Plasma creatine kinase and all-cause mortality in patients on peritoneal dialysis: a multi-center retrospective study

Background: Higher plasma creatine kinase (CK) values are associated with the failure of antihypertensive treatment. However, an association between CK and all-cause mortality in peritoneal dialysis (PD) patients has received little attention. Methods: In this retrospective multicenter study, 2224 incident PD patients with baseline CK values were enrolled from November 1, 2005, to February 28, 2017. All patients with oral statins were excluded and then were divided into four groups [Quartile 1 (<60 U/L), Quartile 2 (60-100 U/L), Quartile 3 (101-179 U/L), and Quartile 4 (>179 U/L)]. The primary endpoint was all-cause mortality. The association between plasma CK values and all-cause mortality was assessed with Cox regression and the Fine and Gray models. Results: Of eligible 1382 patients, 298 (21.6%) patients died during a median 35-month (interquartile range=19-54 months) follow-up period. Patients in Quartile 4 were older (P<0.001), more likely to be male (P<0.001), had a higher prevalence of diabetes (P=0.002), and a history of cardiovascular disease (P=0.005), and higher values of Charlson comorbidity index (P=0.031). All-cause mortality incidence had a signi�cant difference among the four Quartiles (Quartile 1, 16.2%; Quartile 2, 22.2%; Quartile 3, 23.8%; Quartile 4, 24.1%; P=0.043). Quartile 4 had a higher all-cause mortality compared to other groups (Log Rank=10.55, P=0.015). After adjusting for confounding factors, the highest CK quartile had a hazard ratio (HR) for all-cause mortality of 1.72 [95% con�dence interval (CI) 1.31-3.26, P=0.042]. With kidney transplantation or hemodialysis as a competing risk, the Quartile 4 had an HR for all-cause mortality


Background
High creatine kinase (CK), as a signi cant predictor for blood pressure (BP) and the failure of antihypertensive drugs, is associated with the failure of antihypertensive treatment in the general population [1].Previous studies have shown that high CK activity promotes hypertension by enhancing vasoconstriction and renal sodium retention [2][3][4].Additionally, previous studies show that plasma CK levels are associated with decreased in ammation in obesity, whereas higher in ammation is associated with higher obesity-related cardiovascular disease (CVD) [5][6][7].Another study reported that plasma CK levels were inversely and independently associated with C-reactive protein in 454 overweight and obese individuals, supporting the anti-in ammatory effects of plasma CK levels [8].Thus, plasma CK may have bene cial and detrimental effects on prognosis in the general population.
Among peritoneal dialysis (PD) patients, chronic in ammation is a well-recognized nontraditional risk factor that contributes to excessive mortality [9].Therefore, plasma CK supporting the anti-in ammatory effects may also have a bene cial effect on the prognosis of PD patients.However, on the other hand, high plasma CK levels promote hypertension through enhanced vascular contractility and renal sodium retention [2].Thus, these ndings suggested that plasma CK may also have bene cial and detrimental effects on clinical outcomes in PD patients.Therefore, it was di cult to speculate on the association between plasma CK levels and mortality in PD patients.The aim of this study was to evaluate the association between plasma CK levels and mortality in PD patients.

Study Population and Data Collection
We retrospectively conducted a multicenter cohort study of incident patients with PD from four PD centers from November 1, 2005, to February 28, 2017.Patients lacking plasma CK at baseline, with aged < 18 years, or with PD vintage < three months were excluded from this study.We excluded those using statins because statins may increase plasma CK levels.The study was consistent with the ethical principles of the Declaration of Helsinki and was approved by the Human Ethics Committee of each research center.Written informed consent was obtained from all participants.
Baseline demographic data included age, sex, diabetes, a history of CVD, hypertension, Charlson comorbidity index (CCI), and medication use [including calcium channel blockers (CCB), angiotensinconverting enzyme inhibitors/angiotensin receptor blockers (ACEI/ARB), diuretics and β-blockers].Clinical and biochemical data at the initiation of PD included body mass index (BMI), BP, hemoglobin, albumin, total cholesterol, triglycerides, aspartate aminotransferase (AST), alanine transaminase (ALT), serum uric acid, creatinine, high-sensitivity C-reactive protein (hs-CRP), serum sodium, calcium, phosphorus, and 24hr urine output.The CK values at the initiation of PD were also collected under resting conditions.CVD was de ned as coronary events, arrhythmias, congestive heart failure, cerebrovascular events, or peripheral vascular disease [10].As one of the most commonly used comorbidity models, the comorbidity score was determined based on the CCI [11].
The primary endpoint was all-cause mortality, which was determined by the PD follow-up panels composed of PD primary nurses and professors.If death had two or more potential causes, we generally ascribed the death to the primary cause for hospitalization or the initial presenting condition.If a patient died within three months of transfer to hemodialysis therapy, he or she was not censored because the early mortality was considered to re ect health status during the period of failing PD treatment.All patients were followed up until the cessation of PD, death, or May 31, 2017.The censored data included switching to HD, kidney transplantation, moving to another center, loss to follow-up, or still at four PD centers with a follow-up duration of 8 years.All patients received continuous ambulatory PD treatment.Conventional PD solutions (Dianeal 1.5%, 2.5%, or 4.25% dextrose; Baxter Healthcare, Guangzhou, China), Y sets, and twin bag systems were used in all PD patients.

Statistical Analyses
Eligible patients were strati ed into four groups: Quartile 1<60 U/L, Quartile 2=60-100 U/L, Quartile 3=101-179 U/L, and Quartile 4>179 U/L.Continuous variables were presented as mean ± standard deviation or median and categorical variables as frequency and percentage.Comparisons of variables among groups were undergone using the Chi-squared test, one-way ANOVA, or the Kruskal-Wallis test.The correlations between plasma CK levels and baseline variables were assessed by correlation analysis.Survival times were estimated from Kaplan-Meier curves, and differences in survival probability among groups were assessed using a Log Rank test.The association between CK Quartiles and all-cause mortality was evaluated by the Cox regression model.Unadjusted associations (model 1) were rst examined, followed by adjustments for age, sex, CCI, and the use of medication (model 2).Next, systolic and diastolic BP, hemoglobin, albumin, creatinine, sodium, calcium, phosphorus, and 24-hour urine output were added into model 2 (model 3).

Sensitivity Analyses
The Fine and Gray competing risk model was performed with the covariates included in the Cox regression model, and kidney transplantation or transfer to hemodialysis was considered as the competing event.
The results of the Cox analysis and the Fine and Gray method were presented as the hazard ratio (HR) and the 95% con dence interval (CI).A value of P<0.05 was considered statistically signi cant.Statistical analyses were performed using GraphPad software 8.0 (GraphPad Prism Software Inc., San Diego, California) and the R package (https://www.r-project.org/).

Baseline Patient Characteristics
Of 2224 incident PD patients, 10 patients were younger than 18 years, 84 patients were on PD less than three months, 196 patients were using statins, and 552 patients lacked plasma CK levels at baseline were excluded, with 1382 patients eligible for the present analysis (Figure 1).Of study patients with 51.1±14.8years, 56.9% were male, 23.4% had diabetic, 16.4% had a history of CVD, and 71.9% had hypertension.In addition, 72.9% taken CCB, 31.6% taken ACEI/ARB, 8.2% taken diuretic, and 31.9%taken β-blocker.
Baseline CK levels ranged from 8 to 13585 U/L (median=100 U/L, interquartile range=60-179 U/L).The baseline variables strati ed by Quartiles were shown in Table 1.Patients in Quartile 4 were older (P<0.001),more likely to be male (P<0.001),had a higher prevalence of diabetes (P=0.002), and a history of cardiovascular disease (P=0.005), and higher values of CCI (P=0.031).
Kaplan-Meier estimates of all-cause mortality among different Quartiles were shown (Figure 2).Patients with higher CK Quartiles had a higher cumulative all-cause mortality compared to those lower Quartiles (Log Rank=11.13,P=0.012).Similar results were observed in the male and non-diabetic patients (Log Rank=9.63,P=0.045; Log Rank=12.61,P=0.006).Adjusted HR for all-cause mortality in different Quartiles was shown in Table 2. Quartile 4 was independently associated with all-cause mortality, even after adjusting for demographics, comorbid conditions, and laboratory variables (Quartile 1 as a reference, HR=1.72, 95% CI 1.31-3.26,P=0.042).Subgroups analyzed showed that the Quartile 4 had an independently higher risk of all-cause mortality in males (HR=1.44,95%CI 1.13-2.49,P=0.023) and nondiabetes (HR=1.61,95%CI 1.43-2.52,P=0.019), after adjusting for confounding factors.

Sensitivity Analyses
With renal transplantation or hemodialysis as a competing risk event, the Quartile 4 had an adjusted HR for all-cause mortality of 1.64 (95%CI 1.25-3.48,P=0.046, Table 2) compared to the Quartile 1, after adjusting for confounding factors.Figure 3 showed that with renal transplantation or hemodialysis as a competing risk event, the Quartile 4 had independently higher risk of all-cause mortality in male (HR=1.36,95%CI 1.08-2.55,P=0.029) and non-diabetes (HR=1.53,95%CI 1.40-2.60,P=0.026) group, after adjusting for confounding factors.

Discussion
In this multicenter retrospective study, we found that even after adjustment for baseline variables, higher plasma CK levels were incrementally associated with a higher risk of all-cause mortality in PD patients, especially in those male and non-diabetes PD patients.
Plasma CK tightly binds to ATP-utilizing enzymes, including Ca 2+ -ATPase, myosin ATPase, and Na + /K + -ATPase, to rapidly regenerate ATP from ADP, H + , and phosphocreatine [12].Usually, the release of CK from tissues is proportional to the intracellular CK concentration, a physiological process that occurs without tissue damage, as summarized by Brewster [3].Therefore, the plasma CK of healthy people at rest re ects the tissue CK concentration [3,4,13].However, as lymphatic ow increases with exercise, CK from the interstitial space may enter the circulation rather abruptly and be cleared by the liver in approximately three days [3].An elevation of plasma CK levels is seen following acute myocardial infarction, rhabdomyolysis, intramuscular injections, and strenuous physical activity.A high tissue CK level is thought to result in a phenotype with greater vasoconstriction and enhanced sodium retention by greater ATP-buffering capacity of ATPases involved in ion transport and contractile responses [2,3,12,14].
Previous studies found that relatively high CK is thought to enhance ATP-demanding processes, including resistance to arterial contractility and sodium retention, and to reduce ADP-dependent function in the general population [12, 15][12, 15].Subsequently, in a randomized sample of a multiethnic population in Amsterdam, the Netherlands, CK proved to be an important independent predictor of BP levels and the failure of antihypertensive treatment [3].This study found that after adjusting for age, gender, BMI, and ethnicity, CK was independently associated with BP and with systolic and diastolic BP, increasing by 8.0 and 4.7 mmHg/log CK, respectively.Since then, several other studies have reported that plasma CK levels are associated with the failure of antihypertensive drug treatment [4, 15][4, 15].In addition, plasma CK has been associated with decreased in ammation in obesity, while in ammation is associated with obesity-related CVD [14].Notably, plasma CK along with lean body mass is inversely and independently associated with hs-CRP in overweight and obese individuals, supporting the anti-in ammatory effects of CK.Thus, these ndings suggest that CK may also have bene cial and detrimental effects on the prognosis of patients.Therefore, it is di cult to speculate on the association between CK and death in PD patients.To date, little is known about the association between plasma CK levels and all-cause mortality in PD patients.Our study showed that high plasma CK Quartiles were associated with higher risk of all-cause mortality in PD patients, independent of confounding factors, such as age, sex, CCI, CCB use, ACEIs/ARBs use, β-blocker use and diuretic use.Meanwhile, we found that the results from the competing risk model were consistent with the Cox regression models.Additionally, similar results were also observed in the male and those without diabetes.These ndings suggested that monitoring plasma CK levels may be bene cial for improving the prognosis of PD patients, especially in the male and those without diabetes.A possible explanation for the described association between total plasma CK levels and mortality was that plasma CK was associated with the failure of antihypertensive therapy, which has been linked to higher mortality [16].There is increasing evidence that high plasma CK levels are thought to enhance ATP-demanding processes, including resistance to arterial contractility and sodium retention.In our study, higher plasma CK levels were associated with a higher risk of all-cause death.Our results showed that there was a signi cantly positive relationship between the levels of plasma CK and BP.It is well known that high BP is associated with high all-cause mortality [16,17].Although the lowest Quartile of CK (<60 U/L) was associated with a lower risk of all-cause death, the mean hemoglobin and albumin levels in the Quartile 1 were signi cantly lower than those in the Quartile 4. It is well known that lower hemoglobin and albumin, as markers of malnutrition, are associated with a higher risk of all-cause mortality [18].
Additionally, lower plasma CK may be considered a marker of malnutrition and relatively low muscle mass, which has been linked to a higher risk of all-cause mortality [17,19].Thus, these ndings suggested that the adverse effect of the failure of antihypertensive treatment of higher CK levels on allcause mortality may be stronger than the protective effect of anti-in ammatory of higher CK levels on allcause mortality, which may lead to increased all-cause mortality in the present study.Therefore, mechanisms of the effect of plasma CK on mortality should be investigated in future studies, and the management of plasma CK may improve the clinical prognosis of PD patients.
The strengths of this study are the multicenter nature of the study, a large number of patients, the ability to adjust for signi cant risk factors for all-cause mortality, and sensitive analysis of competing risk model.There are some limitations in the present study.First, due to the multicenter design of the cohort, there were some variations in the ascertainment and validation of the endpoint.However, it was a multicenter study; therefore, center-speci c effects may be excluded.Second, the retrospective nature of the study allows us to establish associations but not causal relationships.As with all retrospective studies, a potential limitation is that the associations may be in uenced by confounding by other risk factors.Because of the restriction of sample size, we did not adjust for all factors associated with higher mortality.Therefore, the effect of residual confounding cannot be eliminated completely.Due to the retrospective data, we are not capable of getting precise home BP data because of patient's bias.We will prospectively observe the longitudinal association between plasma CK levels and BP of these patients.Third, guidelines of dialysis do not regularly recommend lipid-regulating treatment of hyperlipidemia and CVD events in PD patients.Thus, even though PD patients developed CVD events, we usually do not prescribe statins for these patients.This was why patients with statins accounted for < 10% of total PD patients.Nonetheless, the ndings are mainly applicable in statin naïve subjects.Future study should focus on the relationship between CK levels and statins in PD patients.Fourth, whether plasma CK can be cleared via PD remains unknown.Thus, whether the patients with higher CK had greater clearance needs that were not met by the CAPD remains unknown.We will further observe the association between dialysis dose and plasma CK levels.Lastly, the changes in variables and treatments overtime during the follow-up were not included.

Conclusions
In conclusion, we found an independent association between higher plasma CK levels at the commencement of PD and a higher risk of all-cause mortality in PD patients, especially in males and those without diabetes.These ndings suggest that monitoring plasma CK levels may be bene cial for improving the prognosis of PD patients, and clinicians could use plasma CK as a valuable biomarker for mortality in PD patients.

Declarations
Ethics approval and consent to participate: The study was consistent with the ethical principles of the Declaration of Helsinki and was approved by the Human Ethics Committee of the Second A liated Hospital of Guangzhou Medical University, Zhujiang Hospital of Southern Medical University, Jiujiang No. 1 People's Hospital, A liated Sixth People's Hospital, Shanghai Jiao Tong University, and the First A liated Hospital of Nanchang University.Written informed consent was obtained from all participants.
Consent for Publication: All authors have approved the submitted version.All authors have agreed both to be personally accountable for the author's own contributions and to ensure that questions related to the accuracy or integrity of any part of the work, even ones in which the author was not personally involved, are appropriately investigated, resolved, and the resolution documented in the literature.

Figures
Figure 1 The ow chart showed how patients were selected for the present study.CK, creatine kinase.
Adjusted hazards ratio for all-cause mortality.Hazard ratios for Quartile 4 relative to Quartile 1 (adjusted in multivariable models).CVD, cardiovascular disease.

Table 2 .
Adjusted hazards ratio for all-cause mortality in different CK quartiles.kidney transplantation or transfer to hemodialysis as a competing risk.Model 1: unadjusted.Model 2: adjusted for age, sex, CCI, and use of medication.Model 3: model 2 adjusted for systolic and diastolic BP, hemoglobin, albumin, creatinine, sodium, calcium, phosphorus and 24-hr urine output.CK, creatine kinase; CCI, Charlson comorbidity index. *