Obesity is associated with increased morbidity but not mortality in critically ill patients

Obesity is associated with increased morbidity but not mortality in critically ill patients

Sakr Y. Intensive Care Med 2008;34:1999–2009

Clinical Question

  • In patients admitted to the ICU do overweight/obese patients compared with normal weight patients have a greater risk of morbidity and mortality?


  • Planned post-hoc analysis of a previous trial: The SOAP Study
  • Prospective, observational cohort study
  • Consecutive sampling
  • Pre-printed data collection forms
  • Follow up until death, hospital discharge or for 60 days
  • A priori statistical plan
  • Multivariate Cox proportional hazard model with time to in-hospital death right censored at 60 days
  • Variables included in the Cox regression analysis were age, gender, co-morbid diseases, SAPS II and SOFA scores on admission, the type of admission (medical or surgical), the presence of sepsis, and the need for mechanical ventilation or renal replacement therapy during the ICU stay[short key points]


  • 198 European intensive care units
  • Data collection: 1st-15th May 2002


  • Inclusion: All patients >15yrs old admitted to ICU
  • Exclusion: ICU Stay <24hrs AND only admitted for routine post-operative care/monitoring
  • 3147 patients enrolled, Body Mass Index (BMI) reported in 2878


  • 4 sub-groups
    • Underweight (BMI <18.5 kg/m2), n=120
    • Overweight (BMI 25–29.9 kg/m2), n=1047
    • Obese (30–39.9 kg/m2), n=424
    • Very obese (≥40 kg/m2), n=81


  • Normal weight: (BMI (18.5–24.9 kg/m2), n=1206
  • For all patients: Body mass index calculated from actual measurement or ‘best clinical estimate’
  • Baseline characteristics:
  • Significant age difference between:
    • Overweight (63.6 years) and obese (63.5 years) patients, compared with normal weight patients (58.4 years), p<0.01
      Underweight patients (52.6 years) compared with normal weight patients, p<0.01
  • Very obese (61.7%) and under weight (53.2%) patients were significantly more commonly female compared with normal weight patients (38.1%)
  • Comparing underweight patients vs. normal weight patients
  • The SOFA score on admission was significantly lower: 3.7 vs. 5.1, p<0.01
  • The incidence of respiratory failure on admission was significantly lower: 12.5% vs. 20.8%, p<0.01. (In comparison the incidence in obese patients was 25.9% and 24.7% in very obese.)
  • BMI unrelated to incidence of diabetes, heart failure or cancer in the study sample


  • Primary outcome: not defined
  • Pre-determined outcomes:
    • Mortality
      • ICU and hospital mortality was unrelated to BMI
        SAPS II adjusted odds ratio for hospital mortality was not significantly increased in any of the weight categories compared with normal weight in a subgroup analysis stratifying patients according to sex, age, the presence of sepsis, septic shock, or shock due to any cause, the type of admission (medical or surgical), and ICU length of stay of more or less than 2 days
        Statistically significant increased incidence of ICU-acquired infections in obese (10%, p<0.01) and very obese (12%, p<0.05) groups compared to normal weight (9%). However, incidence of sepsis syndromes similar among groups
        Significantly higher rate incidence of respiratory organ failure in obese patients (46%) and very obese patients (47%) compared with normal weight (40%), p<0.05
    • Length of stay: very obese patients, compared with normal weight patients, showed a trend towards longer
    • ICU stay, median (IQ): 4.1 (1.8–12.1) vs. 3.1 (1.7–7.2), days, p = 0.06
    • Hospital stay: 14.3 (8.4–27.4) vs. 12.3 (5.1– 24.4), days, p = 0.08
    • Mean SOFA scores during ICU stay were lower in under weight vs. normal weight patients (3.5 vs. 4.5, p<0.01)
  • Post-hoc analysis
    • In patients who were mechanically ventilated for more than 2 days, trend towards a higher mortality in underweight patients compared with those with normal BMI: SAPS II adjusted odds ratio 1.84 (95% CI: 0.94–3.61, p=0.077)

obesity1 2

Authors’ Conclusions

  • BMI does not have a significant impact on mortality in European ICU patients, although patients with higher BMI developed ICU-acquired infections more frequently


  • Large heterogenous population representing normal intensive care admissions (good external validity)
  • Interesting epidemiological data collected regarding obese patients on intensive care
  • BMI recorded in majority of patients (91%)
  • Very little loss to follow up/low attrition bias (although not clearly stated)


  • Use of BMI:
    • No record of how many patients had estimated weight and height and allowing ‘estimated BMI’ may seriously effect internal validity
      BMI may not be an appropriate measure of obesity in critically unwell patients (recent weight change, volume depletion/overload)
  • Authors acknowledge there may be logistical issues with discharge of very obese patients from intensive care which may inappropriately increase length of stay
  • Small subgroup size of very obese patients (n=81)
  • No power calculation
  • Did not analyse for incidence of venous thromboembolism, for which obesity is a known risk factor
  • Selection bias: obese patients with more comorbidities referred to intensive care may not have been admitted. Therefore the sample may be a more medically and functionally fit cohort of obese patients (note no significant increase in incidence of diabetes and heart failure in obese groups)
  • Not registered on

The Bottom Line

  • The study is unable to conclusively say whether obesity significantly impacts on morbidity and mortality in ICU patients due to methodological problems impacting on its internal validity and lack of clarity whether sufficiently powered to detect an effect. I will continue to prognosticate and manage patients based on underlying disease process and co-morbidities rather than BMI

External Links


Summary author: @IntenseMedic
Summary date: 20th December 2015
Peer-review editor: @davidslessor

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.