The Pulmonary Artery Catheter - friend or foe?


Use of the pulmonary artery catheter is controversial. It is considered by some to be the gold standard for haemodynamic monitoring and therapy and by others to be both unnecessary and dangerous. This case highlights both its usefulness and problems.

Clinical problem

A 40 year old lady was admitted to hospital with abdominal pain and vomiting. She had a past medical history of ulcerative collitis necessitating a previous pan-proctocolectomy and ileostomy. She had had a number of admissions since with subacute bowel obstruction. Over 5 days of conservative management her symptoms gradually settled. She then developed hypoxic respiratory failure and hypotension. She was admitted to ICU with a diagnosis of severe sepsis secondary to a hospital acquired pneumonia.


Initial treatment on ICU consisted of broad spectrum antibiotics (Piperacillin-Tazobactam), fluid resuscitation and CPAP via facemask. Her respiratory failure worsened and she was intubated the following day. She rapidly developed multi organ failure (MOF) requiring noradrenaline to maintain mean arterial pressure (MAP). A high lactate (6 meq/l) prompted an exploratory laparotomy which revealed adhesions accounting for small bowel obstruction but nothing else abnormal. After theatre her lactate increased further to around 10 and with ongoing acute kidney injury, renal replacement therapy (RRT) (CVVHDF) was commenced. A pulmonary artery catheter (PAC) was sited to guide fluid resuscitation and vasopressor/inotrope use. Based on readings of a low cardiac index and high SVRI dobutamine was commenced in addition to noradrenaline, the dose of which was reduced. A dose of frusemide was given to ‘offload’ the heart based on readings of a high central venous pressure (CVP) and high pulmonary artery occlusion pressure (PAOP) (presumed to reflect LAP as no history of pulmonary hypertension) combined with a low cardiac index (CI) and high systemic vascular resistance index (SVRI). Refractory hypoxaemia prompted commencement of high frequency oscillatory ventilation (HFOV). Activated protein C was started 12h post-operatively (it was stopped 36h later after her platelets fell to 14). Over the next 24h lactate remained high at 8-9. SvO2 was measured at 60-65%. A reassessment of cardiovascular status was then made. A focussed ECHO confirmed global moderately impaired LV function. CI was measured at 1.9 l/min/m2. PAOP was 17cm H2O. It was not possible to assess systolic pressure variation (SPV), pulse pressure variation (PPV) or stroke volume variation (SVV) as HFOV was ongoing. Passive leg raising resulted in an increase in MAP from 75 - 95 with an increase in CVP of 3 and on this basis a fluid challenge was given. Hartmanns was syringed in until there was a sustained rise in CVP of 3. The total volume infused was 350mls. MAP increased from 78 to 115, CI measured before and afterwards increased 21% from 1.9 to 2.3. SvO2 20 minutes later had increased from 65 to 76%. Noradrenaline was stopped and dobutamine continued. Further fluid challenges were given as the crystalloid redistributed. Lactate only improved slightly despite the improvement in cardiac output and high dose CVVHDF (36ml/kg/h total dose) which was put down to high dose dobutamine. 48 hours later dobutamine and RRT were no longer required and weaning from mechanical ventilation commenced. The PAC was unable to be withdrawn and it was noticed on imaging that it was knotted in the right ventricle. It was successfully removed by an interventional radiologist.


The PAC measures pressures, cardiac output (by thermodilution) and SvO2. From these measurements many other variables can be calculated. No monitoring device has ever been shown to improve outcomes on its own. This includes pulse oximetry
1. Like these monitors the PAC will only improve outcome if the monitored data are used correctly for therapy. In order for it to be of any use and not cause any harm, it has to be inserted safely, the data has to be obtained correctly, this data interpreted correctly and put into clinical context and then the appropriate therapy instituted.

Problems related to the catheter :

The complications identified by the PAC-man 2 and Escape 3 trials were site haematoma, arterial puncture, arrhythmias, catheter related blood stream infection, catheter knotting and pulmonary infarction and haemorrhage. In this case catheter knotting could have lead to serious consequences - a sternotomy by a cardiothoracic surgeon was thankfully avoided by the success of an interventional radiologist.

Problems with data collection and interpretation:

There are a number of problems in obtaining accurate data from a PAC. Besides this a number of studies have highlighted the inability of clinicians to correctly interpret the data from PACs
4. PACs are now used with less and less frequency in the UK with resulting implications for training in their use. In this case the patient was given frusemide based on a clinicians interpretation of data from the PAC when it was clearly demonstrated afterwards that the patient was fluid responsive. This could have resulted in significant harm.

Preload is end diastolic wall stress and not end diastolic pressure or end diastolic volume. Studies have conclusively shown that PAOP is unrelated to fluid responsiveness.

Cardiac Output:
Cardiac output data depends on the phase of the respiratory cycle during which the injectate is given as positive pressure ventilation produces variations in ventricular stroke volume.
Tricuspid regurgitation is significant in around half of critically ill ventilated patients. This will render thermodilution techniques inaccurate.

Derived data (such as SVR):
Derived data are calculated by combinations of measured variables, all of which may have their own error. Thus errors are multiplied.
Impedence, and not resistance, is mathematically correct for the cardiovascular system as oscillatory flow is occurring, but is practically difficult to measure, so resistance is used as a surrogate for the opposition to flow.
SVR and afterload are not the same. Noradrenaline has been shown to increase SVR while reducing afterload.
The normal range of SVR is derived from normal patients. Haematocrit, plasma protein concentration and temperature all affect blood rheology and therefore SVR readings.
Stroke work index (MAP x SVI) is inaccurate with a ventricle with reduced inotropy and compliance.

Improved outcome:
Studies have consistently failed to show an improvement in outcome with the use of the PAC. These include large prospective randomised controlled trials, meta-analyses and a Cochrane review. 7
There is some weak evidence that more severely ill patients may benefit from a PAC (observational cohort study and retrospective analysis
Observational studies suggested increased harm with PACs but these have been refuted by subsequent prospective randomised controlled trials. 8

Lessons learnt

The complications associated with the PAC combined with the failure of trials to prove its usefulness plus the availability of non invasive cardiac output monitors have prompted some authors to advocate it being consigned to the history books. The evidence seems to suggest neither harm nor benefit. What is important is remembering that the PAC is merely a monitor, its risks and benefits should be evaluated on an individual basis and correct understanding of its use is essential. My opinion based on the evidence and my personal experience is that less invasive methods of monitoring such as pulse contour analysis and echocardiography (which have both been shown to have good correlation with PAC data
9) have a better risk benefit ratio.


Pedersen T, Pedersen P, Moller AM. Pulse oximetry for perioperative monitoring. Cochrane Database Syst Rev 2001(2):CD002013.

Harvey S, Harrison DA, Singer M, Ashcroft J et al. PAC-Man study collaboration: Assessment of the clinical effectiveness of pulmonary artery catheters in management of patients in intensive care (PAC-Man): a randomised controlled trial. Lancet 2005; 366: 435-36.

Binanay C, Califf RM, Hasselblad V et al. Evaluation study of congestive heart failure and pulmonary artery catheterization effectiveness: the ESCAPE trial. JAMA 2005; 294: 1625-33.

Ilberti TJ et al. A multicentre study of physicians knowledge of the pulmonary artery catheter. Pulmonary artery catheter study group. JAMA 1990;264(22):2928-2932

Wigfull, J. et al. Contin Educ Anaesth Crit Care Pain 2005 5:84-88

Lang R, Borow K, Neumann A, Janzen D. Systemic vascular resistance: an unreliable index of left ventricular afterload.
Circulation 1986; 74: 1114–23

Harvey S, Young D, Brampton W et al. Pulmonary artery catheters for adult patients in intensive care. Cochrane Database of Systematic Reviews 2006; 3: Art. No.:CD003408. DOI: 10.1002/14651858.CD003408.pub2.

V Zochios, S Gopal. The current role of the pulmonary artery catheter in critical care: a case report and review of the literature. JICS 2009. Volume 10, Number 4

CL Colebourn, V Barber, JB Salmon, JD Young. The accuracy of diagnostic and haemodynamic data obtained by TTE in critically ill adults. JICS 2008 Volume 9, Number 2