Effects of hemodialysis on blood fatty acids

Abstract Omega‐3 (n‐3) fatty acids have beneficial cardiovascular effects, perhaps also in chronic kidney disease (CKD) patients. A low omega‐3 index is an independent cardiovascular risk factor in end‐stage renal disease (ESRD) dialysis patients. However, the plasma measurements invariably ignore circulating blood cells, including the preponderant erythrocytes (RBCs). We measured fatty acids (HPLC‐MS lipidomics) in all components of the circulating blood, since RBC n‐3 fatty acid status has been linked to cardiovascular disease and mortality. We studied 15 healthy persons and 15 CKD patients undergoing regular hemodialysis treatments. While total fatty acid levels differed significantly in RBCs from healthy controls and CKD patients, the hemodialysis treatment had no effect on plasma or RBC fatty acid levels. No changes occurred in the percentage of eicosapentaenoic acid (C20:5 n‐3, EPA) and docosahexaenoic acid (C22:6 n‐3; DHA) (omega‐3 quotient) in RBC membrane fatty acids. Nonetheless, hemodialysis treatments increased plasma levels of various total fatty acids, namely C12:0, C14:0, C16:0, C20:2 n‐6, C20:4 n‐6, and C22:6 n‐3 (DHA), while plasma levels of free fatty acids were unchanged. These data suggest that despite significant changes in fatty acids signatures between healthy persons and CKD patients, hemodialysis does not alter RBC n‐3 fatty acid status, including the omega‐3 quotient. The dialysis treatment per se does not appear to be responsible for a lower omega‐3 index in CKD patients.

The two most important n-3 polyunsaturated fatty acids (PUFA) are eicosapentaenoic acid (C20:5 n-3, EPA) and docosahexaenoic acid (C22:6 n-3, DHA). However, the impact of the individual fatty acids for the predicting risk is not clearly elucidated. In recent randomized, double-blind, placebo-controlled trials, dietary n-3 fatty acid supplementation (3-6 g daily) improved coronary atherosclerosis (Schacky, Angerer, Kothny, Theisen, & Mudra, 1999), but had (1 g daily) no cardiovascular benefit in initially healthy adults or in diabetic patients (ASCEND Study Collaborative Group et al., 2018;Manson et al., 2019). Two more recent large-scale trials showed that dietary EPA (C20:5 n-3) (4 g daily) is effective for prevention of major coronary events in hypercholesterolemic patients (Yokoyama et al., 2007) and reduces cardiovascular events in patients with established CVD, including in those with diabetes mellitus and other risk factors (Bhatt et al., 2018). Current evidence suggests that circulating n-3 PUFA is also associated with lower risk of cardiovascular events and mortality in ESRD patients undergoing regular hemodialysis treatment (Chowdhury et al., 2014;Khor et al., 2018).

| METHODS
The Charité University Medicine institutional review board on the use of humans in research approved the study, and written informed consent was obtained. The study was duly registered: ClinicalTrials.gov, Identifier: NCT03857984; F I G U R E 1 Schematic illustration of the hypothetic influence of hemodialysis associated with blood passing through the dialyzer, oxidative stress, uremia and red blood cell (RBC)-endothelial interactions affecting relative EPA + DHA content in RBC, that is, omega-3 quotient, which is the percentage of EPA + DHA in red cell fatty acid lipids institutional review board, EA2/113/08. Recruitment was primarily via person-to-person interview. Prior to participation in the study, 15 healthy volunteers (6 male and 9 female) and 15 CKD patients (7 male and 8 female) undergoing regular hemodialysis treatment signed informed consent forms which outlined the treatments to be taken and the possible risks involved (Table 1). We included patients who had a history of renal failure and requiring hemodialysis/hemofiltration, aged ≥18 years, and were able to consent in writing. Exclusion criteria for healthy control group were age <18 years, chronic illness requiring specific medications (such as corticosteroids, for instance), pregnancy, inability to follow simple instructions, or debilitating medical conditions. No healthy control subjects were taking medications. The patients included in the group CKD were diagnosed for the following conditions: diabetes mellitus (4 patients), hypertension (3 patients), membranous glomerulonephritis (2 patients), ADPKD (autosomal-dominant polycystic kidney disease) (1 patient), other, or unknown (5 patients).

Red blood cell (RBC) Hemodialysis
Venous blood was collected in each healthy subject by subcutaneous arm vein puncture in the sitting position. In the group of dialyzed patients (CKD group), all the blood samples were collected on the fistula arm right before beginning of the dialysis (starting of the HD, pre-HD) and at the end of the dialysis (5-15 min before termination, post-HD). Patients underwent thrice-weekly dialysis, which lasted from 3 hr 45 min to 5 hr, based on high flux AK 200 dialyzers (Gambro GmbH). All samples were analyzed for RBC fatty acids status and plasma fatty acids. RBCs were separated from EDTA blood by centrifugation, and fatty acids in RBCs or plasma were determined by high-performance liquid chromatography-mass spectrometry (HPLC-MS) described in Fischer et al., (2014); Gollasch, Dogan, Rothe, Gollasch, and Luft (2019).
We performed sample size calculation for a difference in means in omega-3 quotients. We found that our study would require a sample size of 9 (number of pairs) to achieve a power of 80% and a level of significance of 5% (two sided), for detecting a mean of the differences of 2.1 (Harris, Del Gobbo, and Tintle (2017)) between pairs, assuming the standard deviation of the differences to be 1.7 (Fischer et al., 2014;Gollasch et al., 2019). A 2.1% increase in omega-3 index is associated with a 15% risk reduction for fatal coronary heart disease relative to the mean level . Thus, in our sample size calculation, statistical significance and clinical relevance both were taken into account.
Descriptive statistics were calculated, and variables were examined for meeting assumptions of normal distribution without skewness and kurtosis. We used the Shapiro-Wilk test to determine whether the data were normally distributed. In order to determine statistical significance, two-tailed t test or Mann-Whitney test were used to compare values of CKD versus control groups. Homogeneity of variances was asserted using Levene's test. Paired t test or paired Wilcoxon test was used to compare pre-HD versus post-HD values. The .05 level of significance (p) was chosen. All data are presented as mean ± SD. All statistical analyses were performed using SPSS Statistics software (IBM Corporation) or All-Therapy statistics beta (AICBT Ltd).

| Clinical characteristics
The clinical and demographic characteristics of the patients and control subjects (Table 1) show that age between ESRD patients and the healthy subjects was not different. The body mass indices between the two groups were also not different. The ESRD patients had diabetes mellitus, hypertension, membranous glomerulonephritis, autosomal-dominant polycystic kidney disease (ADPKD), and other or unknown disease, as underlying causes. Major cardiovascular complications

PAD (n) 3
Note: Data are presented as mean ± SD or frequencies.
T A B L E 2 Comparison of blood fatty acids between control subjects versus hemodialysis (HD) patients before hemodialysis (n = 15 each) including cardiovascular and cerebrovascular events, and peripheral artery disease are given. Subjects were Caucasians, with the exception of one Black and one Asian subject in each group.
T A B L E 2 (Continued) of free fatty acids within RBC showed increased values for C20:5 n-3 (EPA) + C22:6 n-3 (DHA) and increased total free fatty acids in ESRD patients, compared to controls (Table  3B). The omega-3 quotient of free fatty acids in erythrocytes of CKD patients did not differ from the quotient measured in control subjects, despite higher free C20:5 n-3 (EPA) plus C22:6 n-3 (DHA) levels and lower total levels of free fatty acids (C12:0-C24:1 n-9) in the RBCs of CKD patients.
We also inspected the effects of hemodialysis on omega-3 quotients (Table 5). C20:5 n-3 (EPA) + C22:6 n-3 (DHA) values in RBCs have not increased (p = .053), while no effects were also observed on total fatty acids or [C20:5 n-3 (EPA) + C22:6 n-3 (DHA)]/ total fatty acids in RBCs (Table  5A). Similarly, no effects were observed on the sum of RBC free fatty acids or omega-3 quotient of free fatty acids in RBCs (Table 5B). Thus, no fatty acid-level variations were found in RBCs in response to hemodialysis. Furthermore, no changes occurred also in the RBC omega-3 quotient in response to hemodialysis. Together, the findings indicate that ESRD is associated with an altered RBC fatty acid status, that is, individual signature, which has no effect on the erythrocyte n-3 fatty acid quotient. Moreover, hemodialysis treatment is insufficient to change the RBC fatty acid signature of ESRD patients.

| DISCUSSION
We evaluated blood fatty acids in normal controls and ESRD patients. Also, we studied the effects of the hemodialysis treatment, which could be deleterious. Unique is the fact that T A B L E 3 Omega-3 quotients in control subjects versus hemodialysis (HD) patients before hemodialysis (n = 15 each)

T A B L E 5 Effects of hemodialysis on omega-3 quotients (n = 15 each)
A. Omega-3 quotient of RBC total fatty acids in the CKD patients before (pre-HD) hemodialysis and at cessation (post-HD) of hemodialysis.

B. Omega-3 quotient of plasma free fatty acids of the CKD patients before (pre-HD) hemodialysis (HI) and at cessation (post-HD) of hemodialysis (HII)
Fatty acid (µg/ml)

T A B L E 4 (Continued)
we included the erythron in the analysis, rather than merely plasma values. The issue is important since the RBC mass (>40% of the circulating blood) is important to n-3 homeostasis and metabolism. Since ESRD patients die on dialysis within 5 years, the hypothesis that the treatment (in-and-ofitself) could be injurious seems reasonable. Our study investigated effects on n-3 fatty acids, and we encompassed all of the components in the circulating blood.
To our knowledge, our study is the first study to assess the impact of single hemodialysis treatment on individual RBC fatty acids using large-scale lipidomics. Although we did not confirm our hypothesis that RBC fatty acids, including the omega-3 quotient, vary during hemodialysis, we observed significant differences in RBC fatty acid status, that is, specific fatty acid signatures, between ESRD patients and control subjects. However, the omega-3 quotient did not vary between CKD patients and healthy volunteers. Finally, hemodialysis treatment did not induce increased mobilization of individual free fatty acids into plasma or erythrocytes, but caused greater rate of oxidation of total fatty acids, which accumulated in the circulating blood during hemodialysis.
Dietary omega-3 fatty acids modulate the profile of eicosanoids in humans primarily via the cytochrome P450 (CYP)epoxygenase pathway, which could mediate cardioprotective and vasodilatory effects of n-3 fatty acids (Fischer et al., 2014). Recent results demonstrate that CYP enzymes efficiently convert C20:5 n-3 (EPA) and C22:6 n-3 (DHA) to bioactive epoxy and hydroxy metabolites that could mediate some of the beneficial cardiovascular effects of dietary n-3 fatty acids (Arnold et al., 2010). Thus, pharmacological interventions targeting the CYP-eicosanoid pathway could offer promising new options for cardiovascular disease risk and management. An alternative therapeutic approach is to focus on supplementation of individual fatty acids. As such, recent data show that dietary C20:5 n-3 (EPA, 4 g daily, REDUCE-IT trial) is effective for prevention of major coronary events in hypercholesterolemic patients (Yokoyama et al., 2007). Dietary C20:5 n-3 supplementation is also effective for prevention of cardiovascular events in patients with established cardiovascular disease or with diabetes and other risk factors (Bhatt et al., 2018). Our results demonstrate that RBC fatty acids, including the omega-3 quotient, do not vary during hemodialysis. The results are supported by studies, which did not use large-scale lipidomics but rather focused on individual fatty acids and used smaller number of patients (Friedman, Siddiqui, & Watkins, 2008;Peuchant et al., 1994;Peuchant, Salles, Vallot, Wone, & Jensen, 1988;Taccone-Gallucci et al., 1989). Our results support the concept that the omega-3 quotient is strongly affected by diet, for example, C22:6 n-3/C20:5 n-3 fatty acid (DHA/EPA)-rich diet (Begum, Belury, Burgess, & Peck, 2004;Fischer et al., 2014;Saifullah et al., 2007), but not hemodialysis treatment itself. The baseline differences in RBC C20:5 n-3 (EPA), C18:3 n-6, and C20:3 n-6 levels observed in CKD patients may contribute to increased cardiovascular risk in ESRD.

| Saturated fatty acids
Surprisingly, we did detect increases in total C12:0, C14:0, C16:0, and C18:0 plasma levels (besides C20:2 n-6 and C20:4 n-6) in response to hemodialysis. Friedman et al. analyzed C18:0/C16:0 fatty acids and observed no change of C18:0, but a decrease of C16:0 fatty acid in response to a single hemodialysis (Friedman et al., 2008). The reasons for these changes and the discrepancies are unknown. The mechanism by which hemodialysis raises levels of individual saturated (C12:0, C14:0, C16:0 and C18:0) and n-6 (C20:2 n-6 and C20:4 n-6) fatty acids in plasma is not known. Since long-chain PUFA cannot be synthesized endogenously in appreciable amounts, accelerated release into plasma could be a possible explanation. Consistently, the observed accumulation of fatty acids in plasma in did not occur at the expense of free fatty acids. In support of this notion, hemodialysis treatment has been shown to upregulate lipoprotein lipase and phospholipase A2 activity, both of which produce fatty acids from triglycerides and phospholipids, respectively (Friedman et al., 2008;Watkins, Li, Allen, Hoffmann, & Seifert, 2000). The more pronounced changes observed within the blood plasma, as compared with the RBC compartment, are not unexpected since plasma is considered more dynamic with respect to fatty acid flux (Friedman et al., 2008).

| CONCLUSIONS
Our results suggest that hemodialysis treatment does not change the levels of RBC n-3 fatty acid status of ESRD patients in the systemic circulation. Our study revealed significant differences in total and free RBC fatty acid status between ESRD patients and control subjects, although the omega-3 quotients did not vary between both groups. We also found that hemodialysis did not induce increased mobilization of individual free fatty acids into plasma or erythrocytes, but caused greater rate of oxidation of total fatty acids, which accumulated (C12:0, C14:0, C16:0, C20:2 n-6 and C20:4 n-6) in the circulating blood during hemodialysis treatment. Further studies are needed to elucidate the individual fatty acids altered in CKD for their cardiovascular risk predictions.