Electrolyte and renal disorders in patients with newly diagnosed glioblastoma

Background Electrolyte disturbances and altered renal function have been linked to the prognosis of critically ill patients and recently also of cancer patients. Little is known about the prevalence and prognostic impact of electrolyte and renal disorders in patients with glioblastoma (GBM), the most frequent malignant primary brain tumor. This study aimed to assess electrolyte and renal disorders in GBM patients and evaluate their effect on patients’ outcome. Methods


Introduction
Glioblastoma (GBM) is the most frequent primary malignant brain tumor in adults. Prognosis remains poor with a median survival of 17-25 months despite multimodal therapies (1). Still, survival varies signi cantly among individuals and estimation of patients' prognosis evolves into a highly relevant aspect in the treatment of patients with GBM. Therapies comprise tumor resection or stereotactic biopsy followed by combined radio-/chemotherapy or solely one of both adjuvant modalities, according to the patients' clinical status, age, etc. Several prognostic factors have been identi ed in GBM, like patients' age, clinical performance status, methylation of O6-methylguanin-DNA-methyltransferase gene promotor (MGMT), or mutation of the isocitrate-dehydrogenase gene 1(IDH1) (2)(3)(4)(5). These parameters allow an early estimation of patients' prognosis after GBM diagnosis. However, the identi cation of other relevant survival markers, particularly of those assessable prior to surgery, is of eminent clinical importance and might be helpful during initial treatment planning.
Electrolyte disorders and altered renal function have long been linked to patients' prognosis in various diseases. Aberrations in serum electrolyte levels were shown to be associated with a higher mortality in intensive care units in critically ill adult patients (6,7). Several recent studies additionally suggest that electrolyte disorders might be associated with a worse prognosis in a variety of malignancies (8), e.g. in colorectal cancer (9), lung cancer (10) and lymphoma (11,12). To date, data on prevalence and eventual clinical impact of preoperative disturbances in serum electrolytes and renal parameters in GBM patients is scarce.
This study aimed to assess routinely investigated electrolytes and renal parameters in GBM patients at admission and to evaluate the impact of electrolyte and renal disorders on patients' survival.

Study population
In this retrospective longitudinal cohort study all patients (aged ≥ 18 years) with newly diagnosed and histologically con rmed GBM, that were treated between July 2005 and December 2018 in the University Hospital Essen were included. After surgery (microsurgical tumor resection or tumor biopsy), patients were transferred to adjuvant therapies (chemotherapy, radiation), or in accordance with patients' willingness, to best supportive care. The study was approved by the Institutional Ethics Committee.

Data management
The primary objective of the study was to evaluate the association between routinely analyzed electrolytes and kidney function parameters with initial demographic/clinical characteristics and survival of GBM patients. As outcome endpoints, overall survival (OS) and 1-year survival (1-YS) were assessed.
Preoperative serum blood samples were routinely collected 1-2 days prior to surgery. For electrolytes, sodium, chloride, potassium, and calcium were investigated. Kidney function was assessed using creatinine, urea, and glomerular ltration rate (GFR) GFR

Electrolytes and kidney function in GBM
In this cohort of GBM, electrolyte disorders occurred in 275 (30.6%) of GBM patients, whereas renal disorders were more frequent, affecting 544 patients (60.4%). In 74 patients (8.2%), more than one electrolyte parameter was deranged. The most frequent electrolyte disorder was hyperchloremia (13%), followed by hyponatremia (6.7%). A restricted GFR occurred in 29% of patients, whereas hyperuricemia was present in 48% of patients. Frequencies of electrolyte and renal disorders are shown in Fig. 1.
Associations of each electrolyte disorder with patients' baseline parameters were tested. Here, frequent associations with patients' age, preoperative clinical performance, and previous comorbidities were detected. For electrolyte disorders, no signi cant association with renal disorders were found. Detailed information on associations of electrolyte and renal disorders with the above-mentioned parameters are presented in Table 2. 3.3. Association of electrolyte disorders with GBM outcome
Patients with hyponatremia, hypochloremia or hypocalcemia had a shorter OS and survived less frequently one year. All results of univariate analyses are shown in Table 3. Survival analyses using Kaplan-Meier curves and the log-rank test con rmed poorer survival in GBM patients with hypochloremia (p < 0.001), hyponatremia (p = 0.026) and hypocalcemia (p < 0.001). The Kaplan-Meier curves are shown in Fig. 2A Moreover, the impact of the number of electrolyte disorders on OS was also investigated. Here, a signi cant association between the number of electrolyte disorders and OS was found (p < 0.001). This association is further supported by Kaplan-Meier analysis (log-rank test p < 0.001) as shown in Fig. 2D an urea level ≤ 20 mg/dl (p < 0.001, 7.3 vs. 10.6 months). A GFR < 60 ml/min/1.73 m 2 was also associated with a shorter median OS (p < 0.001, 6.7 vs. 10.2 months). Additionally, patients with a GFR < 60 ml/min/1.73 m 2 showed less frequently a survival of more than 1 year (p = 0.001). This association was also found in patients with a urea serum level > 20 mg/dl (p < 0.001). Altered creatinine levels were not associated with OS or 1YS.
Survival analyses applying the log-rank test and Kaplan-Meier curves revealed a signi cantly worse survival in patients with a serum urea level > 20 mg/dl (p < 0.001) and in patients with a GFR < 60 ml/min/1.73 m 2 (p < 0.001) (see Supplemental Figure S1).

Multivariate analysis
Multivariate analyses did not reveal a serum urea level > 20 mg/dl or a restricted GFR as independent predictive factors for 1-YS and OS in binary logistic and Cox regression analyses respectively (see Supplemental Table S1).

Discussion
In cancer patients, electrolyte and acid-base disorders are reported in more than 50% (8). Moreover, kidney function and electrolyte homeostasis have been linked to patients' prognosis for several types of cancer (8,10,11). This study identi ed hypochloremia in GBM as an independent prognostic marker for 1-YS and OS, whereas all other investigated electrolytes and parameters of kidney function failed to display a consistent relation to patients' outcome.
Chloride is the main anion in plasma and interstitial uid and carries a signi cant role in maintaining serum osmolarity and acid-balance. There are several possible causes for hypochloremia in patients, like chloride loss via vomiting or diarrhea, use of diuretics, or excess water gain due to infusion of hypotonic solutions. During the last years, chloride serum levels gained increasing attention in the intensive care of critically ill patients and also in cancer patients (13). In intensive care units, hypochloremia occurs in up to one-third of all individuals (14)(15)(16). This study found hypochloremia in 6.3% of patients with GBM in the preoperative routine investigation of electrolytes. This study cohort mainly comprises patients that are admitted to hospital for elective surgery of intracranial lesions, whereas the studies mentioned-above investigated critically ill patients in intensive care units. This might explain the higher rate of hypochloremia in the studies that were mainly conducted on patient collectives in intensive care units.
Several studies additionally addressed the association between hypochloremia and patients' outcome. Kimura et al. detected a signi cantly higher mortality in patients with hypochloremia after elective thoracic or abdominal surgery (16). Similar results revealed a study of 106,505 adult patients undergoing noncardiac surgery. Here, patients with preoperative hypochloremia showed a signi cant increased 90-days mortality compared to patients with normochloremia. Patients with hypochloremia also had a higher risk of postoperative acute kidney injury (17). These results are in agreement with several investigations reporting a poorer outcome for patients with hypochloremia (14,(18)(19)(20) whereas Thongprayoon et al. only found hospital-acquired hyperchloremia in 39,298 patients to be associated with increased in-hospital mortality (21). There is only sparse evidence for the in uence of hypochloremia on the outcome of cancer patients.
One recent study retrospectively collected clinical data and electrolytes parameters from all cancer patients treated over one year. In 25,881 patients, hypochloremia occurred in 24.5% of patients. The authors found a higher in-hospital mortality in patients with electrolyte disorders compared to patients with normal electrolytes (8). In another investigation among 5,089 patients with colorectal cancer, hypochloremia was associated with a worse overall survival and shorter disease-free survival (9). This study detected hypochloremia to be an independent prognostic factor for OS and 1-YS in GBM patients and is well in line with previous results. So far, there are no investigations that speci cally address hypochloremia and its prognostic impact on GBM survival. One recent study analyzed the predictive value of hyponatremia in 200 GBM patients. Similar to the results of this study, the authors could not demonstrate an association between hyponatremia and patients' outcome (22). In this cohort of GBM patients, arterial hypertension, diabetes, a higher age, and a poor preoperative clinical status characterized the cohort of patients with hypochloremia. This observation is in line with previous studies, that also reported renal dysfunction, hypertension, and diabetes as risk factors for electrolyte disorders (8).
Hypochloremia might be a surrogate marker for patients with a speci c risk pro le that is associated with a poorer outcome. Petnak et al. investigated chloride levels of adult patients at discharge and found both hypochloremia and hyperchloremia to be associated with an increased risk of one-year mortality (23). In 18,825 critically ill adult patients, a uctuation of chloride serum levels during the rst 72 hours after admission to the intensive care unit was associated with increased 30-day mortality (6). So far, there are no studies on patients with GBM that included the evaluation of hypochloremia during the course of the disease and the in uence of an early correction of hypochloremia on patients' outcome.

Limitations
Page 8/17 The limitations of this study are mainly due to its retrospective design. The interpretation of the results is limited due to in part incomplete data, that carry the risk of inaccuracy. Furthermore, treatment strategies in our cohort were heterogenous and several factors, that might in uence patients' renal function and electrolytes could not be incorporated in the analysis, e.g. medication at admission, repetitive vomiting, etc.

Conclusion
Despite the limitations mentioned above, this study is based on a large cohort of GBM patients and their corresponding serum laboratory parameters at admission. Hypochloremia was identi ed as an independent prognostic factor making serum chloride levels a promising preoperative biomarker in GBM. The results of this study will have to be con rmed in a prospective study including multiple centers. Additionally, the role of hypochloremia during the adjuvant therapies and the effect of adjustment of chloride serum levels on survival should be elucidated. The study was approved by the Institutional Ethics Committee, University of Essen (15-6504-BO and 15-6505-BO). As this is an retrospective analysis, consent on participation is not applicable.

Consent for publication
The manuscript does not contain any individual person's data.

Availability of data
The data that support the ndings of this study are available from the corresponding author upon reasonable request.

Competing interests
The authors state that there are no con icts of interest, ethical adherence or any nancial disclosures.   Prevalence of electrolyte and renal disorders in the analyzed GBM cohort. Preoperative serum lab values were available for: sodium, potassium and creatinine in 860 patients; chloride and urea in 857 patients; calcium in 717 and glomerular ltration rate (GFR) in 645 patients. Reference ranges with appropriate units are shown in white bars. The prevalence of aberrations below (light grey bars) and above (dark grey bars) the range limits is reported in percentages.

Figure 2
Survival analysis using Kaplan-Meier curves for hyponatremia (A), hypochloremia (B) and hypocalcaemia (C). Figure 2D shows survival analysis for combined electrolyte disorders.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. TableS1.docx