Prognostic significance of neutrophil-to-lymphocyte ratio in prostate cancer: evidence from 16,266 patients

This study was aimed to investigate the prognostic value of neutrophil-to-lymphocyte ratio (NLR) in patients with prostate cancer (PCa). A meta-analysis including 14 publications (15 cohorts) with 16,266 patients was performed to evaluate the association between NLR and overall survival (OS), progression-free survival (PFS)/recurrence-free survival (RFS) in PCa using hazard ratio (HR) and 95% confidence intervals (95% CI). The combining data showed that increased NLR predict poor OS (HR = 1.38, 95%CI: 1.22–1.56) and PFS/RFS (HR = 1.24, 95%CI 1.05–1.46) in PCa. Stratified analysis by PCa type, sample size, ethnicity and NLR cut-off value revealed that NLR showed consistent prognostic value in metastatic castration-resistant prostate cancer (mCRPC) patients and predict poor PFS/RFS in Asians, but not in Caucasians. These statistical data suggested that increased NLR could predict poor prognosis in patients with PCa.


Results
Selection and characteristics of included eligible studies. The selection process was shown in Fig. 1.

Sensitivity analysis. Omitting any single study by turn, sensitivity analysis demonstrated that the combined
HRs for OS and PFS/RFS did not significantly alter (Fig. 4).
Publication bias. Evaluation of publication bias using Begg's test (p < 0.05 was considered as statistical significant) demonstrated that there was no significant publication bias in OS and PFS/RFS studies (p = 0.119 and p = 0.251, respectively)( Fig. 5).

Discussion
In the present study, we aimed to explore the prognostic value of NLR in patients with PCa.  Growing evidence has indicated that inflammatory response could be heavily involved in the occurrence and development of different cancer types [26][27][28] . Studies revealed that inflammation-related neutrophils and immunocytes including lymphocytes were indispensable participants in tumorigenesis 29 . Inflammation exerts an important role in tumor formation and development through facilitating angiogenesis, proliferation and protecting tumors from apoptosis. By secreting a variety of chemokines, tumor cells could attract pro-inflammatory cells  into tumor microenvironment, subsequently, an array of cytokines produced by neutrophils stimulate tumor cells growth 30 . Of these inflammatory parameters reflecting the systemic inflammatory response, an increased neutrophil-to-lymphocyte ratio (NLR) has been found valuable to predict clinical outcome of cancer patients 31 . Additionally, NLR is obtained from routine blood test, making it an easily available and reliable marker. A large amount of studies have showed the correlation between increased pretreatment NLR and poor prognosis in different malignant tumors including gastric cancer, colorectal cancer, breast cancer, prostate cancer, soft-tissue sarcoma and non-small cell lung cancer 17,27,[32][33][34][35] .
Our results demonstrated that elevated NLR was in associated poor OS and PFS/RFS in patients with PCa, which was in accordance with the results from meta-analysis with other cancer types 34,36,37 . We have noted that a recently published work investigated the prognostic value of NLR in various solid tumors 11 . However, in that meta-analysis 11 , only three studies concerning patients with castration resistant prostate cancer were included. In the present work, we included patients with both localized PCa and mCRPC in an adequately sufficient data. Moreover, subgroup analysis was performed, which could provide detailed information for clinical management. More importantly, we found that NLR has adequate prognostic value for OS and PFS/RFS in mCRPC. Although chemotherapy and hormone therapy were used, mCRPC commonly occurred after the treatment after few years. Therefore, the NLR measurement could provide valuable prognostic information for mCRPC and be helpful for optimal treatment strategies selection. In addition, our results showed increased NLR predicted poor PFS/RFS in Asians, but not in Caucasians, which could be attributed to the ethnicity heterogeneity. Notably, corticosteroids were one of the palliative treatment options in patients with mCRPC for 30 years 38 . Intake of corticosteroids had immunosuppressive effects, which could influence the value of NLR 21 . The prognostic value of NLR in PCa should be evaluated after adjustment on potential confounders including use of corticosteroids. However, only one 21 of the included studies provided the relevant data, thus the analysis could not be conducted in the current meta-analysis due to insufficient data. Interestingly, Lorente et al. 21 found that NLR had independent prognostic value on OS regardless of corticosteroids usage using multivariable analysis. However, a recent work 39 found that, in patients with melanoma, elevated NLR and corticosteroids before week 1 were associated with poorer OS in univariate analysis, however, in multivariate analysis, elevated NLR remained an independent prognostic factor whereas the prognostic efficiency of intake of corticosteroids disappeared. Since corticosteroids were widely used drugs in mCRPC treatment, further studies should take into account intake of corticosteroids when performing survival analysis. Furthermore, more large scale studies are needed to explore the effects of use of corticosteroids on survival.
Although our study was the first meta-analysis concerning NLR in PCa prognostication, there were several limitations need to be addressed. First, vast majority of included publications employed samples of Caucasian ethnicity, thus the evaluation of OS and RFS/PFS in Asians might be derived by chance because of sample insufficiency. Second, Shafique et al. 16 recruited all types of PCa patients that could not be classified as any PCa types,  which may contribute to heterogeneity when subgroup analysis and meta-regression were performed. Third, only publications in English were included, which could cause language bias for selection. Further investigations are needed to address the above-mentioned shortcomings.
In summary, our study demonstrated that elevated NLR predict poor OS and PFS/RFS in patients with PCa. Increased NLR showed consistent prognostic value in mCRPC patients and predict poor PFS/RFS in Asians, but not in Caucasians. The present findings could provide implications for clinical management of patients with PCa and further investigations involving large sample size and more ethnic backgrounds are needed.

Methods
Literature search. A comprehensive literature searching of Pubmed, Embase and Web of Science database was conducted. The search strategy included the combinations of the following key words: (NLR OR neutrophil-lymphocyte ratio OR neutrophil-to-lymphocyte ratio) AND (prostate cancer OR prostate carcinoma OR prostatic neoplasms OR PCa) AND (prognosis OR survival OR outcome OR recurrence). The last search was updated on October 17th, 2015. We also manually checked the reference list to identify additional publications. The published language was limited to English.
Inclusion and exclusion criteria. Inclusion criteria for publication selection were as follows: (i) the diagnosis of PCa for patients was histopathologically confirmed; (ii) the value of NLR was obtained for blood sample testing; (iii) investigated the association of NLR with PFS, RFS or OS; (iv) provided HRs and 95% CIs for NLR in OS and (or) PFS/RFS, or HRs and 95%CIs could be calculated according to the raw data provided in the article; (v) defined the cut-off value of increased NLR; (vi) published in English language.
Major exclusion criteria were as follows: (i) letters, editorials, review articles; (ii) failed to provide data of interest or insufficient data to estimate HRs and 95%CIs; (iii) failed to identify the cut-off value for elevated NLR; (iv) animal studies and irrelevant studies.
Data extraction. Two investigators (XB,G and XS,G) independently gathered information from each eligible study. The following data was extracted: surname of the first author, study country, year of publication, ethnic origin of the subjects, sample size, subtype of PCa, treatment method, cut-off value defining elevated NLR and HRs with corresponding 95% CIs for PFS/RFS and(or) OS. Discrepancies between the two investigators were settled by discussion.
Statistical analysis. The HR and the corresponding 95%CI were used to assess the prognostic efficiency of NLR on PCa. HR and 95% CI were directly extracted from each single study, if provided, or calculated according to the methods clarified by Tierney 40 et al. Cochran's Q test and Higgins I-squared statistic were applied to test the heterogeneity of pooled data. I 2 > 50% and p < 0.1 indicated significant heterogeneity and random-effects model was adopted to combine the effective value. I 2 < 50% and p > 0.1 were considered as no heterogeneity and a fixed-effects model was then adopted. Sources of inter-study heterogeneity were explored using subgroup analysis and meta-regression. Begg's funnel plot was used to evaluate publication bias. Sensitivity analyses were carried out to access the robustness of the results. All p values were two-tailed, and statistical significance level was set at p < 0.05. All statistical analyses were conducted using STATA 12.0 software (STATA Corporation, College Station, TX).