The impact of vesicoureteral reflux (VUR) on renal allograft outcomes is debatable, with small cohort studies reporting controversial results. The objective of this retrospective study was to evaluate long-term clinical effects of early VUR in a large cohort of kidney transplant patients. Posttransplantation voiding cystourethrography was used to evaluate 646 consecutive kidney transplant recipients before discharge. The study endpoints included VUR grade, death-censored graft or patient survival, renal function, proteinuria and occurrence of urinary tract infections (UTIs). Of the 646 recipients, 263 (40.7%) were diagnosed with VUR. VUR grade II was most common (19.8%), followed by grades III (10.2%), I (7.9%) and IV (2.8%). VUR was less common in transplantations performed by experienced compared to inexperienced surgeons (36% vs. 48%; p = 0.004). VUR did not affect death-censored graft or patient survival and was not associated with proteinuria or occurrence of UTIs. Patients with VUR had a lower eGFR at 1 year after transplantation than did patients without VUR (60 vs. 52 mL/min/1.73 m2; p = 0.02), although this difference was not observed at 3 and 5 years after transplantation. We conclude that early VUR, a common finding among renal transplant patients, may not have a meaningful impact on long-term transplant outcomes.
Long-term graft survival is a primary goal of renal transplantation. Although short-term survival rates have significantly improved over the last several decades, recent registry analyses indicate little-to-no decrease in long-term graft attrition (). Various immunological and nonimmunological factors affect graft survival (2011). Urological complications related to transplant surgery can negatively influence graft performance (2001, 2003, 2001). Despite use of meticulous ureteroneocystostomy techniques, including submucosal tunneling of the ureter, vesicoureteral reflux (VUR) remains a frequent finding after renal transplantation (2003, 2010). The actual impact of VUR on graft survival is the subject of an ongoing debate. Studies evaluating voiding cystourethrographies (VCUGs) in asymptomatic renal transplant recipients have revealed VUR rates of up to 86% (2007, 1990, 1982). Conflicting data have been reported regarding the associations between VUR and renal function (1993, 1993, 1994). Some reports suggest that urinary tract infections (UTIs) compromise renal allograft survival (1983, 1977, 2008, 2005), but the impact of VUR on the risk of developing UTI is less certain (2004). In addition, the effect of VUR on graft and patient survival remains unclear, and studies evaluating short-term survival rates have revealed inconsistent results (2009, 2007, 1997, 2008). The aim of this study was to investigate the effect of early posttransplantation VUR on long-term clinical outcomes.
Patients and Methods
In this retrospective study, we included 646 of 1536 consecutive adult patients who underwent kidney transplantation at the Medical University of Vienna between January 1999 and December 2007 (follow-up until July 2011). Criteria for the study inclusion were (i) age ≥18 years and (ii) availability of per-protocol VCUGs performed after transplantation and before discharge. The patient baseline data and immunosuppressive therapy are shown in Table 1.
Table 1. Baseline characteristics and immunosuppression
Comparing study patients with nonincluded subjects transplanted during the same time period, we found small but significant differences regarding proportion of living donor transplantation (66 [10.3%] vs. 133 [14.9%], p = 0.007) and donor age (median 51 [41–61] vs. 48 [37–59] p = 0.001), respectively. There were, however, no significant differences regarding other baseline variables, such as age, gender, cold ischemia time, HLA mismatch, current panel-reactive antibody (PRA) levels (data not shown).
Surgical techniques and assessment of VUR
According to our clinical protocol, ureteral implantation was performed in an antireflux manner using extravesical submucosal tunneling. Ureteral anastomosis was performed using running absorbable sutures as described earlier (1993). Ureteral stents were not routinely used. Foley removal was routinely scheduled on postoperative day 5 after transplantation.
All included patients underwent VCUG prior to first discharge after transplantation as part of our clinical protocol. The median interval between renal transplantation and VCUG was 24 days (IQR: 14–29). The VCUG study results were reviewed and graded by an independent radiologist and urologist (1985).
The severity of VUR was graded according to the International Reflux Committee Study (1985): grade I, reflux into nondilated ureter; grade II, reflux into renal pelvis and calyces without dilation; grade III, reflux with mild-to-moderate dilation and minimal blunting of fornices; grade IV, reflux with moderate ureteral tortuosity and dilation of pelvis and calyces; grade V, reflux with gross dilation of ureter, pelvis, and calyces, loss of papillary impressions and ureteral tortuosity.
The level of training and experience of the surgeon in each case was analyzed. Surgeons who had performed less than 30 transplantations were categorized as inexperienced and surgeons who had performed more than 30 were considered experienced.
Documented study endpoints included date of graft loss and patient death, occurrence of biopsy proven T cell or C4d-positive antibody-mediated rejection (C4d staining on paraffin sections) according the definitions of the Banff classification (2008), estimated glomerular filtration rate (eGFR) calculated according to the Mayo Clinic equation (2004) (patients receiving dialysis were considered as having an eGFR of 5 mL/min/1.73 m2), and protein excretion (patients receiving dialysis were not included in the statistical analysis). We defined significant proteinuria as a protein/creatinine ratio greater than 500 mg/g and in some patients, as protein excretion greater than 0.5 g/L or 0.5 g/24 h and/or a positive dipstick urinalysis. In addition, our analysis included evaluation of urine culture results obtained at routine follow-up visits as well as upon admission for presentation of clinical symptoms consistent with UTI within the first year after transplantation. UTIs were diagnosed based on positive urine cultures with more than 105 colony-forming units of pathogenic organism per milliliter of urine. Patients with two or more episodes of UTI during the follow-up were considered to have recurrent UTIs.
Continuous data were given as the median and the interquartile range (IQR; range from the 25th to the 75th percentile). Discrete data were presented as counts and percentages. Chi-square tests or, if appropriate, exact tests were used to compare groups of categorical data. For comparisons of continuous data, the nonparametric Mann–Whitney U-test was used. Kaplan–Meier analysis was used to calculate graft and patient survival, and the Mantel Cox log-rank test was used to compare survival between groups. Multivariate models (logistic regression analysis or Cox regression analysis) were used to determine the independent effect of VUR on transplant outcome parameters. Such models included cofounding variables unequally distributed between groups or confounders considered to have an impact on the end point. A two-sided p-value of <0.05 was considered statistically significant. Statistical calculations were performed using SPSS for Windows, version 12.0 (SPSS Inc., Chicago, IL, USA).
The study was approved by the ethics committee of the Medical University of Vienna (Research ethics reference 1078/2012).
Overall, 263 (40.7%) of the 646 kidney transplant recipients were diagnosed with posttransplantation VUR before discharge. Fifty-one (7.9%) patients had grade I VUR; 128 (19.8%), grade II; 66, (10.2%) grade III and 18 (2.8%), grade IV VUR. Over the entire period 397 (62%) transplantations were performed by experienced surgeons and 249 (38%) were done by inexperienced surgeons.
As shown in Table 1, VUR was less common in live donor transplants than in deceased donor transplants (6% vs. 46%; p = 0.004). There were also imbalances in IL-2 receptor antibody induction therapy (p = 0.001) and retransplantation (p = 0.001). The presence or absence of VUR was significantly influenced by the surgeon's experience level at the time of transplantation. VUR was less common in transplantations performed by experienced compared to inexperienced surgeons (36% vs. 48%; p = 0.004).
Effect of VUR on UTI incidence
The overall incidence of UTI during the first year of follow-up was 30.2%, with 4% of the patients showing more than 2 infection episodes. We found no significant relationship between diagnosis of VUR and UTI incidence. Simple UTI was diagnosed in 24.7% of patients with VUR and 27.2% of patients without VUR (p = 0.78). Recurrent UTIs were noted in 4.2% (with VUR) vs. 3.9% (without VUR) of the enrolled patients (p = 0.67).
Effect of VUR on graft function and rejection
As shown in Figure 1A, patients with VUR had significantly lower eGFR levels at 1 year after transplantation than did patients without VUR (52 vs. 60 mL/min/1.73 m2; p = 0.02). In a multivariate model including all relevant confounders (living donor, female sex, retransplantation, HLA-mismatch, initial IL2 antibody induction, T cell- or C4d-positive antibody-mediated rejection, recipient age, donor age and cold ischemic time [CIT]), VUR still predicted lower eGFR (HR, 2.06 [CI, 1.22–3.48]; p = 0.007). However, analysis of renal function at 3 and 5 years showed no significant differences (Figure 1A). Moreover, we could not detect any difference between patients with and without VUR with respect to the occurrence of proteinuria (Figure 1B). Finally, VUR was not associated with T cell-mediated rejection (32% [with VUR] vs. 26% [without VUR]; p = 0.07) or antibody-mediated rejection (20% [with VUR] vs. 17% [without VUR]; p = 0.32).
Effect of VUR on graft and patient survival
Kaplan–Meier analysis revealed a trend toward lower death-censored graft survival in patients with VUR (p = 0.09; Figure 2). However, in multivariate Cox regression analysis (included confounders: living donor, female sex, retransplantation, HLA-mismatch, initial IL2 antibody induction, T cell- or antibody-mediated [C4d-positive] rejection, recipient age, donor age and CIT), no significant effect of VUR was observed (HR, 1.26 [CI, 0.83–1.88] p = 0.26). Comparison of different VUR categories (no VUR vs. VUR grades 1 + 2 vs. VUR grades 3 + 4) also showed no significant differences with respect to graft survival (Figure 3) or patient survival (not shown).
VUR is a common finding in patients after renal transplantation (2010). However, the actual impact of this condition has been a subject of ongoing debate (2009, 2008, 1997, 1993, 1994, 1977, 1978). To the best of our knowledge, this study is the largest series of deceased and living donor kidney allograft recipients in which the effect of early VUR on the incidence of UTI, rejection rates, and graft and patient survival was analyzed.
In our cohort of 646 patients, who underwent standardized per-protocol VCUG as part of our clinical routine, we found a 40.7% overall rate of reflux. This is in accordance with previously published data reporting an incidence of VUR as high as 86% in asymptomatic patients after renal transplantation (1977, 2008). However, previous studies performed VCUG merely per-indication or in selected patient cohorts. Thus, direct comparison with such data is difficult.
Even though the antireflux ureteroneocystostomy technique was used for all our patients, the number of patients presenting with VUR was high. The exact mechanisms underlying VUR development after a renal transplantation using ureteroneocystostomy in an antireflux manner remain to be elucidated. Cash et al. evaluated the impact of surgeons’ experience level on functional outcomes after renal transplantation and found no significant difference between experienced and inexperienced surgeons (2012). However, their analysis did not include an evaluation of VUR. In our study the surgeons’ experience level showed no impact on patient or graft survival, but we could demonstrate that it significantly influenced VUR rates. An interesting finding, for which we have no good explanation, was that patients receiving IL-2 antibody induction showed a significantly lower rate of reflux. A role of alloimmune mechanisms favoring VUR, which may be prevented by more intense immunosuppression, remains speculative.
In our cohort, the overall incidence of urinary tract infections during the first year of follow-up was 30%. Recurrent infection episodes were seen in only 4% of patients. Although there is consensus that recurrent UTIs have a potential for unfavorable renal allograft outcome, the association of VUR with UTIs is less certain (2010). Previously published studies postulated VUR as a possible risk factor for developing recurrent UTIs (2008). Ohba et al. (2004) found that VUR-related pyelonephritis could be an important long-term complication in the survival of renal allografts. However, in their study, VCUG for determining VUR was available for only seven patients.
Among transplant patients, a markedly increased risk for UTIs appears to derive from VUR into the native kidney, rather than into the transplanted kidney (2005, 1998). In the study of Chuang et al. (2005), preexisting VUR increased the relative risk for developing UTIs in the transplant population. Similarly, Erturk et al. (1998) reported a high incidence of UTIs over a mean period of 54 months among renal transplant patients with a history of VUR. In contrast, studies analyzing VUR into the transplanted kidney found no increased risk for UTI (1993, 1994, 2010).
Our data showed that UTIs, both solitary and recurrent, were not influenced by the presence or absence of VUR. Nevertheless, UTIs are frequent after renal transplantations (2008) and may adversely impact renal allograft function, even in the absence of VUR (2004, 2007). Dupont et al. (2010) evaluated whether the presence of VUR confers an increased risk for UTIs and renal allograft scarring. Of the patients with a history of recurrent UTI, 65% showed no evidence of VUR. We could not confirm the potential negative effect of UTI on long-term outcome in this study, as no effect was seen on graft or patient survival. This could be explained by a low incidence of recurrent UTIs compared to that reported in other studies (2007), close postoperative monitoring, and aggressive treatment.
The main objective of our study was to evaluate the impact of early VUR on short- and long-term clinical outcome of renal transplantation, including graft and patient survival. Currently, the evidence available is poor and mainly derived from very small and underpowered studies, thus limiting data interpretation. In a small cohort series of 37 renal transplant patients, Favi et al. (2009) found no effect of VUR on graft survival or renal function at 5 years, as indicated by VCUG. The reported VUR incidence of 41% (15/37 patients) is in accordance with our findings. However, this study included no patients with VUR higher than grade II. Furthermore, only patients with grafts surviving more than 2 years who had an UTI episode in their posttransplant course were included, making this cohort with a limited number of patients even less comparable. In another small series of 74 renal transplant recipients, no influence of VUR on graft function could be demonstrated (2008). However, data on long-term graft function and survival were not available in this study. In both studies, VCUGs were only performed for clinical indications (2009, 2008).
Our data showed that long-term patient and graft survival and renal function did not appear to be influenced by the presence of VUR. Interestingly, eGFR levels at 1 year after transplantation were significantly lower in patients with VUR than in patients without VUR (p = 0.02). This finding was confirmed in a multivariate model including all relevant confounders, but the difference was no longer seen at 3 and 5 years. We further evaluated patients with VUR according to the severity of reflux in order to determine whether the finding at 1 year was related to a higher grade of reflux; no differences were found.
We recognize that our study has several limitations. A considerable proportion of patients could not be included because routine VCUG was not available. This lack of data may have led to some bias. Indeed, a missing-data analysis revealed small but significant imbalances in donor type and age. Another drawback of our study is that VCUG was performed early after transplantation, and follow-up VCUGs were not available. The ureteroneocystotomy continues to heal for at least 6–8 weeks after transplantation, and reflux identified early postoperative may later disappear or reappear. In light of the earlier finding that VUR incidence may change with time after surgery and considering that the ureteroneocystotomy continues to heal for at least 6–8 weeks after transplantation, one may argue that VUR identified early after transplantation could disappear or reappear during subsequent follow-up (1990, 2000). Further limitations are that the recording of UTIs was limited to the first year after renal transplantation and that for some of the parameters tested, including death-censored graft survival for cases with grade III and IV reflux, the study might have been underpowered.
This is the largest series of consecutively performed per-protocol VCUGs early after renal transplantation. The incidence of VUR was high, even though antireflux procedures were used for implantation. However, early VUR had no meaningful effect on death-censored graft survival or overall patient survival. Thus, the necessity of per-protocol VCUGs and the utility of performing antireflux anastomosis have to be questioned.
The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.