Investigators drew on data from over 1000 patients with an average age of 71 years who had experienced an acute ischemic stroke (AIS). All were enrolled in the multicenter National Institutes of Health (NIH) Field Administration of Stroke Therapy – Magnesium (FAST-MAG) trial, a randomized study of prehospital magnesium therapy in stroke patients, reported in 2015 in the New England Journal of Medicine.
The trial failed to show an effect of intravenous magnesium sulfate started in the ambulance on 90-day disability outcomes in patients with suspected AIS, although it suggested such prehospital treatment was at least feasible.
For this analysis, researchers used data from FAST-MAG to examine the relationship between BMI and 90-day outcomes.
Results showed that the 3-month risk for mortality declined linearly with higher BMI in all groups, but especially in those who were overweight or obese.
On the other hand, the risk for disability had a U-shaped relation to BMI, in which those who were underweight or severely obese had a higher risk for disability.
“We found an inverse linear relationship between BMI and mortality, wherein mortality decreased as BMI increased, although there was a U-shaped curve when it came to disability — with higher BMI associated with more disability,” lead author Zuolu Liu, MD, stroke fellow, University of California-Los Angeles Medical Center, told Medscape Medical News.
The findings will be presented at the American Academy of Neurology (AAN) 2019 Annual Meeting, to be held May 4-10 in Philadelphia.
A survival advantage for those with higher BMI has been observed for diverse acute illnesses, and called the “obesity paradox,” the authors write.
However, prior stroke studies of this paradox have been underpowered, and have “yielded mixed results,” they note.
The investigators set out to “delineate the relationship between BMI and 90-day outcomes in a large cohort of patients with AIS,” with 3 outcomes: death, disability or death, and low stroke-related quality of life.
Univariate and multivariate models were used to adjust for 12 additional prognostic variables.
“We used data from the registry of the NIH’s FAST-MAG clinical trial because it was a large cohort and we had access to the patient registry,” Liu said.
The average age of the 1033 patients was 71 years (±13), of whom roughly half (45.1%) were women. The median BMI was 27.5 (±5.6).
Increase Nutritional Reserve?
When BMI was used as a continuous variable, the risk for death declined linearly with higher BMI (unadjusted P = .02, adjusted P = .004).
Adjusted odds ratios (OR, 95% Confidence Interval) for mortality across the BMI categories are shown in the following:
Underweight 1.67 (0.57 – 4.88)
Normal weight: 1.00
Overweight: 0.85 (0.53 – 1.36)
Obese: 0.54 (0.29 – 1.04)
Severely obese: 0.38 (0.16 – 0.88)
By contrast, the risk for disability had a U-shaped relation to BMI (quadratic P = .02).
ORs for disability or death declined through the categories of underweight, normal, overweight, and obese, (1.19, 1.00, 0.78, and 0.72, respectively) but rose in the severely obese (0.96). Adjustment for other prognostic factors attenuated the relationship (P = .27).
Trends for low stroke-related quality of life were similar but nonsignificant.
She suggested that a potential mechanism for this finding is that that a person with normal weight, or who is overweight or obese, might have more nutritional reserve, compared with those who are underweight.
On the other hand, “if you are underweight or severely obese, your ability to engage in post-stroke therapies might be limited, which in turn would limit your ability to recover and contribute to greater disability,” Liu said.
“No Direct Implications”
Commenting on the study for Medscape Medical News, Edward Boyko, MD, professor of medicine and adjunct professor of epidemiology, University of Washington, Seattle, said that obesity does not have some sort of protective effect.
“Rather, it is likely that lower BMI might be due to chronic conditions or lifestyle habits such as smoking, which would result in poorer outcomes after a stroke; and obesity may be a marker for fewer comorbid suggestions,” said Boyko, who was not involved with the study.
Liu agreed that it is currently “too early to give practical take-home findings.”
Further research is needed to elucidate the nature and mechanism of this “paradoxical finding,” she said.
American Academy of Neurology (AAN) 2019 Annual Meeting: Poster 3.3-031. To be presented May 7, 2019.