Pressure Waveform Analysis

PulseCO – Pressure Waveform Analysis

History of Arterial Pressure Waveform Analysis

Arterial blood pressure is maintained by ensuring an adequate cardiac output (blood flow) and appropriate vascular resistance (SVR).

Mathematically the average/ mean arterial pressure MAP = CO*SVR. Cardiac output units of flow are litres per minute. The cardiac output is the product of the stroke volume ejected per beat times the heart rate (beats per minute). The first arterial blood pressure measurement was made in 1733 (by Hale).

Continuous arterial blood pressure is now routinely measured in millions of patients per year. Blood pressure monitoring provides a very useful early warning of shock i.e. inadequate blood flow and tissue perfusion. Diagnosing shock and low blood pressure status in surgery allows earlier intervention and resuscitation. Deciding on the correct intervention requires knowledge of the determinants of the low blood pressure.

Is the low blood pressure a consequence of heart failure, low blood volume or a change in vascular resistance to blood flow? Concurrently measuring the cardiac output and the arterial blood pressure allows the judgment of a more appropriately choice of corrective intervention. Suitable interventions could include administering fluids – to fill up the vascular volume, giving drugs to support the heart’s strength of contraction or administration of vasoactive drugs to reduce or increase arterial resistance.

The ease of measuring arterial pressure has meant that the clinical diagnosis and management of the adequacy of the circulation has until this day been very pressure centric.

Arterial pressure has been used as a surrogate measure of cardiac output. Ideally, diagnostic and clinical decisions concerning the circulation would be based on a full understanding of the drivers of clinically relevant hypo- and hypertensive events observed while monitoring arterial pressure.

The goal has been to simultaneously measure cardiac output, blood pressure vascular and systemic resistance.

Early Methods of Deriving Cardiac Output From The Arterial Blood Pressure

Windkessel and Pulse Contour Approaches

Not surprisingly the search has been on for a simple method that could provide continuous cardiac output measurements in patients during surgery and in shock. Adolf Fick made the first accurate measurement of cardiac output in 1870. Unfortunately, measuring the cardiac output proved to be not as easy as measuring arterial blood pressure. The Fick method proved to be cumbersome and not very practical for most clinical situations.

Otto Frank (1899) was the first person to develop a mathematical model for describing the circulation and the relationship between blood pressure and blood flow. He formulated the concept of the two-element Windkessel model (air chamber/elastic reservoir), which described the shape of the arterial pressure waveform in terms of resistance and compliance elements.

His model draws similarities between the heart and the arterial system to a closed hydraulic circuit comprised of a water pump connected to a mixed air and water filled chamber. As water is pumped into the chamber, the water compresses the air in the pocket and simultaneously pushes water out of the chamber, back to the pump. The compression of air simulates the elasticity and extensibility (arterial compliance) of the aorta, as blood is pumped into it from the heart.

The resistance that the water overcomes to leave the Windkessel, simulates the resistance to flow encountered by the blood as it flows through the major and minor arteries/ arterioles to the capillaries. Functionally, the Windkessel/ elastic reservoir effect of the major arteries dampens the fluctuation in blood pressure that would otherwise occur if blood was pumped into a stiff and less compliant arterial system.

During diastole the stored blood in the aorta and its associated driving pressure, assists in the maintenance of a continuous blood flow away from the heart to the periphery, thereby ensuring consistent end organ perfusion pressure and oxygen transport. This circulation model was improved by Broemser (1930) who proposed a modified 3-element model, which added a second resistive element to the 2-element Windkessel model between the pump and the air chamber, thereby simulating the resistance to blood flow encountered across the aortic valve.

Frank’s circulatory model and its subsequent refinement led to a number of attempts to derive the stroke volume from more complex Windkessel-like models of the circulation (Kouchoukos, 1970 & Wesseling 1993). These techniques became collectively known as “Pulse Contour” methods of deriving stroke volume and cardiac output from the arterial pressure.


Broemser Ph, et. al. Ueber die Messung des Schlagvolumens des Herzens auf unblutigem Weg. Zeitung für Biologie. 1930;90:467-507.
Burattini R, Di Salvia PO. Development of systemic arterial mechanical properties from infancy to adulthood interpreted by four-element Windkessel models. J Appl Physiol. 2007;103:66–79.
Erlanger J, Hooker DR. An experimental study of blood-pressure and of pulse-pressure in man. Johns Hopkins Hospital Reports 1904;12:145-378.
Fick A. Ueber die Messung des Blutquantums in den Herzventrikeln. Sitzungsberichteder Physiologisch-Medizinosche Gesellschaft zu Wurzburg 1870; 2:16.
Frank O. Die Grundform des arteriellen Pulses. Z Biol 1899;37:483–526.
Stephen Hales Statical Essays: Haemastaticks, 1733.
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(6):425-435.
Langewouters GJ, Wesseling KH, Goedhard WJA. The pressure dependent dynamic elasticity of 35 thoracic and 16 abdominal human aortas in vitro described by a five component model. J Biomech. 1985;18:613–620.
Remington JW, Noback CR, Hamilton WF, Gold JJ. Volume elasticity characteristics of the human aorta and prediction of the stroke volume from the pressure pulse. Am J Physiology. 1948;153:298-308.
Wesseling KH, Jansen JR, Settels JJ et al. Computation of aortic flow from pressure in humans using a nonlinear, three element model. J Appl Physiol 1993;74:2566–2573.

PulseCO Method

The PulseCO™ algorithm method is based on the principles of conservation of mass and power. Stroke volume is calculated from an analysis of the stroke volume induced pulsatile change in the pressure waveform. It is important to note that this is not a Pulse Contour or modified Windkessel approach, but rather a different and novel non-morphological approach.

The PulseCO™ method overcomes the limitations of the more historic approaches that were reported to be compromised by the variable contribution of the arterial reflected wave and resistance changes that can occur in peripheral arteries.

Briefly, the PulseCO™ analysis transforms the arterial waveform from pressure to a volume equivalent through a compliance and aortic volume correction maneuver. Autocorrelation of the volume waveform derives the beat period (heart rate) and input pulsatile volume change i.e. stroke volume. Cardiac output is derived by multiplying the stroke volume by the heart rate.

The LiDCOplus and the LiDCOrapid monitors both use identical implementations of the PulseCO™ algorithm. The algorithm derives the pre-calibration cardiac output i.e. CO Algorithm (COa) in exactly the same way in both products.

The COa is then calibrated i.e. made more accurate by scaling with a calibration factor (CF). This is achieved by entering the CO known (COk) into the monitor. The CF is calculated in both products in exactly the same way CF = COk/COa.

The LiDCOrapid can also derive the CF via a nomogram, which calculates the CF from the patient’s age, height and weight. In this case, the displayed cardiac output is then equal to the CF x COa.


Problems addressed that limited the precision of morphology (pulse pressure/systolic area) approaches
• Morphology-based approaches (Windkessel & Pulse Contour) may have difficulties finding the systolic area;
• Reflected pressure waves – can move into the systolic part of the waveform with vasoconstriction & further away with vasodilation increasing or decreasing the area of pressure waveform analysed;
• Frequent recalibration – after a significant change in hemodynamics is required
Strengths of the PulseCO™ autocorrelation approach;
• The power and energy components are conserved within the beat data despite vasoconstriction or dilation, so changes in waveform morphology & shape do not affect the autocorrelation measurement of stroke volume;
• Arterial line damping/ changes in frequency response do not affect the measurement;
• Frequent recalibration – after a significant change in hemodynamics – is not required.

PulseCO Comparative Studies – Accuracy/Precision

What is the precision/trending ability of the core PulseCO™ pulse power/autocorrelation algorithm to follow changes in stroke volume & cardiac output?

The software code of the core LiDCO pressure waveform algorithm (PulseCO), as used in the original PulseCO™, LiDCOplus and now LiDCOrapid Monitors, has remained completely unchanged since the launch of the first PulseCO™ monitor in 2001. This means that all the 100 or so papers and abstracts published on the performance of the core PulseCO™ software are comparable and still relevant.

Acceptable limits of precision have been defined by Critchley & Critchley (1999), who stated that if a new (cardiac output) method is to replace an older, established method, the new method should have errors not greater than the older reference method. Over the last 10 years the precision of LiDCO’s core PulseCO™ algorithm to trend changes in stroke volume has been evaluated in a wide number of challenging clinical situations – these include: general surgical patients (Heller et al. 2002), high cardiac outputs (Hallowell & Corley et al. 2005), hyperdynamic liver transplantation patients (Costa et al. 2007), off-pump (Missant and Wouters 2007) and on-pump cardiac surgery (Wilde et al. 2007; ), post pediatric heart transplantation (Kim et al. 2006), post-operative care (Pitman et al. 2005, Hamilton, Huber and Jessen, 2002); pre-eclampsia (Dyer et al. 2011), congestive heart failure (Kemps et al. 2008, Mora et al. 2011) and general intensive care (Mills et al. 2010, Brass et al. 2011, Cecconi et al. 2010, Smith et al. 2005).
As can be seen (table below) the 95% confidence limits (Bland Altman statistics) reported in these studies range between 17% – 30% in adults and are within the original precision limits proposed by Critchley and Critchley (1999). Therefore, these results obtained from a variety of clinical situations demonstrate that the PulseCO™ algorithm has sufficient precision to follow cardiac output changes without recalibration and is equivalent in precision to bolus thermodilution.

The data shown above in the table are derived from the statistical analysis of repeated paired measures at differing cardiac outputs from multiple patients. The statistics used assumes each pair of measurements are fully independent of each other (conventional method). In fact, they are not fully independent as they are actually multiple measures taken within a single patient.

Thus the conventional method of analysis (Bland Altman/Critchley) is by far the most commonly used way of comparing the accuracy of two methods of measurement of cardiac output. Another, and perhaps more clinically relevant way of looking at the data is to examine the way that changes in cardiac output from each of the sequential PulseCO™ measurements compares to the changes in sequential dilution cardiac output control measurements.

This is known as the consecutive change method of analysis. This form of analysis was used in the paper published by Wilde et al. 2007. Some of their results are shown below. The advantage of this consecutive change analysis method is that it can be used to quantify the concordance (another way of looking at the trending ability) between the two techniques. It can be seen that 88% of the relative change comparison points are seen to be falling in the upper right and lower left quadrants of the plot.

This analysis allowed Wilde et al. to state that 88% of all changes in cardiac output of greater than 0.5l/min were similarly detected by the two methods i.e. PulseCO and the control thermodilution measurements.

PulseCO References

Comparison against Electromagnetic flow probe & Fick
1. Marquez J, McCurry K, Severyn D, Pinsky M. Ability of Pulse Power, Esophageal Doppler and Arterial Pulse Pressure to Estimate Rapid Changes in Stroke Volume in Humans. Crit Care Med. 2008;36(11):3001 – 3007.
2. Kemps H, Thijssen E, Schep G, Sleutjes B,De Vries W, Hoogeveen A, Wijn P, Doevendans P. Evaluation of two methods for continuous cardiac output assessment during exercise in chronic heart failure patients. J Appl Physiol. 2008;105:1822-1829.
Comparison against pulmonary artery thermodilution
3. Hamilton TT, Huber LM, Jessen ME. PulseCO: A Less-Invasive Method to Monitor Cardiac Output From Arterial Pressure After Cardiac Surgery. Ann Thorac Surg. 2002;74:S1408-12
4. Pittman J, Bar Yosef S, SumPing J, Sherwood M, Mark J. Continuous cardiac output monitoring with pulse contour analysis: A comparison with lithium indicator dilution cardiac output measurement. Crit Care Med. 2005;33(9):2015-2021.
5. Missant C, Rex S, Wouters P. Accuracy of cardiac output measurements with pulse contour analysis (PulseCO) and Doppler echocardiography during off-pump coronary artery bypass grafting. European Journal of Anaesthesiology. 2008;25(3):243-248
6. Costa MG, Della Rocca G, Chiarandini P, Mattelig S, Pompei L, Barriga MS, Reynolds T, Cecconi M, Pietropaoli P. Continuous and intermittent cardiac output measurements in hyperdynamic conditions: pulmonary artery catheter versus lithium dilution technique. Intensive Care Med. 2007. DOI 10.1007/s00134-007-0878-6.
7. Wyffels P, Sergeant P, Wouters P. The value of pulse pressure and stroke volume variation as predictors of fluid responsiveness during open chest surgery. Anaesthesia. 2010;65:704:709j.
8. Dyer R, Piercy J, Reed A, Strathie G, Lombard C, Anthony J, James M. Comparison between pulse waveform analysis and thermodilution cardiac output determination in patients with severe pre-eclampsia. Br J Anaest. 2011;106(1)77–81.
9. De Wilde RBP, Schreuder JJ, van den Berg PCM, Jansen JRC. An evaluation of cardiac output by five arterial pulse contour techniques during cardiac surgery. Anaesthesia. 2007;62:760-768
10. Kirwan C, Smith J, Lei K, Beale R. A comparison of two calibrated continuous arterial pressure waveform based measurements of cardiac output over 24 hour. Crit Care Med. 2005;33(12)Suppl:208-S.A56.

PulseCO & Fluid Management

Assessing Fluid Responsiveness

Prompt treatment of hypovolemia is necessary to sustain blood pressure, blood flow & tissue perfusion. Excessive volume administration may result in edema formation and impaired tissue perfusion, with consequent organ dysfunction and increased risk of morbidity and even death.

Un-monitored attempts at central blood volume expansion can be dangerous.
Because blood volume cannot be reliably assessed clinically, the likelihood of hemodynamic response (preload responsiveness) and the actual magnitude of the response to a volume challenge (cardiac output/ stroke volume response) have replaced the use of static volume parameters such as; mean arterial pressure, central venous pressure (CVP) or pulmonary capillary wedge pressure (PCWP).
Arterial preload responsiveness parameters have been calculated manually from arterial blood pressure displayed monitors since at least the late 1960’s. Automated display of; systolic pressure variation (SPV), pulse pressure variation (PPV%), stroke volume variation (SVV%) and stroke volume response (∆SV) are now increasingly used to guide fluid management in the arterial line patient and their protocolized use has been shown to lead to improvements in outcome. (Lopes et al. 2007).

The preload response parameters reported by the LiDCOrapid are all mathematically derived from the PulseCO™ algorithm’s primary measured parameters. They have been independently validated to be appropriately sensitive, specific and precise. When used as diagnostic aids in a fluid management protocol they have been associated with improved outcome in high-risk surgery and transplantation applications.

PulseCO SVV% & PPV% Method

Studies validating the precision, sensitivity, and utility of LiDCO’s pulse power algorithm (PulseCO™) derived fluid responsiveness and stroke volume response parameters in different patient populations.

Fluid Responsiveness

Cardiac surgery: Belloni et al. 2007 showed that PPV% and SVV% before fluid challenge were significantly higher in the fluid responder group than in the non-responder group. Patients with PPV% and SVV% of 12% were responsive to fluid challenges, whereas patients who did not increase their cardiac index in response to fluid challenge had PPV% and SVV% < 10%. A significant correlation among baseline PPV% & SVV% and change in cardiac index was also shown; patients with a high value of PPV% & SVV% at baseline responded to volume loading with a larger increase in cardiac index.

Therefore, LiDCO’s baseline PPV% & SVV% provided useful information about the patients’ position on an individual Starling curve. The greater the PPV% & SVV%, the more likely the patient was to be responsive to ventricular fluid loading. Finally, in the fluid responder group, the mean PPV% & SVV% were significantly lower after the fluid challenge compared with baseline values, indicating a change in their volume status after fluid administration.

The authors concluded:

“PCO-LiDCO technology used in this study provided reliable measurements of PPV%, SVV%, and SPV. A further advantage of this technique was that it only required peripheral intravenous and arterial access, both of which are commonly placed for perioperative care.”

Wyffels et al. 2010 also examined the performance of the LiDCO monitor PPV% and SVV% in a cardiac surgery population. They investigated the ability of pulse pressure variation and stroke volume variation to predict fluid responsiveness during mechanical ventilation in 15 patients undergoing open chest surgery by comparing their respective correlations with cardiac output changes induced by a passive leg raise. Under closed chest conditions, both pulse pressure variation and stroke volume variation correlated well with the induced cardiac output changes (r = 0.856, P = 0.002 and r = 0.897, P = 0.0012, respectively). Their data show that LiDCO’s pulse pressure variation and stroke volume variation are valid predictors of fluid responsiveness under closed chest conditions. In contrast, the static parameters of CVP or wedge pressure did not predict fluid response.

Bariatric surgery: Avery et al. 2010 assessed the LiDCO preload responsiveness parameters in bariatric surgery patients. ROC curve analysis gave predicted cut-off limits of 9.5% for SVV% (AUC = 0.900), 15.5% for PPV% (AUC=0.912). There was no difference between SVV% and PPV% (p=0.69). Sensitivity and specificity for SVV% (100 & 75) and PPV% (100 & 85) were similar and acceptable, whereas that of HR (87 & 35) and MAP (75 & 60) were less reliable.

Their conclusion was:

“PulseCO™ derived Stroke Volume Variation predicts fluid responsiveness in the morbidly obese with consistent sensitivity and specificity. This compares favorably with studies in non-obese patients.”

Intensive care: Cecconi et al. 2009 assessed the LiDCO/PulseCO™ preload responsiveness parameters in post-operative intensive care patients on IPPV with 8ml/kg tidal volumes and PEEP of 5cm H20. Only PPV% and SVV% were able to give cutoff values with sensitivity and specificity >50%. ROC curve analysis gave predicted cut-off limits of 11% for SVV% (Sens: 72%, Spec: 79%) and for PPV% >12% (Sens: 86%, Spec: 70%).

Their conclusion was:

“SVV, PPV and SPV of PulseCO are good predictors of Fluid Responsiveness in fully sedated and mechanically ventilated patients in the intensive care.”

Effects of vasoactive drugs on PPV%/ SVV%. A key question for pressure waveform analysis methods has been their robustness in the presence of vasoactive administration – Hadian et al. 2010 investigated the effects of vasoactive therapy on LIDCO’s PPV% and SVV%. Seventy-one paired events were studied – 38 vasodilators, 10 vasoconstrictor & 14 inotrope administrations. As expected, vasodilator therapy increased PPV% and SVV% from 13% to 17% and 9 to 15% respectively (P < 0.001). Whereas, increasing inotropes or vasoconstriction did not alter PPV% and SVV%. The authors’ conclusion was “ Thus, SVV% and PPV% can be used to drive fluid resuscitation algorithms in the setting of changing vasoactive drug therapy”.

High-risk general surgery: Richard et al. 2010 compared measurements of arterial pulse pressure variation (PPV) from the LiDCOrapid to the same parameter displayed by the Philips Intellivue MP50 monitors on patients during high-risk abdominal surgery.

Their conclusions were that the:

“LiDCOrapid PPV measurement demonstrated a high degree of correlation with the Philips Intellivue MP50 PPV values. Use of either monitor should generate accurate measurements of continuous PPV to guide fluid resuscitation.”

Stroke volume response

No fluid responsiveness parameter is 100% accurate in predicting the magnitude of the fluid response. The increase in stroke volume needs to be> 10% for a 200 ml infusion, or the hemodilution effects will override the flow increase resulting in a net decrease in oxygen delivery. Therefore a very necessary safety check is for the clinician to monitor the fluid response itself (% change in blood flow/stroke volume) to the fluid administration. Trended stroke volume response (∆SV) should be displayed in order that the user can reassure them that the response is proportionate to the volume of fluid infused.

The LiDCOrapid monitor shows both trended changes in SVV% or PPV% and ∆SV in response to fluid challenges to increase stroke volume – see below for an example:


The absolute precision of the PulseCO™ algorithm for discerning the small stroke volume changes necessary to detect small changes in SVV% and ∆SV was investigated at Pittsburgh University by Marquez et al. 2008. The PulseCO™ derived stroke volume changes following an inferior vena cava occlusion were compared to flow changes measured with an aortic electromagnetic flow probe.

These investigators showed elegantly that the PulseCO™ algorithm can precisely follow small acute/fast changes in stroke volume. Another US-based group at Columbia University, NY (Dizon et al. 2010) showed that the core PulseCO™ algorithm could discriminate < 5% changes of stroke volume in heart failure patients undergoing biventricular pacemaker resynchronisation.

They further showed that patients with a > 5% improvement of stroke volume performed better at 2-month follow-up clinics in terms of length of 6 min walk, and improvement in their echocardiographic dyssynchrony profile.


1. Belloni L, Pisano A, Natale A, Piccirillo M, Piazza L, Ismeno G, De Martino G. Assessment of Fluid-Responsiveness Parameters for Off-Pump Coronary Artery Bypass Surgery: A Comparison Among LiDCO, Transesophageal Echocardiography, and Pulmonartery Catheter. J Cardiothorac Vasc Anesth. 2008;22(2):243-8
2. Murugan R, Venkataraman R, Wahed A, Elder M, Carter M, Madden N, Kellum J. Preload responsiveness is associated with increased interleukin-6 and lower organ yield from brain-dead donors. Crit Care Med. 2009;37(8):2387-2393.
3. Jain A, Dutta A. Stroke Volume Variation as a Guide to Fluid Administration in Morbidly Obese Patients Undergoing Laparoscopic Bariatric Surgery. Obes Surg. 2010. DOI 10.1007/s11695-009-0070
4. Abdel-Galil K, Craske D, McCaul J. Optimisation of intraoperative haemodynamics: early experience of its use in major head and neck surgery. Br J Oral and Maxillofac Surg. 2010;48(3):189-191.
5. Brass P, Mills E, Latza J, Peters J, Berendes E. LiDCOrapid and PiCCOplus preload response parameter validation study. Proceedings of 31st International Symposium on Intensive Care and Emergency Medicine. 2011;(Suppl 1):P62 doi:10.1186/cc9481
6. Hadian M, Severyn D, Pinsky M. The effects of vasoactive drugs on pulse pressure and stroke volume variation in postoperative ventilated patients. J Critical Care. 2010;(26)3:328.e1-8.
7. Wyffels P, Sergeant P, Wouters P. The value of pulse pressure and stroke volume variation as predictors of fluid responsiveness during open chest surgery. Anaesthesia. 2010;65:704-709
8. Marquez J, McCurry K, Severyn D, Pinsky M. Ability of Pulse Power, Esophageal Doppler and Arterial Pulse Pressure to Estimate Rapid Changes in Stroke Volume in Humans. Crit Care Med. 2008;36(11)3001-3007.
9. Dizon J, Quinn TA, Cabreriza S, Wang D, Spotnitz H, Hickey K, Garan H. Real-time Stroke Volume Measurements for the Optimization of Biventricular Pacing Parameters. Europace. 2010;12(9):1270-4.
10. Richard K, Novak M, Quill T, Cannesson M, Koff M. Functional Hemodynamics during High-Risk Abdominal Surgery: Are All Monitors Created Equal? Presentation A997 at the ASA, 2010 San Diego
11. Avery S, Mills E. Jonas M, Margarson M. PulseCO derived stroke volume variation for prediction of fluid responsiveness in the morbidly obese. Presentation A996 at the ASA, 2010 San Diego