Li W, Katzmarzyk PT, Horswell R, Zhang Y, Zhao W, Wang Y, et al. Body Mass Index and Stroke Risk Among Patients With Type 2 Diabetes Mellitus. Stroke. 2014
It sounds like a paradox, and it probably is. Much like my recent trip to the movie theater to watch Interstellar (without spoiling the movie), I’m having a hard time making sense of how things can happen in the future without ever having happened in the past. I’m being told that 1+1 = infinity. This is the stuff of black holes, singularities and Tesseracts!
Much work has been done regarding the evaluation of obesity and all cause and CVD mortality in diabetic patients. Results have wildly varied from positive, inverse, U-shape to no associations. Most notably however is the recurring result where obesity pre-disposes to a certain chronic disease state, yet the more obese you are, the more you are protected against that disease. This phenomenon is called the Obesity Paradox and has been seen time and time again in different studies. Li et al jump into this spaceship of a concept and add a twist; they evaluate obesity in stroke among diabetic patients.
From 1997-2013 patients from LSU health care services, with diabetes diagnosed by ICD-9 code within 5 years, were included in this study. Patients with CAD, smoking history or incomplete data sets were excluded to avoid potential confounders. Over 29,000 patients were included. ICD-9 codes were also used in determination of stroke as an endpoint for these patients. BMI was evaluated as both a continuous and categorical variables. During a follow up period of 8.3 years, 2,883 subjects developed stroke (2,821 ischemic and 109 hemorrhagic). Hazard ratios associated with different levels of BMI at baseline 18.5-24.9 [reference group], 25-29.9, 30-34.5, 35-39.9, >40) were 1.00, 0.86, 0.83, 0.76 and 0.70 for total stroke. The same patterns were apparent individually for both ischemic stroke and hemorrhagic stroke. Several factors including HTN, A1c, GFR, age, sex ,race were controlled for, yet there still remained an inverse relationship.
Many have argued that the obesity paradox is not true and offer up some of the following reasons for refuting such a claim: The sample sizes in the reports of the obesity paradox are generally small; however Li et al have a fair strong sample size to evaluate. There are others who argue that chronic disease states cause wasting and so naturally those who are more severely effected by a disease will have decreased weight/BMI. In this argument, BMI is determined by the illness, and not the other way around. Stroke however unlike other chronic processes comes suddenly; and thus one could argue that this claim is not valid. Ultimately, however, the most powerful argument against the obesity paradox a broad sweeping umbrella of an argument that is not interested in explaining any specifics. It is a mathematical argument that reminds us that statistical techniques show association between factors; not cause effect. Yet, if this is the case, how do we explain seeing this paradox over and over again? One is left to believe that there must be a specific reason.
For the record, I absolutely loved Interstellar. The complexity of the concepts and its inability to be flushed out on screen caused some to criticize it for being sloppy. Sometimes though we don’t want all of life’s answers laid out on a silver platter; it’s more exciting to think through concepts and revel in the potential explanations. The same can be said of the obesity paradox. We may OR we may not ever know the reason; maybe it’s not a paradox after all. But we should let this simmer a little longer and try to entertain why and how it exists. And in the process we may discover broader implications of science and health. But until we figure it out, I wouldn’t endorse eating donuts to protect against stroke just yet.