- Top of page
- Patients and methods
The use of sirolimus (SRL) in orthotopic liver transplantation (OLT) has been controversial after experimental data suggested an increased risk of hepatic artery thrombosis (HAT). To assess the safety and efficacy of SRL as de novo immunosuppression in OLT recipients. Outcomes of 252 OLT patients who received SRL were compared with outcomes of 291 OLT recipients who received calcineurin inhibitor in a retrospective study. Primary outcomes of this study were: patient- and graft survivals, vascular, biliary, wound complications and rejection rates. Secondary outcomes were: postoperative infection rate, bone marrow and renal function and changes of lipid levels. Patient- and graft survivals, rejection and infection rates were similar. In the SRL group, HAT occurred in 1.2%, biliary complications in 19.4%, and incisional hernias in 9.1%. In the control group the incidence of HAT was 5.8% (P = 0.004), biliary complications 18.5% (P = NS) and incisional hernias 7.2% (P = NS). Patients on SRL experienced significantly higher levels of serum triglycerides but fewer acute cellular rejections. Bone marrow and renal functions were similar in both the groups. Our findings would suggest that SRL is safe and effective for very selected OLT recipients. Randomized controlled trials are necessary to confirm our results.
- Top of page
- Patients and methods
The use of SRL in OLT patients has not been approved in the United States because a phase II controlled study has suggested an increased risk of HAT, graft loss and patient death. Nevertheless, up to 15% of OLT recipients in the USA receive SRL within the first year after surgery as it is often used as a CNI-sparing agent, especially in cases of CNI toxicity . Currently, the experience of using SRL in OLT recipients early after surgery seems limited to a few transplant centers and there is only a modest body of literature on its effectiveness and safety as a de novo immunosuppression medication in OLT recipients.
The primary aim of this study was to assess the outcomes of a large cohort of recipients who received SRL in the first 24 h after OLT and continued for at least 6 months afterwards. The comparison of the results was made with a control group of patients treated with standard CNI therapy at the same centers and during the same period. To our knowledge, this is the largest observational study assessing the outcomes of de novo SRL immunosuppression after OLT and most likely it would not be repeated after the FDA issued the black box label [7,13,14,28,29]. Watson  et al. were the first to describe the use of SRL as primary immunosuppression in OLT. Chang et al.subsequently explored the feasibility of converting stable OLT recipients affected by CNI toxicity from CyA or TAC to SRL and reported no significant adverse events in the process . Other authors described similar experiences and observed benefits of using SRL such as prevention of renal dysfunction, reduced neurotoxicity and reduced steroid-resistant allograft rejections [30–33]. Nevertheless, the experience with SRL in OLT recipients in many transplant centers has been modest [34,35] with the majority of programs in the United States, Europe and other countries viewing the use of CNIs as essential for the success of OLT [36,37]. TAC, and less so CyA, are the most commonly used CNI immunosuppressant drugs for solid organ transplant recipients and provide excellent graft- and patient survival rates . However, renal dysfunction or failure , hypertension, diabetes, hyperlipidemia and osteoporosis are major side-effects directly related to the time of exposure and blood concentration of these drugs [39–41].
At the University of Alberta and the University of Colorado, the use of SRL was implemented in a substantial number of patients undergoing OLT prior to the FDA black box label and in selected patients thereafter to reduce side-effects related to CNI therapy, or as an antitumor agent in those with hepatocellular carcinoma [28,42,43]. As SRL has a different safety profile than CNI, the possible advantages of its use must be weighed against a range of other side-effects such as peripheral swelling, joint pain, wound infections, hyperlipidemia, oral and gastrointestinal ulcerations , dermatitis, interstitial pneumonitis [45–47] and possibly vascular thrombosis.
In our experience, patients on SRL had survival and rejection rates, graft functions and infection rates comparable to the control group and confirmed the results of other previous smaller observational studies [14,34,48,49]. One of the most controversial aspects of using SRL in OLT recipients is the potential higher risk of graft loss because of HAT . Our findings did not confirm an increased thrombotic risk for these patients and supported the results reported by other investigators who used SRL after OLT [14,28,50]. It is known that HAT complicates 4–15% of OLT and occurs more frequently after pediatric OLT [51,52]. Several technical and congenital conditions have been found to increase the risks of HAT: dissection of the hepatic arterial wall, technical imperfections with the anastomosis, celiac artery stenosis, hypercoagulable state, transplantation for primary sclerosing cholangitis, aberrant arterial anatomy, back table arterial reconstruction of the allograft, and high-resistance microvascular arterial outflow caused by rejection or severe ischemia-reperfusion injury (IRI) [53–56]. In our experience, thrombotic vascular complications occurred in 5.5% of patients on SRL and in 9.2% of patients on CNI (P = NS). Contrary to our expectations, HAT occurred in 1.2% of patients on SRL and in 5.8% of patients on CNI (P = 0.004). One of the possible explanations of these findings is that the group treated with CNI had a significantly higher percentage of patients with autoimmune diseases that are well known to be associated with hypercoagulability and risk of thrombosis. Unfortunately, the small number of patients who developed this complication in each group did not allow us to perform any further statistical analysis to explore this hypothesis further .
Over the last two decades, patient- and graft survival rates have improved significantly. Recent literature reports that 75–85% of all individuals undergoing OLT are alive at 5 years independent of the immunosuppression regimen used [57–59]. The longer life-expectancy of these patients demand a careful evaluation of the spectrum of all the long-term side-effects of immunosuppression as they have become the main cause of death with functioning grafts. Epidemiological studies have shown that chronic kidney disease is a predisposing factor for higher morbidity and mortality in transplant recipients [60,61] as well as in the general population . Therefore, immunosuppressive medications without renal toxicity are very attractive. Published clinical studies indicate improved renal function after conversion from CNI to SRL in the first 1–6 months after renal or OLT [33,63,64]. In our study, both groups experienced improved renal function during the first year after transplant and relative stability during the following years possibly because of the resolution of preoperative hepato-renal syndrome. It is important to note that these outcomes were seen in patients transplanted in the pre-MELD era, a time when renal dysfunction was much less common in OLT recipients, and perhaps partially explaining the excellent long-term renal function observed in both groups of patients.
Sirolimus (SRL) has been associated with a negative effect on wound healing [65–68] because of its antifibrotic effects . As in previous reports [70,71], we observed a rather high incidence of incisional hernias requiring surgical revision associated with the use of SRL. Severe perioperative wound infections occurred in 5.8% of patients and wound dehiscences in 0.8%. Symptomatic incisional hernias requiring surgical repair occurred in 50 of 252 patients (19%) during the first 5-year post OLT vs. 14% in the control group. At both transplant centers, it was felt that there was no need to discontinue SRL therapy in the pre- and postoperative period for incisional hernia repairs. Although this difference did not reach statistical power, the results of this study support the finding of other authors who have reported up to 15–17% of patients with incisional hernias requiring surgical therapy after OLT [34,51,52] and some delayed wound healing with SRL .
There is limited knowledge regarding the incidence of bile duct complications because of poor healing in patients undergoing OLT and treated with SRL. In this study, the rate of bile duct anastomotic complications was 19.4%, which is comparable to patients treated with CNI . More importantly, few of the adverse events involving the bile duct anastomosis had long-term effects that contributed to graft failure.
In contrast to the previous clinical trials of OLT recipients treated with SRL-based immunosuppression, we saw no increased risk of infection in this group. Overall, the rate of infection in patients on SRL was quite low: bacteremia 11%, pneumonia 10.3%, fungal infection 0.4%, CMV 2.6%, and PTLD 0.4%. None was higher than the CNI-treated control group, nor do the infections rates appear higher in comparison to published experience . These findings might be attributable to the lower level of SRL used at our transplant centers, or the combination of lower SRL blood levels and the fact that other immunosuppressive medications such as steroids were tapered relatively quickly. Contrary to other observational studies of solid organ transplant recipients where the incidence of SRL-induced pneumonitis ranged from 2%  to 11% [73,74], we did not observe any case. We suspect that this might be as a result of several possibilities. One possible explanation is attributable to the difficulty of diagnosing interstitial pneumonitis resulting from SRL in the early postoperative period as the clinical and radiological presentations are similar to infective pneumonia that occurs much more frequently . Another explanation might be that during the period of this study, the association between exposure to SRL and interstitial pneumonitis had not yet become established and therefore the clinical diagnosis of this condition might have been misinterpreted as infective pneumonia.
Dyslipidemia is perhaps the most common metabolic side-effect of SRL and it may not be dose-dependent . Clinical studies in renal transplantation have demonstrated that up to 80% of patients on SRL have hypercholesterolemia [77,78]. Similar to the renal transplant literature, in OLT recipients treated with SRL, the incidence of hypercholesterolemia has been reported to be as high as 44% . In comparison to previous studies, the lipid profile of our patients showed modest levels of hyperlipidemia that might have been attributable to the combination of maintaining lower blood levels of SRL, and/or the reduced use of steroids in addition to the introduction of statins and/or fibrate medications for all patients with elevated serum lipid profile.
Sirolimus (SRL)-associated bone marrow suppression is attributable to inhibition of specific cytokines and vascular growth factors [68,79,80]. In our study, SRL did not significantly affect platelet or white blood counts; they remained in the range of 160–190 000/mm3 and 4000–6000/mm3 respectively over time without frequent need for granulocyte stimulating growth factor injection. Similarly, the level of hematocrit remained quite stable (31–40%) although oral iron supplementation was prescribed for all patients who had shown evidence of suppressed erythropoietic function. Anemia was less commonly a problem in Denver, likely a result of the altitude-induced higher baseline level of hematocrit.
As in many retrospective observational studies, there are several inherent weaknesses of this study mostly attributable to the lack of randomization of patients. Although the immunosuppression protocols used for subjects treated with SRL were similar at both medical centers, selection and treatment bias could not be avoided without random allocation. As in every retrospective analysis there is the risk of introducing sampling, selection and other bias as the groups of patients assembled for the study differ in ways other than the factors under investigation. Although, no significant baseline differences were seen between recipients on CNI-based immunosuppression or SRL-based one, there was heterogeneity in etiology of cirrhosis, pre- and postoperative care provided at the two participating medical centers. For example, in both centers, the protocols employed for the use of SRL were started early after OLT and continued for at least six consecutive months. On the other hand, the use of steroids and the blood level of immunosuppressive medications were not uniform. Another important limitation of this study is that the majority of patients were enrolled prior to the introduction of the MELD scoring system for the allocation of cadaveric grafts. The mean MELD score at the time of transplantation for patients enrolled in this study was only 15, this being the average MELD score for patients transplanted in the pre-MELD era and this value is significantly lower than the average score of patients undergoing OLT in recent years [81,82]. Higher MELD scores are associated with more advanced liver and other organs dysfunction. Therefore, the incidence of pre- and post-transplant renal insufficiency, infections, and wound healing problems observed in these two cohorts may apply only to patients with relatively preserved hepatic and renal function.
Despite these limitations, our study has the strength of being one of the largest observational studies on the use of SRL in OLT recipients treated in North America. After the black box label, the experience of SRL in OLT has been quite modest and most likely a similar study will not be feasible any time in the near future. Our findings can not overturn the results of randomized trials, but they suggest that for patients with moderate MELD scores at the time of OLT, the use of SRL alone or in conjunction with low-dose CNI may be safe and effective as rejection rates, graft losses and patient survivals were similar in both the groups. Nevertheless, we recognize the limitations of our study, and wish that new randomized controlled trials are performed to test the effectiveness and safety of promising new m-TOR inhibitors for OLT recipients.