Skip to main content

Research Repository

Advanced Search

The effect of intermittent exercise on mechanical loading and biomarkers of bone metabolism

Evans, Will James


Will James Evans


Massimiliano Ditroilo


Introduction: Bone is a metabolically active tissue which plays a multifunctional role within the body. Animal models have provided evidence demonstrating that dynamic and unaccustomed mechanical loads imposed by gravity and/or muscle activation which exceed the customary strain stimulus are anabolic for bone. Moreover, because the bone’s response to multiple cycles diminishes over time, the inclusion of rest periods between loading cycles has been shown to augment the osteogenic response by allowing the mechanosensory system to re-initialise. As such, rest-inserted intermittent exercise might offer a favourable environment for bone adaptation beyond traditional continuous exercise. Despite this many still engage in continuous aerobic exercise training such as continuous running, which might be deleterious to bone. The use of interval or intermittent running has been proposed as an alternative to reduce bone fatigue and maximise bone accrual more so than continuous exercise. However, evidence for the benefits of intermittent load-bearing exercise on bone adaptation is confined to animal models. Therefore, the aims of the thesis were (1) to investigate the effect of intermittent exercise of varying exercise-to-rest durations with a fixed ratio on the mechanical loading dose, assessed via components of the ground reaction force (GRF), and the osteogenic index, and (2) and to establish the magnitude of effect of intermittent load-bearing exercise on changes in bone tissue, using bone turnover markers compared to a non-exercising control condition.

Study 1 The Force 3 non-motorised treadmill (NMT) was used throughout the thesis because of its ability to measure vertical (vGRF) and anterior horizontal (hGRF) GRF continuously during exercise, and its capacity to reflect intermittent movement patterns which are not as easily replicable on a motorised treadmill (MT). However, a preliminary study to investigate; (1) the cardiorespiratory responses to running on an NMT compared to an MT, and (2) establish appropriate reference speeds to dose intermittent performance for subsequent protocols on the NMT was required. Therefore, the aim of study one was to establish the validity of peak cardiorespiratory responses to intermittent and continuous graded exercise tests (GXTs) on a NMT compared to an MT. When a continuous GXT is performed on the NMT a similar maximal oxygen uptake can be achieved (𝑉̇𝑂2π‘šπ‘Žπ‘₯) compared to that achieved on the MT (P = 1.00, d = 0.01, trivial). However, there was a reduction in peak heart rate (P = 0.0001, d = 0.9, moderate), and peak speed is reduced by ~30%. When an intermittent GXT (15 x 15 s) is performed on a NMT, a similar 𝑉̇𝑂2π‘šπ‘Žπ‘₯ can be achieved (P = 0.701, d = 0.16, trivial), together with smaller reductions in peak HR (P = 0.170, d = 0.34, small), and peak speed (P = 0.009, d = 0.8, moderate) compared to a continuous MT running. Conversely, 30 x 30 s had a statistically significant reduction in 𝑉̇𝑂2π‘šπ‘Žπ‘₯ (P = 0.04, d = 0.37, small) and peak HR (P = 0.0001, d = 0.57, small) compared to the Cont-MT. As such, the shorter 15 x 15 s GXT was used to obtain reference speeds for subsequent studies (Studies 2-4).

Study 2 The osteogenic potential of impact exercise can be quantified from the analysis of kinetic (peak vGRF, load rate and vertical impulse) variables with greater peak vGRF, load rate and vertical impulse reflecting (indirectly) the magnitude and rate of the mechanical load applied to bone tissue. Our aim was to establish how manipulating the duration and frequency of exercise-to-rest intervals might change the mechanical loading environment using a fixed 1:1 ratio. Twelve healthy active males performed five 45 min intermittent running protocols on the NMT. Experiment 1: three of the intermittent protocols differed in their exercise-to-rest durations (5 s intervals [5s-Int], 20 s intervals [20s-Int] and 80 s intervals [80s-Int]). Experiment 2: three of the protocols differed in the rate of acceleration and deceleration but matched for the exercise-to-rest duration (20 s by 2 s intervals [20s2s-Int], 20 s by 4 s [20s4s-Int] and 20 s by 6 s [20s6s-Int]). The primary outcome measures for experiment 1 & 2 were mean and peak vGRF, vertical impulse, load rate and the intra-step variability assessed via the coefficient of variance (%CV) of the kinetic and kinematic data. There was no statistical difference between conditions for impulse (P = 0.175), maximum load rate (LR) (P = 0.104) or average load rate (ALR) (P =0.345). Peak vGRF data were statistically different between conditions (P = 0.023) with the 5s-Int being greater than the 80s-Int (P = 0.022). There was a statistical effect of condition on all CV data for vertical impulse (P = 0.0001), load rate (P = 0.0001), vGRF (P = 0.0001), kVert (P =0.0001) and kLeg (P = 0.0001) with the 5s-Int & 20s-Int being higher than the 80s-Int. The similarity in the mean GRF data are likely due to the higher loads generated during higher speeds being counteracted by the lower loads at the lower speeds. The variability in the data are caused by the variation between the high and low speeds.

Study 3 The magnitude, rate and frequency of the mechanical load are proportional to the amount of bone adaptation. These components were considered in isolation in study 2. However, the components can be combined into one mathematical algorithm, the osteogenic index (OI), to assess the osteogenic potential of the exercise. However, because we demonstrated no statistical effect on mean loading due to the higher loads being counteracted by the lower loads, it is unlikely that the traditional OI can distinguish between the variable loading environments of more intermittent exercise. A novel approach has been developed which incorporates the magnitude and rate of the loading dose across a frequency spectra. The number of loading segments of a particular exercise can then be categorised into the magnitude, intensity and frequency. As yet this method has only been utilised with accelerometers and only during steady-state conditions. Experiment one: There was a statistically significant difference between conditions for the OI_fft (P = 0.0001). The OI_fft was highest for the 80s-Int being 28% greater than the 5s-Int and 23% greater than the 20s-Int. There was a statistically significant 3-way interaction for condition*frequency band*intensity (P = 0.012) with the more intermittent conditions having a higher loading dose at higher frequencies. Experiment two: There was a significant main effect for condition for the mean differences of the OI_fft between conditions (P = 0.033). There was no significant effect of condition for the OI_BW (P = 0.572). There was no significant 3-way interact for condition*frequency band*loading intensity (P = 0.870). When the magnitude, intensity and frequency of multiple loading segments are considered, intermittent locomotion allows the individual to obtain higher magnitudes of loading dose, shifting towards a higher frequency band. It is unclear whether the intermittent protocol might offer a more favourable loading environment due to the more variable loading patterns with greater magnitudes of load at higher frequencies.

Study 4: Whilst brief continuous load-bearing exercise increases bone remodelling in favour of resorption, it is unclear how intermittent exercise effects acute bone remodelling. Our aim was to investigate the effect of varying degrees of intermittent exercise, with a fixed exercise-to-rest ratio, on acute bone remodelling, as measured by bone turnover biomarkers. It was hypothesised that the more intermittent protocol would result in a greater bone turnover compared to the non-exercising control and less intermittent conditions. The same exercise protocols from study 2 and 3 were used. Venous blood samples were collected at the same time of day following a 12 h fast at baseline, 1 h, 2 h and 24 h post-exercise. Carboxyterminal crosslinked telopeptide (CTX-1) and procollagen type 1 amino terminal propeptide (P1NP) were used as markers of bone resorption and formation, respectively. There was a significant main effect for time (P = 0.0001), condition (P = 0.032) and a significant condition by time interaction (P = 0.001) for CTX-I. At 1 h the 20s-Int and 5s-Int were higher than the control condition (20s-Int: P = 0.0001, 5s-Int: P = 0.010). There was no significant condition by time interactions for P1NP. The results confirm that bone remodelling is stimulated acutely by load-bearing exercise. Very short and short interval intermittent exercise results in greater bone turnover compared to longer interval intermittent exercise.

Conclusion: The effect of different exercise-to-rest intervals, and therefore different frequency of intermittency did not have an effect on the mean vGRF, load rate, vertical impulse or osteogenic index. However, as expected there was a greater intra-step variability in these measures leading to a more variable mechanical loading environment which might offer a more favourable loading environment for bone. Indeed, when bone tissue turnover was assessed using traditional bone turnover markers all exercise conditions demonstrated an increase in bone resorption compared to a non- exercising control condition at 1 h. However, only the short and very short intervals were statistically elevated above control. Therefore, very short and short intermittent exercise might cause greater bone resorption in the prevailing hours following exercise which could stimulate an increase bone formation.


Evans, W. J. (2016). The effect of intermittent exercise on mechanical loading and biomarkers of bone metabolism. (Thesis). University of Hull. Retrieved from

Thesis Type Thesis
Deposit Date Mar 7, 2017
Publicly Available Date Feb 23, 2023
Keywords Sports sciences
Public URL
Additional Information Department of Sport, Health and Exercise Science, The University of Hull
Award Date Nov 1, 2016


Thesis (4.5 Mb)

Copyright Statement
Β© 2016 Evans, Will James. All rights reserved. No part of this publication may be reproduced without the written permission of the copyright holder.

You might also like

Downloadable Citations