HEART DISEASE. THE OBESITY PARADOX. IS THERE A J-CURVE

I just another interesting paper. Is there a J-curve in patients with obesity and chronic illnesses like CAD? Is there such a thing as the obesity paradox?
Dr Oskar Angeras and colleagues from the University of Gotenberg analyse the data from the SCAAR ( Swediah Coronary Angiogram and Angioplasty Registry ), and published their findings in the 5th Sept issue of the European Heart Journal.
In their study, they included 65,000 patients who had an acute coronary syndrome, and all received coronary angiogram. Following the coronary angiogram, they either received medical therapy, PCI or CABG. He divided these 65,000 patients into 9 BMI categories. The reference category ( normal BMI was BMI 21-23.5Kgm/M2. These patients were followed for 21months.
They found that for those who received PCI, those in the BMI 18 and below fared the worse. Those with BMI 26.5 - 28 fared the best. There was still mortality reduction in those with BMI between 28 - 35. Once the BMI reaches 35, those patients fared badly. The BMI did not seem to affect those who underwent CABG. -
The message seemed to be that too thin is bad, too fat ( BMI greater than 35 ) is bad. Overweight with BMI greater than 26 and less than 35 is still OK. Once BMI goes beyond 35, the mortality rises.
We cannot understand this. It seem to go against conventional medical wisdom. In fact there were other studies which suggest that if you are overweight with a chronic illness, reducing your weight may increase your moratilty from that illness. Strange.
Is there an "Obesity paradox"? Does a bit of obesity protect us? Is there a J-curve, be obese but not too much? Should we ask patients who are overweight, not to lose weight? Now we have more questions than answers.
Obviously, we need more data, on such an important subject.
In the meantime, eat, drink and be merry. All in moderation of course.