Comparison of Patient Outcome According to Renal Replacement Modality after Renal Allograft Failure

Abstract

The aim of this study is to investigate the clinical course of patients with failed allograft according to the type of renal replacement modality. Three hundred sixty-eight patients with failed allograft were included. Of these, 233 patients started hemodialysis (HD-PSKT), 64 patients started peritoneal dialysis (PD-PSKT), and 71 patients underwent second transplantation (ReKT). At baseline, age, sex, laboratory findings, and comorbidity did not differ significantly among three groups. Chronic rejection was the most common cause of allograft failure (81.6%) followed by acute rejection (10.7%). During the observation period, 96 patients died. The most common cause of death was cardiovascular disease (39.6%) followed by infection (34.4%) and malignancy (8.3%). Infection was important cause of death within 10 years from allograft failure, but cardiovascular disease and malignancy occupied significant portion of death after 10 years from allograft failure. Significant difference was not found among the three groups in the cause of allograft failure and the cause of death. The patient outcome was better in the ReKT than in the other two groups and it did not differ significantly between the PD-PSKT and HD-PSKT. In multivariate analysis, old age, hypoalbuminemia, and high comorbidity were proved to be the independent risk factors for mortality and the ReKT was still significantly superior to the HD-PSKT and PD-PSKT after adjustment for other confounding factors. In conclusion, second transplantation may result in survival benefit, and proper management of nutrition and comorbidity may help to improve outcome in patients with failed allograft.

• hemodialysis
• allograft failure
• kidney transplantation
• peritoneal dialysis

INTRODUCTION

With the increase of end-stage renal disease (ESRD) patient population, the number of kidney transplantation (KT) has also markedly increased. Although long-term allograft survival has improved over the years, a single transplant seems not to be sufficient for many patients. Thus, many patients need to return to dialysis treatment after allograft failure.

According to a European study, the patients restarting dialysis after allograft failure accounted for 4% of the total previously transplanted patients in 1995.¹ Another report showed that 4.8% of the patients starting dialysis in 2004 had already undergone transplantation and 20% of the patients on the kidney transplant waiting list have previously had a failed renal allograft in the United States.² In Korea, the cumulative KT number is 10,119 from 1969 to 2007.³ However, it is estimated that a significant portion of these patients have suffered from allograft failure. Thus, a considerable portion of newly developed ESRD patients may come from transplantation because of allograft failure.

It is postulated that transplant-associated conditions, for example, immune suppression and alloimmune reaction to allograft, can affect the clinical course of the patients after allograft failure. Therefore, they should be regarded as a separated patient's population from transplant-naïve ESRD patients.^4,5 However, data on the patient outcome after allograft failure are limited and confounding.⁶ In this regard, we evaluated the clinical course of patients with failed allograft and assessed the clinical outcome according to renal replacement modality (RRT).

SUBJECTS AND METHODS

From March 1969 to December 2008, 1690 cases of KT have been performed in Seoul St. Mary's Hospital. Allograft failures were confirmed in 640 cases. Inclusion criteria were patients with allograft failure and who survived for more than 3 months after allograft failure. Two hundred and thirty-three patients were excluded because they died with a functioning graft or died within 3 months after allograft failure. Thirty-nine patients with second allograft failure were also excluded to exclude the bias induced by previous transplant history; hence, 368 patients were included in our study. We retrospectively reviewed the medical records of those patients. This study was approved by the Institutional Review Board of Seoul St. Mary's Hospital.

Public health insurance system developed in 1978 in Korea. In this study, only four patients took KT before 1978; hence all patients except them took medical care for dialysis or KT through public health insurance. We divided the patient population into three RRT groups, namely, hemodialysis (HD-PSKT), peritoneal dialysis (PD-PSKT), and second transplantation (ReKT) group based on the type of RRT performed after allograft failure. All patients who took second transplantation during observation period were classified as ReKT group irrespective of the duration from allograft failure to second transplantation. All patients continued to take immune suppressant for several months with reduced dose after allograft failure. In detail, we maintained cyclosporine 50–100 mg or tacrolimus 1–2 mg and prednisone 5–10 mg and stopped it at 3–6 months after allograft failure.

Baseline characteristics such as gender and age at allograft failure were collected. Allograft failure was defined as to start dialysis or take second transplantation. We investigated the duration in which the patients were on dialysis and the dialysis modality before KT, the duration of primary functioning allograft, and the cause of allograft failure. For the patients who died during the follow-up period, the cause of death was investigated. We also collected the laboratory results such as hemoglobin (g/dL), serum creatinine (mg/dL), and serum albumin (g/dL) level within 5 days prior to starting dialysis or second transplantation. Combined diseases at allograft failure were investigated and those were used to assess the comorbidity by using the Stoke study scoring system (Davies Score).^7,8 We compared the variables mentioned above among three RRT groups (HD-PSKT, PD-PSKT, ReKT).

In this study, the patient's survival duration was measured from the date of allograft failure to the date of patient death, loss of follow-up, conversion to another modality, and the end of the observation period (1 April 2009). We compared the calculated patient survival rate among three groups. Using multivariate analysis, the significant prognostic factors to predict patient outcome after allograft failure were assessed.

STATISTICS

Statistical analysis was performed using SPSS version 15.0 (SPSS Inc., Chicago, IL, USA). Continuous variables were compared by the Student's t-test or one-way analysis of variance. For categorized variables, the chi-square frequency analysis was used. Patient survival rate was calculated using the Kaplan–Meier method and we used the log-rank test to compare the survival rates of each group. Cox proportional hazard model was used for the multivariate analysis to evaluate mortality risk factors. The results were presented as mean ± standard deviation. The results were considered significant when the p-value was below 0.05.

RESULTS

Baseline Characteristics of the Patients with Failed Allograft

Out of the 368 patients with failed allograft, 233 patients (63.3%) belonged to the HD-PSKT; 64 patients (17.4%) to the PD-PSKT; and 71 patients (19.3%) to the ReKT group. Age at allograft failure was 40.9 ± 11.1 years and 264 (71.7%) patients were male. The most common cause of underlying renal disease was glomerulonephritis (75.3%) followed by diabetes mellitus (DM; 15.5%). Pre-transplant dialysis duration was 17.3 ± 24.7 months and the duration of functioning allograft was 77.8 ± 62.1 months. Allograft nephrectomy was done in 11 cases at allograft failure because of the graft intolerance symptom such as persistent gross hematuria, fever, or tenderness on the allograft. Most were living related donor transplantation (66.8%). The most common cause of allograft failure was chronic rejection (81.6%) followed by acute rejection (10.7%) and recurrence of primary renal disease (2.0%). Above variables did not differ significantly among three RRT groups (Table 1).

Table 1. Nonimmunologic characteristics of the patients with failed allograft

	Total (n = 368)	HD-PSKT (n = 233)	PD-PSKT (n = 64)	ReKT (n = 71)	p
Age at allograft failure (year)	40.9 ± 11.1	41.3 ± 11.2	41.8 ± 11.7	39.5 ± 7.5	0.21
Male (%)	264 (71.7)	169 (72.5)	40 (62.5)	55 (77.5)	0.14
Underlying renal disease
Glomerulonephritis	277 (75.3)	180 (77.2)	48 (75)	49 (69)
Diabetic nephropathy	57 (15.5)	35 (15.0)	6 (9.4)	16 (22.5)
Hypertension	15 (4.1)	8 (3.4)	4 (6.2)	3 (4.2)	0.52
Polycystic kidney disease	3 (0.8)	1 (0.4)	1 (1.6)	1 (1.4)
Others	3 (0.8)	2 (0.8)	1 (1.6)	(0.0)
Unknown	13 (3.5)	7 (3.0)	4 (6.2)	2 (2.8)
Graft survival duration (month)	77.8 ± 62.1	74.9 ± 63.0	83.3 ± 64.0	82.2 ± 57.2	0.51
Pre-transplant dialysis (month)	17.3 ± 24.4	17.5 ± 21.1	21.4 ± 25.7	12.2 ± 33.7	0.13
Follow-up after graft failure (month)	95.0 ± 76.0	79.6 ± 70.6	65.5 ± 4.9	135.1 ± 76.0	0.00
Transplant nephrectomy (%)	11 (3.0)	6 (2.6)	2 (3.0)	3 (4.2)	0.29

Notes: HD-PSKT, patients who started hemodialysis after allograft failure; PD-PSKT, patients who started peritoneal dialysis after allograft failure; ReKT, patients who underwent second transplantation.

Laboratory Findings and Comorbidity at Renal Allograft Failure

At the time of allograft failure, hemoglobin was 9.1 ± 1.9 g/dL, serum creatinine (Scr) was 7.1 ± 3.3 mg/dL, and serum albumin was 3.4 ± 0.5 g/dL (Table 2). DM was the most commonly combined disease in this study population at allograft failure (24.2%) and peripheral vascular disease was the second common disease (9.3%). Out of 89 patients with DM, 32 patients (35.9%) were new onset diabetes after transplantation (NODAT). According to the Davies Score, 213 patients (57.9%) were considered as low-risk, 132 patients (35.8%) as medium-risk, and 9 patients (2.4%) as high-risk patients. The laboratory findings and distribution of the Davies Score did not differ significantly among three RRT groups (Table 3).

Table 2. Immunologic characteristics of patients with failed allograft

	Total (n = 368)	HD-PSKT (n = 233)	PD-PSKT (n = 64)	ReKT (n = 71)	p
Donor type, n (%)
Living related donor	246 (66.8)	152 (65.2)	40 (62.5)	54 (76.1)	0.50
Living unrelated donor	104 (28.3)	65 (27.9)	22 (34.4)	17 (23.9)	0.52
Deceased donor	18 (4.9)	16 (6.9)	2 (3.1)	0 (0.0)	0.22
HLA mismatch number	2.9 ± 1.2	2.9 ± 1.2	3.0 ± 1.1	2.7 ± 1.0	0.33
Main immune suppressant, n (%)
Cyclosporine, n (%)	275 (74.7)	173 (74.2)	45 (70.3)	57 (80.2)	0.59
Tacrolimus, n (%)	40 (10.9)	27 (11.6)	8 (12.5)	5 (7.0)	0.47
Azathioprine, n (%)	53 (14.4)	33 (14.1)	11 (17.2)	9 (12.7)	0.42
Cause of graft failure, n (%)
Chronic rejection	300 (81.6)	192 (81.6)	51 (81.3)	57 (80.3)	0.82
Acute rejection	39 (10.7)	24 (10.3)	6 (9.4)	9 (12.5)	0.79
Accelerated rejection	5 (1.4)	4 (1.4)	1 (1.5)	0	0.26
Recurrence of primary disease	7 (1.7)	4 (1.4)	2 (3.0)	1 (1.4)	0.39
Hyperacute rejection	2 (0.6)	2 (0.9)	0	0	0.18
Others	15 (3.8)	8 (3.6)	3 (4.5)	4 (8.5)	0.30

Notes: HD, hemodialysis; PD, peritoneal dialysis; HD-PSKT, patients who started hemodialysis after allograft failure; PD-PSKT, patients who started peritoneal dialysis after allograft failure; ReKT, patients who underwent second transplantation.

Table 3. The laboratory findings and comorbidity status at renal allograft failure

	Total (n = 368)	HD-PSKT (n = 233)	PD-PSKT (n = 64)	ReKT (n = 71)	p
Hemoglobin (g/dL)	9.1 ± 1.9	8.9 ± 1.9	9.4 ± 1.9	9.2 ± 1.9	0.21
Serum creatinine (mg/dL)	7.1 ± 3.3	7.3 ± 3.5	6.2 ± 2.8	7.4 ± 2.6	0.08
Serum albumin (g/dL)	3.4 ± 0.5	3.4 ± 0.5	3.3 ± 0.5	3.5 ± 0.5	0.07
DM, n (%)	89 (24.2)	54 (23.2)	14 (21.9)	21 (29.6)	0.49
Malignancy, n (%)	13 (3.6)	7 (3.0)	5 (7.8)	1 (1.4)	0.10
IHD, n (%)	17 (4.7)	11 (4.7)	5 (7.8)	1 (1.4)	0.21
PVD, n (%)	34 (9.3)	24 (10.3)	7 (10.9)	3 (4.2)	0.26
LVD, n (%)	13 (3.6)	12 (5.2)	1 (1.6)	0 (0)	0.08
SCVD, n (%)	1 (0.3)	1 (0.4)	0 (0)	0 (0)	0.87
Other disease, n (%)	24 (6.6)	20 (8.6)	2 (3.1)	2 (2.8)	0.21
Davies score
Low risk (=0), n (%)	220 (59.8)	138 (59.2)	37 (57.8)	47 (66.2)	0.88
Medium risk (=1 or 2), n (%)	139 (37.8)	89 (38.2)	25 (39.1)	23 (32.4)	0.88
High risk (≥3), n (%)	9 (2.4)	6 (2.6)	2 (3.1)	1 (1.4)	0.79

Notes: DM, diabetes mellitus; IHD, ischemic heart disease; PVD, peripheral vascular disease; LVD, left ventricular dysfunction; SCVD, systemic connective tissue disorder; HD, hemodialysis; PD, peritoneal dialysis; HD-PSKT, patients who started hemodialysis after allograft failure; PD-PSKT, patients who started peritoneal dialysis after allograft failure; ReKT, patients who underwent second transplantation.

Renal Replacement Modality before Kidney Transplantation

Before KT, 287 patients (77.9%) were on HD, 54 patients (14.8%) were on PD, and 27 patients (7.4%) took preemptive KT. The distributions of dialysis modality before KT show significant differences among three groups. In HD-PSKT and ReKT group, the proportion of HD before KT was more than 80%. However, in PD-PSKT group, patients who were on PD before transplantation more frequently continued this dialysis modality (Figure 1). With respect to the dialysis modality between the allograft failure and second transplants, 77.5% (55/71) of the ReKT group patients underwent HD, but only 2.8% (2/71) of the patients underwent PD during this period. The interval between the allograft failure and second transplantation was 27.3 ± 27.3 months. Fourteen patients (19.7%) underwent second transplantation directly after allograft failure.

Figure 1. Comparison of pre-transplant renal replacement moda- lities according to the renal replacement modality groups after allograft failure. The HD-PSKT group and ReKT group show a similar distribution of renal replacement modality before transplantation. However, in the PD-PSKT group, significantly more patients were on PD before transplantation.

Notes: HD, hemodialysis; PD, peritoneal dialysis; Preemptive, patients who underwent transplantation without dialysis; HD-PSKT, patients who started hemodialysis after allograft failure; PD-PSKT, patients who started peritoneal dialysis after allograft failure; ReKT, patients who underwent second transplantation; HD-PSKT versus ReKT, p = 0.39; HD-PSKT versus PD-PSKT, p < 0.05; PD-PSKT versus ReKT, p < 0.05.

The Cause of Death after Renal Allograft Failure

Ninety-six patients died during the observation period. Cardiovascular disease (CVD) (39.6%) was the most common cause, followed by infection (34.4%) and malignancy (8.3%) (Table 4). The distribution of the causal disease for death shows significantly different pattern according to the duration from the allograft failure. For example, infection was a major cause of death within 10 years after allograft failure, but the proportion significantly decreased after 10 years. Malignancy accounted for less than 5% of the deaths within 10 years after allograft failure, but it became important cause of death after 10 years from allograft failure (Figure 2). The distribution of the causal disease for death did not differ significantly among three RRT groups (Table 4).

Table 4. The cause of death in patients after allograft failure

	Total (n = 96)	HD-PSKT (n = 63)	PD-PSKT (n = 19)	ReKT (n = 14)	p
Cardiovascular, n (%)	38 (39.6)	27 (42.9)	6 (31.6)	5 (35.7)	0.53
Infection, n (%)	33 (34.4)	23 (36.5)	7 (36.8)	3 (21.4)	0.54
Malignancy, n (%)	8 (8.3)	4 (6.3)	1 (5.3)	3 (21.4)	0.16
Liver, n (%)	3 (3.1)	3 (4.8)	0 (0)	0	0.44
Gastrointestinal, n (%)	6 (6.3)	3 (4.8)	2 (10.5)	1 (7.1)	0.65
Social, n (%)	1 (1)	0	0	1 (7.1)	0.05
Unknown, n (%)	7 (7.3)	3 (4.8)	3 (15.8)	1 (7.1)	0.27

Notes: HD, hemodialysis; PD, peritoneal dialysis; HD-PSKT, patients who started hemodialysis after allograft failure; PD-PSKT, patients who started peritoneal dialysis after allograft failure; ReKT, patients who underwent second transplantation.

Figure 2. Proportions of the main causes of death according to the interval from allograft failure. Note the sudden decrease in the proportion of infection and the increase in the proportion of malignancy after 10 years.

Patient Survival Rate and Prognostic Factors after Allograft Failure

The overall patient survival rate after allograft failure was 93%, 79%, and 63% at 1, 5, and 10 years, respectively. The ReKT group (1, 5, 10 years: 96%, 91%, 86%, respectively) was superior to the HD-PSKT (1, 5, 10 years: 89%, 70%, 57%, respectively) and PD-PSKT groups (1, 5, 10 years: 98%, 85%, 65%, respectively) in patient survival (ReKT vs. HD-PSKT, ReKT vs. PD-PSKT, p < 0.05, respectively). But no significant differences existed between the HD-PSKT and PD-PSKT groups (p = 0.653) (Figure 3). Univariate analysis revealed that age, serum albumin level, Davies Scores at allograft failure, and RRT group were the significant predictors for patient mortality. Multivariate analysis using Cox hazard proportional model revealed that high-risk group according to the Davies Score, age greater than 50 years, serum albumin level, and HD-PSKT or PD-PSKT were still independent risk factors for death when adjusted for gender, graft survival duration, transplant nephrectomy, donor type, HLA mismatch number, hemoglobin, and serum creatinine level (Table 5).

Figure 3. Comparison of patient survival rate according to the RRT group after allograft failure. The ReKT group shows a superior outcome as compared to the PD-PSKT and HD-PSKT groups. There are no significant differences between PD-PSKT and HD-PSKT groups.

Notes: RRT, renal replacement modality; HD-PSKT, patients who started hemodialysis after allograft failure; PD-PSKT, patients who started peritoneal dialysis after allograft failure; ReKT, patients who underwent second transplantation (ReKT vs. PD-PSKT, p = 0.009; ReKT vs. HD-PSKT, p = 0.00; PD-PSKT vs. HD-PSKT, p = 0.32).

Table 5. Variables affecting patient outcome after renal allograft failure (Cox proportional hazards analysis)

	Univariate			Multivariate^a
	HR	95% CI	p	HR	95% CI	p
Age at allograft failure (>50 year)	2.31	1.04–3.82	0.00	1.71	1.04–2.82	0.04
Male	0.64	0.54–1.64	0.84	–	–	–
Pre-transplant dialysis duration	0.99	0.98–1.01	0.19	–	–	–
Graft survival duration	1.00	0.99–1.00	0.15	–	–	–
Deceased donor	1.299	0.76–2.21	0.07	–	–	–
HLA mismatch number	0.98	0.32–2.51	0.75	–	–	–
Transplant nephrectomy	2.51	0.87–4.51	0.45	–	–	–
Cause of allograft failure	0.835	0.68–1.23	0.12
Davies Score
Low risk (reference)	1			1
Medium risk	1.59	1.27–2.95	0.00	1.68	0.90–3.06	0.07
High risk	4.26	2.34–13.55	0.00	5.51	1.28–9.11	0.00
Presence of DM	2.49	1.15–4.02	0.00	1.82	1.14–3.89	0.03
NODAT	1.23	0.92–1.65	0.16	–	–	–
Hemoglobin (g/dL)	1.02	0.89–1.15	0.89	–	–	–
Serum creatinine (mg/dL)	0.91	0.82–1.02	0.12	–	–	–
Serum albumin (g/dL)	0.43	0.23–0.80	0.00	0.68	0.27–0.91	0.02
RRT group
HD-PSKT (reference)	1			1
PD-PSKT	0.76	0.43–1.32	0.33	0.88	0.49–1.49	0.59
ReKT	0.22	0.09–0.54	0.00	0.23	0.09–0.57	0.01

Notes: HR, hazard ratio; CI, confidence index; DM, diabetes mellitus; NODAT, new onset diabetes after transplantation; HD, hemodialysis; PD, peritoneal dialysis; HD-PSKT, patients who started hemodialysis after allograft failure; PD-PSKT, patients who started peritoneal dialysis after allograft failure; ReKT, patients who underwent second transplantation.

^aAdjusted by gender, graft survival duration, transplant nephrectomy, donor type, HLA mismatch number, hemoglobin, and serum creatinine level.

DISCUSSION

With the increase in the number of cumulative renal transplants, patients with failed allograft have formed a significant portion of the total ESRD patients. However, enough data have not been available on the outcome and clinical course of those patients.^6,9 In our study, 63.3% of the patients returned to HD, 17.4% returned to PD, and 19.3% underwent a second transplant after allograft failure. In 2007, 67.4% of the total ESRD patients were on HD, 15.7% of the patients were on PD, and 20.8% of the patients had undergone transplantation in Korea.³ Thus, it seems that the distribution of RRT in patients with failed allograft is similar to that of transplant-naïve ESRD patients in Korea.

Most patients with failed allograft already had an experience of dialysis. Thus, we assumed that the type of dialysis modality before KT may be associated with the distribution of RRT after allograft failure. Actually in the comparison of dialysis modality before KT among three RRT groups, the proportion of patients who were on PD among the PD-PSKT group was significantly higher compared to that of the HD-PSKT and ReKT. Actually, among patients who were on PD before KT, nearly half of them returned to the PD-PSKT after allograft failure (data were not presented). This finding suggests that dialysis modality before KT may affect the type of dialysis modality after allograft failure, and this is more dominant in patients who were on PD.

CVD, infection, and malignancy were the three most common causes of death after allograft failure and it was commonly observed in three RRT groups. CVD was the most important cause of death during the entire period and the proportion showed increasing tendency with time from allograft failure. In contrast, infection was the important cause of death within 10 years from allograft failure, and it abruptly decreased after 10 years from allograft failure. Malignancy-related death was so rare in the early period after allograft failure. But it showed gradual increase with time, and it accounted for more than 30% of deaths developed after 10 years from allograft failure.

The sudden decrease of infection-related deaths after 10 years from allograft failure does not mean that the incidence of infection decreased. Rather, those findings suggest that it was not the cause of death during that period because the prevalence and severity of CVD or malignant disease increased further. The traditional risk factors for CVD and malignancy increase with aging and prolonged duration of dialysis.^10–12 In addition, prolonged use of immune suppressant can increase the risk for CVD and malignancy.¹³ Hence, we postulated that various factors were involved in CVD- or malignancy-related deaths increasing with time from allograft failure.

In contrast to patients with functioning graft, nonimmunologic factors have been known to be more important than immunologic factors in patients with failed allograft.¹⁴ It is thought that with the chronic fibrotic change of renal allograft, alloimmune reaction to grafted kidney diminished, hence immunologic factors could not affect patient outcome any more. Stepwise reduction of immune suppressant may be associated with the less importance of immunologic factor as well. Actually, the significant mortality factors in this study were comorbidity, hypoalbuminemia, old age, and RRT group, all of which were the nonimmunologic factors.^4,14–16

Previously, the comorbidity calculated by Davies scoring system has been shown to predict mortality in transplant-naïve ESRD patients and patients with failed allograft.^7,8,17 In this study, it was the most powerful predictor as well. Low serum albumin is significantly associated with malnutrition and inflammation and has known to be the important risk factor for death in the patients receiving dialysis.^18–20 The results of this study also have shown that hypoalbuminemia is an independent mortality risk factor in patients with failed allograft. Among the components of this scoring system, DM was the most commonly combined disease at renal allograft failure and moreover the presence of DM alone could predict patient mortality after allograft failure.

About one-third of DM was NODAT in this study. It has been reported to be associated with the graft and patient survival in renal transplant recipients.²¹ However, it could not predict mortality in patients with failed allograft in this study maybe because of the relatively low prevalence of NODAT in this study. Therefore, further investigation is required about the effect of NODAT on the patient mortality after allograft failure. Recently, Ayus et al. reported that transplant nephrectomy was reported to be associated with improved survival.²² We also evaluated the effect of transplant nephrectomy but it was not significantly associated with patient outcome. Owing to the limited patient number, it could not be concluded in this study.

The mortality rate of the patients who returned to dialysis after allograft failure have been reported to be higher than that of transplant-naïve ESRD patients who are on the waiting list.²³ Loss of the protective effect of the transplanted kidney has been suggested as a cause of this increased mortality.⁴ However, the patient outcome in our study is somewhat better than that in previous reports about Western population.^15,16,23,24 Furthermore, this is higher than the result of transplant-naïve ESRD patients who underwent HD or PD in Korea.²⁵ Ten-year survival rate of transplant-naïve PD patients was 48.2% and it was 54% in HD patients, somewhat lower than those of this study.^26,27

This divergence could be accounted by the inclusion criteria of this study. In this study, patients who died with functioning graft or died within 3 months of the allograft failure were excluded, because we intended to evaluate the effect of each renal replacement method on long-term survival. Exclusion of those patients may result in better patient outcome in this study. This could account for the relatively younger age of this study population as well. The mean age of all patients who underwent graft failure was 44.6 ± 10.7 years at transplantation and 53.2 ± 11.7 years at graft failure. This suggests that exclusion of those patients may decrease the mean age of this study population and it may partly contribute to the better survival rate of the patient population as well.

In transplant-naïve ESRD patients, transplantation is definitely superior to dialysis and no significant difference was found between HD and PD in patient survival in previous reports.^28–30 In patients with failed allograft, accumulating evidences showed that second transplantation could result in improved survival compared to the patients who remained on dialysis after allograft failure and no differences were found between PD and HD as well.^15,28,31,32 However, most previous studies did not compare three renal replacement modalities in the same study population. They mostly focused on the comparison of patient outcome between HD and PD.^17,33–35 In the comparison between second transplantation and dialysis they did not consider PD and HD as separate modality.^15,36

In this regard, we included not only the HD-PSKT and PD-PSKT groups but also the ReKT group in the comparison of RRT. The duration between the first and second transplantation was included in the survival duration of the ReKT group because this period was known not to be associated with the patient outcome.³⁷ We could not find any significant difference between HD-PSKT and PD-PSKT groups in patient survival. But the ReKT group showed a significantly better outcome compared to other two groups. Furthermore, we used multivariate Cox regression analysis to exclude the effect of confounding factors. Even after adjustment by other significant mortality risk factors, the ReKT was still superior to HD-PSKT and PD-PSKT. Therefore, we concluded that a second transplantation can give survival benefit to patients with failed grafts without the effect of other mortality factors.

In conclusion, CVD, infection, and malignancy were the most important causes of death irrespective of RRT after allograft failure. Therefore, prevention and surveillance of those disorders must be done in patients with failed allograft. Second transplantation was still superior modality compared to HD and PD in patient outcome after allograft failure. Comorbidity and hypoalbuminemia were independent mortality risk factors and proper management for them may help to improve patient outcome.

Declaration of interest: The authors report no conflicts of interest. The author alone is responsible for the content and writing of the paper.