Horticultural activity predicts later localized limb status in a contemporary pre-industrial population

Abstract

Objectives: Modern humans may have gracile skeletons due to low physical activity levels and mechanical loading. Tests using pre-historic skeletons are limited by the inability to assess behavior directly, while modern industrialized societies possess few socio-ecological features typical of human evolutionary history. Among Tsimane forager-horticulturalists, we test whether greater activity levels and, thus, increased loading earlier in life are associated with greater later-life bone status and diminished age-related bone loss.

Materials and methods: We used quantitative ultrasonography to assess radial and tibial status among adults aged 20+ years (mean ± SD age = 49 ± 15; 52% female). We conducted systematic behavioral observations to assess earlier-life activity patterns (mean time lag between behavioural observation and ultrasound = 12 years). For a subset of participants, physical activity was again measured later in life, via accelerometry, to determine whether earlier-life time use is associated with later-life activity levels. Anthropometric and demographic data were collected during medical exams.

Results: Structural decline with age is reduced for the tibia (female: -0.25 SDs/decade; male: 0.05 SDs/decade) versus radius (female: -0.56 SDs/decade; male: -0.20 SDs/decade), which is expected if greater loading mitigates bone loss. Time allocation to horticulture, but not hunting, positively predicts later-life radial status (β_Horticulture = 0.48, p = 0.01), whereas tibial status is not significantly predicted by subsistence or sedentary leisure participation.

Discussion: Patterns of activity- and age-related change in bone status indicate localized osteogenic responses to loading, and are generally consistent with the logic of bone functional adaptation. Nonmechanical factors related to subsistence lifestyle moderate the association between activity patterns and bone structure.

Keywords: Tsimane; bone functional adaptation; mechanical loading; physical activity.

Figure 1

Generalized additive models for men (A) and women (B) of the effect of age on radial and tibial SOS (radius [men]: edf=3.82, p<0.001, adj. R²=0.10; tibia [men]: edf=2.19, p<0.001, adj. R²=0.05; radius [women]: edf=5.13, p<0.001, adj. R²=0.45; tibia [women]: edf=4.65, p<0.001, adj. R²=0.12). Upper and lower lines of each curve represent local 95% confidence intervals for the thin plate spline. X-axis marks indicate unique ages in the sample (range: 20-85).

Figure 2

Effect of earlier-life time allocation to work (in black or gray) and leisure (red) on later-life radial (A) and tibial (B) SOS (n=116 individuals). Effect sizes (95% CIs) represent the difference in SOS for individuals in the upper quartile of task-specific time allocation compared to all others. Least-squares regressions are weighted by number of behavioural scans/person. For models of radial SOS, controls include age and time lag between behavioural observation and ultrasound (mean±SD=12±1 years); additional controls for models combining sexes (all except hunt, where n=59 men) include age², sex and sex*age². For models of tibial SOS, controls include age, height, weight and time lag; additional controls for models combining sexes include sex, sex*age and sex*height.