Metabolic Effect of Breaking Up Prolonged Sitting With Stair Climbing Exercise Snacks

Hossein Rafiei; Kosar Omidian; Étienne Myette-Côté; Jonathan Peter Little


Med Sci Sports Exerc. 2021;53(1):150-158. 

In This Article


Study Design

A randomized crossover design was used involving two 9-h experimental trials: i) sedentary (SED; participants were sitting on a chair throughout the experimental trial and asked to minimize their movement) and ii) stair "snacks" (SS; ascending three flights of stairs at a brisk speed (15–30 s) once every hour (×8)) with identical meals. A third condition (sedentary low carbohydrate) was also included in the randomization schedule because the original study was designed and funded to measure saliva insulin responses across the day, and a low-insulin response comparator condition was deemed necessary for this research question. The study was registered with three conditions in the clinical trial registry, but because the low-carbohydrate condition was not included in the hypothesis tested herein, only data from the two relevant SED and SS conditions are reported. The exact same high-glycemic index meals of a peanut butter jam sandwich with 400 mL of orange juice from concentrate (~530 kcal; 97 g carbohydrate, 11 g protein, 11 g fat) were provided for breakfast, lunch, and dinner in each condition. Meals in both conditions were matched for calories and were provided at 0, 180, and 360 min. A wash-out period of 3–7 d was chosen to eliminate any potential carryover effects. The study was approved by the University of British Columbia Clinical Research Ethics Board (ID H17-01747) and was registered on (NCT03374436). The study conformed to the standards set by the Declaration of Helsinki. All participants provided written informed consent before data collection.


Participants were recruited through distribution of posters, by e-mail, and by word of mouth on the University of British Columbia campus (Kelowna, British Columbia, Canada). The main objectives were to determine the effects of stair snacks in young healthy-weight (HW) men (n = 12; age, 22.8 ± 4.3 yr; body mass index (BMI), 24.3 ± 2 kg·m−2; study 1) and in adults with overweight/obesity (OW; n = 11, characterized with elevated waist circumference >88 cm for women and >102 cm for men; age, 50.2 ± 14.3 yr; BMI, 35.1 ± 6.4 kg·m−2; study 2). Study 1 was designed as a pilot test of stair climbing snacks and included only male participants to minimize the potential influence of menstrual cycle on insulin sensitivity and glucose tolerance. Based on the successful feasibility of performing eight stair climbing snacks across 9 h, study 2 began after study 1 was completed and included both male and female participants, which facilitated easier recruitment of participants with overweight/obesity characterized with elevated waist circumference and to be inclusive of both sexes. Inclusion criteria for study 1 were as follows: 1) BMI of 18.5–24.9 kg·m−2 and 2) age of 18–35 yr. Inclusion criteria for study 2 were as follows: 1) waist circumference ≥88 cm for women or ≥102 cm for men (elevated waist circumference in our study was defined according to World Health Organization classification of the waist circumference cutoff points made for overweight or obesity, and association with disease risk) and 2) age of 18–69 yr. Exclusion criteria for both studies were as follows: 1) previous diagnosis of diabetes; 2) currently taking insulin, oral hypoglycemic drugs, or any medications affecting blood glucose; 3) diagnosed CVDs; 4) current smoker; 5) allergy to eggs or peanuts; 6) undertaking serious (>5 d·wk−1 exercise training; 7) medical/orthopedic conditions that would limit physical activity; and 8) individuals following a special diet such as ketogenic diet or being vegan. Among female participants in study 2, five were postmenopausal and three were premenopausal (two with intrauterine device and one on birth control pills). Premenopausal female participants were tested in the follicular phase of the menstrual cycle (days 3–9 after menstruation).

Baseline Testing

On the first visit, participants provided written informed consent and were interviewed by a registered dietician to collect data about lifestyle habits and medical conditions to confirm eligibility. Then, anthropometric measurements (height, weight, waist circumference, and hip circumference) were measured. Waist circumference and hip circumference were measured in duplicate to the nearest 0.5 cm using standard procedures.

Standardization of Diet and Activity Before Study

Before experimental trials, participants were interviewed by a study dietician to confirm they were not on a specific diet (e.g., vegan, keto, etc.) and had not lost or gained weight recently. Energy intake for the day before each trial was controlled by asking the participants to fill out a 24-h food recall a day before visit 1 and instructing them to reproduce the exact same diet on the day before subsequent testing days. This was confirmed by a dietician in the morning before running each trial. We did not specifically prescribe energy intake on the day before the trial in order to increase external validity and reduce participant burden, but each participant ate the same foods (type and quantities) before each trial, which was guided and verified by the study dietician. Energy intake on the day of the trial was not adjusted between SS and SED, as the energy expenditure during the short (~15–30 s) stair climbing was, by design, very low and it would not have been difficult or impractical to accurately determine the energy expended by the stair climbing exercise in the protocol. Participants were instructed to abstain from alcohol and refrain from exercising on the day before each experimental trial visit, which was confirmed by self-report and by measuring steps taken using an activity tracker (Mio Slice watch, Canada). Steps recorded were only available in the HW group because of technical issues with the watch in the OW group. Self-reported sleep hours on the night before each visit were also recorded. Compliance with activity standardization, confirmation of sleep hours, and dietary control were confirmed by the dietician upon arrival to the laboratory for each experimental visit.

Experimental Trials

Participants arrived at the laboratory between 7:00 and 8:30 AM after an overnight fast (≥10 h). An intravenous catheter (BD Nexiva; Becton Dickinson, Franklin Lakes, NJ) was inserted into an antecubital vein for repeat blood sampling. Samples were drawn into tubes containing ethylenediamine tetraacetic acid (EDTA-K2) at time 0 (fasting) and every 30 min for a total 19 blood samples across 540 min. Samples were immediately centrifuged at 1550g for 15 min at 4°C and kept in −80°C freezer before batch analyses (details hereinafter). In all trials, participants remained seated in a chair for 9 h working on a computer, watching TV, or reading. Participants were allowed to walk (~10 m) from the laboratory to use the bathroom. Meals were provided just after the first morning blood sample (time 0) and at 180 and 360 min. Water was provided ad libitum. The aforementioned wrist watch monitor was used to track steps and heart rate over the 9-h laboratory visit. In the SS condition, participants walked (~25 m) to a stairwell adjacent to the laboratory and were instructed to ascend three flights of stairs without skipping steps (55 steps) as quickly and safely as possible, every hour beginning at 60 min, for a total of eight stair snacks. The stair climbing snacks at 180 and 360 min were performed immediately before meal consumption. Participants with OW were asked to climb the stairs at a self-selected challenging pace given that stair climbing at a sprint pace was not feasible for some participants. Stair climbing was supervised by a research technician who recorded RPE (category-ratio 0–10 scale), total time of each stair climb, and heart rate immediately after each stair snack. Details of the experimental protocol are presented in Figure 1.

Figure 1.

Overview of the study design. Meal, high-glycemic index meal.


Participants were randomized to complete the intervention trials using a Williams latin square design and an online randomizer (

Biochemical Analyses

Blood metabolites were analyzed using commercially available kits as follows: plasma glucose (glucose hexokinase; Pointe Scientific Inc., Canton, MI), plasma nonesterified fatty acids (NEFA; HR Series; Wako Diagnostics, Mountain View, CA), and plasma TG (Pointe Scientific Inc., Canton, MI) were analyzed on a Chemwell 2910 automated analyzer (Awareness Technologies, Palm City, FL). Plasma insulin (human insulin enzyme-linked immunosorbent assay; Crystal Chem, Elk Grove Village, IL) was analyzed on an iMark Microplate absorbance reader (Bio-Rad, Hercules, CA). All assays except TG were run in duplicate following the manufacturer's instructions. Intra-assay coefficients or variance averaged 3.2% for plasma glucose, 5.9% for plasma insulin, and 3.3% for NEFA.

Statistical Analyses

All values are reported as the mean (SD). The study was designed as a pilot study aiming to enroll 12 participants in each group. Sample size calculations were performed for the primary outcome of insulin AUC based on previous studies reporting 15%–20% reductions in postprandial insulin in activity break versus sedentary conditions[10,22] yielding effect sizes of d = 1.0. With a conservative correlation of r = 0.5 between repeated measures, a two-tailed α of 0.05, and power of 0.80, it was estimated that 10 participants would be required to detect a significant difference in the primary outcome. The total (9 h) AUC for plasma insulin, glucose, NEFA, and TG were calculated according to the trapezoid method using baseline of zero (Prism version 8.0; GraphPad Software Inc.). The positive incremental AUC was calculated using the trapezoid rule with baseline subtraction. AUCs in SED versus SS condition were analyzed using t-test and reported as main analyses. The main analyses for insulin and glucose were supplemented with a linear mixed-effects model with fixed repeated-measures effects of meal (breakfast, lunch, dinner), condition, and their interaction and a random effect of participants to explore the difference between 3-h AUC for each meal. Significant interactions between meal and condition were followed up with pairwise comparisons examining each meal between conditions, whereas a main effect of meal was followed up with pairwise comparisons using Bonferroni adjustments with conditions collapsed. Because the two studies were run at different times and the HW group included only male participants, the two groups were analyzed in separate models. Before all statistical testing, normality and skewness were assessed by Q–Q plots and data were natural log transformed when required. Cohen d effect sizes were calculated for pairwise comparisons using the method that was previously defined for repeated measures.[23] In this method, mean and SD of the pairwise groups and their correlation were taken into account. Significance was set at P < 0.05.