Within-subject variability in low back load in a repetitively performed, mildly constrained lifting task

A repeated-measures in vivo experiment.

To describe within-subject variability of spinal compression in repetitive lifting.

Epidemiology and failure mechanics suggest that peak loads may be more predictive of injury than average loads. Nevertheless, biomechanical studies usually focus on the latter.

Ten healthy males performed 360 lifts in 1 hour of a 45-L crate, weighted with a stable 10-kg mass on 1 day and with an unstable mass (10 kg of water) on another day. The maximum compression force in each lift was estimated, using a simple inverse dynamics model and a single equivalent muscle model.

The individual distributions of maximum compression force were slightly skewed to the right (average skewness 0.67). Median and 95th percentile values were used to characterize the distribution. The median (50th percentile) compression ranged from 3375 to 6125 N, and from 3632 to 6298 N in the stable and unstable load conditions, respectively. The within-subjects peak (95th percentile) compression forces were from 405 to 1767 N and from 526 to 2216 N, respectively, higher than the median values. The peak values differed significantly between conditions, whereas the difference in medians did not reach significance. Only a limited trendwise (fatigue-related) variance could be demonstrated.

Peak spinal compression by far exceeds median compression in repetitive lifting and can be affected by task conditions independently from the median. Therefore, the variability of spinal loads needs to be taken into consideration when analyzing and redesigning tasks that can cause spinal injuries.