Research Review 


by Lindsay Buchanan, RD 


Why this article? 

Our first research review comes from dietitian Lindsay Buchanan who recently attended Dr. Sharma's 5A's of Obesity Workshop in Halifax. After hearing Dr. Sharma speak about the "obesity staging system" used in his Edmonton clinic, she wanted to read more about the research behind this interesting tool.


Research Article:

Padwal, R.S., Nicholas, M.P., Allison, D.B. & Sharma, A.M. (2011) Using the Edmonton obesity staging system to predict mortality in a population-representative cohort of people with overweight and obesity. Canadian Medical Association Journal, 183(14), E1059-E1066.


Type of Study:

Retrospective analysis



Body Mass Index (BMI) is commonly used as a measure to classify excess adiposity and is consistently associated with morbidity and mortality in observational studies. However, BMI also has well-cited limitations such as not differentiating between differences in body composition e.g. an individual’s percentage of lean muscle versus fat. The authors of this paper identify that BMI also does not directly reflect “…the presence of underlying obesity-related comorbidity, reduced quality of life or diminished functional status – elements that are widely considered to be critically important to the clinical assessment of patients with excess body weight” (pp E1059-E1060). They propose a new system of evaluating mortality in overweight and obese patients which incorporates the presence of comorbidities – the Edmonton Obesity Staging System (EOSS). 



Data collection

  • National Health and Human Nutrition Examination Surveys (NHANES) III 1988-1994 and 1999-2004.

  • The National Death Index and information from public mortality files up to December 31st, 2006.

  • Ethnicity was categorized as (1) non-Hispanic white, (2) non-Hispanic black, (3) Mexican-American and (4) other.

  • Smoking status was either (1) smoker, (2) former smoker and (3) never smoked.

  • Weight, height and waist circumference was measured by trained examiners. Health history was recorded including risk factors such as blood pressure, serum lipids, fasting glucose levels etc.


Sample population

  • To be eligible individuals had to be non-pregnant adults (> 20 y.o.), classified as overweight or obese, who had no history of cancer, and were randomized to the morning sessions of the mobile examination centre (AM sessions allowed for fasting glucose measurements).

  • In total 8143 individuals from NHANES III and NHANES 1999-2004 (4373 and 3770) were eligible. Final sample population size was 7967 (4367 and 3600).



  • Kaplan-Meier plots and Cox proportional hazards regression.

  • Separate baseline hazards were fit for each race/ethnicity and smoking status. Individuals who reported never smoking were analyzed independently to eliminate potentially confounding effects.


Key Findings:

The EOSS was a stronger predictor of increasing mortality risk independent of BMI, hypertriglycerideic waist and metabolic syndrome.

Bottom Line:

The EOSS appears to be a more precise method of assessing overweight and obesity risk in comparison to BMI and can be a useful tool when developing and prioritizing health interventions. The results of this research indicate that the best predictor of mortality is not soley weight but examining the whole picture of health which includes the presence of other risk factors, chronic disease, end-organ damage and disabilities. In other words, obesity on its own may increase the RISK of mortality but not necessarily the severity. Further research addressing limitations of the EOSS would be a helpful addition to this literature.


Strengths and Limitations:



  • Size of the sample population.

  • National representation within the sample population.

  • Detailed information on height, weight, waist circumference, fasting glucose, blood pressure, serum blood lipids, and health history.

  • Adjusting for hypertriglycerideic waist and metabolic syndrome. 



  • Some comorbidities within the EOSS have been arbitrarily assigned e.g. diabetes without differentiation between type 1 and type 2. Further data examining the burden of illness would be helpful when establishing the weighting of specific comorbidities.

  • Due to insufficient information on specific comorbidities ( e.g. sleep apnea, obesity hypoventilation syndrome, gastroesophageal reflux disease) and psychological data - these were not included in the final EOSS scores.

  • This retrospective study only examines mortality as part of the scoring system without incorporating elements related to quality of life. .

  • DNSS facebook

© 2018 by The Dietitians Network of Nova Scotia.