Quantitative estimation of extravascular lung water volume and preload by dynamic 15O-water positron emission tomography

AIMS: Left ventricular filling pressure (preload) can be assessed by pulmonary capillary wedge pressure (PCWP) during pulmonary arterial catheterization (PAC). An emerging method [pulse indexed contour cardiac output (PICCO)] can estimate preload by global end-diastolic volume (GEDV) and congestion as extravascular lung water (EVLW) content. However, no reliable quantitative non-invasive methods are available. Hence, in a porcine model of pulmonary congestion, we evaluated EVLW and GEDV by positron emission tomography (PET). The method was applied in 35 heart failure (HF) patients and 9 healthy volunteers. METHODS AND RESULTS: Eight pigs were studied. Pulmonary congestion was induced by a combination of beta-blockers, angiotensin-2 agonist and saline infusion. PAC, PICCO, computerized tomography, and 15O-H2O-PET were performed. EVLW increased from 521 ± 76 to 973 ± 325 mL (P < 0.001) and GEDV from 1068 ± 170 to 1254 ± 85 mL (P < 0.001). 15O-H2O-PET measures of EVLW increased from 566 ± 151 to 797 ± 231 mL (P < 0.001) and GEDV from 364 ± 60 to 524 ± 92 mL (P < 0.001). Both EVLW and GEDV measured with PICCO and 15O-H2O-PET correlated (r2 = 0.40, P < 0.001; r2 = 0.40, P < 0.001, respectively). EVLW correlated with Hounsfield units (HU; PICCO: r2 = 0.36, P < 0.001, PET: r2 = 0.46, P < 0.001) and GEDV with PCWP (PICCO: r2 = 0.20, P = 0.01, PET: r2 = 0.29, P = 0.002). In human subjects, measurements were indexed (I) for body surface area. Neither EVLWI nor HU differed between chronic stable HF patients and healthy volunteers (P = 0.11, P = 0.29) whereas GEDVI was increased in HF patients (336 ± 66 mL/m2 vs. 276 ± 44 mL/m2, P = 0.01). CONCLUSION: The present study demonstrates that 15O-H2O-PET can assess pulmonary congestion and preload quantitatively. Hence, prognostic information from 15O-H2O-PET examinations should be evaluated in clinical trials.