SEARCH

SEARCH BY CITATION

Keywords:

  • C-reactive protein;
  • hyponatremia;
  • molecular targeted therapy;
  • prognostic factor;
  • renal cell carcinoma

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Conflict of interest
  9. References

Objectives:  Hyponatremia is reported to be associated with poor survival in localized renal cell carcinoma and metastatic renal cell carcinoma treated with immunotherapy. However, there are no reports on the relationship between hyponatremia and prognosis of metastatic renal cell carcinoma treated with molecular targeted therapy. We evaluated the prognostic significance of hyponatremia in metastatic renal cell carcinoma treated with molecular targeted therapy as first-line therapy.

Methods:  We retrospectively analyzed a database comprising 87 patients treated from April 2008 to July 2011 with sorafenib or sunitinib as first-line therapy for metastatic renal cell carcinoma. Patients were divided into three groups according to serum sodium level: severe hyponatremia (≤134 mEq/L), mild hyponatremia (135–137 mEq/L) and normal natremia (≥138 mEq/L).

Results:  Median cancer-specific survival time was 8.8 months in the patients with severe and mild hyponatremia, and 32.6 months in the patients with normal natremia (P < 0.001). Multivariate analysis showed severe and mild hyponatremia to be significantly associated with cancer-specific survival (hazard ratio 6.228; 95% confidence interval 2.161–17.947, P = 0.001; hazard ratio 3.374; 95% confidence interval 1.294–8.798, P = 0.013), respectively. Neutrophilia and high C-reactive protein level (C-reactive protein ≥1.0 mg/dL) were significant prognostic factors to predict inferior cancer-specific survival. In Harrell's concordance index calculation, hyponatremia could significantly improve the predictive accuracy for estimation of survival probability (P = 0.028).

Conclusions:  Hyponatremia (<138 mEq/L), neutrophilia and high C-reactive protein levels seem to represent significant predictive factors for cancer-specific survival in metastatic renal cell carcinoma patients treated with molecular targeted therapy as first line therapy. Furthermore, hyponatremia might be significantly associated with chronic inflammation and tumor aggressiveness.


Abbreviations & Acronyms
CI =

confidence interval

CRP =

C-reactive protein

CSS =

cancer-specific survival

ECOG PS =

Eastern Cooperative Oncology Group Performance Status

eGFR =

estimated glomerular filtration rate

HR =

hazard ratio

IL-6 =

interleukin-6

LDH =

lactate dehydrogenase

LLN =

lower limit of normal

mRCC =

metastatic renal cell carcinoma

MSKCC =

Memorial Sloan-Kettering Cancer Center

RCC =

renal cell carcinoma

TNM =

tumor–nodes–metastasis

ULN =

upper limit of normal

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Conflict of interest
  9. References

The prognosis of mRCC is poor, with a 5-year survival rate of <10%.1 Two groups of targeted agents, such as vascular endothelial growth factor-targeted therapies and mammalian target of rapamycin inhibitors, have revolutionized the treatment of mRCC and have largely displaced immunotherapy as the first-line standard of care.2–4 Several prognostic factors for patients with mRCC have been identified to improve the selection of patients for the immunotherapy and molecular targeted therapy.5–7 In previous studies, systematic inflammatory responses, such as neutrophilia, thrombocytosis and elevated CRP, were closely correlated with the prognosis of patients with mRCC.

Serum sodium is an easily obtained and routinely measured plasma electrolyte, but hyponatremia is an often underdiagnosed and untreated electrolyte disturbance. In previous reports, hyponatremia has been shown to predict an adverse outcome in liver cirrhosis,8 congestive heart failure9 and community-acquired pneumonia.10 Furthermore, hyponatremia is reported to be associated with poor survival in several carcinomas, such as hepatocellular carcinoma,11 gastric cancer12 and small cell lung cancer.13 In terms of RCC, hyponatremia has been associated with poor survival in localized RCC and mRCC treated with immunotherapy.14–16 However, there are no reports on the relationship between hyponatremia and the prognosis of mRCC treated with molecular targeted therapy.

The aim of the present study was to evaluate the prognostic significance of hyponatremia in mRCC treated with molecular targeted therapy as first-line therapy. Furthermore, we assessed whether consideration of hyponatremia could improve the predictive ability of other prognostic factors including neutrophilia, platelets and CRP level.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Conflict of interest
  9. References

Patients

We retrospectively analyzed a database comprising 87 patients treated from April 2008 to July 2011 with sorafenib or sunitinib as first-line therapy for mRCC at Osaka University Graduate School of Medicine and its affiliated hospitals listed in the acknowledgments. The initially diagnosed tumors were staged according to the American Joint Committee on Cancer (2002) cancer staging classification,17 and the patients, characteristics, including laboratory findings, were evaluated at the time of drug administration according to modified MSKCC risk groups.5,18

Clinical features evaluated were age, sex, ECOG PS, hemoglobin, corrected serum calcium, serum LDH, time from diagnosis to treatment, prior nephrectomy, number of organs involved in metastasis, TNM stage, modified MSKCC risk groups, neutrophiles, platelets, eGFR,19 CRP level and the molecular targeted therapy administered. The study was approved by an institutional review board of Osaka University, which provided the necessary institutional data-sharing agreements before initiation of the study. CSS time was calculated from the date of initiation of first-line therapy until death or the date of the patient's last follow-up visit. The CSS rates were calculated by the Kaplan–Meier method.

Statistical analysis

The primary aim was to evaluate the prognostic significance of hyponatremia in mRCC treated with molecular targeted therapy as first-line therapy and analyze other prognostic factors including the previously reported clinical features. In accordance with both a cut-off value of hyponatremia frequently used in the literature (serum sodium ≤134 mEq/L)20 and a more conservative definition of hyponatremia (serum sodium ≤137 mEq/L) provided by Kumar and Berl in a cutting-edge article on this issue,21 patients were divided into three groups according to serum sodium level: severe hyponatremia (≤134 mEq/L), mild hyponatremia (135–137 mEq/L) and normal natremia (≥138 mEq/L), as previously reported.22 The other clinical features were defined by sex (male or female), ECOG PS (0 or 1,2,3), hemoglobin (<the LLN or ≥LLN), corrected serum calcium (≤the ULN or >ULN), serum LDH (≤1.5 × ULN or >1.5 × ULN), time from diagnosis to treatment (<1 year or ≥1 year), prior nephrectomy (yes or no), number of organs involved in metastasis (single or multiple), bone metastasis (yes or no), TNM stage (I, II, III, or IV), neutrophiles (≤ULN or >ULN), platelets (≤ULN or >ULN), eGFR (<60 mL/min/1.73 m2 or ≥60 mL/min/1.73 m2), CRP level (<0.3 mg/dL, ≥0.3 to <1.0 mg/dL, or ≥1.0 mg/dL) and molecular targeted therapy (sunitinib or sorafenib). Associations between serum sodium concentration and clinicopathological features were evaluated with Fisher's exact test and logistic regression analysis. We used the Cox regression model to calculate the HR for univariate and multivariate analysis. Prognostic factors related to CSS were analyzed with Cox regression analysis using a step-wise forward selection with P < 0.05 as the criterion for model entry or stay for multivariate analysis.

In addition, likelihood-ratio tests were carried out to evaluate whether the model with established risk factors improved after the addition of severe and mild hyponatremia, as Harrell described. A value of P < 0.05 was considered statistically significant. Statistical analysis was carried out with the Statistical Package for the Social Sciences software, version 16.0 (SPSS, Chicago, IL, USA) and stata ic, version 12.0 (STATA Corporation, College Station, TX, USA).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Conflict of interest
  9. References

Patient characteristics

The clinical and pathological characteristics of the 87 study patients with mRCC are shown in Table 1. Two patients were lost to follow up. The range of serum sodium was 127–145 mEq/L, showing that this cohort did not include any patients with severe hyponatremia (<125 mEq/L) or severe hypernatremia (>155 mEq/L).21 Serum sodium levels were <135 mEq/L in nine (10.3%) patients, 135–137 mEq/L in 19 (21.8%) patients and ≥138 mEq/L in 59 (67.9%) patients. The histology of 52 patients was clear cell carcinoma and that of 13 patients was non-clear cell carcinoma. The histology of 22 patients was unknown. In univariate analysis, the status of hemoglobin, ECOG PS, LDH, neutrophilia, the presence of bone metastasis and platelets were significantly associated with the serum sodium concentration. In multivariate analysis, the status of the presence of bone metastasis and neutrophilia was significantly associated with severe and mild hyponatremia (Tables 1,2). None of the other variables, including the status of prior nephrectomy, eGFR and CRP level, showed a significant association with the serum sodium concentration.

Table 1. Clinical characteristics of 87 patients receiving molecular targeted therapy as first-line therapy according to order of serum sodium concentration (mEq/L)
VariablesTotal (n = 87)Serum sodium concentration status P
Severe hyponatremia (n = 9)Mild hyponatremia (n = 19)Normal natremia (n = 59)
Na < 135135 ≤ Na < 138138 ≤ Na
Age (years)38–87 (median 67.0)41–84 (median 71.0)54–79 (median 69.0)38–87 (median 65.0)0.540
Sex     
 Male61814390.353
 Female261520 
ECOG PS     
 05439420.030
 1.2.33361017 
Hemoglobin     
 <LLN58915340.019
 ≥LLN290425 
Corrected serum calcium     
 >ULN91350.683
 ≤ULN6881446 
 Unknown10028 
Lactate dehydrogenase     
 >1.5 ULN10361<0.001
 ≤1.5 ULN7561356 
 Unknown2002 
Time from diagnosis to treatment     
 <1 year2843210.195
 ≥1 year5951638 
Prior nephrectomy     
 Yes63712440.582
 No242715 
No. metastatic sites     
 Single4778320.258
 Multiple3721025 
 Unknown3012 
Bone metastasis     
 Yes32313160.005
 No556643 
TNM stage     
 I, II, III2836190.996
 IV5961340 
Modified MSKCC risk group     
 Favorable1202100.226
 Intermediate5051035 
 Poor18468 
 Unknown7016 
Neutrophiles     
 >ULN6411<0.001
 ≤ULN8051758 
 Unknown1010 
Platelets     
 >ULN181890.032
 ≤ULN6981150 
eGFR (mL/min/1.73 m2)     
 ≥6034410200.326
 <60535939 
CRP (mg/dL)     
 ≥1.040611230.355
 0.3≤, <1.0161411 
 <0.3302424 
 Unknown1001 
Tyrosine kinase inhibitor     
 Sorafenib49511330.988
 Sunitinib384826 
Follow-up time (months)0.5–35.9 (median 10.3)0.5–10.3 (median 4.2)0.6–35.9 (median 6.6)0.9–35.5 (median 13.5)0.002
Table 2. Clinical characteristics correlated with severe and mild hyponatremia (<138 mEq/L) with logistic regression analysis
VariablesMultivariate
Exponent95% CI P
Presence of bone metastasis5.6921.969–16.457 0.001
Neutrophilia25.4222.543–254.138 0.006

Prognostic factors of CSS in univariate and multivariate analysis

Overall median CSS time was 32.6 months (95% CI 17.4–47.7; Fig. 1). Median CSS time was 4.2 months (95% CI 0.0–12.5) in the patients with severe hyponatremia, 9.3 months (95% CI 1.9–16.6) in the patients with mild hyponatremia and 32.6 months (95% CI 25.9–39.2) in the patients with normal natremia (≥138 mEq/L; P < 0.001; Fig. 2). Both severe hyponatremia (HR 2.915; 95% CI 1.289–6.590; P = 0.010) and mild hyponatremia (HR 7.390; 95% CI 2.877–18.981; P < 0.001) were associated with poor prognosis in univariate analysis (Table 3). Even when stratified according to modified MSKCC risk of intermediate and poor group, the prognosis of the patients with normal natremia was significantly better than that of the patients with severe hyponatremia and mild hyponatremia, respectively (P = 0.05 and P = 0.004) (Fig. 3a,b). In other clinical features, poor ECOG PS, anemia, elevated LDH, TNM stage IV, neutrophilia, impaired renal function and high level of CRP were associated with poor prognosis.

image

Figure 1. Probability estimates of CSS in 87 patients receiving molecular targeted therapy.

Download figure to PowerPoint

image

Figure 2. Probability estimates of CSS rate in 87 patients in three groups based on serum sodium concentration (severe hyponatremia [≤134 mEq/L], mild hyponatremia [135–137 mEq/L] and normal natremia [≥138 mEq/L].

Download figure to PowerPoint

Table 3. Univariate analysis of prognostic factors for CSS
VariablesCategoryUnivariate
Exponent95% CI P
Age (continuously coded) 1.0070.963–1.0530.759
SexMaleRef.  
Female0.7720.344–1.7320.531
ECOG PS0Ref.  
1, 2, 33.0981.523–6.3010.001
Hemoglobin≥LLNRef.  
<LLN3.7451.309–10.7120.014
Corrected serum calcium≤ULNRef.  
>ULN2.4520.987–6.0960.053
Lactate dehydrogenase≤1.5 ULNRef.  
>1.5 ULN6.7382.735–16.604<0.001
Treatment form diagnosis to treatment≥1 yearRef.  
<1 year2.3441.010–5.4360.047
Prior nephrectomyYesRef.  
No2.4361.176–5.0450.017
Number of metastatic sitesSingleRef.  
Multiple1.3150.938–1.8430.113
TNM stageI, II, IIIRef.  
IV2.4521.005–5.9850.049
Serum sodium concentration (mEq/L)≥138Ref.  
135–1377.3902.877–18.981<0.001
<1352.9151.289–6.5900.010
Neutrophiles≤ULNRef.  
>ULN6.2622.739–14.317<0.001
Platelets≤ULNRef.  
>ULN1.4740.654–3.3220.350
eGFR (mL/min/1.73 m2)≥60Ref.  
<602.0661.018–4.1910.044
Tyrosine kinase inhibitorSunitinibRef.  
Sorafenib1.3240.609–2.8750.479
CRP (mg/dL)≥1.0Ref.  
0.3≤, <1.00.1400.051–0.388<0.001
<0.30.1130.026–0.4960.004
image

Figure 3. Probability estimates of CSS rate in two groups based on serum sodium concentration in 87 patients stratified according to modified MSKCC (a) intermediate and (b) poor risk group.

Download figure to PowerPoint

Multivariate analysis showed severe hyponatremia and mild hyponatremia to be significantly associated with CSS (HR 3.374; 95% CI 1.294–8.798, P = 0.013; HR 6.228; 95% CI 2.161–17.947, P = 0.001), respectively (Table 4). The patients with normal natremia had a significantly superior prognosis of CSS compared with those with both severe hyponatremia and mild hyponatremia. In other clinical features, neutrophilia and a high level of CRP were found to be significant prognostic factors to predict inferior CSS (Table 4).

Table 4. Multivariate analysis of prognostic factors for CSS
VariablesCategoryMultivariate
Exponent95% CI P
Serum sodium concentration (mEq/L)≥138Reference  
135–1376.2282.161–17.9470.001
<1353.3741.294–8.7980.013
CRP (mg/dL)≥1.0Ref.  
0.3≤, <1.00.2310.081–0.6630.006
<0.30.1110.023–0.5380.006
Neutrophiles≤ULNRef.  
>ULN3.5971.046–12.3640.042

Predictive values of severe and mild hyponatremia

Table 5 shows the results of Harrell's concordance index (c-index) calculations for Cox regression analysis. The predictive accuracy calculated with the c-index improved with the addition of severe and mild hyponatremia to the multivariate model, significantly increasing the C-statistic from 0.759 to 0.820 (increase of 0.061; 95% CI 0.007–0.116, P = 0.028).

Table 5. Harrell's concordance index for Cox regression models predicting CSS
VariablesExponent95% CI P
Two risk factors0.7590.636–0.882Reference
(CRP [≤1.0 mg/dL] and neutrophiles [>ULN])
Two risk factors0.8200.706–0.9340.028
(CRP [≤1.0 mg/dL] and neutrophiles [>ULN] + severe and mild hyponatremia (<138 mEq/L)

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Conflict of interest
  9. References

We retrospectively examined whether serum sodium concentration was a significant prognostic factor for mRCC treated with molecular targeted therapy. We found that severe and mild hyponatremia (<138 mEq/L) together with neutrophilia and high level of CRP were the significant factors to predict inferior CSS. Furthermore, hyponatremia could improve the predictive ability of the other prognostic factors.

Serum sodium has been neglected as a significant clinical feature, despite being one of the most frequently obtained blood tests in daily clinical practice. Hyponatremia is reported to be associated with poor prognosis of several carcinomas, including hepatocellular carcinoma,11 gastric cancer12 and RCC.14–16 In terms of RCC, hyponatremia is reported to be associated with poor prognosis of localized and metastatic RCC treated with immunotherapy. Vasudev et al. reported that low preoperative sodium concentration (median 139 mEq/L) might be an important factor associated with reduced survival in patients with RCC, including localized RCC.14 In mRCC patients treated with immunotherapy, Jeppessen et al. reported that the hyponatremia group (<136 mmol) had inferior overall survival time (median 5.5 months) compared with the normal natremia group (median 18.6 months).15 The present study is the first, to our knowledge, to show by univariate and multivariate analysis that severe and mild hyponatremia (<138 mEq/L) is a poor prognostic factor of mRCC treated with molecular targeted therapy as first-line therapy. In previous reports, there was no commonly accepted definition of hyponatremia, and various cut-off points for serum sodium concentration were used. We used Kumar and Berl's conservative definition of hyponatremia (serum sodium ≤137 mEq/L)21 in the present study.

One of the mechanisms inducing hyponatremia depends on the level of chronic inflammation. In experimental and clinical studies, chronic inflammation might lead to production of IL-6 at inflammatory lesions, which might induce neutrophilia and stimulate excessive release of antidiuretic hormone resulting in hyponatremia and hypocalcemia.23–26 In the patients with malignant tumors, Mastorakos et al. reported that the administration of IL-6 resulted in a fast dependent rise in vasopressin concentration,27 and it was suggested that IL-6 might contribute to hyponatremia. In mRCC, several systematic inflammatory responses, such as neutrophilia, thrombocytosis7,28 and a high level of CRP,28–30 were reported to be significant prognostic factors, and chronic inflammation might be an important reaction reflecting tumor aggressiveness. Regarding the relationship between serum IL-6 and mRCC, anemia,31 CRP31 and bone metastasis32 were reported to be closely correlated with IL-6 concentration. Furthermore, IL-6 was reported to be frequently upregulated and implicated in the ability of cancer cells to metastasize to bone in numerous cancer types including RCC.33 In the present study, severe and mild hyponatremia was significantly associated with anemia, high level of LDH, neutrophilia, thrombocytosis and the presence of bone metastasis in univariate analysis. Although we did not examine serum IL-6, hyponatremia in mRCC might be closely correlated with chronic inflammation.

Other explanations for low sodium might be renal impairment,34 poor adrenal gland function35 and use of diuretics.36 In the present study, there was no statistical correlation between the status of natremia and nephrectomy, eGFR or adrenal metastasis, and as in previous reports, renal impairment was not found to be a significant prognostic factor in multivariate analysis.37 Further investigation is required to determine the definition of hyponatremia that would be a prognostic factor of mRCC, and further studies are requires to rule out the mechanisms behind low serum sodium.

In the present study, high CRP level (≥1.0 mg/dL) and neutrophilia were the significant prognostic factors predicting inferior CSS times. Previously, several reports showed that CRP is a significant prognostic factor for metastasis and mortality in RCC, including Japanese patients.28–30 Furthermore, CRP was reported to be a predictive marker in immunotherapy for RCC.38 Neutrophilia is also well known as an important prognostic factor of mRCC treated with molecular targeted therapy, and it is used as one factor of a prognostic model in the age of targeted therapy.7 In the present study, the five patients with neutrophilia, a high level of CRP and hyponatremia had the worst CSS times, and they all died within 1.4 months (median 0.87 months; 95% CI 0.37–1.37; data not shown). Neutrophilia and a high level of CRP, as well as hyponatremia, are important markers reflecting tumor aggressiveness in the era of molecular targeted therapy for mRCC, and further confirmatory studies are warranted to create a new prognostic model that includes CRP level and the status of natremia.

The present study had limitations that should be acknowledged. In Japan, patients could not be treated by mammalian target of rapamycin inhibitors until January 2010, and 18 patients with modified MSKCC poor risk were included in the present study. Furthermore, the number of patients in the present study was small and the study was retrospective. Further investigations will be necessary to determine the definition of hyponatremia and to confirm the present results.

In conclusion, we found that low serum sodium concentration (<138 mEq/L) was a significant factor to predict inferior CSS together with poor neutrophilia and high level of CRP. Furthermore, hyponatremia might be significantly associated with neutrophilia and the presence of bone metastasis reflecting chronic inflammation including serum high level of IL-6 and tumor aggressiveness. Although the mechanisms behind hyponatremia require further research, hyponatremia can be easily and readily determined and might be an important prognostic factor.

Acknowledgment

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Conflict of interest
  9. References

The authors thank the investigators, their staff and the affiliated institutions for their important contributions to this study in The Osaka Renal Cell Carcinoma Clinical Study Collaboration: Dr Susumu Miyoshi (Osaka Rosai Hospital); Dr Seiji Yamaguchi (Osaka General Medical Center Hospital); Dr Yutaka Yasunaga, Dr Toshitsugu Oka (National Hospital Organization Osaka National Hospital); Dr Masatoshi Mukai, Dr Norio Meguro (Toyonaka Municipal Hospital); Dr Hironori Nomura, Dr Kiyomi Matsumiya (Osaka Police Hospital); Dr Satoko Fukuda, Dr Tsuneo Hara (Ikeda Municipal Hospital); Dr Kenji Nishimura (Hyogo Prefectural Nishinomiya Hospital); Dr Toshiaki Yoshioka (Sumitomo Hospital); Dr Nobukazu Murosaki, Dr Masato Honda (Kinki Central Hospital of the Mutual Aid Association of Public School Teachers); Dr Daizo Oka, Dr Nobumasa Fujimoto (Osaka Koseinenkin Hospital); Dr Minoru Koga, Dr Hideki Sugao (Minoh Municipal Hospital); Dr Yasuyuki Kojima (Inoue Hospital); Dr Miyaji Kyakuno (Osaka Seamen's Insurance Hospital); and Dr Takahiro Yoshida, Dr Mototaka Sato, Dr Koji Hatano, Dr Hiroshi Kiuchi (Osaka University Graduate School Of Medicine).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Conflict of interest
  9. References