Background
More people die from cardiovascular disease (CVD) than from any other cause [
1]. Since physical activity (PA) is important for cardiovascular health, all adults are recommended to perform ≥150 min of moderate or ≥75 min of vigorous PA weekly, or some combination of the two [
2]. The PA intensity can be expressed as absolute or relative. Absolute intensity is quantified using work energy expenditure (i.e. metabolic equivalents-of-task or METs), while relative intensity is determined relative to individual cardiorespiratory fitness (CRF) (i.e. peak oxygen uptake or VO
2peak) and differs for the unfit compared to fit individuals [
3].
Accelerometers are often used to objectively assess population adherence to PA recommendation [
4,
5]. Accelerometer output is given in counts [
6]. However, the count thresholds used to define moderate-to-vigorous physical activity (MVPA) and assess PA recommendation adherence, are based on absolute intensity and are derived from physical exertion of healthy young to middle-aged adults [
7]. These thresholds could have low validity in those with low CRF (i.e. older adults). As the CRF declines with age, it results in changes in relative effort required to perform PA [
8,
9] and for many unfit older adults, absolute light intensity PA (<3 METs) requires moderate relative effort, while absolute vigorous intensity PA (6–9 METs) is often unattainable [
8]. Those not meeting the absolute PA recommendation may be meeting the relative PA recommendation, likely resulting in underestimation of PA-recommendation adherence in this population.
It is estimated that 5.4 million people in the United Kindom would attain or exceed vigorous relative intensity (>70% of VO
2peak) by walking at ≈4.8 km/h [
10]. Therefore, even low absolute PA, if performed at high relative intensity, has potential to benefit many by improving CRF, which is a powerful predictor of mortality [
11,
10]. However, relative PA assessment in a population is hindered by methodological limitations [
12] and until recently, the only available relative thresholds were derived from physical exertion of young to middle-aged healthy adults. Further, methodology associated with application of these thresholds is rather complex, limiting their use to smaller studies [
13].
Since PA recommendation is also given in relative intensity [
2,
14], it may be valuable, in populations of varying CRF (i.e. older adults) to measure MVPA using recently published relative thresholds derived from physical exertion of the older adults [
15]. Furthermore, it is not known if absolute or relative thresholds quantify MVPA that better associates with metabolic syndrome (MetS) [
16], which was found to associate with CVD- and all-cause mortality in the older adults [
17,
18] or what role the two components of MVPA, moderate (MPA) and vigorous (VPA) physical activity, play.
The aim of this study, therefore, was to compare the proportion of the older adults meeting relative versus absolute PA recommendation and to determine the method which quantifies PA that better associates with MetS.
Discussion
The main finding of the current study was that a significantly higher proportion of our older adults met the relative versus absolute PA recommendation. Those below the relative PA recommendation were more likely to have increased prevalence of MetS compared to those below the absolute PA recommendation. Importantly, not meeting the relative VPA recommendation was associated with higher likelihood of MetS, while no such association was observed in those below the absolute MPA, VPA or MVPA recommendation.
To our knowledge, this is the first study utilizing accelerometers to objectively assess relative PA in a large population of older adults using thresholds stratified for sex and CRF, and investigating how the relative MVPA and its two components, MPA and VPA, associate with MetS. As individuals age, their CRF declines, producing a change in the relative intensity of effort required for PA [
8,
9]. Many individuals with low CRF, including the older adults, require moderate relative effort for low absolute PA and can rarely reach absolute MPA, MVPA or VPA [
8].
Previous studies showed accelerometer thresholds to vary with age, with older individuals having lower accelerometer count output. This difference was ascribed to variation in CRF, which stressed the importance of relative intensity when assessing PA in populations of different ages and CRF levels [
26]. Ozemek et al. were the first to illustrate the importance of CRF in accelerometer PA assessment when they showed an unfit younger individual to have a significantly lower MVPA threshold compared to a more fit older individual [
13]. Notably, they found the MVPA counts to be significantly correlated to CRF, explaining ca. 30% of the variability, while only ca. 1% could be explained by age and BMI. However, Ozemek’s approach to relative PA assessment is based on physical exertion of younger adults and requires complex methodology, making its use cumbersome in larger studies [
13]. In our study we utilized thresholds derived from physical exertion of older adults to quantify relative PA [
15]. These thresholds are adjusted for CRF and sex and are feasible to use in large studies [
15].
Higher proportion of our participants met the absolute MVPA recommendation compared to other studies on older Norwegians [
27‐
29]. This difference could be ascribed to fewer study participants [
27‐
29], wider age interval [
27‐
29], and different methodology [
27‐
29] of these studies. For instance, these studies utilized the Troiano [
27‐
29], while we used the Freedson threshold [
27‐
29] for absolute MVPA assessment and they applied an earlier (30 min of daily MVPA) recommendation, while we used the current Norwegian PA recommendation [
27‐
29]. Notably, by applying the earlier PA recommendation in a preliminary analysis of our study, we obtained a similar result, with 23% of our population meeting the absolute MVPA recommendation.
The lack of association between absolute PA recommendation adherence and MetS in the current study may be attributable to absolute thresholds being too high for this population. Some older adults may have difficulties reaching absolute thresholds and as a result fail to meet the absolute PA recommendation. In fact, results of our absolute VPA analysis show that out of 509 men and 567 women included in our study, only two individuals, both men, managed to meet the absolute VPA recommendation. One could argue that expecting unfit individuals, such as some older adults, to accumulate sufficient time at an absolute intensity higher than their maximal capacity (e.g. VO
2peak ~ 3METs) in order to meet the PA recommendation is unrealistic. Relative thresholds, on the other hand, are individualized in terms of CRF and sex. Indeed, it has been shown that the least fit individuals from the Generation 100 cohort are also least likely to adhere to both absolute and relative PA recommendation compared to the moderately fit and highly fit individuals [
30]. Approximately 30% of our population fail to meet the relative PA recommendation. This is important information as it may allow researchers to identify the least active individuals when designing strategies to increase PA participation in older adults. Our results encourage researchers to consider PA intensity in both relative and absolute terms when utilizing accelerometers in assessment of PA, especially in populations with varying degrees of fitness and physical functionality. Thus, meeting the relative PA recommendation for older adults may not only prove a realistically achievable goal, but may be beneficial for health.
The strength of our study is that it uses objective measures of absolute and relative PA, and relates them to objectively measured health indicator in a large sample of older adults. Furthermore, CRF of our population was assessed objectively. However, relative MPA did not perform well in our study. This was likely due to 2-min interruption allowance during the 10-min-bout data processing. Actilife defines interruptions as “minutes outside of the minimum and maximum count levels” [
31]. These interruptions do not distinguish peaks from drops. Unlike the MVPA and VPA (with only drops counted as interruptions), the MPA threshold is defined as a range where peaks and drops crossing the range are added up, resulting in higher number of discounted bouts.
The extensive analysis of non-participants in our study revealed presence of a selection bias, with the participants reporting higher PA, education and better health than non-participants [
19]. Our participants were likely fitter than non-participants. Nevertheless, our population was diverse and included healthy as well as older adults with co-morbidities [
19]. Our population is a good representation of the general older Norwegian population, with similar co-morbidity prevalence as described in the 2015 Norwegian Institute of Public Health report [
19,
32]. The thresholds for relative PA analysis were sex and CRF stratified, and created specifically for our population. However, the external generalizability of our findings to populations of different ages and ethnicities is limited, and the cross-sectional nature of our study prevents us from establishing causality. Therefore, longitudinal studies on more diverse populations with hard endpoints such as morbidity and mortality are needed.
Acknowledgements
The cardiopulmonary tests were performed at the core facility NeXt Move, Norwegian University of Science and Technology. We thank the Clinical Research Facility at St. Olavs Hospital for outstanding assistance during the testing and all staff for the contribution to data collection.