Is preoperative serum lactate dehydrogenase useful in predicting the outcomes of patients with upper tract urothelial carcinoma?

Abstract Background Lactate dehydrogenase (LDH) has been proved to be associated with clinical outcomes in various carcinomas; however, limited evidence was available in upper urinary tract urothelial carcinoma (UTUC). Thus, the aim of this study was to evaluate the prognostic impact of LDH in UTUC. Patients and methods A cohort of 668 patients WERE retrospectively included between 2003 and 2016. Kaplan‐Meier method and Cox proportional hazards regression models were used to evaluate the association of LDH with overall survival (OS), cancer‐specific survival (CSS), disease recurrence‐free survival (RFS), and metastasis‐free survival (MFS). The cutoff level of LDH was set at 220 U/L for the upper limit of normal. Results Kaplan‐Meier plots showed the group with elevated LDH had significant poor OS (P = 0.003), CSS (P = 0.005), and RFS (P = 0.005), but not MFS (P = 0.099). However, multivariate Cox analysis suggested that LDH was not an independent predictor for CSS (HR 1.50, 95%CI: 0.87‐2.59), OS (HR 1.56, 95%CI: 0.94‐2.58), RFS (HR 1.33, 95%CI: 0.83‐2.12), or MFS (HR 1.16, 95%CI: 0.79‐1.71). Albumin, globulin, and HBDH were also not related to survival outcomes of UTUC patients in multivariate analysis, while higher alkaline phosphatase was associated with worse CSS and OS, and higher white blood cells contributed to poor CSS and RFS. In subgroup analysis, results found higher LDH was associated with poor OS in patients with localized disease (pT ≤ 2) (HR 4.03, 95%CI: 1.37‐11.88). Conclusion The preoperative LDH was not an independent prognostic factor for patients with UTUC, while elevated LDH was proved to be correlated with worse OS in patients with localized disease.


| Patients
A total of 710 patients with UTUC received RNU with bladder cuff excision treatment between 2003 and 2016 in West China Hospital. The clinicopathological data including age, gender, anemia, perioperative blood transfusion, tumor side and location, size, tumor grade, TNM classification, lymph node status, surgical margin status, multifocality, concomitant variant histology (CVH), lymphovascular invasion (LVI), tumor architecture, and adjuvant therapy were collected. Patients with preoperative infection, liver diseases, fever, other tumors, with the previous cystectomy for invasive bladder cancer, or who received the treatment of neoadjuvant or adjuvant chemotherapy or radiotherapy before surgery were excluded. The LDH value and other serum biochemical patterns of each patient were extracted from the most recent routine examines within 1 month before surgery. Patients with missing LDH value were also excluded from our cohort. Lymph node dissection was not routinely performed. Eventually, 668 patients were enrolled in this retrospective study. The study was approved by the Ethics Committee of West China Hospital, and the methods were carried out in accordance with the approved guidelines. For this type of study, informed consent is not required.

| Pathological evaluation
All RNU specimens were, respectively, re-elevated by two separately specific pathologists according to standard procedures. The 2010 American Joint Committee of Cancer TNM classification and the WHO International Society of Urological Pathology consensus classification were used to evaluate the tumor stage and grade, respectively.

| Cutoff value selection
The cutoff value of LDH was determined as the upper limit of normal range from West China Hospital, which was 220 U/L. Higher than the cutoff value was considered as a high level of LDH. Other serum biochemical markers including alpha-hydroxybutyrate dehydrogenase (HBDH, >180 U/L vs ≤180 U/L), alkaline phosphatase (ALP, >90 vs ≤90 U/L), albumin (ALB, >35 vs ≤35 g/L), globulin (GLB, >30 vs ≤30 g/L), and white blood cells (WBC, >8.3 vs ≤8.3*10 9 /L) were also included in analysis.

| Follow-up strategies
Patients were followed every 3-4 months for the first year after surgery according to the guideline, semiannually for the second and third year, and annually thereafter, or as clinically indicated with urinary cytology and excretory urography of the contralateral upper urinary tract, and routine check-ups that included history, physical examination, blood laboratory tests, and chest radiography. If clinically indicated, selective bone scan and chest/abdomen CT/MRI were elevated.
Disease recurrence was defined as local recurrence in the operating field, lymph node spread and/or distant metastasis that had not been found in the preoperative examination. Specifically, the tumor found in the urinary bladder or contralateral upper urinary tract after surgery was not regarded as tumor relapse.

| Statistical analysis
Continuous variables were analyzed using Student's t test, and categorical variables were elevated using the chi-squared test or Fisher's exact test. Probabilities of overall survival (OS), cancer-specific survival (CSS), disease recurrence-free survival (RFS), and metastasis-free survival (MFS) were estimated using the Kaplan-Meier method, and the log-rank test was used to assess differences. Univariate and multivariable Cox's proportional hazards regression models were used to evaluate the relationships between variables and OS, CSS, RFS, and MFS. Risk factors with a P value <0.15 in the univariate analysis were included in the multivariate analysis model. Hazard ratios (HRs) with their 95% CIs were used to assess the strength of the individual variables. All reported P values were two-sided with statistical significance set at P < 0.05. Statistical analyses were performed using IBM SPSS Statistics version 22.0 (IBM Corp., Armonk, NY, USA).

| RESULT
The clinicopathological features of the patient cohort included in this study are shown in Table 1 respectively. The 2-year OS and 5-year OS were 69.2% and 55.2%, respectively. The CSS at the second and fifth year was 72.9% and 61.3%, respectively. Kaplan-Meier plots showed the group with high serum LDH level had significant poor OS (P = 0.003), CSS (P = 0.005), and RFS (P = 0.005) compared with that in the group with normal serum LDH values ( Figure 1A-C), but not with respect to MFS (P = 0.099) ( Figure 1D). Univariate Cox analysis showed that high preoperative serum LDH level was a poor prognostic factor for CSS (HR 1.54, 95%CI: 1.00-2.38), OS (HR 1.48, 95%CI: 1.00-2.18), RFS (HR 1.60, 95%CI: 1.12-2.29), and MFS (HR 1.63, 95%CI: 1.04-2.54) ( Table 2). Also alkaline phosphatase, white blood cells, and globulin were significantly correlated with CSS, OS, RFS, and MFS. Albumin was only associated with CSS and OS. After controlling for the effects of standard clinicopathological features, multivariate Cox analysis showed that serum LDH value was no longer an independent predictor for CSS (HR 1.50, 95%CI: 0.87-2.59), OS (HR 1.56, 95%CI: 0.94-2.58), RFS (HR 1.33, 95%CI: 0.83-2.12), or MFS (HR 1.16, 95%CI: 0.79-1.71) ( Table 3). Moreover, albumin, globulin, and HBDH were also not related to survival outcomes of UTUC patients. However, the higher level of alkaline phosphatase, anemia, tumor architecture, and CVH was shown as independent predictors for CSS and OS. In addition, results showed that tumor stage and grade, tumor size, and transfusion were significant prognostic factors for CSS, OS, RFS, and MFS. Tumor site and adjuvant therapy were also independent predictors for MFS (Table 3).

| DISCUSSION
Previous studies have demonstrated that high serum LDH level was associated with prognosis in several malignancies, such as renal cell carcinoma, melanoma, prostate cancer, squamous cell cancer, nasopharyngeal and colorectal cancer. 5,7 Limited evidence had shown high serum LDH level was related to unfavorable prognosis in patients with UTUC. 6 In our study, the results suggested that high LDH level was associated with gender, perioperative blood transfusion, and tumor grade, which was also partially in agreement with that of previous published study showing that the high LDH level might reflect heavier tumor burden in gastric cancer. 6,8 Although many studies found LDH was a good prognostic predictor for cancer patients, the exact mechanism was still unclear. Lactate dehydrogenase, which regulated by hypoxia-inducible factor-1 alpha (HIF-1a), is involved in the glycolytic pathway and catalyzes the conversion of pyruvate and lactate coupled with the conversion of NADH and NAD+. Thus, high LDH level could reflect the oncogenic aerobic glycolysis or the Warburg effect which can promote the malignant transformation and survival of cancer cells. 3,6,9 However, the results of our cohort suggested that high serum LDH level was not associated with overall CSS, OS, RFS, or MFS in UTUC patients, which was consistent with the results reported by Kluth et al. 10 Also, subgroup analysis in this study showed that serum LDH level could not affect survival outcomes of patients with the low-grade disease or high-grade UTUC. But interestingly, we found elevated serum LDH contributed to poor OS in patients with localized UTUC, which was totally different from the result of Zhang et al who reported elevated LDH was correlated with worse OS in patients having advanced disease. Although the cutoff value of LDH in Zhang et al 5,6 was set as 245 U/L, a little higher than that in our study, a recent meta-analysis found different LDH cutoff value did not affect the HR for survival outcomes. More importantly, 100 UTUC patients with only 10 cases having high LDH level were included in the study conducted by Zhang et al 6 and also they did not adjust other confounder biomarkers. These limitations significantly reduced the significance of their results.
In our cohort, we also found albumin, globulin, and HBDH were not associated with prognosis of UTUC patients, while higher alkaline phosphatase was found to be associated with worse CSS and OS in all patients and higher globulin was associated with poor CSS and RFS in cases with localized disease separately. In contrast, Kluth et al reported lower serum albumin level contributed to higher mortality after disease recurrence and Sheth et al found lower albumin was associated with worse RFS and OS without adjusting the impact of other confounder biomarkers. 10,11 Moreover, they did not exclude patients with liver diseases which could affect albumin levels and other biochemical indexes such as globulin, aspartate aminotransferase, and bilirubin. 10 Previous evidence found increased alkaline phosphatase level in patients with kidney disease and liver cancer. 12,13 Kluth et al reported alkaline phosphatase was not a prognostic predictor in patients with recurrent disease, but another study including patients with high-grade disease reported ALP ≥ 116 U/L was associated with adverse RFS and OS in univariate analysis and an AA score based on the cumulative number of alterations in albumin and alkaline phosphatase was proved to be an independent predictor of RFS and OS in multivariate analysis. 10,11 Sheth et al also suggested white blood cells were not a prognostic predictor in high-grade patients, while our cohort found higher white blood cells to be associated with worse CSS and RFS, which was in line with Kluth et al reported higher white blood cells were correlated with a higher after disease recurrence. 10,11 Recent studies also figured out that white blood cells were associated with worse CSS in patients with bladder cancer in univariate analysis. 14 In our study, the LVI was detected in only 99 (14.8%) patients, which was significantly lower than the majority of published studies. 15 Like Rink et al and some other previous studies reported, we also found the presence of LVI was not a risk factor for oncologic outcomes. [16][17][18][19] In addition, previous studies proved that LVI was not linked to oncologic outcomes in patients already had LN metastases. 20 As lymphadenectomy was a level C evidence recommended by guideline, it was not routinely performed in our cohort. 1 78.4% of patients were not staged with a lymph node dissection in our study. Thus, the prognostic impact of LVI may be underestimated in our cohort. Our results also showed that the surgical margin status was not an independent prognostic factor. Similarly, Kim et al 21 also found the margin status was not associated with disease-free survival or CSS in UTUC. Although the EAU guideline suggested that positive surgical margin (PSM) was an independent predictor of disease recurrence, we noticed the reference the guideline cited reported that PSM was only related to MFS but not associated with CSS or RFS. 22 In addition, some studies found the prognostic value of surgical margin only in the univariate analysis. 22 Moreover, Colin et al 22 also explained that in organ-confined disease (≤pT2), the most of PSM was caused by surgical mistake instead of the high invasive ability of cancer due to the dissection too close to the ureter when performing distal ureterectomy. Thus, the findings of surgical margins at tumor site, especially at the ureteral location, should be explained with caution. In our study, 10 cases with PSM were at a ≤pT2 stage. At last, many clinicopathological parameters were included in our multivariate models, so other factors may overpower the PSM effect. Some limitations of this study should be mentioned. First, the retrospective nature of this study may cause a selection bias. In addition, as there is no consensus on the lymphadenectomy pattern for UTUC and the benefits of lymphadenectomy remain uncertain, lymphadenectomy was not routinely conducted and the extent of lymph node dissection was not standardized. Moreover, the data of smoking were not available in this study; however, one most recent systematic review study found that smoking was not associated with OS in UTUC patients. 23 Finally, our cohort only included patients from our single center, and the results should be validated by well-designed prospective multi-institution studies in the future.

| CONCLUSION
Our study has found that preoperative LDH, albumin, globulin, and HBDH were not associated with survival outcomes in patients with UTUC, although higher LDH was found to be associated with poor OS in cases with localized disease. Moreover, higher alkaline phosphatase was proved to be independently correlated with worse CSS and OS, and also higher white blood cells were an independent predictor for CSS and RFS.