Declaration of interest: Both authors have helped develop and have published on the Modelflow method (mentioned in reference 8). The second author holds a patent on that method.
Theoretically, tissue flow should be maintained at the lowest possible arterial perfusion pressure to minimize afterload and myocardial energy expenditure.1 In the monitoring and measurement of cardiovascular parameters in critically ill patients, however, cardiac output plays a subservient role to arterial blood pressure. Since blood pressure is the principal cardiovascular parameter monitored by the body via baroreceptors, this seems physiological. Furthermore, continuous cardiac output has proved extremely difficult to monitor.
Intermittent methods such as indicator dilution are widely used, but for monitoring purposes cardiac output should be available continuously, like arterial pressure.2 The paper by Linton and Linton in this issue3 offers a solution to the monitoring problem by deriving cardiac output from arterial pressure via a cascade of chained models. Such so-called pulse contour methods attempt to determine cardiac stroke volume from characteristics of the arterial pressure pulse. With a first publication in 1904,4 pulse contour preceded Korotkoffs auscultation paper by 1 year. Both ideas were embraced almost immediately, yet clinical acceptance of the pulse contour and auscultatory methods has differed. Why? Because pulse contour methods are based on solid physical principles and less solid physiological models and they involve substantial computations, all factors that do not ease clinical acceptance.
Major concerns about pulse contour analysis
Non-linearity
If it was possible to know the degree to which the aorta complies to a 1 mm Hg pressure increment and to measure its reponse, then the stroke volume ejected into the aorta could be computed by measuring the associated increase in pressure.4 For instance, if this compliance was 2 ml per 1 mm Hg pressure increment, then a pressure rise of 40 mm Hg would correspond to a stroke volume of 40x2=80 ml.
The aortas compliance at low distending pressure is substantial but at higher pressures compliance progressively decreases, resisting overstretching of the aortic wall. Such non-linear behaviour effectively prevents any simple approach to pulse contour estimates of stroke volume.
Typically, for a 50-yr-old patient, aortic compliance might be 3, 1.1 and 0.5 at distending pressures of 50, 100 and 150 mm Hg, respectively. Thus the same 40 mm Hg pulse pressure equates to stroke volumes of 120, 44 and 20 ml, depending on pressure. These values also change with the patients age; non-linearity is more pronounced in the elderly.
The significance of aortic non-linearity in relation to pulse contour measurement was appreciated as early as 1928,5 and that of age dependency in 1954,6 but these two factors were only measured with great precision in 1984.7 To our knowledge, the variables have been taken into account fully when measuring aortic flow by the Modelflow method only.8
Radial artery pressure
The pressure that determines stroke volume is proximal aortic pressure, which is not routinely available. More peripherally recorded pressure waves can be substituted if they bear a known, fixed, and wide bandwidth relationship to aortic pressure. This is almost the case for arterial pressures measured at sites along the arteries in the arm (see references cited by Linton and Linton3). The usable physiological bandwidth is amore than adequate15 Hz, even to the finger, but a strong resonance is present which must be carefully compensated for by an antiresonance.9
Damped waveforms
Intra-arterial pressure is measured predominantly with catheter-transducer systems. With continuous flush devices these systems have become very reliable. Nevertheless, in view of the exposure to blood which this technique produces, a non-invasive method that monitors the quality of the recorded pressure pulse10 is much needed; it should also signal at an early stage the presence of inadequate recordings.3 Non-invasive measurement of finger arterial pressure has no significant instrumental damping. Its innovative use with the Modelflow method in an ICU environment has been investigated recently.11 12
Inadequate pulse detection
To measure stroke volume, each individual heart beat must be identified in the continuous pressure tracing. Digital computer programs can be designed that hardly ever miss or create a beat, even in the presence of severe arrhythmias or damped pressure curves or when recorded peripherally, such as in the radial artery. Yet, such algorithms are not always implemented.13
Absolute cardiac output
The dependence of the properties of the aorta on age, gender, distending pressure and arteriosclerosis are now understood in great detail, but still we are left with aortic diameter at maximal pressure as the single parameter that may vary up to ±40% from the population average.7 11 The effect is that absolute levels of cardiac output cannot be determined with certainty, although changes in cardiac output can be tracked with precision. These problems have been recognized since 1904. Böger and Wezler14 were the first to propose a once-only calibration of pulse contour cardiac output by another method such as indicator dilution. It was recently shown15 that this calibration method may be applicable even after 48 h of monitoring in ICU patients.
Minor concerns
Aortic pathology
A patent aortic valve is required to prevent any back flow that is not modelled. In addition, aortic aneurysms modify the properties of the aorta in unpredictable ways. Changes with sympathetic outflow in smooth muscle tone in the aortic wall may also modify these properties. The amount of aortic smooth muscle, however, is small and major effects have not been reported.
Body position
The properties of the aorta change with distending pressure and so depend on the patients posture. Yet, between the supine and erect positions, offsets in young adult subjects appear small.16 Obese patients may show differences in pulse contour-estimated cardiac output due to abdominal compression when moved from the supine to the prone position, however.
Conclusion
Since the major concerns about continuous measurement of cardiac output now have solutions, methods that include these solutions will probably prove reliable and sufficiently precise in tracking changes in cardiac output in critically ill patients. Linton and Linton3 go a long way in this direction, but early reports cannot be fully accepted before well controlled17 clinical studies have established the facts in context. Fully non-invasive cardiac output monitoring from arterial pressure could at last be in sight.
J. J. van Lieshout
K. H. Wesseling
Academic Medical Centre
Meibergdreef 9
1105 AZ Amsterdam
The Netherlands
References
1 Feddersen K, Aren C, Nilsson NJ, Radegran K. Cerebral blood flow and metabolism during cardiopulmonary bypass with special reference to effects of hypotension induced by prostacyclin. Ann Thorac Surg 1986; 41: 395400[Abstract]
2 Gardner RM. Continuous cardiac output: how accurately and how timely? Crit Care Med 1998; 26: 13023[ISI][Medline]
3 Linton NWF, Linton RAF. Estimation of changes in cardiac output from the arterial blood pressure waveform in the upper limb. Br J Anaesth, 2001; 86: 48696
4 Erlanger J, Hooker DR. An experimental study of blood pressure and of pulse-pressure in man. Johns Hopkins Hosp Rep 1904; 12: 145378
5 Liljestrand G, Zander E. Vergleichende Bestimmungen des Minutenvolumens des Herzens beim Menschen mittels der Stickoxydulmethode und durch Blutdruckmessung. Z Ges Exp Med 1928; 59: 10522
6 Starr I. Clinical tests of the simple method of estimating cardiac stroke volume from blood pressure and age. Circulation 1954; 9: 66481[ISI]
7 Langewouters GJ, Wesseling KH, Goedhard WJA. The static elastic properties of 45 human thoracic and 20 abdominal aortas in vitro and the parameters of a new model. J Biomech 1984; 17: 42535[ISI][Medline]
8 Wesseling KH, Jansen JRC, Settels JJ, Schreuder JJ. Computation of aortic flow from pressure in humans using a nonlinear, three-element model. J Appl Physiol 1993; 74: 256673[Abstract]
9 Karamanoglu M, ORourke MF, Avolio AP, Kelly RP. An analysis of the relationship between central aortic and peripheral upper limb pressure waves in man. Eur Heart J 1993; 14: 1607[Abstract]
10 Prentza A, Wesseling KH. Catheter-manometer system damped blood pressures detected by neural nets. Med Biol Eng Comput 1995; 33: 58995[ISI][Medline]
11 Hirschl MM, Binder M, Gwechenberger M et al. Noninvasive assessment of cardiac output in critically ill patients by analysis of the finger blood pressure waveform. Crit Care Med 1997; 25: 190914[ISI][Medline]
12 Popovich MJ, Hoffman WD. Noninvasive cardiac output monitoring. Crit Care Med 1997; 25: 17834[ISI][Medline]
13 Rödig G, Prasser C, Keyl C, Liebald A, Hobbhahn J. Continuous cardiac output measurement: pulse contour analysis vs thermodilution technique in cardiac surgical patients. Br J Anaesth 1999; 82: 52530
14 Böger A, Wezler K. Zur physikalischen Bestimmung absoluter Werte des Schlag- und Minutenvolumens des Herzens beim Menschen. Arch Exp Path Pharmacol 1937; 184: 25375
15 Jellema WT, Wesseling KH, Groeneveld ABJ, Stouteribeek CP, Thijs LG, van Lieshout JJ. Continuous cardiac output in septic shock by simulating a model of the aortic input impedance. A comparison with bolus injection thermodilution. Anesthesiology 1999; 90: 131728[ISI][Medline]
16 Harms MPM, Wesseling KH, Pott F et al. Continuous stroke volume monitoring by modelling flow from non-invasive measurement of arterial pressure in humans under orthostatic stress. Clin Sci 1999; 97: 291301[ISI][Medline]
17 Siegel LC, Pearl RG. Noninvasive cardiac output measurement: troubled technologies and troubled studies. Anesth Analg 1992; 74: 7902[ISI][Medline]