The large volume of crystalloid given during continuous hemofiltration is redistributed throughout the interstitium, and lots of lymph is produced

Many liters of hemofiltration 'replacement solution' exit the circulation and percolate through the tissues.

While continuous venovenous hemofiltration is thought of as a water removal technology, it is sometimes employed in a euvolemic mode (i.e., with no net loss of body water), such as in the support of a child with acute liver failure and encephalopathy. In euvolemia or in negative fluid balance, one might assume there should be no increase in extravascular pressure and therefore no increase in lymphatic flow. In fact, however, increased lymph flow is probably the source (actually, the supplier) of much of the water circulated through the hemofilter, even in euvolemic mode. The induced movement of body water by any means may increase lymphatic flow.

We can come to this conclusion by following the movement of albumin. Immediately following water loss induced by a diuretic, furosemide, the steady state return rate of albumin (from the interstitium into the plasma) is increased. Albumin is carried into the plasma passively by an increase in the lymphatic flow of water, i.e., by convection.

A similar phenomenon is seen with experimental euvolemic plasmapheresis. Through an increase in lymph flow, serum albumin concentration is restored (or perhaps we should say 'maintained', as the lymph flow probably increases immediately as serum proteins are being removed by plasmapheresis).

And finally, when saline is infused intravenously while holding the mean arterial pressure and venous pressure constant, albumin can be detected moving from the vascular system into the lymph. In a healthy person this can only occur if water is moving in the same direction, as healthy capillaries do not leak albumin.

It is tempting to think of hemofiltration as a simple closed loop of ultrafiltrate production balanced by an infusion of 'replacement solution'. In reality, infused replacement solution takes a more circuitous route. Crystalloid enters the circulation, percolates through the interstitium, and then is taken up by lymphatics, thus returning to the circulation.


Henriksen JH, Parving HH, Lassen NA, Winkler K. Increased transvascular escape rate and lymph drainage of albumin in pigs during intravenous diuretic medication. Relations to treatment in man and transport mechanisms. Scand J Clin Lab Invest 1982; 42: 423-9.

Wallace JR, Bell DR. Comparison of protein lymph flux and extravascular uptake in skin during increased venous pressure. Am J Physiol 1992; 263: H895-902.

Gandhi RR, Bell DR. Influence of saline infusions on transvascular protein transport. Am J Physiol 1992; 262: H443-50.