Dinosaurs still roam the earth, I know this. They are slowly and surely retiring, though. I trained in the days where Manny Rivers and the Surviving Sepsis Campaigns pushed for Early-Goal Directed Therapy were king. Hey, there was a reason why it was so successful, EGDT decreased in-hospital mortality from 46.5% to 30.5% in that single centered study where the patients were watched over like hawks, etc. I’m not here to take apart that study. I am not also not saying that Central Venous Pressure (CVP) is useless. It does have its uses such as trending in certain patient populations such as those in the CVICU. But one should never use the isolated value to determine management of the patient. Even in the case of suspicion of tamponade in a CVICU patient you are also seeing the CI/CO decrease and equalization of pressures in the cardiac chambers.
I have described what fluid responsiveness is over the last several posts. I have written several posts about the matter. Whether you look at stroke volume, cardiac index, cardiac output, pulse-pressure variation, stroke volume variation, end-tidal CO2, echo variables such as VTI. Fluid responsiveness is giving a patient either a bolus of fluid or performing a passive leg raise in a matter that one could assess that the increase in cardiac preload is going to yield an increase in cardiac output. Simple as that. You know what doesn’t tell whether there’s an increase in cardiac output or stroke volume? The CVP.
One thing is clinical gestalt where people interpret a CVP and think they know what they’re doing, another thing is actual statistics and data. The PROCESS, PROMISE, and ARISE trials from 2015 and 2016 have shown us that we’ve gotten significantly better at identifying and treating sepsis to the extent where we are no longer routinely measuring CVP’s and ScVO2’s. We, as a specialty, have moved on to bigger and better things, as we should, given that the 1959 study that introduced CVP/RAP (right atrial pressure) likely would not have been published today due to it’s lack of statistical evaluation amongst other things.
I’m not going to get into the nitty gritty statistics on how Paul Marik conducted this meta-analysis where they looked at CVP in relation to a change in stroke volume index/cardiac index. They did find that the area under the curve was just 0.56 for ICU patients. Dust off your statistics cap, team, and AUC of 0.56 isn’t good. As a reference 0.5 means that you can’t tell whether an assessment works or not. 0.7-.8 is acceptable, 0.8-0.9 is good, and >0.9 is outstanding. I linked a paper so you could teach yourself how to interpret the receiver operating curve/area under the curve below.
To add insult to injury in the case of CVP, they also checked a correlation coefficient in this meta-analysis. I’ll review the stats for your here as well. The correlation coefficient for the study was 0.18. 0.00-0.30 is a negligible correlation. Also not good. You’d want a correlation coefficient >0.5 to at least be moderately positive. That’s not the case here.
Two years later, Eskesen and friends conducted a systematic review of 1148 patients over 51 studies and had similar findings. Regardless of the group CVP, whether it was <8, 8-12, or >12, the AUC was not above the “acceptable” level of 0.7. Heck, not even the upper arm of the confidence interval showed that. The table in that paper is pretty impressive of how pointless CVP is as a predictor of fluid responsiveness. I would post it here but I don’t want to get dinged by the copyright gods.
All in all, does CVP predict fluid responsiveness? The answer is a whole bunch of NOPE. The data proves it. Marik goes on to say that CVP should not be used to guide fluid therapy. I largely agree with him. I just feel that in medicine it’s hard to talk in absolutes. Please read the article for yourself and don’t trust me.
Does an increase or change in Central Venous Pressure (CVP) indicated fluid unresponsiveness?
Some clinicians have proposed trending CVP and the change in CVP to indicate fluid responsiveness and unresponsiveness. This, however, does not have much data to support it. Hamzaoui, et al looked at this in a recently published a prospective observational study in Critical Care Medicine in June 2020. They analyzed 50 patients and took 57 sets of measurements. As the “gold standard” of sorts to compare CVP, they used Velocity Time Integral (VTI). VTI is an echocardiogram measurement. Going down the path of explaining it is beyond this section of the post, see below in the echocardiogram section. Just be aware that it is very good for assessing stroke volume which helps calculate cardiac output (CO) and therefore fluid responsiveness, hence the authors using it as the baseline.
Rather than providing the patients with IV fluids and potentially cause harm by overloading patients, the authors perform passive leg raising (PLR) on the patients. They defined an increase in CVP ≥ 5mmH2O or 4mmHg as a stopping point for more fluids. This means that the patient would not be fluid responsive moving forward. A positive PLR was defined as an increase in VTI ≥ 10%.
The authors found the neither baseline CVP nor ΔCVP were any good for predicting lack of fluid responsiveness. This means DON’T USE IT. The area under the curve was 0.54 (p=0.6) and 0.59 (p=0.25) respectively. An acceptable AUC is at least 0.7 as a point of reference. Yet another example of why you should not use CVP to guide fluid resuscitation. Consider other alternatives. The authors concluded that “a marked increase in CVP during PLR does not indicate that the PLR test is negative and hence that preload responsiveness is unlikely.”
Marik PE, Cavallazzi R. Does the central venous pressure predict fluid responsiveness? An updated meta-analysis and a plea for some common sense. Crit Care Med. 2013;41(7):1774‐1781. doi:10.1097/CCM.0b013e31828a25fd
Link to Abstract
Eskesen TG, Wetterslev M, Perner A. Systematic review including re-analyses of 1148 individual data sets of central venous pressure as a predictor of fluid responsiveness. Intensive Care Med. 2016;42(3):324‐332. doi:10.1007/s00134-015-4168-4
Link to Abstract
Citation in case you don’t remember what the Area Under the Curve/Receiver Operating Curve is:
Mandrekar JN. Receiver operating characteristic curve in diagnostic test assessment. J Thorac Oncol. 2010;5(9):1315‐1316. doi:10.1097/JTO.0b013e3181ec173d
Link to FREE PDF
Citation for Correlation Coefficient
Mukaka MM. Statistics corner: A guide to appropriate use of correlation coefficient in medical research. Malawi Med J. 2012;24(3):69‐71.
Link to FREE PDF
Hamzaoui O, Gouëzel C, Jozwiak M, et al. Increase in Central Venous Pressure During Passive Leg Raising Cannot Detect Preload Unresponsiveness [published online ahead of print, 2020 Jun 8]. Crit Care Med. 2020;10.1097/CCM.0000000000004414. doi:10.1097/CCM.0000000000004414
Link to Abstract
How to support my work:
I have written “The Vasopressor & Inotrope Handbook: A Practical Guide for Healthcare Professionals,” a must-read for those in the field! You have several options to get a copy while supporting my endeavors. If you’re in the US, you can order directly from me with a special touch: A SIGNED COPY. Please note that I handle these shipments personally, so I appreciate your patience.
For quicker delivery or international orders, I recommend purchasing through Amazon. It is also available for KINDLE. When you use these affiliate links, I earn an additional commission at no extra cost to you, which is a great way to support my work.
Stay tuned for the audiobook version, which will soon be available on Audible. It’s an exciting addition that I’m eager to share with you!
Although great care has been taken to ensure that the information in this post is accurate, eddyjoe, LLC shall not be held responsible or in any way liable for the continued accuracy of the information, or for any errors, omissions or inaccuracies, or for any consequences arising therefrom.