Albumin-globulin Ratio for Prediction of Long-term Mortality in Lung Adenocarcinoma Patients

Non-small cell lung cancer (NSCLC), which includes a broad range of lung tumors encompassing squamous cell carcinomas and adenocarcinomas as the major subtypes, is the most common cause of cancer related death (Jemal et al., 2010). In patients with lung adenocarcinoma, only a few parameters are described as a prognostic factors such as stage of disease, tumor metabolism, high GLUT1 expression and performance status (de Geus-Qei et al., 2007; Andersen et al., 2011). Serum total proteins contain albumin, globulins and other inflammatory proteins such as C-reactive protein (CRP), interleukins and tumor necrosis factors (McPherson et al., 2006). Albumin and globulins play a pivotal role in the inflammatory process (Azab et al., 2013). Serum albumin is also an objective parameter that reflects the degree of long-term nutrition status (Laky et al., 2007). The globulins play an important role in immunity and chronic inflammation and reflects a cumulative exposure of different cytokines (Azab et al.,


Introduction
Non-small cell lung cancer (NSCLC), which includes a broad range of lung tumors encompassing squamous cell carcinomas and adenocarcinomas as the major subtypes, is the most common cause of cancer related death (Jemal et al., 2010).In patients with lung adenocarcinoma, only a few parameters are described as a prognostic factors such as stage of disease, tumor metabolism, high GLUT1 expression and performance status (de Geus-Qei et al., records; (4) patient had complete follow-up data.(5)Stage 3 or 4 patients at the diagnosis.Exclusion criterias were as follows: (1) patient underwent surgery; (2) patients who had preexisting liver disease; (3) patients with chronic inflammatory disease including autoimmune disorder and infection.A total of 240 patients matched the inclusion and exclusion criteria.Patients information, disease stage and laboratory datas prior to therapy were retrieved from the patient records.The AGR was calculated using the equation AGR=albumin/(total protein-albumin).AGR values of patients are ranked from lowest to highest and the total number of patients were devided into three equal tertiles according to the AGR values (1 th tertile was contain 80 patients with the lowest AGR values and 3 rd tertile was contain 80 patients with the highest AGR values).Furthermore, patients were devided two groups by value of AGR 1.01 according to the ROC analysis (see statistic section).Group 1 consisted of 107 patients and Group 2 consisted of 133 patients.Albumin values of patients also ranked from lowest to highest and the total number of patients were divedid in to three equal tertiles according to the albumin levels like AGR tertiles.Survival analysis was used to evaluate the prognostic value of AGR, AGR tertiles, albumin and low-high groups of AGR.The primary endpoint was all-cause mortality.The patients were followed through January 2014 for the primary endpoint, which was obtained from the hospital cancer registry.The study was approved by the institutional review board.
The tumor size (T), lymph node status (N), presence of metastasis (M) and the American Joint Committee on Cancer (AJCC) stage for each patient were obtained from patients data in our cancer registry (Grene et al., 2002).

Statistic
Statistical analysis was performed using the SPSS 19.0 software package.The continuous variables (according to the normality of the distribution) were presented as means±standard deviations or median with minimummaximum levels.Categorical variables were presented as percentages.AGR tertiles were analyzed using either the chi-square test for categorical variables or the One-way Anova test for continuous variables.Receiver operating characteristics (ROC) analysis was performed.The best cutoff value was determined and the sensitivity and specificity at that point were determined (Figure 1).Survival rates were analyzed using the Kaplan-Meier method and were compared using the logrank test.A Univariate analyse was performed for each variable that may potential predictors of mortality using the Cox proportion-hazards model.Those variables, which found tobe statistically significant (p<0.05) were included in a multivariate Cox regression model.Statistical significance was defined as p<0.05.

Comparision of AGR tertiles according to the baseline characteristics
Level of ALT was significantly lower in 1 st AGR tertile than in 2 nd AGR tertile (p=0.021).BMI was significantly lower in 1 st AGR tertile than in 3 rd AGR tertile (p=0.005).Serum albumin level and AGR were significantly lower, serum globulin level was significantly higher in 1 st AGR tertile than 2 nd and 3 rd tertiles (p<0.001 for all, Table 1).
Table 2 presents the association between baseline characteristics and overall survival in patients with lung adenocarcinoma (univariate Cox analysis).In our study, high levels of aspartate transaminase and white blood cell were associated with higher risk of mortality (p=0.002 and p=0.019 respectively).In addition, low levels of albumin and AGR were associated with higher risk of mortality (p=0.001 and p=0.021 respectively).Also low level of hemoglobin was associated with higher risk of mortality  but not statistically significant (p=0.051).Lower mortality was noted in patients with 3 rd tertile of AGR compare with 1 st tertile of AGR.The mortality rate of 2 nd and 3 rd tertile of AGR was not significantly different.The hazard ratio of the patients in the 1 st AGR tertile increased by 1.680 (p=0.026) in comparison with those in the 3 rd AGR tertile (Table 2).

According to the ROC curve analysis
The predictive value of AGR for mortality from lung adenocarcinoma (sensitivity of 63% and specificity of 50%, area under the ROC curve=0.570) was 1.01 (Figure .2).We devided all patients by value of AGR 1.01.Group 1 consisted of 107 patients and Group 2 consisted of 133 patients.The mean survival time for low AGR group was 12.1 (95% CI:8.551-11.449) and for high AGR group was 17.4 months (95% CI:11.307-166939)(p=0.001).Kaplan-Meier curves showed significantly higher survavial rates in the Group 2 (higer AGR group) than in the Group 1 (lower AGR group) (Figure 1).

Multivariate analysis
Table 3 presents the multivariate analysis by Cox proportional hazard models.We included the variables that associated with survival status in univariate Cox analyses.High levels of white blood cell and low levels of albumin and AGR remained an independent predictors of mortality (Table3).Furthermore low tertile of AGR (become 1 st AGR tertile) and bad performance status were an independent predictors of mortality (Table 3).

Discussion
The main finding of our study was that low levels of pretreatment albumin and AGR are an independent and significant predictor of long-term mortality in patients with lung adenocarcinoma.Previous studies showed that a low AGR was predictive for poor survival in breast cancer (Azab et al., 2013), colorectal cancer (Azab et al., 2013) and nasopharyngeal carcinoma (Du et al., 2014 ), To the best of our knowledge, this is the first study to specifically focus on the predictive value of the AGR in lung adenocarcinoma.
Serum albumin is produced by the liver and maintains osmotic pressure, buffers blood pH and plays a role in the transport of hormones, fatty acids and other compounds.Serum albumin has antioxidant function, acts as a free radical scavenger, play a role in hemostasis of calcium and steroid hormones, and has inhibitor effect on cancer cell lines (Laursen et al., 1990;Sonnenschein et al., 1996).Serum albumin is used to assess patient nutritional status, disease progression, severity and prognosis in hospital (Laursen et al., 1990;Sonnenschein et al., 1996).
Hypoalbuminemia is frequent in hospitalized patients and malnutrition, chronic inflammation, nephrotic syndrome and liver disease suppress albumin synthesis (Yeun et al., 1998).Some studies showed that chronic   inflammation may play a more important role than malnutrition (Gabay et al., 1999;Cueto et al., 2001;Kaysen et al., 2002).Hypoalbuminemia has a strong predictive value on mortality and morbidity regardless of its cause (Gatta et al., 2012).Up to now, hypoalbuminemia has been reported as an independent negative prognostic factor for survival in several cancers including lung (Gupta et al., 2010), gastric (Onate-Ocana et al., 2007), colorectal (Azab et al., 2013) and breast (Gupta et al., 2010).Also low serum albumin level was shown to be an important prognostic factor for tumor recurrence and correlate with poor survival in NSCLC patients (Jin et al., 2013;Yildirim et al., 2013).Similar to these results, in our study, we observed that patients with decreased albumins had poor survival compared to those with high albumin in patients with lung adenocarcinoma.In our study, while patients with BMI < 18 kg/m 2 were considered as malnutrition, only 8 patients in our group had BMI <18 kg/m 2 .Thus, we propose that the prognostic value of low albumin and AGR could solely be explained by chronic inflammation, and malnutrition does not play an important role in our study patients.
Chronic inflammation markers (a part of the calculated globulins) play an important role in the proliferation, progression, development and metastasis of tumor cells (Mantovani et al., 2008;Gatta et al., 2012;Guo et al., 2013;Unal et al., 2013).They also significantly associated with angiogenesis, risk of reccurence and survival in patients with cancer.(Gabay et al., 1999;Mantovani et al., 2008;Yildirim et al., 2013).In several studies, high inflammatory parameters (CRP, alpha and gamma globulins, neutrophil lymphocyte ratio (NLR), complement C3-C4) were found and associated with reduced overall survival in patients with different cancers including lung cancers (Shimada et al., 2003;Canna et al., 2005;Kaya et al., 2013;Zheng et al., 2013;Dai et al., 2014;Soliman et al., 2014).Increasing synthesis of acute phase proteins and interleukins and tumor necrosis-related hypoxia may be the mechanisms underlying this increased inflammation (McMillan et al., 2009).In our study, level of white blood cell was an independent predictor of mortality in patients with lung adenocarcinoma (Table 3).So, we showed that the association between chronic inflamation and mortality, similar to the studies so far.
Previous studies reported that pretreatment AG) is an independent and significant predictor of long-term mortality in breast, colorectal and nasopharyngeal cancer patients (Azab et al., 2013, Du et al., 2014).Using this ratio is very plausible because it contain both important parameters that predicting long term mortality in cancer patients: albumin and globulin (which contain chronic inflammation parameters).We propose that AGR can be used as prognostic marker in patients with lung adenocarcinoma.Additionally, we suggest that AGR is superior to serum albumin in predicting mortality in cancer population for a couple of reasons.First, the low AGR is a combination of two important predictors of adverse outcome, low albumin and high globulin.So, AGR might predict mortality better than the other two parameters (albumin and nonalbumin proteins).Second, some conditions such as dehydration and fluid retention may affect the serum consantration of albumin and globulin but AGR would be unaffected by these conditions.In our study, low levels of AGR was a predictor of poor outcome in patients with lung adenocarcinoma patients.
The major limitations of this study were the retrospective single-center design and lack of measurement of specific inflammatory markers such as CRP and cytokine levels.
In conclusion, pretreatment low AGR is an independent predictor of poor survival in patients with lung adenocarcinoma.Calculation of AGR is easily accessible and cheap to use laboratory method for predicting mortality in patients with lung adenocarcinoma.AGR may be more useful than albumin because of AGR would be unaffected by some conditions that affect albumin and nonalbumin proteins and AGR is a superior prognostic factor in its incorporation of both variables.Further studies are needed to include AGR and other prognostic factors in risk stratification models.DOI:http://dx.doi.org/10.7314/APJCP.2014.15.15.6449 Albumin-globulin

FigureFigure 2 .
Figure 1.The Receiver Operating Characteristic (ROC) Curve for Albumin-globulin Ratio for Predicting Lung Adenocarcinoma Mortality and Kaplan-Meier Survival Curves Stratified by AGR Groups which was Determined by ROC Curve Analysis (p=0.001) Ratio Prediction of Long-term Mortality in Lung Adenocarcinoma Patients 0

Table 1 . Baseline Characteristics of the Patients by AGR Tertiles.
COPD: Chronic obstructive pulmonary disease, AGR: albumin-globulin ratio, BMI: body mass index, WBC: white blood cell.p 1 : difference between 1 st tertile vs 2 nd tertile, p 2 = difference between 1 st tertile vs 3 th tertile.Albumin-globulin Ratio Prediction of Long-term Mortality in Lung Adenocarcinoma Patients *

Table 2 . Hazard Ratios of Baseline Characteristics for All-cause Mortality in Patients with Lung Adenocarcinoma (Univariate Analysis)
*AGR: Albumin-globulin ratio

Table 3 . Cox Proportional Multivariate Hazard Models for All-cause Mortality in Patients with Lung Adenocarcinoma
*AGR: Albumin-globulin ratio