A Simple Prediction Rule for All-Cause Mortality in Bariatric Surgery Eligible PatientsThursday, October 17, 2013
Regular readers will be quite familiar with our previous work on the Edmonton Obesity Staging System (EOSS), which ranks bariatric patients on a five-point ordinal scale based on the presence and severity of functional, mental and medical health problems.
As we showed in our analyses of several large datasets, individuals at higher EOSS stages are at far greater risk of all-cause mortality than individuals at lower EOSS stages. Interestingly enough, we found that BMI levels contribute little, if anything, to the actual mortality risk of these individuals – apparently, all that counts is how “sick” you are, not how “big” you are.
While EOSS is gradually winding its way around the globe towards greater popularity and acceptance, my colleagues and I now publish an even simpler rule for predicting all-cause mortality in bariatric patients – the paper was just released online at JAMA-Surgery.
In this paper, we studied over 15,000 individuals from the United Kingdom General Practice Research Database (GPRD), a population-representative primary care registry, who met current eligibility criteria for bariatric surgery (BMI, ≥35.0 alone or 30.0-34.9 with an obesity-related comorbidity) between January 1, 1988, through December 31, 1998.
We used binary logistic regression to construct a parsimonious model and a clinical prediction rule for 10-year all-cause mortality.
The final model, which included age, type 2 diabetes mellitus, current smoking, and male sex had a concordance or C-statistic of 0.768.
Based on this model, we developed a simply clinical prediction rule, scoring into 4 tiers with 10-year all-cause mortality ranging from was 0.2% in tier 1 to5.2% in tier 4.
Although BMI significantly (albeit poorly) predicted mortality, it did not add much to the model in terms of discrimination or calibration.
Thus, our findings show that all-cause 10-year mortality in obese individuals eligible for bariatric surgery can be estimated using a simple 4-variable prediction rule based on age, sex, smoking, and diabetes mellitus.
Once again (as in EOSS), body mass index was not an important mortality predictor.
These findings may have important consequences for prioritization of patients for bariatric surgery, at least if one chooses to prioritize individuals with the highest mortality risk – these would be older men with type 2 diabetes, who smoke (whereby, one would assume that they would immediately stop smoking if nothing else).
However, we also realize that mortality risk is only one consideration that goes into deciding who will benefit the most from surgery.
Certainly, severe obesity is associated with numerous other important consequences on mental, physical and economic health, which, although not lethal, can well make life quite unpleasant. The positive impact of bariatric surgery on these problems is well documented and probably as (if not more) important to people living with obesity than simply staying alive.
Nevertheless, to clinicians and administrators the message is clear – BMI alone is not a good predictor of health and certainly not a predictor of mortality. Prioritization systems based on BMI should be abandoned.