Oh, Overweight People Live Longer; Another New Study Another Old Myth.

The myth that overweight people actually live longer than their normal weighted counterparts gets trotted out regularly. The most current one, referenced below, has been the source of a lot of triumphal crowing that our obsession with being overweight is somehow, just wrong.

Association of All-Cause Mortality With Overweight and Obesity Using Standard Body Mass Index Categories A Systematic Review and Meta-analysis. Katherine M. Flegal, PhD; Brian K. Kit, MD; Heather Orpana, PhD; Barry I. Graubard, PhD

In this world, we choose information to suit ourselves, don’t we? Well the PPP is no different and so with regards to this study, I choose alternative information to suggest that all is not that which it seems in this matter. As I said, this is one of those debates that rears its ugly head from time to time. Last release of a similar study was in 2013; the study was more or less identical to the current study and here is a simple, elegant explanation as to why it is wrong! Turns out to be all to do with methodology, of course.

Perhaps if I repost these two diagrams, that will be of assistance. The reference for them is as above : Monday’s medical myth: overweight people live longer

A graphical representation of the current study, showing mid-points of the broad BMI categories and the resultant death rates against the reference category. In the current study, the ‘overweight’ BMI category recorded death rates of 94 against every 100 deaths in the normal weight category. This was reported by the authors to be statistically significant.
if you do nothing different except to split each BMI category into half, plotting two data points for each category, things change, and there is no reduced mortality rate for the overweight group.

In a nutshell, the current study and previous studies, use a broad range of the Body Mass Index categories, there are really only 6, these being underweight, normal weight, overweight and 3 classes of ‘obesity’. In the ‘standard’ categories of BMI there is about a 5 point range, eg ‘overweight’ is defined as being in the range 25–30 [25 to 29.9 in precise terms]. In this range, the overweight people show a statistically lower mortality ration than their normal weight colleagues. However, when you split this range into BMI increments of 2.5, the mortality rates differ, to the extent that even overweight people, BMI of 25–30 register a higher mortality rate than normal weight people. So if you do nothing different with the data save for splitting the BMI categories into half [eg the ‘overweight category, 25- 29.9 is split into two categories, 25–27.5 and 27.6–29.9] and graph them as separate data points, they fall above the mortality rates for normal weight people. Please read the study linked above to understand. And remember, nothing else changes, just the range of the 6 categories.

Below I give a plain English explanation of the current study, together with some more of its pitfalls. I have been able to do this because the complete study is available, FREE! on the internet, so a big up to those that made this happen.

Terms first. The authors used the generally accepted categories described above of overweightness according to the BMI. They examined some 97 studies in total. The summary table below shows the BMI range, the label [category name] associated with the BMI range and what I have called the ‘mortality ratio’ which just references the other categories against the normal weight category set to 100 for convenience. This is a tabular representation of the graph above, using the actual data from the study:

Their categories:

BMI — — — — — — Category — — — — — -Mortality Ratio

18- 25 — — — — — Normal weight — — — -100 [baseline]

25–30 — — — — — Overweight — — — — — 94

30–35 — — — — — Grade 1 obesity — — — -95

35 -39 — — — — — Grade 2 obesity } — — 129

40 + — — — — —Grade 3 obesity }

So in plain English their findings basically say you can be overweight, by some kilograms, and your overall risk of dying is less than someone of normal weight. We have seen above how this is really a statistical aberration. For me the big-ticket item is that if you are really really overweight, your risk of mortality is ‘significantly’ greater, near on 30% greater.

That’s the study in plain English. Here is my two-point commentary.

Point 1: So what? Lost in the headlines is the main point; if you are really obese, then the chances of dying before your time is very much higher. The target population in terms of risk factors is always going to be the more obese people. So I would rather the headline around this study to read:

New Meta Analysis Confirms the Grave Risk of Early Death for Morbidly Obese People.

Point 2: This is not the get out of jail card that the headlines and much of the associated commentary around them suggests for those people who are overweight. Here’s why.

1] As the graphs above show, the BMI measure is pretty much useless, too broad brushed in terms of category size. As soon as you narrow the categories down, the data changes such that the finding that overweight people living longer is shown for what it as, a statistical artefact.

2] The BMI is not really a good indicator of problematic weight in the lower ranges anyway, since it does not take into account built, muscle density etc etc. You can have an elevated BMI and be quite thin.

3] The only variable examined in the current study is mortality rate; this also is too broad-brushed. Put frankly, you can be extremely unhealthy to the point where your life lacks quality and still not be dead. The relative health of the different cohorts in this study was not examined. Quality of life is a much more meanigful measure of the impact of weight than is the death rate. To examine real impact of weight we need much more specific measures. These will include:

· More fine grained physical measures including percentage of body fat and its distribution, in particular, visceral fat which seems to be an excellent predictor of doom.

· Clever measures of health such as prescription rate for medications, actual medication use, hospitalization rates, attendance rates at medical practitioners, pathology testing rates and such like.

· Psycho-social measures such as self-reports of happiness, participation in activities and so on examine the impact of weight on mental health.

Dr Google will reveal many studies which draw a direct relationship between increased weight and the above factors. Yes, I get that I am no different in pointing out that there is always research to back up any point of view. I am just interested in presenting this alternative view.

 1] The current study contributes little to what we already know. You cannot draw from this study any conclusion that being overweight is a protective factor for longer life.

2] Increased weight, even quite small increases, are correlated with adverse health outcomes, such as increased accessing of health services when compared to people who are not overweight.

3] Eat less, eat more plants. The simplest way to stack the odds of health a little more in your favour.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.