Associations of mid-to-late-life inflammation with late-life mobility and the influences of chronic comorbidities, race, and social determinants of health: The Atherosclerosis Risk in Communities Study

Abstract

Background: Relationships of midlife inflammation with late-life mobility and influences of chronic health conditions, race, and social determinants of health (SDoH) on these relationships are poorly understood.

Methods: Among 4758 community-dwelling participants (41% men, 20% Black), high-sensitivity C-reactive protein (hsCRP) was measured over 20+ years: in midlife at study visit 2 (V2: 1990-1992, 47-68 years); at V4 (1996-1998, 53-74 years); and with concurrent late-life 4-m gait speed at V5 (2011-2013, 67-88 years, mean 75 years). SDoH measures included race, the national-rank area deprivation index, education, and income. We examined associations of late-life gait speed with midlife hsCRP (V2 continuous and clinically high ≥3 mg/L), with 20-year hsCRP history from midlife (V2-V5 average continuous hsCRP and clinically high ≥3 mg/L) and with inflammation accumulation (visits and years with high hsCRP). Regression models adjusted for demographic, cardiovascular, and SDoH measures; effect modification by the presence of other common chronic conditions (obesity, diabetes, hypertension) and race were examined, with and without accounting for SDoH.

Results: High midlife hsCRP was associated with slower late-life gait speed, even among those without chronic conditions in midlife: -4.6 cm/s (95% CI: -6.4, -2.8). Importantly, sustained high hsCRP was associated with a 20-year slowing of -10.0 cm/s (-14.9, -5.1) among those who never experienced obesity, diabetes, or hypertension over the 20-year period. Associations were similar between Black participants, -3.8 cm/s (-6.9, -0.7) and White participants -3.3 (-4.5, -2.2) per interquartile range of midlife hsCRP; effect modifications by chronic conditions and race were unsupported throughout. Results were robust to accounting for SDoH or otherwise; however, worse SDoH was associated with higher inflammation and slower gait speed in both Black and White participants.

Conclusions: Inflammation in midlife may contribute to clinically meaningful late-life slowing of gait speed, even among otherwise healthy-appearing adults and regardless of race and socioeconomic disadvantage. Regular monitoring and interventions for inflammation may be warranted from midlife.

Keywords: chronic disease; gait speed; geriatrics functional assessment; inflammation; midlife; social determinants of health.

Figure 1:

Associations of Inflammation with Late-Life Gait Speed by Chronic Condition^a Status for (A) High hsCRP^b in Mid-Life (20-years earlier), (B) High 20-year average hsCRP, (C) Number of visits with High hsCRP a. Chronic conditions included obesity, hypertension and/or diabetes. b. High CRP=high sensitivity c-reactive protein ≥3 mg/L. c. Difference comparing 3 versus 1 High CRP visits. d. Estimates from linear regression models adjusted for age, sex, race, alcohol use, smoking status, coronary heart disease, stroke, anti-inflammatory medications, aspirin use, education

Figure 2.

Area Deprivation Index (ADI) Distributions^a and Associations with (A) Gait Speed^b and (B) 20-Year Average hsCRP^c,d by Race a. Distributions of National ADI depicted with histograms and kernel density estimates by race b. Associations for Gait speed represent absolute differences in gait speed per Inter Quartile Range (IQR) of ADI (IQR=25) c. hsCRP: high sensitivity C-reactive protein d. Associations for hsCRP represent relative differences per IQR of ADI (IQR=25) e. Estimates from regression models adjusted for age, sex, current alcohol use, current smoking status, coronary heart disease, stroke, anti-inflammatory medication use, education

Figure 3:

Associations of Inflammation with Late-Life Gait Speed by Race and Chronic Condition (CC) Status^a for (A) Mid-Life hsCRP^b levels (20-years earlier), (B) 20-year average hsCRP levels, (C) Years of High hsCRP Accumulation (over 20-years) a. Chronic conditions included obesity, hypertension and/or diabetes. b. hsCRP: high sensitivity C-reactive protein c. Estimates from linear regression models adjusted for age, sex, race, alcohol use, smoking status, coronary heart disease, stroke, anti-inflammatory medications, aspirin use, education, income and ADI d. Midlife (visit 2) income and ADI measures were used for adjusted associations in panel A; 20-year average income and 20-year average ADI measures were used for adjusted associations in panels B & C e. IQR: Inter Quartile Range