Preoperative Serum Lactate Dehydrogenase Level Predicts Progression and Prognosis in Patients with Glioma

Background: To evaluate the value of serum Lactate Dehydrogenase (LDH) level in predicting recurrence and the overall survival (OS) of glioma patients. Methods: A total number of 216 patients with glioma in our institution were retrospectively recruited to analyze the relationship between preoperative serum LDH level and prognosis. Results: Overall, the average age of patients was 43.58±17.22 years old; 53.7% (116 of 216) of the enrolled patients were male. Multivariate analysis revealed that serum LDH level (odds ratio [OR]=0.97, 95% con�dence interval [CI]=0.96-0.98, P<0.001) and World Health Organization (WHO) grade (grade II: OR=19.64, 95%CI=5.56-69.35, P<0.001; grade III: OR=19.50, 95%CI=7.08-53.73, P<0.001; grade IV: OR=15.23, 95%CI=4.94-46.97, P<0.001) were signi�cant and independent of 1-year Progression-free survival (PFS) after adjusting for confounders. The predictive performance of serum LDH level was represented with area under curve (AUC) = 0.741, 95%CI=0.677-0.798. Multivariate Cox analysis revealed that LDH level (hazard ratio [HR]=2.56, 95%CI=1.59-4.15, P<0.001) and WHO grade (grade II: HR=4.58, 95%CI=0.56-37.23, P=0.155; grade III: HR=16.35, 95%CI=2.16-123.80, P=0.007; grade IV: HR=42.13, 95%CI=5.83-304.47, P<0.001) remained associated with survival at 2-year follow-up. At 3-year follow-up, lymphocyte count (HR=0.68, 95%CI=0.51-0.91, P=0.008), LDH level (HR=2.21, 95%CI=1.40-3.49, P=0.001), and WHO grade (grade II: HR=1.44, 95%CI=0.44-4.68, P=0.543; grade III: HR=4.99, 95%CI=1.68-14.87, P=0.004; grade IV: HR=16.96, 95%CI=6.13-46.93, P<0.001) remained associated with survival in multivariate Cox analysis. Conclusion: Our study demonstrated that preoperative serum LDH level could serve as a reliable indicator for predicting prognosis of glioma patients. Further multicenter studies are still required to verify our �ndings.


Introduction
Glioma is a common type of intracranial malignant tumor, accounting for 15.1% of central nervous system tumor.There are approximately 30 thousand people die from glioma worldwide every year 1 .The clinical behaviors of glioma are highly variable and complex, including hemiplegia, epilepsy, and vomiting.The maximum safe surgical resection combining with radiation therapy and chemotherapy is the current standard therapy which improved the prognosis of those patients.Nonetheless, survival duration remains short, and the median survival time of glioblastoma patients is only 15 months 2−4 .High-grade glioma is characterized by high mortality rates and rapid progression.Moreover, it is di cult to ensure the quality of life in patients with glioma due to high rate of operative complications.Predicting prognosis in patients with glioma is both a di cult and important task for clinicians.
Glioma cells are in a microenvironment which is characterized by chronic in ammation and mediates tumor progression.The progression and development of glioma are closely related to the in ammatory state and immune reaction of the body.Therefore, the in ammatory biomarkers which re ect host in ammation status may also indicate the prognosis of glioma.In recent studies, in ammatory markers in peripheral blood including neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), and lymphocyte-monocyte ratio (LMR) have been con rmed closely related to the prognosis of patients with glioma 5−10 .Lactate Dehydrogenase (LDH) is serum in ammatory marker which has been applied to the prediction of pneumonia 11−13 .However, it remains unknown whether LDH could be used as a prognostic predictor in patients with glioma.
Virtually almost all patients with glioma will undergo routine blood tests including serum LDH level before surgery.Those tests could evaluate the clinical state of patients and may serve as surrogate biomarkers for some speci c changes caused by pathological characteristics of the tumor.Due to their standardization, readily availability, and cheapness, it is extremely attractive that the routine blood tests could function as proxy biomarkers for glioma and prognosticators for patients.In our study, we aimed to evaluate the value of serum LDH level in predicting recurrence and the overall survival (OS) of glioma patients.

Patient Characteristics
The demographics and baseline characteristics of the 216 patients were presented in Table 1.In this study, all patients were divided into a recurrence group (n = 77, 35.60%) and a non-recurrence group (n = 139, 64.40%).Overall, the average age of patients was 43.58 ± 17.22 years old; 53.7% (116 of 216) of the enrolled patients were male and 46.30% of them were female.Age, leukocyte count, neutrophil count, lymphocyte count, NLR, LMR, serum sodium level, serum potassium level, serum LDH level, World Health Organization (WHO) grade were signi cantly different between the two groups.The difference of Karnofsky Performance Status (KPS) for the two groups was not signi cant, P = 0.057.Utilizing ROC analysis, the prognostic value of preoperative LDH level and WHO grade were presented in Fig. 1.The best cut-off value of preoperative LDH level for predicting 1-year PFS was 179 U/L.The predictive performance was represented with AUC = 0.741 (95%CI = 0.677-0.798),sensitivity = 74.8%,speci city = 63.6%, and Youden index = 0.385.Based on the best cut-off value of grade III, the predictive performance of WHO grade (AUC = 0.810, 95%CI = 0.751-0.860,sensitivity = 84.4%,speci city = 77.0%,and Youden index = 0.614) was also calculated by ROC analysis (Fig. 1).DeLong's test indicated that the AUC of preoperative LDH level was comparable with that of WHO grade (Z = 1.363,P = 0.170).

Univariate and Multivatiate
In the univariate Cox hazard regression analysis (Table 4)  4).

Discussion
In adults, glioma is the most common primary intracranial tumor.It is characteristic of high invasion, high recurrence risk and short survival especially in glioblastomas.Once tumor relapse occurs, many patients could not live independently which causes a considerable burden to the patient and society.Finding effective measures to delay the disease progression and improve the survival time has been the focus of intense research.Understanding the crucial markers that play a role in glioma malignant behaviors may assist clinicians in educating patients about prognosis and monitoring therapeutic e cacy.Because of this, a reliable prognostic predictor in glioma has long been appreciated.In recent years, histopathological, immunohistochemical, and molecular features has attracted growing attention for its tremendous potential on suggesting the treatment sensitivity and predicting the prognosis 14 .However, molecular detections are complex, expensive, and time-consuming, a convenient, cheap, and routine available serum biomarker is more popular in clinical uses.
LDH is known to be a crucial glycolytic enzyme that exists widely in multiple organs.Moreover, it is also a nonspeci c in ammation marker and could have an increased concentration in in ammatory process 15 .
When tissues are injured, the degree of injury can be sensitively re ected by the serum LDH level 16 .
Elevated LDH could indicate an active status of in ammation.Indeed, as a well-known marker of in ammation, LDH has been widely recognized for their predictive performance in pneumonia [11][12][13] .In patients with melanoma, non small cell lung carcinoma, metastatic breast cancer, and metastatic pancreatic cancer, elevated LDH has been demonstrated to be correlated with poor prognosis [17][18][19][20] .However, the prognostic value of LDH in glioma patients has not received the attention it deserves.
In our study, we evaluated the prognosis of glioma patients by utilizing 1-year PFS, 2-year OS, and 3-year OS.We found that the preoperative serum LDH level of recurrence group was signi cantly higher than that of non-recurrence group.The same result was achieved in the postoperative serum LDH level between the two groups.In addition, the postoperative serum LDH level was signi cantly higher than preoperative serum LDH level, but with similar ΔLDH (P = 0.289).These results suggested that the surgical effect was near-equal to the recurrence group and non-recurrence group.Meanwhile, they also con rmed that the basic level of serum LDH was determined by their own in ammatory status which further affected their prognosis.Malignant tumor could evoke both local and systemic in ammatory responses, which has been recognized as cancer-related in ammation 22 .Moreover, immune response and in ammation response are closely correlated with the occurrence and development of malignant tumor.The interaction between glioma cells and their microenvironment has received growing attention over the past years.
In ammatory mediators and in ammatory cytokines in the tumor microenvironment could in uence glioma progression by participating the metabolism and apoptosis in glioma cells 23 .The in ammation status has already been proven to affect proliferation rate in vivo and in vitro 24 .Also, it has been shown that alteration of the tumor microenvironment could induce the malignancy progression of low-grade glioma under the regulation of in ammatory cytokines 25 .Therefore, peripheral blood in ammatory biomarkers may provide novel predictors for disease progression and prognosis.
It has been constantly reported that peripheral blood in ammatory biomarkers have predictive value for prognosis of glioma patients.Weng et al proposed that NLR combining with contrast enhancement were clearly associated with the proliferation potency of glioma 7 .He et al reported that pretreatment NLR and albumin-to-gamma-glutamyl transferase ratio aid in predicting diagnosis of grade III oligodendroglial gliomas 8 .Wang et al con rmed the prognostic value of preoperative in ammation markers including NLR and PLR in glioblastomas 5 .However, the reliability of peripheral blood in ammation biomarkers for predicting prognosis of glioma is still in dispute.In a previous study, recently reported by Mass et al, routine blood test including various types of in ammation biomarkers could not independently predict survival in patients with glioblastoma 26 .
LDH was not valued equally with those mentioned in ammatory markers by clinicians.But in fact, the basic scienti c research has clari ed the intimate relationship between LDH and glioma.In glioma cells, overexpression of LDH gene could promote their proliferation, invasion, and glycolysis 27,28 .These experimental ndings suggested that glioma patients with abnormally elevated LDH level were more likely to suffer from a dismal prognosis due to their robust state of in ammation, stronger ability for proliferation, and more aggressive behavior.They were consistent with our study and also provided an objective experimental basis for the further exploration of LDH in glioma.In our study, serum LDH level was con rmed as an independent prognostic predictor in patients with glioma for the rst time.It showed good performance both in predicting 2-year survival rate and 3-year survival rate which provided a reliable means to assess the risk of death in glioma patients.Also, it con rmed the crucial role of LDH gene playing in glioma cells in turn and provided new insight into individualized precision therapy for patients.
There were also several inherent limitations apply to our study as well.First, as a single-center, retrospective, and observational study, insu cient sample size could cause selection bias and some interference to the study.Second, as a nonspeci c biomarker of in ammation, other unidenti ed physiological and pathophysiological alterations in the organism may lead to abnormal elevation of serum LDH level.Third, the measurement of LDH level was only performed once before surgery, which may increase the potential bias.Therefore, the predictive value of serum LDH level for the prognosis of glioma patients still needs to be further validated. In

Radiological Examination and Follow-up Evaluation
Two neuroradiologists blinded to the patient information independently analyzed the preoperative and postoperative magnetic resonance imaging (MRI) or computed tomography (CT) images.For each case, preoperative MRI or CT scan was performed to evaluate the diameter of tumor and peritumor edema.
After surgery, a follow-up contrast-enhanced CT scan or MR imaging was performed at the rst month.
The subsequent intervals of follow-up were 3 months at the rst year, 6 months at the second year, and 12 months at the next year.Progression-free survival (PFS) in this study was de ned as the interval from surgery to relapse con rmed on the imaging.The recurrence group contained patients with growing residual tumor or new lesions detected during the rst year follow-up period.The other patients without those lesions were included into the non-recurrence group.Overall survival was de ned as the interval from surgery to death or last follow-up.At the end point, the 2-year survival rate and 3-year survival rate of patients were estimated.

Statistical Analysis
Statistical analysis was performed using SPSS 17.0 statistical software (SPSS, Inc., Chicago, Illinois, USA).P value less than 0.05 was considered as statistically signi cant.Continuous variables, presented as mean ± standard deviation, were analyzed by 2-sample t test.They were expressed as median (interquartile range) and analyzed by non-parametric test if they did not meet normal distribution.Categorical variables, described as frequency (percentage), were compared by Pearson χ 2 test or Fisher exact test.
Univariate logistic regression analysis was used for analyzing the risk factors of 2-year PFS and 3-year PFS.Variables with P < 0.10 in univariate analysis were included for further multivariate analysis.Backward stepwise multivariate regression analysis incorporating all those variables with univariate association was performed to create the nal model.The predictive value of factors was evaluated by the receiver operating characteristic (ROC) curve analysis.The predictive performance of predictor was presented with its sensitivity, speci city, and Youden index.DeLong's test was used to compare the predictive performance in different predictors by area under the curve (AUC).Kaplan-Meier curve analysis was performed to assess 2-year survival rate and 3-year survival rate after surgery.The log-rank test was used to calculate P value for comparing survival curves.The univariate and multivariate Cox proportional regression models were utilized to evaluate the prognostic signi cance of variables.

Table 3
Univariate and Multivariate Cox Hazard Regression Analysis of 2-year OS with Possible Predictive Factors Univariate and Multivatiate Survival Analysis in Patients with and without Elevated Preoperative LDH at 3year Follow-up OS: overall survival; HR: hazard ratio; CI: con dence interval; WBC, white blood cell; NEU, neutrophil; LYM, lymphocyte; NLR, neutrophil-lymphocyte ratio; LMR, lymphocyte-monocyte ratio; LDH, Lactate dehydrogenase; WHO, World Health Organization.

Table 4
Univariate and Multivariate Cox Hazard Regression Analysis of 3-year OS with Possible Predictive Factors In the univariated logistic regression analysis, age, KPS score, white blood cell count, neutrophil count, lymphocyte count, NLR, LMR, serum sodium level, serum potassium level, serum LDH level, and WHO grade were all signi cantly associated with 1-year PFS.-year survival rate were statistically correlated with the abnormally elevated LDH level (P<0.001).Subgroup analysis based on WHO grade got equally meaningful results.Not only that, multivariate Cox analysis revealed that both serum LDH level and WHO grade were independent prognostic indicator of 2year survival rate and 3-year survival rate in glioma patients.Patients with abnormally elevated LDH level had 2.56-fold risk of death than those with normal LDH level at 2-year follow-up.The risk was 2.21 times higher than patients with normal LDH level when the length of follow-up was extended to 3 years after surgery.In addition, the risk of death in patients with high-grade glioma was quite high at 2-year follow-up ( 95%CI = 0.751-0.860),sensitivity = 84.4%,speci city = 77.0%,and Youden index = 0.614.DeLong's test indicated that the predictive performance of serum LDH level was comparable to glioma WHO grade (Z = 1.363,P = 0.170).In addition, the serum LDH level showed an excellent predictive performance in patients with low-grade glioma (AUC = 0.934, 95%CI = 0.866-1.000,sensitivity = 100.0%,speci city = 74.0%,and Youden index = 0.740).The Kaplan-Meier curves analysis suggested that the higher 2-year survival rate and 3 conclusion, our study demonstrated that preoperative serum LDH level could serve as a reliable indicator for predicting prognosis of glioma patients.Abnormally elevated LDH level may suggest rapid proliferation and progression in glioma cells.The generalizability and convenience of this serum biomarker make it clinically feasible to play a role in almost all hospitals.It also has practical clinical implications for the preoperative management and individualized treatment of glioma patients.Further multicenter studies with large sample size and long-term follow-up are still required to verify our ndings.January 2019, 216 patients who met the inclusion criteria and exclusion criteria with glioma at the First A liated Hospital of Fujian Medical University were retrospectively recruited.It was approved by local ethics committee of the First A liated Hospital of Fujian Medical University which waived the requirement of informed consent due to the retrospective design and conformed to the ethical guidelines of the Declaration of Helsinki.The inclusion criteria are as follows: procedures within one week before surgery.The interval of postoperative reexamination was from 1 week to 1 month after surgery.The normal values of LDH are from 120 to 250 U/L.ΔLDH was calculated by subtracting preoperative LDH from postoperative LDH.
Patient data including general demographics, past medical history, surgical schemes(gross-total resection, subtotal resection, or biopsy), treatment regimens ( concurrent chemoradiotherapy, CCRT), and other related data were collected.All patients underwent general preoperative blood tests by extracting fasting venous blood in the morning.The routine blood examinations were underwent by standard laboratory test