Neutrophil to lymphocyte ratio can predict overall survival in patients with stage II to III colorectal cancer

Abstract

The neutrophil-to-lymphocyte ratio (NLR) is a prognostic inflammatory marker in colorectal cancer (CRC), however, little is known for its prognostic role in stage II to III CRC patients underwent curative resection. This study was aimed to investigate prognostic role of NLR in stage II to III CRC patients underwent surgery. Medical records of 1378 newly diagnosed CRC patients between June 2006 and March 2020 were reviewed. Data were collected by using electronic medical record software. Survival rate were analyzed using the Kaplan–Meier method. The cutoff values of NLR in stage II to III CRC patients were defined by maximally selected log-rank statistics. Multivariable cox proportional-hazard models were performed to find risk factors associated with overall survival (OS) in stage II to III CRC patients underwent surgery. Among 1378 CRC patients enrolled, 910 patients underwent surgery. In entire surgical cohort, age, body mass index (BMI), CEA, carbohydrate antigen 19-9 (CA 19-9), lymphatic invasion, NLR, and albumin-to-globulin ratio (AGR) were found to be risk factors associated with OS (all P < .05). In stage II to III CRC patients underwent curative resection (n = 623), age, BMI, lymphatic invasion, AGR, and NLR were found to be risk factors associated with OS (all P < .05). In the multivariable analysis, CA 19-9 and lymphatic invasion were independent risk factors for OS in entire surgical cohort. In the multivariable analysis for the stage II to III CRC patients, age, BMI, lymphatic invasion and NLR (Hazard ratio = 2.41, 95% confidential interval [CI]: 1.04–5.595, P = .041) were independent risk factors for OS. NLR can be used as a clinically simple and useful parameter for predicting OS in stage II to III CRC patients undergoing curative resection, however, its optimal cutoff value should be further evaluated.

1. Introduction

Colorectal cancer (CRC) is the third most diagnosed cancer worldwide.^[1] In Korea, CRC was the third most common cancer for male and female, and the second most common reason of cancer-related death.^[2] By the 2030, global burden of CRC is expected to increase by 60%, with more than 2.2 million new cases and an estimated 1.1 million deaths.^[1] Furthermore, the burden of CRC will increase as the older population grows.^[3] Currently, many efforts to find a simple prognostic factor have been ongoing, because about half of CRC patients experience disease recurrence and metastasis.^[4]

It is well known that an inflammation plays a significant role in the development and progression of malignancies.^[5] Neutrophils are an integral part of innate immune response and suppress the activity of cytotoxic T cells, which may promote tumor progression, and lymphocytes are important in cell-mediated adaptive immune response. Therefore, the balance between neutrophil and lymphocyte may be associated with prognosis of malignancies. Recently, for example, the potential diagnostic and prognostic role of neutrophil-to-lymphocyte ratio (NLR) has been published in CRC patients.^[6,7] presurgical NLR > 4.7 showed a poor prognostic factor for 5-year survival, disease free survival and overall survival (OS) in stage II CRC patients underwent curative resection.^[6] However, previous studies were limited as they had no standard cutoff of NLR, heterogenous patients, and variable strength of association between NLR and OS in CRC patients. Other inflammatory markers, such as lymphocyte-to-C-reactive protein ratio (LCR), albumin-to-globulin ratio (AGR), and C-reactive protein-to-albumin ratio (CAR), are also emerging as new prognostic inflammatory markers. However, only few studies evaluated their role in CRC patients,^[8] and no studies their role in stage II to III CRC patients underwent curative resection.

Therefore, this study aimed to investigate the potential prognostic role of NLR, LCR, AGR, and CAR for OS in stage II to III CRC patients underwent curative resection.

2. Patients and Methods

2.1. Data source and study population

We retrospectively investigated the medical records of patients with newly diagnosed CRC at Kyung Hee University Hospital at Gangdong, Seoul, Korea, between June 2006 and March 2020. Patients satisfying any of the following criteria were excluded from this analysis: age younger than 25 years; history of inflammatory bowel disease, familial adenomatous polyposis, or Lynch syndrome; patients with <1-year follow-up period after diagnosis of CRC. Patients were also excluded from this analysis, if they had inadequate data in the electronic medical records. A total of 1378 CRC patients were analyzed for this study based on the above inclusion and exclusion criteria (Fig. 1). Patients were waived for informed consent due to retrospective study design. This study was approved by the Institutional Review Board of Kyung Hee University Hospital at Gang Dong (KHNMC IRB 2020-10-004).

Figure 1.: A flow diagram for the study participants. For a colorectal cancer (CRC) cohort, 1378 patients with CRC were enrolled after excluding 604 patients. Surgical cohort included 910 patients with CRC who underwent surgery. Among them, 623 patients with stage II to III CRCs underwent curative resection. CRC = colorectal cancer, IBD = inflammatory bowel disease.

Clinical and pathological data were collected by using electronic medical record software from the patient records. The data included were the followings: demographic data; body mass index (BMI); family history of CRC; preoperative laboratory data; surgical treatment; survival; pathological data, such as tumor location, size and differentiation; pathological tumor-node-metastasis staging; number of lymph nodes harvested; lymphatic, vascular or neural invasion. In this study, the American Joint Committee on Cancer tumor-node-metastasis staging was used; 6^th edition^[9] was adopted between 2006 and 2009 and 7^th edition^[10] was adopted between 2010 and 2020. The tumor location was defined as right-sided (cecum, ascending colon, hepatic flexure, and transverse colon) and left-sided (splenic flexure, descending colon, sigmoid colon, and rectum). For the preoperative laboratory data, following laboratory variables were collected before operations: protein, albumin, hemoglobin, white blood cells with differentiation including lymphocyte and neutrophil, platelet, C-reactive protein (CRP), carcinoembryonic antigen (CEA), and carbohydrate antigen 19-9 (CA 19-9). OS was calculated from the date of CRC diagnosis to the date of death from any cause and censored at the last follow-up examination.

2.2. Definition of variables

NLR is estimated by dividing the absolute neutrophil count by the absolute lymphocyte count from a complete blood count with differential count.^[6,7] For the other inflammatory prognostic factors, LCR was calculated by lymphocyte count (numberx10³/μL)/CRP (mg/dL),^[8] AGR was calculated by serum albumin (mg/dL)/serum globulins (mg/dL), and CAR was calculated by CRP (mg/dL)/albumin (mg/dL).

2.3. Statistical analyses

All data are presented as the mean ± standard deviation or the median and interquartile range (the 25th and 75th percentiles), if distributions were normal or skewed, respectively. For time-to event analyses, survival estimates were analyzed using the Kaplan–Meier method. The cutoff values of NLR, LCR, CAR, and AGR in CRC patients were defined by maximally selected log-rank statistics.^[11] The cutoffs of other variables were defined as 65 years for age (old age criteria by World Health Organization), 25 kg/m² for BMI (obesity criteria in Asian adult by World Health Organization^[12]), 5 ng/mL for CEA (upper normal limit of CEA), and 60 U/mL for CA 19-9 (cut off level to predict survival rate in patients with CRC^[13]). To analyze how much each risk factor is related to survival, multivariable cox proportional-hazard models were used using the risk factors that were significant (P < .05) or relatively significant (P < .1) in the univariable analysis. Hazard ratios (HRs) and 95% confidential interval (CI) were measured using cox proportional-hazard analysis. All P values were 2-tailed, and P value < .05 was considered statistically significant. Statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC) and R studio Version 1.2.1335.

3. Results

In total, 1378 patients (767 males and 611 females) with CRC were enrolled after excluding 604 patients due to age younger than 25 years (n = 4), history of inflammatory bowel disease (n = 4) and insufficient patient data (n = 596). The mean age was 63.7 ± 12.0 years and 4.1% of them had a family history of CRC in a first-degree relative. Among them, 910 patients underwent surgery for CRCs.

3.1. Patient characteristics of surgical cohort

Table 1 shows the clinical and pathological characteristics of 910 surgical patients. The mean age was 65.1 ± 11.8 years, 502 patients were male (55.1%), and 4.0% of them had a family history of CRC. During median follow-up of 1639.7 days, 53 patients were dead (5.8%). With respect to tumor locations, left-sided tumors (61.1%) were dominant than right-sided tumors (29.6%). For pathological tumor-node-metastasis stages, patients of stage II to III (68.5%) were dominant, followed by stage I (22.7%), and stage IV (6.6%). Most (86.2%) patients had well or moderately differentiated adenocarcinoma, but 11.7% of patients had poorly differentiated adenocarcinoma, signet ring cell cancer or mucinous cancers. Mean number of regional lymph node metastasis was 2.0, and mean number of lymph node harvest during operation was 25.7.

Table 1 - Clinical and pathological characteristics of patients with colorectal cancer who underwent operations.

Clinical and pathological characteristics	Results
Clinical characteristics
Age at diagnosis (yr), mean ± SD	65.1 ± 11.8
Sex (male), n (%)	502 (55.1)
Positive family history of colorectal cancer, n (%)	36 (4.0)
Body mass index (kg/m2), mean ± SD	23.4 ± 3.6
Duration of follow-up (d), median [IQR]	1639.7 [795.8–1456.5]
Death during follow-up, n (%)	53 (5.8)
Pathological characteristics
Location of tumor, n (%)
Right-sided	269 (29.6)
Left-sided	556 (61.1)
Overlapping	80 (8.8)
Unspecified	5 (0.6)
pTNM stage, n (%)
Stage I	207 (22.7)
Stage II–III	623 (68.5)
Stage IV	60 (6.6)
Stage, unspecified	20 (2.2)
Differentiation,* n (%)
Well/moderate adenocarcinoma	778 (86.2)
Signet ring cell, mucinous, poorly differentiated	106 (11.7)
Others	19 (2.1)
Lymph node status
No. of regional lymph node metastasis, mean ± SD	2.0 ± 4.2
No. of lymph node harvest, mean ± SD	25.7 ± 13.4
Lympho-vascular and neural invasion, n (%)
Lymphatic invasion	261/892 (30.2)
Vascular invasion	50/864 (5.8)
Neural invasion	80/796 (10.1)

IQR = interquartile range, pTNM = pathological tumor-node-metastasis, SD = standard deviation.

^*Seven (0.8%) patients were missed for differentiation of colorectal cancer.

3.2. Risk factors for the survival rate in surgical cohort

Table 2 shows the risk factors for the survival rates of patients with CRC who underwent operations (n = 910). In surgical cohort, BMI, CEA, CA 19-9, and lymphatic invasion were found to be risk factors of 5-year survival rate (all P < .05) (Table 2). In addition, age, BMI, CEA, CA 19-9, lymphatic invasion, NLR, and AGR were found to be risk factors of OS in surgical cohort (all P < .05). NLR ≥ 3.15 showed a poor prognosis of OS in CRC patients who underwent operations (P = .021, Fig. 2A).

Table 2 - Risk factors for the survival rates of patients with colorectal cancer who underwent operations.

Variables	Number (%)	5-yr survival rate		Overall survival rate
		(95% CI)	P	(95% CI)	P
Clinical variables
Age (yr)			.109		.004
25–64	423 (46.5)	0.95 (0.93–0.98)		0.94 (0.92–0.97)
≥65	487 (53.5)	0.92 (0.89–0.95)		0.78 (0.69–0.88)
Sex			.505		1.000
Male	502 (55.2)	0.94 (0.92–0.97)		0.87 (0.82–0.93)
Female	408 (44.8)	0.93 (0.90–0.96)		0.87 (0.80–0.95)
BMI (kg/m2)			<.001		.003
<25	635 (70.0)	0.92 (0.89–0.94)		0.84 (0.79–0.90)
≥ 25	272 (30.0)	0.98 (0.96–1.00)		0.95 (0.89–0.99)
CEA (ng/mL)			.036		.030
<5	594 (69.9)	0.95 (0.93–0.98)		0.89 (0.84–0.95)
≥5	256 (30.1)	0.90 (0.85–0.95)		0.86 (0.79–0.93)
CA19-9 (U/mL)			.042		.004
<60	738 (93.1)	0.95 (0.93–0.97)		0.88 (0.83–0.93)
≥60	55 (6.9)	0.78 (0.64–0.96)		0.78 (0.64–0.96)
Pathological variables
Location of tumor			.752		.800
Right-sided	269 (32.6)	0.94 (0.90–0.97)		0.85 (0.73–0.98)
Left-sided	556 (67.4)	0.94 (0.92–0.97)		0.88 (0.84–0.92)
Lymphatic invasion			.001		<.001
No	603 (69.8)	0.96 (0.94–0.98)		0.89 (0.84–0.94)
Yes	261 (30.2)	0.87 (0.81–0.92)		0.80 (0.71–0.91)
Differentiation			.383		.400
Well/moderately diff.	778 (88.0)	0.94 (0.92–0.96)		0.87 (0.83–0.92)
Poorly diff.	106 (12.0)	0.90 (0.83–0.99)		0.85 (0.74–0.98)
Lymph node harvest (number)			.944		.500
<12	68 (7.6)	0.94 (0.87–1.00)		0.94 (0.87–0.99)
≥12	824 (92.4)	0.94 (0.92–0.96)		0.86 (0.82–0.92)
Inflammatory variables
Neutrophil/lymphocyte ratio			.537		.021
<3.15	579 (64.1)	0.94 (0.92–0.97)		0.90 (0.85–0.95)
≥3.15	324 (35.9)	0.93 (0.90–0.96)		0.83 (0.76–0.92)
Lymphocyte (×103)/CRP ratio			.296		.050
≥9.08	175 (28.7)	0.94 (0.89–0.99)		0.85 (0.79–0.91)
<9.08	435 (71.3)	0.91 (0.87–0.94)		0.94 (0.89–0.99)
Albumin/globulin ratio			.153		.010
≥1.67	72 (7.9)	0.95 (0.89–1.00)		0.88 (0.83–0.92)
<1.67	836 (92.1)	0.94 (0.92–0.96)		0.85 (0.72–1.00)
CRP/albumin ratio			.285		.060
<0.06	211 (34.4)	0.93 (0.89–0.98)		0.93 (0.89–0.98)
≥0.06	402 (65.6)	0.91 (0.87–0.95)		0.85 (0.78–0.91)

BMI = body mass index, CA 19-9 = carbohydrate antigen 19-9, CEA = carcinoembryonic antigen, CI = confidential interval, CRP = C-reactive protein.

Figure 2.: Kaplan–Meier survival curves of patients with colorectal cancer who underwent operations according to neutrophil-to-lymphocyte ratio (NLR). Higher NLR had a worse overall survival than lower NLR in the entire surgical cohort (A) and stage II to III surgical cohort (B).

3.3. Risk factors for the survival rate in stage II to III surgical cohort

As the surgical cohort has included heterogenous CRC patients from stage I to IV, risk factors for the survival rate were analyzed for only stage II to III CRC patients underwent curative resection (n = 623) (Fig. 1). BMI and lymphatic invasion were found to be risk factors of 5-year survival rate for stage II to III CRC patients underwent curative resection (both P < .05, Table 3). In addition, age, BMI, lymphatic invasion, and NLR were found to be risk factors for stage II to III CRC patients underwent curative resection (all P < .05). NLR ≥ 3.15 showed a poor prognosis for OS in patients with stage II to III CRCs underwent curative resection (P = .005, Fig. 2B).

Table 3 - Risk factors for the survival rates of patients with colorectal cancer in stage II to III who underwent curative resection (n = 623).

Variables	Number (%)	5-yr survival rate		Overall survival rate
		(95% CI)	P	(95% CI)	P
Clinical variables
Age (yr)			.056		.005
25–64	273 (43.8)	0.96 (0.93–0.99)		0.95 (0.92–0.98)
≥ 65	350 (56.2)	0.92 (0.88–0.95)		0.76 (0.66–0.89)
Sex			.310		.900
Male	344 (55.2)	0.95 (0.92–0.98)		0.86 (0.79–0.93)
Female	279 (44.8)	0.92 (0.88–0.96)		0.86 (0.77–0.96)
BMI (kg/m2)			<.000		.010
<25	454 (73.0)	0.92 (0.89–0.95)		0.84 (0.77–0.91)
≥25	168 (27.0)	0.99 (0.97–1.00)		0.93 (0.85–1.00)
CEA (ng/mL)			.543		.600
<5	388 (66.3)	0.95 (0.92–0.98)		0.87 (0.80–0.95)
≥5	197 (33.7)	0.93 (0.89–0.98)		0.89 (0.81–0.96)
CA19-9 (U/mL)			.188		.100
<60	509 (92.5)	0.95 (0.93–0.97)		0.87 (0.81–0.94)
≥60	41 (7.5)	0.85 (0.71–1.00)		0.85 (0.71–1.00)
Pathological variables
Location of tumor			.477		.300
Right-sided	199 (31.9)	0.96 (0.93–0.99)		0.86 (0.74–0.99)
Left-sided	424 (68.1)	0.93 (0.90–0.96)		0.86 (0.81–0.92)
Lymphatic invasion			.008		.002
No	402 (64.8)	0.96 (0.94–0.99)		0.88 (0.81–0.94)
Yes	218 (35.2)	0.88 (0.83–0.94)		0.84 (0.76–0.94)
Differentiation			.842		.900
Well/moderately diff.	530 (85.5)	0.94 (0.91–0.96)		0.86 (0.80–0.92)
Poorly diff.	90 (14.5)	0.94 (0.89–1.00)		0.88 (0.77–1.00)
Lymph node harvest (number)			.471		.800
<12	25 (4.0)	0.88 (0.72–1.00)		0.88 (0.72–1.00)
≥12	598 (96.0)	0.94 (0.92–0.96)		0.86 (0.80–0.92)
Inflammatory variables
Neutrophil/lymphocyte ratio			.230		.005
<3.15	381 (61.7)	0.95 (0.92–0.98)		0.90 (0.83–0.97)
≥3.15	237 (38.3)	0.92 (0.88–0.96)		0.79 (0.70–0.90)
Lymphocyte (×103)/CRP ratio			.064		.060
≥9.08	105 (24.5)	0.96 (0.93–1.00)		0.82 (0.75–0.91)
<9.08	318 (75.2)	0.91 (0.87–0.95)		0.96 (0.93–1.00)
Albumin/globulin ratio			.105		.200
≥1.67	46 (7.4)	0.91 (0.75–1.00)		0.91 (0.75–1.00)
<1.67	576 (92.6)	0.93 (0.91–0.96)		0.86 (0.80–0.92)
CRP/albumin ratio			.105		.300
<0.06	129 (30.3)	0.95 (0.91–0.99)		0.95 (0.91–0.99)
≥0.06	297 (69.7)	0.91 (0.87–0.95)		0.83 (0.75–0.92)

BMI = body mass index, CA 19-9 = carbohydrate antigen 19-9, CEA = carcinoembryonic antigen, CI = confidential interval, CRP = C-reactive protein.

3.4. Multivariable analysis for predictors of OS

Using the multivariable analysis, CA 19-9 (HR = 3.11, 95% CI: 1.14–8.50, P = .026) and lymphatic invasion (HR = 3.01, 95% CI: 1.45–6.224, P = .003) were independent risk factors for OS in entire surgical cohort (Table 4). In a surgical cohort with stage II to III CRC, age (HR = 1.05, 95% CI: 1.01–1.09, P = .03), BMI (HR = 0.23, 95% CI: 0.05–0.96, P = .045), lymphatic invasion (HR = 2.50, 95% CI: 1.17–5.34, P = .017) and NLR (HR = 2.41, 95% CI: 1.04–5.595, P = .041) were independent risk factors for OS (Table 5).

Table 4 - Univariable and multivariable analysis for predictors of overall survival in surgical cohort.

	Univariable analysis			Multivariable analysis
Variables	HR	95% CI	P value	HR	95% CI	P value
Age (25–64 vs ≥65)	2.28	1.28–4.08	.005	1.73	0.78–3.82	.172
Sex (female vs male)	1.02	0.58–1.78	.946	-	-	-
BMI (<25 vs ≥25)	0.30	0.12–0.71	.005	0.39	0.14–1.14	.086
Location (right- vs left-sided)	0.94	0.51–1.71	.845	-	-	-
CEA (<5 vs ≥5 ng/mL)	1.87	1.03–3.39	.037	0.91	0.41–2.04	.828
CA19-9 (<60 vs ≥60 U/mL)	3.11	1.38–6.99	.006	3.11	1.14–8.50	.026
Lymphatic invasion (no vs yes)	3.25	1.87–5.66	<.001	3.01	1.45–6.22	.003
Poorly diff. (no vs yes)	1.38	0.62–3.07	.428	-	-	-
NLR (<3.15 vs ≥3.15)	1.88	1.09–3.27	.023	1.59	0.75–3.35	.217
LCR (<9.08 vs ≥9.08)	2.33	0.97–5.57	.057	1.18	0.46–3.05	.728
AGR (<1.67 vs ≥1.67)	0.27	0.08–0.92	.037	0.42	0.07–2.27	.314
CAR (<0.06 vs ≥0.06)	1.75	0.83–3.70	.142	-	-	-

AGR = albumin-to-globulin ratio, BMI = body mass index, CAR = CRP-to-albumin ratio, CEA = carcinoembryonic antigen; CA19-9, carbohydrate antigen 19-9, CI = confidence interval, LCR = lymphocyte-to-CRP ratio, NLR = neutrophil-to-lymphocyte ratio, OR = odd ratio.

Table 5 - Univariable and multivariable analysis for predictors of overall survival in stage II to III CRC patients who underwent curative resections.

	Univariable analysis			Multivariable analysis
Variables	HR	95% CI	P value	HR	95% CI	P value
Age (25–64 vs ≥65)	1.06	1.03–1.09	<.001	1.05	1.01–1.09	.003
Sex (female vs male)	1.05	0.56–1.99	.875	-	-	-
BMI (<25 vs ≥25)	0.30	0.11–0.83	.021	0.23	0.05–0.96	.045
Location (right- vs left-sided)	1.19	0.87–1.62	.283	-	-	-
CEA (<5 vs ≥5 ng/mL)	1.20	0.59–2.46	.614	-	-	-
CA19-9 (<60 vs ≥60 U/mL)	2.26	0.79–6.46	.130	-	-	-
Lymphatic invasion (no vs yes)	2.61	1.37–4.98	.003	2.50	1.17–5.34	.017
Poorly diff. (no vs yes)	1.07	0.41–2.76	.883	-	-	-
NLR (<3.15 vs ≥3.15)	2.45	1.28–4.67	.006	2.41	1.04–5.59	.041
LCR (<9.08 vs ≥9.08)	3.00	0.91–9.94	.072	1.00	0.99–1.00	.999
AGR (<1.67 vs ≥1.67)	0.22	0.05–0.92	.037	0.98	0.40–2.42	.966
CAR (<0.06 vs ≥0.06)	1.65	0.67–4.08	.275	-	-	-

AGR = albumin-to-globulin ratio, BMI = body mass index, CAR = CRP-to-albumin ratio, CEA = carcinoembryonic antigen; CA19-9, carbohydrate antigen 19-9, CI = confidence interval, LCR = lymphocyte-to-CRP ratio, NLR = neutrophil-to-lymphocyte ratio, OR = odd ratio.

4. Discussion

In this study, we evaluated the risk factors of 5-year survival and OS for the prognosis of CRC, which are the main therapeutic oncologic outcomes, for entire and stage II to III surgical cohort. The notable finding of this study is that NLR is a predictor of OS in stage II to III CRC patients underwent curative resection. In our study, preoperative NLR ≥ 3.15 is associated with poorer OS in patients with stage II to III CRC patients (HR = 2.41, 95% CI: 1.04–5.595, P = .041). Although many studies have been published for the prognostic role of NLR in CRC,^[14–17] it is desirable to compare their prognostic role in homogenous CRC patients who underwent similar treatment. From this perspective, the prognostic role of NLR was validated in more homogenous CRC patients: stage I CRC,^[18] stage IIA CRC,^[19] and stage II CRC.^[6] Our study was unique it that prognostic role of NLR was evaluated in stage II to III CRC patients who underwent curative resection. In our study, however, CAR, LCR, and AGR has no prognostic role in the stage II to III CRC patients who underwent curative resection. CRC has been the leading cause of death in patients with CRC, but, noncancer causes of death account for an increasing proportion over time in patients with CRC.^[20,21] After 10 years following nonmetastatic CRC diagnosis, non-cancer deaths accounted for 71.9% of all deaths in 475,771 patients with CRC.^[20] In addition, only 10.4% of deaths were attributed to CRC, 15.3% were attributed to other cancers and 34.2% were secondary to heart disease in 302,345 patients with CRC.^[21] In this regard, prognostic factor to predict OS after surgical resection for CRC, such as NLR in this study, is warranted.

In a systemic review from 19 studies with 10,259 patients,^[14] NLR > 5.0 is associated with poorer long-term survival in patients with CRC. But this review was limited, as significantly heterogenous surgical and medical managements were included in enrolled studies. In addition, they included all studies with variable NLR cutoffs (2.4~5.0) and sample sizes (115~5336). Another systemic review including 16 studies showed that high NLR predicted poorer OS (HR: 1.813, 95% CI: 1.499–2.193) in all CRC patients.^[15] A recent meta-analysis with 16 studies including 5897 patients also showed that high NLR was associated with poor OS (HR: 1.66, 95%CI: 1.36–2.02, P < .001).^[16] Largest meta-analysis from 71 studies including 32,788 patients also showed that high NLR was associated with poor outcomes in CRC patients.^[17] Correcting for publication bias, however, true effect size to HR was only 1.57 (95% CI 1.39–1.78; P < .0001) for OS in CRC patients. In our study, however, we failed to show the prognostic role of NLR in entire surgical cohort. It may make sense as our entire surgical cohort included patients with variable medical and surgical treatments for stage I to IV CRCs.

In multivariable analysis of our study, CA 19-9 and lymphatic invasion were independent risk factors for OS in entire surgical cohort, which was consistent with our previous study.^[13] In addition, age, BMI, and lymphatic invasion were independent risk factors for OS in stage II to III CRC patients who underwent curative resection, which were also consistent with previous studies.^[22–24] Among prognostic factors, the association of BMI and CRC survival was more complex than other risk factors. In 6 prospective studies participating in the Genetics and Epidemiology of Colorectal Cancer Consortium, BMI 25.0 to 29.9 kg/m² was associated with decreased mortality for stage II to III CRC patients, compared with normal BMI.^[23] BMI ≥ 30 kg/m² was associated with increased mortality for stages II CRC, but decreased mortality for stages III CRC.^[23] In our study, we classified BMI as only 2 groups with 25 kg/m² cutoff, and BMI 25.0 to 29.9 kg/m² or ≥30 kg/m² were not separated in our study. The relationship between BMI and mortality in stage II to III CRC patients may warrant further studies.

As another potential prognostic inflammatory markers, CAR, LCR, and AGR was also evaluated in this study. Japanese studies showed that CAR was associated with OS in 627 CRC patients who underwent surgery,^[25] and LCR was an independent prognostic factor for OS in 477 patients with CRC.^[8] Another Japanese study also showed that AGR was an independent prognostic factor for OS in 941 patients with CRC.^[26] In an American study with 534 CRC patients, lowest AGR tertile was an independent predictor of 4-year survival rate compared with highest AGR tertile.^[27] However, they were limited with variable cutoff values, different study population, and inconsistent results.^[8,25–27] In our study, CAR, LCR, and AGR were not associated with OS in CRC patients and stage II to III CRC patients who underwent surgery. Our negative finding may warrant further studies for their prognostic roles for stage II to III CRC patients who underwent curative resection.

We concede that 1 of the limitations of this study is that it was a single-center, retrospective study, which may limit the generalization of our findings. However, we can fully assess detailed clinical and pathological information thorough comprehensive review of medical record by 1 author (YC). We excluded patients followed-up <1 year as they had insufficient survival data, which may well reflect the natural course of CRC. We used different cutoff values for inflammatory markers, but we adopted optimal cutoff levels with maximally selected rank statistics. Finally, we didn’t evaluate the prognostic factors associated with the relapse-free survival, as we have no relapse data.

5. Conclusion

NLR can be used as a clinically simple and useful prognostic parameter in stage II to III CRC patients who underwent curative resection. However, NLR deserves to be explored further for its optimal cutoff value as a prognostic marker in stage II to III CRC patients. CAR, LCR, and AGR, did not show any prognostic roles in stage II to III CRC patients who underwent curative resection.

Acknowledgments

We thank Lee SH for her statistical analysis. Editage® (www.editage.co.kr) provided assistance in the English revision of the manuscript through premium service.

Author contributions

Conceptualization: Yerim Cho, Jae Myung Cha.

Supervision: Su Bee Park, Jin Young Yoon, Min Sub Kwak, Jae Myung Cha.

Writing – original draft: Jae Myung Cha.

Writing – review & editing: Jae Myung Cha.