Validity of GT3X and Actiheart to estimate sedentary time and breaks using ActivPAL as the reference in free-living conditions

• GT3X and Actiheart showed poor validity to estimate ActivPAL’ sedentary time.
• Both methods showed poor validity to estimate ActivPAL’ breaks in sedentary time.
• Adding HR did not result in improvements for the sedentary behavior estimation.
• Actiheart had a higher rate of lost data and negative feedback from participants.
• GT3X can be a more suitable/feasible alternative to estimate sedentary patterns.

Sedentary time, specifically sitting/reclining, is a risk factor for many non-communicable diseases and premature mortality. Inclinometers have been used as a valid measurement of sedentary time and its patterns; however, there is a lack of information regarding the validity of alternative accelerometry and heart rate methods. The validity of GT3X and Actiheart in estimating changes in daily sedentary time and breaks, during free-living settings, using ActivPAL as the reference was examined. A crossover randomized control trial of an intervention that aimed to reduce ∼3 h/day of sitting time included 10 overweight/obese adults (37–65 years). Participants had a total of 74 valid days for the three devices (29 controls; 45 interventions). For ActivPAL, sedentary time was measured directly based upon posture (sitting/reclining); Actiheart, the presumed MET cutpoint for sedentary time (<1.5 METs) based on accelerometry + heart rate; GT3X, the traditional <100 counts min−1. A break in sedentary time was defined as when the participants were above the aforementioned cutoffs. GT3X overestimated and Actiheart underestimated sedentary time (bias = 135 min; bias = −156 min, respectively) and both methods overestimated breaks in sedentary time (bias = 78; bias = 235 breaks, respectively). The GT3X method was in better agreement with the ActivPAL sedentary time (r2 = 0.70; concordance correlation coefficient (CCC) = 0.56) than the Actiheart (r2 = 0.24; CCC = 0.31). The present results highlight the magnitude of potential errors in estimating sedentary time and breaks from common alternative methods other than ActivPAL. Because misclassification errors from the commonly used surrogates are potentially large, this raises concern that alternative methods used in many epidemiological observations may have underestimated the true effects caused by too much sitting (ClinicalTrials.govID:NCT02007681).