Introduction

Lung cancer continues to be a serious global health problem, with 900,000 annual new cases in men and 330,000 in women (Ferlay et al. 2001; Parker et al. 1997). Small-cell lung cancer (SCLC) accounts for ∼20% of all bronchogenic carcinomas (Boring et al. 1994). Disease extent at diagnosis is the most prominent prognostic factor of SCLC (Albain et al. 1990; Paesmans et al. 2000). Two-year survival rate of patients with extensive disease (ED) does not exceed 5%, whereas, it is 20–40% in patients with limited disease (LD) (Thatcher et al. 2005). Although, SCLC has the unique behavior such as paraneoplastic syndromes, the neuroendocrine properties are considered to be most likely responsible for it (Williams 1997). In general, tumors of the neuroendocrine origin are capable of producing arginine-vasopressin (AVP). The mRNA for AVP has been detected in SCLC (Oh MS 2002). Generally, the secretion of AVP is down-regulated by the low-serum level (<1.0 pg/ml) in the situation of low plasma osmotic pressure (<280 mOsm/kg) (Zerbe et al. 1980). But in syndrome of inappropriate antidiuretic hormone secretion (SIADH) accompanied with malignancy such as SCLC, AVP is often excessively secreted in spite of low plasma osmotic pressure. The ectopically produced AVP binds to receptors in the kidney, causing retention of free water, which resulted in hyponatremia and low-osmotic pressure (Oh MS 2002). Hyponatremia, occasionally experienced in SCLC patients, has been previously reported to affect the survival of SCLC marginally (Sagman et al. 1991; Kawahara et al. 1997). However, the correlation of SIADH with the prognosis of SCLC is controversial (Tai et al. 2006; List et al. 1986; Harper et al. 1982). On the other hand, there are several cases whose serum-sodium level was normal despite elevation of serum AVP level (Johnson et al. 1997), and a third of patients with SCLC developing hyponatremia had no AVP activity in their plasma (Kamoi et al. 1987; Bliss Jr et al. 1990; Johnson et al. 1997). Then serum level of AVP is recently considered as a tumor marker for SCLC (North 1991). The purpose of the present study is to evaluate the serum level of AVP as a prognostic factor of patients with ED-SCLC.

Materials and methods

Clinical records of patients with histologically or cytologically proven ED-SCLC admitted to the Okayama University Hospital or National Shikoku Cancer Center Hospital between January 1981 and December 2001 were retrospectively investigated in this study.

All patients underwent staging work-ups consisting of a chest radiograph, computed tomographic (CT) scans of the chest and abdomen, bone scintigraphy, and magnetic resonance image scans of the brain. Routine blood chemistry included lactate dehydrogenase (LDH), neuron-specific-enolase (NSE), carcino-embryonic-antigen (CEA), and sodium. Disease extent was determined according to the AJCC Cancer Staging Manual (American Joint Committee on Cancer 2002). LD was defined as tumor confined to one hemithorax, mediastinum, and supraclavicular lymphnodes. Tumor invading beyond these sites was defined as ED. Survival was defined as the time between the date of diagnosis and the date of death. Follow-up time for each patient was at least 3 years.

Fifteen pretreatment variables were chosen for analysis. Each variable was used to divide the patients into two groups in order to determine its prognostic significance. The cut-off values used for hematological and biochemical variables were the normal limits of the variable. In our institutions, the normal serum AVP level was ≤6.3 pg/ml. Survival curves were constructed using the Kaplan–Meier method and the difference of survival was assessed with the log-rank test. Multivariate analysis was performed with the Cox proportional hazard model. P < 0.05 was considered to be statistically significant.

Results

A total of 163 patients were investigated in this study. Age was ranging from 29 to 86 with median of 68. Majority of patients (97.6%) had been previously treated with platinum-based doublet or alternating chemotherapy resulting in the objective response rate of 76%.

The results of univariate analysis are summarized in Table 1. Four factors consisting of serum level of LDH, number of metastatic site, performance status (PS) and body weight loss significantly affected the survival. But serum-sodium level did not affect the survival (P = 0.6653, Fig. 1).

Table 1 Univariate analysis
Full size table
Fig. 1
figure 1

Overall survival of patients with ED-SCLC according to serum-sodium level (≥135 mEq/l, n = 141; solid line versus <135 mEq/l, n = 22; dotted line)

Full size image

Table 2 shows the result of multivariate analysis. The factors having prognostic significance in univariate analysis were analyzed in the multiple regression analysis. The independent relative risk (RR) of death was significantly high in patients with elevated LDH level and those with poor PS (≥2).

Table 2 Multivariate analysis using Cox proportional hazard model
Full size table

The data of serum levels of AVP were available in 34 patients. The correlation of serum AVP level with plasma osmotic pressure is summarized in Fig. 2. The median serum AVP level was 1.35 pg/ml (range 0.0–82.3), and the median serum osmotic pressure was 284.9 mOsm/kg. Six patients demonstrated high serum AVP level (>6.3 pg/ml). The results of univariate analysis of prognostic factors including serum AVP level in these 34 patients are summarized in Table 3. Serum AVP level (P < 0.001) and PS (P = 0.0258) were significantly related to the survival, however, the serum-sodium level did not affect the survival (P = 0.1456). The survival curves according to serum AVP level (Fig. 3) shows that survival time of patients having low serum AVP level (≤6.3 pg/ml) is significantly longer than those having high level (>6.3 pg/ml). The patients with high serum AVP level were listed in Table 4. Four of six patients had low serum osmotic pressure (<280 mOsm/kg). In the all patients with high serum AVP level, serum LDH level was elevated.

Fig. 2
figure 2

The correlation of serum AVP level with plasma osmotic pressure (n = 34)

Full size image
Table 3 Univariate analysis in 34 patients whose serum AVP levels were available
Full size table
Fig. 3
figure 3

Overall survival of patients with ED-SCLC according to serum AVP level (≤6.3 pg/ml, n = 28; solid line versus >6.3 pg/ml, n = 6; dotted line)

Full size image
Table 4 The list of patient with high serum AVP level (>6.3 pg/ml)
Full size table

Discussion

Patients with SCLC currently benefit from favorable response to initial treatment. However, the majority of responding patients will subsequently manifest recurrent disease despite the favorable response rates over 80% (Morstyn et al. 1984). Many investigators have previously attempted to identify the factors predictive for the outcome (Singh et al. 2005; Jorgensen et al. 1996; Spiegelman et al. 1989). Disease extent at diagnosis is the most prominent prognostic factor of SCLC (Albain et al. 1990; Paesmans et al. 2000). Numerous former publications concerning prognostic and predictive factors in large populations pointed that PS and serum LDH level were important parameters for predicting survival or response to chemotherapy (Cerny et al. 1987; Sagman et al. 1991; Albain et al. 1990; Osterlind and Anderson 1986; Yip and Harper 2000). Yip and Harper (2000) demonstrated that PS and disease extent were almost uniformly found to be the most important clinical factors and LDH was the most important laboratory factor. Our study also suggested that PS and serum LDH level were the most important variables affecting the survival.

Water and sodium homeostasis is commonly disrupted in patients with lung cancer. Hyponatremia was observed in ∼15–25% of patients with SCLC at presentation (Sagman et al. 1991; Kawahara et al. 1997; Sorensen et al. 1995). Among pretreatment laboratory variables, hyponatremia was reported to affect the survival marginally (Sagman et al. 1991; Kawahara et al. 1997). In this study, although hyponatremia was observed in 13.5% of patients with ED-SCLC, it did not affect the survival. SIADH, one of the most frequently observed paraneoplastic syndromes (Seute et al. 2004), is usually accompanied with hyponatremia. However, correlation of SIADH due to malignancies with the prognosis is not clear and the results of previous studies are controversial. Harper et al. (1982) showed that the survival of SCLC patients developing SIADH was shorter than the others, especially in patients with LD-SCLC. They hypothesized that SIADH was associated with the tumors acquiring drug resistance quickly. In the Vanderbilt experience, however, no relationship was observed between the presence of SIADH and response to chemotherapy or overall survival (List et al. 1986).

Ectopic production of AVP by SCLC cells is considered to play a causal role in the development of hyponatremia, but a third of SCLC patients with hyponatremia had no evidence of AVP secretion (Kamoi et al. 1987; Bliss Jr et al. 1990; Johnson et al. 1997). In this study, a fifth of the SCLC patients with hyponatremia had almost no AVP secretion, and there were several cases with normal serum-sodium levels despite elevation of serum AVP level (Johnson et al. 1997). SCLC is frequently associated with paraneoplastic syndromes because of its neuroendocrine properties (Mason et al. 1997; Dropcho 1995; Ganz et al. 1987). The neuroendocrine properties are also considered to be responsible for metastatic activity (Seute et al. 2004), which suggests that neuroendocrine properties affect the prognosis of SCLC. In our study, high serum AVP level was the variable significantly affecting survival, which indicates that serum AVP level may reflect the degree of neuroendocrine properties. The correlation of serum LDH level with disease extent has been previously reported (Sagman et al. 1991; Kawahara et al. 1997; Cerny et al. 1987; Byhardt et al. 1986). On the other hand, Ganz et al. (1987) demonstrated a significant correlation between disease activity and serum LDH in the serial monitoring of response to therapy in patients with SCLC. Furthermore, Albain et al. (1990) also demonstrated that a normal LDH was the most important predictor of favorable survival in ED patients. Thus, elevated levels of serum LDH may reflect not only disease extent but also tumor bulk. In our study, the serum AVP level was significantly correlated with the serum LDH level (P = 0.046) (Table 5), which suggests that serum AVP level might correlate with tumor bulk. Three of the six cases with high AVP level were not accompanied with hyponatremia. Accordingly, the serum AVP level may become the better indicator of tumor bulk and prognosis than serum-sodium level.

Table 5 The characteristics of patients according to serum AVP level (>6.3 pg/ml vs. ≤6.3 pg/ml)
Full size table

The present study suggests that the serum level of AVP is more useful for predicting prognosis than serum-sodium level. We can apply this marker for the subclassification of SCLC. However, the present study had several limitations, because all the analyses were performed retrospectively. Well-designed prospective trials are warranted.