Cardiovascular Diseases Poster Session
Discussion and Conclusion
In assessing the value of a new method it is customary to compare it with the gold standard. In the measure of cardiac output, the European Society of Cardiology identify dye dilution as the optimal method with which all comparisons should be made (Spiering et al., 1998). However, dye dilution is highly invasive and only appropriate in selected ill patients. Invasive methods are only capable of producing intermittent data and are completely unsuitable for continuous monitoring. As there are no ‘gold standards’ in the non-invasive measurement of cardiac output, new methods have to be assessed by comparison with established techniques.
Our results demonstrate good correlation between ICG and the Portapres, with all parameters being significant at p < 0.01. However, we are reminded by Bland and Altman that correlation coefficients are not accurate in the evaluation of agreement, they merely describe the relation between two variables.
Heart rates recorded by the two methods were almost identical and the mean discrepancy was less than one beat per minute. These readings were also highly consistent, with the 95% limits of agreement being narrow (-9.2 to 7.3 bpm). Here the degree of agreement is acceptable.
The Portapres underestimated MAP by 7.8 mmHg. This represents a real physiological difference as the Portapres measures digital artery, as compared to brachial artery pressure. When relative changes in MAP were compared during exercise a 6% underestimate for Portapres was still occurring, associated with wide limits of agreement (-28.1% to 40.2%). The nature of the protocol made accurate comparisons between the continual beat-to-beat recordings of the Portapres and the interval data from the sphygmomanometer difficult. It would be interesting to see if the two methods corresponded better in a more stable environment with pharmacological manipulation.
Perhaps the most prized parameter of all to the clinician is that of cardiac output. If the heart is thought of as a pump, the output is a measure of its efficiency. It is unfortunately also the most difficult measure on which to obtain accurate information.
The mean difference in CI between the Portapres and ICG was large and clinically significant. In absolute terms the Portapres was underestimating the CI by 2.4 l/min/m2. In the clinical setting an underestimate on this scale could lead to inappropriate changes in therapy. The reason for an error of this magnitude is not immediately obvious. The Portapres software computes an aortic flow waveform from the arterial pressure waveform in the finger. The size of the aortic flow waveform depends on the cross sectional area of the aorta, which is estimated by the software program based on the patient's age, sex, height and weight (Wesseling et al., 1993). Actual differences in aortic root diameter will therefore lead to errors in the estimation of absolute CI. The software manufacturers accept that cardiac output can only be approximately calibrated in absolute terms and are hopeful that a new version of the software (available end 1998), based on a larger patient database, will result in more accurate absolute measurement of cardiac output. In addition, the magnitude of this difference can be explained in part by inaccuracies of the ICG, which is known to overestimate CI by up to 20-30% in healthy young volunteers during exercise (Bernstein DP, 1989).
Theoretically, inaccurate estimation of aortic flow would have no bearing on the ability of the Portapres to monitor change in cardiac output from baseline. Indeed the mean discrepancy between percentage change measurements was only 6.1%. However, the standard deviations of the mean difference were large with limits of agreement ranging from -52.6% to 64.8%. Whilst reproducibility at rest was excellent, it was not possible to assess reproducibility on exercise, due to the dynamic nature of the cardiovascular system after exercise. Therefore, the large differences found between pairs of readings cannot directly be attributed to poor reproducibility of either method. We can speculate that imprecision is more likely to be related to the ICG which is known to be effected by movements of the chest wall. In our protocol, exaggerated thoracic movements on recovery from exercise would effect the reproducibility of the ICG, and is the most obvious explanation for the wide limits of agreement obtained.
The Portapres has a number of attractive features that make it a valuable additional tool to both the clinician and researcher: its ease of application, simplicity of operation, portability and continuity of measurements, beat-to-beat. This study found a good correlation between the Portapres and ICG, with good reproducibility at rest and good agreement between change in cardiac parameters with exercise. However, absolute differences were large and of clinical significance. Further validation of the finger blood pressure monitor is required, in particular a comparison with pharmacological manipulation of parameters would be of interest.
This study was supported by a grant from the High Blood Pressure Foundation.
| Discussion Board | Next Page | Your Poster Session |
|Mills, NL; Spratt, PL; Padfield, DJ; Webb, DJ; (1998). The PORTAPRES In The Non-invasive Assessment Of Dynamic Cardiovascular Function. Presented at INABIS '98 - 5th Internet World Congress on Biomedical Sciences at McMaster University, Canada, Dec 7-16th. Available at URL http://www.mcmaster.ca/inabis98/cvdisease/mills0906/index.html|
|© 1998 Author(s) Hold Copyright|