"Obesity and Muscle Phenotypes as Predictors of Falls in Aging Populations" School of Public Health Department of Epidemiology. 

Committee: 

• Elsa S. Strotmeyer (chair, advisor), EPI
• Iva Miljkovic, EPI
• Samaneh Farsijani, EPI
• Ying Ding, Biostatistics and Health Data Science

Abstract Background: Falls are the leading cause of injury and mortality among adults aged 65 years and older in the United States, with 1 in 4 experiencing at least one fall annually. Age-related changes in body composition, characterized by increased adiposity, ectopic fat deposition, and deterioration in muscle quantity and composition, may play a central role in elevated fall risk. However, few studies have examined varied obesity and muscle phenotypes to determine which predict falls in community-dwelling older adults.

Objectives: This dissertation examined the associations of (1) whole-body, abdominal, and sarcopenic obesity; (2) D3-creatine (D₃Cr) muscle mass and magnetic resonance imaging (MRI)-derived muscle volume and muscle fat infiltration; and (3) computed tomography (CT)-derived fat and muscle area and density, with incident, recurrent, and number of falls in older adults longitudinally.

Methods: Participants included: (1) Falls Cross-Cohort Study (FCCS) data, harmonizing the Health, Aging and Body Composition Study (Health ABC), Cardiovascular Health Study (CHS), the Osteoporotic Fractures in Men (MrOS) Study, and Women’s Health Initiative (WHI) (N=15,054, age 75.8 ± 6.2 years); (2) the Study of Muscle, Mobility and Aging (SOMMA) (N=734, age 76.4 ± 5.0 years); and (3) Health ABC (N=2,372, age 73.6 ± 2.9 years), for the three aims. Generalized estimating equations estimated longitudinal odds ratios (ORs) or incidence rate ratios (IRRs) of incident, recurrent, and number of falls, adjusting for demographic, behavioral, and health status-related covariates.

Results: In FCCS, whole-body, abdominal, non-sarcopenic obese, sarcopenic non-obese, and sarcopenic obesity were associated with higher odds of incident, recurrent, and number of falls over 7.5 ± 4.9 years of follow-up (sarcopenic obesity OR range: 1.37–1.46). In SOMMA, higher D₃Cr muscle mass was associated with lower odds/incidence rates of incident, recurrent, and number of falls over 3.0 ± 1.0 years of follow-up overall (OR/IRR range: 0.73–0.84), while higher MRI muscle volume was associated with lower odds/incidence rates of fall outcomes only in men (OR/IRR range: 0.63–0.78). In Health ABC, higher abdominal visceral, abdominal subcutaneous, and thigh intermuscular fat area were associated with higher odds of incident, recurrent, and number of falls over 9.6 ± 3.1 years of follow-up overall (OR range: 1.07–1.16), while in men, higher thigh intermuscular fat area was associated with higher odds of fall outcomes (OR range: 1.20–1.24) and lower thigh total/quadriceps densities were associated with lower odds of fall outcomes (OR range: 0.75–0.79). In women, higher abdominal visceral fat area was associated with recurrent falls (OR=1.14) and higher abdominal subcutaneous fat density with incident falls (OR=1.08), with no significant thigh-level associations.

Conclusion: These findings identify higher fat quantity and lower muscle strength and mass as predictors of fall risk in older adults, while muscle volume and muscle fat infiltration were associated with fall risk in men only. These findings highlight the importance of considering sex-specific associations and incorporating direct imaging and biomarker measures in body composition research among older adults. These results support interventions targeting abdominal fat reduction and muscle gain to reduce the public health burden of falls.