Clinical Utility of Central Venous Saturation for the Calculation of Cardiac Index in Cardiac Patients

BackgroundMixed venous saturation (MVS) obtained from the distal pulmonary artery (PA) during Swan-Ganz catheterization is the criterion standard for calculating cardiac output (CO) and cardiac index (CI) with the use of the Fick method. We think that calculating CI with the use of central venous saturation (CVS) instead of PA-MVS is both feasible and accurate. Earlier studies were small, enrolled heterogeneous patient populations, and resulted in inconsistent findings.MethodsAll patients undergoing right heart catheterization from January 2011 to January 2012 in our catheterization lab with simultaneous measurements of MVS obtained from the distal PA and CVS obtained from the superior vena cava (SVC) or right atrium (RA) were included. Out of the 902 patients enrolled, we excluded patients (n = 50) who had known cardiac shunt or dialysis fistula, had duplicate medical records, or were septic. We calculated the CI with the use of the assumed Fick method using both MVS (criterion standard) and CVS (SVC or RA saturations) in the remaining 852 patients. We measured the correlation and the agreement between the 2 methods with the use of the Pearson correlation coefficient and Bland-Altman analysis.ResultsTotals of 112 patients with simultaneous PA and RA saturation measurements (group I) and 740 patients with simultaneous PA and SVC saturation measurements (group II) were included. We found an excellent linear correlation between SVC and PA saturation (r = 0.928) and between RA and PA saturation (r = 0.95). There was also an excellent correlation between CI calculated with the use of PA saturation and CI calculated with the use of SVC (r = 0.87) or RA (r = 0.93) saturation. The mean bias of CVS-derived CI compared with MVS-derived CI (criterion standard) was −0.1 (95% limits of agreement [LOA] −1 to +0.77) in the SVC group and −0.006 (LOA −0.68 to +0.69) in the RA group. Patients with low CI had stronger correlation and smaller bias between the 2 methods compared with those with normal or high CI. The presence of baseline hypoxemia, valvular heart disease, or acute coronary syndrome had no significant effect on the correlation or the bias between the 2 methods.ConclusionsIn cardiac patients, CVS can be used as a surrogate to true MVS in the calculation of CI. This method is readily available in patients who have central venous access, and may aid in early goal-directed treatment when cardiogenic shock is suspected.