Objectives: How transplant nephrectomy affects patient survival after return to dialysis is unclear. Here, we compared patient survival after graft loss between patients with and without transplant nephrectomy.
Materials and Methods: We divided 171 patients who received transplant between 2000 and 2015 and had graft loss into 3 groups: 64 had graft failure left in situ (without nephrectomy), 51 had nephrectomy < 3 months posttransplant (early nephrectomy), and 56 patients had nephrectomy > 3 months posttransplant (late nephrectomy). The primary endpoint was patient survival. Risk factors for patient death were also analyzed. Secondary endpoints included relisting for transplant and immunosuppressive agent status.
Results: Patient survival rates at 1, 3, and 5 years posttransplant in those without nephrectomy, early nephrectomy, and late nephrectomy were 92.1% /90.5%/86.6%, 96.0%/89.7%/80.4%, and 100.0% /97.9%/ 95.6%, respectively. Rates in patients with early nephrectomy differed significantly from those with late nephrectomy (P = .005). On multivariate analysis, patient survival was affected by relisting for transplant (hazard ratio 0.17; 95% confidence interval, 0.06-0.41; P < .001) and graft survival duration (hazard ratio 0.36, 95% confidence interval, 0.13-0.93; P = .036). Relisting for transplant occurred in 46.9% of patients without nephrectomy, 56.9% of patients with early nephrectomy, and 51.8% of patients with late nephrectomy. Those with late nephrectomy took 14.7 months after graft loss to relist for transplant, with 7.8 months for those without nephrectomy (P = .039) and 6.3 months for those with early nephrectomy (P = .051). Only 10.9% of those without nephrectomy were immunosuppressive free, which was in contrast to 94.1% and 78.6% of those with early and late nephrectomy, respectively.
Conclusions: After graft failure, patients without nephrectomy did not have inferior survival versus patients who received early or late nephrectomy. Graft survival time and relisting for transplant were associated with patient survival regardless of having transplant nephrectomy.
Key words : Graft failure, Kidney failure, Renal transplantation
The number of renal transplant patients who return to dialysis has been increasing. After graft loss, the nonfunctional graft is either left in situ or surgically removed, despite receiving dialysis or retransplant. Transplant nephrectomy (TNx) is increasingly considered in current clinical practice on the assumption that it improves patient outcomes after graft loss.
There is great variation for the indication of TNx in the literature, with overall incidence reported to be between 4.7% and 43.6%.1-11 Most TNxs are performed for early graft loss (within the first year) versus for late graft loss.12 Transplant nephrectomy has both positive and negative implications on the clinical outcomes of patients.13-15 Advantages include reduction of the cumulative dose of immunosup-pressive agents (ISAs) to control antibiotic/antiviral refractory infections and removal of the alleged chronic inflammatory stimulus, which can result in hypoalbuminemia, erythropoietin-resistant anemia, and elevated C-reactive protein levels.16 In addition, TNx may improve patient survival after graft loss.17 The disadvantages are potential anesthetic and surgical complications, including infection, vascular injury, and hemorrhage.1,3-5,7-11 These complications sometimes can be fatal. Moreover, TNx may have a negative effect on subsequent transplant outcomes and unmasking of HLA antibodies.15,18-20 However, how TNx, in the context of risk factors and surgical complications, affects patient survival has been poorly studied.
In this single-center study, our aim was to compare the clinical course and patient survival between those with TNx and those without TNx after graft loss. Furthermore, patients with TNx were subdivided into those with early TNx as emergency procedure and those with late TNx as elective treatment. The surgical complications of TNx and the risk factors that affected patient outcomes were also assessed.
Materials and Methods
From July 2000 to February 2015, 1314 renal transplants were performed in 1288 patients at The Royal London Hospital (London, UK). Of these transplants, 306 kidneys were lost because of either patient death or graft failure. This study analyzed patient prognosis after graft failure. Exclusion criteria included death with functioning graft (n = 123), preemptive transplant (n = 4), pancreas/kidney transplant (n = 1), and incomplete data (n = 7). After exclusion of these patients, 171 graft loss cases were included in this study. Cases were divided into 3 groups: 64 patients with failed graft left in situ (IS group), 51 patients with TNx < 3 months post-transplant (early TNx group), and 56 patients TNx ≥ 3 months posttransplant (late TNx group). All TNxs were performed at our hospital. The follow-up period was from transplant to patient death, subsequent transplant, or final observation. The median follow-up in all cases was 58.9 months (interquartile range [IQR], 31.1-100.0 mo). The primary endpoint was patient survival. Risk factors for patient death were also analyzed. Clinical course after graft loss, including duration until relisting for transplant, ISA status, and de novo disease, was compared between patient groups. In patients who received TNx, indication and surgical complications were analyzed.
This study was observational and had a longitudinal design. Clinical records were collected since perfor-mance of transplant. Demographic and medical records were obtained from our departmental database (FileMaker Pro 5.5v1; FileMaker Inc, Santa Clara, CA, USA), which was introduced to our unit in July 2000. Data included recipient age at the time of transplant, recipient sex and race/ethnicity, the number of transplants, donor age, donor sex, type of donor, HLA mismatch, type of ISA, cold ischemic time, delayed graft function, acute rejection, graft survival, and patient survival. Moreover, records of the interval from graft loss to relisting for transplant, final ISA status, the incidence of infection after graft loss, the incidence of cardiovascular disease (CVD) after graft loss, and retransplant were reviewed. Regarding data for those with TNx, the timing and indication of TNx, the interval between transplant and TNx, and complications related to TNx were collected. Delayed graft function was defined as requirement for dialysis in the first week after transplant. Graft failure was defined as return to permanent dialysis. Primary nonfunction was defined as requirement for dialysis for more than 12 weeks after transplant.
The immunosuppression regimen evolved per our departmental protocol. All transplant recipients received a calcineurin inhibitor (CNI), which was either cyclosporine or tacrolimus, and mycop-henolate mofetil or azathioprine, with a tapering dose of steroids. Induction therapy comprised antithymocyte globulin or monoclonal anti-CD25 antibody. Cyclosporine or tacrolimus was started on day 0. The dose was adjusted to maintain a cyclosporine trough level of 180 to 220 ng/mL and a tacrolimus trough level of 9 to 12 ng/mL in the first 3 months posttransplant; levels were tapered to 140 to 160 ng/mL and 6 to 8 ng/mL, respectively, beyond 4 to 6 months posttransplant. Steroid protocol included 20 mg/day of oral prednisolone, which was tapered to a maintenance dose of 5 mg/day at 3 months posttransplant. Graft failure within 3 months posttransplant was usually treated by TNx and the discontinuation of all ISAs. When graft failure occurred at 3 months posttransplant or later, mycophenolate mofetil or azathioprine was stopped immediately, in parallel with tapering the dose of CNI at 25% every month until complete withdrawal. Steroid use was withdrawn or discontinued if there was no medical requirement, such as adrenal insufficiency syndrome or planned subsequent transplant.
Indication for TNx was individualized. Transplant nephrectomy was routinely performed for a graft that had failed within 1 month posttransplant. A graft that failed after 1 month posttransplant was removed only if patients had any symptoms or any medical requirements. The clinical indications for TNx were as follows: (1) acute graft loss, (2) suspected ongoing rejection or graft intolerance syndrome (GIS), (3) infection, and (4) miscellaneous reasons. Acute graft loss was defined as insignificant blood flow due to circulatory disorder or vascular complications within 3 months posttransplant. Graft intolerance syndrome was associated with tenderness around graft, hematuria, fever, and chronic infla-mmatory response without infection. The chronic inflammatory response included hypoalbuminemia, erythropoietin resistant anemia, and elevated C-reactive protein levels in the patients. The infection group included TNx due to bacterial, viral, or fungal infection diagnosed by positive culture and/or positive serology. Surgical approach of TNx was normally extracapsular for the early TNx group or intracapsular for the late TNx due to adhesion around the graft.
All values are shown as means and standard deviation (SD), median (IQR), or percentage. Continuous variables were compared using t test, and categorical variables were analyzed using chi-square or Fisher exact test. Quantitative variables were compared using the Mann-Whitney nonpa-rametric test or Steel-Dwass test. Multiple com-parisons were performed by Kruskal-Wallis nonparametric or factorial analysis of variance. Patient survival rates in the 3 groups were calculated by the Kaplan-Meier method and compared using the log-rank test. Univariate and multivariate associations with patient death were analyzed using Cox proportional hazard regression and estimated hazard ratio (HR) with 95% confidence intervals (CI). Factors with P < .05 at univariate analysis were added into the multivariate analysis. The differences were considered statistically significant at P < .05. We used JMP version 12 software (SAS Institute Inc., Cary, NC, USA) for all statistical analyses.
Patients without TNx were older than those with late TNx (43.7 vs 38.4 y; P = .040) (Table 1). However, recipient sex, race/ethnicity, number of transplants, donor age, donor sex, type of donor, and CNI type did not differ among groups. There was a significant difference in HLA mismatch between those without TNx and those with late TNx (2.67 vs 3.43; P = .016). Although there was no difference in cold ischemic time among the 3 groups, as expected, the early TNx group had the highest incidence of delayed graft function (88.2%), whereas those without TNx and late TNx had incidences of 43.8% and 62.5%, respectively. The incidence of acute rejection in the early TNx group was 5.9% because of short graft survival. On the other hand, 51.8% in the late TNx group experienced acute rejection, which was approximately two-fold higher than the group without TNx (23.4%). The most common type of acute rejection was T-cell-mediated rejection and antibody-mediated rejection (AMR), which occurred in 2 patients in the IS group and in 1 patient in the late TNx group. T-cell-mediated rejection with AMR was observed in 7 patients in the late TNx group. Follow-up duration in the early TNx group was significantly shorter (30.4 mo) than in those in the IS and late TNx groups (76.2 and 73.4 months, respectively; P < .001). Graft survival duration (IQR) in the IS group, early TNx group, and late TNx group was 51.1 (24.4-77.8), 0.06 (0.03-0.3), and 21.0 (6.8-45.6) months (all, P < .001), respectively.
Clinical course after graft loss
Relisting for transplant was possible in 46.9% of the IS group, 56.9% of the early TNx group, and 51.8% of the late TNx group (Table 2). The late TNx group took 14.7 months (IQR, 6.9-37.3 mo) after graft loss to relist, which was significantly longer than shown for the IS group (7.8 mo; P = .039). At the end of this study, 94.1% in the early TNx group and 78.6% in the late TNx group were ISA free compared with 10.9% in the IS group. There were no significant differences in the incidence of infection and CVD after graft failure between the 3 groups. Each group had similar rates of subsequent transplant.
Patient survival and risk factors
The overall patient survival rates at 1, 3, and 5 years posttransplant were 95.8%, 93.8%, and 88.7%, respectively. Patient survival rates at 1, 3, and 5 years posttransplant in the IS group were 92.1%, 90.5%, and 86.6% (Figure 1), which was similar to the other 2 groups (vs early TNx, P = .193; vs late TNx, P = .215). However, patient survival rates at 1, 3, and 5 years in the early TNx group were significantly inferior versus those in the late TNx group (96.0%, 89.7%, 80.4% vs 100.0%, 97.9%, 95.6%, respectively; P = .005).
Univariate analysis suggested that patient survival outcome was influenced by patient age at the time of transplant, the duration of graft survival, age at the time of graft loss, and duration before relisting for transplant (Figure 2). Older age at the time of transplant increased the risk of patient death (HR = 1.06; 95% CI, 1.03-1.09; P < .001). Similarly, older age at the time of graft loss increased the risk of death (HR = 1.04; 95% CI, 1.01-1.07; P = .008). However, longer duration of graft survival reduced the risk (HR = 0.78; 95% CI, 0.89-1.27; P < .001). Relisting for transplant strongly improved patient survival (HR = 0.13; 95% CI, 0.05-0.33; P < .001). Although having an in situ nonfunctioning graft was not associated with patient survival, early TNx (< 3 mo posttransplant) was associated with elevated mortality risk (HR = 2.42; 95% CI, 1.02-5.29; P = .044). Interestingly comorbidities, including CVD, surgical complications of TNx, and infection, had no significant association with patient outcome. However, on multivariate analysis, only 2 factors were significantly correlated with patient outcome: relisting for transplant (HR = 0.17; 95% CI, 0.06-0.41; P < .001) and longer duration of graft survival, which resulted in better patient survival (HR = 0.36; 95% CI, 0.13-0.93; P = .036).
Indication for transplant nephrectomy
A total of 51 patients received TNx due to acute graft loss (Table 3). This indication was common for nearly all cases (96%) in the early TNx group. Acute graft loss was further subdivided into 3 reasons: acute circulatory disorder, postoperative bleeding, and surgical technical failure. Vascular thrombosis (66.7%) as acute circulatory disorder was the major indication in the early TNx group, whereas only 5 patients in the late TNx group received TNx due to primary nonfunction. Primary nonfunction sometimes needed prolonged observation until the graft would show adequate graft function after acute tubular necrosis. Transplant nephrectomy due to rejection/GIS was performed to remove the symptomatic failed graft and to withdraw ISA. Fifteen patients received TNx because of refractory infection. Only 1 patient with poor graft function in the early TNx group required TNx to control urinary tract infection. On the other hand, infection was the second most common reason of TNx in the late TNx group (25.0%). In the late TNx groups, TNx was due to urinary tract infection in 9, cytomegalovirus infection in 2, BK virus infection in 2, and infected lymphocele in 1 patient. Of these 14 late TNx patients, 5 had severe graft dysfunction due to infection and the other 9 were already on established dialysis before TNx. There were a total of 5 cases in the miscellaneous reason category. Reasons catego-rized as miscellaneous for TNx included malignancy in 3 patients, with 1 patient in the early TNx group related to donor malignancy (with day 1 TNx after detection of deceased donor gallbladder cancer). The other 2 recipients were in the late TNx group; 1 patient with Kaposi sarcoma and impaired graft function underwent TNx to discontinue ISA to manage the malignancy. The remaining 2 patients in the miscellaneous category were in the late TNx group, which was because of recurrent focal segmental glomerular sclerosis and preparation for subsequent transplant.
Complications of transplant nephrectomy
All surgical complications of TNx are shown in Table 4. Of 107 patients who received TNx, 35 patients (32.7%) had 1 or more surgical complication. The most common complication was wound infection in 13.1%, postoperative fluid collection in 12.1%, and hematoma in 8.4%. In the early TNx group, vascular repair was needed because of the short interval from transplant to TNx. On the other hand, pulmonary complications were observed in only the late TNx group. Two patients in the early TNx group died within 2 months after TNx. The causes of death were cardiac arrest at 30 days after TNx and massive upper gastric intestinal bleeding at 50 days after TNx. The overall incidence of surgical complications was 39.2% in the early TNx group and 26.8% in the late TNx group (P = .171). Blood transfusion was needed for 46 patients (43.0%) during the perioperative period. The early TNx group had a higher rate of blood transfusion than the late TNx group (47.1% vs 39.3%; P = .417).
In this single-center study, the clinical course of renal transplant recipients after graft failure was analyzed. After graft failure, we followed 3 different groups with failed grafts (graft left in situ, early TNx, and late TNx), and we observed different outcomes in each group. Regarding reasons for TNx, early TNx was mostly an unavoidable procedure because of acute vascular thrombosis. Patients who underwent late TNx had a longer time to be relisted for a subsequent transplant. Interestingly, these patients had better patient survival rate than those with early TNx. In general, TNx was associated with high incidence of morbidity, with patients often requiring blood transfusions. Although surgical complications can be avoided with an in situ nonfunctioning graft, patients have the burden of remaining on ISAs.
The absolute indication and optimal timing of TNx still remains controversial, and there is no standard approach. It is not clear whether TNx is appropriate for all patients. The number of TNx procedures has decreased with the introduction of cyclosporine compared with before its introduction.10 Early reports have demonstrated that the major indication for TNx was acute rejection (65.3%-80.9%).1,2 The rate of TNx for chronic rejection is reported to be 3.2% to 57.1%, and this has increased with the introduction of cyclosporine.4,5,8-11 Modern advanced ISAs have reduced rejection and allowed longer graft survival, and the indication for TNx seems to have changed over the past 3 decades. In our study, most early TNx occurred because of acute circulatory disorders, and this procedure was performed to avoid potential future serious morbidity despite symptoms related to a failed graft. Thus, urgent, nonelective TNx is required for early graft loss. We had no TNx procedures due to uncontrollable acute rejection and chronic rejection before graft loss. For late TNx, the procedure is usually needed because of possible rejection/GIS due to decreasing ISAs. Another reason for late TNx is infection due to prolonged ISA, which has a reported incidence of up to 3.7%.5,8-11 In contrast, our study showed a relatively high incidence (13.9%), and persisting infection occurred even beyond those weaned of ISAs. We observed a different clinical course before graft failure between the late TNx and the IS group, with the late TNx group having a higher incidence of acute rejection (both T-cell-mediated rejection and AMR) and needing more ISA. As a result, TNx as a salvage therapy would be required for such intractable rejection/GIS and infection in the late TNx group.
Transplant nephrectomy is a two-edged sword for patients returning to dialysis.13,14 The advantage includes withdrawal of ISAs. Transplant nephrec-tomy also improves the chronic inflammatory response and improves survival of patients who return to dialysis.16,17 On the other hand, there are disadvantages related to loss of residual renal function and the exposure to surgical complications. Overall, TNx is consider a procedure with high risk because of uremia, previous exposure of ISA, and difficulty of the surgery.13 Rate of complications has been shown to range from 9.9% to 60.0%, perhaps due to differences in the definition of “complication.”1-7,9,10 The mortality rate has been reported to be 0.9% to 38.7%.1-5,7,9,10 Setin and associates showed that the total incidence of complications was 48.3%, and urgent TNx had twice the risk of surgical complications than nonurgent TNx.10 Similarly, late TNx (> 1 year posttransplant) had a high incidence of surgical complications and longer hospital stays, as reported by Freitas and associates in a single-center study.21 On the contrary, surgical complication rates were 22.6% with immediate TNx (≤ 2 weeks post-transplant) and 16.1% with delayed TNx (> 2 weeks posttransplant) (P > .05).3 We observed no significant differences in surgical complications between early and late TNx.
Transplant nephrectomy is known as a risk factor for sensitization,22 and a failed graft acts as an anti-HLA antibody sponge, with circulation of antibody increasing after TNx by yet unknown mechanisms. Blood transfusions related to TNx may also add to a further risk of sensitization. Blood transfusion for TNx is frequently needed9 because of preoperative anemia, which occurs because of a chronic inflammatory response in the retained graft and/or renal failure. Moreover, the surgical procedure in a failed graft is often difficult due to adhesions caused by an ongoing inflammatory process, such as rejection and infection, around the graft. Mazzucchi and associates revealed that blood loss and surgical complication rates in late TNx (≥ 2 months posttransplant) were higher than in early TNx (< 2 months posttransplant).9 They found that 44.3% of 70 patients required blood transfusions. However, no significant differences in need for blood transfusion were shown between early and late TNx, which were the same as our results. We have found that early and late TNx were more likely to have long intervals from previous transplant to subsequent transplant due to increasing anti-HLA antibodies.
Withdrawal of ISAs after graft loss is based on the concept for minimal usage. Excess use of ISAs may cause adverse effects, and underuse may trigger acute rejection. Therefore, adjustments in ISAs are needed as long as a failed graft is retained. Smak Gregoor and associates compared between incidence of graft failure in 197 patients with and without low-dose ISA.23 Although low-dose ISA after return to dialysis was associated with patient mortality from both infection and CVD, the incidence of acute rejection with low-dose ISA was comparative to that with discontinued ISA. The ideal clinical management after graft loss is for no symptoms in the failed graft after ISA withdrawal. Although nearly 90% of those in our study on ISAs were on prolonged treatment, we did not find a higher incidence of comorbidities, such as infection and CVD.
Outcomes of TNx have been reported by a number of studies.12,17,18,24,25 In one large study consisting of over 19 000 patients from the United States Renal Data System (USRDS) by Johnson and associates,12 TNx after early graft loss (< 12 mo) was associated with an increased risk of death (HR = 1.13; 95% CI, 1.01-1.26), whereas TNx after late graft loss (≥ 12 mo) was associated with a decreased risk of death (HR = 0.89; 95% CI, 0.83-0.95). They did not clearly explain the reason for this paradoxical finding. In another large study using USRDS, Ayus and colleagues showed the effects of TNx on patient survival after return to dialysis.17 They found that TNx was associated with a 32% reduction in the relative rate of death compared with not receiving TNx. In single-center studies,18,24,25 TNx did not result in inferior patient survival compared with not having TNx. In our study, on multivariate analysis, TNx was not associated with better patient survival, although long-term graft survival and relist for transplant influenced patient survival. Short-term graft survival negatively influenced patient survival in those who had early TNx. On the other hand, patients with late TNx had shorter graft survival than those who had graft left in situ, although patient survival was similar between these groups. Interestingly, graft survival improved when the patient was deemed fit to be relisted for a new kidney transplant, and this represented the most significant factor on our multivariate analysis. Although graft survival was short, relisting was the most important factor affecting patient survival.
There are clearly some limitations in our study, including it being a single-center observational study and having a small sample size. The performance of TNx was based on the individual’s medical condition and the center’s decision and not on randomized selection. We could not collect donor-specific antibodies after graft loss, including before and after TNx unless the patient was relisted. We did not analyze the ISA dosage after graft loss due to lack of data. As a result, any associations between the amount of ISA after graft loss and the symptoms related to the failed graft could not be determined.
In conclusion, our study found that patients with a failed graft did not have inferior survival compared with patients who received early or late TNx. However, graft survival and relisting for new transplant were associated with better patient survival regardless of TNx. Transplant nephrectomy still remains a necessary treatment after graft loss in selected cases to improve patient outcomes. It is important to emphasize that TNx carries significant risk of surgical morbidity, including the need for blood transfusions. Advantages and disadvantages of TNx need to be fully considered. Prospective, multicenter, large studies are recommended to further analyze the association between TNx and patient outcomes.
DOI : 10.6002/ect.2018.0233
From the 1Nephrology and Transplantation Department, The Royal London Hospital,
London, United Kingdom; the 2Nephrology Department, Toho University Faculty of
Medicine, Tokyo, Japan; and the 3Cellular Pathology Department, The Royal London
Hospital, London, United Kingdom
Acknowledgements: The authors have no sources of funding for this study and have no conflicts of interest to declare. *Magdi Yaqoob and Carmelo Puliatti contributed equally to this work.
Corresponding author: Masaki Muramatsu, Nephrology Department, Toho University Faculty of Medicine, 6-11-1, Omori-nishi, Ota-ku, Tokyo 143-8541, Japan
Phone: +81 3 3762 4151
Table 1. Patient Characteristics
Table 2. Clinical Course After Graft Loss
Table 3. Indication for Transplant Nephrectomy
Table 4. Complications of Transplant Nephrectomy
Figure 1. Patient Survival Rate
Figure 2. Risk Factors for Patient Survival