Acute Pyelonephritis Represents a Risk Factor Impairing Long-Term Kidney Graft Function

Authors


* Corresponding authors: Alain Vandewalle and Eric Rondeau, vandewal@bichat.inserm.fr eric.rondeau@tnn.aphp.fr

Abstract

Urinary tract infections (UTIs) and acute pyelonephritis (APN) often occur after renal transplantation, but their impact on graft outcome is unclear. One hundred and seventy-seven consecutive renal transplantations were investigated to evaluate the impact of UTIs and APN on graft function. The cumulative incidence of UTIs was 75.1% and that of APN was 18.7%. UTIs occurred mainly during the first year after transplantation and Escherichia coli, Pseudomonas aeruginosa and Enteroccocus sp. were the most frequent pathogens identified. The risk of developing APN was higher in female (64%) than in male recipients, and was correlated with the frequency of recurrent UTIs (p < 0.0001) and rejection episodes (p = 0.0003). APN did not alter graft or recipient survival, however, compared to patients with uncomplicated UTIs, patients with APN exhibited both a significant increase in serum creatinine and a decrease in creatinine clearance, already detected after 1 year (aMDRD-GFR: APN: 39.5 ± 12.5; uncomplicated UTI: 54.6 ± 21.7 mL/min/1.73 m2, p < 0.01) and still persistent (∼− 50%) 4 years after transplantation. Multivariate analysis revealed that APN represents an independent risk factor associated with the decline of renal function (p = 0.034). Therefore, APN may be associated with an enduring decrease in renal graft function.

Abbreviations: 
APN

acute pyelonephritis

aMDRD-GFR

abbreviated modification of diet in renal disease-glomerular filteration rate formula

CrCI

creatinine clearance

DGF

delayed graft function

PRA

panel reactive antibodies

UTI

urinary tract infection

Introduction

Urinary tract infection (UTI), including asymptomatic bacteriuria, cystitis and pyelonephritis, is the most common form of bacterial infection in renal transplant recipients (1, 2). These infections are thought to be directly attributable to exposure to pathogens during the early postoperative period and to immunosuppressive therapy (3–5). The improvement of surgical procedures, rapid removal of the urethral catheter, and antibiotic prophylaxis have reduced the incidence of UTI during the immediate postoperative period (6, 7), but it is still higher in transplanted patients than in the normal population. Until recently UTI was considered to be relatively benign, but a recent major retrospective study has revealed that late UTIs occurring after renal transplantation (i.e. more than 6 months following surgery) are associated with a significantly increased risk of subsequent death (8). However, this study did not establish whether UTIs were the primary cause of death or were a consequence of serious underlying illness. In addition, most of the studies did not distinguish between simple UTIs and pyelonephritic UTIs and the impact of acute pyelonephritis (APN) on immediate and long-term renal graft function is still debated (9–11). Also, the impact of APN on chronic allograft nephropathy, which is the most prevalent cause of renal transplant failure in the first decade posttransplant, is not clear. Although the incidence of early APN (i.e. during the 3 months following surgery) appears to be significantly detrimental to graft outcome (11), the long-term consequences of APN on graft survival have not been clearly established (8, 11, 12). Therefore, the question remains as to whether the occurrence of APN following renal transplantation may significantly impair renal function and thereby promote graft loss. The aim of the present study is to characterize the frequency of causative microorganisms responsible for simple or pyelonephritic UTI, to analyze the factors associated with APN and to evaluate the impact of APN on renal function in a group of adult patients who had kidney transplant, with a follow-up time of 4 years.

Materials and Methods

Patients

This retrospective study was conducted on 172 consecutive patients who had received a renal allograft (177 kidney transplants) from January 2000 to December 2005 at Tenon hospital (Paris, France). The median follow-up period was 665 days and the percentage of graft loss was 16.9%. In all patients, a de novo insertion of the ureter into the bladder was performed. A ureteral stent catheter was inserted, and removed by cystoscopy 4 to 6 weeks later. A urethral catheter was also inserted, and then systematically removed 5 days after surgery.

Immunosuppressive treatment

Several different protocols were used for the induction and maintenance of immunosuppression. For induction therapy, the monoclonal anti-IL-2 receptor antagonist basiliximab (Simulect®, Novartis Pharma, Rueil Malmaison, France) was used in 138 renal grafts, and rabbit ATG (Thymoglobulin®, Imtix-Sangstat, Lyon, France) were used in 26 renal grafts. Thirteen patients did not receive any induction therapy. Corticosteroids were given to all the patients. A single 500 mg dose of methylprednisolone (Solumedrol®, Pharmacia, St Quentin-en-Yvelines, France) was given intraoperatively, with 1 mg/kg/day prednisone thereafter (Cortancyl®, Novartis Pharma, Rueil Malmaison, France), and was progressively decreased to a dose of 5 mg/day. Steroid treatment in 66% of the patients, was associated with a combination of mycophenolate mofetil (Cellcept®, Roche, Neuilly-sur-Seine, France) followed by cyclosporin treatment (Neoral®, Novartis Pharma), which was started when the serum creatinine fell below 2.5 mg/dL, at latest day 7 posttransplantation. The dose of cyclosporin was adjusted to provide drug blood level between 100 ng/mL and 200 ng/mL. Fourteen patients were treated with a combination of prednisone, mycophenolate mofetil, and sirolimus (Rapamune®, Wyeth-Lederlé, Paris, France). Thirty-four patients (19.2%) received prednisone, mycophenolate mofetil, and tacrolimus (Prograf®, Astellas Pharma Inc., Levallois Perret, France). The dose of tacrolimus was adjusted to obtain trough blood levels of 5–15 ng/mL. Five patients received only prednisone and mycophenolate mofetil.

Antibiotic prophylaxis

All the patients received antibiotic prophylaxis consisting of 2 g of intravenous cefazolin intraoperatively. Postoperative antibiotic prophylaxis consisted of trimethoprim-cotrimoxazole (Bactrim®, Roche, Neuilly-sur-Seine, France), administered during the first 3 months to prevent bacterial infection and Pneumocystis carinii pneumoniae. Due to intolerance to Bactrim®, three patients received inhalation of pentamidine isethionate on a monthly basis (Pentacarinat® nebulizer solution, Sanofi-Aventis, Paris, France). All subjects also received oral fluconazole (Triflucan®, Pfizer, Orsay, France) daily and prophylactic treatment against cytomegalovirus (CMV) consisting of ganciclovir (Cymevan®, Roche, France) or valganciclovir (Rovalcyte®, Roche, France) during the first 3 months, except for the CMV-negative recipients who received a CMV-negative kidney. None of the patients received prophylactic treatments such as lactobacillus or cranberry products, thought to prevent fimbriated E. coli from adhering to uroepithelial cells (13, 14). Eight patients developed CMV disease. This was defined as the association of fever with one or more of the following clinical symptoms or biological abnormalities: leukopenia, gastrointestinal disease, pancreatitis, hepatitis, pneumonitis, myalgia, arthralgia and/or nephritis. Virological proof of CMV infection was obtained in all patients by antigenemia or rapid viremia tests. The patients who had developed CMV infection were treated with intravenous ganciclovir.

Laboratory investigation

Blood samples and urinary midstream specimens were obtained twice weekly during the first 3 months, weekly during the following three months, and approximately monthly after over a period of 4 years. Creatinine clearance (CrCl) was calculated using the Cockcroft and Gault formula (15) and the abbreviated MDRD formula [aMDRD = 1.86 × (serum creatinine, mg/dL)−1.154× (age)−0.203× (0.742 for woman; × 1.21 for African American)] (16).

Episodes of UTI and APN

The diagnosis of UTI was reached on the basis of a urinary bacteria count of more than 104 colony forming units (cfu)/mL combined with a urinary leukocyte count of more than 104/mL, or of a single urinary bacteria count of more than 105 cfu/mL. Urinalysis and urine culture were performed in all patients irrespective of symptoms at the frequency described above. Additional urine culture and urinalysis was performed when patients exhibited symptoms suggesting of UTI. UTI was referred to as recurrent if it followed the resolution of a previous UTI and/or involved a reinfection or relapse. APN was defined as a combination of a positive diagnosis of UTI and fever and with one or more of the following clinical or biological symptoms: graft pain, chills, cystitis, dysuria. An association of two antimicrobial agents, based on the antimicrobial susceptibility of the infecting organism, were used, first intravenously until apyrexia was achieved, and then orally for at least 3 weeks. All patients exhibiting APN were hospitalized and monitored at least twice a week during approximately 1 month. Urological evaluation was only carried out for patients who developed a second episode of APN. Patients with uncomplicated UTIs were generally orally treated using single antibiotherapy during 5–7 days. For all patients with uncomplicated UTIs, plasma levels of drugs and laboratory investigations were generally performed the day of the diagnosis and the 2 weeks following, until clinical symptoms and bacteriological abnormalities were resolved.

Acute rejection episodes

Acute rejection episodes were suspected in the case of an elevation of the serum creatinine and were confirmed by graft biopsy examination. Anti-donor HLA specific antibodies in the serum were not performed systematically, except for patients with humoral rejection diagnosed by histological examination of graft biopsy sample. In most cases, the patients with acute rejection were not tested for C4d staining. In a few cases (eight patients), biopsy sampling was not technically possible, and these ‘intention-to-treat’ episodes also were taken into account. The treatment of acute rejection consisted of intravenous corticosteroid (Solumedrol®) boluses for five consecutive days, followed by ATG in the event of corticosteroid resistance manifested as the persistence or exacerbation of the elevated serum creatinine after the last bolus, and the absence of histological improvement.

Parameters studied and statistical analyses

A number of factors listed in tables were analyzed. Body mass index (BMI) and proteinuria were analyzed after 3 months and 1 year following transplantation. CrCl was determined at 3 and 6 months, and 1, 2, 3 and 4 years after renal transplantation. Analyses were performed on the entire cohort of patients and in three groups of patients: Group 1, patients who did not develop UTI (-UTI); Group 2, patients with uncomplicated UTI (+UTI, −APN); Group 3, patients with pyelonephritic UTI (+APN). In patient Group 3, only serum creatinine and CrCl values measured after the first episode of APN were taken into account. For each of the continuous variables studied, we compared the three groups using the Kruskall Wallis test, and when a p value was <0.05, comparison between two groups was performed using the Mann-Whitney U test. For categorical variables, χ2 was used. Serum creatinine and glomerular filtration rate (GFR) values were compared between the three groups at various times following transplantation. Data are reported as mean values ± SD, and a p value <0.01 was considered to be significant. A backward stepwise logistic regression analysis was also performed to determine the independent risk factors for APN. Variables with a p value <0.2 in the univariate analysis were included in the multivariate model. Graft and patient survival analyses were performed using the Kaplan-Meier method, and statistical differences between curves were assessed using the log rank test. Backward stepwise logistic regression was also performed to identify the potential independent risk factors associated with the decline of GFR during the first year posttransplantation. For this purpose, we used a binary variable. The limit between the upper third and fourth quartiles of serum creatinine variations (Δcreat) between year 1 and month 3 posttransplantation was 19 μmol/L. Such Δcreat increase was considered as clinically representative of degradation of renal graft function. Therefore, preoperative, operative and postoperative potential risk factors were analyzed for their association using this criteria on 107 patients for which data were available at month 3 and year 1 posttransplantation. For univariate statistics, χ2 (for categorical variables) and Mann-Whitney's U-tests (for continuous variables) were used. Variables with a p value <0.2 in the univariate analysis were included in this backward stepwise logistic regression. Differences were considered to be significant for p value <0.05. Statistical analysis was performed with StatView™ software. Multivariate analysis using the Cox proportional hazard model analysis was also used to determine relevant risk factors for graft lost on the entire cohort of transplant patients.

Results

Frequency of urinary tract infections and causative microorganisms

The demographic characteristics of the three groups of transplanted patients analyzed are shown in Table 1. After transplantation, most of the patients (75.1%) had at least one episode of UTI during the five years of follow-up. UTIs occurred more often in female than in male recipients. Table 2 recapitulates the diversity and percentages of causative microorganisms identified in urine cultures during the first episode of UTI or during second and subsequent UTIs. A wide range of other Gram-negative and Gram-positive rods, which had not been detected during the first episode of UTI, were also identified as the single causative agent (Table 2). The proportion of polymicrobial UTI (two germs in most cases, most probably the consequence of contamination of the sample) was about 8%, whether during the first or the subsequent episodes of UTI (Table 2).

Table 1.  Differences in demographic factors linked to the presence or absence of APN
 TotalGroup 1Group 2Group 3p-value
  1. Total = whole population of transplanted patients; Group 1 = patients who never had UTI; Group 2 = patients who did have uncomplicated UTI; Group 3 = patients who did experience APN. Diabetes, urological and polycystic kidney diseases are not included in the Table, as they were only 1 to 4 of initial disease in each group. BMI = body mass index; Anti-PRA = panel reactive antibodies; MMF = mycophenolate mofetil; DGF = delayed graft function; Post-T = posttransplantation.

  2. *Values are given as means.

Number of transplants1774410825 
Recipient age (year*)46.543.947.944.8NS
Ethnic group (Caucasian/African/Other)90/45/4222/10/1255/31/2213/4/8NS
Male recipients117347490.0017
Anti-PRA > 20 %14563NS
Hepatitis C virus seropositivity194114NS
BMI at 3 months post-T*24.624.424.724.6NS
BMI at 1 year post-T* Initial end-stage renal disease24.924.625.423.3NS
Glomerulonephritis6623310NS
Nephroangiosclerosis and others111217515NS
Donor age (year*)51.752.250.855.3NS
Male donors104256613NS
Living donor247143NS
Primary transplantation148349618NS
Cold-ischemia time (h*)20.318.120.622.9NS
 Immunosuppressive treatment NS
Anti IL-2 receptor138338619NS
ATG268126NS
Cyclosporine/MMF/prednisone117297117NS
Sirolimus/MMF/prednisone14194NS
Tacrolimus/MMF/prednisone3410213NS
Number of patients with DGF4515246NS
Patients with acute rejection551327150.0003
CMV disease8152NS
Post-graft diabetes174103NS
Proteinuria at month-3 post-T (g/24h*)0.540.590.570.29NS
Proteinuria at year-1 post-T (g/24h*)0.730.730.611.26NS
Table 2.  Causative microorganisms identified in urine of renal transplant recipients during the first and recurrent episodes of UTI over 5 years
 First episode of UTISecond and more episodes of UTI
Number of positive cultures(%)Number of positive cultures(%)
  1. *Others corresponded to rare pathogens (≤1%) detected during the first episode of UTI (such as Klebsiella planticola or Haemophilus influenzae), and second or more UTI episodes (such as Comamonas sp, Staphylococcus aureus or Aeromonas hydrophila). Polymicrobial UTIs corresponded to UTI with positive urine culture for two germs and in one case to three germs in the group of second and more episodes of UTIs.

Causative microorganisms Gram-negative rods148 499 
Escherichia coli42(28.4)126(25.1)
Pseudomonas aeruginosa22(14.9)80(16.0)
Enterobacter cloacae9(6.1)20(4.0)
Klebsiella pneumoniae3(2.0)29(5.8)
Klebsiella oxytoca2(1.3)13(2.6)
Serratia marcescens2(1.3)10(2.0)
Stenotrophomonas maltophilia2(1.3)5(1.0)
Citrobacter freundii1(<1.0)16(3.2)
Proteus mirabilis10(2.0)
Morganella morganii Gram-positive cocci6(1.2)
Enterococcus sp35(23.6)92(18.4)
Staphylococcus nonaureus17(11.5)48(9.6)
Streptococcus sp4(2.7)11(2.2)
 Others*9(6.1)33(6.6)
Total number of UTIs136 459 
Polymicrobial UTIs12(8.8)37(8.1)

Figure 1 displays the timing of the occurrence of the first episode of UTI following renal implantation. P. aeruginosa, coagulase-negative Staphylococci or E. cloacae were frequently detected in the urine within the first 3–5 weeks following surgery. Enterococcus sp and E. coli were also frequently detected during the first six weeks and 12 weeks, respectively, after transplantation (Figure 1). The microorganisms responsible for the first episode of UTI during the first three months following renal graft (excluding the natural resistance of P. aeruginosa to Bactrim®) were frequently found to be resistant to trimethoprim-cotrimoxazole (E. coli, 84%; E. cloacae, 67%; noncoagulase Staphylococcus, 86%; Enterococcus sp, 46%). Most of the UTIs (74%) occurred during the first year following transplantation, mainly during the first three months (81.9%). Thereafter, the proportion of UTIs significantly decreased during the second (35.7%) to fourth (21.5%) year following transplantation.

Figure 1.

Time to occurrence of the first episode of UTI depending on the microorganism identified in the urine culture. Time values are means ± SD from (n) positive urine culture.

Incidence of acute graft pyelonephritis in renal transplant recipients

Among 133 renal transplant patients who experienced single or recurrent episodes of UTI, 25 patients (18.7%; 9 men, 16 women) had 41 typical episodes of APN (one episode of APN: 14 patients; 2 APN episodes: 6 patients; 3 episodes: 5 patients). The risk of developing APN was closely correlated to the frequency of recurrent episodes of UTI (p < 0.0001). Among patients with recurrent APN, vesicoureteric reflux and dilatation of the upper urinary tracts were detected in three cases. Figure 2A depicts the cumulative incidence of APN as a function of time after renal engraftment. About half of APN (51%) occurred within the first six months following transplantation. Where E. coli represented the main uropathogen (51%) responsible for about half of the APN, the percentages of the other causative microorganisms were similar to those of pathogens identified in urine cultures over the follow-up period. At 30 months following renal graft, graft survival was 85.1% in patients without APN and 76.5% with APN. Kaplan-Meier analysis of graft survival indicated that the occurrence of single or recurrent episodes of APN did not significantly influence graft survival in patients with or without APN (Figure 2B). However, Kaplan-Meier analysis stratified by gender revealed larger differences in graft survival in female than in male transplant patients (Figure 2C and D). Due to the relative limited number of transplanted patients, it cannot be excluded that the differences might have been significant with a larger sample size. Despite this restriction, APN also did not significantly affect graft survival in patients with APN compared to patients without APN (graft survival rates after 5 years: patients without APN: 76%; patients with APN: 75.6%).

Figure 2.

Frequency of APN and Kaplan-Meier analysis of graft survival after transplantation. (A) Cumulative incidence of APN (expressed in % of total renal grafts). (B–D) Kaplan-Meier analysis was performed on the whole population of patients (male + female) with (+APN) and without APN (-APN) (B) and the same group of patients stratified by gender (C and D).

Factors associated with acute pyelonephritis

We then analyzed the various groups of transplant patients to find out which factors were associated with the occurrence of APN (Table 1). Significant differences were found between the three groups for the gender of the recipient (p = 0.0017), acute rejection episodes (p = 0.0003) and in patients who developed recurrent UTIs (p < 0.001). The multivariate analysis also indicated that female recipient (risk ratio: 5.143, 95%CI: 1.862–14.206, p = 0.0017), acute rejection episodes (risk ratio: 3.84, 95%CI: 1.371–10.790, p = 0.0105), and the number of UTIs (risk ratio: 1.177, 95%CI: 1.061–1.306, p = 0.0021) represented independent risk factors associated with APN.

Influence of acute pyelonephritis on long-term graft function

To find out whether uncomplicated UTIs and/or APN could impair renal graft function, the variations in serum creatinine and CrCl values were analyzed in transplanted patients over the posttransplant follow-up period (Figure 3). The time of occurrence of APN and rejection episodes temporally associated with APN on the evolution of serum creatinine levels for individual patients are illustrated in Figure 3A. In two cases, APN with rejection was followed by a dramatic increase in serum creatinine. In all other patients, serum creatinine frequently tend to increase progressively after APN (Figure 3A). As a result, mean serum creatinine values were significantly greater (p < 0.01) one year after renal graft in patient Group 3 (2.01 ± 0.42 mg/dL) than in patient Group 1 (1.59 ± 0.51 mg/dL) and patient Group 2 (1.60 ± 0.63 mg/dL). The levels of CrCl did not differ significantly in the three groups of patients for the first three months after surgery (Figure 3B). One year after transplantation, CrCl was significantly lower (p < 0.01) in patient Group 3 with APN than in the other patient Groups 1 and 2 with or without uncomplicated UTIs, respectively (aMDRD; Group 1: 56.4 ± 20.5; Group 2: 54.6 ± 21.7; Group 3: 39.5 ± 15.5 mL/min/1.73 m2). No differences in the levels of CrCl were observed in the six patients who experienced rejection episodes less than 3 months before or after the episode of APN as compared to those of patients from Group 3 who did not (aMDRD; 41.9 ± 22.1 vs. 38.6 ± 13.4 mL/min/1.73 m2). Thereafter, CrCl values always remained significantly lower in patient Group 3 with APN than in the other two groups (Figure 3B). Four years after renal transplantation, the mean CrCl value for the patient Group 3 was about 50% lower (p < 0.01) than that found for patient Group 1 who did not develop UTI or patient Group 2 who exhibited uncomplicated UTIs (Figure 3B). Similar results were found using the C&C method (not shown). To better assess the deleterious consequence of APN on renal graft function, CrCl levels measured just before (time 0), six months and one year after the occurrence of the first episode of APN in patient Group 3 were compared with corresponding CrCl values of patients with no or simpe UTI (Groups 1 and 2). The time 0 chosen for Groups 1 and 2 corresponded to the mean time delay of occurrence of the first APN episode in patients Group 3 (i.e. 6.5 months posttransplantation). The results confirmed that the occurrence of APN was associated with significant delayed decrease in CrCl (Figure 3C). Table 3 outlines the risk factors associated with the decline of renal graft function. Univariate analysis identified the following risk factors: proteinuria at 1 year posttransplantation, diabetes mellitus, acute rejection episodes and APN. Multivariate analysis revealed that APN represented the only independent risk factor (p = 0.0034) associated with the decline in renal function (Table 3). Although of borderline significance, acute rejection did not appear to be a significant risk factor (p = 0.0643), suggesting that APN represents a factor associated with renal function degradation independently of acute rejection. Furthermore, the risk of decline of the renal graft function increased by 3.4-fold in patient Group 3 with APN. Multivariate analysis using the Cox model on the whole cohort of transplanted patients indicated that the factors associated with increased graft loss were: donor age (p = 0.035), graft origin (deceased or live donor) (p = 0.0173) and aMDRD at year 1 posttransplantation (p = 0.037). Neither APN nor acute rejection episodes appeared to be independent factors of graft loss.

Figure 3.

Influence of APN on serum creatinine and creatinine clearance (CrCl) in renal grafts. (A) Individual variations of serum creatinine levels in patient Group 3 with APN. Arrows indicate the time of occurrence of APN. Open circles correspond to serum creatinine values measured during the APN episode. Stars indicate the rejection episodes temporally (before or after) associated with APN. (B) Variations of CrCl values measured using the aMDRD formula in patient Group 1 (no UTI, –UTI); patient Group 2 (uncomplicated UTI, +UTI –APN) and patient Group 3 (+APN). (C) Variations in mean CrCl values in patients Group 3 (+APN) before, and 6 months and 1 year after the first episode of APN and compared to corresponding values from patients with no UTI (–UTI) or uncomplicated UTI (+UTI –APN). * p < 0.01 Group 3 versus Groups 1 and 2.

Table 3.  Univariate and multivariate analysis of factors associated with a decline of renal graft function between 3 months and 1 year after renal transplantation
Univariate analysisPatient groups Δcreat ≤ 19μmol/L (n= 81)Δcreat > 19 μmol/L (n= 26)p
  1. APN = acute pyelonephritis; post-T = posttransplantation; Δcreat = difference in serum creatinine between year 1 and month 3 posttransplantation.

  2. *Number of patients in each group.

  3. **Values are given as means.

Male recipients*5917NS
Ethnic groups (Africans vs. others*)187NS
Recipient age at transplantation (year**)45.644.8NS
Deceased donors*6923NS
Donor age (year**)48.951.8NS
Male donor*5217NS
Anti-PRA >20%*53NS
Hepatitis C virus seropositivity*64NS
Duration of cold-ischemia (h**)20.120.6NS
Patients with DGF*116NS
Hospitalization for CMV*61NS
BMI at month 3 post-T**24.623.1NS
Proteinuria at month 3 post-T (g/24 h**)0.290.69NS
BMI at year 1 post-T**25.124NS
Proteinuria at year-1 post-T (g/24 h**)0.461.440.071
Antidyslipidemia treatment at month 3 post-T*196NS
Acute rejection*26130.192
Diabetes*330.1404
Antihypertensive treatment at month 3 post-T*5816NS
APN*1990.040
Multivariate analysisRR95%CIp
APN3.391.19–17.900.0035

Discussion

Patients who have undergone renal transplantation are highly susceptible to UTIs, and in the present study 18.7% of the patients with UTIs went on to develop APN, which is a potent risk factor for the deterioration of renal graft function. Renal grafts are particularly susceptible to the direct and/or indirect consequences of UTI. There is greater risk of bacterial invasion of the transplanted kidney, because of the immunosuppression and vulnerability of the graft that follow surgical manipulation (3–5). Intraoperative ureteral stents and bladder catheter may also increase the risk of UTIs in transplanted patients (5, 12).

The cumulative incidence of UTIs differs widely from one study to another, ranging from 6% to 86% (11, 17–19). In the present study, UTIs occurred in about 75.1% of the renal recipients over a 5-year period. Such a high incidence of UTI may be explained by the frequency of urine sampling, and the criteria used to define UTI. Furthermore, the relatively high percentage of uropathogens resistant to the trimethoprim-cotrimoxazole may explain, at least in part, the relative high incidence of UTI and APN during the 3 months following transplantation and raised the question of a more adapted antibiotic prophylactic treatment.

The factors associated with an increased risk of UTI after renal transplantation are still controversial. We show that among risk factors, female sex and acute rejection episodes are associated with a higher incidence of APN. Giral et al. (11) also reported that female recipients had a higher risk than the male recipients of developing APN. UTI is not uncommon in patients with renal failure secondary to diabetic nephropathy, and the immunosuppressive effect of corticosteroid therapy and steroid-induced diabetes are both thought to increase the susceptibility to infection. However, in our study, underlying kidney disease, such as diabetic nephropathy or polycystic kidney disease, did not appear to influence the incidence of UTI and APN in renal grafts (1). Likewise, the different immunosuppressive regimens had no significant impact on the occurrence of UTIs. The role of the immunosuppressive regimen as an infection risk factor remains controversial. Recent randomized trials comparing combinations of two different doses of sirolimus plus cyclosporin (20) or tacrolimus versus cyclosporin for primary immunosuppression (21, 22) led to the conclusion that these immunosuppressive treatments have no direct influence on the occurrence of infections. However, mycophenolate mofetil-based primary immunosuppression has been recently reported to be associated with a greater incidence of APN (12).

There is also no consensus about the impact of UTIs on graft rejection. It is generally agreed that the increased incidence of infections is proportional to the intensity of immunosuppressant therapy (23–25). Müller et al. (23) reported that transplanted patients displaying chronic rejection had more UTIs than those displaying no apparent signs of rejection. In contrast, Giral et al. (11) reported that rejection episodes do not constitute a risk factor for APN. In the present study, we found a statistical significant association between acute rejection episodes and APN. The rapid onset of APN (six cases) following acute rejection episodes suggests that the pyelonephritis is a direct consequence of more intensive immunosuppressant therapy. On the other hand, the occurrence of three acute rejection episodes rapidly following APN could also suggest that the renal infection may have triggered an immunostimulatory response, which in turn increased the risk of acute rejection. Other similar occurrences of acute rejection shortly after APN have been reported recently (26). However, in most cases the APN was not directly or chronologically linked to acute rejection episodes. Since not all patients with acute rejection develop UTI, we cannot rule out the possibility that variations in the host's immune function may also determine susceptibility to UTIs (25,27). Further studies are needed to investigate the immunological status of the transplanted patients who are particularly susceptible to APN.

The consequences of UTI on long-term graft function are poorly defined. Early APN occurring during the first 3 months following renal transplantation is detrimental to graft outcome independently of acute rejection episodes (11). Later onset UTI has been generally considered to be relatively benign. However, a recent retrospective study on a large cohort of primary renal transplant recipients reveals that even UTIs occurring some considerable time after renal transplantation, can contribute significantly to subsequent mortality and graft loss (8). Kaplan-Meier analysis indicates that single or recurrent episodes of APN did not significantly alter middle-term graft survival. However, the occurrence of APN appears to be associated with a decline of renal graft function. Whether APN represents an independent risk factor for long-term renal graft dysfunction remain to be determined. The results from the multivariate analysis also suggest that APN represent an independent risk factor of the degradation of the renal graft function, independently of acute rejection episodes, which however are more frequent in the group of pyelonephritic patients. However, the present study has some limitations. This is a single center study and the cohort of patients analyzed remains limited. Although all transplanted patients were screened, this is a retrospective study, and it cannot be fully excluded that some data are missing. Moreover, our data are associative by nature and one also cannot exclude that some unidentified factor(s) associated to APN would be the real culprit in the decline of renal graft function. Although the statistical analyses revealed association between APN and the decline of CrCl, they also show, but do not explain, the association between APN and acute rejection episodes. Further studies should therefore be required to confirm on a larger scale, and also to better understand, the links between APN, acute rejection episodes, and the decline of renal graft function. Still, the present findings suggest that APN has a delayed negative impact on renal graft function. More specific attention to prevention and monitoring of recurrent UTIs and/or APN can be expected to be of benefit for the long-term preservation of renal grafts.

Acknowledgments

This work was supported in part by the INSERM and by an Interface INSERM-Assistance Publique-Hôpitaux de Paris (AP-HP) grant (to AV). We thank Mrs. M. Richard for her precious help in data collection.

Ancillary