Today I cast a critical eye over a recent paper from Professor Tatara’s team in Hyogo, Japan. DOI: 10.1111/aas.13548. It is a fine example of the sort of physiology experiment that can be performed in clinical practice to advance our understanding of acute body water distribution changes.
The study population was 56 consecutive patients (age range, 20-80 years) who were scheduled to undergo major abdominal surgery for cancer. Professor Wernermann (Karolinska Institute, Sweden) recruited 10 apparently similar patients aged 65.7 ± 5.6 years, scheduled for major pancreatic or oesophageal surgery, to a more detailed investigation to elucidate the time pattern of plasma albumin changes and calculated extravasation of albumin (the albumin shift). Norberg et al. Critical Care (2016) 20:113
A really important piece of information in such studies, rarely provided, is whether patients were upright and ambulant before reclining on the operating table. As we see in the Swedish cohort, plasma albumin concentration measured at the time an arterial line is first sited and before anaesthesia is induced is high (mean circa 32 g/l). It is likely the legacy of high capillary pressures in the legs while the subject was upright driving hyperfiltration of plasma water to the interstitial fluid (ISF) of the legs. As the patient reclines and relaxes on the operating table, the capillary pressures in the legs suddenly become much lower and a transient absorption of ISF to the plasma is initiated. The plasma protein concentration and the peripheral blood haematocrit fall quickly. We see in the Swedish cohort that plasma albumin concentration is still falling as anaesthesia is induced and as surgery is commenced (<30 g/l). Intravenous infusions commence, further diluting plasma proteins and red cells as the other dilution influences continue. Nadir albumin concentrations are typically achieved by the end of surgery (<20 g/l).
The Japanese patients had their first blood sample drawn after induction of anaesthesia and at a time when the patients were hypotensive. There is no steady-state baseline value here. The mean plasma albumin concentration at that point was 37 g/l. An infusion of around half a litre of hydroxyethyl starch in 15 minutes reduced the plasma albumin to 30 g/l. The dilution effect of the infusion was immediate; the likely elevation of capillary pressure by the infusion would have tended to increase filtration, but taking longer to manifest itself.
The Japan team claim very reasonably that “it is plausible that lowered capillary hydrostatic pressure in the S group compared to the P group, attenuated fluid filtration from plasma into the interstitium. This scenario may also explain the decrease of ΔPVHb with time (ie, 15 vs 30 minutes) in the P group, which likely resulted from increased transcapillary fluid filtration.”
It is important to recall in this example that the presumed ability of a phenylephrine infusion to increase capillary pressure is explained by increased stressed blood volume elevating mean systemic and arterial pressure; if the predominant effect had been arteriolar constriction in a normotensive subject with little capacity to further increase the stressed volume then capillary pressure would have fallen while the arterial pressure rose. This is another example of the vital concept of Context Sensitivity. Vasoconstrictors can be used to support plasma volume expansion from hypovolaemia (below the J point) but they will oppose plasma volume expansion in euvolaemia or hypervolaemia (at and above the J point).