The impact of preoperative serum C-reactive protein on the prognosis of patients with upper urinary tract urothelial carcinoma treated surgically


Kazutaka Saito, Department of Urology, Graduate School, Tokyo Medical and Dental University, 1-5-45, Yushima, Bunkyo-ku, Tokyo, 113–8519, Japan.



To assess the impact of preoperative C-reactive protein (CRP) levels on the prognosis in patients with upper urinary tract (UUT) urothelial carcinoma (UC) primarily treated surgically, as it is increasingly recognized that a systemic inflammatory response is associated with the prognosis for patients with various malignancies.


The clinical records of 130 patients treated surgically for UUT-UC were reviewed retrospectively. An elevated CRP was defined as >0.5 mg/dL. Actuarial survival curves were calculated by Kaplan–Meier method, with the difference between curves evaluated using the log-rank test. A multivariate analysis was used to identify prognostic factors, with Cox’s proportional hazard model.


The median (range) follow-up was 47 (3–190) months. The preoperative serum CRP level was elevated in 24 patients (23%). There were significant associations between CRP level and haemoglobin concentrations, pathological T stage, tumour grade, lymph node involvement and lymphovascular invasion. The 5-year disease-specific and recurrence-free survival rates of 24 patients with elevated CRP were significantly worse than those of the 106 with no CRP elevation (both P < 0.001). On multivariate analysis, preoperative CRP level, pathological T stage and lymph node involvement were significant prognostic factors for disease-specific and recurrence-free survival.


This study indicated that an elevated preoperative CRP level predicts a poor survival in patients with UUT-UC.


upper urinary tract


urothelial carcinoma


C-reactive protein


lymphovascular invasion


lymph node involvement.


The incidence of upper urinary tract urothelial carcinoma (UUT-UC) has increased over the last 20 years in the USA [1], although UUT-UC is relatively uncommon, accounting for 5–7% of all urothelial malignancy [2]. In general, the prognosis for UUT tumours is poorer than that for bladder tumours. A disease-specific 5-year survival rate is 60–80% after surgical management of organ-confined renal pelvic UC [3]. Hall et al.[4] reported that 27% of patients had recurrent disease within a few years after surgery. Although systemic adjuvant chemotherapy with methotrexate, vinblastine, doxorubicin and cisplatin is widely applied for treating UC [5,6], its moderate to severe side-effects make it difficult for many patients. Specific indicators for the prognosis of the patients should be evaluated to provide a better therapeutic approach [7].

Previous reports show that pT stage, pathological grade, tumour location, lymph node involvement (LNI), lymphovascular invasion (LVI) and surgical procedure are prognostic factors for UUT-UC [4,8–11]. However, all of these are postoperative factors. To our knowledge, no preoperative biomarker that can predict prognosis in patients with UUT-UC has been fully elucidated. Identifying preoperative prognostic factors, including a serum biomarker, would allow a better therapeutic approach.

It was reported that the presence of a systemic inflammatory reaction is associated with a poor prognosis in various malignancies. The serum level of C-reactive protein (CRP), an acute-phase reactant, has been reported as a prognostic indicator for some malignancies [12]. We recently reported that the serum CRP level is an independent prognostic factor in patients with localized RCC [12]. However, it has not been clarified whether the serum CRP level is of prognostic value in patients with UUT-UC. In the present study, we investigated the impact of preoperative CRP level on the prognosis in patients with UUT-UC treated surgically in one institution.


In all, 164 consecutive patients with newly diagnosed UUT-UC were treated surgically at our institution between 1990 and 2005: five with distant metastasis at diagnosis were excluded from the study and eight were lost to follow-up within 3 months after surgery; 17 were also excluded because their preoperative CRP levels were unavailable. Three patients with pyelonephritis and another patient with chronic hepatitis, both of which affect serum CRP levels, were also excluded. Consequently, the remaining 130 patients were included in the present study. Nephroureterectomy with removal of the bladder cuff was performed in 124 patients (95%), while partial ureterectomy was used in six selected patients (5%). The regional lymph nodes were dissected in patients with enlarged nodes on evaluation before or during surgery; an extended lymphadenectomy was not used routinely.

Nephroureterectomy specimens were macroscopically and microscopically examined to determine the extent of invasion, the degree of lymph node metastasis and LVI. Tumour stage and grade were determined according to the TNM system [13] and the WHO grading system [14], respectively.

The follow-up after surgery included urinary cytology and cystoscopy every 3 months for 2 years, and then every 6 months. CT and chest X-rays, and/or MRI were used every 6 months for 5 years and annually thereafter.

Serum CRP was measured by latex agglutination using a CRP-L kit (Mitsubishi Kagaku Iatron Co. Ltd, Tokyo, Japan). We defined patients with a serum CRP of >0.5 mg/dL as elevated, as previously reported [12].

Continuous values are expressed as the median (interquartile range) unless indicated otherwise. Associations were evaluated using Fisher’s exact test for nominal valuables and one-way anova for continuous variables. Overall, disease-specific and recurrence-free survival curves were calculated by the Kaplan–Meier method, and the curves compared for clinicopathological variables using the log-rank test. Preoperative factors included age, sex, laterality of the tumour, haemoglobin concentration and serum CRP level. Age was divided at 70 years according to the median age of the patients. Postoperative factors were pT stage, LNI, tumour grade, LVI and the location of the tumour (presence or absence of ureteric involvement). Factors related to survival were analysed by Cox proportional hazards regression models. In a multivariate analysis, pT, tumour grade and LNI were divided into two groups (pTa-2 vs pT3/4; tumour grade G1/2 vs G3; and LNI, positive vs negative, respectively). For all analyses, differences were considered significant at P < 0.05.


The clinicopathological characteristics of the 130 patients are summarized in Table 1; the disease-specific and recurrence-free survival curves for all patients are shown in Fig. 1. During the follow-up of 47 (3–190) months, 55 of the 130 patients (42%) died. The 5- and 10-year overall survival rates were 61% and 48%, respectively. During the follow-up 34 of the 130 patients (26%) died from disease; the 5- and 10-year disease-specific survival rates were 70% and 68%, respectively. The disease recurred in 59 (45%) of the 130 patients during the study period; the primary recurrence site was the bladder in 24 (41%), distant in the lung, bone, liver and lymph nodes in 21 (36%), local (tumour bed, soft tissue, or lymph nodes) in 11 (19%) or UUT in three (5%). The 5- and 10-year recurrence-free survival rates were 52% and 46%, respectively.

Table 1.  Variables in patients with UUT-UC treated surgically
Variable/categoryTotalN (%) of patients with CRPP
  • *

    patients with serum CRP of >0.5 mg/dL.

Age, years
 <70 65 53 (50)12 (50) 
 ≥70 65 53 (50)12 (50)1.000
 Male 88 72 (68)16 (67) 
 Female 42 34 (32) 8 (33)0.901
 Left 66 56 (53)10 (42) 
 Right 63 49 (46)14 (58) 
 Bilateral  1  1 (1) 00.522
Haemoglobin, g/dL
 ≥12 88 79 (75) 9 (37) 
 <12 42 27 (26)15 (63)0.001
 a−2 63 57 (54) 6 (25) 
 3/4 67 49 (47)18 (75)0.013
 negative106 98 (93) 8 (8) 
 positive 24  11 (46)13 (54)<0.001
 1/2 67 62 (59) 5 (21) 
 3 63 44 (42)19 (79)0.001
 negative 66 62 (59) 4 (17) 
 positive 64 44 (42)20 (83)<0.001
Ureteric involvement
 negative 55 45 (43)10 (42) 
 positive 75 61 (58)14 (58)1.000
Surgical procedure
 nephroureterectomy124102 (96)22 (92) 
 partial ureterectomy  6  4 (4) 2 (8)0.336
Figure 1.

Overall (red), disease-specific (blue) and recurrence-free (yellow) survival curves for the entire cohort. The number of patients at risk is shown on the horizontal axis at 0, 5 and 10 years.

The serum CRP level was elevated in 24 patients (23%); the haemoglobin concentration in this group was significantly lower than in those with normal CRP levels (P < 0.001), and the pT and grade were higher (P < 0.001 and P = 0.004, respectively). The frequency of LNI and LVI in the elevated CRP group was significantly higher than in the normal group (both P < 0.001).

The frequency of disease recurrence in the elevated CRP group (75%, 18/24) was significantly higher that that in patients with no CRP elevation (39%, 41/106; P = 0.001). As shown in Fig. 2a, the 5-year recurrence-free survival rate of the 24 patients in the elevated CRP group (21%) was significantly worse than that of the other 106 patients (59%, P < 0.001).

Figure 2.

a, Recurrence-free and b , disease-specific survival curves according to CRP level in patients with UUT-UC; elevated CRP (red) and normal CRP group (blue). The numbers of patients at risk are shown on the horizontal axis at 0, 5 and 10 years.

Recurrence-free survival rates, including bladder recurrence, were compared according to the factors before and after surgery in Table 2. Univariate analysis showed that CRP level, pT LNI and LVI (all P < 0.001) were significantly associated with recurrence-free survival. The multivariate Cox proportional hazards regression model showed that CRP (P = 0.023), pT (P = 0.013) and LNI (P = 0.002) were significant and independent factors for recurrence-free survival. The hazard ratio (95% CI) of the elevated CRP was 1.45  (1.05–1.97). The results were no different even if superficial bladder cancer recurrence was not considered in calculating the recurrence-free survival rate.

Table 2.  Univariate and multivariate analyses of factors predicting recurrence-free and disease-specific survival
Variable/categoryUnivariate, PMultivariate (full model) PMultivariate (reduced model) HR (95% CIP
 Age, years, ≥ 70 vs <700.259
 Sex, male vs female0.547
 Laterality, left vs right0.601
 Haemoglobin, g/dL, <12 vs ≥120.714
 CRP, mg/dL, ≥ 0.5 vs <0.5<0.0010.0331.45 (1.05–1.97)0.023
 pT, pT3/4 vs pTa−2<0.0010.0401.42 (1.09–1.91)0.013
 Grade, 3 vs 1/20.436 
 LNI, +ve vs −ve<0.0010.0041.69 (1.22–2.31)0.002
 LVI, +ve vs −ve0.0020.578
 Ureteric involvement, yes vs no0.174
 Surgical procedure, partial ureterectomy vs nephroureterectomy0.061 
 Age, years, ≥ 70 vs <700.0300.0941.51 (1.07–2.18)0.019
 Sex, male vs female0.141
 Laterality, left vs right0.417
 Haemoglobin, g/dL, <12 vs ≥ 120.0030.301
 CRP, mg/dL, ≥ 0.5 vs <0.5<0.0010.0441.78 (1.21–2.68)0.004
 pT, pT3/4 vs pTa-2<0.0010.0961.77 (1.19–2.68)0.014
 Grade, 3 vs 1/20.0250.816 
 LNI, +ve vs −ve<0.0010.0022.25 (1.49–3.37)<0.001
 LVI, +ve vs −ve<0.0010.151
 Ureteric involvement, yes vs no0.0090.458
 Surgical procedure, partial ureterectomy vs nephroureterectomy0.0070.193

The frequency of disease-specific death in patients with elevated CRP (75%, 18/24) was significantly higher that in the normal CRP group (15%, 16/106; P < 0.001). In the elevated CRP group, all of the patients ultimately died from the disease after recurrence. The 5-year disease-specific survival rate of the 24 patients with elevated CRP (25%) was significantly worse than that of the other 106 (80%, P < 0.001; Fig. 2b).

The disease-specific survival rates were also compared according to factors before and after surgery (Table 2). Univariate analysis showed that age (P = 0.030), preoperative haemoglobin concentration (P = 0.003), CRP, pT, LNI (all P < 0.001), grade (P = 0.025), LVI (P < 0.001), ureteric involvement (P = 0.009) and surgical procedure (P = 0.007) were significantly associated with disease-specific survival. Multivariate Cox proportional hazards regression models showed that age (P = 0.019), CRP (P = 0.004), pT (P = 0.014) and LNI (P < 0.001) were significant and independent predictors of disease-specific survival, with CRP a preoperative predictor of disease-specific survival. The hazard ratio (95% CI) of elevated CRP was 1.78  (1.21–2.68).


In the present study, we show for the first time that a preoperative serum biomarker, CRP, is an independent prognostic factor in patients with UUT-UC. A high CRP level before surgery, with the conventional pathological prognostic indicators, pT and LNI, can predict a poor prognosis in patients with UUT-UC [4,11].

In the present patients, an elevated CRP was associated with progressive tumour characteristics, including high tumour stage and grade, positive LNI and LVI. Therefore, an elevated CRP would predict more aggressive and progressive disease. Furthermore, three-quarters of patients with a high CRP level died from their disease and elevated CRP was an independent preoperative risk factor for disease-specific death. Thus, CRP, an acute-phase inflammatory product, could be a new prognostic marker in UUT-UC. As CRP is a not a specific biomarker it cannot be used in patients with other diseases in which CRP might be elevated, e.g. inflammatory or cardiovascular disease. Our result should be evaluated in larger confirmatory studies.

It is increasingly recognized that disease progression depends on a complex interaction of the tumour and the inflammatory response of the host [15,16]. Indeed, the serum level of CRP is frequently elevated in patients with cancer, and it was reported as a prognostic indicator in various malignancies, including colon, lung, breast, liver, oesophagus and ovary [17–22]. In genitourinary malignancy, we reported that elevated CRP predicts a poor prognosis in patients with RCC [14]. To our knowledge there has been only one study showing that elevated serum CRP is associated with a poor prognosis in patients with UC of the bladder [23].

The mechanisms by which a systemic inflammatory response, as shown by elevated CRP, affects cancer-specific survival are currently unclear. However, it is known that as a part of the systemic inflammatory response to the tumour there is a release of inflammatory cytokines and growth factors, which not only stimulates tumour growth [24] but also produces profound catabolic effects on the host metabolism [25]. CRP liver production is strongly induced by cytokines such as interleukin-1, TNF and interleukin-6. Experimental studies showed that UC cells can produce interleukin-6 [26], and moreover interleukin-6 is recognized as a growth promoter in UC [26]. These findings support the view that a systemic inflammatory response of the host indicates an aggressive nature of UUT-UC.

In conclusion, this is the first report that elevated CRP is associated with a poor prognosis in patients with UUT-UC. The serum CRP level is a simple, reliable, inexpensive and reproducible marker. Thus, the preoperative CRP level is a useful adjunct that can be associated in routine practice with preoperative risk evaluation in patients with UUT-UC.


None declared.