Strong Heart Study and 2) availability of MRI data from the CDCAI Study. Thus, our cross-sectional sample included 465 eligible cohort survivors. Participants who reported prior stroke (n = 33) or heart disease (including heart attack and heart surgery; n = 70) at the CDCAI visit were excluded from analyses. This resulted in an analytical sample of 363 persons (1 excluded participant reported both stroke and heart disease). Data collection LTL was measured by quantitative polymerase chain reaction from blood samples collected at the baseline visit of the Strong Heart Study (1989–1991) (29). In brief, LTL was assessed as the ratio of telomeric products (T) to single-copy genes (S), yielding the T/S ratio. The rationale behind use of this method is that longer telomeres generate more product in a polymerase chain reaction that uses primers specific to telomeric DNA. The quantitative polymerase chain reaction intraassay coefficient of variation was 4.5% and the interassay coefficient of variation was 6.9%, indicating very good assay performance. Values of LTL greater than 3 standard deviations above the mean (n = 6) were considered technical artifacts; however, participants with these LTL values were already excluded from the analytical sample. Therefore, the final analytical sample comprised 363 persons, and continuous LTL (T/S ratio) was used in all inferential analyses. VBI and brain atrophy were defined on the basis of cranial MRI data (28). Using standardized protocols, study neuroradiologists scored MRI scans for the presence and location of infarcts (including lacunar type) and hemorrhages; graded the degree of sulcal loss, ventricle enlargement, and WMH, each on a 10-point scale from 0 (best) to 9 (worst); and estimated software-processed volumes of hippocampus, total WMH, and total brain mass, which were standardized as percentage of intracranial volume. Age (in years), sex, and study site (Northern Plains, Southern Plains, or Southwest) were assessed at the baseline visit of the Strong Heart Study in 1989–1991. During the CDCAI visit in 2010–2013, additional data were collected on sociodemographic characteristics (years of education, annual family income, Native (tribal) language speaking capacity), health behaviors (tobacco smoking history, alcohol use patterns), and clinical factors. Education, income, Native (tribal) language speaking capacity, smoking, and alcohol use were reported on a self-administered questionnaire. All study variables and categories are defined in Table 1. Table 1. Characteristics of American Indian Participants in a Study of Leukocyte Telomere Length and Brain Aging (n = 363), United States, 2010–2013 Characteristic  No. of Persons  Mean (SD) LTLa  Study site       Northern Plains  201  1.12 (0.46)   Southern Plains  120  1.39 (0.54)   Southwest  42  1.36 (0.58)  Sex       Female  263  1.28 (0.53)   Male  100  1.11 (0.46)  Age at examination, years       60–69  131  1.31 (0.52)   70–74  125  1.20 (0.51)   75–79  62  1.17 (0.51)   ≥80  45  1.23 (0.52)  Household income, dollars/year       <15,000  178  1.23 (0.50)   15,000–35,000  115  1.23 (0.50)   >35,000  67  1.28 (0.59)  Education       Some high school or less  67  1.16 (0.43)   High school graduation  91  1.24 (0.51)   Any college  144  1.25 (0.53)   College graduation  61  1.29 (0.58)  Native language speaking capacityb       Not at all  104  1.23 (0.51)   A little  112  1.29 (0.52)   Moderately  47  1.38 (0.58)   Very well  100  1.12 (0.46)  Alcohol usec       Never drinker  85  1.25 (0.50)   Former drinker  202  1.27 (0.54)   Light drinker  26  1.12 (0.50)   Current drinker  31  1.11 (0.45)  Lifetime cigarette smoking, pack-years       Never smoker  147  1.23 (0.52)   <10  84  1.26 (0.48)   10–25  63  1.18 (0.55)   >25  69  1.28 (0.52)  Body mass indexd       18–24 (normal)  60  1.20 (0.49)   25–29 (overweight)  108  1.25 (0.50)   ≥30 (obese)  192  1.25 (0.55)  Quartile of LDL cholesterol, mg/dL       <75  85  1.20 (0.47)   75–99  108  1.22 (0.52)   100–139  122  1.21 (0.53)   ≥140  43  1.43 (0.53)  Clinical hypertensione       Yes  283  1.25 (0.53)   No  80  1.19 (0.45)  Clinical diabetes mellitusf       Yes  169  1.26 (0.54)   No  194  1.21 (0.50)  Clinical chronic kidney diseaseg       Stage 1  48  1.16 (0.39)   Stage 2  222  1.23 (0.55)   Stage 3  85  1.31 (0.48)   Stage 4  6  1.36 (0.68)   Stage 5  2  1.06 (0.37)  Quartile of C-reactive protein, mg/L       <1.7  95  1.19 (0.42)   1.8–3.4  84  1.20 (0.53)   3.5–6.5  90  1.19 (0.54)   >6.5  89  1.33 (0.53)  Characteristic  No. of Persons  Mean (SD) LTLa  Study site       Northern Plains  201  1.12 (0.46)   Southern Plains  120  1.39 (0.54)   Southwest  42  1.36 (0.58)  Sex       Female  263  1.28 (0.53)   Male  100  1.11 (0.46)  Age at examination, years       60–69  131  1.31 (0.52)   70–74  125  1.20 (0.51)   75–79  62  1.17 (0.51)   ≥80  45  1.23 (0.52)  Household income, dollars/year       <15,000  178  1.23 (0.50)   15,000–35,000  115  1.23 (0.50)   >35,000  67  1.28 (0.59)  Education       Some high school or less  67  1.16 (0.43)   High school graduation  91  1.24 (0.51)   Any college  144  1.25 (0.53)   College graduation  61  1.29 (0.58)  Native language speaking capacityb       Not at all  104  1.23 (0.51)   A little  112  1.29 (0.52)   Moderately  47  1.38 (0.58)   Very well  100  1.12 (0.46)  Alcohol usec       Never drinker  85  1.25 (0.50)   Former drinker  202  1.27 (0.54)   Light drinker  26  1.12 (0.50)   Current drinker  31  1.11 (0.45)  Lifetime cigarette smoking, pack-years       Never smoker  147  1.23 (0.52)   <10  84  1.26 (0.48)   10–25  63  1.18 (0.55)   >25  69  1.28 (0.52)  Body mass indexd       18–24 (normal)  60  1.20 (0.49)   25–29 (overweight)  108  1.25 (0.50)   ≥30 (obese)  192  1.25 (0.55)  Quartile of LDL cholesterol, mg/dL       <75  85  1.20 (0.47)   75–99  108  1.22 (0.52)   100–139  122  1.21 (0.53)   ≥140  43  1.43 (0.53)  Clinical hypertensione       Yes  283  1.25 (0.53)   No  80  1.19 (0.45)  Clinical diabetes mellitusf       Yes  169  1.26 (0.54)   No  194  1.21 (0.50)  Clinical chronic kidney diseaseg       Stage 1  48  1.16 (0.39)   Stage 2  222  1.23 (0.55)   Stage 3  85  1.31 (0.48)   Stage 4  6  1.36 (0.68)   Stage 5  2  1.06 (0.37)  Quartile of C-reactive protein, mg/L       <1.7  95  1.19 (0.42)   1.8–3.4  84  1.20 (0.53)   3.5–6.5  90  1.19 (0.54)   >6.5  89  1.33 (0.53)  Abbreviations: eGFR, estimated glomerular filtration rate; LDL, low-density lipoprotein; LTL, leukocyte telomere length; SD, standard deviation. a LTL was assessed as the T/S ratio (ratio of telomeric products (T) to single-copy genes (S)). b Native (tribal) language speaking capacity was categorized according to responses on the questionnaire. c Categories of alcohol use were based on self-reported recency of use and typical use, as follows: never drinker; former drinker, with last drink more than 1 year before; light drinker, with last drink more than 1 month before; and current drinker, with last drink within the past month. d Weight (kg)/height (m)2. e Clinical hypertension was defined as systolic blood pressure/diastolic blood pressure ≥130/90 mm Hg or use of antihypertensive medication. f Clinical diabetes was defined as fasting glucose concentration ≥126 mg/dL or use of antidiabetic medication. g Chronic kidney disease was based on clinical criteria: stage 1, eGFR ≥90 mL/minute; stage 2, eGFR 60–89 mL/minute; stage 3, eGFR 30–59 mL/minute; stage 4, eGFR 15–29 mL/minute; stage 5, eGFR <15 mL/minute (end-stage renal disease). Table 1. Characteristics of American Indian Participants in a Study of Leukocyte Telomere Length and Brain Aging (n = 363), United States, 2010–2013 Characteristic  No. of Persons  Mean (SD) LTLa  Study site       Northern Plains  201  1.12 (0.46)   Southern Plains  120  1.39 (0.54)   Southwest  42  1.36 (0.58)  Sex       Female  263  1.28 (0.53)   Male  100  1.11 (0.46)  Age at examination, years       60–69  131  1.31 (0.52)   70–74  125  1.20 (0.51)   75–79  62  1.17 (0.51)   ≥80  45  1.23 (0.52)  Household income, dollars/year       <15,000  178  1.23 (0.50)   15,000–35,000  115  1.23 (0.50)   >35,000  67  1.28 (0.59)  Education       Some high school or less  67  1.16 (0.43)   High school graduation  91  1.24 (0.51)   Any college  144  1.25 (0.53)   College graduation  61  1.29 (0.58)  Native language speaking capacityb       Not at all  104  1.23 (0.51)   A little  112  1.29 (0.52)   Moderately  47  1.38 (0.58)   Very well  100  1.12 (0.46)  Alcohol usec       Never drinker  85  1.25 (0.50)   Former drinker  202  1.27 (0.54)   Light drinker  26  1.12 (0.50)   Current drinker  31  1.11 (0.45)  Lifetime cigarette smoking, pack-years       Never smoker  147  1.23 (0.52)   <10  84  1.26 (0.48)   10–25  63  1.18 (0.55)   >25  69  1.28 (0.52)  Body mass indexd       18–24 (normal)  60  1.20 (0.49)   25–29 (overweight)  108  1.25 (0.50)   ≥30 (obese)  192  1.25 (0.55)  Quartile of LDL cholesterol, mg/dL       <75  85  1.20 (0.47)   75–99  108  1.22 (0.52)   100–139  122  1.21 (0.53)   ≥140  43  1.43 (0.53)  Clinical hypertensione       Yes  283  1.25 (0.53)   No  80  1.19 (0.45)  Clinical diabetes mellitusf       Yes  169  1.26 (0.54)   No  194  1.21 (0.50)  Clinical chronic kidney diseaseg       Stage 1  48  1.16 (0.39)   Stage 2  222  1.23 (0.55)   Stage 3  85  1.31 (0.48)   Stage 4  6  1.36 (0.68)   Stage 5  2  1.06 (0.37)  Quartile of C-reactive protein, mg/L       <1.7  95  1.19 (0.42)   1.8–3.4  84  1.20 (0.53)   3.5–6.5  90  1.19 (0.54)   >6.5  89  1.33 (0.53)  Characteristic  No. of Persons  Mean (SD) LTLa  Study site       Northern Plains  201  1.12 (0.46)   Southern Plains  120  1.39 (0.54)   Southwest  42  1.36 (0.58)  Sex       Female  263  1.28 (0.53)   Male  100  1.11 (0.46)  Age at examination, years       60–69  131  1.31 (0.52)   70–74  125  1.20 (0.51)   75–79  62  1.17 (0.51)   ≥80  45  1.23 (0.52)  Household income, dollars/year       <15,000  178  1.23 (0.50)   15,000–35,000  115  1.23 (0.50)   >35,000  67  1.28 (0.59)  Education       Some high school or less  67  1.16 (0.43)   High school graduation  91  1.24 (0.51)   Any college  144  1.25 (0.53)   College graduation  61  1.29 (0.58)  Native language speaking capacityb       Not at all  104  1.23 (0.51)   A little  112  1.29 (0.52)   Moderately  47  1.38 (0.58)   Very well  100  1.12 (0.46)  Alcohol usec       Never drinker  85  1.25 (0.50)   Former drinker  202  1.27 (0.54)   Light drinker  26  1.12 (0.50)   Current drinker  31  1.11 (0.45)  Lifetime cigarette smoking, pack-years       Never smoker  147  1.23 (0.52)   <10  84  1.26 (0.48)   10–25  63  1.18 (0.55)   >25  69  1.28 (0.52)  Body mass indexd       18–24 (normal)  60  1.20 (0.49)   25–29 (overweight)  108  1.25 (0.50)   ≥30 (obese)  192  1.25 (0.55)  Quartile of LDL cholesterol, mg/dL       <75  85  1.20 (0.47)   75–99  108  1.22 (0.52)   100–139  122  1.21 (0.53)   ≥140  43  1.43 (0.53)  Clinical hypertensione       Yes  283  1.25 (0.53)   No  80  1.19 (0.45)  Clinical diabetes mellitusf       Yes  169  1.26 (0.54)   No  194  1.21 (0.50)  Clinical chronic kidney diseaseg       Stage 1  48  1.16 (0.39)   Stage 2  222  1.23 (0.55)   Stage 3  85  1.31 (0.48)   Stage 4  6  1.36 (0.68)   Stage 5  2  1.06 (0.37)  Quartile of C-reactive protein, mg/L       <1.7  95  1.19 (0.42)   1.8–3.4  84  1.20 (0.53)   3.5–6.5  90  1.19 (0.54)   >6.5  89  1.33 (0.53)  Abbreviations: eGFR, estimated glomerular filtration rate; LDL, low-density lipoprotein; LTL, leukocyte telomere length; SD, standard deviation. a LTL was assessed as the T/S ratio (ratio of telomeric products (T) to single-copy genes (S)). b Native (tribal) language speaking capacity was categorized according to responses on the questionnaire. c Categories of alcohol use were based on self-reported recency of use and typical use, as follows: never drinker; former drinker, with last drink more than 1 year before; light drinker, with last drink more than 1 month before; and current drinker, with last drink within the past month. d Weight (kg)/height (m)2. e Clinical hypertension was defined as systolic blood pressure/diastolic blood pressure ≥130/90 mm Hg or use of antihypertensive medication. f Clinical diabetes was defined as fasting glucose concentration ≥126 mg/dL or use of antidiabetic medication. g Chronic kidney disease was based on clinical criteria: stage 1, eGFR ≥90 mL/minute; stage 2, eGFR 60–89 mL/minute; stage 3, eGFR 30–59 mL/minute; stage 4, eGFR 15–29 mL/minute; stage 5, eGFR <15 mL/minute (end-stage renal disease). Clinical factors were measured by study staff or by laboratory assay from blood and urine samples collected during the CDCAI visit. These included body mass index (weight (kg)/height (m)2), C-reactive protein concentration (mg/L), low-density lipoprotein cholesterol concentration (mg/dL), systolic and diastolic blood pressure (mm Hg), fasting glucose concentration (mg/dL), and glomerular filtration rate (mL/minute), estimated by means of the 2012 Chronic Kidney Disease Epidemiology Collaboration equation (30, 31). Clinical conditions (kidney disease, hypertension, hyperlipidemia, diabetes) are shown in Table 1. Analyses We calculated summary statistics for LTL (T/S ratio) according to categories of selected participant characteristics as of the time of the CDCAI visit (n = 363). We calculated Spearman nonparametric coefficients (ρ) and P values for correlations between continuous LTL and continuous, graded, and binary MRI findings. We created box plots and scatterplots of continuous LTL in association with MRI findings, with horizontal axes chosen to maximize visual clarity. We constructed multivariable regression models and examined exponentiated (β) coefficients to estimate measures of risk between LTL (exposure) and MRI findings (outcome) by using Poisson regression for graded variables with incidence rate ratios, logistic regression for binary variables with odds ratios, and linear regression for volumetric variables with risk ratios. All point estimates are reported with 95% confidence intervals. We conducted adjustments for regression models in series. Model 1 included age, sex, and study site. Model 2 added alcohol, smoking, C-reactive protein, and low-density lipoprotein cholesterol. Model 3 added education, income, and body mass index. We structured these adjustment schematics to first adjust for a priori confounders based on their established associations with both exposure and outcome (model 1, model 2). Additional hypothetical confounders, with an established association with exposure but an unclear association with outcome (or vice versa), were added in model 3 in order to empirically evaluate the influence of adding these factors to the primary models, to avoid overfitting and unnecessary loss of degrees of freedom. Model 4 added adjustment and interaction terms for hypertension, diabetes mellitus, or stage of chronic kidney disease (CKD). The reasoning behind these structured adjustments was that age, sex, lipid levels, inflammation, alcohol use, and smoking are likely or known determinants of LTL, justifying their inclusion a priori as confounders. Education and income were also included a priori as proxy variables for socioeconomic status. Study site was included as well, based on possible site-specific differences in genetic control of LTL (32). Clinical factors such as hypertension, diabetes, and kidney disease are either hypothetical mediators (e.g., downstream of the clinical effects of LTL attrition) or confounders (e.g., upstream of the clinical effects of both LTL attrition and VBI) of the primary associations of interest. Therefore, each clinical factor was examined separately using a term for interaction with the exposure of interest (LTL). Secondary analyses stratified the models by study site because of significant age differences noted among participants across study sites. All statistical analyses were conducted with Stata, version 14 (StataCorp LP, College Station, Texas). RESULTS The study population was predominantly female and aged 70 years or older, with an overall age range of 64–93 years (Table 1). Annual household income was generally low, with nearly half of participants having an income below $15,000, although more than half of the participants had at least some college education. Many participants were bilingual, speaking both English (a cohort eligibility criterion) and their Native (tribal) language moderately well to very well. Most participants either never smoked or had limited smoking experience, and most either never used alcohol or had limited alcohol consumption. Obesity, diabetes, and hypertension were generally common. Low-density lipoprotein cholesterol concentration and CKD stage were predominantly within the normal range, but with large variance. Overall, analytical population characteristics were similar to those of the larger CDCAI Study population from which the analytical sample was drawn (27). The overall distribution of LTL (T/S ratio) was nearly normal, with a slight rightward skew, although there was some variability along the spectrum of values, probably due to small numbers (see Web Figure 1, available at https://academic.oup.com/aje). LTL was generally shorter for male, low-income, low-education, alcohol-using, hypertensive, and diabetic participants; LTL was generally longer for dyslipidemic and overweight or obese participants (Table 1). There was no clear relationship of LTL with bilingual status or smoking history, and there was possibly a nonlinear relationship with older age and stage of CKD. LTL measurements also revealed site-specific variations. The Northern Plains site had somewhat lower mean LTL values, the Southern Plains site had somewhat higher mean LTL values, and the Southwest site had intermediate LTL values. Spearman nonparametric correlation coefficients for correlations between LTL and MRI findings demonstrated significant negative associations between LTL and sulcal grade (ρ = −0.16), ventricle grade (ρ = −0.23), and WMH volume (ρ = −0.25) and positive associations with hippocampal volume (ρ = 0.12). Shorter telomeres corresponded to more extreme MRI findings for each factor (Web Table 1). Box plots of LTL according to WMH, sulcal, and ventricle grade (Figure 1, left column) showed similar trends, with lower LTL ranges among the higher grades. The trends for ventricle and sulcal grades were most evident. Small numbers in the highest grades led to imprecise estimates of the range of LTL. Scatterplots of LTL according to WMH, hippocampus, and total brain volumetric estimates (Figure 1, right column) repeated the results from correlation tests, with larger mean values and greater variability in WMH volumes corresponding to shorter LTL, but without any evidence of trend for hippocampal or total brain volumes. Figure 1. View largeDownload slide Association of leukocyte telomere length (LTL), defined as T/S ratio, with brain grades and volumes from cranial magnetic resonance imaging (MRI) in a study of LTL and brain aging among American Indians, United States, 2010–2013. Left-hand column: horizontal bar graphs of LTL according to graded findings from MRI, including white matter hyperintensity (WMH) grade (A), sulcal grade (C), and ventricle grade (E). (Bars, 95% confidence intervals.) Right-hand column: scatterplots with quadratic polynomial regression (solid lines) and 95% confidence intervals (shaded gray areas) for LTL (T/S ratio), with structural volumes defined by MRI as percentage of intracranial volume, including WMH volume (B), hippocampal volume (D), and total brain volume (F). T/S, ratio of telomeric products (T) to single-copy genes (S). Figure 1. View largeDownload slide Association of leukocyte telomere length (LTL), defined as T/S ratio, with brain grades and volumes from cranial magnetic resonance imaging (MRI) in a study of LTL and brain aging among American Indians, United States, 2010–2013. Left-hand column: horizontal bar graphs of LTL according to graded findings from MRI, including white matter hyperintensity (WMH) grade (A), sulcal grade (C), and ventricle grade (E). (Bars, 95% confidence intervals.) Right-hand column: scatterplots with quadratic polynomial regression (solid lines) and 95% confidence intervals (shaded gray areas) for LTL (T/S ratio), with structural volumes defined by MRI as percentage of intracranial volume, including WMH volume (B), hippocampal volume (D), and total brain volume (F). T/S, ratio of telomeric products (T) to single-copy genes (S). Regression models examining independent associations between LTL and MRI findings (Tables 2–4) demonstrated significant associations for ventricle grade (incidence rate ratio = 0.9, 95% confidence interval (CI): 0.8, 0.9; P = 0.003) and WMH volume (risk ratio = 0.9, 95% CI: 0.8, 1.0; P = 0.015). Results from models 1–3 were not substantively different, although the statistical tests for sulcal grade did not exclude chance as a potential explanation for the associations observed in models with the most adjustments. Interpretation of the direction and magnitude of these associations is as follows: Per unit longer LTL (in units of T/S ratio), ventricle and sulcal grades were 10% lower and WMH volumes were 10% smaller, independently of age, sex, study site, alcohol, smoking, low-density lipoprotein cholesterol, education, income, or body mass index. In more general terms: For each clinical factor, shorter telomeres were associated with more extreme MRI findings. Table 2. Associations Between Leukocyte Telomere Length (per Unit of T/S Ratio) and Graded Findings From Cranial Magnetic Resonance Imaging Among American Indians, United States, 2010–2013 Type of MRI Findinga  Model 1b  Model 2c  Model 3d  IRR  95% CI  P Value  IRR  95% CI  P Value  IRR  95% CI  P Value  White matter intensities  0.9  0.8, 1.0  0.201  0.9  0.8, 1.1  0.301  0.9  0.8, 1.1  0.327  Sulcal loss  0.9  0.9, 1.0  0.033  0.9  0.9, 1.0  0.056  0.9  0.9, 1.0  0.084  Ventricle enlargement  0.9  0.8, 0.9  0.001  0.9  0.8, 0.9  0.002  0.9  0.8, 0.9  0.003  Type of MRI Findinga  Model 1b  Model 2c  Model 3d  IRR  95% CI  P Value  IRR  95% CI  P Value  IRR  95% CI  P Value  White matter intensities  0.9  0.8, 1.0  0.201  0.9  0.8, 1.1  0.301  0.9  0.8, 1.1  0.327  Sulcal loss  0.9  0.9, 1.0  0.033  0.9  0.9, 1.0  0.056  0.9  0.9, 1.0  0.084  Ventricle enlargement  0.9  0.8, 0.9  0.001  0.9  0.8, 0.9  0.002  0.9  0.8, 0.9  0.003  Abbreviations: CI, confidence interval; IRR, incidence rate ratio; T/S ratio, ratio of telomeric products (T) to single-copy genes (S). a The degree of white matter intensities, sulcal loss, and ventricle enlargement was graded on a 10-point scale from 0 (best) to 9 (worst). b Model 1 adjusted for age, sex, and study site. c Model 2 adjusted for model 1 variables plus alcohol, smoking, C-reactive protein, and low-density lipoprotein cholesterol. d Model 3 adjusted for model 2 variables plus education, income, and body mass index. Table 2. Associations Between Leukocyte Telomere Length (per Unit of T/S Ratio) and Graded Findings From Cranial Magnetic Resonance Imaging Among American Indians, United States, 2010–2013 Type of MRI Findinga  Model 1b  Model 2c  Model 3d  IRR  95% CI  P Value  IRR  95% CI  P Value  IRR  95% CI  P Value  White matter intensities  0.9  0.8, 1.0  0.201  0.9  0.8, 1.1  0.301  0.9  0.8, 1.1  0.327  Sulcal loss  0.9  0.9, 1.0  0.033  0.9  0.9, 1.0  0.056  0.9  0.9, 1.0  0.084  Ventricle enlargement  0.9  0.8, 0.9  0.001  0.9  0.8, 0.9  0.002  0.9  0.8, 0.9  0.003  Type of MRI Findinga  Model 1b  Model 2c  Model 3d  IRR  95% CI  P Value  IRR  95% CI  P Value  IRR  95% CI  P Value  White matter intensities  0.9  0.8, 1.0  0.201  0.9  0.8, 1.1  0.301  0.9  0.8, 1.1  0.327  Sulcal loss  0.9  0.9, 1.0  0.033  0.9  0.9, 1.0  0.056  0.9  0.9, 1.0  0.084  Ventricle enlargement  0.9  0.8, 0.9  0.001  0.9  0.8, 0.9  0.002  0.9  0.8, 0.9  0.003  Abbreviations: CI, confidence interval; IRR, incidence rate ratio; T/S ratio, ratio of telomeric products (T) to single-copy genes (S). a The degree of white matter intensities, sulcal loss, and ventricle enlargement was graded on a 10-point scale from 0 (best) to 9 (worst). b Model 1 adjusted for age, sex, and study site. c Model 2 adjusted for model 1 variables plus alcohol, smoking, C-reactive protein, and low-density lipoprotein cholesterol. d Model 3 adjusted for model 2 variables plus education, income, and body mass index. Table 3. Associations Between Leukocyte Telomere Length (per Unit of T/S Ratio) and Lesion Findings From Cranial Magnetic Resonance Imaging Among American Indians, United States, 2010–2013 Type of MRI Finding  Model 1a  Model 2b  Model 3c  OR  95% CI  P Value  OR  95% CI  P Value  OR  95% CI  P Value  Any infarcts (≥3 mm)  0.7  0.5, 1.1  0.167  0.7  0.5, 1.2  0.215  0.7  0.4, 1.2  0.174  Lacunar infarcts  0.7  0.4, 1.2  0.205  0.8  0.5, 1.3  0.384  0.8  0.4, 1.3  0.319  Hemorrhages (any)  0.9  0.4, 2.3  0.851  1.0  0.4, 2.4  0.978  1.0  0.4, 2.5  0.968  Type of MRI Finding  Model 1a  Model 2b  Model 3c  OR  95% CI  P Value  OR  95% CI  P Value  OR  95% CI  P Value  Any infarcts (≥3 mm)  0.7  0.5, 1.1  0.167  0.7  0.5, 1.2  0.215  0.7  0.4, 1.2  0.174  Lacunar infarcts  0.7  0.4, 1.2  0.205  0.8  0.5, 1.3  0.384  0.8  0.4, 1.3  0.319  Hemorrhages (any)  0.9  0.4, 2.3  0.851  1.0  0.4, 2.4  0.978  1.0  0.4, 2.5  0.968  Abbreviations: CI, confidence interval; OR, odds ratio; T/S ratio, ratio of telomeric products (T) to single-copy genes (S). a Model 1 adjusted for age, sex, and study site. b Model 2 adjusted for model 1 variables plus alcohol, smoking, C-reactive protein, and low-density lipoprotein cholesterol. c Model 3 adjusted for model 2 variables plus education, income, and body mass index. Table 3. Associations Between Leukocyte Telomere Length (per Unit of T/S Ratio) and Lesion Findings From Cranial Magnetic Resonance Imaging Among American Indians, United States, 2010–2013 Type of MRI Finding  Model 1a  Model 2b  Model 3c  OR  95% CI  P Value  OR  95% CI  P Value  OR  95% CI  P Value  Any infarcts (≥3 mm)  0.7  0.5, 1.1  0.167  0.7  0.5, 1.2  0.215  0.7  0.4, 1.2  0.174  Lacunar infarcts  0.7  0.4, 1.2  0.205  0.8  0.5, 1.3  0.384  0.8  0.4, 1.3  0.319  Hemorrhages (any)  0.9  0.4, 2.3  0.851  1.0  0.4, 2.4  0.978  1.0  0.4, 2.5  0.968  Type of MRI Finding  Model 1a  Model 2b  Model 3c  OR  95% CI  P Value  OR  95% CI  P Value  OR  95% CI  P Value  Any infarcts (≥3 mm)  0.7  0.5, 1.1  0.167  0.7  0.5, 1.2  0.215  0.7  0.4, 1.2  0.174  Lacunar infarcts  0.7  0.4, 1.2  0.205  0.8  0.5, 1.3  0.384  0.8  0.4, 1.3  0.319  Hemorrhages (any)  0.9  0.4, 2.3  0.851  1.0  0.4, 2.4  0.978  1.0  0.4, 2.5  0.968  Abbreviations: CI, confidence interval; OR, odds ratio; T/S ratio, ratio of telomeric products (T) to single-copy genes (S). a Model 1 adjusted for age, sex, and study site. b Model 2 adjusted for model 1 variables plus alcohol, smoking, C-reactive protein, and low-density lipoprotein cholesterol. c Model 3 adjusted for model 2 variables plus education, income, and body mass index. Table 4. Associations Between Leukocyte Telomere Length (per Unit of T/S Ratio) and Volumetric Findings From Cranial Magnetic Resonance Imaging Among American Indians, United States, 2010–2013 Type of MRI Finding  Model 1a  Model 2b  Model 3c  RR  95% CI  P Value  RR  95% CI  P Value  RR  95% CI  P Value  WMH volumed  0.9  0.8, 1.0  0.005  0.9  0.8, 1.0  0.009  0.9  0.8, 1.0  0.015  Hippocampal volumed  1.0  1.0, 1.0  0.527  1.0  1.0, 1.0  0.441  1.0  1.0, 1.0  0.527  Brain volumed  1.7  0.7, 4.2  0.221  1.5  0.6, 3.8  0.347  1.5  0.6, 3.8  0.374  Type of MRI Finding  Model 1a  Model 2b  Model 3c  RR  95% CI  P Value  RR  95% CI  P Value  RR  95% CI  P Value  WMH volumed  0.9  0.8, 1.0  0.005  0.9  0.8, 1.0  0.009  0.9  0.8, 1.0  0.015  Hippocampal volumed  1.0  1.0, 1.0  0.527  1.0  1.0, 1.0  0.441  1.0  1.0, 1.0  0.527  Brain volumed  1.7  0.7, 4.2  0.221  1.5  0.6, 3.8  0.347  1.5  0.6, 3.8  0.374  Abbreviations: CI, confidence interval; RR, risk ratio; T/S ratio, ratio of telomeric products (T) to single-copy genes (S); WMH, white matter hyperintensities. a Model 1 adjusted for age, sex, and study site. b Model 2 adjusted for model 1 variables plus alcohol, smoking, C-reactive protein, and low-density lipoprotein cholesterol. c Model 3 adjusted for model 2 variables plus education, income, and body mass index. d Volumes are expressed as percentage of intracranial volume. Table 4. Associations Between Leukocyte Telomere Length (per Unit of T/S Ratio) and Volumetric Findings From Cranial Magnetic Resonance Imaging Among American Indians, United States, 2010–2013 Type of MRI Finding  Model 1a  Model 2b  Model 3c  RR  95% CI  P Value  RR  95% CI  P Value  RR  95% CI  P Value  WMH volumed  0.9  0.8, 1.0  0.005  0.9  0.8, 1.0  0.009  0.9  0.8, 1.0  0.015  Hippocampal volumed  1.0  1.0, 1.0  0.527  1.0  1.0, 1.0  0.441  1.0  1.0, 1.0  0.527  Brain volumed  1.7  0.7, 4.2  0.221  1.5  0.6, 3.8  0.347  1.5  0.6, 3.8  0.374  Type of MRI Finding  Model 1a  Model 2b  Model 3c  RR  95% CI  P Value  RR  95% CI  P Value  RR  95% CI  P Value  WMH volumed  0.9  0.8, 1.0  0.005  0.9  0.8, 1.0  0.009  0.9  0.8, 1.0  0.015  Hippocampal volumed  1.0  1.0, 1.0  0.527  1.0  1.0, 1.0  0.441  1.0  1.0, 1.0  0.527  Brain volumed  1.7  0.7, 4.2  0.221  1.5  0.6, 3.8  0.347  1.5  0.6, 3.8  0.374  Abbreviations: CI, confidence interval; RR, risk ratio; T/S ratio, ratio of telomeric products (T) to single-copy genes (S); WMH, white matter hyperintensities. a Model 1 adjusted for age, sex, and study site. b Model 2 adjusted for model 1 variables plus alcohol, smoking, C-reactive protein, and low-density lipoprotein cholesterol. c Model 3 adjusted for model 2 variables plus education, income, and body mass index. d Volumes are expressed as percentage of intracranial volume. Models including adjustment and interaction terms for potential effect modifiers revealed some evidence for interaction with stage of kidney disease, but not with diabetes or hypertension (Table 5). Significant P values for the interaction term suggested that white matter grade, infarcts, and brain volume might be mediated by the presence of the most extreme categories of CKD—stage 4 (glomerular filtration rate <30 mL/minute) and stage 5, also known as end-stage renal disease (glomerular filtration rate <15 mL/minute). Nevertheless, statistical power was insufficient to conduct stratified analyses by CKD stage, since only 8 participants altogether were characterized as stage 4 or 5. Table 5. P Values From Tests of Interaction With Comorbidity in Regression Analyses of the Association Between Leukocyte Telomere Length and Findings From Cranial Magnetic Resonance Imaging Among American Indians, United States, 2010–2013a Type of MRI Finding  Comorbid Condition  Chronic Kidney Diseaseb  Hypertensionc  Diabetesc  Stage 2  Stage 3  Stages 4 and 5  White matter grade  0.351  0.567  0.636  0.492  0.632  Sulcal grade  0.930  0.134  0.536  0.335  0.577  Ventricle grade  0.846  0.249  0.407  0.474  0.139  Any infarcts  0.389  0.660  <0.001  0.674  0.543   Lacunar infarcts  0.556  0.933  <0.001  0.755  0.635  Hemorrhages  0.287  0.968  0.023  0.204  0.929  WMH volumed  0.425  0.742  0.199  0.622  0.241  Hippocampal volumed  0.502  0.502  0.288  0.840  0.103  Brain volumed  0.263  0.734  0.025  0.924  0.627  Type of MRI Finding  Comorbid Condition  Chronic Kidney Diseaseb  Hypertensionc  Diabetesc  Stage 2  Stage 3  Stages 4 and 5  White matter grade  0.351  0.567  0.636  0.492  0.632  Sulcal grade  0.930  0.134  0.536  0.335  0.577  Ventricle grade  0.846  0.249  0.407  0.474  0.139  Any infarcts  0.389  0.660  <0.001  0.674  0.543   Lacunar infarcts  0.556  0.933  <0.001  0.755  0.635  Hemorrhages  0.287  0.968  0.023  0.204  0.929  WMH volumed  0.425  0.742  0.199  0.622  0.241  Hippocampal volumed  0.502  0.502  0.288  0.840  0.103  Brain volumed  0.263  0.734  0.025  0.924  0.627  Abbreviations: CKD, chronic kidney disease; WMH, white matter hyperintensity. a Results were adjusted for age, sex, study site, smoking, alcohol use, low-density lipoprotein cholesterol, C-reactive protein, education, income, and body mass index. b Models comparing the given CKD stage with CKD stage 1. c Models comparing the presence of the given clinical condition with its absence. d Volumes are expressed as percentage of intracranial volume. Table 5. P Values From Tests of Interaction With Comorbidity in Regression Analyses of the Association Between Leukocyte Telomere Length and Findings From Cranial Magnetic Resonance Imaging Among American Indians, United States, 2010–2013a Type of MRI Finding  Comorbid Condition  Chronic Kidney Diseaseb  Hypertensionc  Diabetesc  Stage 2  Stage 3  Stages 4 and 5  White matter grade  0.351  0.567  0.636  0.492  0.632  Sulcal grade  0.930  0.134  0.536  0.335  0.577  Ventricle grade  0.846  0.249  0.407  0.474  0.139  Any infarcts  0.389  0.660  <0.001  0.674  0.543   Lacunar infarcts  0.556  0.933  <0.001  0.755  0.635  Hemorrhages  0.287  0.968  0.023  0.204  0.929  WMH volumed  0.425  0.742  0.199  0.622  0.241  Hippocampal volumed  0.502  0.502  0.288  0.840  0.103  Brain volumed  0.263  0.734  0.025  0.924  0.627  Type of MRI Finding  Comorbid Condition  Chronic Kidney Diseaseb  Hypertensionc  Diabetesc  Stage 2  Stage 3  Stages 4 and 5  White matter grade  0.351  0.567  0.636  0.492  0.632  Sulcal grade  0.930  0.134  0.536  0.335  0.577  Ventricle grade  0.846  0.249  0.407  0.474  0.139  Any infarcts  0.389  0.660  <0.001  0.674  0.543   Lacunar infarcts  0.556  0.933  <0.001  0.755  0.635  Hemorrhages  0.287  0.968  0.023  0.204  0.929  WMH volumed  0.425  0.742  0.199  0.622  0.241  Hippocampal volumed  0.502  0.502  0.288  0.840  0.103  Brain volumed  0.263  0.734  0.025  0.924  0.627  Abbreviations: CKD, chronic kidney disease; WMH, white matter hyperintensity. a Results were adjusted for age, sex, study site, smoking, alcohol use, low-density lipoprotein cholesterol, C-reactive protein, education, income, and body mass index. b Models comparing the given CKD stage with CKD stage 1. c Models comparing the presence of the given clinical condition with its absence. d Volumes are expressed as percentage of intracranial volume. Because the Southern Plains site showed a steeper decline in LTL with advancing age, secondary analyses were stratified by study site (Web Table 2). Findings demonstrated that the primary results for ventricle grade and WMH volume were concentrated among these participants. Newly significant associations were also detected for white matter grade at the Southwest site (incidence rate ratio = 0.76, 95% CI: 0.62, 0.93; P = 0.007), with an interpretation similar to that of the primary models, as discussed above. DISCUSSION We conducted cranial MRI in a multiregional sample of elderly American Indians to assess the association of LTL with MRI-defined brain disease. Independent of sociodemographic factors, key health behaviors, and clinical comorbidity, we found a significant association between shorter LTL and the presence of ventricle enlargement and extensive WMH. These results suggest that accelerated cellular aging, as measured by shortened telomeres, may reflect the burden of WMH, white matter disease, and ventricle enlargement, independently of other predictors of brain disease. Morbidity was most extreme among participants whose LTL revealed the steepest declines with advancing age. Previous studies have also found associations between VBI in older adults and extreme LTL attrition, WMH, and brain atrophy (13, 14). However, to our knowledge, our findings are the first to define such associations in an understudied minority population with reference to a broad range of MRI findings of VBI and brain atrophy. Kidney dysfunction might be an effect modifier for these associations. Accounting for interactions between LTL and stage of kidney disease revealed a potential association between LTL and the presence of white matter disease, infarcts, and measures of cerebral atrophy for participants with CKD stages 4 and 5. However, our stratified analyses by stage of CKD had limited statistical power because of small numbers in certain strata. Therefore, evidence for effect modification should be interpreted with caution. Nevertheless, these preliminary indications call for closer examination of brain aging in the context of CKD and other conditions causing or closely correlating with loss of kidney function. A key limitation of our study was the cross-sectional nature of data collection, which involved only 1 measure of LTL in 1989–1991 and 1 cranial MRI in 2010–2013. This limitation has 3 important implications. First, we cannot infer a temporal sequence from these data. Nevertheless, VBI or cerebral atrophy is unlikely to be the cause of telomere shortening, although an unknown causal factor might account for both the LTL measurements and the MRI results. Moreover, our findings take the form of prevalence ratios. We cannot interpret these estimates as reflecting associations between LTL and the risk of subsequent VBI, because the baseline examination in 1989–1991 could not detect prevalent subclinical disease. Therefore, we had no way to ensure that participants who might have experienced such disease at baseline were excluded from our analytical sample. Second, with only a single measure of LTL, we could not differentiate participants who had short telomeres at birth from those who experienced high rates of attrition over their lifetimes. However, this source of measurement error is likely to have been nondifferential and thus to have contributed to the risk of type II error, or the possibility of false-negative findings. In addition, telomeres are rarely measured in newborns, so any clinical implications of cross-sectional telomere measurements in later life will likely be subject to similar inferential limitations. Even if clinicians cannot determine the source of observed LTL shortening, comparisons may still be made with a norm or standard; in such cases, LTL could function as a useful biomarker of premature cellular aging. Third, the length of time elapsed since the initial LTL measurement might have resulted in survivorship bias. For example, telomeres might have continued to deteriorate after measurement in 1989–1991. If we assume that LTL attrition is causal, participants with shorter LTL in the earlier examination might have experienced more extreme aging-related disease or mortality, ensuring their absence from the participant pool for the present study. This scenario would have differentially removed the most extreme cases, thereby attenuating the discoverability of our results, contributing further to type II error and leading to false-negative findings. In such a case, our current findings would actually be an underestimate. Notably, we had inadequate statistical power to incorporate inverse probability weighting in our analytical models, so we could not establish whether additional MRI findings might be implicated in the observed association with LTL. Research involving a shorter elapsed time between LTL assay and MRI examination would be more conclusive in this respect. Our findings indicate a potential relationship linking accelerated cellular aging to VBI and central (and possibly cortical) brain atrophy. Whether this association is related to cognitive decline or functional loss remains to be determined. To advance pathological models of aging-related brain disease and its sequelae, future research should examine neurocognitive and functional outcomes in relation to telomere length. Although LTL is a well-supported marker of cellular aging and inflammatory processes, it is not a proxy for systemic aging. Our results suggest that telomere attrition might reflect more advanced brain pathologies, but additional research is needed to extend our understanding beyond the leukocyte to the potential contribution of allopathic load or the utility of some composite score of chronic disease and aging. By assessing telomere length and perhaps kidney function, such a score might be useful in better prediction of premature aging. Our sociodemographic data show that American Indians in the CDCAI sample were more likely to report a college education than previous subsets of the Strong Heart Study cohort described in the literature (26). This outcome might have resulted from some form of selection bias. Perhaps college-educated American Indians were more inclined to consent to genetic testing (a requirement of CDCAI Study participation) than those with lower educational levels; or perhaps more education was correlated with better health and longer life, resulting in a sample of survivors who were better educated than their peers who passed away. However, adjustment for education did not affect our analytical results, so the available evidence does not indicate a major role for education in our findings. The observed differences in LTL among study sites might have been due to age. The ratios of women to men were similar at all sites, but mean age differed by about 1.5 years across sites, and the Southern Plains site showed a steeper decline in LTL with advancing age than did the Northern Plains site (data not shown). The observed nonlinear associations between LTL and older age (U-shape) and CKD stage and the unexpected directionality of the association with body mass index categories may also reflect some contribution of survival or assay availability selection in this cohort. SHS participants who survived from study baseline (1989–1991) and who also participated in the CDCAI Study (2010–2013) were, on average, 6 years younger and less likely to be male than those who only participated in the SHS baseline examinations (Web Table 3). SHS survivors were also slightly more likely to be current alcohol users or to have dyslipidemia and less likely to have diabetes and diabetic nephropathy or hypertension at SHS baseline. Finally, SHS survivors were less likely to have had a stroke between the baseline and CDCAI examinations, but there was no difference in the likelihood of myocardial infarction. This study represents, to our knowledge, the first consideration of potential associations linking telomere length with the presence of WMH and ventricular enlargement in a sample of American Indians. It also offers the first indication that renal function might be an effect modifier for these associations. Our findings can help to clarify the relationship between cell senescence and the risk and severity of brain disease in American Indians, and by extension in other minority populations with a similar burden of risk factors. Directions for future research include examining the interaction of telomere shortening with loss of cognitive and physical function and determining the resulting implications for interventions to prevent premature aging. ACKNOWLEDGMENTS Author affiliations: Initiative for Research and Education to Advance Community Health, Elson S. Floyd College of Medicine, Washington State University, Seattle, Washington (Astrid M. Suchy-Dicey, Clemma J. Muller, Dedra Buchwald); Department of Radiology, School of Medicine, University of Washington, Seattle, Washington (Tara M. Madhyastha, Dean Shibata); Texas Biomedical Research Institute, San Antonio, Texas (Shelley A. Cole); Department of Epidemiology, College of Public Health and Health Professions, University of Florida, Gainesville, Florida (Jinying Zhao); Department of Neurology, School of Medicine, University of Washington, Seattle, Washington (W. T. Longstreth, Jr.); and Department of Epidemiology, School of Public Health, University of Washington, Seattle, Washington (W. T. Longstreth, Jr.). This work was supported by funding from the National Heart, Lung, and Blood Institute (grants U01HL41642, U01HL41652, U01HL41654, U01HL65520, U01HL65521, R01HL109315, R01HL109301, R01HL109284, R01HL109282, R01HL109319, and R01HL093086) and the National Institute on Aging (grant P50AG005136). We thank all of the Strong Heart Study participants and communities. The opinions expressed in this paper are those of the author(s) and do not necessarily reflect the views of the Indian Health Service. Conflict of interest: none declared. Abbreviations CDCAI Cerebrovascular Disease and Its Consequences in American Indians CI confidence interval CKD chronic kidney disease LTL leukocyte telomere length MRI magnetic resonance imaging VBI vascular brain injury WMH white matter hyperintensities REFERENCES 1 Aviv A. Leukocyte telomere length, hypertension, and atherosclerosis: are there potential mechanistic explanations? Hypertension . 2009; 53( 4): 590– 591. Google Scholar CrossRef Search ADS PubMed  2 Eitan E, Hutchison ER, Mattson MP. Telomere shortening in neurological disorders: an abundance of unanswered questions. Trends Neurosci . 2014; 37( 5): 256– 263. Google Scholar CrossRef Search ADS PubMed  3 Kappei D, Londono-Vallejo JA. Telomere length inheritance and aging. Mech Ageing Dev . 2008; 129( 1-2): 17– 26. Google Scholar CrossRef Search ADS PubMed  4 Zhao J, Miao K, Wang H, et al.  . Association between telomere length and type 2 diabetes mellitus: a meta-analysis. PLoS One . 2013; 8( 11): e79993. Google Scholar CrossRef Search ADS PubMed  5 Fitzpatrick AL, Kronmal RA, Kimura M, et al.  . Leukocyte telomere length and mortality in the Cardiovascular Health Study. J Gerontol A Biol Sci Med Sci . 2011; 66( 4): 421– 429. Google Scholar CrossRef Search ADS PubMed  6 Hammadah M, Al Mheid I, Wilmot K, et al.  . Telomere shortening, regenerative capacity, and cardiovascular outcomes. Circ Res . 2017; 120( 7): 1130– 1138. Google Scholar CrossRef Search ADS PubMed  7 Peng H, Mete M, Desale S, et al.  . Leukocyte telomere length and ideal cardiovascular health in American Indians: the Strong Heart Family Study. Eur J Epidemiol . 2017; 32( 1): 67– 75. Google Scholar CrossRef Search ADS PubMed  8 Fitzpatrick AL, Kronmal RA, Gardner JP, et al.  . Leukocyte telomere length and cardiovascular disease in the Cardiovascular Health Study. Am J Epidemiol . 2007; 165( 1): 14– 21. Google Scholar CrossRef Search ADS PubMed  9 Ding H, Chen C, Shaffer JR, et al.  . Telomere length and risk of stroke in Chinese. Stroke . 2012; 43( 3): 658– 663. Google Scholar CrossRef Search ADS PubMed  10 Willeit P, Willeit J, Brandstätter A, et al.  . Cellular aging reflected by leukocyte telomere length predicts advanced atherosclerosis and cardiovascular disease risk. Arterioscler Thromb Vasc Biol . 2010; 30( 8): 1649– 1656. Google Scholar CrossRef Search ADS PubMed  11 Zee RY, Castonguay AJ, Barton NS, et al.  . Relative leukocyte telomere length and risk of incident ischemic stroke in men: a prospective, nested case-control approach. Rejuvenation Res . 2010; 13( 4): 411– 414. Google Scholar CrossRef Search ADS PubMed  12 Fyhrquist F, Silventoinen K, Saijonmaa O, et al.  . Telomere length and cardiovascular risk in hypertensive patients with left ventricular hypertrophy: the LIFE Study. J Hum Hypertens . 2011; 25( 12): 711– 718. Google Scholar CrossRef Search ADS PubMed  13 Wikgren M, Karlsson T, Söderlund H, et al.  . Shorter telomere length is linked to brain atrophy and white matter hyperintensities. Age Ageing . 2014; 43( 2): 212– 217. Google Scholar CrossRef Search ADS PubMed  14 King KS, Kozlitina J, Rosenberg RN, et al.  . Effect of leukocyte telomere length on total and regional brain volumes in a large population-based cohort. JAMA Neurol . 2014; 71( 10): 1247– 1254. Google Scholar CrossRef Search ADS PubMed  15 Kota LN, Bharath S, Purushottam M, et al.  . Reduced telomere length in neurodegenerative disorders may suggest shared biology. J Neuropsychiatry Clin Neurosci . 2015; 27( 2): e92– e96. Google Scholar CrossRef Search ADS PubMed  16 Liu M, Huo YR, Wang J, et al.  . Telomere shortening in Alzheimer’s disease patients. Ann Clin Lab Sci . 2016; 46( 3): 260– 265. Google Scholar PubMed  17 Cai Z, Yan LJ, Ratka A. Telomere shortening and Alzheimer’s disease. Neuromolecular Med . 2013; 15( 1): 25– 48. Google Scholar CrossRef Search ADS PubMed  18 Schürks M, Buring J, Dushkes R, et al.  . Telomere length and Parkinson’s disease in men: a nested case-control study. Eur J Neurol . 2014; 21( 1): 93– 99. Google Scholar CrossRef Search ADS PubMed  19 Endicott AA, Taylor JW, Walsh KM. Telomere length connects melanoma and glioma predispositions. Aging (Albany NY) . 2016; 8( 3): 423– 424. Google Scholar CrossRef Search ADS PubMed  20 Willeit P, Raschenberger J, Heydon EE, et al.  . Leucocyte telomere length and risk of type 2 diabetes mellitus: new prospective cohort study and literature-based meta-analysis. PLoS One . 2014; 9( 11): e112483. Google Scholar CrossRef Search ADS PubMed  21 Peng H, Zhu Y, Yeh F, et al.  . Impact of biological aging on arterial aging in American Indians: findings from the Strong Heart Family Study. Aging (Albany NY) . 2016; 8( 8): 1583– 1592. Google Scholar CrossRef Search ADS PubMed  22 Haycock PC, Heydon EE, Kaptoge S, et al.  . Leucocyte telomere length and risk of cardiovascular disease: systematic review and meta-analysis. BMJ . 2014; 349: g4227. Google Scholar CrossRef Search ADS PubMed  23 Harwell TS, Oser CS, Okon NJ, et al.  . Defining disparities in cardiovascular disease for American Indians: trends in heart disease and stroke mortality among American Indians and whites in Montana, 1991 to 2000. Circulation . 2005; 112( 15): 2263– 2267. Google Scholar CrossRef Search ADS PubMed  24 Rosamond W, Flegal K, Friday G, et al.  . Heart disease and stroke statistics—2007 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation . 2007; 115( 5): e69– e171. Google Scholar CrossRef Search ADS PubMed  25 Stansbury JP, Jia H, Williams LS, et al.  . Ethnic disparities in stroke: epidemiology, acute care, and postacute outcomes. Stroke . 2005; 36( 2): 374– 386. Google Scholar CrossRef Search ADS PubMed  26 Lee ET, Welty TK, Fabsitz R, et al.  . The Strong Heart Study. A study of cardiovascular disease in American Indians: design and methods. Am J Epidemiol . 1990; 132( 6): 1141– 1155. Google Scholar CrossRef Search ADS PubMed  27 Suchy-Dicey AM, Shibata D, Best LG, et al.  . Cranial magnetic resonance imaging in elderly American Indians: design, methods, and implementation of the Cerebrovascular Disease and Its Consequences in American Indians Study. Neuroepidemiology . 2016; 47( 2): 67– 75. Google Scholar CrossRef Search ADS PubMed  28 Suchy-Dicey AM, Shibata DK, Madhyastha TM, et al.  . Findings of vascular brain injury and structural loss from cranial magnetic resonance imaging in elderly American Indians: the Strong Heart Study. Neuroepidemiology . 2017; 48( 1–2): 39– 47. Google Scholar CrossRef Search ADS PubMed  29 Zhao J, Zhu Y, Lin J, et al.  . Short leukocyte telomere length predicts risk of diabetes in American Indians: the Strong Heart Family Study. Diabetes . 2014; 63( 1): 354– 362. Google Scholar CrossRef Search ADS PubMed  30 Inker LA, Shaffi K, Levey AS. Estimating glomerular filtration rate using the chronic kidney disease-epidemiology collaboration creatinine equation: better risk predictions. Circ Heart Fail . 2012; 5( 3): 303– 306. Google Scholar CrossRef Search ADS PubMed  31 Inker LA, Schmid CH, Tighiouart H, et al.  . Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med . 2012; 367( 1): 20– 29. Google Scholar CrossRef Search ADS PubMed  32 Zhu Y, Voruganti VS, Lin J, et al.  . QTL mapping of leukocyte telomere length in American Indians: the Strong Heart Family Study. Aging (Albany NY) . 2013; 5( 9): 704– 716. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png American Journal of Epidemiology Oxford University Press

Telomere Length and Magnetic Resonance Imaging Findings of Vascular Brain Injury and Central Brain Atrophy: The Strong Heart Study

Loading next page...
 
/lp/ou_press/telomere-length-and-magnetic-resonance-imaging-findings-of-vascular-39GC8Bzd6R
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
0002-9262
eISSN
1476-6256
D.O.I.
10.1093/aje/kwx368
Publisher site
See Article on Publisher Site

Abstract

Abstract Telomeres are repeating regions of DNA that cap chromosomes. They shorten over the mammalian life span, especially in the presence of oxidative stress and inflammation. Telomeres may play a direct role in cell senescence, serving as markers of premature vascular aging. Leukocyte telomere length (LTL) may be associated with premature vascular brain injury and cerebral atrophy. However, reports have been inconsistent, especially among minority populations with a heavy burden of illness related to vascular aging. We examined associations between LTL and magnetic resonance imaging in 363 American Indians aged 64–93 years from the Strong Heart Study (1989–1991) and its ancillary study, Cerebrovascular Disease and Its Consequences in American Indians (2010–2013). Our results showed significant associations of LTL with ventricular enlargement and the presence of white matter hyperintensities. Secondary models indicated that renal function may mediate these associations, although small case numbers limited inference. Hypertension and diabetes showed little evidence of effect modification. Results were most extreme among participants who evinced the largest decline in LTL. Although this study was limited to cross-sectional comparisons, it represents (to our knowledge) the first consideration of associations between telomere length and brain aging in American Indians. Findings suggest a relationship between vascular aging by cell senescence and severity of brain disease. aging, brain, central brain atrophy, magnetic resonance imaging, telomere length, vascular aging, vascular brain injury Telomeres are DNA regions comprising thousands of TTAGGG repeats that protect the ends of mammalian chromosomes from shortening during mitotic cell replication (1). Telomere shortening reflects the inability of replication enzymes to copy lagging ends of DNA strands, as well as the presence of arbitrary DNA damage from oxidative stress and inflammation (2). Such attrition eventually triggers cell senescence, a state in which the cell can no longer replicate and ultimately dies (1). Therefore, measures of leukocyte telomere length (LTL) reflect the accumulated cellular impact of stress and inflammation. Shorter LTL and more rapid LTL loss independently predict cardiovascular disease morbidity and mortality, as well as incident diabetes (3–7). Data from the Cardiovascular Health Study suggest that a difference of approximately 1,000 base pairs in LTL is associated with a 3-fold increase in risk of myocardial infarction, even after accounting for conventional risk factors (8). Attrition of LTL may also be associated with increased risk of vascular brain injury (VBI). Studies examining LTL in relation to overt VBI, such as stroke, have produced inconsistent results (2, 5, 8–12). One study with 100 Europeans identified a potential association between shorter LTL and measures of cortical atrophy and white matter hyperintensities (WMH), although statistical power and generalizability were limited (13). A larger study with white, black, and Hispanic participants demonstrated that LTL was associated with brain and hippocampal volumes, including white matter volume, with shorter telomeres corresponding to smaller volumes (14). Other studies have established associations between telomere attrition and dementia (15), Alzheimer disease (16, 17), Parkinson disease (18), Huntington disease (15), adult glioma (19), diabetes (20), arterial stiffness (21), and cardiovascular disease (6, 7, 22). However, no prior study has examined the association of LTL attrition with subclinical VBI in a minority population with a high incidence of cerebrovascular disease; nor has any research with a health disparities population assessed a wide range of findings from cranial magnetic resonance imaging (MRI)—including WMH, white matter grade, infarcts, hemorrhages, sulcal widening, ventricle enlargement, overall brain volume, and hippocampal volume. American Indians experience a particularly high incidence and prevalence of diseases related to vascular aging, such as coronary heart disease, stroke, hypertension, and diabetes (23–25). Examining LTL in relation to a range of measures of covert, imaging-defined VBI in this population might therefore offer more sensitive etiological insights than existing studies. In addition, given the possibility that many cerebral morbidities related to telomere attrition might share a similar biological mechanism, studies of the factors associated with degenerative and vascular brain abnormalities could elucidate important mechanistic pathologies. The results of such work could inform future preventive or therapeutic research among American Indians and other health disparities populations with similar exposure patterns. Therefore, we conducted a cross-sectional study to investigate whether VBI and cerebral atrophy, as measured by cranial MRI, were associated with shorter LTL in a community sample of American Indians. METHODS Setting The Strong Heart Study assembled the largest population-based cohort of elderly American Indians ever recruited, with 4,549 study participants enrolled between 1989 and 1991 (26). Participants were 45–74 years of age at enrollment and represented 13 tribal communities across 3 major geographic regions, including the Northern Plains, Southern Plains, and Southwest. In follow-up efforts, several waves of longitudinal data were collected over the next 2 decades. The Cerebrovascular Disease and Its Consequences in American Indians (CDCAI) Study, also known as the Strong Heart Stroke Study, was a follow-up examination of survivors from the original cohort conducted between 2010 and 2013. It involved the administration of cranial MRI and assessments of cognitive function (27). Details on methods used in the Strong Heart Study (26) and the CDCAI Study (27), including informed consent procedures and MRI examinations (28), have been previously published. Selection Eligibility criteria for the present study included 1) availability of LTL data from the Strong Heart Study and 2) availability of MRI data from the CDCAI Study. Thus, our cross-sectional sample included 465 eligible cohort survivors. Participants who reported prior stroke (n = 33) or heart disease (including heart attack and heart surgery; n = 70) at the CDCAI visit were excluded from analyses. This resulted in an analytical sample of 363 persons (1 excluded participant reported both stroke and heart disease). Data collection LTL was measured by quantitative polymerase chain reaction from blood samples collected at the baseline visit of the Strong Heart Study (1989–1991) (29). In brief, LTL was assessed as the ratio of telomeric products (T) to single-copy genes (S), yielding the T/S ratio. The rationale behind use of this method is that longer telomeres generate more product in a polymerase chain reaction that uses primers specific to telomeric DNA. The quantitative polymerase chain reaction intraassay coefficient of variation was 4.5% and the interassay coefficient of variation was 6.9%, indicating very good assay performance. Values of LTL greater than 3 standard deviations above the mean (n = 6) were considered technical artifacts; however, participants with these LTL values were already excluded from the analytical sample. Therefore, the final analytical sample comprised 363 persons, and continuous LTL (T/S ratio) was used in all inferential analyses. VBI and brain atrophy were defined on the basis of cranial MRI data (28). Using standardized protocols, study neuroradiologists scored MRI scans for the presence and location of infarcts (including lacunar type) and hemorrhages; graded the degree of sulcal loss, ventricle enlargement, and WMH, each on a 10-point scale from 0 (best) to 9 (worst); and estimated software-processed volumes of hippocampus, total WMH, and total brain mass, which were standardized as percentage of intracranial volume. Age (in years), sex, and study site (Northern Plains, Southern Plains, or Southwest) were assessed at the baseline visit of the Strong Heart Study in 1989–1991. During the CDCAI visit in 2010–2013, additional data were collected on sociodemographic characteristics (years of education, annual family income, Native (tribal) language speaking capacity), health behaviors (tobacco smoking history, alcohol use patterns), and clinical factors. Education, income, Native (tribal) language speaking capacity, smoking, and alcohol use were reported on a self-administered questionnaire. All study variables and categories are defined in Table 1. Table 1. Characteristics of American Indian Participants in a Study of Leukocyte Telomere Length and Brain Aging (n = 363), United States, 2010–2013 Characteristic  No. of Persons  Mean (SD) LTLa  Study site       Northern Plains  201  1.12 (0.46)   Southern Plains  120  1.39 (0.54)   Southwest  42  1.36 (0.58)  Sex       Female  263  1.28 (0.53)   Male  100  1.11 (0.46)  Age at examination, years       60–69  131  1.31 (0.52)   70–74  125  1.20 (0.51)   75–79  62  1.17 (0.51)   ≥80  45  1.23 (0.52)  Household income, dollars/year       <15,000  178  1.23 (0.50)   15,000–35,000  115  1.23 (0.50)   >35,000  67  1.28 (0.59)  Education       Some high school or less  67  1.16 (0.43)   High school graduation  91  1.24 (0.51)   Any college  144  1.25 (0.53)   College graduation  61  1.29 (0.58)  Native language speaking capacityb       Not at all  104  1.23 (0.51)   A little  112  1.29 (0.52)   Moderately  47  1.38 (0.58)   Very well  100  1.12 (0.46)  Alcohol usec       Never drinker  85  1.25 (0.50)   Former drinker  202  1.27 (0.54)   Light drinker  26  1.12 (0.50)   Current drinker  31  1.11 (0.45)  Lifetime cigarette smoking, pack-years       Never smoker  147  1.23 (0.52)   <10  84  1.26 (0.48)   10–25  63  1.18 (0.55)   >25  69  1.28 (0.52)  Body mass indexd       18–24 (normal)  60  1.20 (0.49)   25–29 (overweight)  108  1.25 (0.50)   ≥30 (obese)  192  1.25 (0.55)  Quartile of LDL cholesterol, mg/dL       <75  85  1.20 (0.47)   75–99  108  1.22 (0.52)   100–139  122  1.21 (0.53)   ≥140  43  1.43 (0.53)  Clinical hypertensione       Yes  283  1.25 (0.53)   No  80  1.19 (0.45)  Clinical diabetes mellitusf       Yes  169  1.26 (0.54)   No  194  1.21 (0.50)  Clinical chronic kidney diseaseg       Stage 1  48  1.16 (0.39)   Stage 2  222  1.23 (0.55)   Stage 3  85  1.31 (0.48)   Stage 4  6  1.36 (0.68)   Stage 5  2  1.06 (0.37)  Quartile of C-reactive protein, mg/L       <1.7  95  1.19 (0.42)   1.8–3.4  84  1.20 (0.53)   3.5–6.5  90  1.19 (0.54)   >6.5  89  1.33 (0.53)  Characteristic  No. of Persons  Mean (SD) LTLa  Study site       Northern Plains  201  1.12 (0.46)   Southern Plains  120  1.39 (0.54)   Southwest  42  1.36 (0.58)  Sex       Female  263  1.28 (0.53)   Male  100  1.11 (0.46)  Age at examination, years       60–69  131  1.31 (0.52)   70–74  125  1.20 (0.51)   75–79  62  1.17 (0.51)   ≥80  45  1.23 (0.52)  Household income, dollars/year       <15,000  178  1.23 (0.50)   15,000–35,000  115  1.23 (0.50)   >35,000  67  1.28 (0.59)  Education       Some high school or less  67  1.16 (0.43)   High school graduation  91  1.24 (0.51)   Any college  144  1.25 (0.53)   College graduation  61  1.29 (0.58)  Native language speaking capacityb       Not at all  104  1.23 (0.51)   A little  112  1.29 (0.52)   Moderately  47  1.38 (0.58)   Very well  100  1.12 (0.46)  Alcohol usec       Never drinker  85  1.25 (0.50)   Former drinker  202  1.27 (0.54)   Light drinker  26  1.12 (0.50)   Current drinker  31  1.11 (0.45)  Lifetime cigarette smoking, pack-years       Never smoker  147  1.23 (0.52)   <10  84  1.26 (0.48)   10–25  63  1.18 (0.55)   >25  69  1.28 (0.52)  Body mass indexd       18–24 (normal)  60  1.20 (0.49)   25–29 (overweight)  108  1.25 (0.50)   ≥30 (obese)  192  1.25 (0.55)  Quartile of LDL cholesterol, mg/dL       <75  85  1.20 (0.47)   75–99  108  1.22 (0.52)   100–139  122  1.21 (0.53)   ≥140  43  1.43 (0.53)  Clinical hypertensione       Yes  283  1.25 (0.53)   No  80  1.19 (0.45)  Clinical diabetes mellitusf       Yes  169  1.26 (0.54)   No  194  1.21 (0.50)  Clinical chronic kidney diseaseg       Stage 1  48  1.16 (0.39)   Stage 2  222  1.23 (0.55)   Stage 3  85  1.31 (0.48)   Stage 4  6  1.36 (0.68)   Stage 5  2  1.06 (0.37)  Quartile of C-reactive protein, mg/L       <1.7  95  1.19 (0.42)   1.8–3.4  84  1.20 (0.53)   3.5–6.5  90  1.19 (0.54)   >6.5  89  1.33 (0.53)  Abbreviations: eGFR, estimated glomerular filtration rate; LDL, low-density lipoprotein; LTL, leukocyte telomere length; SD, standard deviation. a LTL was assessed as the T/S ratio (ratio of telomeric products (T) to single-copy genes (S)). b Native (tribal) language speaking capacity was categorized according to responses on the questionnaire. c Categories of alcohol use were based on self-reported recency of use and typical use, as follows: never drinker; former drinker, with last drink more than 1 year before; light drinker, with last drink more than 1 month before; and current drinker, with last drink within the past month. d Weight (kg)/height (m)2. e Clinical hypertension was defined as systolic blood pressure/diastolic blood pressure ≥130/90 mm Hg or use of antihypertensive medication. f Clinical diabetes was defined as fasting glucose concentration ≥126 mg/dL or use of antidiabetic medication. g Chronic kidney disease was based on clinical criteria: stage 1, eGFR ≥90 mL/minute; stage 2, eGFR 60–89 mL/minute; stage 3, eGFR 30–59 mL/minute; stage 4, eGFR 15–29 mL/minute; stage 5, eGFR <15 mL/minute (end-stage renal disease). Table 1. Characteristics of American Indian Participants in a Study of Leukocyte Telomere Length and Brain Aging (n = 363), United States, 2010–2013 Characteristic  No. of Persons  Mean (SD) LTLa  Study site       Northern Plains  201  1.12 (0.46)   Southern Plains  120  1.39 (0.54)   Southwest  42  1.36 (0.58)  Sex       Female  263  1.28 (0.53)   Male  100  1.11 (0.46)  Age at examination, years       60–69  131  1.31 (0.52)   70–74  125  1.20 (0.51)   75–79  62  1.17 (0.51)   ≥80  45  1.23 (0.52)  Household income, dollars/year       <15,000  178  1.23 (0.50)   15,000–35,000  115  1.23 (0.50)   >35,000  67  1.28 (0.59)  Education       Some high school or less  67  1.16 (0.43)   High school graduation  91  1.24 (0.51)   Any college  144  1.25 (0.53)   College graduation  61  1.29 (0.58)  Native language speaking capacityb       Not at all  104  1.23 (0.51)   A little  112  1.29 (0.52)   Moderately  47  1.38 (0.58)   Very well  100  1.12 (0.46)  Alcohol usec       Never drinker  85  1.25 (0.50)   Former drinker  202  1.27 (0.54)   Light drinker  26  1.12 (0.50)   Current drinker  31  1.11 (0.45)  Lifetime cigarette smoking, pack-years       Never smoker  147  1.23 (0.52)   <10  84  1.26 (0.48)   10–25  63  1.18 (0.55)   >25  69  1.28 (0.52)  Body mass indexd       18–24 (normal)  60  1.20 (0.49)   25–29 (overweight)  108  1.25 (0.50)   ≥30 (obese)  192  1.25 (0.55)  Quartile of LDL cholesterol, mg/dL       <75  85  1.20 (0.47)   75–99  108  1.22 (0.52)   100–139  122  1.21 (0.53)   ≥140  43  1.43 (0.53)  Clinical hypertensione       Yes  283  1.25 (0.53)   No  80  1.19 (0.45)  Clinical diabetes mellitusf       Yes  169  1.26 (0.54)   No  194  1.21 (0.50)  Clinical chronic kidney diseaseg       Stage 1  48  1.16 (0.39)   Stage 2  222  1.23 (0.55)   Stage 3  85  1.31 (0.48)   Stage 4  6  1.36 (0.68)   Stage 5  2  1.06 (0.37)  Quartile of C-reactive protein, mg/L       <1.7  95  1.19 (0.42)   1.8–3.4  84  1.20 (0.53)   3.5–6.5  90  1.19 (0.54)   >6.5  89  1.33 (0.53)  Characteristic  No. of Persons  Mean (SD) LTLa  Study site       Northern Plains  201  1.12 (0.46)   Southern Plains  120  1.39 (0.54)   Southwest  42  1.36 (0.58)  Sex       Female  263  1.28 (0.53)   Male  100  1.11 (0.46)  Age at examination, years       60–69  131  1.31 (0.52)   70–74  125  1.20 (0.51)   75–79  62  1.17 (0.51)   ≥80  45  1.23 (0.52)  Household income, dollars/year       <15,000  178  1.23 (0.50)   15,000–35,000  115  1.23 (0.50)   >35,000  67  1.28 (0.59)  Education       Some high school or less  67  1.16 (0.43)   High school graduation  91  1.24 (0.51)   Any college  144  1.25 (0.53)   College graduation  61  1.29 (0.58)  Native language speaking capacityb       Not at all  104  1.23 (0.51)   A little  112  1.29 (0.52)   Moderately  47  1.38 (0.58)   Very well  100  1.12 (0.46)  Alcohol usec       Never drinker  85  1.25 (0.50)   Former drinker  202  1.27 (0.54)   Light drinker  26  1.12 (0.50)   Current drinker  31  1.11 (0.45)  Lifetime cigarette smoking, pack-years       Never smoker  147  1.23 (0.52)   <10  84  1.26 (0.48)   10–25  63  1.18 (0.55)   >25  69  1.28 (0.52)  Body mass indexd       18–24 (normal)  60  1.20 (0.49)   25–29 (overweight)  108  1.25 (0.50)   ≥30 (obese)  192  1.25 (0.55)  Quartile of LDL cholesterol, mg/dL       <75  85  1.20 (0.47)   75–99  108  1.22 (0.52)   100–139  122  1.21 (0.53)   ≥140  43  1.43 (0.53)  Clinical hypertensione       Yes  283  1.25 (0.53)   No  80  1.19 (0.45)  Clinical diabetes mellitusf       Yes  169  1.26 (0.54)   No  194  1.21 (0.50)  Clinical chronic kidney diseaseg       Stage 1  48  1.16 (0.39)   Stage 2  222  1.23 (0.55)   Stage 3  85  1.31 (0.48)   Stage 4  6  1.36 (0.68)   Stage 5  2  1.06 (0.37)  Quartile of C-reactive protein, mg/L       <1.7  95  1.19 (0.42)   1.8–3.4  84  1.20 (0.53)   3.5–6.5  90  1.19 (0.54)   >6.5  89  1.33 (0.53)  Abbreviations: eGFR, estimated glomerular filtration rate; LDL, low-density lipoprotein; LTL, leukocyte telomere length; SD, standard deviation. a LTL was assessed as the T/S ratio (ratio of telomeric products (T) to single-copy genes (S)). b Native (tribal) language speaking capacity was categorized according to responses on the questionnaire. c Categories of alcohol use were based on self-reported recency of use and typical use, as follows: never drinker; former drinker, with last drink more than 1 year before; light drinker, with last drink more than 1 month before; and current drinker, with last drink within the past month. d Weight (kg)/height (m)2. e Clinical hypertension was defined as systolic blood pressure/diastolic blood pressure ≥130/90 mm Hg or use of antihypertensive medication. f Clinical diabetes was defined as fasting glucose concentration ≥126 mg/dL or use of antidiabetic medication. g Chronic kidney disease was based on clinical criteria: stage 1, eGFR ≥90 mL/minute; stage 2, eGFR 60–89 mL/minute; stage 3, eGFR 30–59 mL/minute; stage 4, eGFR 15–29 mL/minute; stage 5, eGFR <15 mL/minute (end-stage renal disease). Clinical factors were measured by study staff or by laboratory assay from blood and urine samples collected during the CDCAI visit. These included body mass index (weight (kg)/height (m)2), C-reactive protein concentration (mg/L), low-density lipoprotein cholesterol concentration (mg/dL), systolic and diastolic blood pressure (mm Hg), fasting glucose concentration (mg/dL), and glomerular filtration rate (mL/minute), estimated by means of the 2012 Chronic Kidney Disease Epidemiology Collaboration equation (30, 31). Clinical conditions (kidney disease, hypertension, hyperlipidemia, diabetes) are shown in Table 1. Analyses We calculated summary statistics for LTL (T/S ratio) according to categories of selected participant characteristics as of the time of the CDCAI visit (n = 363). We calculated Spearman nonparametric coefficients (ρ) and P values for correlations between continuous LTL and continuous, graded, and binary MRI findings. We created box plots and scatterplots of continuous LTL in association with MRI findings, with horizontal axes chosen to maximize visual clarity. We constructed multivariable regression models and examined exponentiated (β) coefficients to estimate measures of risk between LTL (exposure) and MRI findings (outcome) by using Poisson regression for graded variables with incidence rate ratios, logistic regression for binary variables with odds ratios, and linear regression for volumetric variables with risk ratios. All point estimates are reported with 95% confidence intervals. We conducted adjustments for regression models in series. Model 1 included age, sex, and study site. Model 2 added alcohol, smoking, C-reactive protein, and low-density lipoprotein cholesterol. Model 3 added education, income, and body mass index. We structured these adjustment schematics to first adjust for a priori confounders based on their established associations with both exposure and outcome (model 1, model 2). Additional hypothetical confounders, with an established association with exposure but an unclear association with outcome (or vice versa), were added in model 3 in order to empirically evaluate the influence of adding these factors to the primary models, to avoid overfitting and unnecessary loss of degrees of freedom. Model 4 added adjustment and interaction terms for hypertension, diabetes mellitus, or stage of chronic kidney disease (CKD). The reasoning behind these structured adjustments was that age, sex, lipid levels, inflammation, alcohol use, and smoking are likely or known determinants of LTL, justifying their inclusion a priori as confounders. Education and income were also included a priori as proxy variables for socioeconomic status. Study site was included as well, based on possible site-specific differences in genetic control of LTL (32). Clinical factors such as hypertension, diabetes, and kidney disease are either hypothetical mediators (e.g., downstream of the clinical effects of LTL attrition) or confounders (e.g., upstream of the clinical effects of both LTL attrition and VBI) of the primary associations of interest. Therefore, each clinical factor was examined separately using a term for interaction with the exposure of interest (LTL). Secondary analyses stratified the models by study site because of significant age differences noted among participants across study sites. All statistical analyses were conducted with Stata, version 14 (StataCorp LP, College Station, Texas). RESULTS The study population was predominantly female and aged 70 years or older, with an overall age range of 64–93 years (Table 1). Annual household income was generally low, with nearly half of participants having an income below $15,000, although more than half of the participants had at least some college education. Many participants were bilingual, speaking both English (a cohort eligibility criterion) and their Native (tribal) language moderately well to very well. Most participants either never smoked or had limited smoking experience, and most either never used alcohol or had limited alcohol consumption. Obesity, diabetes, and hypertension were generally common. Low-density lipoprotein cholesterol concentration and CKD stage were predominantly within the normal range, but with large variance. Overall, analytical population characteristics were similar to those of the larger CDCAI Study population from which the analytical sample was drawn (27). The overall distribution of LTL (T/S ratio) was nearly normal, with a slight rightward skew, although there was some variability along the spectrum of values, probably due to small numbers (see Web Figure 1, available at https://academic.oup.com/aje). LTL was generally shorter for male, low-income, low-education, alcohol-using, hypertensive, and diabetic participants; LTL was generally longer for dyslipidemic and overweight or obese participants (Table 1). There was no clear relationship of LTL with bilingual status or smoking history, and there was possibly a nonlinear relationship with older age and stage of CKD. LTL measurements also revealed site-specific variations. The Northern Plains site had somewhat lower mean LTL values, the Southern Plains site had somewhat higher mean LTL values, and the Southwest site had intermediate LTL values. Spearman nonparametric correlation coefficients for correlations between LTL and MRI findings demonstrated significant negative associations between LTL and sulcal grade (ρ = −0.16), ventricle grade (ρ = −0.23), and WMH volume (ρ = −0.25) and positive associations with hippocampal volume (ρ = 0.12). Shorter telomeres corresponded to more extreme MRI findings for each factor (Web Table 1). Box plots of LTL according to WMH, sulcal, and ventricle grade (Figure 1, left column) showed similar trends, with lower LTL ranges among the higher grades. The trends for ventricle and sulcal grades were most evident. Small numbers in the highest grades led to imprecise estimates of the range of LTL. Scatterplots of LTL according to WMH, hippocampus, and total brain volumetric estimates (Figure 1, right column) repeated the results from correlation tests, with larger mean values and greater variability in WMH volumes corresponding to shorter LTL, but without any evidence of trend for hippocampal or total brain volumes. Figure 1. View largeDownload slide Association of leukocyte telomere length (LTL), defined as T/S ratio, with brain grades and volumes from cranial magnetic resonance imaging (MRI) in a study of LTL and brain aging among American Indians, United States, 2010–2013. Left-hand column: horizontal bar graphs of LTL according to graded findings from MRI, including white matter hyperintensity (WMH) grade (A), sulcal grade (C), and ventricle grade (E). (Bars, 95% confidence intervals.) Right-hand column: scatterplots with quadratic polynomial regression (solid lines) and 95% confidence intervals (shaded gray areas) for LTL (T/S ratio), with structural volumes defined by MRI as percentage of intracranial volume, including WMH volume (B), hippocampal volume (D), and total brain volume (F). T/S, ratio of telomeric products (T) to single-copy genes (S). Figure 1. View largeDownload slide Association of leukocyte telomere length (LTL), defined as T/S ratio, with brain grades and volumes from cranial magnetic resonance imaging (MRI) in a study of LTL and brain aging among American Indians, United States, 2010–2013. Left-hand column: horizontal bar graphs of LTL according to graded findings from MRI, including white matter hyperintensity (WMH) grade (A), sulcal grade (C), and ventricle grade (E). (Bars, 95% confidence intervals.) Right-hand column: scatterplots with quadratic polynomial regression (solid lines) and 95% confidence intervals (shaded gray areas) for LTL (T/S ratio), with structural volumes defined by MRI as percentage of intracranial volume, including WMH volume (B), hippocampal volume (D), and total brain volume (F). T/S, ratio of telomeric products (T) to single-copy genes (S). Regression models examining independent associations between LTL and MRI findings (Tables 2–4) demonstrated significant associations for ventricle grade (incidence rate ratio = 0.9, 95% confidence interval (CI): 0.8, 0.9; P = 0.003) and WMH volume (risk ratio = 0.9, 95% CI: 0.8, 1.0; P = 0.015). Results from models 1–3 were not substantively different, although the statistical tests for sulcal grade did not exclude chance as a potential explanation for the associations observed in models with the most adjustments. Interpretation of the direction and magnitude of these associations is as follows: Per unit longer LTL (in units of T/S ratio), ventricle and sulcal grades were 10% lower and WMH volumes were 10% smaller, independently of age, sex, study site, alcohol, smoking, low-density lipoprotein cholesterol, education, income, or body mass index. In more general terms: For each clinical factor, shorter telomeres were associated with more extreme MRI findings. Table 2. Associations Between Leukocyte Telomere Length (per Unit of T/S Ratio) and Graded Findings From Cranial Magnetic Resonance Imaging Among American Indians, United States, 2010–2013 Type of MRI Findinga  Model 1b  Model 2c  Model 3d  IRR  95% CI  P Value  IRR  95% CI  P Value  IRR  95% CI  P Value  White matter intensities  0.9  0.8, 1.0  0.201  0.9  0.8, 1.1  0.301  0.9  0.8, 1.1  0.327  Sulcal loss  0.9  0.9, 1.0  0.033  0.9  0.9, 1.0  0.056  0.9  0.9, 1.0  0.084  Ventricle enlargement  0.9  0.8, 0.9  0.001  0.9  0.8, 0.9  0.002  0.9  0.8, 0.9  0.003  Type of MRI Findinga  Model 1b  Model 2c  Model 3d  IRR  95% CI  P Value  IRR  95% CI  P Value  IRR  95% CI  P Value  White matter intensities  0.9  0.8, 1.0  0.201  0.9  0.8, 1.1  0.301  0.9  0.8, 1.1  0.327  Sulcal loss  0.9  0.9, 1.0  0.033  0.9  0.9, 1.0  0.056  0.9  0.9, 1.0  0.084  Ventricle enlargement  0.9  0.8, 0.9  0.001  0.9  0.8, 0.9  0.002  0.9  0.8, 0.9  0.003  Abbreviations: CI, confidence interval; IRR, incidence rate ratio; T/S ratio, ratio of telomeric products (T) to single-copy genes (S). a The degree of white matter intensities, sulcal loss, and ventricle enlargement was graded on a 10-point scale from 0 (best) to 9 (worst). b Model 1 adjusted for age, sex, and study site. c Model 2 adjusted for model 1 variables plus alcohol, smoking, C-reactive protein, and low-density lipoprotein cholesterol. d Model 3 adjusted for model 2 variables plus education, income, and body mass index. Table 2. Associations Between Leukocyte Telomere Length (per Unit of T/S Ratio) and Graded Findings From Cranial Magnetic Resonance Imaging Among American Indians, United States, 2010–2013 Type of MRI Findinga  Model 1b  Model 2c  Model 3d  IRR  95% CI  P Value  IRR  95% CI  P Value  IRR  95% CI  P Value  White matter intensities  0.9  0.8, 1.0  0.201  0.9  0.8, 1.1  0.301  0.9  0.8, 1.1  0.327  Sulcal loss  0.9  0.9, 1.0  0.033  0.9  0.9, 1.0  0.056  0.9  0.9, 1.0  0.084  Ventricle enlargement  0.9  0.8, 0.9  0.001  0.9  0.8, 0.9  0.002  0.9  0.8, 0.9  0.003  Type of MRI Findinga  Model 1b  Model 2c  Model 3d  IRR  95% CI  P Value  IRR  95% CI  P Value  IRR  95% CI  P Value  White matter intensities  0.9  0.8, 1.0  0.201  0.9  0.8, 1.1  0.301  0.9  0.8, 1.1  0.327  Sulcal loss  0.9  0.9, 1.0  0.033  0.9  0.9, 1.0  0.056  0.9  0.9, 1.0  0.084  Ventricle enlargement  0.9  0.8, 0.9  0.001  0.9  0.8, 0.9  0.002  0.9  0.8, 0.9  0.003  Abbreviations: CI, confidence interval; IRR, incidence rate ratio; T/S ratio, ratio of telomeric products (T) to single-copy genes (S). a The degree of white matter intensities, sulcal loss, and ventricle enlargement was graded on a 10-point scale from 0 (best) to 9 (worst). b Model 1 adjusted for age, sex, and study site. c Model 2 adjusted for model 1 variables plus alcohol, smoking, C-reactive protein, and low-density lipoprotein cholesterol. d Model 3 adjusted for model 2 variables plus education, income, and body mass index. Table 3. Associations Between Leukocyte Telomere Length (per Unit of T/S Ratio) and Lesion Findings From Cranial Magnetic Resonance Imaging Among American Indians, United States, 2010–2013 Type of MRI Finding  Model 1a  Model 2b  Model 3c  OR  95% CI  P Value  OR  95% CI  P Value  OR  95% CI  P Value  Any infarcts (≥3 mm)  0.7  0.5, 1.1  0.167  0.7  0.5, 1.2  0.215  0.7  0.4, 1.2  0.174  Lacunar infarcts  0.7  0.4, 1.2  0.205  0.8  0.5, 1.3  0.384  0.8  0.4, 1.3  0.319  Hemorrhages (any)  0.9  0.4, 2.3  0.851  1.0  0.4, 2.4  0.978  1.0  0.4, 2.5  0.968  Type of MRI Finding  Model 1a  Model 2b  Model 3c  OR  95% CI  P Value  OR  95% CI  P Value  OR  95% CI  P Value  Any infarcts (≥3 mm)  0.7  0.5, 1.1  0.167  0.7  0.5, 1.2  0.215  0.7  0.4, 1.2  0.174  Lacunar infarcts  0.7  0.4, 1.2  0.205  0.8  0.5, 1.3  0.384  0.8  0.4, 1.3  0.319  Hemorrhages (any)  0.9  0.4, 2.3  0.851  1.0  0.4, 2.4  0.978  1.0  0.4, 2.5  0.968  Abbreviations: CI, confidence interval; OR, odds ratio; T/S ratio, ratio of telomeric products (T) to single-copy genes (S). a Model 1 adjusted for age, sex, and study site. b Model 2 adjusted for model 1 variables plus alcohol, smoking, C-reactive protein, and low-density lipoprotein cholesterol. c Model 3 adjusted for model 2 variables plus education, income, and body mass index. Table 3. Associations Between Leukocyte Telomere Length (per Unit of T/S Ratio) and Lesion Findings From Cranial Magnetic Resonance Imaging Among American Indians, United States, 2010–2013 Type of MRI Finding  Model 1a  Model 2b  Model 3c  OR  95% CI  P Value  OR  95% CI  P Value  OR  95% CI  P Value  Any infarcts (≥3 mm)  0.7  0.5, 1.1  0.167  0.7  0.5, 1.2  0.215  0.7  0.4, 1.2  0.174  Lacunar infarcts  0.7  0.4, 1.2  0.205  0.8  0.5, 1.3  0.384  0.8  0.4, 1.3  0.319  Hemorrhages (any)  0.9  0.4, 2.3  0.851  1.0  0.4, 2.4  0.978  1.0  0.4, 2.5  0.968  Type of MRI Finding  Model 1a  Model 2b  Model 3c  OR  95% CI  P Value  OR  95% CI  P Value  OR  95% CI  P Value  Any infarcts (≥3 mm)  0.7  0.5, 1.1  0.167  0.7  0.5, 1.2  0.215  0.7  0.4, 1.2  0.174  Lacunar infarcts  0.7  0.4, 1.2  0.205  0.8  0.5, 1.3  0.384  0.8  0.4, 1.3  0.319  Hemorrhages (any)  0.9  0.4, 2.3  0.851  1.0  0.4, 2.4  0.978  1.0  0.4, 2.5  0.968  Abbreviations: CI, confidence interval; OR, odds ratio; T/S ratio, ratio of telomeric products (T) to single-copy genes (S). a Model 1 adjusted for age, sex, and study site. b Model 2 adjusted for model 1 variables plus alcohol, smoking, C-reactive protein, and low-density lipoprotein cholesterol. c Model 3 adjusted for model 2 variables plus education, income, and body mass index. Table 4. Associations Between Leukocyte Telomere Length (per Unit of T/S Ratio) and Volumetric Findings From Cranial Magnetic Resonance Imaging Among American Indians, United States, 2010–2013 Type of MRI Finding  Model 1a  Model 2b  Model 3c  RR  95% CI  P Value  RR  95% CI  P Value  RR  95% CI  P Value  WMH volumed  0.9  0.8, 1.0  0.005  0.9  0.8, 1.0  0.009  0.9  0.8, 1.0  0.015  Hippocampal volumed  1.0  1.0, 1.0  0.527  1.0  1.0, 1.0  0.441  1.0  1.0, 1.0  0.527  Brain volumed  1.7  0.7, 4.2  0.221  1.5  0.6, 3.8  0.347  1.5  0.6, 3.8  0.374  Type of MRI Finding  Model 1a  Model 2b  Model 3c  RR  95% CI  P Value  RR  95% CI  P Value  RR  95% CI  P Value  WMH volumed  0.9  0.8, 1.0  0.005  0.9  0.8, 1.0  0.009  0.9  0.8, 1.0  0.015  Hippocampal volumed  1.0  1.0, 1.0  0.527  1.0  1.0, 1.0  0.441  1.0  1.0, 1.0  0.527  Brain volumed  1.7  0.7, 4.2  0.221  1.5  0.6, 3.8  0.347  1.5  0.6, 3.8  0.374  Abbreviations: CI, confidence interval; RR, risk ratio; T/S ratio, ratio of telomeric products (T) to single-copy genes (S); WMH, white matter hyperintensities. a Model 1 adjusted for age, sex, and study site. b Model 2 adjusted for model 1 variables plus alcohol, smoking, C-reactive protein, and low-density lipoprotein cholesterol. c Model 3 adjusted for model 2 variables plus education, income, and body mass index. d Volumes are expressed as percentage of intracranial volume. Table 4. Associations Between Leukocyte Telomere Length (per Unit of T/S Ratio) and Volumetric Findings From Cranial Magnetic Resonance Imaging Among American Indians, United States, 2010–2013 Type of MRI Finding  Model 1a  Model 2b  Model 3c  RR  95% CI  P Value  RR  95% CI  P Value  RR  95% CI  P Value  WMH volumed  0.9  0.8, 1.0  0.005  0.9  0.8, 1.0  0.009  0.9  0.8, 1.0  0.015  Hippocampal volumed  1.0  1.0, 1.0  0.527  1.0  1.0, 1.0  0.441  1.0  1.0, 1.0  0.527  Brain volumed  1.7  0.7, 4.2  0.221  1.5  0.6, 3.8  0.347  1.5  0.6, 3.8  0.374  Type of MRI Finding  Model 1a  Model 2b  Model 3c  RR  95% CI  P Value  RR  95% CI  P Value  RR  95% CI  P Value  WMH volumed  0.9  0.8, 1.0  0.005  0.9  0.8, 1.0  0.009  0.9  0.8, 1.0  0.015  Hippocampal volumed  1.0  1.0, 1.0  0.527  1.0  1.0, 1.0  0.441  1.0  1.0, 1.0  0.527  Brain volumed  1.7  0.7, 4.2  0.221  1.5  0.6, 3.8  0.347  1.5  0.6, 3.8  0.374  Abbreviations: CI, confidence interval; RR, risk ratio; T/S ratio, ratio of telomeric products (T) to single-copy genes (S); WMH, white matter hyperintensities. a Model 1 adjusted for age, sex, and study site. b Model 2 adjusted for model 1 variables plus alcohol, smoking, C-reactive protein, and low-density lipoprotein cholesterol. c Model 3 adjusted for model 2 variables plus education, income, and body mass index. d Volumes are expressed as percentage of intracranial volume. Models including adjustment and interaction terms for potential effect modifiers revealed some evidence for interaction with stage of kidney disease, but not with diabetes or hypertension (Table 5). Significant P values for the interaction term suggested that white matter grade, infarcts, and brain volume might be mediated by the presence of the most extreme categories of CKD—stage 4 (glomerular filtration rate <30 mL/minute) and stage 5, also known as end-stage renal disease (glomerular filtration rate <15 mL/minute). Nevertheless, statistical power was insufficient to conduct stratified analyses by CKD stage, since only 8 participants altogether were characterized as stage 4 or 5. Table 5. P Values From Tests of Interaction With Comorbidity in Regression Analyses of the Association Between Leukocyte Telomere Length and Findings From Cranial Magnetic Resonance Imaging Among American Indians, United States, 2010–2013a Type of MRI Finding  Comorbid Condition  Chronic Kidney Diseaseb  Hypertensionc  Diabetesc  Stage 2  Stage 3  Stages 4 and 5  White matter grade  0.351  0.567  0.636  0.492  0.632  Sulcal grade  0.930  0.134  0.536  0.335  0.577  Ventricle grade  0.846  0.249  0.407  0.474  0.139  Any infarcts  0.389  0.660  <0.001  0.674  0.543   Lacunar infarcts  0.556  0.933  <0.001  0.755  0.635  Hemorrhages  0.287  0.968  0.023  0.204  0.929  WMH volumed  0.425  0.742  0.199  0.622  0.241  Hippocampal volumed  0.502  0.502  0.288  0.840  0.103  Brain volumed  0.263  0.734  0.025  0.924  0.627  Type of MRI Finding  Comorbid Condition  Chronic Kidney Diseaseb  Hypertensionc  Diabetesc  Stage 2  Stage 3  Stages 4 and 5  White matter grade  0.351  0.567  0.636  0.492  0.632  Sulcal grade  0.930  0.134  0.536  0.335  0.577  Ventricle grade  0.846  0.249  0.407  0.474  0.139  Any infarcts  0.389  0.660  <0.001  0.674  0.543   Lacunar infarcts  0.556  0.933  <0.001  0.755  0.635  Hemorrhages  0.287  0.968  0.023  0.204  0.929  WMH volumed  0.425  0.742  0.199  0.622  0.241  Hippocampal volumed  0.502  0.502  0.288  0.840  0.103  Brain volumed  0.263  0.734  0.025  0.924  0.627  Abbreviations: CKD, chronic kidney disease; WMH, white matter hyperintensity. a Results were adjusted for age, sex, study site, smoking, alcohol use, low-density lipoprotein cholesterol, C-reactive protein, education, income, and body mass index. b Models comparing the given CKD stage with CKD stage 1. c Models comparing the presence of the given clinical condition with its absence. d Volumes are expressed as percentage of intracranial volume. Table 5. P Values From Tests of Interaction With Comorbidity in Regression Analyses of the Association Between Leukocyte Telomere Length and Findings From Cranial Magnetic Resonance Imaging Among American Indians, United States, 2010–2013a Type of MRI Finding  Comorbid Condition  Chronic Kidney Diseaseb  Hypertensionc  Diabetesc  Stage 2  Stage 3  Stages 4 and 5  White matter grade  0.351  0.567  0.636  0.492  0.632  Sulcal grade  0.930  0.134  0.536  0.335  0.577  Ventricle grade  0.846  0.249  0.407  0.474  0.139  Any infarcts  0.389  0.660  <0.001  0.674  0.543   Lacunar infarcts  0.556  0.933  <0.001  0.755  0.635  Hemorrhages  0.287  0.968  0.023  0.204  0.929  WMH volumed  0.425  0.742  0.199  0.622  0.241  Hippocampal volumed  0.502  0.502  0.288  0.840  0.103  Brain volumed  0.263  0.734  0.025  0.924  0.627  Type of MRI Finding  Comorbid Condition  Chronic Kidney Diseaseb  Hypertensionc  Diabetesc  Stage 2  Stage 3  Stages 4 and 5  White matter grade  0.351  0.567  0.636  0.492  0.632  Sulcal grade  0.930  0.134  0.536  0.335  0.577  Ventricle grade  0.846  0.249  0.407  0.474  0.139  Any infarcts  0.389  0.660  <0.001  0.674  0.543   Lacunar infarcts  0.556  0.933  <0.001  0.755  0.635  Hemorrhages  0.287  0.968  0.023  0.204  0.929  WMH volumed  0.425  0.742  0.199  0.622  0.241  Hippocampal volumed  0.502  0.502  0.288  0.840  0.103  Brain volumed  0.263  0.734  0.025  0.924  0.627  Abbreviations: CKD, chronic kidney disease; WMH, white matter hyperintensity. a Results were adjusted for age, sex, study site, smoking, alcohol use, low-density lipoprotein cholesterol, C-reactive protein, education, income, and body mass index. b Models comparing the given CKD stage with CKD stage 1. c Models comparing the presence of the given clinical condition with its absence. d Volumes are expressed as percentage of intracranial volume. Because the Southern Plains site showed a steeper decline in LTL with advancing age, secondary analyses were stratified by study site (Web Table 2). Findings demonstrated that the primary results for ventricle grade and WMH volume were concentrated among these participants. Newly significant associations were also detected for white matter grade at the Southwest site (incidence rate ratio = 0.76, 95% CI: 0.62, 0.93; P = 0.007), with an interpretation similar to that of the primary models, as discussed above. DISCUSSION We conducted cranial MRI in a multiregional sample of elderly American Indians to assess the association of LTL with MRI-defined brain disease. Independent of sociodemographic factors, key health behaviors, and clinical comorbidity, we found a significant association between shorter LTL and the presence of ventricle enlargement and extensive WMH. These results suggest that accelerated cellular aging, as measured by shortened telomeres, may reflect the burden of WMH, white matter disease, and ventricle enlargement, independently of other predictors of brain disease. Morbidity was most extreme among participants whose LTL revealed the steepest declines with advancing age. Previous studies have also found associations between VBI in older adults and extreme LTL attrition, WMH, and brain atrophy (13, 14). However, to our knowledge, our findings are the first to define such associations in an understudied minority population with reference to a broad range of MRI findings of VBI and brain atrophy. Kidney dysfunction might be an effect modifier for these associations. Accounting for interactions between LTL and stage of kidney disease revealed a potential association between LTL and the presence of white matter disease, infarcts, and measures of cerebral atrophy for participants with CKD stages 4 and 5. However, our stratified analyses by stage of CKD had limited statistical power because of small numbers in certain strata. Therefore, evidence for effect modification should be interpreted with caution. Nevertheless, these preliminary indications call for closer examination of brain aging in the context of CKD and other conditions causing or closely correlating with loss of kidney function. A key limitation of our study was the cross-sectional nature of data collection, which involved only 1 measure of LTL in 1989–1991 and 1 cranial MRI in 2010–2013. This limitation has 3 important implications. First, we cannot infer a temporal sequence from these data. Nevertheless, VBI or cerebral atrophy is unlikely to be the cause of telomere shortening, although an unknown causal factor might account for both the LTL measurements and the MRI results. Moreover, our findings take the form of prevalence ratios. We cannot interpret these estimates as reflecting associations between LTL and the risk of subsequent VBI, because the baseline examination in 1989–1991 could not detect prevalent subclinical disease. Therefore, we had no way to ensure that participants who might have experienced such disease at baseline were excluded from our analytical sample. Second, with only a single measure of LTL, we could not differentiate participants who had short telomeres at birth from those who experienced high rates of attrition over their lifetimes. However, this source of measurement error is likely to have been nondifferential and thus to have contributed to the risk of type II error, or the possibility of false-negative findings. In addition, telomeres are rarely measured in newborns, so any clinical implications of cross-sectional telomere measurements in later life will likely be subject to similar inferential limitations. Even if clinicians cannot determine the source of observed LTL shortening, comparisons may still be made with a norm or standard; in such cases, LTL could function as a useful biomarker of premature cellular aging. Third, the length of time elapsed since the initial LTL measurement might have resulted in survivorship bias. For example, telomeres might have continued to deteriorate after measurement in 1989–1991. If we assume that LTL attrition is causal, participants with shorter LTL in the earlier examination might have experienced more extreme aging-related disease or mortality, ensuring their absence from the participant pool for the present study. This scenario would have differentially removed the most extreme cases, thereby attenuating the discoverability of our results, contributing further to type II error and leading to false-negative findings. In such a case, our current findings would actually be an underestimate. Notably, we had inadequate statistical power to incorporate inverse probability weighting in our analytical models, so we could not establish whether additional MRI findings might be implicated in the observed association with LTL. Research involving a shorter elapsed time between LTL assay and MRI examination would be more conclusive in this respect. Our findings indicate a potential relationship linking accelerated cellular aging to VBI and central (and possibly cortical) brain atrophy. Whether this association is related to cognitive decline or functional loss remains to be determined. To advance pathological models of aging-related brain disease and its sequelae, future research should examine neurocognitive and functional outcomes in relation to telomere length. Although LTL is a well-supported marker of cellular aging and inflammatory processes, it is not a proxy for systemic aging. Our results suggest that telomere attrition might reflect more advanced brain pathologies, but additional research is needed to extend our understanding beyond the leukocyte to the potential contribution of allopathic load or the utility of some composite score of chronic disease and aging. By assessing telomere length and perhaps kidney function, such a score might be useful in better prediction of premature aging. Our sociodemographic data show that American Indians in the CDCAI sample were more likely to report a college education than previous subsets of the Strong Heart Study cohort described in the literature (26). This outcome might have resulted from some form of selection bias. Perhaps college-educated American Indians were more inclined to consent to genetic testing (a requirement of CDCAI Study participation) than those with lower educational levels; or perhaps more education was correlated with better health and longer life, resulting in a sample of survivors who were better educated than their peers who passed away. However, adjustment for education did not affect our analytical results, so the available evidence does not indicate a major role for education in our findings. The observed differences in LTL among study sites might have been due to age. The ratios of women to men were similar at all sites, but mean age differed by about 1.5 years across sites, and the Southern Plains site showed a steeper decline in LTL with advancing age than did the Northern Plains site (data not shown). The observed nonlinear associations between LTL and older age (U-shape) and CKD stage and the unexpected directionality of the association with body mass index categories may also reflect some contribution of survival or assay availability selection in this cohort. SHS participants who survived from study baseline (1989–1991) and who also participated in the CDCAI Study (2010–2013) were, on average, 6 years younger and less likely to be male than those who only participated in the SHS baseline examinations (Web Table 3). SHS survivors were also slightly more likely to be current alcohol users or to have dyslipidemia and less likely to have diabetes and diabetic nephropathy or hypertension at SHS baseline. Finally, SHS survivors were less likely to have had a stroke between the baseline and CDCAI examinations, but there was no difference in the likelihood of myocardial infarction. This study represents, to our knowledge, the first consideration of potential associations linking telomere length with the presence of WMH and ventricular enlargement in a sample of American Indians. It also offers the first indication that renal function might be an effect modifier for these associations. Our findings can help to clarify the relationship between cell senescence and the risk and severity of brain disease in American Indians, and by extension in other minority populations with a similar burden of risk factors. Directions for future research include examining the interaction of telomere shortening with loss of cognitive and physical function and determining the resulting implications for interventions to prevent premature aging. ACKNOWLEDGMENTS Author affiliations: Initiative for Research and Education to Advance Community Health, Elson S. Floyd College of Medicine, Washington State University, Seattle, Washington (Astrid M. Suchy-Dicey, Clemma J. Muller, Dedra Buchwald); Department of Radiology, School of Medicine, University of Washington, Seattle, Washington (Tara M. Madhyastha, Dean Shibata); Texas Biomedical Research Institute, San Antonio, Texas (Shelley A. Cole); Department of Epidemiology, College of Public Health and Health Professions, University of Florida, Gainesville, Florida (Jinying Zhao); Department of Neurology, School of Medicine, University of Washington, Seattle, Washington (W. T. Longstreth, Jr.); and Department of Epidemiology, School of Public Health, University of Washington, Seattle, Washington (W. T. Longstreth, Jr.). This work was supported by funding from the National Heart, Lung, and Blood Institute (grants U01HL41642, U01HL41652, U01HL41654, U01HL65520, U01HL65521, R01HL109315, R01HL109301, R01HL109284, R01HL109282, R01HL109319, and R01HL093086) and the National Institute on Aging (grant P50AG005136). We thank all of the Strong Heart Study participants and communities. The opinions expressed in this paper are those of the author(s) and do not necessarily reflect the views of the Indian Health Service. Conflict of interest: none declared. Abbreviations CDCAI Cerebrovascular Disease and Its Consequences in American Indians CI confidence interval CKD chronic kidney disease LTL leukocyte telomere length MRI magnetic resonance imaging VBI vascular brain injury WMH white matter hyperintensities REFERENCES 1 Aviv A. Leukocyte telomere length, hypertension, and atherosclerosis: are there potential mechanistic explanations? Hypertension . 2009; 53( 4): 590– 591. Google Scholar CrossRef Search ADS PubMed  2 Eitan E, Hutchison ER, Mattson MP. Telomere shortening in neurological disorders: an abundance of unanswered questions. Trends Neurosci . 2014; 37( 5): 256– 263. Google Scholar CrossRef Search ADS PubMed  3 Kappei D, Londono-Vallejo JA. Telomere length inheritance and aging. Mech Ageing Dev . 2008; 129( 1-2): 17– 26. Google Scholar CrossRef Search ADS PubMed  4 Zhao J, Miao K, Wang H, et al.  . Association between telomere length and type 2 diabetes mellitus: a meta-analysis. PLoS One . 2013; 8( 11): e79993. Google Scholar CrossRef Search ADS PubMed  5 Fitzpatrick AL, Kronmal RA, Kimura M, et al.  . Leukocyte telomere length and mortality in the Cardiovascular Health Study. J Gerontol A Biol Sci Med Sci . 2011; 66( 4): 421– 429. Google Scholar CrossRef Search ADS PubMed  6 Hammadah M, Al Mheid I, Wilmot K, et al.  . Telomere shortening, regenerative capacity, and cardiovascular outcomes. Circ Res . 2017; 120( 7): 1130– 1138. Google Scholar CrossRef Search ADS PubMed  7 Peng H, Mete M, Desale S, et al.  . Leukocyte telomere length and ideal cardiovascular health in American Indians: the Strong Heart Family Study. Eur J Epidemiol . 2017; 32( 1): 67– 75. Google Scholar CrossRef Search ADS PubMed  8 Fitzpatrick AL, Kronmal RA, Gardner JP, et al.  . Leukocyte telomere length and cardiovascular disease in the Cardiovascular Health Study. Am J Epidemiol . 2007; 165( 1): 14– 21. Google Scholar CrossRef Search ADS PubMed  9 Ding H, Chen C, Shaffer JR, et al.  . Telomere length and risk of stroke in Chinese. Stroke . 2012; 43( 3): 658– 663. Google Scholar CrossRef Search ADS PubMed  10 Willeit P, Willeit J, Brandstätter A, et al.  . Cellular aging reflected by leukocyte telomere length predicts advanced atherosclerosis and cardiovascular disease risk. Arterioscler Thromb Vasc Biol . 2010; 30( 8): 1649– 1656. Google Scholar CrossRef Search ADS PubMed  11 Zee RY, Castonguay AJ, Barton NS, et al.  . Relative leukocyte telomere length and risk of incident ischemic stroke in men: a prospective, nested case-control approach. Rejuvenation Res . 2010; 13( 4): 411– 414. Google Scholar CrossRef Search ADS PubMed  12 Fyhrquist F, Silventoinen K, Saijonmaa O, et al.  . Telomere length and cardiovascular risk in hypertensive patients with left ventricular hypertrophy: the LIFE Study. J Hum Hypertens . 2011; 25( 12): 711– 718. Google Scholar CrossRef Search ADS PubMed  13 Wikgren M, Karlsson T, Söderlund H, et al.  . Shorter telomere length is linked to brain atrophy and white matter hyperintensities. Age Ageing . 2014; 43( 2): 212– 217. Google Scholar CrossRef Search ADS PubMed  14 King KS, Kozlitina J, Rosenberg RN, et al.  . Effect of leukocyte telomere length on total and regional brain volumes in a large population-based cohort. JAMA Neurol . 2014; 71( 10): 1247– 1254. Google Scholar CrossRef Search ADS PubMed  15 Kota LN, Bharath S, Purushottam M, et al.  . Reduced telomere length in neurodegenerative disorders may suggest shared biology. J Neuropsychiatry Clin Neurosci . 2015; 27( 2): e92– e96. Google Scholar CrossRef Search ADS PubMed  16 Liu M, Huo YR, Wang J, et al.  . Telomere shortening in Alzheimer’s disease patients. Ann Clin Lab Sci . 2016; 46( 3): 260– 265. Google Scholar PubMed  17 Cai Z, Yan LJ, Ratka A. Telomere shortening and Alzheimer’s disease. Neuromolecular Med . 2013; 15( 1): 25– 48. Google Scholar CrossRef Search ADS PubMed  18 Schürks M, Buring J, Dushkes R, et al.  . Telomere length and Parkinson’s disease in men: a nested case-control study. Eur J Neurol . 2014; 21( 1): 93– 99. Google Scholar CrossRef Search ADS PubMed  19 Endicott AA, Taylor JW, Walsh KM. Telomere length connects melanoma and glioma predispositions. Aging (Albany NY) . 2016; 8( 3): 423– 424. Google Scholar CrossRef Search ADS PubMed  20 Willeit P, Raschenberger J, Heydon EE, et al.  . Leucocyte telomere length and risk of type 2 diabetes mellitus: new prospective cohort study and literature-based meta-analysis. PLoS One . 2014; 9( 11): e112483. Google Scholar CrossRef Search ADS PubMed  21 Peng H, Zhu Y, Yeh F, et al.  . Impact of biological aging on arterial aging in American Indians: findings from the Strong Heart Family Study. Aging (Albany NY) . 2016; 8( 8): 1583– 1592. Google Scholar CrossRef Search ADS PubMed  22 Haycock PC, Heydon EE, Kaptoge S, et al.  . Leucocyte telomere length and risk of cardiovascular disease: systematic review and meta-analysis. BMJ . 2014; 349: g4227. Google Scholar CrossRef Search ADS PubMed  23 Harwell TS, Oser CS, Okon NJ, et al.  . Defining disparities in cardiovascular disease for American Indians: trends in heart disease and stroke mortality among American Indians and whites in Montana, 1991 to 2000. Circulation . 2005; 112( 15): 2263– 2267. Google Scholar CrossRef Search ADS PubMed  24 Rosamond W, Flegal K, Friday G, et al.  . Heart disease and stroke statistics—2007 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation . 2007; 115( 5): e69– e171. Google Scholar CrossRef Search ADS PubMed  25 Stansbury JP, Jia H, Williams LS, et al.  . Ethnic disparities in stroke: epidemiology, acute care, and postacute outcomes. Stroke . 2005; 36( 2): 374– 386. Google Scholar CrossRef Search ADS PubMed  26 Lee ET, Welty TK, Fabsitz R, et al.  . The Strong Heart Study. A study of cardiovascular disease in American Indians: design and methods. Am J Epidemiol . 1990; 132( 6): 1141– 1155. Google Scholar CrossRef Search ADS PubMed  27 Suchy-Dicey AM, Shibata D, Best LG, et al.  . Cranial magnetic resonance imaging in elderly American Indians: design, methods, and implementation of the Cerebrovascular Disease and Its Consequences in American Indians Study. Neuroepidemiology . 2016; 47( 2): 67– 75. Google Scholar CrossRef Search ADS PubMed  28 Suchy-Dicey AM, Shibata DK, Madhyastha TM, et al.  . Findings of vascular brain injury and structural loss from cranial magnetic resonance imaging in elderly American Indians: the Strong Heart Study. Neuroepidemiology . 2017; 48( 1–2): 39– 47. Google Scholar CrossRef Search ADS PubMed  29 Zhao J, Zhu Y, Lin J, et al.  . Short leukocyte telomere length predicts risk of diabetes in American Indians: the Strong Heart Family Study. Diabetes . 2014; 63( 1): 354– 362. Google Scholar CrossRef Search ADS PubMed  30 Inker LA, Shaffi K, Levey AS. Estimating glomerular filtration rate using the chronic kidney disease-epidemiology collaboration creatinine equation: better risk predictions. Circ Heart Fail . 2012; 5( 3): 303– 306. Google Scholar CrossRef Search ADS PubMed  31 Inker LA, Schmid CH, Tighiouart H, et al.  . Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med . 2012; 367( 1): 20– 29. Google Scholar CrossRef Search ADS PubMed  32 Zhu Y, Voruganti VS, Lin J, et al.  . QTL mapping of leukocyte telomere length in American Indians: the Strong Heart Family Study. Aging (Albany NY) . 2013; 5( 9): 704– 716. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

American Journal of EpidemiologyOxford University Press

Published: Feb 23, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off