The preoperative geriatric nutritional risk index (GNRI) is an independent prognostic factor in elderly patients underwent curative resection for colorectal cancer

The world is becoming longer-lived, and the number of elderly colorectal cancer patients is increasing. It is very important to identify simple and inexpensive postoperative predictors in elderly colorectal cancer patients. The geriatric nutritional risk index (GNRI) is a marker of systemic nutrition and is associated with poor survival in various kinds of cancers. A few reports have investigated recurrence factors using preoperative GNRI with CRC (colorectal cancer) patients. This study aimed to investigate whether preoperative GNRI is associated with recurrence-free survival (RFS) and overall survival (OS) in elderly patients with CRC. This study retrospectively enrolled 259 patients with Stage I–III CRC who were more than 65 years old and underwent curative surgery at a single institution in 2012–2017. We classified them into low GNRI (RFS: ≤ 90.5, OS ≤ 101.1) group and high GNRI (RFS: > 90.5, OS > 101.1) group. Multivariable analyses showed low GNRI group was an independent risk factor for 3-year RFS (P = 0.006) and OS (P = 0.001) in the patients with CRC. Kaplan–Meier analysis showed 3-year RFS and 3-year OS were significantly worse in the low GNRI group than in high GNRI group (p = 0.001, 0.0037). A low-preoperative GNRI was significantly associated with a poor prognosis in elderly CRC patients.


Kaplan-Meier curve of GNRI in elderly.
Survival analyses were performed between low GNRI group and high GNRI group according to cutoff value of GNRI. Statistically significant differences between the two groups were revealed by Kaplan-Meier curves on both 3-year RFS (P < 0.0001) and 3-year OS (P < 0.004), indicating a potential prognostic value of GNRI. The 3-year RFS were 62.1% for the low GNRI group, 82.1% for the high GNRI group, respectively ( Fig. 2A). Furthermore, according to the TNM staging stratification analysis, the patients with a low GNRI group were closely associated with poor prognosis stages I + II and III (P = 0.0003, p = 0.046; Fig. 2B,C). The TNM staging analysis was performed by adding Stage I and Stage II due to the small number of Stage I. The 3-year OS were 85.4% for the low GNRI group, 95.3% for the high GNRI group, respectively (Fig. 3A). In OS, the patients with a low GNRI group were closely associated with poor prognosis stages I + stage II and III (P = 0.040, p = 0.017; Fig. 3B,C).
Comparison with other nutritional indicators using ROC curve. ROC analysis was performed using PNI, GPS, and CONUT scores, which are nutritional markers that have been reported to be associated with cancer recurrence, and AUC was calculated. As a result, AUC (area under the curve) had the highest GNRI with GNRI of 0.661, PNI of 0.621, GPS of 0.595, and CONUT of 0.643 (Fig. 4), GNRI was the best predictor of RFS in cases with CRC.

Discussion
Many studies have reported that nutrition-related factors and host immunity have a strong impact on the prognosis of cancer patients 7,8 . The GNRI was firstly reported that simple and accurate tool for predicting the risk of morbidity and mortality in hospitalized elderly patients 9 . The GNRI was strongly associated with mortality in elderly hospitalized patients and in patients with various cancers [10][11][12][13] . In our study, a survival analysis of stage I-III CRC patients who underwent curative surgery revealed that patients with low GNRI had significantly worse 3-year RFS than those with high GNRI. Similarly, in the 3-year OS, the prognosis was poor in the low GNRI group. The GNRI was also an independent risk factor for 3-year RFS and 3-year OS in multivariate analysis. The underlying mechanism by which the low GNRI group results in poor prognosis among colorectal cancer patients undergoing curative surgery is unknown. Two factors can be inferred for the poor prognosis of the low GNRI group. The GNRI is composed of serum albumin levels and body weight (actual body weight [ABW]/ IBW) and represents malnutrition.
First, cancer patients are prone to malnutrition, showing a reduced anabolic response to nutritional support. Anabolic resistance refers to the resistance to assimilation in which protein synthesis in muscle tissue does not occur normally after ingesting nutrients such as amino acids due to surgery, trauma, chronic debilitating diseases,  Table 5. Correlation between colorectal cancer stage and GNRI status.  www.nature.com/scientificreports/ aging, etc. 14 . This also occurred in CRC patients, and it has been reported that a blunted reaction of muscle protein synthesis was observed in CRC patients after injection of the amino acid mixture 15 . Second, albumin synthesis may be suppressed in patients with CRC. There are GPS (Glasgow Prognostic Score) and PNI (prognostic nutritional index) in the score of nutritional evaluation using albumin. The GPS is a score using serum albumin level and CRP (C-reactive protein). It has been reported that when the serum albumin level is low, GPS becomes high and the prognosis was poor in postoperative patients with CRC 16 . The PNI is a score calculated using lymphocyte count and serum albumin level. Tominaga et al. reported a poor prognosis for patients with postoperative CRC with low preoperative PNI 17 . Thus, low serum albumin levels have been reported to have a poor prognosis. Hypoalbuminemia induces an impaired immune response, and immunity had influence on cancer prognosis 18 . Additionally, a low serum albumin levels was associated with elevated inflammatory cytokines such as interleukin-1, and interleukin-6, tumor necrosis factor-alpha, CRP which may lead to the progression of CRC 19 . Therefore, a low GNRI may reflect impaired tumor immunity which may cause cancer progression. The TNM stage has been widely used as the most applicable postoperative staging evaluation system for various cancers worldwide, and it plays an important guiding role in postoperative follow-up and treatment for CRC patients 20,21 . However, it is often reported that there is significant survival heterogeneity among CRC patients with the same TNM stage, and that the TNM stage is inadequate in individual prognosis prediction 22,23 . This may be because the TNM stage only classifies patients according to postoperative pathological results but does not include the patient's own nutritional status. In recent years, we have focused on the tumor environment from the tumor itself, especially the nutritional and inflammatory status of the patient 7,23,24 . By classifying colorectal cancer patients by stage and using GNRI, the ability to discriminate prognosis was improved. Therefore, we believe that GNRI can effectively complement the TNM stage and play an important role in assessing the individual prognosis of CRC patients.

Low GNRI High GNRI P-value Low GNRI High GNRI P-value
This study has some limitations. First, this study was retrospective in design and included patients from a single institution. Overcoming potential biases in observational studies requires controlled randomized controlled trials comparing each GNRI risk group. Second, this study has undergone surgery for a variety of colorectal cancers and does not take into account differences between surgical procedures. Third, there is no consensus regarding the GNRI cut-off value, and this makes it difficult to use the GNRI in clinical settings. We selected the GNRI herein by using a ROC analysis. The GNRI is a non-specific marker of nutrition, which implies that another systemic disease can affect the GNRI. Our study findings need further review and validation in more CRC patients.

Conclusions
Our study provided novel evidence for the clinical relevance and potential feasibility of GNRI as a prognostic biomarker in CRC. Assessment of our developed GNRI could identify patients with elderly CRC who have a poor prognosis.

Patients and methods
Patient selection. Stage I-III CRC diagnosed based on the 8th edition of the United States Joint Commission on Cancer (AJCC) 25   www.nature.com/scientificreports/ Other nutritional markers (Prognostic Nutritional Index: PNI, GPS, CONUT score). PNI is a nutritional index proposed by Onodera et al. 27 . And is calculated using serum albumin and total lymphocyte count. PNI = 10 × Alb + 0.005 × total lymphocyte count. Initially, it was reported as a risk predictor of perioperative complications, later it was reported that evaluation of preoperative PNI was useful as a predictor of prognosis in cancer patients 28 . Glasgow Prognostic Score (GPS) was published by McMillan et al. In 2003. This is a classification using the nutritional index proposed for the first time in non-small cell lung cancer 29 . It was reported to be a better prognostic marker than classification based on stage and performance status. The CONUT score is used as a nutritional evaluation index calculated by scoring albumin level, total lymphocyte count, and total cholesterol level 30 . We have previously reported that CONUT score is useful as a predictor of prognosis after colorectal cancer surgery 25 .
Survival follow-up. Surgical resection was defined as curative when there was no evidence of tumor recurrence and the distant metastases were histologically and macroscopically complete. Patients were followed up every 3 months for the first 3 years, every 6 months for the next 2 years. At each follow-up, all patients underwent physical examination and measurements of serum CEA (carcinoembryonic antigen) and CA19-9 (carbohydrate antigen . They also underwent colonoscopy 1-2 years after surgery (rectal cancer was every year after surgery). Thoraco-abdominal computed tomography scans were usually taken every 6 months. Recurrence was defined as the appearance of a radiological, clinical, and/or pathological diagnosis of cancer cells that were local or distant from their original location.
Determination of cut-off values. The cut-off value for the GNRI was defined according to the receiveroperating characteristic (ROC) curve analysis with Youden's index for the survival, and for BMI 22, for CEA (5 ng/ml) and CA19-9 (37 U/ml) were the upper limit of the normal range in our institute.
Statistical analysis. Differences in categorical variables were examined using a chi-square test or Fisher's exact test. Relapse-free survival (RFS) was calculated from the date of the patient underwent surgery to that of recurrence or death, overall survival (OS) was calculated from the date of the patient underwent surgery to that of death, using the Kaplan-Meier method. Univariate and multivariate analyses were performed using a Cox proportional hazards regression model for RFS and OS. Multivariate analyses were performed using the factors that were significant in univariate analyses. Clinical variables that were considered for univariate and multivariate analyses, in addition to the target GNRI, were previously identified confounding factors with an impact on the prognosis with CRC: sex, age at the diagnosis, histology, pathological T stage (T1/2 or T3/4), lymphnode metastasis (present or absent), BMI (≥ 22 or < 22), CEA levels (< 5.0 vs. ≥ 5.0 ng/mL), CA-19-9 levels (< 37 vs. ≥ 37 U/mL). Probability (p)-values ≤ 0.05 were considered significant. All statistical analyses were performed using JMP 15 software (SAS, Cary, NC, USA).
Human and animal rights. All procedures performed in this study involving human participants were in accordance with the 1964 Helsinki declaration and its latest amendments and comparable ethical standards. These authors do not perform a study with animals.
Ethics approval. This study has been approved by Teikyo University comittee (Registration Number; 19-153).

Consent to participate. A written informed consent was obtained from all individual participants
included in the study.

Data availability
All data generated or analysed during this study are included in this article. Further enquiries can be directed to the corresponding author. www.nature.com/scientificreports/