Thursday, 22 May 2014

Measuring CVP? A plea for some common sense.

There was an updated meta-analysis recently published looking at the utility of CVP measurement to predict fluid responsiveness. You can probably guess the answer. If the CVP is low then give them some fluid and they get better... right? Not so fast.

These authors from Virginia and Kentucky performed a typical but good quality systematic review. They were looking for clinical trials that reported the correlation coefficient or area under the curve (AUC), between CVP and change in cardiac performance following an intervention that altered cardiac preload. 43 studies were eventually selected for data abstraction. The vast majority were small studies conducted in an ICU or operating room and none were done in an emergency department. Most patients got a small bolus of colloid. Cardiac output was measured using various means.

Results? 57% of patients were fluid responders. The summary AUC was 0.56. This is as good as flipping a coin. The summary correlation coefficient was a terrible 0.18. (1.0 would be considered perfect and 0 is no correlation.)

The authors conclude that there is no data to support the widespread practice of using CVP to guide fluid therapy and it should be abandoned.

This looks to be quite definitive. The Achilles heel of a meta-analysis is negative reporting bias. One should be quite cautions of “positive” meta-analysis, but one that tells you something doesn’t work is believable.

 I would most certainly not place a central line for the sole purpose of CVP measurement. Many would argue that they only real reason for placing one would be for administration of noradrenalin (norepinephrine).

With the widespread uptake of point-of-care ultrasound I can’t imagine relying on a CVP measurement to guide therapy. One only needs to non-invasively take a quick look at the IVC (and probably the lungs and heart while you are there) to get an idea of fluid status. Sure, this is far from perfect. But it probably works well at the extremes (when it’s important) and will certainly avoid the morbidity of central line placement.



Marik PE, Cavallazzi R. Does the Central Venous Pressure Predict Fluid Responsiveness? An Updated Mea-Analysis and a Plea for Some Common Sense. Crit Care Med 2013;41:1774-1778.

No comments:

Post a Comment