TY - JOUR T1 - 142 Estimations of plasma volume status in patients with chronic heart failure: a useful tool for diagnosing and treating congestion? JF - Heart JO - Heart SP - A107 LP - A109 DO - 10.1136/heartjnl-2021-BCS.139 VL - 107 IS - Suppl 1 AU - Joe Cuthbert Y1 - 2021/06/01 UR - http://heart.bmj.com/content/107/Suppl_1/A107.abstract N2 - Introduction Neurohormonal activation in patients with chronic heart failure (CHF) causes plasma volume expansion which, if untreated, leads to overt venous congestion. Plasma volume status (PVS) can be estimated using formulae based on a patient’s sex, weight, haemoglobin and haematocrit. Such non-invasive methods to assess congestion may be useful, particularly in the wake of the COVID-19 pandemic. We compared the clinical value of two measures of PVS in a cohort of unselected outpatients with CHF (The Hull LifeLab).Methods Patients with an echocardiogram, N-terminal pro B-type natriuretic peptide (NTproBNP) and complete data on signs and symptoms were evaluated (n=3505). CHF was defined as signs and symptoms of the disease and either left ventricular systolic dysfunction (LVSD) worse than mild, or LVSD mild or better and raised N-terminal pro-B-type natriuretic peptide (NTproBNP) levels (>125 ng/L). Two formulae to estimate PVS were used: (a) Hakim formula (based on estimations of actual and ideal plasma volume calculated from weight and haemoglobin - H-PVS); and (b) Duarte formula (calculated from haematocrit and haemoglobin – D-PVS) (figure 1). Patients were split into quartiles of H-PVS and D-PVS. Variance inflation factor (VIF) was used to assess co-linearity between both measures and other variables. Outcome measures were all cause mortality, and mortality or heart failure hospitalisation. Multivariable Cox regression, Harrel’s C-statistic and Akaike’s Information Criterion (AIC) were used to assess the prognostic utility of each measure of PVS.Results Patients in the highest quartile (most congested) of D-PVS or H-PVS had higher NTproBNP levels, were more likely to have severe symptoms and signs, and be prescribed a loop diuretic than those in lower quartiles (table 1 and 2). Patients in the highest quartile paradoxically weighed much less than those in those lowest quartiles. H-PVS and D-PVS were strongly correlated with one another (r=0.79; P<0.001) but had only modest positive correlations with log[NTproBNP] (r=0.32 and r=0.24 respectively; P<0.001). There was extreme co-linearity between D-PVS and haemoglobin in linear regression models with other continuous variables (VIF = 14). Higher H-PVS (hazard ratio (HR) = 1.01 (95% confidence interval (CI) = 1.00 – 1.02); P=0.002) or D-PVS (HR = 1.08 (95% CI = 1.03 – 1.14); P=0.002) was associated with greater risk of all-cause mortality or hospitalisation with heart failure. Neither the Harrel’s C-statistic nor AIC of outcome models for 1 year mortality including either weight or haemoglobin were improved by the addition of H-PVS or D-PVS (figure 2).Abstract 142 Figure 1 Abstract 142 Figure 2 View this table:Abstract 142 Table 1 Patient characteristics by quartile of D-PVSView this table:Abstract 142 Table 2 Patient characteristics by quartile of H-PVSConclusions Changes in weight or haemoglobin in patients with CHF are not always due to changes in plasma volume. Despite apparent associations with disease severity, H-PVS and D-PVS are just surrogates of the variables from which they are calculated and not reliable measures of congestion. It is likely neither has any clinical utility.Conflict of Interest None ER -