To the Editor,
We thank Dr Füzéki and colleagues for their thoughtful letter regarding our meta-analysis of the association of light intensity physical activity (LIPA) and adult cardiometabolic health and mortality . The letter raised a number of important issues that we faced during the preparation of our meta-analysis.
Firstly, we agree that how LIPA is defined is crucial for all subsequent steps of the review and meta-analysis. We did adopt a very strict definition of LIPA in terms of metabolic cost (≥1.5 <3 MET) and the equivalent operational definitions of heart rate, % of V02 max and rate of perceived exhaustion. However, as Füzékiand colleagues rightly pointed out, there is no absolute standardization of LIPA. As boundaries between intensity bands are very subjective and not exact thresholds, there is no hard evidence to suggest that any one classification is better than another. Consequently, we did not want to reference only one classification of LIPA but rather include all definitions to harvest as wide an evidence base as possible.
The biggest issue we encountered when identifying literature did not result from our definition of LIPA but from differences in how authors use the word ‘light activity’. For example, a large number of studies including the STRIDE study recommended by Füzékiand colleagues refer to a low volume of activity as light activity. As Füzékiand colleagues rightly pointed out, volume is not a measure of intensity. Similarly, the type and nature of activity is bound to provide only a rough approximation of usual intensity. As a result, we only included studies that provided a quantitative value of intensity for light activity which resulted in the exclusion of some yoga studies that reported only the type of activity and not the intensity of the activity.
Several issues do exist in the translation of self-report PA instruments to MET-intensities, but this is equally the case with objective measures as well. These translational issues are likely to have contributed to the heterogeneity of results that we reported. However, we chose not to expand on the poor accuracy of self-reported measure in our manuscript, because there is a very large body of literature covering this.
As stated in our manuscript (page 2), we only included one study per dataset and selected the study with the highest quality score. Consequently, we only included one study from the NHANES dataset in each section of the analyses. The recent review by Fuzeki et al  was very useful in summarizing the different findings derived from a single data source, albeit an important one (NHANES). Although we see value in this approach, we feel that it would have been inappropriate for the aims of our meta-analysis to include replication analyses from the same dataset that corresponds to one source of evidence.
We agree that, ideally, we would have used a dichotomous selection criterion to separate “healthy” participants from non-healthy participants. However, finding cohorts of adults and older adults who are entirely free of disease is unrealistic. Our approach was pragmatic, logical, and highly congruent. We used dichotomy and excluded rehabilitation trials as these studies would have included only people with diagnosed cardiovascular events and cancer whose physical activity behaviour is likely to be severely affected by major medical interventions, such as heart surgery or chemotherapy. Such concerns are much less relevant for participants who “have some form of metabolic impairment (eg, obesity or type 2 diabetes)”.
Finally, we agree that the choice of confounding variables is important, as this is one of the common limitations of meta-analyses. Unlike studies investigating sedentary behaviour, studies that focus on LIPA do not always adjust for MVPA. Further, very few studies in the vast evidence base on MVPA have adjusted their analyses for LIPA.
In summary, we hope that our meta-analysis will provide impetus for further research that will address the various limitations in the field. The letter of Füzéki and colleagues provides valuable, complementary insights to our review and will assist in guiding future meta-analyses that investigate the effects of LIPA on health.
Sebastien Chastin PhD
Emmanuel Stamatakis PhD
Mark Hamer PhD
Philippa Dall PhD
Marieke De Craemer PhD
Jelle van Cauwenberg PhD
Katrien De Cocker PhD
1 Chastin SFM, Craemer M de, Cocker K de, et al. How does light-intensity physical activity associate with adult cardiometabolic health and mortality? Systematic review with meta-analysis of experimental and observational studies. Br J Sports Med2018.Published Online First: 25 April 2018. doi: 10.1136/bjsports-2017-097563
2 Füzéki E, Engeroff T, Banzer W. Health Benefits of Light-Intensity Physical Activity: A Systematic Review of Accelerometer Data of the National Health and Nutrition Examination Survey (NHANES). Sports Med2017;47(9):1769–93.