It has now been 50 years since the first pediatric liver transplant, and the success of these transplants—patient and graft survival—still largely depends on our ability to maintain artery patency. Although many interventions, medical and technical, have been implemented, the truth remains that a high percentage of arteries continue to become occluded and thus compromise patient outcomes. Early thrombosis often leads to graft compromise and failure, and this requires reoperation and retransplantation. The prevalence and clinical relevance of late hepatic artery thrombosis (HAT) remain unclear.
In this issue of Liver Transplantation, Kivelä et al. present a cross-sectional study of 34 liver transplant patients less than 18 years old over a 10-year period (starting in 1987) with a median follow-up of 9.5 years (interquartile range = 4.0-16.4 years). Ninety-nine pediatric patients originally underwent transplantation in that time frame. Thirty-two percent of these patients were deceased. Thirty-four of the surviving 66 patients were available to undergo magnetic resonance imaging (MRI) with contrast for an evaluation of HAT. Interestingly, they report that 44% of these 34 patients had evidence of late arterial thrombosis, with late thrombosis defined as thrombosis 1 month after the date of transplantation; this was much higher than the previously reported rates (1.2%-2.7%).[3-6] As the discussion makes clear, this could represent 23% of patients in the best case scenario and 71% in the worst case scenario because only 52% of the whole population underwent imaging. By comparing the MRI findings with ultrasonography, laboratory, biopsy, and clinical data, the authors have done a comprehensive job of analyzing these patients at time points surrounding MRI. They bring attention to the possibility that late HAT is wildly underdiagnosed during the standard follow-up of pediatric patients. Our 2 most pressing concerns regarding these data are (1) whether the late HAT rate of 44% is accurate for long-term pediatric transplant survivors and (2) what the clinical application of this knowledge should be as we go forward.
One of the biggest concerns in accepting this high percentage for late thrombosis is that the MRI technique described in the article's methodology was not consistently angiographic. Without angiography, it may be hard to draw convincing conclusions about HAT, especially without any indication of flow issues on a similarly timed ultrasound examination. It is well known that ultrasound is not the gold standard for the detection of HAT; however, changes in flow can suggest a need for a closer investigation. It is surprising that there was no suggestion of intrahepatic changes in flow on ultrasound for these patients. In addition, if we agree that ultrasound accuracy is as poor as suggested at this institution (40% sensitivity), it is possible that a percentage of these patients had unrecognized early HAT if ultrasound was the only imaging performed in the early posttransplant period.
For comparison, in the year 2008 at our institution, Lucile Packard Children's Hospital at Stanford, 40 transplants were performed in children less than 18 years of age (43% were <1 year of age). Early HAT was diagnosed in 1 child, who required retransplantation (2.5%). The 5-year graft survival rate for this cohort was 92.5%. There have been no patients diagnosed with late HAT so far. In light of Kivelä et al.'s study, it might be assumed that late HAT has been severely underdiagnosed at our institution, presumably because, just as at other centers, computed tomography and magnetic resonance angiography imaging is not routinely performed for arterial evaluation. However, according to a review, 19 of these patients underwent computed tomography with contrast (37%) or MRI with contrast (63%). The reason for obtaining the images was cancer surveillance, hepatoblastoma or posttransplant lymphoproliferative disorder, biliary tract evaluation, or something else. All of these images were taken 1 month or more after transplantation, and there was documentation of hepatic artery patency for all of these images except for 1 image for which the contrast timing did not allow comment. Two other patients underwent computed tomography scans with documentation of artery patency, but the scans were performed within the first month after transplantation. Either there is a vast difference in imaging sensitivity between our centers or a difference in the management of HAT prevention or late thrombosis is occurring primarily after 5 years of follow-up. Although the variation did not reach statistical significance, the median time to follow-up MRI after transplantation in Kivelä et al.'s study was 5.8 years for those without late HAT and 15.1 years for those with late HAT; this suggests that late thrombosis is more prevalent 5 years after transplantation. Conversely, the adult transplant literature shows that the median time to late thrombosis is 6 months.
This article suggests that MRI might be the right modality for monitoring these patients in the long run. We agree that magnetic resonance angiography (not MRI) may be the best modality for analyzing artery patency in children because of its accuracy and minimal radiation exposure; however, the question remains whether routine monitoring should be done at all even if we accept a 44% rate of late HAT. Kivelä et al. have shown that patients with late thrombosis have no significant differences in histological or enzymatic markers. This argues against any form of more stringent monitoring. As they point out in their discussion, the long-term outcomes of late HAT are not well defined. In addition, in contrast to acute arterial thrombosis, when late HAT is currently diagnosed, there is no intervention that can change its clinical course. There is also no imaging modality that can suggest a clinical course or prognosis once the diagnosis is made. The identification of collateral vessels may provide reassurance that arterial flow is maintained in a graft, but this has not been proven to prevent biliary complications in the future, nor does having late HAT guarantee biliary complications to come. With increasing health care costs, it is hard to argue that an imaging modality significantly more costly than both computed tomography and ultrasound should be routinely used to identify an entity without a treatment or prognosis.
Notably, Kivelä et al. have brought to our attention that the incidence of late HAT may be significantly higher than expected. In the future, determining an accurate incidence of late HAT, the time of thrombosis, and the clinical impact could be crucial to improving long-term graft survival. With advances in immunosuppression and surgical techniques, the long-term outcomes of pediatric liver transplantation are improving, and new ideas for prolonging graft survival need to be constantly considered. Especially for our pediatric patients, who have the most potential for longer life expectancies, we need to aim for graft survival that exceeds our current expectations. However, we will not know the real impact of late HAT or the need for surveillance until its long-term outcomes are studied. The population of Kivelä et al.'s study with reproduced image findings may provide invaluable data regarding late HAT's long-term influence on the field.