“One unifying biomarker, molecule, behavior, treatment, or approach is unlikely to solve all the problems of [acute kidney injury].”
TO date, we have no single biomarker for acute kidney injury (AKI) though many have been tested. Failed biomarkers include urinary cystatin-C, urinary neutrophil gelatinase–associated lipocalin, α & π glutathione S-transferases, urinary I-type fatty acid–binding protein, and urinary hepicidin. Combination of urinary tissue inhibitors of metalloproteinase-2 and insulin-like growth factor–binding protein-7 may hold promise. In this issue of Anesthesiology, Kertai et al.1 associate a decrease in platelet number after surgery with AKI. The impulse is to assume that a decrease in platelet count signals microvascular thrombosis, but platelets may be an indicator cell for other dysfunctions.
Acute kidney injury is complex, highly investigated, prevalent (30 to 51%), and responsible for morbidity. The interactions of risk factors (age, diabetes mellitus, hypertension, preexisting kidney dysfunction, IV dye loads, and many more) in conjunction with hemodynamic perioperative factors, regional blood flow, multiple medications (vasoconstrictors among others), and inflammatory reactivity are highly interactive. We need actionable events to intervene, not just avoidance of risk factors.
Medicine has attacked complex processes with compartmentalized dissective thinking. Systems analysis is difficult and possible once the building blocks allow for computer computational models. Hemostasis is an example of biologic complexity, into which platelets play a key role. Computer models of hemostasis contain hundreds of nonlinear, simultaneous reactions and cell responses, feeding forward and backward. Coagulation is not your mother’s old extrinsic/intrinsic cascade anymore! Far reaching consequences with the actions of thrombin is an example of complexity. Thrombin activates multiple cell lines and proteins as well as turns on genes and ramps up inflammation. Existing computer models of coagulation/hemostasis have not been fed the myriad changes of cardiopulmonary bypass (CPB).
Platelets are inflammatory cells with multiple actions, through gated responses of 20 to 50 known pathways for interactions with their environment, inflammation, and the microcirculation. We focus upon platelets as the keys to hemostasis/thrombosis overlooking their larger role. In reality, platelets form the first-line inflammatory defenses (first phagocytize invading pathogens). Platelets are important in normal endothelial cell (EC) health, glycocalyx maintenance, shear stress signaling, erythrocyte lipid/prostanoid transfer, vascular tone, growth, and regeneration. Platelet function is a catch-all term bantered about with little regard for depth of meaning. Thrombocytopenia (TCP) in response to CPB is the norm, not the exception.
Cardiopulmonary bypass is an inflammatory systemic insult not unlike sepsis. TCP is a hallmark of AKI in a range of processes. Viral infection (hepatitis and hantavirus infection), bacterial infection (sepsis), preeclampsia/eclampsia, trauma, heat stroke, reperfusion injury, radiology dye load, even certain snake envenomation described TCP associated with AKI. What these situations have in common is EC dysfunction, inflammation, and dysregulation of homeostasis. The accompanying retrospective study cannot define causation although heparin-induced thrombocytopenia (HiT) is discussed. When HiT has been studied in CPB, 30 to 50% of patients by days 4 to 10 produce antibodies. In the bivalirudin CPB studies, antibody formation was decreased and adverse outcomes were fewer than in standard heparin-anticoagulated CPB. Perhaps, undetected heparin antibody has a role in the TCP and AKI but that has not yet been studied. If TCP does not resolve or it worsens within 24 to 48 h after CPB, HiT should be investigated. The data reported here show that the TCP is most often early (days 1 and 2), not when antibody seems to peak (days 4 to 10). That does not mean that HiT/platelet factor-4 heparin antibody is not at least partially responsible. We simply do not know from these data.
Cardiopulmonary bypass causes a systemic inflammatory reaction that is multifactorial, highly variable, and the model of biologic complexity. Unfractionated heparin (UFH) used in CPB is itself a cause of TCP. UHF is naturally found in mast cell granules as a chelator of histamine and calcium. When released from mast cells, in response to ex vivo invasion, UFH promotes neutrophil movement. UFH has synergy with heparan (the long chain polysaccharide actively maintained on the endothelial glycocalyx). Different in structure and sulfhydryl charges, UFH and heparan are distinct. UFH liberates heparan connected to syndecan tethers. Heparan binds interleukin-6, interleukin-8, and tissue necrosis factor-α but importantly all stimulate leukocytes to release cytokines—why? The slime coat of Escherichia coli are mucopolysaccharides with synergy to heparan. Heparan, similar to UFH, does not circulate in plasma. EC biology and the interface of the glycocalyx with flowing plasma and platelets are important. Ischemia–reperfusion leads to shedding (50 to 90%) of the glycocalyx. Imbedded enzymes are lost during shedding and the ECs shift from being anti- to proinflammatory/prothrombotic. Xanthine oxidase is held in the glycocalyx, the loss of which (caused by UFH) changes the ability of the cell surface to control oxidative stresses. Albumin, an antioxidant, is in the glycocalyx and with CPB albumin levels drop. Antithrombin is held in the glycocalyx, interacts with heparan, and is dramatically reduced by both UFH and CPB. Exogenous administration of albumin during CPB is promising, and antithrombin improves EC recovery in sepsis, eclampsia, and CPB. The administration of antithrombin in extra corporeal membrane oxygenation decreases the amount of TCP. The interactions of EC and CPB are complex and not well investigated to date.
A culture of blaming the CPB machine for inflammation exists and is simplistic and outdated; such discussions should be abandoned by informed physicians. Thrombin generation during CPB comes predominantly from EC, tissue release, and less from contact activation. In extra corporeal membrane oxygenation compared with CPB, there is less thrombin generation, no tissue destruction, and few microbubbles. UFH binds only plasma-free thrombin not that held on cells (the majority of thrombin). The insults to ECs may be key to understanding the biologic complexity of CPB inflammation. Renal ECs (highly specialized) have roles in blood filtering, secretory, and other functions of the kidney. In kidney rejection, renal EC glycocalyx is destroyed, similar to what is seen with ischemia–reperfusion or insults of CPB.
Microvascular air bubbles (universal in CPB) destroy EC, as do low blood flow, shear stress, ischemia, and high potassium (cardioplegia). Oxidative stress is particularly bad for EC. CPB decreases the concentration of circulating antioxidants. Nadir platelet count in this study was associated with more transfusion. The difference in median transfusion utilization is large and compelling, but is it a cause of TCP and AKI or an effect of bleeding and unstable patients? Transfusion has been widely reported to be associated with AKI. Transfusion of stored erythrocyte carries a large free iron load dependent on the length of storage as well as the amount transfused. Ferritin and hepicidin levels are reduced during CPB. Free iron is toxic to EC, and the use of CPB with cardiotomy suction causes hemolysis (free iron/heme). Transfusion also has cytokine loads, low pH, and lipid cellular microparticles (inflammatory platelet activators) and blocks the microcirculation. Stiffened, dysfunctional erythrocyte (banked blood) create abnormal shear stresses that make EC shed their glycocalyx when exposed to transfusion. Erythrocyte transfusions are associated with worse bleeding, not less. Perhaps the increased bleeding is due to high levels of platelet-activating factor found in the banked blood. Platelet-activating factor creates dysfunctional platelets, neutrophil activation, and multisystem organ failure including lung dysfunction (EC capillary leak). Relationships seem to be everywhere, all pointing toward microvascular EC dysfunction and dysregulated biologic complexity.
In conclusion, the accompanying article is rich with observations. The platelets are probably not the problem, but an indicator of poorly buffered inflammation. The platelets tell us to embrace biologic complexity to not look for simple solutions. They tell us to look upstream, perhaps all the way to the very fundamentals of CPB (UFH, antioxidants, and many more). Biologic complexity is modified by physician behavior, culture, and vagaries—i.e., heparin–protamine protocols, length of bypass, and transfusion. One unifying biomarker, molecule, behavior, treatment, or approach is unlikely to solve all the problems of AKI. However, in depth research, embracing the beauty of biocomplexity may allow hypothesis-driven projects to sprout from this important observational study. We need to look at what is causing the platelets to react so dramatically. It seems to Kertai et al. that the interaction of EC and platelets might reveal a secret, if only we can ask the probing questions.
The author is not supported by, nor maintains any financial interest in, any commercial activity that may be associated with the topic of this article.