Utility of CVP measurement in the ICU

This issue was vaguely touched upon in Question 14 from the first paper of 2001, "What are the determinants of central venous pressure? How may its measurement guide patient management?"  Some of this is already covered in the chapter on the information derived from the central venous pressure waveform. Those waveform findings are undoubtedly useful (for example, when there is impressive tricuspid regurgitation the abnormal CVP trace is an obvious and memorable feature). But the measurement itself: how may it guide patient management? A more recent Question 8 from the first paper of 2014 asks more directly of the candidates, "Discuss the role of CVP monitoring". This requires a little bit more thinking.

Extensive discussion of central venous pressure measurement and monitoring take place in the Haemodynamic Monitoring section. For the time-poor exam candidate who is untrusting of resources with "Deranged" or "Fast Lane"in the title, an excellent overview exists in this book chapter by Smith, Grounds and Rhodes;  one's reading can be safely limited to this single reference.

It is also quite easy to find articles about how bad the CVP is in predicting this or that, but it is much more difficult to find anybody writing anything in its defence.  Sheldon Magder's 1998 article in Intensive Care Medicine is probably the best one of those, even if it is almost 20 years old, and slightly off the topic. The same author also produced a similar piece in 2006, arguing that the major limiting factor for the utility of CVP measurement is a failure to acknowledge the potential errors of measurement. 

Anyway. First, briefly let us think about the waveform:

Information from waveform analysis

The CVP waveform can be a diagnostic aide. It can identify valvular or atrial pathology, and therefore has "utility" in the ICU. The waveform can give information regarding the following physiological features:

• Atrial rhythm
• Atrial capture in response to pacing
• Presence of atrial activity in SVT (i.e. "is it just a rapid AF?")
• Characteristic waveforms are seen in the following settings:
  • AF: absent a waves
  • SVT: absent a waves
  • Juctional rhythm, VT, complete heart block: cannon fused ac waves
  • Tricuspid regurgitation: fused cv waves
  • Triciuspid stenosis: prominent a wave
  • Pericardial constriction or poor RV compliance: bifid CVP wave
  • Cardiac tamponade: prolonged y descent
• Characteristic venous waveform (and pressure) can be used to confirm central venous catheter position (i.e. that it is not in the artery).

Information from the amplitude, i.e. the central venous pressure value

Central venous pressure may be reflective of right atrial pressure, which may be elevated in the following settings:

• Differential diagnosis of shock:
  Some aetiologies of shock can frequently be associated with a high central venous pressure; these include the following:
  • Right heart failure
  • Congestive heart failure
  • Massive PE
  • Cardiac tamponade
  • High PEEP
  • Constrictive pericarditis
• Other aetiologies (eg. septic shock, haemorrhagic shock) would normally present with a low CVP; or rather, it would be very unusual to find these shock states presenting with a high CVP.

However, CVP is not diagnostic of these pathological states. In recent and not-so-recent history, the CVP has undergone a major change, and is no longer viewed as a particularly useful parameter (whereas previously it was used extensively). In order to render this  into a well-organised exam-oriented discussion of CVP, I have formatted my response to Question 8 in the form of a stereotypical "critically evaluate" answer.

Thus:

Rationale for central venous pressure monitoring

• When measured in the superior vena cava, the CVP is related to the right atrial pressure, and therefore represents the filling pressure of the right side of the heart.
• The filling pressure of the right heart may be correlated with ventricular preload.
• Manipulation of ventriular preload is desirable, as it may improve cardiac output
• Thus, CVP may be useful as an indicator of whether the ventricular preload can be usefully manipulated, i.e. it may be predictive of fluid responsiveness. This statement begs a forceful counter-argument, which is presented later. 

Arguments for routine CVP monitoring

• CVP transducers display waveforms; even if the pressure measurement itself is not viewed as meaningful the CVP waveform is still useful (as discussed above)
• Central lines are ubiquitous, and transducing them is a relatively risk-free process, which means one can measure the CVP without any additional danger or discomfort to the patient.
• The pressure from the line can be used to confirm correct line position.
• The magnitude of the pressure can be used to discriminate between differential diagnoses of shock states.
•  

Arguments for use of central venous pressure as a therapeutic target

• In many parts of the world the CVP is the only advanced method of  monitoring preload or ventricular filling pressure.
• In a certain population of patients, CVP may still be useful as a means of predicting fluid responsiveness. These may include patients with stable (healthy unchanging) right heart function and right ventricular compliance.
• CVP may be an important variable to target when the patient has recently had surgery on a major central venous structure, or on the liver. A high CVP in these situations may be associated with a greater risk of bleeding.

Limitations of central venous pressure as a therapeutic target

Arguments from theoretical physiology:

• Even when measured in the superior vena cava, the CVP often does not truly correspond to the pressure distending the right atrium at the end of diastole.
• The most relevant pressure in terms of right atrial filling is the pressure at the onset of the c-wave, and this is not the pressure which will be displayed by the transducer.
• The most important pressure which dilates the right ventricle is not even the right atrial pressure, but the ventricular transmural pressure, which is a function of intrathoracic pressure as well as the right atrial pressure. As it is impossible to accurately determine at any given moment what the intrathoracic pressure is (or how much of it is transmitted transmurally) the right ventricular filling pressure can never be accurately known anyway, at least not without cannulating the ventricle.

Arguments from experimental physology:

• CVP does not correlate well with cardiac index (Ishihara et al, 2000)
• CVP does not correlate well with stroke volume index (Michard et al, 2003)
• CVP correlates poorly with those preload-related haemodynamic indices which do correlate well with cardiac index and stroke volume, such as the ITBVI and LVEDVI. (Diebel et al, 1992) 
• Changes in CVP correlate poorly with changes in stroke volume in response to volume loading (Gödje et al, 1998)

Arguments from clinical evidence:

• Probably the best summary of these arguments comes from the recent meta-analysis by Marik and Cavallazzi (2013), which has abundantly trashed the use of CVP for the prediction of fluid responsiveness ("This approach to fluid resuscitation should be abandoned", the authors concluded).