When I was in college, I had the luxury of being able to
stay incredibly active with collegiate Ultimate Frisbee. And as cliché as it
sounds to be an undergrad at Western, playing Ultimate in the rain in
Bellingham (which we now know is the wettest place in the lower 48 http://www.seattlepi.com/local/connelly/article/Bellingham-tops-list-of-cities-with-least-sunshine-4954055.php),
our team had some ~15 hrs/week of practice, including weekend tournaments all
across the region. It was an exhausting sport! Throughout college, I don’t know
if I ever thought that I could ‘over eat’. At that time, I basically believed
that the more I ate, the better I felt. And even though I look back on this
time with a bit of dread, remembering my hands freezing off when the turf field
froze over and the snow didn’t melt for months, I think of this time as a
luxurious occasion where I literally felt that anything I ate was perfectly
okay.
But, as other collegiate athletes have corroborated, going
from athlete to non-athlete is awful. You go from eating anything and
everything you can possibly get your hands on to, to looking at your gut and
thinking, “huh?” For me, the ‘eat everything and anything’ attitude maintained
for a couple of years after college. I continued to play Ultimate Frisbee for a
year or two, playing for a year with UW while in grad school, then for a club
team over the summers, and my body consumed the energy about as fast as I could
put it in. But at some point, when I realized I needed to focus on graduate
school, when it seemed like I had less time available for the sport, and when I
continually fought off injuries, I found myself lacking the energy and
excitement to get back into it. I continued to play every once in a while and I
got to play for an awesome club team in Germany when I was in Heidelberg (http://www.heidees.de/pages/heidees-news.php),
but my competitive training continued to dwindle. Although it was, and
sometimes still continues to be, a little painful to stop playing, I feel that
I’m learning I need to plan for my health and fitness for the long haul, and training
for Frisbee wasn’t going to be sustainable for me. Kudos to those who’ve made
it work!
But, when I felt that I fully stepped off the Frisbee field,
and I stopped burning thousands of calories a day from training, I did not
become less hungry. A normal daily diet for my years of doing sports still
seemed…normal. And as predictably as an overfilled glass of water will spill over
its edge, I found myself gaining weight, something I never felt was even an
option before. What ensued was a continual battle, that I know many, many,
people are also fighting, where eating became a thought process. I found myself
no longer able to just eat ‘everything and anything’. So, after years of
sports, I joined the many who actually have to think about what they eat, so lame!
So, with wanting to think about what I eat, and trying to
eat a balanced and ‘healthy’ diet, I find myself asking, what is healthy? To
me, health is a vague concept. What does it mean to be healthy? Can one be
healthy because they lack disease? Or does it also imply a certain trajectory
of disease free survival for the long term? I sometimes just feel healthier if
I eat at an organic co-op, does that make me healthier? And why do I feel so unhealthy
if I eat, even a relatively low fat, low caloric meal, at a fast-food
restaurant? And how on earth can Trader Joe’s, selling to an organic and health
conscious market, get away with calling this coconut oil? http://www.amazon.com/Trader-Joes-Organic-Virgin-Coconut/dp/B007UWNBYS.
This is a side tangent, but an ‘oil’ implies that it’s liquid at room
temperature, this is solid saturated fats, just as bad as any butter or animal
fat, but it feels pretty healthy, right?
In my mind, ‘being healthy’ implies a trajectory of
long-term survival in the absence of disease. The aspect of ‘feeling healthy’
is when you feel good doing it, and it may include some Trader Joe’s coconut
oil, because it’s delicious!
So how can we measure how healthy we are? Of course there
a many clinical tests in combination with clinical symptoms that a physician
can use to elucidate the presence of a disease. And of course there are
clinical measures that are good a predicting disease risk (such as blood
pressure, hemoglobin A1C levels, blood lipids, etc…). And of course there are
functional measures that can assess general physical fitness and overall wellbeing
(i.e. how well one can walk, exercise, etc…). But the simplest measure that can
actually be fairly good at predicting risk of disease, and can be measured by
anyone in his or her own home, is body weight.
When I last went to the doctor for an annual checkup, after
they measured my height, and took my weight, they quickly reported my BMI (body
mass index, kg/m2) to be 25.0. As I expected it to be roughly around
25, this wasn’t a surprise to me. But I knew that 25 was the cut-off between
what is considered ‘normal weight’ (18.5 – 24.9) and ‘over weight’ (25-30),
with obese being 30+. So by just a smidgen, I was considered ‘over weight’. But
no one at the clinic seemed to care! With such a daunting clinical condition as
‘over weight’, there has got to be some real evidence to say that I am entering
into a category that has an increased risk of mortality, right? Should I be
trying to drop that weight to join my ‘normal weight’ counterparts? Can I
calculate the number of years I’ll lose by now entering into a trajectory of a
health span as an ‘over weight’ person? Well, I don’t quite think that I’m ‘over
weight’ in a true sense, although I should definitely try and maintain an
active life full of exercise with a diet rich in fruits and vegetables for many
reasons, and perhaps my weight will come down a little. But seeing the BMI in
action made me want to look more into what evidence we actually have for using
the BMI, and try and answer, how bad is being ‘over weight’ or even obese?
Looking into a few recent studies that have conducted large
meta-analyses looking into the association between BMI and mortality,
it becomes clear that there is some real evidence that being obese (BMI, 30+)
increases risk of mortality. In addition, there is real evidence that being
underweight (BMI, <18.5) increases risk of mortality. These data appear to
show what is considered a J-shaped curve, where, starting at an underweight
BMI, as BMI increases mortality drops to an ‘ideal’ level, then mortality risk
continues to increase as BMI continues onwards towards the obese range. But
what is the ideal BMI? Is there a specific BMI we should all be shooting for?
And how normal is the ‘normal’ BMI of 18.5 to 24.9?
In an extremely comprehensive analysis, published by
Gonzalez et al. in the NEJM in 2009 http://www.nejm.org/doi/full/10.1056/NEJMoa1000367,
the complex relationship between BMI and risk of mortality becomes clear. For
this study, the median BMI in 1.46 million white adults of their study
population was 26.2. So right away, I’m starting to think that a 'normal' BMI of 18.5 to
24.9, isn’t so ‘normal’. The crux of their data is the non-linear relationship
between mortality risk and BMI, as shown below.
Looking at the data, it becomes clear that those in the
study population with a BMI of 17.5 seemed to be a part of a category that had
a significant increase in mortality risk, upwards of a 2-fold increase in risk
to the referent population being those with a BMI between 22.5 and 24.9. The
risk of mortality drops as BMI increases, to what appears to be an ‘ideal’ BMI
of something around 23-24 in both men and women. When BMI goes above this
‘ideal’, the risk in mortality doesn’t seem to dramatically increase right
away. In both men and women, when BMI reached 26, there appeared to be a 6-9%
increase in mortality risk in non-smokers. But then it begins to really climb,
reaching a 44% increase in mortality risk with those with BMIs of 31, and over
100% increase in risk for those with BMIs above 36 and even greater for those
with BMIs above 40.
In almost complete agreement with the Gonzalez study, the
Prospective Studies Collaboration, within the British Heart Foundation and
Oxford published in Lancet their
meta-analysis from 57 studies, looking at BMI and mortality in 900,000 adults http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(09)60318-4/fulltext.
Their main findings of BMI and mortality risk are shown in the figure below.
What they report is that mortality risk is lowest for those
with a BMI between 22.5 and 25. Where again, this J-shaped curve appears, where
risk of mortality is higher for those with BMIs below 22.5, and increases for
those above 25. But really, mortality risk for those with BMIs between 20 and
roughly 27.5 appear to be nearly the same, likely small increases in risk for
those with 20-22.5 and for those 25-27.5.
So looking at these data, it makes sense why a recent study
by Flegal et al. in JAMA in 2013 http://jama.jamanetwork.com/article.aspx?articleid=1555137,
found a significant decrease in mortality risk for those with BMIs between 25
and 30 compared to those with BMIs between 18.5 and 24.9. Their table showing
hazard ratios (HRs) and BMI is shown below.
This publication got a lot of press, and people seemed to
feel that there must be something that’s missing, but when looking at the data
before it, these results aren’t too surprising. We know that mortality risk is
higher for those with BMIs lower than 22.5, and we know that it begins to climb
again for those with BMIs above 25, but not at a very fast rate. So if you
combine the people with BMIs between 18.5 and 24.9, you’re going to catch the
people with the higher risk of mortality with BMIs below 22.5 and the
group with BMIs between 25 and 30, includes those with higher risk between 27.5 and 30, but
combined, the overall risk is lower for the 'over weight' group compared to the 'normal weight' group.
So what can we learn from these studies? I think we can
safely say that an ‘ideal’ BMI is between 22.5 and 25, and if one is a little
lower than that, or a little higher than that, there isn’t really any strong
evidence to say that risk of mortality increases. But there does consistently
show strong evidence for increased mortality for those with low BMIs and with
high BMIs. Overall, I think the ‘normal’ BMI range of 18.5 to 24.9, is a little
off, and should probably be more like 22-27, perhaps this will change in the
future.
But of course, this analysis is severely limited by the main
problem with these studies: BMI is a measure that is adequate at best to
predict mortality risk. We all know the examples of the incredibly athletic and
muscular person who is not obese, but his/her BMI puts them into a category
they don’t want to be in.
The reality is that in these studies, those individuals with
high BMIs have a higher risk of mortality due to cardiovascular disease,
diabetes, renal disease, cancer, etc., diseases often times due to improper
metabolic health. Metabolic health refers to really how our body deals with
energy and our diet, such as our blood sugar and insulin sensitivity and our
blood lipids and cholesterols. We know that these factors play important causal
roles in disease progression and it’s important to get an idea as to how we’re
doing on these fronts. What the BMI hopes to assess is metabolic health, and what
we really want to know is what is someone’s chance of developing disease due to
these improper metabolic controls? And unfortunately, BMI is only a crude and
limited measure.
This is why a call for better metrics for metabolic health was
made recently in the journal Science http://www.sciencemag.org/content/341/6148/856.full,
where the authors discuss how we now know that there are people with normal as
well as obese BMIs who are metabolically healthy, just as there are people who
have normal as well as obese BMIs who are metabolically unhealthy.
In the figure above, risk of mortality is shown by the
elongated triangle at the top, where you can see that risk of mortality is
greatest for those who are obese and metabolically unhealthy. But
interestingly, someone who is of ‘normal’ BMI but metabolically unhealthy has a
higher risk of mortality than someone who is obese but metabolically healthy.
What this highlights is the lack of mortality risk
prediction one can get from BMI alone, and of course, any good physician will
work with the patient to understand other metabolic factors that are greater
predictors of risk, but what can we do at home? When I step on the scale, and
reveal my BMI, am I appropriately gauging where I am ‘health’ wise? I don’t
think my body weight will tell me everything, but I think if I work to keep my
diet and exercise in check, in combination with keeping an eye on such things
as cholesterols, blood lipids, blood sugars, and blood pressure, my body weight
may help me keep a frame of reference as to how I’m doing, and for now, my BMI
of 25 doesn’t sound too bad.
Chad Weldy
