Impact of baseline diabetes and obesity statuses on outcomes post-ischemic stroke

Proposal
1331

Title of Proposed Research

Lead Researcher

Colleen Bauza

Affiliation

Medical University of South Carolina in Public Health Sciences

Funding Source

None

Potential Conflicts of Interest

None

Data Sharing Agreement Date

28 January 2016

Lay Summary

Obesity and diabetes have continued to increase in prevalence since 1980 in the United States. Based on the 2011-2012 National Health and Nutrition Examination Survey (NHANES), 68.5% of US adults were considered overweight or obese, and 34.9% were obese. According to the American Diabetes Association, 29.1 million (9.3%) Americans had diabetes in 2012. While diabetes is considered a group of diseases marked by high levels of blood glucose resulting from issues in how insulin is produced and/or works, Type II diabetes (non-insulin dependent diabetes) accounts for 95% of all diagnosed cases of diabetes.
Obesity and diabetes afflict a significant number of Americans and increase their risk of multiple negative health-related outcomes. Yet, both obesity and type II diabetes are considered to be preventable and modifiable risk factors. While it is well known that obesity and diabetes are associated with an increased risk of ischemic stroke, there is a lack of consensus regarding the impact of obesity and diabetes on stroke related outcomes.
The Prevention Regimen for Effectively avoiding Second Stroke (PRoFESS) trial provides the ideal large cohort in which to investigate the relationship of obesity and diabetes with post primary stroke outcomes.

Objective: To describe the effect that obesity and diabetes have on post stroke outcomes.
Primary aims:
Aim 1: To describe the relationship between obesity, diabetes, and the combination on post-stroke mortality.
Aim 2: To describe the relationship between obesity, diabetes, and the combination on recurrent vascular events post-stroke.
Aim 3: To describe the relationship between obesity, diabetes, and the combination on re-hospitalization post-stroke.

The proposed study follows a prospective cohort design where the cohort of ischemic stroke subjects enrolled in the PRoFESS trial was prospectively followed for approximately four years to obtain outcomes. Time to mortality, time to first vascular event post-stroke, and time to first re-hospitalization post-stroke will be explored via Cox-proportional hazards modeling for all aims.

There is a paucity of information concerning the joint impact of obesity and diabetes on stroke-related outcomes; this research will be able to build upon existing knowledge and extend our understanding of factors influencing the post stroke trajectory. Results from this study will provide evidence that post-stroke outcomes are associated with obesity and diabetes. Moreover, this research will suggest that reducing the burden of these risk factors may diminish poor outcomes following stroke.

Study Data Provided

[{ "PostingID": 1970, "Title": "BI-9.159", "Description": "PRoFESS - Prevention Regimen For Effectively Avoiding Second Strokes" }]

Statistical Analysis Plan

As previously stated, the overall objective is to illustrate the extent to which misclassification and measurement error are present in diagnosing obesity measured by BMI, WC, and WHR using data from the PRoFESS cohort. The distribution of each anthropometric measure (i.e., BMI, WC, WHR) will be graphically assessed using a QQ-plot. Categorical baseline characteristics will be presented as a number and proportion, and continuous baseline characteristics will be shown as a mean and standard deviations or as a median and interquartile range. The coefficient of variation (CV), a measure of relative reliability,21 will be calculated for the following anthropometric measures from the PRoFESS cohort: body weight, height, BMI, WC, WHR. Additionally, the CV of the same anthropometric measures from the NHANES III survey cohort will also be calculated. The anthropometric measures from the NHANES III cohort will be considered reference measures. The CV for each anthropometric variable from the PRoFESS cohort will be compared to the CV for each anthropometric variable from the NHANES III cohort. Specifically, the CV of an anthropometric measure from a specific cohort with the lowest value will be considered a more reliable measure than the anthropometric measure from the other cohort.21The classic indices of diagnostic performance (i.e., sensitivity, specificity, positive predictive value, negative predictive value, and the missclassification rate) will be calculated for the anthropometric measures of WC and WHR in comparison to BMI. Due to the number of expert panels (e.g., US Preventive Services Task Force, American Heart Association/American Stroke Association, World Health Organization (WHO)) advocating the use of BMI to screen for obesity22,23,1,24, BMI will be considered the ‘gold standard' for this illustration. The following categorizations will be used to determine the diagnostic performance of WC and WHR in comparison to BMI: obesity (BMI = 30 kg/m2)25, abdominal obesity (WC > 102 cm (men) and > 88 cm (women))25, and WHR = 0.53 (men) and = 0.54 (women).26Assessing the classic indices of diagnostic performance of a method in comparison to an imperfect reference standard, such as BMI, may lead to bias.17 As such, it is necessary to also assess the ability of the other anthropometric measures as well as the imperfect reference standard to predict a clinically meaningful outcome, such as all-cause mortality, in order to determine which anthropometric measure best predicts the outcome of interest. Logistic regression models will be utilized to estimate the odds ratio (OR) and 95% confidence interval (95% CI) for a one standard deviation (SD) increase in each anthropometric measure in relation to all-cause mortality. As BMI, WC, and WHR have different units, z-scores will be calculated for each anthropometric measure. Specifically, z-scores will be calculated for all subjects in addition for each gender. Gender-stratified logistic regression models will be fit to analyze the association between each anthropometric measure and all-cause mortality adjusted for potential confounders. Potential confounders will include age, race/ethnicity, qualifying stroke neurological severity, ischemic stroke sub-type, baseline systolic blood pressure, hypertension, treatment assignment, history of congestive heart failure, history of atrial fibrillation, history of coronary artery disease, history of previous stroke or TIA, history of myocardial infarction, smoking status, alcohol consumption, and average physical activity prior to qualifying stroke.To determine the independent predictive value of each anthropometric measure for all-cause mortality, receiving operating characteristic curve (ROC) analysis will be used to determine the area under the curve (AUC). The AUC ranges from 0 to 1.0, with 0 indicating perfectly inaccurate discrimination and discrimination and 1.0 indicating perfect discrimination.27 In general, an AUC of 0.5 suggests no discrimination, 0.7-0.8 is considered acceptable discrimination, 0.8-0.9 was considered excellent discrimination, and = 0.9 was considered outstanding discrimination.28 To ensure that each model is not poorly calibrated, calibration of each model will be analyzed by the Hosmer-Lemeshow (HL) test. A model will be considered poorly calibrated if the p-value of the corresponding HL ?2 is less than 0.05. The anthropometric measure that has the highest AUC that is also well calibrated will be considered the best anthropometric measure that best discriminates between those that die and those that do not die from all-causes. Further, the Bland Altman29 approach will be employed to demonstrate the extent to which WC and WHR agree with BMI using the Bland-Altman limits of agreement analysis. This method estimates the bias between the mean difference (or bias) of two quantitative measures by constructing limits of agreement using the mean and standard deviation of the differences between the two measures.29,30 Agreement will be assessed using Bland Altman plots and calculating limits of agreement. Similar to the previous illustration, the anthropometric measures will be standardized for all subjects in addition for each gender. For the purposes of this illustration, wide limits of agreement were considered 2 z-scores or more, between 1.5 and 1.99 z-scores as wide, and less than 1.5 z-scores as reasonable agreement. Overall, these illustrations will highlight the extent to which misclassification and measurement error are present in diagnosing obesity using data from the PRoFESS cohort. Further, this study will highlight the need for future studies to note the accuracy of its measurements, specifically of their exposures. Overall, results of this study will highlight the need for future studies to collect additional measures of obesity as well as to report on the precision of these measures in order to accurately estimate the effect of obesity on outcomes following an ischemic stroke.A.7. Publication PlanThis amendment is intended to be part of my dissertation. I intend on publishing the results in a peer reviewed journal by the end of 2018. Potential peer-reviewed journals where I plan submitting the manuscript include, but not limited to: Annals of Epidemiology, Preventative Medicine Reports, American Journal of Preventative Medicine, American Journal of Public Health, BMC Public Health, and Scientific Reports.References1. Mozaffarian D, Benjamin E, Go A, et al. Heart disease and stroke statistics-2015 update: a report from the American Heart Association. Circulation. 2015;131:e29-322.2. Allen CL, Bayraktutan U. Risk factors for ischaemic stroke. Int J Stroke. 2008;3(2):105-116. doi:10.1111/j.1747-4949.2008.00187.x.3. Sacco R, Benjamin E, Broderick J, et al. Risk factors. Stroke. 1997;28:1507-1517.4. Calle E, Thun M, Petrelli J, Rodriguez C, Al. E. Body mass index and mortality in a prospective cohort of US adults. N Engl J Med. 1999;341:1097-1105.5. Adams K, Schatzkin A, Harris T, Kipnis V, Al. E. Overweight, obesity , and mortality in a large prospective cohort of persons 50 to 71 years old. N Engl J Med. 2006;355:763-778.6. Olsen TS, Dehlendorff C, Petersen HG, Andersen KK. Body mass index and poststroke mortality. Neuroepidemiology. 2008;30(2):93-100. doi:10.1159/000118945.7. Doehner W, Schenkel J, Anker SD, Springer J, Audebert H. Overweight and obesity are associated with improved survival, functional outcome, and stroke recurrence after acute stroke or transient ischaemic attack: Observations from the tempis trial. Eur Heart J. 2013;34(4):268-277. doi:10.1093/eurheartj/ehs340.8. Andersen KK, Olsen TS. The obesity paradox in stroke: Lower mortality and lower risk of readmission for recurrent stroke in obese stroke patients. Int J Stroke. 2015;10(1):99-104. doi:10.1111/ijs.12016.9. Vemmos K, Ntaios G, Spengos K, et al. Association between obesity and mortality after acute first-ever stroke: The obesity-stroke paradox. Stroke. 2011;42(1):30-36. doi:10.1161/STROKEAHA.110.593434.10. Towfighi A, Ovbiagele B. The impact of body mass index on mortality after stroke. Stroke. 2009;40(8):2704-2708. doi:10.1161/STROKEAHA.109.550228.11. Bell CL, LaCroix A, Masaki K, et al. Prestroke factors associated with poststroke mortality and recovery in older women in the women's health initiative. J Am Geriatr Soc. 2013;61(8):1324-1330. doi:10.1111/jgs.12361.12. Skolarus LE, Sanchez BN, Levine DA, et al. Association of body mass index and mortality after acute ischemic stroke. Circ Cardiovasc Qual Outcomes. 2014;7(1):64-69. doi:10.1161/CIRCOUTCOMES.113.000129.13. Oesch L, Tatlisumak T, Arnold M, Sarikaya H. Obesity paradox in stroke-myth or reality? A systematic review. PLoS One. 2017;12(3):e0171334.14. Stevens J, Bradshaw PT, Truesdale KP, Jensen MD. Obesity paradox should not interfere with public health efforts. Int J Obes (Lond). 2014;39(April):80-81. doi:10.1038/ijo.2014.60.15. Dixon J, Egger G, Finkelstein E, Kral J, Lambert G. “Obesity paradox” misunderstands the biology of optional weight throughout the life cycle. Int J Obes. 2015;39:82-84.16. Abdelaal M, le Roux CW, Docherty NG. Morbidity and mortality associated with obesity. Ann Transl Med. 2017;5(7):161. doi:10.21037/atm.2017.03.107.17. Trikalinos TA, Balion CM. Chapter 9: Options for Summarizing Medical Test Performance in the Absence of a “Gold Standard.” J Gen Intern Med. 2012;27(Suppl 1):67-75. doi:10.1007/s11606-012-2031-7.18. Reitsma JB, Rutjes AWS, Khan KS, Coomarasamy A, Bossuyt PM. A review of solutions for diagnostic accuracy studies with an imperfect or missing reference standard. J Clin Epidemiol. 2009;62(8):797-806. doi:10.1016/j.jclinepi.2009.02.005.19. Schiller I, Smeden M Van, Hadgu A, Libman M, Reitsma B. Bias due to composite reference standards in diagnostic accuracy studies. 2016;(November 2015). doi:10.1002/sim.6803.20. Carroll RJ. Measurement Error in Epidemiologic Studies. In: Wiley StatsRef: Statistics Reference Online. John Wiley & Sons, Ltd; 2014. doi:10.1002/9781118445112.stat05178.21. Strike P. Statistical Methods in Laboratory Medicine. Oxford: Butterworth-Heinemann; 1991.22. Meschia JF, Bushnell C, Boden-Albala B, et al. Guidelines for the primary prevention of stroke: A statement for healthcare professionals from the American heart association/American stroke association. Stroke. 2014;45(12):3754-3832. doi:10.1161/STR.0000000000000046.23. Jensen MD, Ryan DH, Apovian CM, et al. 2013 AHA/ACC/TOS guideline for the management of overweight and obesity in adults: A report of the American College of cardiology/American Heart Association task force on practice guidelines and the obesity society. Circulation. 2014;129(25 SUPPL. 1). doi:10.1161/01.cir.0000437739.71477.ee.24. National Institutes of Health. The Practical Guide: Identification, Evaluation, and Treatment of Overweight and Obesity in Adults.; 2000. http://www.who.int/nutrition/topics/FFA_summary_rec_conclusion.pdf.25. WHO. Waist Circumference and Waist-Hip Ratio: Report of a WHO Expert Consultation. World Heal Organ. 2008;(December):8-11. doi:10.1038/ejcn.2009.139.26. Swainson MG, Batterham AM, Tsakirides C, Rutherford ZH, Hind K. Prediction of whole-body fat percentage and visceral adipose tissue mass from five anthropometric variables. PLoS One. 2017;12(5):e0177175. https://doi.org/10.1371/journal.pone.0177175.27. Rosner B. Fundamentals of Biostatistics. 7th Editio. Belmont: Duxbury Press; 2010.28. Hosmer D, Lemeshow S. Applied Logistic Regression. 3rd Editio. New York: John; 2000.29. Martin Bland J, Altman D. Statistical Methods for Assessing Agreement Between Two Methods of Clinical Measurement. Lancet. 1986;327(8476):307-310. doi:10.1016/S0140-6736(86)90837-8.30. Giavarina D. Understanding Bland Altman analysis. Biochem Medica. 2015;25(2):141-151. doi:10.11613/BM.2015.015.

Publication Citation

Bauza, Colleen; Martin, Renee; Yeatts, Sharon D.; Borg, K; Magwood, Gayenell; Selassie, Anbesaw; and Ford, Marvella E. (2018) Determining the joint effect of obesity and diabetes on all-cause mortality and cardiovascular-related mortality following an ischemic stroke.
10.1155/2018/4812712

Bauza, Colleen; Martin, Renee; Yeatts, Sharon D.; Borg, K; Magwood, Gayenell; Selassie, Anbesaw; and Ford, Marvella E. (2018) Determining the joint effect of obesity and diabetes on functional disability at 3-months and on all-cause mortality at 1-year following an ischemic stroke.
10.1186/s12902-018-0255-1