Inverse association of transferrin saturation with mortality risk in chronic kidney disease

Background: Transferrin saturation (TSAT) is an indicator of iron deficiency or overload, but its relationship with mortality in patients with different stages of chronic kidney disease (CKD) is unclear. We investigated the association of TSAT with mortality in CKD patients. Methods: In 479 CKD patients (97 CKD3-4 patients, 298 CKD5 non-dialysis patients and 84 peritoneal dialysis patients; median age 58 years, 67% males, 33% cardiovascular disease, CVD, and 29% diabetes), biomarkers of iron status (plasma iron, TSAT, transferrin and ferritin), systemic inflammation (high sensitivity C-reactive protein, hsCRP, and interleukin-6, IL-6) and nutritional status were assessed. During median follow-up of 35.6 months, 139 (29%) patients died, and 176 (37%) patients underwent renal transplantation. Patients were stratified into Low (n=157) and Middle and high (n=322) TSAT tertile groups. All-cause and CVD mortality risk were analyzed by competing risk regression with renal transplantation as competing risk. Results: TSAT (median 23%; interquartile range, 17-30%) was negatively associated with presence of diabetes and CVD, body mass index, hsCRP, IL-6, erythropoiesis stimulating agent (ESA) dose, erythropoietin resistance index (ERI) and iron supplementation, and positively associated with hemoglobin, ferritin and s-albumin. In competing risk analysis, low tertile of TSAT was independently associated with increased all-cause mortality risk (sHR=1.74, 95%CI 1.30-2.54) and CVD mortality risk (sHR=1.80, 95%CI 1.02-3.16) after fully adjusting for 1-standard deviation (SD) of age, sex, CKD stages, 1-SD of hemoglobin, 1-SD of ferritin, 1-SD of hsCRP, 1-SD of ESA dose and iron supplementation. Conclusions: Lower TSAT indicating iron deficiency was independently associated with increased mortality risk in CKD patients, underlining that iron status should be considered when evaluating clinical outcomes of CKD patients. by both TSAT and serum ferritin we performed a sensitivity analysis with cutoffs for serum ferritin of < 100 µg/L,< 300 µg/L, < 500 µg/L and ≥ 500 µg/L which demonstrated that a low level of TSAT remained significantly associated with all-cause mortality when applying most of the applied ferritin cutoffs besides serum ferritin ≥ 500 µg/L. These results are in line with results from a study by Eisenga et al. [8] which demonstrated that a low TSAT (< 10%) remained a predictor of mortality when serum ferritin was below 500 µg/L. serum ferritin ≥ µg/L, it may reflect iron overload. An


Introduction
Anemia is common in patients with chronic kidney disease (CKD) and contributes to increased morbidity and mortality [1,2]. Apart from inadequate erythropoietin secretion, a leading cause of anemia in CKD is impaired iron homeostasis [3], commonly presenting as either absolute or functional iron deficiency, affecting 24-85% of patients with CKD [4].
Iron deficiency as such has been shown to be associated with increased risk of mortality in some chronic conditions, including anemia, independently of potential confounders [5,6]. But it is still unclear if iron dysmetabolism, especially iron deficiency alone, could directly impact the clinical outcome in patients with CKD.
Transferrin saturation (TSAT), an index that considers both plasma iron and its main transport protein transferrin, serves as a biochemical marker of iron availability in plasma. Together with plasma ferritin that reflects body iron stores, TSAT is routinely used to monitor iron therapy [7]. In patients with adequate or excessive iron stores as reflected by elevated ferritin levels, a low TSAT is commonly associated with inflammation and hepcidin-induced functional iron deficiency.
Several studies have investigated the association between iron dysregulation and all-cause and cardiovascular mortality using TSAT values in CKD patients [8][9][10]. These studies indicated that low TSAT values are associated with higher mortality in CKD patients, independently of anemia. The level of TSAT that predicts an adverse outcome varied between studies, and therefore the optimal range remains to be defined. Consequently, the prognostic importance of TSAT for all-cause and cardiovascular mortality in patients with various stages of CKD needs to be evaluated further.
The purpose of our study was to evaluate clinical characteristics associated with TSAT levels, factors influencing TSAT in different CKD stages, and the independent potential of TSAT to predict all-cause and CVD mortality in patients with CKD.

Materials And Methods Patients and study design
This study was conducted in 479 clinically stable CKD patients, including 97 CKD3-4 patients, 298 CKD5 non-dialysis (CKD5-ND) patients and 84 CKD5-peritoneal dialysis (CKD5-D) patients. The patients (age 19-87 years) were recruited from three cohorts briefly described below. All patients were followed until renal transplantation or death, or after finishing 60 months of follow-up. There were no patients lost for follow-up.
Patients' characteristics at baseline in these three cohorts are displayed in Table S1. The serum transferrin saturation (TSAT) percentage was calculated by dividing the serum iron level by total ironbinding capacity*100. Patients were divided into two groups based on the tertiles of TSAT: Low tertile group (n = 157), TSAT ranged from (min-max) 5 Anthropometric measurements were obtained at baseline. Body mass index (BMI) was calculated.
Handgrip strength (HGS) was quantified in both hands using a Harpenden Handgrip Dynamometer (Yamar, Jackson, MI, USA). Values for HGS were expressed as percentage of healthy subjects, adjusted for sex. Lean body mass index (LBMI) and fat body mass index (FBMI) were calculated according to the method of Kyle et al. [15] from anthropometric data using four measurements of skinfold thickness [16], and were expressed as kg/m 2 . Nutritional status was evaluated using the subjective global assessment (SGA) questionnaire [17]. Protein energy wasting (PEW) was defined as SGA score > 1 while a score of 0 indicated normal nutritional status. Blood pressure was reported as mean arterial blood pressure (MAP) defined as [diastolic pressure + (systolic pressure-diastolic The erythropoietin resistance index (ERI) was calculated by dividing the weekly weight adjusted epoetin dose by the hemoglobin level at the time of investigation [18].

Biochemical Assessments
After overnight fasting, venous blood samples were collected at baseline. Concentrations of hemoglobin, high-sensitivity C-reactive protein (hsCRP), leucocyte count, serum creatinine, iron, ferritin, transferrin, serum albumin (bromocresol purple), cholesterol, triglycerides, intact parathyroid hormone (iPTH), calcium and phosphate were determined by routine methods at the Department of Laboratory Medicine, Karolinska University Hospital. Interleukin-6 (IL-6) and plasma tumor necrosis factor (TNF) were tested by enzyme-labeled chemiluminescent assay (Immulite, DPC, Los Angeles, CA) at our research laboratory.

Statistical Analyses
Data were expressed as median with interquartile range (IQR) or percentage or sub-distribution hazard ratio (sHR) or crude mortality rate, as appropriate. All tests were two-sided, and statistical significance was set at the level of P < 0.05. Comparisons between two groups were assessed with the non-parametric Wilcoxon test for continuous variables and Chi-square test for nominal variables.
Comparisons between three or more groups were assessed with the non-parametric Kruskal-Wallis test for continuous variables and Chi-square test for nominal variables. Non-parametric Spearman rank test was performed to determine associations between variables. Multivariate logistic regression was used to analyze the factors associated with tertiles of TSAT. Survival during follow-up was analyzed by the competing risk regression model and the cumulative incidence curve [20]. Fine & Gray models were used [21] and adjusted for age, sex, CKD stages, hemoglobin, ferritin, hsCRP, ESA dose and iron supplementation. The sHR for TSAT were calculated with renal transplantation as a competing risk. We also performed restricted cubic spline in the flexible regression model to examine the non-linear association of TSAT level with all-cause mortality [22]. We used four knots to make a smooth curve. Spline curve was expressed as sub-hazard ratio for all-cause mortality, adjusting for age, sex, CKD stages, hemoglobin, ferritin, hsCRP, ESA dose and iron supplementation.
In sensitivity analyses, we used Cox proportional hazard model to further examine the association between lower TSAT and mortality across different subgroups of patients. All statistical analyses were

Univariate Correlations Of Factors Associated With TSAT
Univariate associations between TSAT and other parameters are shown in Table S2. TSAT was negatively associated with DM, CVD, BMI, mean arterial blood pressure, phosphate, leucocyte count, hsCRP, IL-6, use of ESA, ESA dose, iron supplementation and intravenous iron therapy, and positively associated with hemoglobin, ferritin, s-albumin and cholesterol. A negative relationship was found between ERI and TSAT levels (rho=-0.16, P = 0.004, shown in Figure S1). were found to be independently associated with TSAT after adjustments for diabetes, history of CVD and iron supplementation. person-years (shown in Figure S2A). Crude CVD mortality rate/1000 person-years was for the low tertile of TSAT 5.6 (95% CI, 3.9-8.1), middle tertile 2.8 (95% CI, 1.8-4.5), and high tertile 2.4 (95% CI, Figure S2B).

Associations Between TSAT Levels And All-cause Mortality
During follow-up for a median of 35.6 months, 139 (29%) of the patients died and 176 (37%) patients underwent renal transplantation. Multivariate competing risk analysis for all-cause mortality and CVD mortality took renal transplantation into account.
In multivariate competing risk analysis for all-cause mortality, after adjusting for 1-SD of age, sex, CKD stages, 1-SD of hemoglobin, 1-SD of ferritin, 1-SD of hsCRP, 1-SD of ESA dose and iron supplementation, lower TSAT (low tertile of TSAT) was significantly associated with higher all-cause mortality risk when compared to higher TSAT (middle and high tertile of TSAT): sHR, 1.74; 95% CI, 1.20-2.51; P = 0.003 ( Fig. 2A). We performed competing risk analysis for patients with ESA therapy.
After adjustments for the same confounders, the results showed that patients with lower TSAT levels had significantly higher risk of all-cause mortality (sHR, 1.74; 95% CI, 1.16-2.60; P = 0.007; n = 351).
In spline curve analysis using a flexible parametric model of TSAT as a continuous variable in 479 patients, TSAT showed an inverse association with all-cause mortality after adjusting for age, sex, CKD stages, hemoglobin, ferritin, hsCRP, ESA dose and iron supplementation ( Figure S3).
Sensitivity analyses showed that lower TSAT was associated with higher all-cause mortality across different subgroups of age, sex, anemia, inflammation, and dialysis therapy but not in those with ferritin > 500 µg/L, SGA = 0 (i.e., well-nourished) and without ESA therapy (Fig. 3).

Discussion
This study shows that TSAT is inversely associated with mortality risk in CKD patients indicating that iron deficiency contributes to poor clinical outcomes. Iron deficiency exerts adverse effects on the cardiovascular system and has been demonstrated to negatively affect cardiac performance [23][24][25][26].
Previous studies have shown that iron deficiency increases mortality risk, independently of anemia in both heart failure [24,25] and in CKD populations [9]. Our observation that low TSAT associated with increased mortality risk irrespective of hemoglobin status corroborates these findings. The importance of iron status is supported by a randomized clinical trial among 2141 hemodialysis patients showing that high-dose iron therapy administered proactively was superior to a low-dose iron regimen and associated with significantly lower risk of death or nonfatal adverse cardiovascular events [27]. While details for iron therapy were not available, our study suggests that iron status as reflected by TSAT should be considered when evaluating risk factors in various CKD stages, with or without anemia. Whether iron deficiency with or without anemia in CKD should be treated as a distinct entity to improve survival and reduce cardiovascular morbidity deserves further studies.
Since treatment of iron deficiency is commonly guided by both TSAT and serum ferritin levels, we performed a sensitivity analysis with cutoffs for serum ferritin of < 100 µg/L,< 300 µg/L, < 500 µg/L and ≥ 500 µg/L which demonstrated that a low level of TSAT remained significantly associated with all-cause mortality when applying most of the applied ferritin cutoffs besides serum ferritin ≥ 500 µg/L. These results are in line with results from a study by Eisenga et al. [8] which demonstrated that a low TSAT (< 10%) remained a predictor of mortality when serum ferritin was below 500 µg/L.
In our study, 73% patients were on ESA treatment, and TSAT was negatively correlated with ESA dose and ERI. The inverse relation between TSAT and ERI was also presented in a prospective and observational study including 1710 hemodialysis patients [28]. CKD is largely considered as an inflammatory state and inflammation is one of the major causes of ESA resistance, as indicated by independent association of ERI with hsCRP [29]. Systemic inflammation -and CKD -associate with increased concentrations of hepcidin [30]. Raised hepcidin levels further impair the ferroportinmediated release of iron from enterocytes and the reticuloendothelial system [31], leading to high tissue ferritin but inadequate circulating iron available for erythropoiesis. In CKD patients, hyporesponsiveness to ESA can to some extent be explained by functional iron deficiency. A low TSAT level, which reflects either insufficient iron mobilization or depleted iron stores, was observed to be related to higher hsCRP level and ESA dose in our study, indicating interactions between inflammation, iron dysregulation and hypo-responsiveness to ESA.
The interpretation of iron biomarkers is affected by inflammation, a hallmark feature and a wellestablished mortality risk factor in CKD, as inflammation directly affects the levels of most iron biomarkers [32], including TSAT. Inflammation also accelerates the process of atherosclerosis, vascular calcification, and other causes of CVD [33,34]. Thus, when assessing iron biomarkers as risk factors, inflammation is a considerable confounder. In our study, even though hsCRP was strongly associated with TSAT, lower TSAT was still independently associated with both all-cause and CVD mortality risk after adjusting for inflammation. Previous findings suggest as well that mortality risk associated with iron deficiency is independent from inflammation [8,9].
In the present study, 29% patients were diabetic and had significantly lower serum TSAT levels. As presence of diabetes and its related metabolic syndrome are pro-inflammatory conditions, TSAT level may decrease due to a persistent microinflammatory state [35], and this may have contributed to the negative association between TSAT and diabetes Several previous studies have focused on the relationship between iron indicators, especially TSAT, and mortality in CKD and dialysis patients [8][9][10][36][37][38]. In one study including 975 CKD patients, TSAT < 10% was strongly associated with the risk of adverse outcomes [8]. In a pre-dialysis cohort (n = 32,489) the highest risk of mortality was seen among those with the lowest quantile of TSAT (range = 0.36-11%) [9]. In incident dialysis patients with anemia, TSAT ≤ 20% was a significant independent risk factor for adverse clinical outcome [37]. In the present study, low tertile (range = 5.1%-19.2%) was associated with higher risk of both all-cause as well as CVD mortality. So far, the definition of a specific accurate and reliable lower threshold of TSAT that would be useful for predicting adverse clinical outcomes has not been agreed upon; however, as concluded present study and several previous studies there is a robust inverse relationship between TSAT levels and mortality.
As expected, most anemia-related parameters differed between the different stages of CKD. As compared to patients with CKD stage 3-4, the CKD5-ND group showed marked impairment of most parameters, reflecting most likely the decline of renal function and complications linked to this such as inflammation, poor nutritional status and an increased burden of comorbidities. However, most anemia-related parameters improved in CKD5-D patients, such as hemoglobin, iron, ferritin, TSAT and ERI, suggesting that anemia treatment, dialysis therapy per se or factors related to dialysis initiation had a positive effect.
The present study has some limitations that should be considered when interpreting the results. First, this study was observational in nature and cannot prove causality. Second, only baseline demographic and clinical characteristics were used for the analyses; we did not analyze the impact of changes over time of TSAT and other iron parameters on clinical outcomes. Hepcidin levels were not measured in the present cohort. Third, covariates including peptic ulcer disease, gastrointestinal bleed or blood transfusions, which may influence anemia and iron markers were not considered in present study. Fourth, we included different groups of CKD patients. However, the impact of the heterogeneity between the groups was reduced by adjusting for CKD stage in the statistical analyses, and despite inclusion of patients with different stages of CKD, we could demonstrate an independent effect of TSAT on mortality, which may reinforce the conclusion of the study. Strengths of the study include detailed comprehensive phenotyping and that all patients were followed up in a uniform way, with no patients lost to follow up.
In conclusion, low TSAT was inversely and independently associated with increased mortality in CKD patients emphasizing the importance of TSAT when assessing iron status in CKD. Further studies are warranted to elucidate whether serial monitoring, rather than a single measurement of TSAT, can add value to TSAT as a prognosticator of clinical outcomes in CKD.    Adjusted log-HRs of all-cause mortality associated with lower TSAT across subgroups of age, sex, anemia, inflammation, ferritin, nutritional status, ESA therapy and dialysis therapy.

List Of Abbreviations
Models were adjusted for 1-SD of age, sex, CKD stages, 1-SD of hemoglobin, 1-SD of hsCRP, 1-SD of ferritin, ESA therapy and iron supplementation.