The accurate prediction of which recipients of liver transplantation (LT) will do well and which will have serious early complications remains somewhat elusive. Consequently, we frequently identify higher risk categories of LT on the basis of the recipient's diagnosis and Model for End-Stage Liver Disease score, donor variables, operative variables, and graft quality. Despite our experience, we are still frequently surprised by clinical events during the post-LT course. In short, the prediction of severe complications after LT remains more of an art than a science. In this issue of Liver Transplantation, Lesurtel et al. present a new predictive criterion in their article entitled “Low Platelet Counts After Liver Transplantation Predict Early Posttransplant Survival: The 60-5 Criterion.” The authors propose that the platelet count in the early posttransplant period can be used to predict the occurrence of severe complications.
Platelets, in addition to their well-known hemostatic and coagulation functions, are also important in inflammation and ischemia/reperfusion injury after LT. Platelet adherence and aggregation in the reperfused liver graft endothelium can lead to sinusoidal endothelial cell apoptosis and may aggravate the no-reflow phenomenon.[2, 3] Many LT recipients are already thrombocytopenic because of portal hypertension–related splenic sequestration. Reperfusion exacerbates this peripheral thrombocytopenia because of the rapid uptake and aggregation of activated platelets in the liver due to ischemia/reperfusion injury. Platelets are also important in tissue repair and regeneration via serotonin-mediated mechanisms.[5-7] In short, the complexity of platelet function, in concert with the physiology of portal hypertension and transplant ischemia/reperfusion injury, makes thrombocytopenia in the post-LT patient very common.
Information-hungry clinicians love more data to guide decision making, and they enthusiastically employ predictive models for LT and liver surgery.[8-12] In the study under discussion, the authors report on 257 patients who received whole allografts from brain-dead donors for first-time primary transplants. This study proposes a new predictive model called the 60-5 criterion. It analyzes the platelet count during the early post-LT time period and correlates a platelet count < 60 × 109/L on postoperative day 5 with the occurrence of Clavien-Dindo grade IIIB to IV complications and a 2-fold increase in mortality within 90 days of LT.
The retrospective nature of this study does not allow postulation of the mechanism of action of thrombocytopenia and worse outcomes after LT. The authors acknowledge this limitation in their discussion. As such, we are sure that many readers will be left pondering the metaphor of what came first, the chicken or the egg. Is the thrombocytopenia really a cause, or is it simply a marker or maybe a result of other pathophysiological events unfolding and preceding clinical signs and symptoms? This inquiry is beyond the scope of the article, which simply describes an association of platelet counts and adverse outcomes. Regardless, one cannot help but further question the complex interactive mechanisms at work. Unfortunately, the literature is rather sparse on this topic, and we will resist the temptation to speculate here.
The reader will likely notice that the cohort meeting the 60-5 criterion was clearly a sicker group and had a higher average Model for End-Stage Liver Disease score, worse Child-Pugh classification, worse American Society of Anesthesiologists physical status, and more renal dysfunction requiring pre-LT renal replacement therapy (a 2:1 ratio). These variables typically would hint at poorer outcomes after LT. This may suggest a cascade of events resulting in a domino effect of sicker patients having more complicated surgery and an expected complicated post-LT course. Consequently, the patients who met the 60-5 criterion had greater mean intraoperative blood loss (1500 versus 1000 mL) and higher mean transfusion requirements (3 versus 1 U of packed red blood cells). These patients also had on average one 6-pack of platelets transfused, whereas those who did not meet the criterion had approximately 0 platelet transfusions. Interestingly, others have reported that platelet transfusions are independently associated with increased mortality after LT, although the mechanism has been unclear. Furthermore, a report by Chatzipetrou et al. showed that severe post-LT thrombocytopenia, which occurred in 22.1% of their 541 LT cases, was associated with significantly decreased survival.
Despite some inherent design limitations, this study was well modeled with a careful analysis and a logical discussion. However, the authors note that the sample size was relatively small and that the multivariate analysis event rate was low. Therefore, the need to control for confounders can result in a high risk of overfitting. Nevertheless, the subgroup survival analysis was well done and demonstrated that the predictive value for survival was for the early post-LT period only (90 days). The receiver operating characteristic curve analysis with Youden's index demonstrated that the ideal cutoff value for platelet counts was <60 × 109/L on postoperative day 5, and this allowed the development of the proposed 60-5 criterion.
Perhaps the most attractive aspect of this criterion is its simplicity. On postoperative day 5, you look at the platelet count. That is it. No apps, mobile devices, or computers are necessary! That is pretty simple. Also, getting on top of a complication in evolution and intervening even a day earlier might hold a developing problem in check and will likely have a beneficial effect on outcomes. Moreover, it is likely that the majority of these evolving complications will be infectious. Therefore, the clinician may obtain imaging to rule out occult abscesses or technical problems. Additionally, cultures may be collected, or a slowly progressing patient may be started on empiric antibiotics while further investigations are underway.
According to this report by Lesurtel et al., the 60-5 criterion has the potential to evolve into a simple and reliable predictive metric that can provide the clinician with the opportunity to intervene early and theoretically change the postoperative course of the patient. We look forward to prospective clinical investigations that may further strengthen the 60-5 criterion as a prediction tool. Future studies that elucidate the importance of platelet pathophysiology in the early post-LT period are also warranted. A prospective clinical trial with mechanistic/molecular investigations might address both of these questions. Because the authors have a proven track record in executing such trials, we are heartily anticipating any further data that would enhance our understanding of this phenomenon and potentially elucidate therapeutic interventions with the potential to improve outcomes for our patients.
Jay A. Graham and James V. Guarrera
Center for Liver Disease and Transplantation, Department of Surgery, Columbia University College of Physicians and Surgeons, New York, NY