The Swan-Ganz Catheter
What the pulmonary artery catheter measures, what it only calculates, how to read the waveforms as it floats through the heart, and the profiles that tell you which shock you are looking at.
When it is still indicated
The pulmonary artery catheter went from being in half the ICU to being in almost nobody, and the swing went too far in both directions.
Routine use in undifferentiated critical illness does not improve outcomes, and several large trials showed no benefit and a real complication rate. That finding was correct and it should have changed practice, and it did. What it did not establish is that the information is useless, only that measuring something does not treat it, and that a number in the hands of someone who does not know what to do with it changes nothing.
Where it still earns its place
- Cardiogenic shock, especially when you are deciding about mechanical support or a transplant referral. This is the strongest remaining indication, and use in shock has been rising again.
- Shock of unclear cause, where the echo is equivocal and the clinical picture does not fit. Filling pressures with a cardiac output will usually settle it.
- Severe right ventricular failure and pulmonary hypertension, where you need pulmonary vascular resistance rather than an estimate of it, and where fluid decisions are genuinely dangerous in both directions.
- Advanced heart failure evaluation: transplant workup, ventricular assist device candidacy, and titration of inotropes and vasodilators against real numbers.
- Post-cardiotomy management in patients who are not separating from bypass cleanly.
- Mechanical complications of infarction: a step-up in oxygen saturation between the right atrium and pulmonary artery diagnoses a ventricular septal defect, and giant v waves point at acute mitral regurgitation.
Put one in when you have a specific question that the catheter answers and the answer will change what you do. "The patient is sick and I want more data" is how the trials came out negative.
Reading the waveforms as it floats
The catheter tells you where its tip is at every moment, if you watch the trace. Learn these four and you will never be lost.
0 to 8 mm Hg
Low amplitude, gently undulating, with the a, c and v waves of a CVP trace. At roughly 15 to 20 cm from an internal jugular introducer. Inflate the balloon here, and only here.
15 to 30 / 0 to 8
The trace suddenly becomes tall and spiky, with a steep systolic upstroke and a diastolic pressure that falls back near zero. Note that diastolic pressure stays low: this is the feature that separates RV from PA. At roughly 30 to 35 cm. This is where ectopy happens, so move through it deliberately.
15 to 30 / 8 to 15
Systolic pressure is unchanged from the RV, but diastolic pressure steps up and a dicrotic notch appears from pulmonic valve closure. That diastolic step-up is the single most reliable sign you have crossed the valve. At roughly 40 to 45 cm.
6 to 12 mm Hg
The trace flattens abruptly into a low-amplitude venous-looking waveform with a and v waves. At roughly 45 to 55 cm. Deflate the balloon and the PA trace must return. If it does not, the catheter has migrated too far distally and needs to be pulled back.
Inflate to advance. Deflate to withdraw. Withdrawing with the balloon up can damage a valve or knot the catheter.
Never inflate with more than 1.5 mL of air, and stop the moment the waveform wedges even if you have used less. Needing less than 1.25 mL to wedge means the tip has migrated distally.
Never leave it wedged. Two respiratory cycles, take the reading, deflate. A persistently wedged balloon infarcts lung and is the mechanism behind pulmonary artery rupture, which is uncommon and frequently fatal.
Never inject fluid into the balloon port. Air only.
Normal values
| Measurement | Normal range | Notes |
|---|---|---|
| Right atrial pressure / CVP | 2 to 8 mm Hg | Right ventricular preload |
| Right ventricular pressure | 15 to 30 / 2 to 8 | Systolic / end-diastolic |
| Pulmonary artery pressure | 15 to 30 / 8 to 15 | Mean 10 to 20 |
| Pulmonary capillary wedge pressure | 6 to 12 mm Hg | Estimates left atrial pressure |
| Cardiac output | 4 to 8 L/min | Absolute, unindexed |
| Cardiac index | 2.5 to 4.0 L/min/m2 | The number that actually matters |
| Stroke volume index | 33 to 47 mL/m2 | |
| Systemic vascular resistance | 800 to 1200 dynes·s/cm5 | Low in distributive shock, high in cardiogenic |
| Pulmonary vascular resistance | less than 250 dynes·s/cm5 | Under 3 Wood units |
| Mixed venous saturation (SvO2) | 65 to 75 percent | Drawn from the PA port |
The wedge, and the situations where it lies
With the balloon inflated, the catheter occludes a branch pulmonary artery. Flow stops, and you are looking through a static column of blood at the pressure downstream: the pulmonary venous and therefore the left atrial pressure. In the absence of mitral disease that approximates left ventricular end-diastolic pressure, which is what you are actually after.
That chain has several links, and each one can break.
The wedge is only valid if the tip sits in West zone 3, where pulmonary venous pressure exceeds alveolar pressure and there is a continuous column of blood. In zones 1 and 2 you are measuring alveolar pressure instead. High PEEP and hypovolemia both push lung out of zone 3. Practically, the tip should sit below the level of the left atrium, which on a lateral film means the lower half of the chest, and gravity does most of the work for you.
Transmits to the wedge. Roughly a quarter to a half of the applied PEEP shows up in the reading in normal lungs, less in stiff ones. Do not disconnect a patient from PEEP to take a wedge, which derecruits them for a number.
In mitral stenosis the wedge substantially overestimates left ventricular end-diastolic pressure, because of the gradient across the valve. In acute mitral regurgitation, giant v waves inflate the mean and the wedge trace can start to resemble a PA trace, which is a classic source of confusion.
In diastolic dysfunction a normal wedge does not exclude a high end-diastolic pressure, and the relationship between pressure and volume is not the one you assume.
Read at end-expiration, off the printed waveform, and take the a wave. The digital mean averages in the respiratory swing and misleads.
In the absence of pulmonary vascular disease, the pulmonary artery diastolic pressure sits within a few mm Hg of the wedge. If the two track each other on a given patient, you can follow the PA diastolic and stop wedging altogether, which removes the main mechanical risk of the catheter. When they diverge, pulmonary vascular disease has developed and the shortcut no longer applies.
Cardiac output: two ways, and they disagree for reasons
Inject 10 mL of cold saline into the proximal port and measure the temperature change at the distal thermistor. The area under the resulting curve is inversely proportional to flow.
Take three measurements, discard outliers, average the rest. Inject at the same point in the respiratory cycle each time, usually end-expiration. It overestimates output in tricuspid regurgitation, in intracardiac shunts, and in very low output states where the curve is long and flat. Slow injection and warm injectate both produce falsely high values.
Cardiac output = oxygen consumption divided by the arteriovenous oxygen difference. Almost nobody measures oxygen consumption directly, so an assumed value of 125 mL/min/m2 is used, which is where the error creeps in.
More reliable than thermodilution in low output states and in tricuspid regurgitation. Less reliable in patients whose actual oxygen consumption is far from the assumption, which includes anyone febrile, shivering, agitated or septic.
Mixed venous saturation is worth as much as the output itself. An SvO2 below 65 percent means oxygen delivery is not meeting demand, and the causes are the four you can fix: low output, low hemoglobin, low arterial saturation, or high consumption. An SvO2 that is high, above 80 percent, in a shocked patient is not reassuring at all. It means the tissues are not extracting, which is the signature of distributive shock, cyanide toxicity, or a left-to-right shunt.
The calculated numbers
Everything below is derived, not measured. Every error in the inputs propagates.
It is the product of pressure and flow, which is the actual work the heart is doing. Normal is above 1.0 W. In cardiogenic shock, cardiac power output under 0.6 W is among the strongest single predictors of in-hospital mortality in the literature, and it outperforms cardiac index for that purpose. If you have a catheter in a shocked patient, calculate it.
Two more that matter in right heart failure: the pulmonary artery pulsatility index, which is (PA systolic minus PA diastolic) divided by right atrial pressure, where a low value predicts right ventricular failure after a left-sided assist device is placed. And the RA to PCWP ratio, where a value above 0.63 suggests right ventricular failure is the dominant problem rather than a bystander.
Hemodynamic profiles
This is what the catheter is for. Two axes, congestion and perfusion, and four quadrants.
| Profile | CVP | PCWP | CI | SVR | Picture |
|---|---|---|---|---|---|
| Hypovolemic | Low | Low | Low | High | Bleeding, dehydration, third spacing. Cold and dry. |
| Cardiogenic | High | High | Low | High | The classic cold and wet. Pump failure with compensatory vasoconstriction. |
| Distributive | Low or normal | Low or normal | High | Low | Sepsis, anaphylaxis, neurogenic, liver failure. Warm and well perfused until it is not. |
| Obstructive | High | Low or normal | Low | High | Massive PE, tamponade, tension pneumothorax. The CVP and PCWP diverge, which is the tell. |
| Isolated RV failure | High | Low or normal | Low | High | RV infarct, acute pulmonary hypertension. High RA to PCWP ratio. |
| Tamponade | High | High | Low | High | Equalization of diastolic pressures across all four chambers is the signature. |
Divergence of CVP and PCWP narrows the differential fast. Both high is left heart failure or tamponade. CVP high with a normal wedge is right-sided: pulmonary embolism, RV infarct, pulmonary hypertension. Both low is hypovolemia or distributive shock, and the SVR separates those two.
Mixed shock is common and the catheter finds it. The septic patient with a cardiomyopathy has a low SVR and a low cardiac index at once, and treating only one of those makes them worse. A single-axis mental model misses this; the numbers do not.
The older Forrester classification maps the same idea onto infarction, using a wedge above 18 for congestion and a cardiac index below 2.2 for hypoperfusion. Class I is neither, class II is congested but perfusing, class III is hypoperfused but dry, class IV is both, and class IV is the one with the mortality.
Complications
- Arrhythmia during insertion is very common and usually transient ventricular ectopy as the tip crosses the right ventricle. Sustained arrhythmia is uncommon. Right bundle branch block can be induced, which matters enormously if the patient already has a left bundle branch block, because that combination is complete heart block. Have pacing available in that patient.
- Pulmonary artery rupture is rare and often fatal. The risk factors are pulmonary hypertension, age, anticoagulation, distal migration and overinflation of the balloon. Hemoptysis in a patient with a Swan is a pulmonary artery rupture until proven otherwise, and it is an emergency: lie them with the affected side down, secure the airway, stop the anticoagulation, and call for interventional radiology and cardiothoracic surgery immediately.
- Pulmonary infarction from a persistently wedged or distally migrated catheter.
- Knotting, usually from repeated advancement without progress in a dilated right heart. If the catheter will not advance and the waveform is not changing, stop and pull back rather than pushing.
- Infection and thrombosis, which rise with duration. These come out as soon as they are not answering a question.
- Valve damage from withdrawal with the balloon inflated.
- Balloon rupture, suspected when no resistance is felt on inflation and no wedge appears. Stop inflating and do not attempt again. Air embolism is the concern, which is greater in a right-to-left shunt.
The daily checks
Every day on a patient with a PA catheter
Waveform. Is it still a PA trace with the balloon down? A spontaneously wedged trace means distal migration and the catheter needs pulling back now.
Depth. Note the centimeter mark at the introducer and compare it to yesterday.
Film. The tip should be within a few centimeters of the hilum, not out in the periphery.
Balloon. Confirm it is deflated and the syringe is locked open.
Transducer. Level to the phlebostatic axis, and zeroed.
Site. Dressing, erythema, duration.
The real question. What are you doing differently today because of this catheter? If the answer is nothing for two days running, take it out.
Related: arterial lines and CVP for the two waveforms underneath this one, balloon pumps, Impella and ECMO for when the numbers say the heart will not manage alone, and the clinical calculators for the derived values.
This is not medical advice. It is a teaching outline for clinicians and clinicians in training. Insertion, interpretation and management decisions belong to the team caring for the patient. Follow your own institution's protocols and confirm every value before acting on it.