Objectives: Long-term outcomes of kidney transplant recipients with postoperative genitourinary tract infections are not well characterized. In this single center retrospective study, we aimed to investigate the long-term effects of early posttransplant genitourinary infections under a protocol that included 1 month of antibiotic prophylaxis on graft failure and patient outcomes.
Materials and Methods: Electronic medical records of 1752 recipients of kidney-alone transplant between January 2000 and December 2008 were reviewed. Of these, 344 patients had postoperative genitourinary tract infections within 6 months of transplant. Infections included urinary tract infections, recurrent urinary tract infections, and pyelonephritis. All patients received 1-month of treatment with antibiotic prophylaxis for genitourinary infections after graft placement. Kaplan-Meier survival curves and mul-tivariable regression modeling were performed to determine survival outcomes.
Results: In the 344 patients with postoperative infec-tions, the most common cause was Escherichia coli (34.9%). Kaplan-Meier graft survival results showed no significant differences (P = .08) among those with and those without postoperative urinary tract infections; however, patient survival (P = .01) was significantly different. Multivariate analysis demonstrated no significant trend regarding graft failure (hazard ratio: 1.28; 95% confidence interval, 0.95-1.71; P = .09) or patient death (hazard ratio: 1.33; 95% confidence interval, 0.98-1.79; P = .06) in patients with and without genitourinary infections. The major cause of graft failure was infection in the infection cohort (17.4%).
Conclusions: Kidney transplant recipients who develop urinary tract infections within 6 months of transplant may be at increased risk of graft failure or patient death; however, further studies are needed to elucidate the relationship between posttransplant infections and long-term outcomes.
Key words : Graft failure, Pyelonephritis, Renal transplantation, Urinary tract infection
Although kidney transplant provides substantial benefits for the treatment of end-stage renal disease (ESRD), the risks of immunosuppression are sig-nificant. Posttransplant infection remains a serious threat to successful outcomes after kidney transplant despite reductions in infection-related 1-year mor-tality to less than 5%.1,2 Genitourinary infections, specifically urinary tract infections (UTI), are among the most common posttransplant infections, with an incidence rate ranging from 35% to 79%,3-5 and they commonly occur within the first year6 despite antibiotic prophylaxis, with durations of up to 6 months. Risk factors known to be associated with UTIs after kidney transplant include female sex, urinary catheters, diabetes mellitus, allograft trauma, previous UTI, prolonged dialysis before transplant, renal calculi, and urinary reflux.1,7-11 Causative uropathogens primarily consist of Enterococcus species, Enterobacteriae, and Escherichia coli.1,4,7
Studies on posttransplant genitourinary infec-tions and graft function or failure have demonstrated mixed outcomes, with some showing associations12,13 and others not.14-16 Hence, there remains ambiguity regarding the role of genitourinary infections on graft function and patient mortality in posttransplant recipients, specifically with regard to long-term outcomes. In this single center retrospective study, we aimed to address the long-term effects of early posttransplant genitourinary infections under a protocol that included 1-month treatment with antibiotic prophylaxis on graft failure and patient outcomes using both univariate and multivariate survival analyses.
Materials and Methods
We queried the electronic health records at the University of Michigan and identified 1752 adult patients who received kidney transplants between January 2000 and December 2008. Using the EMERSE search program developed at the University of Michigan, we used specific search terms to identify transplant recipients who developed genitourinary infectious complications within 6 months of surgery, including UTI, recurrent UTI (defined as more than 2 UTIs within 6 months), and pyelonephritis. A UTI was defined as escalating symptoms of urinary urgency, frequency, and dysuria clinically and confirmed with a positive urinalysis (2+ leukocyte esterase with or without positive nitrites) and > 100 000 colony-forming units/mL in urine cultures after 72 hours of growth. Urine cultures were followed up, and those with culture-negative results were excluded. Pyelonephritis was diagnosed on symptomatic presentation, which included fevers and dysuria, and was confirmed with positive urinalysis or urine culture and systemic leukocytosis (white blood cell count >12 000) or bacteremia.
The standard protocol in our practice has been to use sulfamethoxazole/trimethoprim or nitrofurantoin as first-line treatment for UTI infections and fluoro-quinolones or amoxicillin as second-line therapy. Duration of treatment is 3 days for women and 7 days for men given the concern for complicated UTI. Patients with pyelonephritis are treated with intra-venous fluoroquinolones with transition to oral therapy for a total of 7 days. During this period, our immunosuppression protocol included tacrolimus, mycophenolate mofetil, and prednisone. Recipients who qualified received thymoglobulin induction depending on immunologic risk or on whether delayed graft function was present. Preoperatively, patients received cefazolin as the antimicrobial of choice.
Foley catheters were removed within 2 days unless dictated to be longer by the surgeon. Ureteral stents were removed within 3 to 4 weeks posttransplant. After transplant, patients received sulfamethoxazole/trimethoprim (single strength tablet) for 30 days for UTI and Pneumocystis jiroveci pneumonia prophylaxis, along with acyclovir or valganciclovir for 3 months for viral prophylaxis.
Graft survival and patient survival were determined for all transplant recipients. Graft failure was defined as graft removal, return to dialysis, or death.17 Follow-up began at the time of transplant and ended at the occurrence of the event of interest, lost to follow-up, or end of the observation period (December 30, 2010). Initial graft and patient survival assessments were performed using Kaplan-Meier analyses, and patient groups were compared using log-rank tests.
Donor, recipient, and transplant-specific variables known to be associated with graft and patient outcomes were obtained from the Scientific Registry of Transplant Recipients. Donor variables included age, sex, race/ethnicity, weight, cause of death, diabetes, hypertension, use of machine perfusion, creatinine level > 1.3 mg/dL, stroke, and expanded criteria. Recipient variables included sex, race/ethnicity, insurance type, weight, body mass index, cold ischemia time, previous organ transplant, diabetes, and donor type. Risk factors were compared between the posttransplant infection group and the control group (those without infection) using contingency table chi-square analysis and one-way analysis of variance for categorical and continuous variables, respectively. Independent risk factors for graft and patient survival were determined using multivariate Cox proportional hazards models. Covariates with P < .20 in the univariate analyses were included in the final multivariate Cox models to determine hazard ratios for graft failure and patient death in those with posttransplant infection and those without. A separate multivariable logistic regression was performed to identify donor or recipient variables that were significantly associated with occurrence of postoperative genitourinary infections. P < .05 was considered statistically significant. Missing data were < 1% for other variables and were imputed when present. All statistical analyses were performed using STATA software (StataCorp, College Station, TX, USA).
Between 2000 and 2008, our center performed 1752 adult kidney transplants. Within this population, 344 patients (19.6%) met the criteria for posttransplant genitourinary infections within 6 months of surgery. Infections included posttransplant UTI, recurrent UTI, and pyelonephritis. The most common causes of infection in our cohort were Escherichia coli (34.9%), Enterococcus species (22.4%), and Pseudomonas species (15.4%) (Table 1).
Donor and recipient variables
Donor, recipient, and transplant-specific characteristics are listed in Table 2. Variables were not statistically different between the posttransplant genitourinary infection group and the control group, except for donor sex (P = .04), donor creatinine level (P = .04), recipient sex (P = .01), recipient insurance type (P = .02), recipient weight (P = .01), previous organ transplant (P = .04), and cause of ESRD (P = .03).
Risk factors for genitourinary infection
Univariate logistic regression analysis was initially performed to identify variables associated with post-transplant infections. Significant variables associated with posttransplant genitourinary infec-tions were donor ethnicity (specifically African American), female recipients, and Medicare insurance. Multivariate logistic regression was performed to identify recipient or donor variables associated with posttransplant genitourinary infection. Significant risk factors associated with posttransplant genitourinary infec-tions included female recipients (odds ratio: 2.42; 95% confidence interval [CI], 1.83-3.21) and Medicare recipients (odds ratio: 1.26; 95% CI, 1.22-2.13). No variables were protective against posttransplant genitourinary infection. Other variables are noted in Table 3.
Kaplan-Meier survival analysis
Kaplan-Meier graft survival rates at 1 and 5 years were 98.1% and 78.1% for those with posttransplant genitourinary infections and 93.1% and 78.6% for those without infection (Figure 1). Log-rank test showed P = .08, indicating no statistical difference between the 2 groups.
Kaplan-Meier patient survival rates at 1 and 5 years were 98.6% and 82.7% for those with posttransplant infections and 96.1% and 83.2% for those without posttransplant genitourinary infections (Figure 2). Log-rank test showed P = .01, demon-strating a statistical difference in patient survival between the 2 groups.
Univariate and multivariate analyses
Univariate analysis was performed for both donor and recipient factors to determine risk factors associated with graft and patient survival. We found no significant differences in graft failure between those with posttransplant infections versus those without infections (hazard ratio [HR] = 1.22; P = not significant). However, with regard to patient survival, there was a significant difference between those with and those without posttransplant infections (HR = 1.46; P = .01). Donors and recipients older than 40 years; donors with hypertension, use of machine perfusion, stroke, and expanded criteria; and recipients with Medicare insurance, older than 50 years old, and diabetes were factors that led to significantly increased risk of graft failure and patient death. Univariate analysis based on recipient uropathogen type demonstrated no significant differences in graft failure or patient death.
The threshold for inclusion in our final multi-variate model was a P value < .20. Based on this criteria, donor age, donor cause of death, presence of diabetes and hypertension in donor, use of machine perfusion, expanded criteria donor, recipient ethnicity, recipient insurance type, recipient age, previous organ transplant, presence of diabetes in recipient, recipient panel reactive antibody, cause of ESRD in recipients, and donor type were imple-mented into our graft failure model. For patient death, the following variables were used: donor age, donor cause of death, presence of hypertension in donor, creatinine > 1.3 mg/dL, stroke in donor, expanded criteria, recipient sex, recipient insurance type, recipient age, presence of diabetes in recipient, recipient panel reactive antibodies, cause of ESRD in recipient, and donor type.
The Cox regression hazards model for graft failure demonstrated no significant trend regarding graft failure between those with (HR: 1.28; 95% CI, 0.95-1.71; P = .09) and those without posttransplant genitourinary infections. Results for other covariates are listed in Table 4. Significant variables in this model included donor machine perfusion, donor stroke, recipient age older than 50, and hypertension as cause of ESRD in recipients.
For patient death, the Cox regression hazards model also showed no significant trend regarding death between those with (HR: 1.33; 95% CI, 0.98-1.79; P = .06) and those without posttransplant genitourinary infections. Results for other covariates and patient death are listed in Table 5. Significant variables in this model included recipient head trauma as cause of death, recipient age older than 50, and diabetes and hypertension as cause of ESRD in recipients.
Cause of graft failure
There were 419 graft failures in our cohort of 1752 transplant recipients. Causes of graft failure were identified for those with posttransplant genitourinary infections and those without. The primary cause of graft failure, aside from patient death, in those without infection was chronic rejection (18.4%). Other causes in the overall group included infection (10.7%), which included recurrent UTI, pyelo-nephritis, and opportunistic infections, acute rejection (3.9%), focal segmental glomerulosclerosis (3.9%) and nephropathy (0.6%), including diabetic and hypertensive. The primary cause of graft failure in those diagnosed with posttransplant infections was infection (17.4%), including recurrent UTI, pyelonephritis, and opportunistic infections; most genitourinary infections occurred at intervals greater than 6 months after transplant. Other causes included chronic rejection (14.1%), acute rejection (3.2%), focal segmental glomerulosclerosis (1.1%), and nephropathy (2.2%). There were no statistical differences seen between the 2 groups (P = .13).
Genitourinary infections, specifically UTIs, are the most common bacterial infection posttransplant, with an incidence rate of between 35% and 79%.3-5 These infections have the potential to impact graft function through acute or chronic kidney injury and can jeopardize both graft and patient survival if systemic sepsis ensues. Furthermore, graft pyelo-nephritis is often difficult to diagnose, as it does not always present with specific symptoms or classic findings of UTI in urinalyses and urine cultures. Hence, there is a likely continuum that makes it difficult to define by clinical presentation or culture data. In our study, the incidence of pyelonephritis was infrequent and could not in itself be used as an endpoint. The impact of these infections on graft and patient survival, however, is poorly understood, with no strong consensus in the literature.
In our study, which is one of the largest single center studies on this topic, we found that genitourinary infections within 6 months of transplant were associated with an increased risk of graft failure and death, which, although not statistically significant, continue to highlight their importance as a potential risk factor. Although our shorter duration of antibiotic prophylaxis treatment may have contributed to their presence, it is noteworthy that other studies have suggested infections to be a greater risk factor than even suggested by our study, and our incidence of infection is comparable to other studies. Our analysis specifically excluded those with negative data; interestingly, there was no significant difference or trend toward worse outcomes when considering all patients treated or suspected of a UTI.
Risk factors known to be associated with UTI after kidney transplant have been extensively studied and include female sex, urinary catheters, diabetes mellitus, allograft trauma, previous UTI, prolonged dialysis before transplant, renal calculi, and urinary reflux.1,7-11 Per our multivariable logistic regression model, we found that female recipients (odds ratio: 2.42; P = .01) were significantly more likely to have postoperative infections than male recipients, which is consistent with the increased risk of UTIs for women in all populations. Patients on Medicare also had an increased risk of infection (odds ratio: 1.62; P = .01). The increased incidence and identification of early postoperative UTIs in recent years may be the result of increased prophylactic ureteral stenting incorporated into surgical protocols.18
Bodro and associates found that lower urinary tract UTIs did not affect graft function.19 In their study, 184 of 867 kidney transplant recipients (21%) developed at least 1 UTI episode at a median of 18 days posttransplant. Urinary tract infections were not associated with graft dysfunction or graft loss; however, acute graft pyelonephritis was an independent risk factor for graft dysfunction and graft loss at 1 year. Giral and associates demonstrated that acute graft pyelonephritis within 3 months of transplant was an independent risk factor for graft loss,12 whereas Pelle and associates demonstrated increased serum creatinine levels and decreased creatinine clearance at 1 year after transplant in patients with pyelonephritis.13 Other studies have not found an association between genitourinary infections and renal graft function.14-16 Chuang and associates found that, in a cohort of 500 kidney recipients, 213 (43%) had 1 or more posttransplant UTI over a mean follow-up of 42 months.8 The authors demonstrated no association with renal graft loss; however, a significant association was shown between UTIs and increased mortality using multivariate regression analysis. Abbott and colleagues analyzed a nationwide cohort using the United States Renal Data System and found that late UTIs (presenting > 6 mo) were significantly associated with subsequent death using Cox regression analysis.20
It has been hypothesized that a bacterial infection activates the immune system, leading to acute or chronic rejection and subsequent impairment of kidney function.19 Others have suggested that acute graft pyelonephritis can cause interstitial scarring that leads to renal impairment.19,21,22 Although not significant, our analysis demonstrated a trend toward poor outcomes, suggesting that both of these may occur to a degree sufficient to cause a negative effect on graft or patient outcomes in the short or long term. Compared with other single center studies, our institution has had an incidence rate of genitourinary infection (19.6%) similar to that shown in other studies (21% to 48%).8,19 In these studies, 6 months of antibiotic prophylaxis was given compared with 1 month at our institution. Although our study was not designed to evaluate the effectiveness of the 1-month duration treatment, both our data and previous studies have suggested that a longer duration of antibiotic prophylaxis may be considered to minimize adverse outcomes related to graft and patient survival.
The primary limitation of our study was its retrospective single center design, which may have limited its generalizability given institutional variances. Second, we identified only 344 patients (19.6%) who had genitourinary infections within 6 months of transplant, which may be attributable to differences in definition. It is also possible that, although our cohort size was large for a single center study, the sample size may not be sufficient to prevent a type II statistical error.
Overall, in our analysis, the impact of UTIs in kidney transplant recipients does not appear to be as significant as we had anticipated. Nevertheless, prompt diagnosis and treatment remain important to optimize outcomes. This information can also be considered in the evaluation of candidates who may be at increased risk for UTIs; although this may be considered a risk factor, such candidates need not be excluded if otherwise suitable.
Kidney transplant recipients who develop UTIs within 6 months of transplant may be at increased risk of short- or long-term graft failure or death; however, further studies are needed to elucidate the relationship between posttransplant infections and long-term outcomes.
DOI : 10.6002/ect.2017.0196
From the 1Department of Internal Medicine, Yale University, New Haven,
Connecticut, USA; and the 2Department of Surgery and 3Department of Urology,
University of Michigan, Ann Arbor, Michigan, USA
Acknowledgements: The authors have no sources of funding for this study and have no conflicts of interest to declare.
Corresponding author: Randall Sung, Taubman Center 2922F, 1500 East Medical Center Drive, Ann Arbor, MI 48109-5331, USA
Phone: +1 734 936 8363
Table 1. Causative Pathogens of Genitourinary Infections
Table 2. Patient Characteristics
Table 3. Risk Factors for Urinary Tract Infection (Multivariate Analyses)
Table 4. Risk Factors for Graft Failure (Multivariate Analyses)
Table 5 continuet). Risk Factors for Graft Failure and Patient Death (Univariate Analyses)
Table 5. Risk Factors for Graft Failure and Patient Death (Univariate Analyses)
Figure 1. Kaplan-Meier Graft Survival at 1 and 5 Years
Figure 2. Kaplan-Meier Patient Survival at 1 and 5 Years