Inter-Individual Responses of Maximal Oxygen Uptake to Exercise Training: A Critical Review

Issues in the determination of 'responders' and 'non-responders' in physiological research

NEW FINDINGS

What is the topic for this review? We discuss the dichotomization of continuous-level physiological measurements into 'responders' and 'non-responders' when interventions/treatments are examined in robust parallel-group studies. What advances does it highlight? Sample responder counts are biased by pre-to-post within-subject variability. Sample differences in counts may be explained wholly by differences in mean response, even without individual response heterogeneity and even if test-retest measurement error informs the choice of response threshold. A less biased and more informative approach uses the SD of individual responses to estimate the chance a new person from the population of interest will be a responder.

ABSTRACT

As a follow-up to our 2015 review, we cover more issues on the topic of 'response heterogeneity', which we define as clinically important individual differences in the physiological responses to the same treatment/intervention that cannot be attributed to random within-subject variability. We highlight various pitfalls with the common practice of counting the number of 'responders', 'non-responders' and 'adverse responders' in samples that have been given certain treatments or interventions for research purposes. We focus on the classical parallel-group randomized controlled trial and assume typical good practice in trial design. We show that sample responder counts are biased because individuals differ in terms of pre-to-post within-subject random variability in the study outcome(s) and not necessarily treatment response. Ironically, sample differences in responder counts may be explained wholly by sample differences in mean response, even if there is no response heterogeneity at all. Sample comparisons of responder counts also have relatively low statistical precision. These problems do not depend on how the response threshold has been selected, e.g. on the basis of a measurement error statistic, and are not rectified fully by the use of confidence intervals for individual responses in the sample. The dichotomization of individual responses in a research sample is fraught with pitfalls. Less biased approaches for estimating the proportion of responders in a population of interest are now available. Importantly, these approaches are based on the SD for true individual responses, directly incorporating information from the control group.

Interindividual variability and individual responses to exercise training in adolescents with obesity

This study investigated the impact of exercise training on interindividual variability and response rates in body composition and cardiometabolic outcomes in adolescents with obesity. Postpubertal males and females (n = 143) were randomly assigned to 6 months of a diet-only control or aerobic, resistance, or combined exercise training. Body composition indices were percentages of body fat mass and lean body mass and waist circumference. Biomarkers of cardiometabolic health were systolic blood pressure and plasma fasting glucose, triglycerides, and high-density lipoprotein cholesterol. Interindividual variability was examined by comparing the standard deviation of individual responses (SD_IR) to a smallest robust change (SRC). The typical error of measurement was used to classify responses. SD_IR exceeded the SRC for percent body fat mass in all exercise groups (SRC = 1.04%; aerobic SD_IR = 1.50%; resistance SD_IR = 1.22%; combined SD_IR = 2.29%), percent lean body mass (SRC = 1.38%; SD_IR = 3.2%,), systolic blood pressure (SRC = 2.06 mm Hg; SD_IR = 4.92 mm Hg) in the resistance group, and waist circumference (SRC = 2.33 cm; SD_IR = 4.09 cm), and fasting glucose (SRC = 0.08 mmol/L; SD_IR = 0.28 mmol/L) in the combined group. However, half of the reported variables (11/21) did not have a positive SD_IR. Importantly, adverse response rates were significantly lower in all 3 exercise groups compared with control for body composition. Although exercise had a small influence on interindividual variability for indices of body composition, the rate of adverse responses did not increase for any outcome. Novelty Interindividual variability and individual responses to exercise training have not been investigated in adolescents with obesity. Six months of exercise training does not increase interindividual variability in adolescents with obesity. Exercise created a positive, uniform shift in responses.

Precision exercise medicine: understanding exercise response variability

There is evidence from human twin and family studies as well as mouse and rat selection experiments that there are considerable interindividual differences in the response of cardiorespiratory fitness (CRF) and other cardiometabolic traits to a given exercise programme dose. We developed this consensus statement on exercise response variability following a symposium dedicated to this topic. There is strong evidence from both animal and human studies that exercise training doses lead to variable responses. A genetic component contributes to exercise training response variability.

In this consensus statement, we (1) briefly review the literature on exercise response variability and the various sources of variations in CRF response to an exercise programme, (2) introduce the key research designs and corresponding statistical models with an emphasis on randomised controlled designs with or without multiple pretests and post-tests, crossover designs and repeated measures designs, (3) discuss advantages and disadvantages of multiple methods of categorising exercise response levels—a topic that is of particular interest for personalised exercise medicine and (4) outline approaches that may identify determinants and modifiers of CRF exercise response. We also summarise gaps in knowledge and recommend future research to better understand exercise response variability.

Research in nutritional supplements and nutraceuticals for health, physical activity, and performance: moving forward 1

This Horizons is part of a series that identifies key, forward-thinking research questions and challenges that need to be addressed. Specifically, this Horizons paper discusses research in nutritional supplements and nutraceuticals for health, physical activity, and performance, and is the product of a discussion by an expert panel that took place in January 2018 prior to the Canadian Nutrition Society Thematic Conference "Advances in Sport Nutrition from Daily Living to High Performance Sport". The objective of this Horizons paper was to identify core considerations for future studies for this research area, and how scientists can be leaders in the field to ensure the best quality science is available for decision makers. It is strongly recommended that the various elements highlighted throughout this Horizons paper will increase the awareness of the significant before-, during-, and after-research due-diligence required to produce research of the highest quality. While it is recognized that many scientists will not be able to meet all of these aspects, it is nonetheless important to consider the points outlined and to recognize that those elements that are not met in studies may be significant limitations. Highlights Research questions that are hypothesis-driven are the strongest, and when combined with careful planning of the study, the result will often be of the best quality. Studies with a strong experimental design help discern between evidence-based findings and those that have not been substantiated.

Gene expression variability in human skeletal muscle transcriptome responses to acute resistance exercise

NEW FINDINGS

What is the central question of this study? Does exercise, independent of random error and within-subject variability, contribute to the variability in gene expression responses to an acute bout of resistance exercise? What is the main finding and its importance? A reanalysis of publicly available microarray data revealed that variability in observed gene expression responses for a subset of genes could be partially attributable to an effect of acute resistance exercise. These finding support the notion that individual responsiveness explains a portion of the variability in observed gene expression responses to acute resistance exercise.

ABSTRACT

The purpose of this study was to use publicly available transcriptomic data to determine whether variability in gene expression responses to an acute bout of acute resistance exercise (ARE) can be attributable to an effect of ARE per se. We examined microarray data from a previous study that collected skeletal muscle biopsies before and 24 h after ARE or a no-exercise time-matched control period (CTL). By subtracting the standard deviation in the observed responses to CTL from ARE, we determined that ARE contributed to the variability in the observed gene expression responses for many (∼31,000), but not all, transcripts included on the Affymetrix Human Genome chips. ARE had a large effect on variability in the observed gene expression responses in 1290 genes that was not attributed to any technical/biological variability associated with repeated measurements. Pathway analysis using WebGestalt revealed that several of these 1290 genes are involved in pathways known to regulate skeletal muscle adaptations to chronic resistance training. These results suggest that variability in the observed gene expression responses for a subset of genes could be partially attributable to an effect of ARE.

A Multi-Center Comparison of O2peak Trainability Between Interval Training and Moderate Intensity Continuous Training

There is heterogeneity in the observed O_2peak response to similar exercise training, and different exercise approaches produce variable degrees of exercise response (trainability). The aim of this study was to combine data from different laboratories to compare O_2peak trainability between various volumes of interval training and Moderate Intensity Continuous Training (MICT). For interval training, volumes were classified by the duration of total interval time. High-volume High Intensity Interval Training (HIIT) included studies that had participants complete more than 15 min of high intensity efforts per session. Low-volume HIIT/Sprint Interval Training (SIT) included studies using less than 15 min of high intensity efforts per session. In total, 677 participants across 18 aerobic exercise training interventions from eight different universities in five countries were included in the analysis. Participants had completed 3 weeks or more of either high-volume HIIT (n = 299), low-volume HIIT/SIT (n = 116), or MICT (n = 262) and were predominately men (n = 495) with a mix of healthy, elderly and clinical populations. Each training intervention improved mean O_2peak at the group level (P < 0.001). After adjusting for covariates, high-volume HIIT had a significantly greater (P < 0.05) absolute O_2peak increase (0.29 L/min) compared to MICT (0.20 L/min) and low-volume HIIT/SIT (0.18 L/min). Adjusted relative O_2peak increase was also significantly greater (P < 0.01) in high-volume HIIT (3.3 ml/kg/min) than MICT (2.4 ml/kg/min) and insignificantly greater (P = 0.09) than low-volume HIIT/SIT (2.5 mL/kg/min). Based on a high threshold for a likely response (technical error of measurement plus the minimal clinically important difference), high-volume HIIT had significantly more (P < 0.01) likely responders (31%) compared to low-volume HIIT/SIT (16%) and MICT (21%). Covariates such as age, sex, the individual study, population group, sessions per week, study duration and the average between pre and post O_2peak explained only 17.3% of the variance in O_2peak trainability. In conclusion, high-volume HIIT had more likely responders to improvements in O_2peak compared to low-volume HIIT/SIT and MICT.

The application of repeated testing and monoexponential regressions to classify individual cardiorespiratory fitness responses to exercise training

PURPOSE

We tested the hypothesis that monoexponential regressions will increase the certainty in response estimates and confidence in classification of cardiorespiratory fitness (CRF) responses compared to a recently proposed linear regression approach.

METHODS

We used data from a previously published RCT that involved 24 weeks of training at high amount-high intensity (HAHI; N = 28), high amount-low intensity (HALI; N = 48), or low amount-low intensity (LALI; N = 33). CRF was measured at 0, 4, 8, 16, and 24 weeks. We fit the repeated CRF measures with monoexponential and linear regressions, and calculated individual response estimates, the error in these estimates (TE_MONOEXP and TE_SLOPE, respectively), and 95% confidence intervals (CIs). Individuals were classified as responders, uncertain, or non-responders based on where their CI lay relative to a minimum clinically important difference. Additionally, responses were classified using observed pre-post-changes and the typical error of measurement.

RESULTS

Comparing the error in response estimates revealed that monoexponential regressions were a better fit than linear regressions for the majority of individual responses (N = 81/109) and mean CRF data (mean TE_MONOEXP:TE_SLOPE; HAHI = 2.00:2.58, HALI = 1.91:2.46, LALI = 1.63:2.18; all p < 0.01). Fewer individuals were confidently classified as responders with linear regressions (N = 29/109) compared to monoexponential (N = 55/109). Additionally, response estimates were highly correlated across all three approaches (all r > 0.92).

CONCLUSIONS

Future studies should determine the type of regression that best fits their data prior to classifying responses. The similarity in response estimates and classification from regressions and observed pre-post-changes questions the purported benefit of using repeated measures to characterize CRF responses to training.

Statistical Considerations for Exercise Protocols Aimed at Measuring Trainability

The individual response to exercise training is of great interest with methods that have been proposed to measure this response reviewed in this paper. However, individual training response estimates may be biased by various sources of variability present in exercise studies, and in particular by within-subject variability. We propose the use of protocols that can separate trainability from within-subject variability.

The magnitude of Yo-Yo test improvements following an aerobic training intervention are associated with total genotype score

Recent research has demonstrated that there is considerable inter-individual variation in the response to aerobic training, and that this variation is partially mediated by genetic factors. As such, we aimed to investigate if a genetic based algorithm successfully predicted the magnitude of improvements following eight-weeks of aerobic training in youth soccer players. A genetic test was utilised to examine five single nucleotide polymorphisms (VEGF rs2010963, ADRB2 rs1042713 and rs1042714, CRP rs1205 & PPARGC1A rs8192678), whose occurrence is believed to impact aerobic training adaptations. 42 male soccer players (17.0 ± 1y, 176 ± 6 cm, 69 ± 9 kg) were tested and stratified into three different Total Genotype Score groups; "low", "medium"and "high", based on the possession of favourable polymorphisms. Subjects underwent two Yo-Yo tests separated by eight-weeks of sports-specific aerobic training. Overall, there were no significant differences between the genotype groups in pre-training Yo-Yo performance, but evident between-group response differentials emerged in post-training Yo-Yo test performance. Subjects in the "high" group saw much larger improvements (58%) than those in the 'medium" (35%) and "low" (7%) groups. There were significant (p<0.05) differences between the groups in the magnitude of improvement, with athletes in the "high" and medium group having larger improvements than the "low" group (d = 2.59 "high" vs "low"; d = 1.32 "medium" vs "low"). In conclusion, the magnitude of improvements in aerobic fitness following a training intervention were associated with a genetic algorithm comprised of five single nucleotide polymorphisms. This information could lead to the development of more individualised aerobic training designs, targeting optimal fitness adaptations.

Changes in Metabolic Syndrome Severity Following Individualized Versus Standardized Exercise Prescription: A Feasibility Study

This study sought to investigate the efficacy of standardized versus individualized exercise intensity prescription on metabolic syndrome (MetS) severity following a 12-week exercise intervention. A total of 38 experimental participants (47.8 ± 12.2 yr, 170.7 ± 8.0 cm, 82.6 ± 18.7 kg, 26.9 ± 6.7 mL·k⁻¹·min⁻¹) were randomized to one of two exercise interventions (exercise intensity prescribed using heart rate reserve or ventilatory threshold). Following the 12-week intervention, MetS z-score was significantly improved for the standardized (−2.0 ± 3.1 to −2.8 ± 2.8 [p = 0.01]) and individualized (−3.3 ± 2.3 to −3.9 ± 2.2 [p = 0.04]) groups. When separating participants based on prevalence of MetS at baseline and MetS z-score responsiveness, there were six and three participants in the standardized and individualized groups, respectively, with three or more MetS risk factors. Of the six participants in the standardized group, 83% (5/6) of the participants were considered responders, whereas 100% (3/3) of the individualized participants were responders. Furthermore, only 17% (1/6) of the participants with MetS at baseline in the standardized group no longer had symptoms of MetS following the intervention. In the individualized group, 67% (2/3) of participants with baseline MetS were not considered to have MetS at week 12. These findings suggest that an individualized approach to the exercise intensity prescription may ameliorate the severity of MetS.

Reliability of Submaximal Yo-Yo Tests in 9- to 16-Year-Old Untrained Schoolchildren

PURPOSE

To examine the reliability of age-adapted submaximal Yo-Yo (Yo-Yo_submax) intermittent tests in untrained schoolchildren aged 9-16 years (n = 139; 72 boys and 67 girls) and within children with high and low percentage of body fat (%BF).

METHODS

Yo-Yo intermittent recovery level 1 children's (YYIR1C), Yo-Yo intermittent endurance level 1 (YYIE1), and Yo-Yo intermittent endurance level 2 (YYIE2) tests were performed 7 days apart by 9- to 11-, 12- to 13-, and 14- to 16-year-old children, respectively. Reliability was tested for Yo-Yo_submax heart rate (HR_submax), peak HR, and maximal distance.

RESULTS

HR_submax typical errors of measurement (TEM) in YYIR1C, YYIE1, and YYIE2 were 2.2% (1.7%-2.9%), 2.4% (1.9%-3.3%), 1.9% (1.6%-2.5%) and 2.4% (1.9%-3.3%), 2.4% (1.9%-3.2%), 1.9% (1.5%-2.4%) for girls and boys, respectively. HR_submax intraclass correlation coefficient values were good to excellent (.62-.87) in all age groups and in schoolchildren of different %BF. TEM for HR_submax ranged from 2.1% to 2.3% in high and low %BF groups. Maximal distance intraclass correlation coefficients were excellent and TEM values ranged from 11% to 12% in both %BF groups. HR_submax was moderately to largely associated (r = -.46 to -.64; P < .002) with Yo-Yo maximal distance across the age groups.

CONCLUSION

Yo-Yo_submax tests are a reliable tool providing useful and sustainable aerobic performance testing in physical education, irrespective of individual %BF.

Inter-individual differences in weight change following exercise interventions: a systematic review and meta-analysis of randomized controlled trials

Previous reports of substantial inter-individual differences in weight change following an exercise intervention are often based solely on the observed responses in the intervention group. Therefore, we aimed to quantify the magnitude of inter-individual differences in exercise-mediated weight change. We synthesized randomized controlled trials (RCTs) of structured, supervised exercise interventions. Fourteen electronic databases were searched for relevant studies published up to March 2017. Search terms focused on structured training, RCTs and body weight. We then sifted these results for those RCTs (n = 12, 1500 participants) that included relevant comparator group data. Standard deviations (SDs) of weight change were extracted, thereby allowing the SD for true inter-individual differences in weight loss to be calculated for each study. Using a random effects meta-analysis, the pooled SD (95% CI) for true individual responses was 0.8 (-0.9 to 1.4) kg. The 95% prediction interval (based on 2SDs) for true inter-individual responses was -2.8 to 3.6 kg. The probability (% chance) that the true individual response variability would be clinically meaningful (>2.5 kg) in a future study in similar settings was 23% ('unlikely'). Therefore, we conclude that evidence is limited for the notion that there are clinically important individual differences in exercise-mediated weight change.

Repeated testing for the assessment of individual response to exercise training

Observed response to regular exercise training differs widely between individuals even in tightly controlled research settings. However, the respective contributions of random error and true interindividual differences as well as the relative frequency of nonresponders are disputed. Specific challenges of analyses on the individual level as well as a striking heterogeneity in definitions may partly explain these inconsistent results. Repeated testing during the training phase specifically addresses the requirements of analyses on the individual level. Here we report a first implementation of this innovative design amendment in a head-to-head comparison of existing analytical approaches. To allow for comparative implementation of approaches we conducted a controlled endurance training trial (1 yr walking/jogging, 3 days/wk for 45 min with 60% heart rate reserve) in healthy, untrained subjects ( n = 36, age = 46 ± 8 yr; body mass index 24.7 ± 2.7 kg/m2; V̇o2max 36.6 ± 5.4). In the training group additional V̇o2max tests were conducted after 3, 6, and 9 mo. Duration of the control condition was 6 mo due to ethical constraints. General efficacy of the training intervention could be verified by a significant increase in V̇o2max in the training group ( P < 0.001 vs. control). Individual training response of relevant magnitude (>0.2 × baseline variability in V̇o2max) could be demonstrated by several approaches. Regarding the classification of individuals, only 11 of 20 subjects were consistently classified, demonstrating remarkable disagreement between approaches. These results are in support of relevant interindividual variability in training efficacy and stress the limitations of a responder classification. Moreover, this proof-of-concept underlines the need for tailored methodological approaches for well-defined problems.

NEW & NOTEWORTHY This work reports a first implementation of a repeated testing training trial for the investigation of individual response. This design amendment was recently proposed to address specifically the statistical requirements of analyses on the individual level. Moreover, a comprehensive comparison of previously published methods exemplifies the striking heterogeneity of existing approaches.

Response heterogeneity: Challenges for personalised medicine and big data approaches in psychiatry and chronic pain

Response rates to available treatments for psychological and chronic pain disorders are poor, and there is a substantial burden of suffering and disability for patients, who often cycle through several rounds of ineffective treatment. As individuals presenting to the clinic with symptoms of these disorders are likely to be heterogeneous, there is considerable interest in the possibility that different constellations of signs could be used to identify subgroups of patients that might preferentially benefit from particular kinds of treatment. To this end, there has been a recent focus on the application of machine learning methods to attempt to identify sets of predictor variables (demographic, genetic, etc.) that could be used to target individuals towards treatments that are more likely to work for them in the first instance. Importantly, the training of such models generally relies on datasets where groups of individual predictor variables are labelled with a binary outcome category - usually 'responder' or 'non-responder' (to a particular treatment). However, as previously highlighted in other areas of medicine, there is a basic statistical problem in classifying individuals as 'responding' to a particular treatment on the basis of data from conventional randomized controlled trials. Specifically, insufficient information on the partition of variance components in individual symptom changes mean that it is inappropriate to consider data from the active treatment arm alone in this way. This may be particularly problematic in the case of psychiatric and chronic pain symptom data, where both within-subject variability and measurement error are likely to be high. Here, we outline some possible solutions to this problem in terms of dataset design and machine learning methodology, and conclude that it is important to carefully consider the kind of inferences that particular training data are able to afford, especially in arenas where the potential clinical benefit is so large.

Reversing the Cardiac Effects of Sedentary Aging in Middle Age-A Randomized Controlled Trial: Implications For Heart Failure Prevention

BACKGROUND

Poor fitness in middle age is a risk factor for heart failure, particularly heart failure with a preserved ejection fraction. The development of heart failure with a preserved ejection fraction is likely mediated through increased left ventricular (LV) stiffness, a consequence of sedentary aging. In a prospective, parallel group, randomized controlled trial, we examined the effect of 2 years of supervised high-intensity exercise training on LV stiffness.

METHODS

Sixty-one (48% male) healthy, sedentary, middle-aged participants (53±5 years) were randomly assigned to either 2 years of exercise training (n=34) or attention control (control; n=27). Right heart catheterization and 3-dimensional echocardiography were performed with preload manipulations to define LV end-diastolic pressure-volume relationships and Frank-Starling curves. LV stiffness was calculated by curve fit of the diastolic pressure-volume curve. Maximal oxygen uptake (Vo₂max) was measured to quantify changes in fitness.

RESULTS

Fifty-three participants completed the study. Adherence to prescribed exercise sessions was 88±11%. Vo₂max increased by 18% (exercise training: pre 29.0±4.8 to post 34.4±6.4; control: pre 29.5±5.3 to post 28.7±5.4, group×time P<0.001) and LV stiffness was reduced (right/downward shift in the end-diastolic pressure-volume relationships; preexercise training stiffness constant 0.072±0.037 to postexercise training 0.051±0.0268, P=0.0018), whereas there was no change in controls (group×time P<0.001; pre stiffness constant 0.0635±0.026 to post 0.062±0.031, P=0.83). Exercise increased LV end-diastolic volume (group×time P<0.001), whereas pulmonary capillary wedge pressure was unchanged, providing greater stroke volume for any given filling pressure (loading×group×time P=0.007).

CONCLUSIONS

In previously sedentary healthy middle-aged adults, 2 years of exercise training improved maximal oxygen uptake and decreased cardiac stiffness. Regular exercise training may provide protection against the future risk of heart failure with a preserved ejection fraction by preventing the increase in cardiac stiffness attributable to sedentary aging.

CLINICAL TRIAL REGISTRATION

URL: https://www.clinicaltrials.gov. Unique identifier: NCT02039154.

Energy compensation after sprint- and high-intensity interval training

Background

Many individuals lose less weight than expected in response to exercise interventions when considering the increased energy expenditure of exercise (ExEE). This is due to energy compensation in response to ExEE, which may include increases in energy intake (EI) and decreases in non-exercise physical activity (NEPA). We examined the degree of energy compensation in healthy young men and women in response to interval training.

Methods

Data were examined from a prior study in which 24 participants (mean age, BMI, & VO₂max = 28 yrs, 27.7 kg•m^-2, and 32 mL∙kg^-1∙min^-1) completed either 4 weeks of sprint-interval training or high-intensity interval training. Energy compensation was calculated from changes in body composition (air displacement plethysmography) and exercise energy expenditure was calculated from mean heart rate based on the heart rate-VO₂ relationship. Differences between high (≥ 100%) and low (< 100%) levels of energy compensation were assessed. Linear regressions were utilized to determine associations between energy compensation and ΔVO₂max, ΔEI, ΔNEPA, and Δresting metabolic rate.

Results

Very large individual differences in energy compensation were noted. In comparison to individuals with low levels of compensation, individuals with high levels of energy compensation gained fat mass, lost fat-free mass, and had lower change scores for VO₂max and NEPA. Linear regression results indicated that lower levels of energy compensation were associated with increases in ΔVO₂max (p < 0.001) and ΔNEPA (p < 0.001).

Conclusions

Considerable variation exists in response to short-term, low dose interval training. In agreement with prior work, increases in ΔVO₂max and ΔNEPA were associated with lower energy compensation. Future studies should focus on identifying if a dose-response relationship for energy compensation exists in response to interval training, and what underlying mechanisms and participant traits contribute to the large variation between individuals.