Dialysis membrane-induced thrombocytopenia

As part of the regular hemodialysis prescription, we as providers have to prescribe the type of membrane we want used for the procedure. This decision is mostly based on what membranes our dialysis unit chooses to buy – though the majority are using synthetic polymer based membranes currently.

The very first dialysis membrane was made from cuprophane, later followed by the development of cellulose-based membranes. Both of these membranes contain many hydroxyl groups, which can intensely activate complement pathways. Newer synthetic polymer based membranes, such as polyacrylonitrile and polysulfone, are less likely to lead to complement activation. It is the degree of complement activation that determines how ‘biocompatible’ a membrane is. Furthermore, complement activation can lead to activation of the coagulation cascade with subsequent thrombosis, an obvious problem at the blood-membrane interface.

I came across an interesting case report of a chronic dialysis patient initiation, which was notable for the development of acute thrombocytopaenia. His mean pre- and post-dialysis platelet counts were 117 ± 22 × 10^-3/μL and 57 ± 27 × 10^-3/μL respectively, whilst dialyzing on a particular polysulfone membrane.

His post-dialysis D-dimer level was 3,748 ng/mL, felt to be consistent with activation of the coagulation system. His complement levels measured post-dialysis were not low - C3 and C4 were 139 and 46 mg/mL respectively. However, we did not have pre-dialysis levels to compare them to. Heparin was withheld, but with no resolution of his low platelet count - Haematology couldn’t find a good reason for it either.

The dialysis membrane was suspected, particularly in light of the elevated D-dimer, and so it was changed to an alternative polysulfone membrane, with a slightly different substructure composition. This resulted in resolution of the platelet count (157 pre and 145 × 10^-3/μL post dialysis) and lowering of the post-dialysis D-dimer level (906 ng/mL).

Activation of the alternate complement pathway has been reported to lead to neutrophil aggregation and stimulation. This, in turn, can predispose to platelet aggregation and activation of the clotting cascade, causing further platelet consumption. Different membrane composition can lead to differing degrees of complement activation and coagulation cascade activation. The development of new cytopaenias in newly initiating dialysis patients should raise the possibility of a membrane-induced reaction. Pre and post measurement of complement and D-dimers may help make the diagnosis. The membrane should be changed and counts followed for resolution.