Sunday, April 1, 2018

Just In Time? Too Late?

The main results from Howden et al., the paper that inspired this blog post. Middle-aged participants who engaged in aerobic training improved VO2max and left ventricular end-diastolic volume, both markers for reduced incidence of heart disease, whereas participants who engaged in flexibility and strength training did not.

Between ages 21 and 56, my level of exercise was inadequate to non-existent. For a few years thereafter, I went from exercise plan to exercise plan trying to get something to stick, until age 61, when I finally restarted cycling on a regular basis, an exercise program which I have maintained fairly well for almost eight years. Thus, a link on Facebook to an article on BBC News with the title of "Middle-aged can reverse heart risk with exercise, study suggests" had its desired effect: it caught my eye and I clicked through and read the article. Reading the article left me concerned: "...the aerobic exercise regimen should be started before the age of 65 when the heart appears to retain 'plasticity' and the ability to remodel itself." Had I waited too long to restart exercising, or had I managed to just squeak under the wire? Could I really believe any of this anyway? As I have blogged here many times, when encountering an article in the popular press on a scientific study, it is important to go back to the original article from a professional, peer-reviewed journal to make sure that its results were accurately reported, and then to evaluate the study described therein to make sure its conclusions are adequately supported, and that I did. I read the original article and determined that it only investigated people between the ages of 51 and 55. The statement that starting exercise later is is less valuable is in the paper, but it is based on comparing the results in this paper to those in an earlier paper from the same team which investigated people between the ages of 68 and 74. In my discussion of these papers, I will first summarize the claims of the papers, discuss the papers critically, describing their limitations, and finally, I will summarize my "bottom line", what results from these two studies I personally find convincing. (The impatient, who are willing to trust my judgement, should feel free to skip to this Bottom Line.)

Summary


Howden et al.1 is the original paper which formed the basis for the article in BBC News. They ask the following question:
"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."

Fujimoto et al.2 is the earlier paper from the same group. They ask, what is to my eye, a functionally equivalent question:
"Healthy but sedentary aging leads to cardiovascular stiffening, whereas life-long endurance training preserves left ventricular (LV) compliance. However, it is unknown whether exercise training started later in life can reverse the effects of sedentary behavior on the heart." 

The group that published these two papers is a cardiology group. I have no expertise in cardiology so cannot comment critically on this part of their work, I simply have to take the authors at their word. Given that, my understanding of what they are saying is that one form of heart failure is associated with a decrease in left ventricular end-diastolic volume or LVEDV. LVEDV decreases with age, but is higher in the elderly who have exercised their whole lives ("master athletes"). The question both papers are asking is, for one of those unfortunates who have not exercised, does exercise later in life reverse the decrease in LVEDV? The hope is that if exercise can increase LVEDV, this will result in a lower probability of the development of this one kind of heart failure and thus increase longevity.

In addition to LVEDV, the authors of both papers measured a number of other parameters. The one that caught my eye was VO2max, because VO2max is probably the most common biomarker for the beneficial effects of exercise on longevity, one that I have discussed quite often on this blog. The authors barely mentioned their results with VO2max, but I will devote significant attention to them.

The reason for considering Howden et al. and Fujimoto et al. together is the assumption that they did roughly equivalent studies on two age groups and got different results, thus the statement in the BBC News article that exercise after age 65 may be too late to reverse heart damage. There are, however, other differences between the papers that weaken this conclusion. A partial comparison of the two papers is in the next figure:


What we note is that the age groups studied is different and that for the younger age group, LVEDV increases significantly, whereas for the older age group, the difference is not statistically significant. We also note that Howden et al. had many more participants, which by itself will increase the statistical significance of their results, a fact I will discuss in the next section. Finally, we note that VO2max, the workhorse biomarker used in a large number of studies as a proxy for health benefits increased in both studies by similar amounts. (Not shown in the figure is that both increases are statistically significant, P=0.0001 for Howden et al. and P=0.001 for the smaller study reported in Fujimoto et al.)

Another difference between the two studies is that they used different exercise plans:

The first number in each column in the above is the number of rides of this kind done per month. The intensity of the ride is given in the column header. Thus, 15 @ 25 min in the Base Pace column means that 15 rides a month are done at Base Pace, where each ride is 25 minutes long. For Intervals, 3 @ 8 x (30 s "on" 90s "off") means that 3 times a month a ride is done consisting of 8 Intervals, where each interval consists of 30 seconds at Interval Pace followed by 90 seconds at Recovery Pace. Base Pace corresponds to the Heart Rate Zone (HR) 3 and 4 that I use on this blog. MSS corresponds to the boundary between HR 4 and HR 5. The "on" part of the intervals are done in HR 5c. Recovery Pace (ridden during the time between intervals and during a recovery ride done the day after intervals, not shown above) are done in Zones 1 and 2. These HRs are so high, I worry that I have misunderstood the authors, I would find it difficult or impossible to complete the workouts above if this is correct. That said, because of these concerns, I read this part of the papers carefully, and can find no other way to interpret what the authors are saying.

The exercise plan in Fujimoto et al. was one year long as compared to Howden et al. which was two years long, but Howden et al. collected results after ten months and all the benefits of exercise had accrued by then, so I only show the results of this first ten months. Fujimoto et al. described their exercise protocol very clearly, so that it was easy for me to put it into the figure above and I am relatively confident I correctly described it. In contrast, I found the description of the exercise plan in Howden et al. ambiguous and confusing, so I am less confident I have accurately described it.

As noted above, I am not a cardiologist so cannot comment on the protocols used in the two papers to measure cardiac parameters, but to my untrained eye, it seemed that the two papers used different protocols.

Finally, there are differences between the papers that concern parts of the studies not relevant to this blog post. Fujimoto et al. also compared active, age matched (elderly) participants, known as master athletes, to the healthy but sedentary participants in their study, and Howden et al. compared the effects of yoga plus strength training to that of the aerobic training. I am  ignoring these parts of the papers.

Critical Analysis 


The two sentences in Howden et al. that inspired the BBC News article are as follows:
"Epidemiological studies show that a measurement of  fitness in middle age is the strongest predictor of future heart failure. Moreover, in observational studies, the dose of exercise associated with reduced heart failure incidence is much higher than that associated with reduced mortality. However, if exercise is started too late in life (ie, after 65 years) in sedentary individuals, there is little effect on LV stiffness [reference to Fujimoto et al.]"

To my eye, both Howden et al. and Fujimoto et al. seem to be well executed studies and the narrow conclusions that the authors draw from their results seem justified. The focus of my critical analysis will thus focus on the comparison between the two papers, made in Howden et al., that is the basis for the assertion that exercise at age 65 and older to too late to protect against one kind of heart failure.

Before starting, note the narrowness of the assertion: the authors do not deny that exercise for those of us older than 65 can decrease mortality overall, but rather that it may be too late to reduce the incidence of one particular disease. In fact, the VO2max results reported in Fujimoto et al. actively affirm the value of exercise for us old folks. That said, is this narrow assertion supported by the results?

The first reservation comes from the fact that these are biomarker studies. What the authors care about is the effect of exercise on the incidence of one kind of heart failure. What the authors measure is LVEDV. How good is the evidence that reducing LVEDV in late middle age (51 to 55) will result in a reduction in heart failure? Note that just because low LVEDV is associated with heart failure does not demonstrate cause and effect. Howden et al. might, in fact, be in a position to test this hypothesis. If they were to follow the 61 participants in this study, the 28 who did yoga and strength training vs the 33 who did aerobic exercise, and if they found that the aerobic group had a lower incidence of heart failure, this would go a long way towards demonstrating the value of exercise in middle age, but of course that study has not (yet) been done. For now, we must accept uncertainty as the inevitable limitation that result from this being a biomarker study.

Even if we accept that the increase in LVEDV caused by exercise results in a reduction in heart failure, the assertion that age 65 is too late for this benefit is suspect. The basis for this assertion is that no improvement in LVEDV was seen in the older participants of Fujimoto et al., whereas improvement in LVEDV was seen in the relatively younger participants of Howden et al. The problem with this assertion is that there are other explanations for the differences observed between the two papers. Firstly, Holden et al. had many more participants than Fujimoto et al., 33 vs 9. Thus, Holden et al. had much more statistical power. There may have been an increase in LVEDV in the older patients in Fujimoto et al., that was hidden by statistical noise, that would have been seen had they looked at more patients. Secondly, if I am correct that the two papers used different techniques for measuring LVEDV, that could also explain the differences. Finally, and most relevant to this blog, the difference between the two papers could be due to different exercise protocols. A lot of the papers I cover here are about precisely that, which exercise protocol produces the the greatest benefit?

Because this is an exercise blog, I'd like to spend a bit more time on the exercise protocols, specifically, how intense they are. (The legend of the figure describing the exercise protocols contains a translation of the intensity scale used in these papers to the heart rate zones I use on this blog.) During the five years I have been writing this blog, I have found the translation of light, moderate, and vigorous exercise intensity recommendations of the medical community to something I can use problematic. That said, I have developed a consensus translation: Light = Heart Rate Zone 1, Moderate = Zone 2, and Vigorous = Zone 4. By that standard, almost all the recommended exercise in these papers would be classified as vigorous. At 5 to 6 hours a week of exercise, this would be more than twice the maximum amount of exercise recommended generally by the medical community. At that level of exercise, I am astonished at the 88% compliance reported.

Bottom Line


The report on Howden et al. in BBC News is overly sensational and in fact adds almost nothing to our understanding of the value of exercise at different ages. Some of this is because the article fails to explain the narrowness of the conclusions of the underlying scientific paper, that it applies to only one kind of heart failure, not overall longevity, and some of this is because of limitations of the paper itself weaken even this very narrow argument. There is extensive evidence in the literature in support of the value of exercise at all ages for increasing longevity, and in fact, when properly interpreted, even the claimed negative result pointed to in the BBC News article supports this conclusion. Thus, even at my ripe old age of 68, I should definitely keep riding.



    1) Howden et al: "Reversing the Cardiac Effects of Sedentary Aging in Middle Age—A Randomized Controlled Trial.
    Implications For Heart Failure Prevention." by Erin J. Howden, Satyam Sarma, Justin S. Lawley, Mildred Opondo, William Cornwell, Douglas Stoller, MD, Marcus A. Urey, Beverley Adams-Huet, and Benjamin D. Levine. Published in Circulation volume 137 [Epub ahead of print]. Year: 2018.
    2) Fujimoto et al: "Cardiovascular Effects of 1 Year of Progressive and Vigorous Exercise Training in Previously Sedentary Individuals Older Than 65 Years of Age" by Naoki Fujimoto, Anand Prasad, Jeffrey L. Hastings, Armin Arbab-Zadeh, Paul S. Bhella, Shigeki Shibata, Dean Palmer, and Benjamin D. Levine. Published in Circulation volume 122 page 1797. Year: 2010.