First, the maths

  • A bolus of 5-10ml cold 5% dextrose into the right atrium should decrease the temperature in the pulmonary artery.
  • The rate of blood flow is  inversely proportional to the change in temperature over time
  • Thus, the mean decrease in temperature is inversely proportional to the cardiac output.
  • The Stewart-Hamilton Equation describes this relationship
Thermodilution measurements of cardiac output with the Stewart-Hamilton formula

These variables are derived from the measured thermodilution curves.

These curves are visibly different in different cardiovascular pathological states:

thermodilution curves for measuring cardiac output

The higher the cardiac output, the faster he blood flow and the shorter and steeper the thermodilution curve. In low cardiac output, the curve is slurred and lazy. Even more so in tricuspid regurgitation.

Some catheter manufacturers have a heating filament near the tip, which means you don't have to bother with intermittent bolus injections; the catheter just does this automatically every 3 or so minutes, and you have a near- continuous cardiac output measurement.

Limitations of thermodilution as a method for measuring cardiac output

This shows good agreement with the Fick method and with the indocyanine green dye dilution method.
However, there is a lot of variability.

    • You should take measurements in expiration.
    • You have to take a mean of 3 measurements.
    • The mean has to be 15% different to the previous mean, otherwise it is within the margin of error.
    • The thermodilution cardiac output can vary by 10% from measurement to measurement without any change in the condition of the patient

Too much injected cold stuff causes underestimation of cardiac output.
Too little injected cold stuff causes overestimation of cardiac output.
Room temperature injectate  produces less accurate readings, but is safer.
Very cold injectate (0-4 degrees) is more accurate, but can induce bradycardia and decreased cardiac output.

Data from this measured concentration curve can then be used to plug into the Stewart-Hamilton equation. George Neil Stewart first described it in 1897, after some animal experiments. His subjects were dogs, and his indicator was sodium chloride.

What are the causes of inaccuracy in thermodilution cardiac output measurement?

  • Catheter is in the wrong position
  • The thermistor tip is up against the wall
  • The respiration is erratic
  • There is an intracardiac shunt
  • Tricuspid regurgitation
  • Cardiac arrhythmia
  • Rapid infusion happening via the IJ line
  • Abnormal hematocrit
  • Slow injectate delivery
  • Injectate not cold enough, or not enough of it


Anna Gawlinski, Measuring Cardiac Output: Intermittent Bolus Thermodilution Method Crit Care Nurse 2004;24:74-78

Toshiaki Nishikawa, Shuji Dohi Errors in the measurement of cardiac output by thermodilution Canadian Journal of Anaesthesia February 1993, Volume 40, Issue 2, pp 142-153