Introduction
Body mass index (BMI) has long been accepted as the main indicator of obesity. Studies have shown that dialysis patients with a lower BMI have poor survival and higher values of BMI seemed to be associated with better survival. This study aims to look at the association of baseline BMI with mid-term survival in a multiethnic South East Asian haemodialysis population.
Method
Patients who joined our institution for haemodialysis programme from 2014 to 2018 were included in this study. BMI was obtained as part of their nutritional assessment during pre-admission clinic. Patients who had major limb amputation (below/through/above knee amputation, prior or during the study period) were excluded in this study. All patients were followed until their termination from HD programme or end of the study (31/12/2018). Patients’ demographic, clinical and nutritional information were obtained from our clinical database. The association of baseline BMI with mid-term all-cause mortality was explored using multivariate Cox regression. Statistical significance was defined by p<0.05.
Result
A total of 2515 incident HD patients were included in the study with 57% male, 71.7% diabetic, 58% Chinese. Their mean age was 62±12years and mean baseline BMI was 24.1±4.9kg/m2. The mean follow up was 2.2±1.4years. Patients were grouped into four categories according to their baseline BMI: underweight (BMI<18.5kg/m2, n=234(9.3%)), normal weight (18.5-24.9kg/m2, n=1373(54.6%)), overweight (25-29.9kg/m2, n=618(24.6%)) and obese (BMI>=30kg/m2, n=290(11.5%)). There was no statistically significant association between baseline BMI and mid-term all-cause mortality after adjusting for confounding factors: age, race, gender, diabetes status, dialysis vintage, acuity status and nutritional status (Underweight: HR=1.31, 95%CI=0.91,1.86, p>0.05; normal weight: reference group; overweight: HR=0.92, 95%CI=0.69,1.24, p>0.05; obese: HR=1.42, 95%CI=0.99,2.04, p>0.05). 
Conclusion
This study found that baseline BMI was not significantly associated with mid-term all-cause mortality. Further studies on the association of time-varying patient’s profile with mortality in HD patients is warranted.