Survival of patients with transitional cell carcinoma of the ureter and renal pelvis in Balkan endemic nephropathy and non-endemic areas of Serbia

Authors


Dejan Dragicevic, Institute of Urology and Nephrology, Clinical Centre of Serbia, Resavska 51, 11000 Belgrade, Serbia.
e-mail: dragidej@eunet.yu

Abstract

An interesting reminder about TCC of the ureter and renal pelvis in Serbia is presented, comparing endemic and non-endemic areas. The authors found similarities between the areas, but also that survival was influenced by female sex, and tumour size, grade and stage.

Several papers in this section describe the value of imaging in urological cancer, not least the importance of PET in prostate cancer. Other articles are presented dealing with other aspects of prostate cancer.

OBJECTIVE

To evaluate the characteristics and survival of patients with upper urinary tract (UUT) transitional cell carcinoma (TCC) in Serbia, followed for ≥5 years or until death.

PATIENTS AND METHODS

From 1998 to 2005 we analysed 114 cases of pathologically confirmed UUT TCC, divided into two groups according to topographical characteristics, and compared their demographic, clinical and pathological characteristics. The influence of various factors on overall 5-year survival of patients with UUT TCC was also tested. The prognostic value of different variables was assessed by univariate and multivariate Cox proportional-hazard models.

RESULTS

The most important change in demographic characteristics of the patients with UUT TCC in Serbia was a similar proportion of patients residing in areas of Balkan endemic nephropathy (BEN) and non-endemic areas. The median (range) follow-up was 67 (46–88) months. The 5-year probability of survival was 51.2 ± 5.8%. There was a significantly lower probability of 5-year survival for patients with a higher histological grade (P = 0.001), higher T stage (P < 0.001) and tumour size >3 cm (P = 0.001) at diagnosis. In this cohort of patients the independent predictors of a poorer outcome of the disease were being female (hazard ratio, HR, 2.2, P = 0.010), tumour size >3 cm (HR 2.8, P = 0.001) and T3 or T4 stages (HR 3.1, P = 0.001).

CONCLUSION

Comparative analysis of the characteristics of UUT TCC between patients from BEN and non-endemic areas of Serbia showed similarities in demographic, clinical and pathological features. Factors that significantly influenced survival of patients with UUT TCC were being female, tumour size and tumour grade and stage.

Abbreviations
UUT

upper urinary tract

BEN

Balkan endemic nephropathy.

INTRODUCTION

Upper urinary tract (UUT) TCCs are uncommon; they account for 5–7% of urothelial tumours, with most being renal pelvic lesions [1]. Ureteric tumours are less common, occurring with a quarter of the incidence of renal pelvic tumours [2]. An unusually high incidence of UUT tumours has been reported from Balkan countries, which have restricted areas of Balkan endemic nephropathy (BEN). BEN was recognized ≈40 years ago as an interstitial non-inflammatory disease of the kidney [3]. In a review of published reports, about two-thirds of patients with UUT TCC in Serbia were from BEN areas [3–6]. Demographic, clinical and pathological characteristics of patients with UUT TCC from BEN areas were reported to be different from those from non-endemic regions, i.e. patients from BEN regions were generally younger, more often had renal failure, and lower grade and stage tumours, and more frequent bilateral and multiple tumours than patients from non-endemic regions [3,4]. However, in the last decade, the incidence of BEN and UUT TCC in BEN regions of Serbia appears to be decreasing [7,8]. However, it remains unknown whether changes in the incidence are also reflected by changes in demographic, clinical and pathological characteristics of UUT TCC. Therefore, demographic, clinical and pathological characteristics of patients with UUT TCC in both BEN and non-endemic areas need to be studied in view of changes in BEN and BEN-associated tumour incidence.

Among demographic, clinical and pathological characteristics of patients with UUT TCC, of particular interest are those shown to influence survival. According to the conclusions of a recent review by Kirkali and Tuzel [2], which analysed data on the survival of patients with UUT TCC in studies conducted between 1989 and 2002, tumour grade and pathological stage appear to be the most significant factors for survival. The data on factors influencing survival of patients with UUT urothelial cancer from BEN and non-endemic areas in Serbia are scarce and refer to periods before the changes in BEN and BEN-associated tumours incidence had been reported [9]. Thus, the question of the role of various patient and/or tumour characteristics influencing prognosis of patients with UUT TCC has also to be answered.

In the present study, we analysed 114 cases of pathologically confirmed UUT TCC that were divided into the two groups according to topographical characteristics, and compared their demographic, clinical and pathological characteristics. The influence of various factors on overall 5-year survival of these patients was also tested.

PATIENTS AND METHODS

We enrolled 129 consecutive patients, admitted to the Institute of Urology and Nephrology, Clinical Centre of Serbia, Belgrade (a Primary National Referral Centre for urological malignancies) from January 1998 to December 2002 with a clinical diagnosis of UUT TCC. After pathological examination, 15 patients were excluded for the following reasons: tuberculosis (two patients), RCC (10) and calculosis (three). We selected 114 patients who all had urothelial tumour on review of the pathological findings. The data collected included: age, sex, BEN or non-endemic area of residence, serum creatinine levels, blood haemoglobin levels, tumour location, presence of synchronous bladder tumour, tumour size determined by ultrasonography, tumour grade according to the WHO system [10], tumour stage according to TNM classification [11], and lymphovascular invasion at diagnosis. Criteria for the areas of BEN were the same as those used in previous studies [3,6–8]. Included in the analysis were patients with permanent residence in BEN or non-endemic areas from their birth to the end of follow-up. The overall survival period was estimated from the time of pathological diagnosis to 31 December 2005, if patients were still alive, or to the date of death, or to the date of the last follow-up for those who were lost to follow-up.

All patients gave written informed consent to participate in the study, which was approved by the Institutional Review Board of the Faculty of Medicine, University of Belgrade.

Survival was analysed for the total cohort of patients, as well as in patients subgrouped according to different categories of variables. We defined the time of pathological diagnosis as time zero and death as the endpoint. The cumulative survival probability was calculated by the Kaplan–Meier method. The log-rank test was used for assessing differences in survival according to different categories of variables [12].

The prognostic value of different variables was assessed by univariate and multivariate regression analyses using the Cox proportional-hazard model [13]. At the first step, univariate analysis of different variables in relation to the other variables was applied. The variables with P < 0.05 in the previous step were entered into a multivariate regression model, to identify those with independent prognostic influence on survival in our patient cohort. A stepwise procedure was used.

RESULTS

The patients’ demographic and clinical profiles are presented in Table 1. In all, 114 patients (52 men and 62 women, male : female ratio of 1 : 1.2) diagnosed with UUT TCC were analysed. The median (range) age was 67 (38–86) years; 49 (42.9%) patients were from BEN areas and 53 (46.5%) from non-endemic areas of Serbia. Anaemia and renal failure were identified in 44 (38.6%) and 76 (33.3%) patients, respectively. The carcinomas were located in the renal pelvis in 37 patients and in the ureter in 30, with a ratio of renal pelvis : ureter incidence at our Institute of 1.2 : 1. There were multiple tumours in 36 patients (32%) and 11 (10%) had bilateral tumours. An association with bladder tumours was identified in 28 patients (25%) and 41 (36%) had tumours of >3 cm. Analysis of the pathological characteristics of the UUT TCC including grade and T stage, showed that most patients had G2 (43%) or G3 tumours (47%) and T2 (33%) or T3 (48%) stage. Lymphovascular invasion was present in 39 (34%) patients.

Table 1. 
The characteristics of 114 patients with UUT tumours
CharacteristicN
  1. SCr, serum creatinine; *Patients who changed area of residence; †Data on tumour size were available for 95 of 114 patients.

Median (range) age, years67 (38–86)
Sex, M/F52/62
From BEN area:
 Yes49
 No53
 Others*12
Anaemia at diagnosis:
 Yes (haemoglobin ≤115 g/L)44
 No (haemoglobin >115 g/L)70
Renal function at diagnosis:
 Normal (SCr < 130 µmol/L)76
 Moderate insufficiency (SCr > 130 µmol/L)31
 Haemodialysis 7
Tumour location:
 Renal pelvis (solitary)37
 Ureter (solitary)30
 Multiple tumours36
 Bilateral 11
Association with bladder tumours:
 Yes28
 No86
Tumour size:
 ≤3 cm54
 >3 cm41
Tumour grade:
 111
 249
 354
Tumour stage:
 T112
 T238
 T355
 T4 9
Presence of lymphovascular invasion:
 Yes39
 No75

The results of the topographical analysis of patients’ characteristics is shown in Table 2; there were no significant differences in age, anaemia, and renal function between patients with UUT TCC from BEN and non-endemic regions. Female patients were predominant in the UUT TCC population from BEN areas compared to patients from non-endemic regions, where both sexes were equally distributed. Comparative analysis of tumour characteristics showed a significant difference in tumour location (P = 0.023), i.e. patients from BEN areas had a higher frequency of ureteric tumours, while those from non-endemic areas had higher proportions of multiple and bilateral tumours. There was no difference in the frequency of other tumour characteristics studied. Thus, analysis of indices of UUT TCC malignant potential showed a similar distribution of various tumour stages (T1-4) and grades (G1-3) in patients from BEN and non-endemic areas.

Table 2.  Comparative analysis of clinical variables between patients with UUT tumours from the BEN region and non-endemic regions
VariableBEN regionNon-endemic regionsP
  1. NS, not significant.

N4953 
Mean (sd) age, years66.1 (9.2)66.2 (10.6)NS
Sex, M:F 1 : 1.7 1.1 : 1NS
N (%):
Male18 (37)28 (53) 
Female31 (63)25 (47) 
Anaemia at diagnosis (haemoglobin ≤115 g/L)24 (49)18 (34)NS
Renal function at diagnosis:  NS
 Normal (SCr < 130 µmol/L)30 (62)36 (68) 
 Moderate insufficiency (SCr > 130 µmol/L)16 (33)13 (24.5) 
 Haemodialysis 3 (5) 4 (7.5) 
Tumour location:  0.023
 Ureter (solitary)18 (37) 8 (15) 
 Renal pelvis (solitary)16 (32)17 (32) 
 Bilateral 4 (8) 7 (13) 
 Multiple 11 (22)21 (40) 
Association with bladder tumours10 (20)18 (34)NS
Mean (sd) tumour size, cm 3.58 (2.09) 3.95 (4.51)NS
N (%):
Tumour grade:  NS
 1 5 (10) 4 (7.5) 
 220 (41)23 (43.5) 
 324 (49)26 (49) 
Tumour stage:  NS
 T1 8 (16) 3 (6) 
 T216 (33)18 (34) 
 T323 (47)27 (51) 
 T4 2 (4) 5 (9) 
Lymphovascular invasion17 (34)16 (30)NS

The median (range) follow-up was 67 (46–88) months. The 5-year probability of survival was 51.2 ± 5.8%. The Kaplan–Meier survival curves were drawn for all variables investigated. Only those variables with a significant difference between their categories are presented in Fig. 1. The 5-year overall survival rates for patients with UUT TCC with histological grade 1, 2 and 3 was 87.5%, 63.1%, and 33.8%, respectively. The 5-year overall survival rates of patients with UUT TCC with tumour stage 1, 2, 3 and 4 was 90.0%, 58.9%, 39.1% and 0%, respectively. As shown, there was a significantly lower probability of 5-year survival for patients with UUT TCC with a higher histological grade (P = 0.001, Fig. 1A), higher T stage (P < 0.001, Fig. 1B) and tumour size of >3 cm (P = 0.001, Fig. 1C) at diagnosis. However, among patients with T3 stage, there was no significant difference in 5-year survival according to histological grade (Fig. 1D).

Figure 1.

Kaplan–Meier estimated overall survival curves for patients with UUT TCC with different tumour grades (A), T stages (B), tumour size (C) and patients with T2 stage divided according to tumour grade (D).

According to the univariate logistic regression analysis of the prognostic factors (Table 3), being female (P = 0.019), a tumour of >3 cm (P = 0.001), higher histological grade (P = 0.001) and T stage (P = 0.001) were significantly associated with a worse prognosis of UUT TCC. All these variables were included in the multivariate regression model, as a final step in the analysis. Being female (hazard ratio, HR, 2.2; 95% CI 1.2–4.1; P = 0.010), a tumour size of >3 cm (2.8; 1.5–5.3; P = 0.001) and T3, 4 stages (3.1; 1.6–6.3; P = 0.001) were independent predictors of a poorer outcome of the disease in our patients.

Table 3.  Univariate analysis of prognostic factors in patients with UUT tumours
VariableHR (95% CI)P
Age (<65 vs ≥65 years)1.7 (0.9–3.2)0.136
Sex (female vs male)2.1 (1.1–3.7)0.019
From BEN area (yes vs no)1.4 (0.8–2.6)0.293
Anaemia at diagnosis (yes vs no)0.6 (0.4–1.2)0.142
Renal function at diagnosis(normal vs failure)1.0 (0.5–1.8)0.910
Tumour location (solitary vs multiple)1.2 (0.7–2.2)0.500
Association with bladder tumours (yes vs no)5.3 (0.7–38.6)0.100
Tumour size (≤3 cm vs >3 cm)3.2 (1.7–5.9)0.001
Tumour grade (grade 1, 2 vs 3)2.9 (1.6–5.5)0.001
Tumour stage (T1, 2 vs T3, 4)3.2 (1.6–6.3)0.001
Presence of lymphovascular invasion (yes vs no)0.7 (0.4–1.2)0.212

DISCUSSION

The results of the present cohort study that included 114 patients with UUT TCC from BEN and non-endemic regions of Serbia showed that the demographic, clinical and pathological characteristics of patients with UUT TCC diagnosed between 1998 and 2002 changed compared to previous studies conducted in the same area. The topographical analysis showed no differences in patient or tumour characteristics between patients from BEN and non-endemic areas. The factors that significantly influenced survival of patients with UUT TCC were being female, tumour size, tumour grade and stage.

The age of the patients with UUT TCC in the present study was similar to that in previous studies in this area [3,6,9] and to studies of patients with UUT TCC in Spain and Taiwan [14,15]. The almost equal distribution of male and female patients is also consistent with the data of several earlier studies from this region [16,17]. However, the sex distribution of patients with UUT TCC in Serbia differs significantly from the trend recently reported in Spain and Taiwan, where male patients were predominant [14,15]. The most important change in the demographic characteristics of patients with UUT TCC in Serbia is the similar proportion of patients residing in regions of BEN and non-endemic regions. In previous studies, most patients with UUT TCC in Serbia were from BEN areas, where the incidence of these tumours was reported to be 100 times higher than in non-endemic regions [3]. Further analysis of the association between the geographical region and the demographic characteristics of the patients with UUT TCC showed no differences in terms of age, while more women were among patients from BEN regions.

In the present study, we also analysed the distinct clinical profile of patients with UUT TCC; the proportion of patients with renal failure in the overall population examined was slightly lower than in previous studies in this region [5,6]. These results can be partly explained in that renal insufficiency often follows UUT TCC in patients from BEN regions, who were less frequent in the present study population than in previous studies. However, the proportion of moderate-to-terminal renal insufficiency in the present study was relatively high. The present results are in agreement with a recent study by Yang [15]. It is also important that in addition to nephritis, renal failure in patients with UUT TCC might occur due to obstruction by the tumour. As expected, the high frequency of renal insufficiency was also followed by a high proportion of anaemic patients in our study. However, in the present topographical analysis there were no differences in renal failure and anaemia between patients from BEN and non-endemic areas. Such data suggest changes in the characteristics of patients with UUT TCC from BEN areas, who were earlier reported to have a higher frequency of renal insufficiency and anaemia than those from non-endemic regions [5].

In the present study, the ratio of renal pelvis : ureteric UUT TCC was lower than in previous reports in Serbia [4–6]. Multiple tumours occurred with the same frequency (32%) as reported in previous studies (23–43%) in our region [4–6]. Similarly, the frequency of bilateral tumours diagnosed was the same (10%) as in earlier studies (8–10%) [4,6]. However, more patients with bilateral tumours were from non-endemic areas. Analysis of indices of UUT TCC malignant potential showed a higher frequency of high-grade and high-stage tumours than previously reported [4–6]. Another interesting phenomenon is that there were no differences between patients from BEN and non-endemic areas in the distribution of any of tumour characteristics tested. These data suggest that the natural history of UUT TCC in patients from BEN areas has changed and that they have more aggressive behaviour than before.

The 5-year survival in the present study was 51.2% and, as expected, advanced stage was associated with significantly worse survival. The rate obtained in the present study appears to be lower than suggested in earlier studies that reported 67.3%[18], 70.5%[19] and 75%[20] overall 5-year survival rates for patients with UUT TCC. Our survival rates according to histological grade and T stage of the tumour correspond with those analysed in a recent review reported by Kirkali and Tuzel [2]. There have been several attempts to explain reasons for the lower survival rate in our region compared to other populations. Thus, it was [8] suggested that a more aggressive behaviour of UUT TCC tumours could be a consequence of specific environmental exposure. An additional explanation might be that tumours in the present study were identified in advanced stages and during the study the region of Serbia was facing a poor economical situation, which affected all levels of health care.

The prognostic factors for UUT TCC have been evaluated by several previous studies: T stage, histopathological grade, lymph node status and vascular invasion were reported as independent prognostic factors [15,21,22]. From univariate analysis in the present study, being female, tumour size, histological grade and T stage at diagnosis were significant prognostic factors. On multivariate analysis, being female (HR 2.2, P = 0.010), tumour size of >3 cm (HR 2.8, P = 0.001) and T3, 4 stages (HR 3.1, P = 0.001) proved to be independent prognostic factors, whereas tumour grade according to the stepwise analysis was not statistically significant. Our results on tumour stage, as a very strong predictor of prognosis, are in accordance with those recently reported by Holmang and Johansson in Sweden [21]. Moreover, in more advanced disease, T stage might influence prognosis more effectively, and the effect that other factors, e.g. histological grade, had on prognosis was concealed. This notion is based on the lack of a significant difference in survival rates between different histological grades in patients with T3 tumours in the present study. Conversely, according to the best of our knowledge, this is the first report indicating a worse prognosis for female patients with UUT TCC. Although the reasons for this remain unknown, interestingly a significantly higher risk for UUT TCC in female patients was reported in several recent studies [15,20]. However, additional studies are necessary to explain such a phenomenon. In the present study, we confirmed that the prognostic impact of tumour size is associated with T stage.

ACKNOWLEDGEMENTS

This work was supported by a grant 145009DJ from the Serbian Ministry of Science and Environmental Protection.

CONFLICT OF INTEREST

None declared. Source of funding: Serbian Ministry of Science Environmental Protection (Grant no. 14500909 DJ).

Ancillary