How to Measure Cardiac Index: Invasive & Non-Invasive Methods

Cardiac index is measured by first obtaining cardiac output and then dividing it by body surface area, and the methods available range from threading a catheter into the pulmonary artery to placing an ultrasound probe on the chest wall. No single technique dominates every clinical scenario. The choice depends on how sick the patient is, whether they are in surgery or on a general ward, and how much accuracy the clinical team needs versus how much risk and invasiveness they can accept.

What Cardiac Index Actually Tells You

Cardiac output is the volume of blood your heart pumps per minute, typically somewhere around 4 to 8 liters in a resting adult. Cardiac index adjusts that number for body size by dividing cardiac output by body surface area. A normal resting cardiac index falls roughly between 2.5 and 4.0 liters per minute per square meter. The reason clinicians care about the indexed value rather than raw output is straightforward: a cardiac output of 4 L/min might be perfectly adequate for a small person and dangerously low for someone much larger. Indexing puts everyone on a comparable scale, at least in theory.

That “in theory” matters. As we’ll see later, the body surface area formulas used for indexing can distort the picture at the extremes of weight, sometimes enough to trigger the wrong treatment decisions.

The Invasive Gold Standard: Pulmonary Artery Catheterization

The pulmonary artery catheter, sometimes still called a Swan-Ganz catheter after its inventors, has been the reference method for cardiac output measurement since the 1970s. A balloon-tipped catheter is floated through a central vein, through the right heart chambers, and into the pulmonary artery. Cardiac output is then measured using thermodilution: a known volume of cold saline is injected into the right atrium, and a thermistor at the catheter tip downstream records the resulting temperature change. A smaller, faster temperature drop means higher blood flow, and vice versa. Dividing the result by body surface area gives you the cardiac index.

Modern pulmonary artery catheters go beyond the classic bolus-injection approach. Contemporary devices use a thermal filament that intermittently heats the blood, allowing near-continuous cardiac output measurement rather than relying on snapshot bolus injections. These catheters can also report right ventricular ejection fraction, end-diastolic volume, and mixed venous oxygen saturation, offering a broader hemodynamic profile from a single device.¹

The catch is risk. A prospective study of 116 catheter insertions found arrhythmias during 90 of them, including premature beats, ventricular tachycardia, and transient right bundle branch block. Although those rhythm disturbances did not lead to lasting harm, the same cohort saw staphylococcal bloodstream infections likely originating from the catheter in about 1.7% of cases and subclavian vein thrombosis in another 1.7%. A postmortem examination in one case revealed perforation of the pulmonic valve.² The procedure demands skilled operators, carries infection and vascular injury risks, and ties the patient to an intensive care setting.

The Fick Principle

Before thermodilution became widespread, the Fick principle was the original gold standard and still serves as the reference method in certain research settings, particularly in pediatric patients. It relies on the relationship between oxygen consumption, the oxygen content of arterial blood, and the oxygen content of venous blood. Measure those three values and you can calculate how much blood the heart must be pumping to deliver that oxygen.

In practice, Fick-based cardiac output requires measuring actual oxygen consumption, which is technically demanding. A study in children confirmed that combining oxygen consumption measurement with the Fick method served as the clinical gold standard against which transpulmonary thermodilution was validated.³ Because it requires pulmonary artery blood sampling and careful gas exchange measurements, the Fick method is even more cumbersome than thermodilution in routine clinical practice. Its main role today is as a calibration benchmark.

Transpulmonary Thermodilution and Pulse Contour Analysis

Transpulmonary thermodilution sits in a middle zone: less invasive than a full pulmonary artery catheter, but still requiring arterial and central venous lines. A cold bolus is injected through a central venous catheter, and the temperature change is detected at the tip of an arterial catheter, typically in the femoral or radial artery, rather than in the pulmonary artery. This approach intermittently measures cardiac output and, crucially, calibrates a pulse contour algorithm that can then provide continuous, real-time tracking between bolus injections.⁴

The continuous tracking part is where things get interesting and where accuracy can slip. Pulse contour analysis estimates cardiac output beat by beat from the shape of the arterial pressure waveform. Some systems require calibration by thermodilution (like the PiCCO device), while others attempt to work without calibration (like the FloTrac/Vigileo system). One head-to-head comparison found that the uncalibrated Vigileo system underestimated cardiac output by a mean of about 0.5 L/min compared with transpulmonary thermodilution, with wide scatter.⁵ A broader evaluation of nine different pulse contour algorithms in cardiac surgery patients found that all of them showed poor accuracy compared with transpulmonary thermodilution.⁶

The clinical takeaway is that calibrated pulse contour systems tend to perform better than uncalibrated ones, but any pulse contour device can drift from reality over time, especially when vascular tone changes rapidly, as happens during sepsis, major surgery, or vasoactive drug adjustments. Periodic recalibration with a thermodilution bolus helps keep the numbers honest.

Transesophageal Echocardiography in the Operating Room

Transesophageal echocardiography places an ultrasound probe in the esophagus, right behind the heart, giving clinicians a close-up, real-time view of cardiac structures and blood flow. It does not break the sterile surgical field, and it provides information about valve function, ventricular volumes, and contractility alongside cardiac output estimates.⁷

Accuracy depends heavily on which flow measurement approach is used. An early validation study in cardiac surgery patients found that cardiac output derived from left ventricular cross-sectional area correlated poorly with thermodilution, with wide limits of agreement spanning nearly 9 L/min. However, when the investigators used pulmonary artery flow velocity measured by transesophageal Doppler instead, correlation jumped dramatically, with a mean difference of just 0.12 L/min and tight limits of agreement.⁸ This means the technique can be quite good when applied skillfully, but operator dependence is real and the measurement approach matters enormously.

Transesophageal echocardiography is considered semi-invasive because the patient must be sedated or under general anesthesia for probe insertion, and there is a small risk of esophageal injury. It is therefore reserved primarily for intraoperative use and selected critical care scenarios.

Transthoracic Echocardiography at the Bedside

For a fully non-invasive approach, transthoracic echocardiography is the most widely available option. The method uses a probe on the chest wall to measure the diameter and blood flow velocity in the left ventricular outflow tract, the channel between the left ventricle and the aorta. By measuring the velocity-time integral of flow through that tract and combining it with the outflow tract’s cross-sectional area, clinicians calculate stroke volume. Multiply by heart rate and you have cardiac output; divide by body surface area and you have cardiac index.⁹

Emergency physicians have been trained to perform these measurements with focused cardiac ultrasound exams.¹⁰ The beauty of the approach is that it requires no catheters, no sedation, and can be done at the bedside in minutes. The weakness is that it provides a snapshot rather than continuous monitoring, it depends on the operator’s ability to get a good acoustic window (difficult in obese patients, those on mechanical ventilation, or patients with chest bandages), and the outflow tract diameter measurement is squared in the calculation, so even small measurement errors get amplified.

A systematic review and meta-analysis comparing echocardiographic cardiac output to thermodilution across many studies found a median bias of just −0.12 L/min, which sounds reassuring. But the spread around that median was wide, with individual studies showing limits of agreement ranging from near-perfect to ±4.72 L/min. The median percentage error across 28 studies was about 24%, and the overall correlation coefficient was 0.83.¹¹ Translation: echocardiography is good enough for many clinical decisions, but not precise enough to replace invasive monitoring in the sickest patients where small differences in cardiac output drive treatment changes.

Bioimpedance and Bioreactance

These are the least invasive of all technologies: electrode patches placed on the skin send small electrical currents through the thorax, and the device estimates stroke volume from how the chest’s electrical properties change with each heartbeat. Impedance cardiography measures changes in the magnitude of electrical impedance, while newer bioreactance devices measure phase shifts in the signal instead, which their developers argue makes them more resistant to noise and motion artifact.¹²

A study testing a bioreactance device in heart failure patients found that the correlation between noninvasively measured cardiac index and directly measured cardiac index was 0.61, identical to the correlation between directly measured cardiac index and peak oxygen consumption.¹³ That is a moderate correlation, acceptable for trend monitoring but hardly a precision instrument.

The reality check came from a study during liver transplantation, one of the most hemodynamically volatile surgeries. Both a pulse-wave-transit-time device and a thoracic impedance device performed poorly against thermodilution, with percentage errors of 69% and 77% respectively.¹⁴ The message is consistent across the literature: bioimpedance and bioreactance devices are useful for non-invasive trending when precision is less critical, but they should not be relied upon for absolute values in unstable patients.

Other Non-Invasive Approaches

Several other techniques round out the non-invasive toolkit. The Ultrasonic Cardiac Output Monitor (USCOM) is a handheld continuous-wave Doppler device placed on the chest that measures flow through either the aortic or pulmonary valve. A study comparing USCOM with oesophageal Doppler during major abdominal surgery found good concordance and high sensitivity for detecting clinically meaningful changes in stroke volume, though the absolute percentage error was about 30%.¹⁵ In patients with single-ventricle heart defects, USCOM cardiac index measurements showed a strong correlation with catheterization-derived values, with a Pearson’s coefficient of 0.89.¹⁶

Finger cuff technologies, such as the ClearSight system, use a volume-clamp method to continuously record arterial pressure waveforms from a finger cuff, then estimate cardiac output by pulse contour analysis.¹⁷ In cardiac surgery patients, the ClearSight system showed a mean bias of about 0.32 L/min compared with pulmonary artery thermodilution under hemodynamically stable conditions, with a percentage error around 25%.¹⁸ That level of accuracy is borderline: the commonly cited threshold for clinical acceptability is a percentage error below 30%, and these devices hover right around that line.

Cardiac MRI using phase contrast imaging can quantify blood flow with high accuracy. In vitro validation showed correlations above 0.97 for both standard and fast cine phase contrast MRI, outperforming Doppler ultrasound.¹⁹ The limitation is obvious: you cannot wheel an MRI scanner to the bedside. Phase contrast MRI is excellent for research and elective diagnostic workups but plays no role in acute hemodynamic monitoring.²⁰

Why the Indexing Step Is Not as Simple as It Looks

Once you have a cardiac output number, dividing by body surface area should be straightforward. It isn’t, especially at the extremes of weight. Multiple body surface area formulas exist, and they can produce substantially different results for the same patient. A mathematical analysis found that among seven commonly used formulas, there was substantial variation in predicted body surface area at height-weight combinations typical of morbid obesity. That variation translates directly into different cardiac index values from the same cardiac output, and in extreme cases, the distortion could suggest inappropriate therapeutic interventions.²¹

The concern extends to pediatric patients as well. A study of children undergoing cardiac procedures cautioned that obesity status is a critical factor in determining body surface area values and that indexed hemodynamic interventions may be inappropriate if the effect of obesity on body surface area formulas is not considered.²² The practical implication is that clinicians treating very obese or very underweight patients need to look at raw cardiac output alongside the indexed number, rather than relying on cardiac index alone to guide treatment thresholds.

Mechanical Ventilation and Other Confounders

Regardless of which technique you use, the number you get is only as reliable as the conditions under which you measure it. Positive pressure ventilation, the standard approach in mechanically ventilated patients, increases pressure inside the chest. That pressure reduces the amount of blood returning to the right side of the heart, lowering right ventricular stroke volume. At the same time, it reduces the workload on the left ventricle by decreasing transmural pressure. These opposing effects mean cardiac output can swing with each breath, and the magnitude of that swing is itself used clinically: dynamic indices like stroke volume variation exploit these cardiopulmonary interactions to predict whether a patient will respond to intravenous fluids.

Other common confounders include arrhythmias (irregular rhythms make thermodilution boluses less reliable and throw off pulse contour algorithms), significant valve disease (regurgitant flow can confuse Doppler-based and thermodilution-based methods differently), and body temperature shifts (thermodilution accuracy depends on the temperature difference between the injectate and blood). None of these make measurement impossible, but they all mean the clinical team needs to interpret the numbers with context rather than taking them at face value.

Measuring Cardiac Index During Exercise

Outside the ICU, cardiac index measurements during exercise testing offer a window into how well the heart responds to increased demand. Cardiopulmonary exercise testing traditionally relies on gas exchange analysis to infer cardiac performance, but several noninvasive techniques can now be paired with exercise protocols to directly measure cardiac output. These include inert gas rebreathing, impedance cardiography, thoracic bioreactance, and transthoracic echocardiography.²³ Adding a direct cardiac output measurement during exercise helps distinguish whether exercise intolerance stems from the heart’s inability to increase output versus problems with oxygen extraction or ventilation.

The challenge with exercise measurements is motion. The patient is pedaling or walking on a treadmill, breathing hard, and sweating, all of which degrade signal quality for bioimpedance devices and make echocardiographic windows harder to obtain. Impedance cardiography during exercise has shown that the first heart sound amplitude increases routinely by four to five times normal, and in some subjects up to 15 times, which illustrates how dramatically the hemodynamic picture shifts during stress.²⁴ The method you choose for exercise cardiac output depends on what compromise between precision, practicality, and continuous versus intermittent readings works best for the clinical question.

Cardiac Index in Pregnancy

Pregnancy is one context where cardiac index monitoring has taken on growing importance, especially in women with pre-existing heart disease. Normal pregnancy increases cardiac output substantially, and the indexed values shift accordingly. A study comparing women with heart disease to healthy controls found that third-trimester cardiac index was similar between the two groups, roughly 4.0 versus 4.3 L/min/m². The revealing finding was in how much cardiac output had increased from baseline: women with heart disease showed a 51% increase in cardiac index over their pre-pregnancy reference, compared with 28% in controls. Strikingly, the odds of a maternal cardiovascular event were inversely associated with peak cardiac index, meaning women whose hearts could not ramp up their output adequately faced higher risk.²⁵

Echocardiography is the dominant measurement tool in pregnancy for obvious reasons: it is noninvasive and carries no radiation risk. Serial transthoracic measurements at different trimesters allow clinicians to track whether the expected rise in cardiac index is occurring. When it falls short, that can signal the need for closer monitoring, medication adjustments, or early delivery planning. This kind of longitudinal trending is where non-invasive methods genuinely shine, even if their single-measurement precision would not satisfy an ICU team managing cardiogenic shock.

The Decline of the Pulmonary Artery Catheter

For all its status as the reference standard, the pulmonary artery catheter has been in decline for decades. By the mid-1980s, it influenced management in over 40% of ICU patients in the United States. Then in 1996, a landmark observational study suggested the catheter might actually harm patients. Subsequent randomized trials failed to demonstrate a survival benefit from its use. The catheter never disappeared entirely, but its role shrank dramatically, pushed by a combination of safety concerns, negative trial results, and the proliferation of less invasive alternatives that could provide “good enough” hemodynamic information for most clinical scenarios.

Today, pulmonary artery catheterization is largely reserved for situations where precise right heart pressures and mixed venous oxygen saturation are needed and cannot be estimated any other way: severe pulmonary hypertension, complex cardiogenic shock, advanced heart failure evaluations, and certain cardiac surgeries. The broader story of cardiac output monitoring over the past 30 years has been a steady migration from one invasive gold standard toward a toolbox of less invasive methods, each with known trade-offs in accuracy, and a clinical culture that increasingly accepts those trade-offs in exchange for reduced risk.