Posted on

Playing The Telephone Game With Research Studies

Recently, I stumbled upon an article that professes to know “The Best Exercise For Aging Muscles.”  My immediate reaction whenever I see the term “best” attached to a complex phenomenon like aging is…

The article references a study from Cell Metabolism titled “Enhanced Protein Translation Underlies Improved Metabolic and Physical Adaptations to Different Exercise Training Modes in Young and Old Humans”.  Whether it’s politics or scientific research, our conversations often overlook important details and contextual elements that would otherwise cloud the lens through which we view the world.  Certainty and superficiality are usually a package deal.  Without reviewing the full text of a study, “science” can play out like a game of telephone.  Research findings may metamorphose quite extensively en route to the end user.  The following analysis compares the above New York Times article with the study it references while seeking to reconcile the two sources of information.  The point here isn’t that the article and the study are good or bad but to demonstrate what caveats better inform the continuous streams of data with which we are continuously bombarded.

Study Design

  • What the article says:
    • “So researchers at the Mayo Clinic in Rochester, Minn., recently conducted an experiment on the cells of 72 healthy but sedentary men and women who were 30 or younger or older than 64. After baseline measures were established for their aerobic fitness, their blood-sugar levels and the gene activity and mitochondrial health in their muscle cells, the volunteers were randomly assigned to a particular exercise regimen.
    • Some of them did vigorous weight training several times a week; some did brief interval training three times a week on stationary bicycles (pedaling hard for four minutes, resting for three and then repeating that sequence three more times); some rode stationary bikes at a moderate pace for 30 minutes a few times a week and lifted weights lightly on other days. A fourth group, the control, did not exercise.”
  • What the study says:
    • “Participants were recruited into two distinct age groups: young (18–30 years) or older (65–80 years) with a goal of an equal number of men and women. The final groups were approximately balanced for sex, and all women in the older group were postmenopausal. Exclusion criteria were structured regular exercise (>20 min, twice weekly), cardiovascular disease, metabolic diseases (type 2 diabetes mellitus, fasting blood glucose > 110 mg/dL, and untreated hypothyroidism or hyperthyroidism), renal disease, high body mass index (BMI > 32 kg/m2 ), implanted metal devices, pregnancy, smoking, and history of blood clotting disorders. Exclusionary medication included anticoagulants, insulin, insulin sensitizers, corticosteroids, sulfonylureas, barbiturates, peroxisome proliferator-activated receptor g agonists, b blockers, opiates, and tricyclic antidepressants.
    • High-intensity interval training consisted of 3 days per week of intervals on an electronically braked cycle ergometer (Monday, Wednesday and Friday) and 2 days per week on motorized treadmill walking (Tuesday and Thursday). The interval protocol was a 10-minute warm-up followed by 4 cycles of 4-minute high intervals (> 90%) with 3- minute rest (pedaling at no load) then a 5-minute cool down. The time per session at high intensity was 16 minutes. The treadmill protocol was a self-selected walking pace (2-4 mph) with a 10-minute warm-up, 45 minutes at incline at 70% VO2 peak then a 5-minute cool down. The resistance protocol was weight training for 60 minutes on 4-days per week of lower (Monday and Thursday) and upper body exercises (Tuesday and Friday). Participants were instructed on proper lifting technique and performed 8-12 repetitions per exercise with one-minute rest between sets. Participants completed two sets of each exercise for week 1, three sets for week 2, and four sets for weeks 3-12. Weights were increased when participants could perform 12 repetitions while maintaining good form. Lower body exercises were leg press, toe raise, lunge, abdominal crunch, leg extension and leg curl. Upper body exercises were chest press, lat pull down, incline chest press, seated row, lateral raise, biceps curl and triceps push down. The combined protocol was 30 minutes of cycling 5-days per week (Monday through Friday) followed by 30 minutes of weight lifting. The cycling protocol was a five-minute warm-up, 20 minutes at 70% VO2 peak, then 5 minutes of cool down. The weight lifting was a 4-day program with lower body (Monday and Thursday) and upper body exercises (Tuesday and Friday). Lower body exercises were leg press, abdominal crunch, leg extension and leg curls. Upper body exercises were chest press, lat pull down, triceps extensions and biceps curls.”

To its credit, the article acknowledges that the study was conducted on healthy subjects.  This qualification is significant because HIIT as defined in the study is probably not advisable for an untrained elderly person with an extensive medical history.  The elderly people who would most benefit from a responsible exercise program are often those being medically treated for the conditions banned by the study.  As the game of telephone continues to play out on various social media platforms, the final message might become something like “HIIT is the best form of training for elderly people.”  Even medical professionals are sometimes guilty of this type of oversimplification.  Clinicians are typically pressed for time and sifting through even a single journal article requires serious commitment.  Finite time and difficult readability aren’t excuses, just a reality that challenges the seamlessness between research and clinical practice.

Details about the various protocols are quoted above because strength training, interval training, and combined training aren’t standardized phenomena.  The specific exercises, volumes, and intensities utilized in different trials ensure that the construct of “strength training”, as an example, is a moving target.  Notice that the HIIT group performed two, hour-long (including warm-up and cool down) incline treadmill-walking sessions per week in addition to the three weekly interval cycling sessions. To be clear, the HIIT group did more than just interval training as also occurred in highly renowned Tabata study.  The HIIT group performed a mix of high(er) and low intensity endurance training, which is effectively what endurance coaches have been doing for decades.  The combined training group performed similar weight training workouts to the resistance only group but presumably at around 60% the volume (30 mins 5x/week vs. 60 mins 4x/week).  The combined training group also cycled 5x/week at a low to moderate intensity for 30 minutes, including warm-up and cool down.  Without even accounting for the interval training sessions performed on the cycle ergometer, the HIIT group performed 80% (2/2.5 hrs) of the combined group’s endurance training volume (as quantified by time).  Moreover, the HIIT group’s lower intensity work was performed on a treadmill, not a cycle ergometer as occurred in the combined group.  Consequently, based on the study design we can’t really draw conclusions about the efficacy of interval training relative to steady state training, which is what some people will inevitably do.  To be fair, such a distinction might not be the point.


  • What The Article Says:
    • “There were some unsurprising differences: The gains in muscle mass and strength were greater for those who exercised only with weights, while interval training had the strongest influence on endurance.
    • But more unexpected results were found in the biopsied muscle cells. Among the younger subjects who went through interval training, the activity levels had changed in 274 genes, compared with 170 genes for those who exercised more moderately and 74 for the weight lifters. Among the older cohort, almost 400 genes were working differently now, compared with 33 for the weight lifters and only 19 for the moderate exercisers.”
  • What The Study Says:
    • “We investigated the extent to which changes in mRNA coincided with phenotypes to further understand the regulation of skeletal muscle changes with age and adaptations to exercise. We performed RNA sequencing on baseline and post-exercise training skeletal muscle biopsies to assess whether transcript levels account for aging or training phenotypes of mitochondria, muscle hypertrophy, and insulin sensitivity. At baseline, when compared to young, 267 gene transcripts were lower and 166 were higher in older people (Figure S5A). Several mitochondrial-, insulin signaling-, and muscle growth-related genes were downregulated with age (Figure S5A). In contrast, among all training regimens, HIIT increased the expression of the largest number of genes in both young and older, especially in mitochondrial, muscle growth, and insulin signaling pathways in older adults (Figures 4A and 4B). In the older, HIIT increased 22 mitochondrial genes, including those involved with translational regulation (ribosomes MT-RNR1 and 2) and mitochondrial tRNA transferase for methionine (MT-TG), leucine (MT-TL1), valine (MT-TV), glycine (MT-TG), and arginine (MT-TR). When compared to HIIT, RT increased 35% and 70% fewer genes in young and old, respectively (Figures 4C and 4D), and CT increased 28% and 84% fewer genes in young and old, respectively (Figures 4E and 4F). These data demonstrate a varied response of gene transcripts based on exercise mode between young and older adults, and the greatest increase was following HIIT in older adults.
    • Next, we determined whether training-induced gene sets are specific to training modes in young and older adults. The young had 274, 74, and 170 genes uniquely increased by HIIT, RT, and CT, respectively (Figure 4G). The older had 396, 33, and 19 genes uniquely increased by HIIT, RT, and CT, respectively (Figure 4H). Taken together, these data show that HIIT induced the largest gene expression change regardless of age. In older adults, the changes in gene expression with HIIT completely A B C D E F G H I K L J Figure 4. Muscle Gene Expression Changes with Exercise Training (A–F) Genes that were differentially expressed following high-intensity interval training (HIIT) in the young (A) or older (B), resistance training in the young (C) or older (D), and combined training in the young (E) or older (F) using an adjusted p value of %0.05 and an absolute fold change of R0.5 were annotated according to their mitochondrial specificity (using MitoCarta) and molecular function (using KEGG). Mito stands for mitochondrial. (G and H) Overlap of genes upregulated with different modes of exercise training in younger (G) and older (H) participants. (I) Overlap of genes that were upregulated with HIIT in older adults and any type of exercise training in younger adults. (J) Gene set enrichment analysis of baseline gene expression differences between young and old participants against genes that were upregulated with HIIT in older participants. Genes that increased expression with age were more likely to increase their expression with HIIT in older participants. (K) A ‘‘universal exercise training response gene set’’ was derived by looking for genes that increased with exercise with an adjusted p value of %0.05 and an absolute fold change of R0.3 in all groups. Gene Ontology (GO) process annotations enriched for this universal exercise training response gene set was derived using MetaCore software configured with an adjusted p value threshold of %0.05. (L) Ingenuity Pathway Analysis (QIAGEN) was used to detect up stream regulators of the ‘‘universal exercise training response gene set.’’ 586 Cell Metabolism 25, 581–592, March 7, 2017 subsumes CT and RT changes. Given that older HIIT produced the largest gene expression change, we assessed whether these genes were unique or overlapping with the younger training groups. One-third of the older HIIT genes (181 out of 553) were also shared by the young HIIT group, and 114 of these were shared with young RT and CT groups (Figure 4I). Another third of older HIIT genes were unique to that group (186 out of 553; Figure 4I). Taken together, these data suggest that a large portion of older HIIT genes is age specific.”

The results of the study were somewhat predictable as the author of the article intimates.  The most significant finding, it seems, is the relative upregulation of gene activity in the elderly HIIT group.  While it’s remarkable that researchers developed the ability to isolate so many individual genes, post-training muscle biopsies MIGHT serve as a proxy for metrics, albeit ones that weren’t measured here, more predictive of function, performance, health, and independence in elderly people.  Moreover, cellular physiology in the muscle, particularly at the mitochondrial level, is but a small piece of the adaptive response to training.  Lots of other genetic and systemic processes are involved, more than can be measured in a single study.

Our genetic understanding of performance is still largely limited in its applicability.  As David Epstein points out in Sports Genea stop watch is currently just as informative, perhaps more so, than a genetic test.  Genetic test results are probably less applicable an outcome measure in sport than even something like VO2 max and the best endurance athletes aren’t handing their medals to slower competitors who possess higher VO2 scores.  See Magness’ “A Brief Rant Against VO2 Max” as some of his points are applicable to this scenario.  What constitutes a “stop watch” for elderly people who exercise is beyond the scope of this analysis but it’s as much a moral, cultural, and political question as it is a physical one.  A muscle biopsy or mitochondrial snapshot does not sufficiently capture the complex nature of this question.  A well-designed study is not equipped to “answer” a multi-disciplinary question so researchers shouldn’t be criticized for science’s inherent limitations.  Consumers of research just need to be careful not to misinterpret highly contextualized data.

Researchers, clinicians, and coaches have different goals and therefore track different metrics.  Genetic profiles and VO2 maxes aren’t “bad” data.  They’re completely relevant when the goal is to learn more about the genetic and aerobic responses to exercise respectively.  Researchers generally don’t make value judgments when they report their findings.  They mainly describe what they observed and confess the limitations of their methodology.  Researchers produce information.  It’s up to the consumers of research to determine what to do with this information.  I’ve yet to encounter a researcher refer to anything as the “best”.


  1. Details matter.  Headlines rarely tell enough of the story.
  2. Researchers are more like reporters than editorialists.  The further detached the original researcher from a messenger, the more likely the information is to be inaccurate or editorialized; just like the telephone game.
  3. The study design does not allow for confident conclusions to be drawn about the effectiveness of interval training compared to steady state training or about combined training compared to training single qualities
  4. It is not clear that this study would necessarily alter the manner in which competent trainers and clinicians are already managing elderly people’s exercise programs.  Moreover, repeating the same protocol 3-5x/week for several months on a stationary bike might not be an effective way to maintain compliance with an exercise program when researchers aren’t paying the trainees.
  5. The goal of a training or exercise program is not necessarily to maximize physiology at the cellular level.  VO2 max testing and muscle biopsies aren’t competitive events.   These types of variables can still be informative provided they are guided by proper intent.  As an example, elderly people are more likely to have a catastrophic outcome in response to a fall so while HIIT might seem superior to strength training under a very particular set of parameters (e.g. a study that focused on mitochondrial gene activity), the importance of strength training to improve muscle mass and bone mineral density in the elderly should not be overlooked.  This study didn’t assess fall risk and conditions like osteoporosis, however, so strength training didn’t substantially alter many of the variables measured by the authors.
  6. Some proportion of high(er) intensity work seems to help maximize the results of a training program irrespective of age
  7. Healthy elderly people should not shy away from high(er) intensity training
  8. One wonders how the results would have varied if the HIIT group performed bouts of even higher intensity work, say <15 seconds bursts closer to an all out effort.  Intensity is a relative term and the HIIT protocol in this study wouldn’t qualify as high intensity work in other settings.
  9. Genetic screening is becoming increasingly sophisticated.  We still don’t sufficiently understand how individual genes interact to influence a construct like “health”.  Genetic reductionism is more possible now but the ability to measure something doesn’t guarantee the relevance of an outcome measure (e.g. VO2 max in endurance athletes).
  10. Always ask how is a particular piece of information actionable
  11. Elderly people aren’t a different species.  Their training shouldn’t look much different from anybody else’s provided they are adequately prepared for the constituent elements of their program.  Here are some examples:

Ace doesn’t just read the headlines…