Seasonal Variation of Nutritional Status and Oxidative Stress in Haemodialysis Patients - Are They Related?

Seasonal variations in body composition and parameters that reect nutritional status are well established in haemodialysis (HD) patients. However, no study has assessed changes in oxidative stress (OS). The objectives of our study were to assess seasonal variations in OS, body composition and other nutritional parameters, as well as their interactions. Seasonal variations in fat tissue mass (FTM), fat tissue index (FTI), adipose tissue mass (ATM), lean tissue mass (LTM), lean tissue index, body cell mass (BCM and overhydration (OH), OS (the blood levels of derivatives of reactive oxygen metabolites (d-ROMs), thiobarbituric reactive substances (TBARS), plasma protein reduced thiol content (THIOLS) and ferric reducing ability of plasma (FRAP) were measured) and other nutritional parameters were assessed in 45 HD patients aged 70 (60.5-76.5) years. those parameters of body lean mass LTM and BCM) and lower proportion of body fat but those correlations were not signicant. These results are similar to the results from previous study of Mehdi Rambod (59)where the inverse association between serum prealbumin and the percentage of total body fat was found. They found that in HD patients with higher prealbumin level was a lower proportion and higher lean body Our results showed signicant seasonal variation in preHD urea level with peak value in January. This is in line with our previous study where pre HD urea level vas signicantly higher in cold months with a value The most likely explanation is change in in dietary intake among HD CRP, hsCRP and leukocytes level did not show signicant seasonal variation with a peak value in January. Similarly to previous study and hsCRP were lower in the summer results may point to a lower risk of infective episodes in summer

Other physiologic and laboratory parameters were also seasonally different. Results showed that mortality differences were related to seasonality of physiologic and laboratory parameters (3).
During the last two decades, OS has become the center of attention as a novel, nontraditional risk factor for in ammation, atherosclerosis and chronic kidney disease (CKD) progression (4). Furthermore, OS is increased at the very early stages of CKD, (5) is augmented in parallel to deterioration of renal function, (6) and is further exacerbated by the HD procedure (7,8). Therefore, HD procedure itself contributes to in ammation and OS stress (8). It is well known that HD is characterized by excessive OS for several reasons. Firstly, multiple comorbidities that usually accompany HD patients like dyslipidemia, hypertension, metabolic syndrome, diabetes mellitus, advanced age, and atherosclerosis trigger prooxidant activity (9). Secondly, in HD patients, antioxidant defense mechanisms are impaired (4). Thirdly, the chronic in ammation that characterizes HD patients is directly linked with OS (10). Also, it is well known that several HD procedure-related factors are implicated in the pathogenesis of OS. Duration of dialysis therapy, iron infusion, anemia, presence of central venous catheter, and bioincompatible dialyzers are several factors triggering the development of OS (9). Finally, the HD procedure per se seems to activate prooxidant mechanisms (11) and every HD session is characterized by further loss of antioxidants molecules (eg. vitamins and trace elements). Convective and diffusive losses of vitamin C during haemodia ltration session is a contributive factor to oxidative stress in HD patients (12). Many dietary restrictions in HD patients as well as the high prevalence of malnutrition may also result in reduced intake of nutritional factors ( eg. vitamin C, D, E) (13) leading to signi cant depletion of antioxidant defense mechanisms.
Protein energy wasting (PEW) is one of the most common comorbidities observed in ESRD on chronic HD (14). It is well known that a multitude of factors can affect the nutritional and metabolic status of patients with ESRD, including decreased dietary nutrient intake, catabolic effects of renal replacement therapy, systemic in ammation, metabolic and hormonal derangements, and comorbid conditions such as diabetes and depression (15). PEW is associated with poor quality of life, complications and increased risk of mortality (16).
Since poor nutritional status is associated with increased death risk in HD patients (17) measuring reliable markers of PEW may lead to timely interventions for individuals at risk. Hypoalbuminemia is currently the most commonly used surrogate of PEW in HD patients and has a strong association with increased mortality in this population of patients (18) even though a low serum albumin appears to be a strong marker of in ammation rather than nutritional status (14). Several studies have advocated the use of serum prealbumin, as a better surrogate of nutritional status in this patient population (14,19).
Body composition has been shown to be an important parameter related to outcome in dialysis patients. Alterations in body composition are very common in dialysis patients, in particular loss of lean tissue mass (LTM) which is induced by PEW and highly associated to morbidity and mortality (20).
Various clinical and laboratory parameters were found to differ between seasons and previous studies have shown seasonal variations in blood pressure level (21), body mass composition and hydration state (22), interdialytic weight gain (3,23), body weight (24), calcium phosphate metabolism (25), clinical and laboratory variables that re ect nutritional status (26) and even cognitive impairment, depression, sleep disorders and quality of life (27) in maintenance HD patients.
To our knowledge there is no study focusing exclusively on seasonal variations of OS in maintenance HD patients. Considering the importance of OS and malnutrition in this population of patients the aim of this study was to de ne more precisely the seasonal variation of OS, body composition, anthropometry and laboratory parameters that re ect nutritional status in maintenance HD patients living in Dalmatia, South Croatia, as well as their interaction.

Study participants
This study was conducted at the dialysis unit of the Department of Nephrology and Dialysis, University Clinical Hospital Centre, in Split. In total, a selected population of 45 adults (15 females and 30 males) on maintenance haemodialysis (HD) patients were included.
Eligible participants were enrolled if they met the following inclusion criteria: (i) aged 18 years or more, (ii) stable duration of maintenance HD of more than six months with a three times weekly dialysis program before study entry, (iii) continuity of HD regimen and (iv) those patients have never changed their modality of dialysis treatment (from peritoneal dialysis to HD). None of the patients received antibiotics, cytotoxic drugs, blood transfusions or corticosteroids during the 3 months prior to and during participation in the study. We also excluded those patients who had an implanted pacemaker or cardioverter de brillator, stents or limb amputation; patients with liver cirrhosis and active underlying malignant disease or active infection at the beginning of the study; and patients who refused to participate in the study.
The ful lment of these criteria was determined by interviewing both participants and their relatives, as well as by reviewing participants' medical records.
The assessment of nutritional status, body mass composition, anthropometric measurements and blood sampling was performed in HD patients before the midweek HD session. Measurements were taken in July, October and January; altogether in 135 HD sessions. Blood was taken just prior to connecting the subject to the dialysis machine and before administering heparin. For determining the post HD blood urea concentration, a blood sample was obtained from the arterial line 2 min after the blood pump was reduced to 50 mL/min (slow-ow technique). All patients were receiving conventional 4 h HD, three times weekly, with bicarbonate dialysate at a ow rate of 500 mL/min and low molecular weight heparin as standard anticoagulation. The dialysis methods and pharmacological therapy have not changed (erythropoietin dose and dose of bone metabolism drugs were changed with dosage adjustment according to clinical guidelines) for the analysed months. High-ux polysulfone dialyzers (Fresenius Medical Care, Bad Homburg, Germany) were used mainly for HD.
All participants were informed of the purpose and nature of the study and provided written consent. The study protocol was accepted by the Hospital Ethics Committee of the University Hospital Centre Split (class 500-03/14-01/40, number: 2181-147-01/06/J.B.-14-2).

Assessment of nutritional status
There is not a single measurement that provides complete assessment of the nutritional status of HD patients. To assess nutritional status in HD patients in this study, anthropometric measurements; the malnutrition in ammation score (MIS); the dialysis malnutrition score (DMS); body mass composition monitoring and laboratory parameters that re ect nutritional status such as serum albumin and prealbumin were used.
The following anthropometric parameters were collected for each study subject: height (cm), dry weight (kg), triceps skin fold (mm), waist circumference (WC), hip circumference (HC) and mid-upper arm circumference (MUAC). Additionally, waist-to-height ratio (WHtR) and BMI were calculated for each study subjects. The subjects stood upright, facing forward with their shoulders relaxed, and measures were taken with exible, nonstretchable measuring tape.
The MIS assessment was performed according to the description by Kalantar-Zadeh et al. (28). The MIS is measured on a 4-point scale and is a quantitative nutrition screening tool consisting of four main parts: patients' related medical history, physical examination, BMI and laboratory parameters. The sum of all components ranges from 0 to 30, and a higher score re ects a more severe degree of malnutrition and in ammation. The MIS was found to signi cantly correlate with hospitalization; mortality; and indices of nutrition, in ammation, and anaemia (29,30). Dialysis Malnutrition Score (DMS) consists of seven features; weight change, dietary intake, GI symptoms, functional capacity, co-morbidity, subcutaneous fat and signs of muscle wasting. Each component has a score from 1 (normal) to 5 (very severe). Thus the malnutrition score (sum of all seven components) is a number between 7 (normal) and 35 (severely malnourished). Lower score denotes tendency towards a normal nutritional status. A higher score is considered to be an indicator of the presence of malnutrition elements i.e. protein energy malnutrition (31).
The assessment of body mass composition was carried out using the Body Composition Monitor portable device (Fresenius Medical Care), which works on the principle of bioimpedance spectroscopy (32). The Body Composition Monitor is a valid method for assessing and monitoring hydration and nutritional status in haemodialysis patients (6). Measurements using the Body Composition Monitor

Oxidative stress
The blood levels of derivatives of reactive oxygen metabolites (d-ROMs) as indicator of protein oxidation and thiobarbituric reactive substances (TBARS) as indicator of lipid oxidation were measured for all study subjects. Total antioxidant capacity, as plasma protein reduced thiol content (THIOLS) and ferric reducing ability of plasma ( FRAP) were also measured.
The d-ROM assay is intended to measure the concentration of total hydroperoxides in serum or heparin plasma. The method was rst described by Alberti et al. (34). It is based on the following principle. In an acidic buffered solution (pH = 4.8), iron ions are released from the serum (plasma) proteins and catalyze the in vitro transformation of hydroperoxides into alkoxyl and peroxyl radicals, which further react with the chromogen N,N-diethyl-p-phenylenediamine. The concentration of the colored complex is directly proportional to the concentration of the hydroperoxides that are present in the sample. The absorbance was measured at 505 nm, and the results are expressed in Caratelli units (CARR U). One CARR U corresponds to 0.08 mg/100 mL H2O2. The characteristics of the assay were evaluated and validated by Verde et al. (35), who reported that the assay is reliable even in patients with hyposideremic anemia.
The total antioxidant capacity, as the FRAP value, was measured. FRAP was measured by the ferric reducing/antioxidant and ascorbic acid (FRASC) assay (36). In this assay, the samples are treated with or without ascorbate oxidase, and antioxidants in the sample are evaluated as reductants of Fe3+ to Fe2+, which is chelated by tripyridyltriazine (TPTZ) to form an Fe2+-TPTZ complex absorbing at 593 nm.
Absorbance was monitored with a UV-Vis spectrophotometer equipped with a six-cell holder and a thermostatically controlled bath. The results were compared with a standard curve prepared daily with different concentrations of vitamin C (ascorbic acid) and expressed as micromoles of vitamin C equivalents. The validity of this method for determination of vitamin C in heparinized human plasma has been shown previously (37).
The assay for plasma protein reduced thiol content (THIOLS) measures the major source of reducing equivalents (or antioxidant capacity) available in the plasma. Thiol groups were assayed according to the method of Ellman (38) as modi ed by Hu (39), as described in Himmelfab et al. (40). Brie y, 2 mL of buffer containing 0.1mol/L Tris,1 mmol/L EDTA, pH 8.2, and 100mL plasma was added to cuvettes, followed by 100uL 10 mmol/L 5dithio-bis (2-nitrobenzoic acid) (DTNB) in methanol. Blanks were run for each sample, prepared as above, with the exception that there was no DTNB in the methanol. Following incubation for 15 minutes at room temperature, sample absorbance was read at 412 nm. Results are expressed as umol/L of glutathione.
TBARS assay is based on reaction of malondialdehyde (MDA), one of end products of lipid peroxidation, with TBA (41). To correct for background absorption, the absorbance values at 572 nm were subtracted from those at 532 nm, which represent the absorption maximum of the TBA:MDA adduct (42). Absorbance was monitored by the above-mentioned UV-spectrophotometer. All measurements were carried out in triplicate. Results were compared with a standard curve prepared daily with different concentrations of MDA and expressed as lmol⁄ L MDA.

STATISTICAL METHODS
To discern the mean effect in the numeric variables of the three measurements,with the signi cance level of 0.05, power 0.95, and the effect size 0.25, the minimum sample size required was 44 subjects (calculated by G * Power 3.1.2, Franz Faul, University in Kiel, Germany). Categorical data are represented by absolute and relative frequencies. Numerical data were described by the median and the limits of the interquartile range. The normality of the distribution of numerical variables was tested by the Shapiro-Wilk test. Differences between numerical variables between measurements were tested by Friedman's test (Post hoc Conover). The correlation between numerical variables was evaluated by Spearman's correlation coe cient ρ (rho). Using multivariate regression analysis (stepwise), we determined the predictors that in uence oxidative stress. All P values were two-sided. The signi cance level was set to Alpha = 0.05. MedCalc Statistical Software version 19.1.7 (MedCalc Software Ltd, Ostend, Belgium; https://www.medcalc.org; 2020) was used in the statistical analysis.

Results
Data were processed on 135 single HD treatments involving a group of 45 MHD patients. Baseline demographic data demonstrated (Table 1) that average age of participants was 70 years (range 60.5-76.5 years), 15 (33.3%) were women, and HD vintage was 27 months (range 16-53.5 months). The average MIS at the beginning of study was 8 (range 6-9) and average DMS was 10 (range 9-12). Other baseline data regarding anthropometric value, body composition parameters, laboratory parameters and parameters of OS for all study participants are shown in Table 1.
OH signi cantly decreased (P=0.004) over season with a peaked value in June. In contrast to body mass composition parameters, neither dry weight nor BMI signi cantly changes over season during 6 months follow up.
Also, signi cant seasonal variations in biochemical parameters in 135 HD session performed on 45 maintenance HD patients were found as shown in Table 3. Signi cant seasonal difference among all study subject were found in uric acid (P=0.03), pre HD urea (P=0.002), albumin (P=0.001) and prealbumin (P=0.001) level. In contrast, neither serum creatinine level, haemoglobin, CRP nor hs-CRP level varied with the season.
Furthermore, seasonal variations of OS parameters during follow up for all study subjects were analysed as shown in Table 4. Signi cantly seasonal variations in d-ROMs (P=0.02) and THIOLS (P=0.02) value were found. Also, results showed signi cantly positive correlation between d-ROMs and hs-CRP (ρ=0.361, P=0.03) ( Figure 1) and signi cantly positive correlation between dialysis duration and TBARS (ρ=0.324, P=0.02).
Correlations of body mass composition parameters with parameters of nutritional status, other anthropometric and biochemical parameters in January are shown in Table 5  To analyze factors that affect body composition parameters we performed multiple regression analysis.

Discussion
This study evaluated seasonal variations of nutritional status, anthropometric, body mass composition and OS parameters during 6-months period and their associations in a group of maintenance HD patients living in a mild Mediterranean climate of the Adriatic coast in Dalmatia, South Croatia, Europe. To our knowledge this is the rst detailed European report of seasonal variation in OS parameters in maintenance HD patients.
In this population of HD patients, we showed signi cant seasonal difference among all study subjects in all body mass composition parameters and anthropometric parameters such are triceps skin fold, WC and WHtR.
First of all, OH signi cantly decreased during 6 months of follow up, especially during rst three months (between June and October). Possible explanations for this nding might be that initial measurement was done at summer month (July) when uid intake is higher as a consequence of increased thirst and higher ambient temperatures and other possibility is that after initial BIA measurement patients were more aware of OH and became more compliant in next three months. Our results are similar to the results of Broers et al. (22) who found that uid overload was highest in spring and summer and lowest in winter.
Therefore, body mass composition parameters that represent fat tissue (FTI, FTM and ATM) showed signi cant seasonal variation during 6 months follow up whereas body mass composition parameters that are predictors of nutritional status (LTI, LTM and BCM) showed signi cant decrease. Results showed that FTI, FTM and ATM were highest in winter (in January) and lowest in summer (in July) whereas LTI, LTM and BCM were lowest during winter (in January) and highest in summer (June). This pattern of body composition changes was already con rmed in previous studies, showing that with dialysis vintage maintenance HD patients experience loss of muscle mass with increasing fat mass and it is called "sarcopenic obesity" (43,44). These results are in line with a recent cohort study in which increase in FTI and decrease in LTI were observed within 2 years following the start of dialysis (43). A combination of these two disorders of body composition leads to a higher risk of mortality more than when each pathology occurs separately (44,45).
As mentioned earlier, anthropometric parameters such are triceps skin fold, WC and WHtR also showed signi cant seasonal differences during follow up. Possible explanation for this ndings in our population of HD patients might be signi cantly increase in fat mass content during follow up because those HD patients with higher value of WC and WHtR had signi cantly higher fat tissue content (higher value of FTI, FTM and BCM) after 6 months of follow up (in January). WC (46,47,48,49) and WHtR (48,49,50) are commonly used in HD patients to assess visceral fat, and the predictability of mortality in HD patients further strengthens the role of WC (47). Obese dialysis patients have both a large WC and high percentage of body fat (51).
In this population of HD patients, we did not show signi cant seasonal differences among all study subject in BMI and dry weight with peak values for both parameters occurring in January. In contrast to results from our study, in previous studies BMI and pre HD weight varied with season but also, like in our study, peaked in January. The difference may be due to older HD patients and shorter HD duration in our study. Body fat mass increase and lean body mass decreased signi cantly during rst year of HD (52). That could be explanation for our ndings, and also because of signi cant increase of fat mass and signi cant decrease of lean mass and OH, BMI and dry weight did not change signi cantly during follow up. The study of Broers et (22) all. supported the fact that BMI in cold months is higher due to fat mass. Next to nutritional intake, it is also possible that low physical activity plays a role in peak value of BMI, signi cantly higher fat mass and lower lean tissue mass in January. It is known that dialysis patients have decreased levels of physical activity, matching a sedentary lifestyle (53). Physical inactivity in dialysis patients is associated with an increased risk for hospitalization and mortality (54,55) and with alterations in body composition and decreased muscle strength (56).
Both albumin and prealbumin levels are sensitive to protein-calorie malnutrition and are negative acutephase proteins, exhibiting decreases in their serum concentration during episodes of in ammation (57,58). Longitudinal measures of albumin or prealbumin concentrations may provide more information about the risk of adverse outcomes on dialysis than single measures of these proteins (59,60). Our results showed statistically signi cant seasonal difference in serum albumin level and serum prealbumin level, with the peak values for albumin occurring in July, and the peak value for prealbumin peaked in January. In contrast to our study in study of longitudinal serum albumin and prealbumin concentrations, (61) found that serum albumin concentrations increased, whereas prealbumin concentrations did not change over time on average. The difference may be due to older HD patients, higher baseline serum albumin, lower prealbumin and CRP level and shorter follow up in our study. Therefore, in our previous study (26) we did not found signi cant seasonal difference nor difference between cold (December and January) and mild (June and September) months in serum albumin level. Possible explanation for this difference might be that in previous study was higher baseline serum albumin level. In our previous study 70.2 % HD patients had baseline serum albumin level ≥ 38 g/L while in present study 51% HD patients had baseline serum albumin level ≥ 38 g/L. According to consensus of the panel of experts of the International Society for renal Nutrition and metabolism serum albumin level < 38 g/L is one of the biochemical parameters that may be indicative of PEW in individuals with kidney disease (14). The interpretation of serum albumin level is complicated since albumin level concentration is the compound resultant of in ammation, nutrition, and uid status (62). Data from previous studies about seasonal variation in albumin level are inconclusive. In contrast to serum albumin, serum prealbumin half-life is relatively short, i.e., 2 to 3 days (63). Hence, it may be a more sensitive indicator of nutritional status than either serum albumin or transferrin (19,64,65). In our study those HD patients with higher prealbumin level had signi cantly higher parameters of body lean mass (LTI, LTM and BCM) and lower proportion of body fat but those correlations were not signi cant. These results are similar to the results from previous study of Mehdi Rambod (59)where the inverse association between serum prealbumin and the percentage of total body fat was found. They found that in HD patients with higher prealbumin level was a lower proportion of body fat and higher proportion of lean body mass.
Our results showed signi cant seasonal variation in preHD urea level with peak value in January. This is in line with our previous study where pre HD urea level vas signi cantly higher in cold months with a peaked value in January (26). The most likely explanation is change in in dietary intake (increased protein intake) in cold months among HD patients. CRP, hsCRP and leukocytes level did not show signi cant seasonal variation with a peak value in January. Similarly to previous study (62,3) CRP and hsCRP were lower in the summer period. These results may point to a lower risk of infective episodes in summer periods.
Of the other laboratory parameters, serum phosphate levels were not signi cantly different between the various season observed in our study. It is important to note no reliable data on phosphate binders were available in this study and that in general, seasonal variation in laboratory parameters are not completely consistent between studies (23,66).
Uric acid level in our study showed signi cantly seasonal differences, uric acid signi cantly decreased from June to October. Possible explanation could be change in food intake. Those patients who consumed more fruits and vegetables had a higher uric acid level which could be attributed to an overall higher intake of food and high-purine products (67). Uric acid is a powerful oxygen radical scavenger in hydrophilic environments, and a study on a large cohort showed that low and not high serum uric acid level predicted all-cause and CV mortality (68). Longitudinal changes in serum uric acid seem to track with changes in nutritional status over time, and these changes are associated with survival of patients on maintenance HD. An increase in serum uric acid levels over time is accompanied by improvement of nutritional status and lower mortality rate (69).
To our knowledge this is the rst study that evaluated seasonal difference in OS stress parameters in HD patients. Our results showed statistically signi cant seasonal difference in level of d-ROMs. d-ROMs level increased from June to January, with peaked value in January. The most likely explanation for signi cantly increase in d-ROMs value during season might be signi cantly increase in fat content during observed seasons. Univariant regression analysis showed that d-ROM was signi cantly in uenced with FTI, FTM and hip circumference. Furthermore, in multivariant analysis most signi cant positive predictor of d-ROMs was hip circumference. It is important to note that in our study those HD patients with higher hip circumference had signi cantly higher fat content (adipose tissue). Also, most signi cant positive predictor of fat tissue indices (FTI, FTM, ATM) was hip circumference and statistically signi cant negative predictor of lean tissue indices (LTM and BCM) was d-ROM. Adipose tissue is a potential source of in ammation in ESRD that is not due to increased adiposity and may contribute to mitochondrial dysfunction in uraemia (70). In line with those nding d-ROMs correlated positively with hsCRP in our population of HD patients. These results suggest correlation between anthropometric value, body mass composition indices and OS. Furthermore, OS might be promoted by fat tissue in HD patients and also could have negative in uence on lean tissue mass.
In addition, we also measured antioxidant defence in serum of our HD population by measuring plasma THIOLS. Signi cant seasonal variations in THIOLS level was found with the peak value in October. Similar pattern of seasonal difference was found for THIOLS and serum albumin and prealbumin level. In multivariate analysis results showed that serum prealbumin level was most signi cant predictor for THIOLS level.
Our results did not show statistically signi cant seasonal difference in level of TBARS. Previous ndings suggested that TBARS levels are increased in the subcutaneous fat tissue of patients with ESRD (71). In contrast to previous study, in our study fat tissue indices ( FTI, FTM, ATM) signi cantly changed over season but TBARS level did not change signi cantly. But, it is important to note that those HD patients with longer dialysis duration had signi cantly higher TBARS which might re ect higher level of OS. The in ammatory status and duration of dialysis treatment are the most important factors relating to oxidative stress in HD patients (72). In line with those ndings our HD patients with longer dialysis duration had signi cantly higher level of TBARS.
Also, total antioxidant capacity expressed as FRAP did not show signi cant seasonal differences. It is well known that fruits and vegetables supply the organism with low-molecular exogenous antioxidants and may contribute to the overall antioxidant capacity of plasma. Furthermore, maintenance HD patients are subject to a number of dietary restrictions and their overall diet frequently does not meet daily caloric and protein requirements (73). Also, the risk of elevated potassium and uid overload may lie behind decreased fruit and vegetable consumption in this group of patients (74). Malnutrition and hypoalbuminemia reduce antioxidant defence (75) and albumin and prealbumin, commonly used nutritional markers, possess antioxidant properties and can account for signi cantly lower FRAP values in HD patients when compared with the healthy controls (67). Despite signi cant seasonal variation in serum albumin, prealbumin and uric acid level our results did not show seasonal variations in FRAP and vitamin C. These ndings suggest that many other components contribute to total antioxidants capacity of plasma in HD patients.
During HD therapy, a signi cant amount of vitamin C is lost and also ascorbic free radicals are formed contributing thus to enhanced OS. Our data did not show signi cant seasonal variation in vitamin C level.
Possible explanation for this ndings might be supplementation of vitamin C as regular treatment in our HD unit for every HD patient.
There was not statistically signi cant difference between season in the DMS and MIS score. These result is in line with result from our previous study (26) were we did not nd seasonal difference in DMS score nor difference between cold and mild months (26). But is important to note that in multivariate analysis MIS was most signi cant positive predictor for FTI and negative predictor for LTI and BCM. Wang WL showed that a high MIS was signi cantly correlated with a low LTI and low FTI (76). The difference may be due to older HD patients, more women and different region and season in our study.
Several limitations need to be acknowledged. One limitation of this study is that the in uence of the dietary habits on nutritional status was not considered. Furthermore, this is single centre observational study with few cases. Also, the study period was nine months and we did not show data for spring season so in this condition a lot of biases such is selection and measured bias might occur. Future multicentre studies on seasonal variation of nutritional status and oxidative stress in maintenance HD patients are needed, with larger number of participants in a different climate region and in a prospective search model and for a longer study period.

Conclusion
We con rmed that seasonal variation in clinical, body mass composition and laboratory parameters that re ect nutritional status are common among maintenance HD patients. Also, to our knowledge this is the rst study that showed seasonal variation in OS parameters. Therefore, our results suggest possible bidirectional correlation between OS and body mass composition parameters, especially those parameters of body composition that re ect fat tissue. It is well known from previous studies that MIS, DMS, serum albumin and in ammatory markers are commonly used predictor of mortality of HD patients. Also, there is a growing body of evidence suggesting that OS predict all-cause and/or CV mortality and adverse CV events in HD patients. These results indicate a need of more regular and systemic approach to careful examination of nutritional status in HD patients with special attention to body mass composition. It is important to note that these seasonal variations might lead to biases in the interpretation of results in clinical studies in which measurement schedules for clinical, biochemical, body mass composition and OS parameters vary during the year.
Based on the results from our previous study (26)  BackgroundThe data obtained in the current study will be available from the corresponding author upon reasonable request after publication of the results on the main research questions.

Con ict of interest
The authors declare that they have no con ict of interest.

Funding
None.

Ethical approval
All procedures performed in this study were in accordance with the ethical standards of University Hospital Center Split and with the 1964 Helsinki declaration and its later amendments.   Figure 1 Correlations between hs C-reactive protein (hs CRP) and derivatives of reactive oxygen metabolites (d-ROMs).