Junkfood Science: Obesity Paradox #15 — No need to stroke out

April 14, 2008

Obesity Paradox #15 — No need to stroke out

I kept waiting for mainstream media to report this study... and waited... and waited. Here’s news that you deserve to hear.

Researchers, led by Dr. Tom Skyhoj Olsen, M.D., Ph.D., at the Stroke Unit, Department of Neurorehabilitation at Hvidovre University Hospital in Hvidovre (Denmark), looked at body mass index and the risk of dying after having a stroke.

Their study was published in the February issue of Neuroepidemiology. As their introduction noted, 40 studies including 240,152 patients with coronary artery disease, have revealed better outcomes for cardiovascular and all-cause mortality among fat patients, while even the greatest obesity was not associated with increased risk of death. Does this include strokes, they wondered? Using a hospital registry which has recorded stroke patients hospitalized in Denmark since 2001, they followed first-time stroke patients for 5 years looking for links between BMI and survival.

The Danish National Indicator Project has been working to include on its registry detailed clinical information on all stroke admissions in Denmark. It includes assessments of the severity of strokes, using the Scandinavian Stroke Scale, and a detailed cardiovascular profile for each patient that includes lifestyle, medical conditions and cardiovascular measures. BMIs were also measured by nurses on admission. For this study, the authors included a representative sample of 13,242 patients with complete data, of the total 39,484 patients in the registry. (The ages and other cardiovascular profiles among those included in this study did not differ notably from those not included.) Stroke diagnoses had been confirmed on CT/MRI scans, with ischemic strokes distinguished from intracerebral hematomas and actual strokes differentiated from transient ischemic attacks. Only a total of 1.1% of the patients were excluded for missing CT/MRI scans (0.4% not done; 0.7% unavailable). Nearly 90% of stroke patients in the registry were treated in hospitals and rehabilitated in special stroke units in Denmark, enabling the consistency of care to facilitate clearer examinations of other confounding factors.

The strengths of this study included its large sample size, its inclusion of stroke patients without selectivity for any particular demographic measure or risk factor, that all the patients had been actually clinically evaluated using the same well-validated stroke criteria and had received consistent care, and that the authors had survival data on virtually all of the patients, with less than 0.2% lost in follow-up.

Before we look at survivals after strokes, the first question that you may wonder about is the demographics of those who had strokes. Were fat people having more strokes to begin with?

No. Fat people, even with the highest BMIs, were not disproportionately having more strokes. In real numbers, for example, among the women stroke patients, 437 had BMIs ≥35 (4.2% of stroke patients). This compared to 5,226 with ‘normal’ weights (50%) and 838 underweight (8%).

The average BMI of people having strokes was 23.

The average age of first-time stroke patients was 72.3 years, just over half (52%) were men, and only 7.8% had very severe strokes with most (60.5%) having had mild strokes.

Those with high BMIs were slightly more likely to be men, married, nonsmokers, non excessive drinkers, diabetics, have hypertension, and to have no prior history of strokes, heart attacks or intermittent arterial claudication.

Despite higher percentages of ‘obese’ and ‘severely obese’ with diabetes and hypertension, not only were the numbers having strokes, but the severity of strokes were linearly related to BMI, inversely — the greatest numbers and the severest strokes were in those with the lowest BMIs.

During the five years of followup, about a third of the patients died (4,841). The highest actual mortality rates (number of deaths per 100 patients), regardless of the cause, were among underweight patients; the lowest mortality rates were among the ‘overweight’.

As the authors emphasized: “Obese and severely obese patients had a lower mortality rate than that of normal weight patients.”


What are the riskiest BMIs?

Since the risk of dying from any cause obviously increases with age, their stratification of risks by age group might be of interest, especially since we so often hear that being fat at younger ages is more dangerous. Among the youngest patients, aged 40 to 65 years, the most ‘severely obese’ had a 40% lower risk of dying than those of ‘normal’ weight. A similarly lower risk was found among those who were ‘overweight.’ Obesity was actually most associated with lower risks among the younger ages.

In contrast, being underweight at these young ages (weights popularly considered ideal by many women) was associated with more than 2.5 times the risk of dying compared to ‘normal’ weight people. These risks at different BMIs were similar among the 65 to 75 year olds. According to the authors, the survival benefits of higher BMIs did not change as the follow-up years went on, either, but were consistent over time.

After the age of 65, BMIs tend to drop with advanced aging, of course. So the authors also adjusted for confounding factors such as age, gender, marital status, risk factors and stroke severity to attempt to isolate the effects of BMI on mortality. Still, the relative risks of dying after having a stroke were 16% lower among ‘obese’ patients compared with ‘normal’ weight patients. And those with low BMIs <18.5 (which would, for example, equate to a 5ft-4in woman weighing 108 pounds) had a 79% higher risk of dying than the ‘obese’ patients, and 63% higher risk than ‘normal’ weight.

“High BMI predicted survival,” they noted. This held regardless of whatever they threw into their analytical mix. The favorable effects of higher BMIs on survival were independent of age, gender, marital status, stoke severity, type of stroke or any health risk factors, said the authors. “In particular, the effect of BMI was the same in smokers and nonsmokers, ischemic and hemorrhagic strokes and within age groups,” they wrote. Even among people who had only mild strokes, those who were underweight were twice as likely to die compared to patients of ‘normal’ weight or the most “obese.” As the authors concluded:

The main finding of our study is that BMI in patients with stroke is inversely related to total poststroke mortality: overweight and obese stroke patients have a lower mortality rate than normal-weight and underweight patients.

While they found that higher BMIs were associated with established “risk factors” for stroke, such as diabetes and hypertension, the risk factors didn’t actually change the facts: that BMIs were inversely related to mortality. Higher BMIs were associated with lower mortalities, whether they examined the data by adjusting for all of the cardiovascular “risk factors” or not. This is an important reminder, that risk factors” are not the same as actual risks, despite the best efforts of some to convince us otherwise. [What are risk factors?]

When no correlation between a condition and outcome can even be found, then continuing to pursue it as having a possible causative relationship is not even credible, and scientists move on. Even looking at population-wide statistics in the U.S., not only have deaths associated with heart disease and strokes been dropping for more than half a century, so, too, have the incidences of strokes - while 'obesity' rates are supposedly rising. "The incidence of stroke has been declining since the early 1950s, before any antihypertensive drugs were invented, and well before any of them were used widely," said Dr. Michael Alderman, M.D., chairman of the department of epidemiology and social medicine, Albert Einstein College of Medicine, Bronx, New York, in a Healthfacts interview about the National High Blood Pressure Education Program.

Dr. Olsen and colleagues also considered the possibility of a “healthy survivor effect” perhaps skewing their finding, and ruled it out. First, by including patients from age 40 years, people who didn’t survive to that age related to being fat are likely too few in number to have any influence on their analysis, they said. Second, the risks of dying within the first days of a stroke were lower among the ‘obese’ and ‘extremely obese’ compared to ‘normal’ weight and this BMI survival benefit didn’t lessen over the years of follow-up. “Hence, an overrepresentation of obese or very obese patients who die from stroke before arriving to hospital is unlikely.”


Putting this study into context with the body of evidence

The authors said it best:

Although our finding may seem controversial, it is not surprising. A recent meta-analysis of 40 studies including 250,152 patients with coronary artery disease also concluded that total mortality was lower in overweight and obese patients, and an overview of 32 studies on mortality risk in elderly people aged ≥65 years showed that BMI in the overweight range was not associated with increased mortality and moderate obesity was only associated with a moderately increased mortality risk. As mean age of our cohort was 72 years, it is interesting to note that studies on elderly patients >70 years of age show an inverse association between BMI and mortality as also seen in our study.

We found that underweight stroke patients clearly had the highest mortality. Lower BMI is highly associated with smoking, but the finding was independent of smoking habits and prevailed also after stratification on smoking- nonsmoking. This finding is consistent with those from other studies. Underweight may be the result of another underlying disease thus increasing mortality risk. However, in studies with the same result as ours, underweight persons had the highest mortality risk even after exclusion of subjects who had died within 2–5 years after inclusion. BMI is a reflection of both body fat and lean mass. Low BMI may be the result of reduced lean mass as a consequence of inactivity because of reduced mental and physical health.

On the other hand, persons with a high BMI may in addition to an increased amount of body fat also have a higher lean mass as a consequence of an active lifestyle and otherwise good mental and physical health.

© 2008 Sandy Szwarc

Bookmark and Share