The Clinical Value of Lipid Abnormalities in Cervical Cancer

Background: To describe the characteristic of plasma lipid prolife among cervical cancer and further evaluate the prognostic signicance of lipid levels in cervical cancer. Methods: We retrospectively reviewed 1,713 patients with cervical cancer in our hospital. Preoperative plasma lipid prole, including cholesterol (TC), triglyceride (TG), high-density lipoprotein cholesterol (HDL), and low-density lipoprotein cholesterol (LDL), of 1,713 cervical cancer patients were compared with 10,397 healthy women. Then, we evaluated the impact of lipids on overall survival (OS) and recurrence-free survival (RFS) in cervical cancer using univariate and multivariate Cox model. Results: While plasma TC, TG and LDL were signicantly higher, HDL was lower in patients with cervical cancer as compared with healthy women. TG was identied as independent predictors for RFS, and OS among patients with cervical cancer. Further stratied by age, patients with higher TG showed a signicantly worse RFS and OS compared with those with lower TG among patients ≥ 50 years old, but not among <50 years old. Conclusion: Cervical cancer was associated with a disordered lipid prole. Hypertriglyceridemia was an independent poor prognostic indicator for cervical cancer, especially for elder patients. Strengthening lipid management may be benecial for improving postoperative OS and RFS in patients with cervical cancer. involvement, lymphovascular space invasion, inltration depth, tumor size and increased pre-operative level of SCC-Ag.


Background
Cervical cancer is the most common gynecologic malignancy with an estimated 596,847new cases and 311,365 deaths in 2018 worldwide [1]. High-risk subtypes of the human papillomavirus (HPV) infection has been well established as the main cause for cervical carcinoma. In the past decade, the use of screening programs and HPV vaccination program has dramatically reduced the incidence of cervical cancer in developed countries. However, these diseases remain a heavy health burden and a major public health problem in the developing world because of the high incidence and poor prognosis of recurrent disease. Therefore, it is imperative to identify some simple and valuable predictor that could identify highrisk cervical cancer patients, poor progression and prognosis, further facilitating novel treatment strategy and improving clinical out coming.
A large body of evidence has shown a relationship between lipid abnormalities and cancer initiation [2].
Lipids crucially contribute to cell proliferation and tumorigenesis, due to its in uence on chemical-energy storage, cellular signaling, cell membranes, and cell-cell interactions [2]. Pre-clinical studies have revealed the ability of adipocytes providing energy for rapid cancer growth and metastasis [3]. A positive correlation was found between serum triglycerides (TG) and risk of endometrial cancer [4]. High levels of total cholesterol (TC) and low density lipoprotein (LDL) increased colorectal cancer risk [5].
Hypertriglyceridemia increased the risk of prostate cancer and promoted the aggressiveness of these disease [6,7]. With regard to cervical cancer, triglyceride was reported elevated among cervical cancer patients in an India population [8]. However, no signi cant difference in any lipid parameters was detected between cervical cancer group and uterine leiomyomas group in a study by Sun et al [9]. Thus, the alterations of plasma lipid pro les in cervical cancer remains a matter of controversy.
Recently, lipid parameters were described as a prognostic factor in various cancers, with contradictory ndings. While increased levels of LDL were reported as a negative prognosis in nasopharyngeal carcinoma [10], higher preoperative LDL was related to an improved 5-year RFS of ovarian cancer [11].
Besides, high density lipoprotein (HDL) was a favorable prognostic indicators of lung cancer and breast cancer [12,13]. So far, the impact of lipid abnormalities in cervical cancer has been poorly investigated. In the current study, we rst evaluated the correlation between lipid parameters and cervical cancer, then further explored the prognostic value of preoperative lipid pro le in a large population of cervical cancer patients treated with radical hysterectomy, and thus, identi ed a reliable and convenient predictor.

Study population
This study included 1,713 patients with pathologically con rmed uterine cervical carcinoma, who underwent radical hysterectomy between January 2008 and December 2018 at the First A liated Hospital of Wenzhou Medical University, China. The following exclusion Criteria was used: (1) Women who received any drugs that impacted lipid metabolism; (2) Patients with chronic diseases that effected lipid levels (i.e., diabetes); 3) Patients received any treatments before serum collection. Information mentioned above was obtained from the electronic medical records. Besides, the control group included 10,397 healthy women. This study was approved by the Ethics Committee of the First A liated Hospital of Wenzhou Medical University and informed consent was signed by the patients before taking part in this study. A detailed review of patient history, general physical examination, pelvic examination (including bimanual pelvic and rectal examinations), preoperative laboratory (plasma lipid pro les, and SCC-Ag ), pathological data (metastasis, stage, and differentiation) of all patients were collected from electronic medical records and reviewed. Detailed clinical data were collected within one week before operation. Preoperative plasma levels of HDL, LDL, TC, TG, and serum squamous cell carcinoma antigen (SCC-Ag) were measured in early morning before surgical operation and immediately measured using a Hitachi 7600-020 automatic biochemical analyzer with the kinetic method [11]. Body mass index (BMI) was calculated as body-weight(kg)/height ( m) 2 .
All of 1,713 cervical cancer patients were classi ed as high-risk, intermediate-risk, and low-risk after postoperative pathological evaluation. High-risk patients de ned as the presence of tumor involvement of the parametria, positive margins, or lymph node metastases [14]. Intermediate risk factors included depth of invasion, lymphovascular space invasion, and tumor size [15]. Patients without high and intermediate risk factors are de ned as low-risk.
Besides, we collected age, BMI, plasma levels of HDL, LDL, TC, TG from 10,397 healthy women as control. Given that lipid levels were affected by confounders such as age and BMI, we matched the healthy women group and the cervical cancer group with age, with every 10 years as a subgroup. Then, by randomly matching the healthy control group and the cancer group with a ratio of 2: 1, we obtained 3,426 healthy women. We then controlled the potential confounders (age and BMI) by regression analysis and analyzed the difference of lipid levels between healthy group and cancer group.
Follow-up examinations were performed every 3 months in the rst 2 years, and then every 6 months for the next 3 years and every 1 year thereafter. Pelvic examination, cytology, the serum concentration of SCC-Ag, and imaging studies including Computed Tomography, Magnetic Resonance Imaging, or Positron Emission Tomography-Computed Tomography, were performed during routine follow up.
The last follow-up date was July 15, 2019. The end points of this study were overall survival (OS) and recurrence-free survival (RFS). Overall survival was determined from the date of surgery to death or last follow-up. Recurrence-free survival (RFS) was calculated from the date of surgery to the tumor recurrence or distant metastasis.

Statistical analysis
Continuous data were presented as median (Q1-Q3) or mean ± standard deviation based on their distribution. and then difference between groups were tested using the Wilcoxon rank-sum test or a standard t test. Categorical data were analyzed using Fisher's exact test or χ2 test. The Kaplan-Meier method was used to calculate survival estimates for OS and RFS. Univariate and multivariate Cox regression analyses were used to evaluate the associations between clinical covariates and survival. A two-tailed P < 0.05 was considered statistically signi cant. Analyses were performed using R(version 3.3).

Characteristics of cervical cancer
Clinical and pathologic characteristics of cervical cancer were summarized in Table 1. This study included 1,713 cervical cancer patients. The median age at diagnosis was 52 years (range: 22-82 years) and the median BMI was 23.24 (19.95, 26.53). The most common histological type was squamous cell carcinoma (85.9%), followed by adenocarcinoma (9.0%). All of patients were diagnosed with FIGO stage I-II. Tumor sizes in 221 patients (22.9%) were ≥ 4cm, 560 (58.3%) were 2-4cm, and 181 (18.8%) were < 2cm. In addition, 274 patients showed lymph node metastasis, 342 cases were lymphovascular space invasion and 40 patients were parametrial involvement. Based on pathology after radical hysterectomy, there were 319 (18.6%) patients with high risk factors, 643 (37.5%) patients with intermediate risk factors.  Figure 1 showed the comparisons of plasma lipid levels between healthy women and patients with cervical cancer. While plasma levels of TC, TG, and LDL were signi cantly higher, plasma HDL levels were signi cantly lower among patients with cervical cancer as compared with healthy women (Fig. 1A).
Given that age may in uence lipid levels, all the individuals were further divided into young (< 50 years) and old groups (≥ 50years). Similarly, patients with cervical cancer showed higher TC, TG, and LDL levels, and lower HDL as compared with healthy women both in young and old groups (Fig. 1B). Subsequently, we matched the healthy group and the cancer group with age, with every 10 years as a subgroup. Then, by randomly matching the healthy group and the cancer group with a ratio of 2: 1, we obtained 3,426 healthy women. Patients with cervical cancer showed higher TC, TG, and LDL levels, and lower HDL levels than healthy women ( Supplementary Fig. 1). The difference of lipid levels, including TC, TG, LDL and HDL, between healthy group and cancer group remained statistically signi cant after further adjustment for age and BMI (P < 0.001, P = 0.001, P < 0.001, P = 0.008, respectively). Thus, compared with healthy women, patients with cervical cancer were coupled to a disordered lipid pro le characterized by higher TG, TC and LDL levels, and lower HDL levels. Initially, univariate analyses were employed to depict the prognostic value of clinicopathological features among patients with cervical cancer. As expected, the factors that were statistically signi cant in predicting poor RFS as well as OS were positive lymph node metastasis, parametrial involvement, late FIGO stage, positive lymphovascular space invasion, deeper in ltration depth, large tumor size and increased pre-operative level of SCC-Ag (Table 2). Next, we explored the prognostic value of the lipid pro le among patients with cervical cancer. Cut-off was determined according to the medium lipid levels of 1,499 cervical cancer patients. Patients were divided into "low" and "high" according to the comparisons between lipid pro le levels with established cut-offs. Table 3 showed the RFS and OS of the patients included in the study according to TC, TG, LDL and HDL levels. As shown in Table 3, patients with high TG levels had a signi cantly worse RFS than those with low TG levels both in univariate and multivariate Cox model. However, null association was detected between TC/LDL/HDL and RFS of cervical cancer.  Table 3). The Kaplan-Meier curves for RFS and OS of the two groups (high TG levels vs low TG levels) were further illustrated in Fig. 2.

Lipid Pro le Levels and survival
Since lipid levels were commonly associated with age, all the individuals were further divided into two groups according to age (< 50 years vs. ≥50 years). Strati ed by age, interestingly, no signi cant association was detected between lipid pro le and RFS/OS of cervical cancer among patients < 50 years old. Nevertheless, patients with higher TG had a signi cantly worse RFS and OS than those with lower TG among patients ≥ 50 years old (Table 4). Cervical cancer patients with high-risk factors showed worse prognosis as comparing with patients with intermediate-risk or low-risk factor, we further analyzed patients with high-risk factors. As shown in Fig. 3 and Fig. 4, higher level of TG was an independent poor prognostic factor among cervical cancer patients with high risk (Fig. 3, Fig. 4).

Discussion
In this study, we explored the clinical signi cation of lipids pro le among cervical cancer patients, and made several important discoveries. Firstly, we comprehensively investigated the difference in serum lipid levels between healthy women and patients with cervical cancer using a relative large number of population, and demonstrated that TC/TG/LDL were elevated, and HDL were decreased in patients with cervical cancer. These results suggest dyslipidemia could be correlation with cervical cancer. Secondly, to our knowledge, this is the rst report to evaluate the prognostic value of lipids pro le in cervical cancer. Our study showed that hyperlipidemia was an independent prognostic factor for both RFS and OS among patients with cervical cancer.
Lipids are essential components of cell membranes, lipid metabolism essentially contributes to tumor cell bioenergetics and biomass formation [16]. Cholesterol and triglycerides are the two main lipids in plasma [17]. LDL and HDL are lipoproteins responsible for cholesterol transportation. While the LDL lead cholesterol to the cells and facilitate the deposition of fat in the vessel, the HDL is on the opposite promotes the removal of excess cholesterol [17]. Epidemiological studies, despite being controversial, reported the correlation between plasma triglyceride levels and the risk for cancer. Higher levels of triglyceride was associated with increased risk of lung, thyroid cancer, renal cancer, prostate cancer, gynecological cancer in a large cohort study in Austria [18]. Similarly, triglycerides were reported to increase the risk for cervix cancer in females, and colon and thyroid cancer in males in a cohort of 22,946 Icelanders [19]. Conversely, lower triglyceride (< 1.70 mmol/L) was related to an evaluated cancer risk in Chinese type 2 diabetes mellitus patients [20]. Our observation of higher lipid prolife in cervical cancer was similar to a previous cohort study among Icelanders, which regarded triglycerides as a positive risk factor for cervix cancer [19]. Further evidence for a relationship with triglyceride comes from India [8]. Raju et al reported TG was elevated in cervical cancer as comparing to healthy control [8]. Therefore, cervical cancer are coupled to a disordered lipid pro le characterized by higher TG, TC and LDL levels, and lower HDL levels. Dyslipidemia could be associated with cervical cancer.
Our current study further explored the prognostic value of hyperlipidemia at diagnosis in cervical cancer. We provided rst evidence that higher plasma triglyceride levels correlated to a worse RFS as well as OS among patients with cervical cancer. This observation was consistent with previous studies in prostate cancer and breast cancer. Elevated serum triglycerides increased the risk of prostate cancer recurrence [21]. Hypertriglyceridemia was correlation with a decreased 5-year OS among patients with triple negative breast cancer [22]. In a recently published, monocentric, retrospective study, Vernieri et al reported higher plasma triglyceride levels correlated with lower progression-free survival in everolimus-treated patients with advanced pancreatic neuroendocrine tumors [23]. These ndings, coupled with evidence that adipocytes provided energy for rapid cancer growth and metastasis in vivo and in vitro [3], indicate that increased utilize of extracellular lipids, or their de novo synthesis, could increase the recurrence risk and result into poor clinical outcomes among cervical cancer. Therefore, we recommend lipid management for patients with cervical cancer, especially those have hyperlipidemia. Accordingly, lipid pro le was recommended routinely evaluated during follow-up Strati cation by age revealed a negative association between hypertriglyceridemia and prognosis in patients ≥ 50 years old rather than in patients < 50 years old, suggesting hypertriglyceridemia was associated with poor outcomes of cervical cancer patients, especially for elder patients. Sub-analysis for cervical cancer patients with high-risk factor, showed that higher level of TG was an independent poor prognostic factor for this population. These sub-analysis further supported preoperative lipid pro le as a promising prognostic predictor in cervical cancer by controlling the confounding data.
Our study has some limitations. As a retrospective study, it was limited availability of collecting the change of lipid pro le during follow-up. This is because serum lipid pro le was not routinely evaluated in all patients during follow-up. A prospective randomized study evaluating the survival bene ts of controlling lipids in primary cervical cancer is warranted.

Conclusions
In conclusion, a disordered lipid pro le characterized by higher TG, TC and LDL levels, and lower HDL levels could be associated with cervical cancer. TG exerted unfavorable in uence either on RFS or on OS of cervical cancer. Hypertriglyceridemia was an independent negative prognostic predictor for cervical cancer. This simple and routinely tested parameters may be a convenient prognostic candidate for cervical cancer in clinic. Strengthening lipid management may be bene cial for improving postoperative OS and RFS in patients with cervical cancer.

Declarations --Acknowledgments
We would like to thank all doctors, nurses, patients, and their family members for their kindness to support our study.
--Consent for publication Not Applicable.
--Ethics approval and consent to participate This study was approved by the Ethics Committee of the First A liated Hospital of Wenzhou Medical University and informed consent was signed by the patients before taking part in this study. All patients agreed to use their clinical data for clinical research before being hospitalized.

--Competing interests
The authors declare that they have no competing interests.
--Availability of data and materials Available under request.
--Authors' contributions: HZ conception and design the study;. AL, WW, LW acquisition of data, AL, MJ and YS analysis of data; AL drafting the manuscript; HZ revising this manuscript. All authors read and approved the nal manuscript.
--Funding information: Not Applicable.  Kaplan-Meier curves for recurrence-free survival (A) and overall survival (B) of patients with cervical cancer according to TG levels. Patients were strati ed into high and low groups according to the cut-off value of 1.49 mmol/L. High TG levels are associated with a poor RFS as well as OS of cervical cancer. P values were estimated using log-rank test.

Figure 3
Multivariate analyses of recurrence-free survival of cervical cancer patients with high-risk factor. P values were estimated using log-rank test.

Figure 4
Multivariate analyses of overall survival of cervical cancer patients with high-risk factor. P values were estimated using log-rank test.