Page 23 - Delaware Medical Journal - July 2016
P. 23
SCIENTIFIC ARTICLE
calculating aortic valve area (AVA) when both techniques had been carried out with our CCL system, (XIMS, Philips Xper Connect, Melbourne, FL). This CCL computer system is widely used and we had presumed that the generally more accurate thermodilution CO would prevail if both had been done.3
METHODS
We studied 100 consecutive patients who underwent right and left heart catheterization from 2009-2012 for assessment of aortic valve areas and had both Fick CO calculated using femoral/radial and pulmonary arterial saturations and thermodilution CO performed for calculation of AVA. In all patients the thermodilution CO used for AVA calculation
was the average of three acceptable measurements of thermodilution CO. The number of saturations done for the Fick CO measurements was at the discretion of the individual operators. We then examined our XIMS system records for each of these procedures and documented the timing when each CO method was performed and reviewed which CO and Interventional cardiologists and CCL technicians were polled about their understanding of the operation of the CO/AVA calculations by the CCL software.
STATISTICS
Correlation of Fick and TD based CO measurements were assessed by linear regression analysis and by variance component analysis. The Pearson correlation (r) derived
from linear regression, assesses the rank ordering of variable values whereas the intra-class correlation (ICC), derived from variance component analysis assesses the degree of agreement between variable values. The ICC is the appropriate statistic for assessing agreement between tests purporting to measure the same thing on the same scale of measurement. The ICC ranges between 0, indicating no agreement, and 1, indicating perfect agreement. Scatter plots of the paired Fick and TD based CO variables included the regression line as well as a 45-degree line of equality along which the paired variables would lie if there was perfect agreement between echo-cath measurements (as would the regression line). Bland-Altman plots were also constructed as complementary assessments of echo-cath measurements.
RESULTS
CCL computer system (Fick is calculated automatically when the appropriate oxygen saturations are entered) became the determinative output for hemodynamic assessment and becomes the second CO type could be selected on the computer during the case and used for calculation, the computer screen would only temporarily display the related AVA and this would not
CO and AVA were calculated and displayed using the estimated Fick method in 32 patients, while TD calculation was used to entry to the program. Once a mixed venous and a systemic
artery pair of saturations were entered into the computer, an estimated Fick CO will be calculated. If a TD is not yet been cardiologists who changed the timing of these measurements from case to case, the determinative CO would change. A poll of 14 interventional cardiologists showed that all but one preferred the TD CO and assumed that the TD CO was used, especially since it showed temporarily on the screen during the case, None of the interventional cardiologists noticed the different calculation timing changed. A poll of 14 interventional cardiologists and nine cath lab technicians showed that none of them were aware of this property of the CCL software.
CO measurement by Fick and TD differ often in our series. The CO differed by as much as 7 L/min., with a weak correlation (R=0.56) and a weak inter class correlation (ICC=0.57). (Figure 2) Similarly the AVA calculated using the two different methods varied as widely as 1 cm2 with similarly weak correlations (R=0.73, ICC=0.72). (Figure 3) A difference in the valve area between the CO methods that would alter the recommendation for valvular replacement surgery (AVA <1.0 cm2) was found in 18 of the 100 cases.
It is possible to reverse the process and make the preferred TD CO the determinative output used for hemodynamic calculations series of 11 mouse clicks that must be done in an exact, non- intuitive sequence. There was no system documentation available refresh the screen, and reenter the second CO numbers to make
Del Med J | July 2016 | Vol. 88 | No. 7
215
