May 25, 2018

expert reaction to four conference posters (unpublished work) on the ‘obesity paradox’, presented at the European Congress on Obesity

Researchers presenting at the European Congress on Obesity in Vienna report unpublished work on the ‘obesity paradox’.

Dr David Carslake, Senior Research Associate, MRC Integrative Epidemiology Unit at the University of Bristol, said:

“These are interesting and mostly well-powered studies showing convincing evidence of an association between high BMI and increased survival within specific patient groups, but we have to remember that such an association does not necessarily mean that high BMI causes the increased survival.  If the high BMI did not cause the increased survival, then patients would get no benefit from increasing their BMI, despite the association.  The Molinger & Gommers abstract provides evidence supporting one plausible causal mechanism – overweight people suffering reduced muscle wastage – but was based on only 26 patients so would benefit from replication in a larger sample.

“There are two main ways in which high BMI could be associated with increased survival in these patient groups which do not involve a causal effect of BMI on survival.  First, unmeasured factors such as smoking behaviour or other diseases could cause both a reduction in BMI and increased mortality.  The authors took steps to adjust for these, but they can never be fully accounted for by the methods they used.  Second, it’s possible that overweight and obese people get admitted to hospital more readily with less severe cases of sepsis, infection or pneumonia.  This is a risk whenever analyses are restricted to specific patient groups, and in this case could give a spurious impression that high BMI was protective.

“It would be unwise to change recommendations about BMI, even in subgroups of patients, without also considering whether BMI affects the chances of getting these conditions in the first place, and what the effect would be on the rates of sickness and death from other causes.  I note that while authors Lin and Wolfshohl point out that BMI could be used for risk stratification of sepsis patients (which does not require causality), in none of these abstracts do the authors call for BMI recommendations to be changed.”

* Abstract titles:

‘The Impact of Body Mass Index on Mortality after Infection: Cohort study of 18,167 patients hospitalized in Denmark’ by S. B. Gribsholt et al. 

‘Obesity with lower comorbidity burden is associated with substantial survival improvement in patients hospitalized with pneumonia’ by S. Chang.

‘The Impact of Being Overweight and Obese on Patients with Sepsis’ by Y. Lin.

‘Effect of obesity in critically ill patients; Muscle Quality as an explanatory outcome for the “Obesity Paradox”’ by J. Molinger et al.

These are conference abstracts / posters that were discussed at the European Congress on Obesity (ECO).  There are no papers as this is not published work.

Declared interests

Dr David Carslake: “I have previously published studies in a related field but have no other interests to declare.”