C-reactive protein is a strong predictor for anaemia in renal cell carcinoma: role of IL-6 in overall survival

Authors


  • Both C.E.F. and M.T. contributed equally to this work

Claudia Falkensammer, Department of Urology, Innsbruck Medical University, Anichstrasse 35, 6020 Innsbruck, Austria – Europe. e-mail: claudia.falkensammer@i-med.ac.at

Abstract

Study Type – Prognostic (case series)

Level of Evidence 4

What’s known on the subject? and What does the study add?

The incidence of renal cell carcinoma (RCC)-associated anaemia has been linked with more aggressive tumours. Metastatic RCC patients with elevated IL-6 levels have significantly decreased haemoglobin concentrations.

We elucidate the association of progression of advanced RCC with anaemia and different factors influencing tumour-associated anaemia. Interleukin-6 emerges as a possible target to treat cancer-related anaemia in metastatic RCC.

OBJECTIVE

• To elucidate the association of progression of advanced renal cell carcinoma with anaemia and investigate factors influencing tumor-associated anaemia.

PATIENTS AND METHODS

• We analyzed different clinical variables to study associations with anaemia in 86 metastatic renal cell carcinoma patients.

• 45 (52%) of patients had already developed anaemia prior to therapy.

RESULTS

• Anaemic patients had an increase in the serum markers C-reactive protein (CRP), IL-6 and erythropoietin (EPO). In addition we observed substantial correlation between IL-6 and CRP serum levels (R = 0.639, P < 0.0001).

• Univariate logistic regression analysis revealed that patients with IL-6 >10 pg/mL had a considerable increase in risk for anaemia (odds ratio 3.86, P= 0.003).

• In addition, patients with CRP >0.7 mg/dL had a very strong increase in risk for anaemia (OR = 14.08, P < 0.0001).

• Stepwise multivariate logistic regression analysis confirmed CRP >0.7 mg/mL as the only independent predictor for anaemia. Cox-regression modeling selected serum IL-6 as the strongest independent prognostic indicator (hazard ratio 3.58, P < 0.0001).

CONCLUSION

• Anaemia depends on serum IL-6, which is a strong inductor of CRP and regulator of the iron-transport. Serum IL-6 may be considered as a target to treat cancer-related anaemia.

INTRODUCTION

Renal cell carcinoma (RCC) is the most common cancer of the kidney and accounts for 2% of all malignancies [1]. In advanced RCC 29% to 88% of the patients are anaemic [2–5]. The incidence of RCC-associated anaemia has been linked with more aggressive tumors [6]. The Union Internationale Contre le Cancer (UICC) and American Joint Committee on Cancer (AJCC) designated anaemia as an established prognostic marker in the consensus report [7]. Nevertheless, few studies have focused on the mechanisms of cancer-induced anaemia in general and these mechanisms have not yet been investigated in RCC.

Anaemia of chronic disease is immune driven [8,9] and is present in several malignancies [10]. Recent work indicates that hypoferremia caused by inflammation depends on IL-6, which is required for the induction of hepcidin [11]. It has been shown that IL-6 is a strong inductor of hepcidin in liver cells [12]. Hepcidin is considered as the principal iron-regulatory hormone and the key mediator of anaemia associated with inflammatory processes [13,14].

Metastatic RCC (mRCC) patients with elevated IL-6 levels have significantly decreased haemoglobin concentrations and administration of IL-6 in patients with refractory cancer induced significant anaemia in most patients with a simultaneous increase in serum neopterin and CRP [15]. Discontinuation of IL-6 administration leads to the resolution of anaemia [16,17].

Furthermore the presence of systemic symptoms as well as increased serum CRP levels highly correlated to serum IL-6 and the serum concentration of CRP during IL-6 administration [18,19]. Considering these associations we retrospectively analyzed different acute phase serum parameters, which may predict anaemia and impact on overall survival.

PATIENTS AND METHODS

PATIENTS

Data obtained as part of routine clinical care were retrospectively analyzed. Patients (n= 86) with mRCC who received dendritic cell-based therapies [20,21] were selected at the Department of Urology at the Medical University of Innsbruck. Patients were selected on the availability of datasets in the patients electronic health record (PowerChart®, Cerner, Kansas City, MO). The study was approved by the local ethical committee (REF. NR. UN3550).

Of the 86 studied subjects 26 patients were female and 69 male with a median age of 59.6 years. The clinical characteristics of the patients with mRCC are reported in Table 1. The diagnosis of RCC was histologically confirmed in all patients following surgery of the primary tumor.

Table 1.  Clinical and demographic characteristics of the patients
CharacteristicAnaemia
No
(N= 41)
Yes
(N= 45)
All Cases
(N= 86)
P
  1. SD, standard deviation. *two tailed Chi-square test, all others Mann Whitney U test.

Median Age (SD)57.5 (8.1)61.5 (10.2)59.6 (9.2)0.318
Sex – no. (%)    
 Female10 (38.5)16 (61.5)260.260*
 male31 (51.7)29 (48.3)60 
Median serum marker (SD)    
 CRP (mg/dL)0.39 (6.53)16.8 (4.47)4.73 (8.34)<0.0001
 IL-6 [pg/mL]5.5 (41.9)125.0(30.4)46.6 (59.7)<0.001
 Erythropoietin12.8 (25.1)35.6 (20.2)22.7 (23.1)0.146
Haemoglobin (g/l) (SD)143.5 (8.7)88.5 (12.1)134.5 (28.3)<0.0001
Lactate dehydrogenase (U/l) (SD)181.0 (58.8)186.5 (59.0)190.0 (58.6)0.742
Pathologic stage – no. (%)    
 unknown4 (44.4)5 (55.6)90.225*
 T17 (43.8)9 (56.3)16 
 T213 (68.4)6 (31.6)19 
 T317 (42.5)23 (57.5)40 
 T40 (0.0)2 (100.0)2 
 unknown20 (51.3)19 (48.7)390.323*
 N015 (51.7)14 (48.3)29 
 N14 (50·0)4 (50.0)8 
 N22 (20.0)8 (80.8)10 
 unknown21 (47.7)23 (52.3)440.449*
 M07 (53.8)6 (46.2)13 
 M113 (44.8)16 (55.2)29 
Nuclear grade (%)    
 unknown3 (42.9)4 (57.1)70.034
 14 (80.0)1 (20.0)5 
 223 (59.0)16 (41.0)39 
 310 (40.0)15 (60.0)25 
 41 (10)9 (90)10 

In each patient the levels of haemoglobin (Hb), CRP, proinflammatory cytokine IL-6, EPO, neopterin and lactate dehydrogenase (LDH) were analyzed. These laboratory parameters were routinely assessed at the Central Medical and Chemical Laboratory of our hospital. Anaemia was defined as Hb levels below 120 g/L in female and below 130g/L in male patients according to WHO cut-off levels. CRP, neopterin, EPO and LDH laboratory cut-off level groups were formed according to the routine analyses of the laboratory. The cut-off level of serum IL-6 was set to 10 pg/mL after a distribution analysis of all IL-6 values.

STATISTICAL ANALYSIS

Univariate and multivariate stepwise logistic regression analysis were performed to calculate the proportion of each variable in predicting anaemia independently with 0.05 for entering and 0.1 for removing a variable in the model. For frequency comparisons Chi-square test and Fisher’s exact test were used. The relationship level of two variables was examined using the Spearman-rho correlation coefficient. A two-sided value of P= 0.05 was considered statistically significant. All calculations were carried out with SPSS software 15.0 (SPSS Inc., Chicago, Illinois).

RESULTS

PATIENTS CHARACTERISTICS

Before treatment start 52% (45 of 86) of patients had already developed anaemia. No significant differences were found regarding age and sex. In addition, CRP and IL-6 levels were significantly higher in anaemic patients with a median CRP of 16.8 mg/dL compared to 0.39 mg/dL (P < 0.0001) and a median IL-6 of 125 pg/mL compared to 5.5 pg/mL (P < 0.001). Interestingly, EPO was increased in anaemic patients with a median of 35.6 mU/mL compared to 12.8 mU/mL in non-anaemic patients, but this did not reach significance (P= 0.146). No significant difference was found between median LDH levels and anaemia (P= 0.742). However, a positive association of anaemia with more aggressive tumors (nuclear grade) was found (P= 0.034) in our patient population (Table 1).

FREQUENCY OF ANAEMIA IN RELATION TO CRP AND IL-6

We observed a positive correlation of IL-6 and CRP serum levels (R = 0.639; P < 0.0001). In addition patients with IL-6 >10 pg/mL had a 2-fold increase in the probability to be affected by anaemia (34.1% and 66.7% respectively, P= 0.011). Patients with CRP >0.7 mg/dL had a 2.8-fold increase in the probability to have an anaemia (31.7% and 86.7% respectively, P < 0.0001) (Table 2).

Table 2.  Frequencies of anaemia in patients in relation to IL-6 and CRP
Anaemia –n (%)IL-6PCRPPTotal
<10 pg/mL>10 pg/mL<0.7 mg/dL>0.7 mg/dL
No27 (65.9)14 (34.1)0.01128 (68.3)13 (31.7)<0.000141
Yes15 (33.4)30 (66.7)  6 (13.3)39 (86.7) 45
Total42 (48.8)44 (51.2) 34 (39.5)52 (60.5) 86

SURVIVAL ANALYSIS

Overall survival was calculated using Kaplan-Meier method. The anaemic study population had a median overall survival of 14.7 months compared to 43.8 months of the non-anaemic group, representing a 3.0-fold decrease in survival (P= 0.001) (Fig. 1). Univariate Cox regression analysis for patients with IL-6 >10 pg/mL revealed an elevated hazard ratio (HR = 3.05, P= 0.0001). All other patients with elevated serum markers had a significant increase in the HR: EPO >20 mU/mL (HR = 2.99, P= 0.001), CRP >0.7 mg/dL (HR = 2.92, P= 0.001), neopterin >10 nmol/L (HR = 2.72, P= 0.002), anaemia (HR = 2.62, P= 0.001), LDH >240 U/L (HR = 2.28, P= 0.005) (Table 3). In the multivariate analysis only IL-6 was retained in the forward conditional selection process (HR = 3.58, P= 0.0001). Likewise, the backward conditional process verified IL-6 >10 pg/mL (HR = 2.6, P= 0.009) as the strongest significant risk factor followed by EPO >20 mU/mL (HR = 2.14, P= 0.024), anaemia (HR = 2, P= 0.035) and age (HR = 0.97, P= 0.08) (Table 3).

Figure 1.

Kaplan-Meier survival plot of overall survival in relation to anaemia.

Table 3.  Cox regression analysis (n= 86)
Factor95% CIHRP
  • *

    Sex adjusted category according to WHO cut-off (female hemoglobin <120 g/L; male hemoglobin <130 g/L).

Univariate   
 Age0.95–0.100.980.306
 Sex0.80–2.491.410.242
 Anaemia*1.04–3.412.620.001
 IL-6 >10 pg/mL1.68–3.053.05<0.0001
 Erythropoietin >20 mU/mL1.58–5.622.990.001
 CRP >0.7 mg/dL1.58–5.832.920.001
 Neopterin >10 nmol/L1.45–5.132.720.002
 LDH >240 U/L1.32–4.672.280.005
Multivariate   
Forward conditional   
 AgeLost in selection process
 Sex
 Anaemia
 IL-6 >10 pg/mL1.82–7.043.58<0.0001
 Erythropoietin >20 mU/mLLost in Selection Process
 CRP >0.7 mg/dL
 Neopterin >10 nmol/L
 LDH >240 U/L
Backward conditional   
 Age0.94–1.000.970.083
 SexLost in selection process
 Anaemia1.05–3.822.000.035
 IL-6 >10 pg/mL1.27–5.352.600.009
 Erythropoietin >20 mU/mL1.10–4.152.140.024
 CRP >0.7 mg/dLLost in selection process
 Neopterin >10 nmol/L
 LDH >240 U/L

LOGISTIC REGRESSION ANALYSIS FOR ANAEMIA

Univariate logistic regression analysis detected a significant increased risk to develop anaemia for: CRP >0.7 mg/dL (OR = 14.08, P < 0.0001), IL-6 >10 pg/mL (OR = 3.86, P= 0.003), neopterin >10 nmol/L (OR = 3.18, P= 0.014) and EPO >20 mU/mL (OR = 3.12, P= 0.042) (Table 4). Interestingly, after performing a stepwise multivariate analysis only CRP >0.7 mg/dL was retained in both stepwise conditional models (forward, OR = 16.0, P= 0.0001 and backward OR = 14.41, P < 0.0001) as a very strong significant predictor. Patients with elevated neopterin levels had also an increased risk for anaemia (OR = 2.86, P= 0.08) (Table 4).

Table 4.  Logistic regression analysis for anaemia
Factor95% CIORP
Univariate   
 Age1.07–0.981.020.362
 Sex0.67–4.371.710.262
 IL-6 >10 pg/mL1.58–9.433.860.003
 Erythropoietin >20 mU/mL1.04–9.263.120.042
 CRP >0.7 mg/dL4.74–41.714.08<0.0001
 Neopterin >10 nmol/L1.26–8.613.180.014
 LDH >240 U/L0.41–4.121.300.66
Multivariate   
Forward conditional   
 AgeLost in Selection process
 Sex
 IL-6 >10 pg/mL
 Erythropoietin >20 mU/mL
 CRP >0.7 mg/dL4.59–55.816.0<0.0001
 Neopterin >10 nmol/LLost in Selection Process
 LDH >240 U/L
Backward conditional   
 AgeLost in Selection process
 Sex
 Anaemia
 IL-6 >10 pg/mL
 Erythropoietin >20 mU/mL
 CRP >0.7 mg/dL4.02–51.614.4<0.0001
 Neopterin >10 nmol/L0.90–9.122.660.075
 LDH >240 U/LLost in selection process

DISCUSSION

Anaemia is often observed in patients with newly diagnosed RCC even before therapeutic intervention. The prevalence of anaemia in advanced RCC is described in 29 to 88% of patients [3–5,22]. Anaemia is also known as a negative prognostic risk factor in RCC [23–25]. As in other cancers, anaemia is more prevalent in advanced than in earlier RCC. The association of anaemia with death from RCC has also been extensively evaluated and accepted by the UICC/AJCC as a valid indicator of prognosis, particularly in the setting of metastatic disease.

Magera et al.[6] studied 1707 patients with clinically confined, unilateral, sporadic clear cell RCC treated with radical nephrectomy. In 584 of 1695 patients (35%) they observed anaemia. In addition, patients with anaemia at presentation were more likely to die from RCC compared with patients with normal or elevated Hb.

We confirm these results in our study population of 86 mRCC patients. A total of 53% were anaemic and had a median overall survival of 14.7 months compared to 43.8 months in the non-anaemic group when performing a Kaplan-Meier analysis of overall survival representing a 3-fold decrease in survival (Fig. 1). Patients at stage III and IV showed a significantly lower haemoglobin compared to patients at stage I and II (P= 0.034) (Table 1).

Further, anaemia may contribute to tumor hypoxia, which has been implicated in the development of more aggressive phenotypes of tumors including RCC. Decreased blood oxygen-carrying capacity secondary to anaemia as well as the generation of abnormal blood vessels and areas of necrosis leads to inadequate tumour oxygenation [10,26]. Tumor hypoxia is one of the key factors in inducing the development of cell clones with an aggressive and treatment-resistant phenotype that leads to rapid progression and poor prognosis [27,28].

Several mechanisms whereby malignancy induces anaemia have been suggested including impairment of iron metabolism and elevations in inflammatory cytokine levels, like TNF-α, interleukin-1, interleukin-6, and interleukin-10 also contributing to reductions in renal erythropoietin production [8].

Elevated serum levels of IL-6 have been correlated with a poor outcome in patients with mRCC. Negrier et al. showed that IL-6 is significantly correlated with progression-free survival and overall survival and that it has prognostic value for overall survival in mRCC [29]. IL-6 is a pleiotropic cytokine with potent pro-inflammatory activities and serves as an autocrine growth factor in RCC [30–33]. IL-6 is produced by a wide variety of normal cells upon stimulation and constitutively by various tumor cell lines [33,34] . RCC cell lines and fresh tumor tissue have been shown to express high IL-6 levels [33,34]. IL-6 expression is induced by a variety of stimuli including cytokines such as IL-1, tumor necrosis factor (TNF), and platelet-derived growth factor (PDGF) [35]. The relationship between IL-6 and hepcidin provides a possible explanation for anaemia during inflammatory processes. IL-6 is an inflammatory acute-phase protein and high levels of this cytokine are accompanied by a marked increase in hepcidin synthesis. It also exerts pyrogenic activity and induces the production of acute-phase proteins including CRP and haptoglobin by liver cells in vitro [36]. Hepcidin’s effects on enterocytes and macrophages lead to hypoferraemia and elevated iron sequestration by macrophages, which in turn lead to iron-limited erythropoiesis and anaemia [11,37].

In this retrospective analysis we show for the first time that CRP is a very strong and independent predictor for anaemia in mRCC. However, for the prediction of survival CRP is lost in the multivariate models and only IL-6, anaemia and EPO remain significant independent factors. CRP belongs to the acute phase proteins and binds to phosphocholine on microbes [38]. It interacts with the complement system [39,40] and shows protective effect in Streptococcus pneumoniae infection [41,42]. CRP itself does not affect anaemia, but IL-6 possibly coinduces CRP with hepcidin. Several studies indicate that fifty percent or more of patients with mRCC have increased levels of circulating IL-6, which correlate to increased CRP levels [43–45]. Olencki et al.[15] showed that the major effects of rhIL-6 is on the acute phase response. CRP levels increased significantly during rhIL-6 therapy compared with baseline. Further, significant anaemia compared with baseline developed in most patients but resolved when rhIL-6 was discontinued [15]. This might explain the strong predicitve capacity of CRP for anaemia in our findings (Fig. 2).

Figure 2.

Model for the main effects of IL-6: Anaemia is induced by hepcidin via IL-6. Neopterin, CRP and other factors are co-induced by IL-6 with no consequence for anaemia. The resulting decrease in Hb stimulates EPO production. Finally, this chain of reactions caused by tumor necrosis, negatively affects survival.

Serum IL-6, CRP and Hb are normally found as reliable predictors in survival analyses in RCC [43,46]. The strong correlation between IL-6 and CRP indicates that CRP is induced by IL-6 confirming the dependency of CRP from IL-6 [47,48]. Necrosis present in metastases may be responsible for elevated serum IL-6 levels, since necrosis but not apoptosis has been shown to be accompanied by IL-6 production [49]. IL-6 has pleiotropic effects and is known to induce hepcidin, which leads to hypoferremia and elevated iron sequestration by macrophages [50]. Subsequently, anaemia emerges as a consequence of the iron-deprived environment.

Hypoxia and iron-limiting factors are known to induce EPO production [51]. Thus, anaemic conditions lead to endogenous EPO production, which can be considered a counter measure aimed at the restoration of normal levels of Hb. In our patients the endogenous EPO levels are obviously not sufficient to overcome IL-6 induced anaemia. Therefore, at least in the forward Cox model EPO serum level was lost in the selection process. Likewise, in the logistic regression analysis EPO is lost in the selection process. This indicates that in the causal chain EPO depends on anaemia, which is induced by IL-6 (Fig. 2).

Kamai et al.[52] studied 32 RCC patients and observed a correlation of serum IL-6 level with metastasis, higher tumor grade and higher stage. Further, a higher serum CRP level was associated with poorly differentiated RCC and metastatic disease. The serum hepcidin level was positively correlated with the serum levels of IL-6 and CRP. In Kamai`s study a Kaplan-Meier analysis suggested that higher hepcidin expression by RCC was significantly related to shortened overall survival.

Summarizing, CRP significantly predicted anaemia in the logistic regression analysis in the present study. Concerning survival, IL-6 and CRP were both significant predictors with equivalent HRs in the univariate Cox-regression model. In both multivariate models only IL-6 remained as significant predictor again indicating interdependency of these two factors. Thus anaemia is probably not the only effect of IL-6, which negatively affects survival. Other unknown factors regulated by IL-6 seem to impair patient survival. Taken together, inhibition of IL-6 or hepcidin may be an attractive therapeutic target to treat anaemia in mRCC patients and enhance patients quality of life and survival.

ACKNOWLEDGEMENT

This work was supported by a grant of the kompetenzzentrum medizin tirol (kmt) awarded to MT. This work was also supported by Tilak GmbH, the holding company of Innsbruck Hospital. We also thank G. Bartsch, head of the Department of Urology, for continuous support.

CONFLICT OF INTEREST

None declared.

AUTHORSHIP

CEF, RR and MT were responsible for conception and design, data analysis, interpretation and manuscript writing and revision/editing. RR, MT, CEF and NL were responsible for provision of study materials of patients. RR, CEF and NL were responsible for collection and assembly of data. RR performed statistical analysis.

Ancillary