Introduction
Obesity is associated with lower natriuretic peptide (NP) levels, which has implications for the diagnosis of heart failure (HF). It might be that the lower NP is due to obese subjects becoming symptomatic with lesser degrees of left ventricular dysfunction than normal weight subjects. We therefore explored the relation between body mass, left ventricular dysfunction and NT-proBNP in a large cohort of patients being investigated for possible HF. Methods
We report on 2866 patients referred with suspected HF to a community heart failure clinic between January 2001 and September 2023. All patients had BMI, NT-proBNP, and covariate baseline values (age, sex, creatinine, left ventricular ejection fraction (LVEF), New York Heart Association (NYHA) class, and cardiac rhythm). Patients were categorised as 'lean' BMI <25 kg/m2, 'overweight' BMI 25–29.9 kg/m2, and 'obese' BMI ≥30 kg/m2. NYHA class III and IV were combined to ensure comparable sample sizes. Results
The mean age was 71 ± 11 years, 63% male, mean BMI was 29 ± 6 kg/m2, and mean LVEF was 47 ± 14%. Around one-third had LVEF ≤40%. There were fewer patients with LVEF ≤40% in higher BMI categories (p <0.05). A quarter were in NYHA class III-IV. The median NT-proBNP level was 633 (170 – 1957) ng/L. Hypertension was common, especially in higher BMI categories (p <0.05). The median creatinine level was 95 (79 – 119) μmol/L. There was an inverse relation between NT-proBNP and BMI: median (25th and 75th centile) NT-proBNP was: 1269 (324 – 3451) ng/L, 643 (171 – 1886) ng/L, and 402 (137 – 1244) ng/L in lean, overweight, and obese groups respectively (p <0.001). Within each NYHA class, there was an inverse relation between log NT-proBNP and LVEF across different BMI categories, with similar slope coefficients (-0.064 to -0.040, p <0.001). However, for any given LVEF, log NT-proBNP level was lower in higher BMI groups. Average intercepts on a plot with LVEF on the X axis and log NT-proBNP on the Y were 9.3, 8.7, and 8.2 kg/m2 in lean, overweight, and obese categories respectively (p <0.001). There was an inverse relation between BMI as a continuous variable and NT-proBNP in the multivariable regression model (p <0.001). Each one unit increase in BMI (1 kg/m2) was associated with a 3% lower NT-proBNP level (p <0.001). Each SD increase in BMI (5.7 kg/m2 in our sample) was associated with a 10% lower NT-proBNP level (p <0.001). Age, sex, creatinine, LVEF, presence or absence of NYHA III-IV, and presence or absence of atrial fibrillation were also independent predictors of NT-proBNP levels (p <0.001). Among all the independent predictors identified, BMI had the weakest association with NT-proBNP. Conclusion
In patients referred for suspected HF, there is an inverse relation between BMI and NT-proBNP. The effect remains clinically significant even when relevant confounders are considered. Conflict of Interest
None