Keiichi Ito md phd, Department of Urology, National Defense Medical College, 3-2 Namiki, Tokorozawa, Saitama 359-8513, Japan. Email: firstname.lastname@example.org
Aim: C-reactive protein (CRP) elevation is reportedly a prognostic factor in patients with renal cell carcinoma (RCC). Thrombocytosis has recently been reported also to be a prognostic factor in RCC and, like CRP, to be related to inflammatory cytokines such as interleukin-6. The aim of this study was to evaluate the importance of both thrombocytosis and CRP elevation in tumor recurrence and prognosis for patients with RCC.
Methods: The clinical records of 178 patients who underwent radical nephrectomy were reviewed. Thrombocytosis was defined as a platelet count ≥350 000/mm3, and CRP elevation was defined as a CRP level ≥1.0 mg/dL. Disease-free survival and cause-specific survival rates were calculated. Independent predictors for recurrence and prognosis were determined.
Results: Patients with thrombocytosis and patients with elevated CRP levels had significantly higher pathological T stage, clinical stage, tumor size, histological grade, and percentage of microvascular invasion than did patients without THC and patients with CRP levels <1.0 mg/dL, respectively. There was a significant correlation between platelet counts and CRP levels. Multivariate analysis showed that distant metastasis, tumor size, grade 3 components, and CRP elevation were independent predictors for prognosis but thrombocytosis was not. In N0M0 RCC patients, tumor size, microvascular invasion, and CRP elevation were independent predictors for recurrence. CRP elevation and tumor size were independent predictors for prognosis.
Conclusions: Platelet count and CRP level are strongly correlated in patients with RCC, but only CRP elevation is an independent predictor for recurrence and prognosis.
Patients with renal cell carcinoma (RCC) sometimes develop an inflammatory paraneoplastic syndrome that is combined with the elevation of acute phase reactants. The levels of these reactants, including that of C-reactive protein (CRP), and erythrocyte sedimentation rate are significantly correlated with the survival rates of patients with RCC,1–3 and CRP elevation has recently been reported to be an important predictor of the prognosis for patients with RCC.4 CRP is a biomarker of survival in patients with metastatic RCC treated with interleukin (IL)-2 based immunotherapy5 and preoperative CRP levels predict the impact of cytoreductive nephrectomy on the prognosis for patients with metastatic RCC.6 Because preoperative CRP is easily measured, it is clinically important to evaluate the significance of CRP elevation.
Platelet count is another easily measured parameter, and thrombocytosis (THC) has also been recently reported to be a significant predictor of the prognosis for RCC patients. It has a significant influence on the cause-specific survival of patients after radical nephrectomy (RNx).7,8 The prognosis for patients with metastatic RCC who show THC is significantly worse than that for patients who do not show THC9 and THC is also a prognostic factor for patients with localized RCC.10
The reactive THC and CRP elevation in RCC are both caused by the production of inflammatory cytokines such as IL-6.11–13 When metastatic RCC is treated with IL-2, serum IL-6 levels and CRP levels are correlated and the survival of patients with detectable IL-6 is shorter than that of patients with undetectable IL-6.14 RCC patients with elevated serum IL-6 frequently have higher platelet counts and higher CRP levels than do those without elevated IL-6,12 and elevated serum IL-6 elevation is also significantly associated with paraneoplastic THC in patients with RCC.11 Both THC and CRP elevation are thus associated with IL-6 production due to RCC, although other cytokines such as vascular endothelial growth factor (VEGF) are also reported to be related to THC in RCC patients.15
The aim of the present study was to evaluate whether RCC recurrence and the prognosis for patients with RCC are similarly impacted by CRP elevation and THC, and to determine whether and how strongly CRP elevation and THC are independent predictors of recurrence and prognosis.
We reviewed the records of the 295 patients at our institution with RCC treated by RNx between 1985 and 2003, and found that both serum CRP levels and platelet counts were available for 178 of these patients. The mean age of these 178 patients (51 female and 127 male) was 59.3 ± 0.9 years. Eighty-three patients underwent right RNx and 95 patients underwent left RNx. All patients were evaluated postoperatively every 3–6 months for the first 5 years and every 6–12 months thereafter to monitor local recurrence and metastasis. Follow-up examinations consisted of physical examination, chest radiography, abdominal and chest computed tomography (CT), blood tests, and if indicated, radionuclide bone scanning. Follow-up intervals were calculated from the date of RNx to the last recorded follow-up (range 1–232 months, mean 44.5 months). The pathological tumor stage and histological grading was determined according to the TNM classification system.16 Regional lymphadenectomy was performed when lymph node swelling was detected by preoperative radiological examination or at operation. Disease-free survival was evaluated using the date at which local recurrence or metastatic disease was identified, and cause-specific survival was evaluated using the date of death due to disease progression or the last follow-up date.
Factors evaluated are listed in Tables 1 and 2. All preoperative factors were measured within 1 month before RNx. THC was defined by a platelet count ≥350 000/mm3. Preoperative CRP elevation was defined by a CRP level ≥1.0 mg/dL. It is difficult to determine an appropriate cutoff value for defining CRP elevation, so we simply used the 1.0-mg/dL cutoff value used in previous studies.6,17 Both platelet count and CRP level were measured in the laboratory of National Defense Medical College Hospital.
Table 1. Characteristics of patients with or without thrombocytosis (THC)
Variables of different groups were compared using the Mann–Whitney U-test. The independence of fit of categorical data was analyzed by the chi-squared test. Values shown as mean ± SE or n.
60.6 ± 1.6
58.8 ± 1.1
Side of tumor location (right/left)
Tumor size (cm)
8.5 ± 0.5
5.1 ± 0.3
Pathological T stage (1–2/3–4)
Pathological T stage
Lymph node metastasis (–/+)
Distant metastasis (–/+)
Presence of grade 3 (–/+)
Microvascular invasion (–/+)
Platelet count (× 104/mm3)
35.7 ± 1.9
23.7 ± 0.6
Results are presented as the mean ± standard error (SE). Variables of different groups were compared using the Mann–Whitney U-test. The independence of fit of categorical data was analyzed by the chi-squared test. The correlation between variables was analyzed using Spearman’s rank correlation coefficients. Survival curves were constructed by the Kaplan–Meier method, and the differences assessed using the log-rank test. Cox’s proportional hazard regression model was used for univariate and multivariate analyses. For all tests a P-value less than 0.05 was considered to indicate statistical significance.
Patient characteristics and pathological findings other than histological types are listed in Table 1. The predominant histological type was clear cell in 140 patients, granular cell in 29, papillary in 4, chromophobe in 3, spindle cell in 1, and unclassified in 1. Thirteen of the 178 patients (7.3%) had metastases in surgically resected lymph nodes, and 28 patients (15.7%) had distant metastasis that were confirmed by pathological diagnosis at operation or were obvious by preoperative radiological examinations.
The mean preoperative CRP level was 2.1 ± 0.3 mg/dL (range: 0.3–20.7 mg/dL) and the mean platelet count was 272 900 ± 7940 mg/dL (range: 86 000–853 000/mm3). Preoperatively, 27 patients (15.2%) had THC and 53 patients (29.8%) had CRP levels ≥1.0 mg/dL.
Patients with THC (Table 1) and patients with CRP levels ≥1.0 mg/dL (Table 2) had larger tumors, higher pathological T stages, higher percentages of lymph node metastasis and distant metastasis, higher histological grades and higher percentages of microvascular invasion (MVI) than did patients without THC and patients with CRP levels <1.0 mg/dL, respectively. The preoperative CRP levels and the platelet counts showed a significant correlation (P < 0.0001) when analyzed using Spearman rank correlation coefficient (Fig. 1).
The cause-specific survival rates of patients with THC were significantly lower than those of patients without THC (P < 0.01) (Fig. 2a), and the cause-specific survival rates of patients with CRP levels ≥1.0 mg/dL were significantly lower than those of patients with CRP <1.0 mg/dL (P < 0.0001) (Fig. 2b).
Cox’s proportional hazard model analysis was used to evaluate factors predicting cause-specific patient survival. Univariate analysis showed that pathological T stage, lymph node metastasis, distant metastasis, tumor size, presence of grade 3 component, MVI, THC, and CRP elevation were significantly associated with cause-specific survival (P < 0.05). Multivariate Cox’s proportional hazard model analysis revealed that distant metastasis (P < 0.0001), tumor size (P = 0.0054), presence of grade 3 component (P = 0.0427), and CRP elevation (P = 0.0008) were significant predictors of cause-specific survival (Table 3).
Table 3. Cox’s proportional hazards regression analysis for predicting cause-specific survival in patients with renal cell carcinoma
Relative risk ratio 95% CI
By 1 cm tumor size increase.
CRP, C-reactive protein; MVI, microvascular invasion; N, lymph node involvement; M, distant metastasis; pT, pathological T stage; THC, thrombocytosis.
We also used Cox’s proportional hazard model analysis to evaluate factors predicting disease recurrence and cause-specific survival in patients without either lymph node metastasis or distant metastasis (N0M0 patients). Univariate analysis demonstrated that T stage, tumor size, presence of grade 3 component, MVI, THC, and CRP elevation were significantly associated with both disease-free survival (P < 0.05) (Table 4) and cause-specific survival (P < 0.05) (Table 5). Multivariate Cox’s proportional hazard model analysis showed tumor size (P = 0.0003), MVI (P = 0.0011), and CRP elevation (P = 0.0003) to be significant predictors of disease-free survival (Table 4). Tumor size (P = 0.0451) and CRP elevation (P = 0.0067, odds ratio: 15.625) were found to be significant predictors of cause-specific survival in N0M0 patients (Table 5).
Table 4. Cox’s proportional hazards regression analysis for predicting disease-free survival in patients with N0M0 renal cell carcinoma
The patients were classified into four groups according to CRP levels and platelet counts: group 1, THC– and CRP <1.0 mg/dL; group 2, THC+ and CRP <1.0 mg/dL; group 3, THC– and CRP ≥1.0 mg/dL; and group 4, THC+ and CRP ≥1.0 mg/dL. Cause-specific survival was evaluated for each group. Patients in group 2 had no recurrence during the follow-up period. With (group 4) or without THC (group 3), patients with CRP ≥1.0 mg/dL had a significantly lower cause-specific survival rate than those with CRP <1.0 mg/dL (groups 1 and 2). Cause-specific survival did not differ significantly between group 3 and group 4 (P > 0.05). These results indicated that CRP level predicted survival in patients with RCC more strongly than platelet count did.
In the present study, both THC and CRP elevation were prognostic factors in univariate analysis predicting tumor recurrence in patients with RCC and the prognosis for those patients. Platelet counts and CRP levels were strongly correlated with each other. Thus, there appeared to be common causes, such as IL-6 elevation, for the THC and CRP elevation in patients with RCC. Multivariate analysis with other prognostic factors revealed that CRP ≥1.0 mg/dL was one of the independent predictors for both recurrence and prognosis. THC, in contrast, was not an independent predictor. Thus, CRP was found to be a more important predictor for recurrence and prognosis than THC. Furthermore, with or without THC, patients with CRP ≥1.0 mg/dL had a poorer prognosis than did those with CRP <1.0 mg/dL. Although previous studies have shown the prognostic significance of THC for RCC patients,7–10 the present study is the first comparing THC and elevated CRP levels in RCC patients and showing that CRP is the more important predictor.
The authors of some previous reports have defined THC as a platelet count ≥400 000,7–10 but O’Byrne et al. defined it as a platelet count ≥337 000,15 and we defined it as a platelet count ≥350 000. We did so because the median platelet count in all 178 patients in the present study was 254 500 and 66.7% of the patients with platelet counts between 350 000 and 400 000 had CRP ≥1.0 mgdL, 55.6% of those had tumor sizes ≥7 cm, and more than half of these patients appeared to have some degree of RCC-associated inflammation.
C-reactive protein levels and platelet counts were significantly correlated. More than 85% of the patients with THC had CRP levels ≥1.0 mg/dL and the average CRP level in patients with THC was 7.4 mg/dL. Because these two factors were strongly correlated, some factors such as IL-6 are probably involved in both situations. There were, however, patients with CRP ≥1.0 mg/dL but without THC and patients with CRP <1.0 mg/dL but with THC. THC and CRP elevation are thus not always correlated. In addition to IL-6, factors such as VEGF and platelet-derived growth factor (PDGF) may be involved in THC. VEGF has been suggested to be related to THC in patients with RCC.15
In the present study, the presence of distant metastasis, a larger tumor size, CRP elevation, and the presence of grade 3 histology were independent predictors for prognosis. Large tumors and tumors associated with CRP elevation or with poorly graded histology might contain renal cancer cells with aggressive biological activities. RCC associated with CRP elevation would produce multiple cytokines including IL-6, would be stimulated through the receptors of those cytokines, and would proliferate using signaling pathways such as the nuclear factor (NF)-κB activation pathway.18 A tumor with aggressive biological activity will progress more rapidly and threaten the patient’s life sooner.
We also evaluated the predictors for recurrence and prognosis in patients with N0M0 RCC. Gogus et al. reported that THC was a significant predictor for predicting the prognosis for patients with localized RCC when a Kaplan–Meier survival curve is used.10 In our analysis of patients with N0M0 RCC, CRP elevation predicted recurrence and prognosis more strongly than THC did. It is also interesting that in patients with localized RCC, MVI is one of the independent predictors for recurrence but not for prognosis. For prognosis, tumor size and CRP elevation are stronger predictors than MVI. MVI would be expected to be related to metastasis. When tumors recur in RCC patients, however, recurrent tumors with aggressive biological activity such as tumors associated with CRP elevation or large tumors will progress rapidly and threaten the patient’s life.
C-reactive protein has recently been reported to be a predictor for the response to IL-2 in patients with metastatic RCC.5 When metastatic RCC is treated with IL-2 immunotherapy, survival is better for patients with low or normal levels of CRP. In metastatic RCC cytoreductive nephrectomy is beneficial to patients whose pretreatment serum CRP is elevated.6 In metastatic RCC patients with CRP ≥1.0 mg/dL, serum immunosuppressive acidic protein decreased significantly and natural killer cell activity increased significantly after cytoreductive nephrectomy, suggesting that nephrectomy may be beneficial in patients having elevated serum CRP levels before surgery.17 It is important that this simple factor, CRP, was an independent predictor for recurrence and prognosis in patients with RCC.
In conclusion, platelet counts and CRP levels are strongly correlated in patients with RCC. In multivariate analysis, CRP elevation is an independent predictor for prognosis but THC is not. Elevated levels of CRP are also an independent predictor for recurrence and prognosis in patients with N0M0 RCC. CRP level is thus a clinically important factor predicting tumor recurrence in RCC patients and predicting their prognosis.