Etude des marqueurs iNnovants de la VOLémie
Congratulations to Bernard Vigue for conceiving the ENVOL study, and to his colleagues for painstakingly collecting the data. The majority of patients had primary intracranial pathology which makes one concerned about relevance in a non-neuro setting, but I shall put that to one side. The ‘optimal control of blood volume without fluid overload’ is of course what the Revised Starling Equation and Glycocalyx Model paradigm is all about, so I would be thrilled if the team have truly discovered ‘biomarkers to help characterize fluid overload status.’
Fluid Overload was defined as cumulative gain of more than 4 litres of water or 36 g sodium, said to be equivalent to 4 litres isotonic salt solution. I would therefore presume that if a patient presented 4 litres or 36 g sodium below her premorbid state, she would be classified as overloaded once the presenting deficit was corrected. The collected data were interpreted within an Adjacent Buckets Twigley Hillman paradigm of extracellular fluid distribution, so it is not surprising that plasma volume as estimated by red cell dilution was only loosely correlated to plasma volume as estimated by albumin dilution, r squared 0.75. No endothelial glycocalyx in this paradigm!
Hold on to your hats for the next finding; patients had lower-than-normal total blood volume on D2 and D7. That’s right; after apparently successful circulatory resuscitation, the TBV was subnormal. The majority of patients were >20% below normal TBV and so were classified as hypovolaemic, though we presume that their haemodynamic state was not suggestive of hypovolaemia. So what should you treat? The calculated blood volume or the measured haemodynamic state? There are those who would increase the calculated blood volume by plasma and/ or RBC transfusion even though blood pressures and heart rate are satisfactory. For an example, see Jacob, Chappell, Hofmann-Kiefer et al 2012. I am not among them.
So which biomarker can we utilise to warn us of fluid overload? The Winner among the measured candidate molecules was pro-adrenomedullin. Notice, however, that the measured plasma concentration was log-transformed to make the statistical test positive. I have used this trick myself in the long-historic past when studying clonidine premedication, and have never been entirely comfortable that it is OK to draw conclusions about the actual concentration rather than its logarithm. Notice also that not a single measurement of pro-adrenomedullin concentration in this study was within the normal range (< 0.39 nmol/l). Anyway, we are assured that when [pro-adrenalmedullin] > 0.865 nmol/l the sensitivity and specitivity for sodium overload are 0.62 and 0.85. If water overload is of more interest to you, then the threshhold concentration is 1.1 nmol/l and the sensitivity and specitivity are 0.83 and 0.60. I’d rather examine the patient.