Inflammatory markers predict MIA syndrome progression and cardiovascular disease outcomes in maintenance dialysis patients

Abstract

Objective

To investigate associations between inflammatory markers and malnutrition-inflammation-atherosclerosis (MIA) syndrome in dialysis patients and evaluate their predictive value for cardiovascular events.

Methods

This retrospective cohort study enrolled 197 maintenance dialysis patients (2017–2022). Inflammatory markers (CRP, IL-6, TNF-α), nutritional parameters, and atherosclerosis indicators were collected at baseline, 6, 12, and 24 months. Multivariate regression and Cox models assessed associations with MIA syndrome and cardiovascular outcomes.

Results

Elevated inflammatory markers were significantly associated with MIA syndrome severity and cardiovascular events (P < 0.001). High-inflammation patients had higher MIA scores (8.7 ± 2.1 vs. 6.4 ± 1.9, P < 0.001). CRP correlated negatively with albumin (r=-0.41) and positively with carotid intima-media thickness (r = 0.36, both P < 0.001). Hemodialysis patients showed higher CRP levels (8.2 ± 4.4 vs. 6.1 ± 3.7 mg/L, P = 0.011) and cardiovascular event rates (46.3% vs. 29.2%, P = 0.027) than peritoneal dialysis patients. Inflammatory changes correlated with MIA progression (r = 0.37, P < 0.001). After adjustment, CRP (HR = 1.17, 95% CI: 1.08–1.26), IL-6 (HR = 1.12, 95% CI: 1.04–1.20), and TNF-α (HR = 1.14, 95% CI: 1.05–1.23) independently predicted cardiovascular events.

Conclusion

Inflammatory markers independently associate with MIA syndrome and cardiovascular events in dialysis patients. Hemodialysis patients demonstrate greater inflammatory burden than peritoneal dialysis patients, with inflammation potentially mediating this difference. Findings support inflammatory monitoring for risk stratification and targeted interventions.

Introduction

Chronic kidney disease (CKD) has become a major global public health concern, with many patients eventually progressing to end-stage renal disease (ESRD), which requires maintenance dialysis to sustain life. Although advancements in dialysis technology have extended patient lifespans, cardiovascular disease (CVD) remains the leading cause of mortality among ESRD patients [1–6]. In recent years, malnutrition-inflammation-atherosclerosis (MIA) syndrome has garnered widespread attention as a critical mechanism underlying the high mortality rates observed in dialysis patients. MIA syndrome involves a complex interplay among malnutrition, chronic inflammation, and atherosclerosis, which not only significantly elevates the risk of CVD but is also closely associated with increased all-cause mortality [7–9].

Current research indicates that inflammation plays an important role in the development of MIA syndrome. Inflammatory markers such as C-reactive protein (CRP), interleukin-6 (IL-6), and tumor necrosis factor-alpha (TNF-α) have been demonstrated to be closely associated with poor prognoses in CKD patients [10–12]. However, existing studies have focused predominantly on data from single time points, with limited longitudinal data on the dynamic changes in inflammatory markers and their relationship with MIA syndrome progression. Additionally, while different dialysis modalities (hemodialysis vs. peritoneal dialysis) are known to differ in their inflammatory profiles [13–17], the extent to which inflammation mediates the relationship between dialysis modality and patient outcomes remains incompletely understood.

Hemodialysis and peritoneal dialysis may affect inflammatory responses through distinct biological mechanisms. Hemodialysis involves frequent extracorporeal circulation and exposure to dialysis membranes, which may trigger complement activation and cytokine release. In contrast, peritoneal dialysis utilizes the peritoneal membrane as a more physiological interface, potentially resulting in a different inflammatory profile. However, whether these differences in inflammatory burden contribute to differential clinical outcomes between dialysis modalities remains unclear, and the potential mediating role of inflammation in this relationship has not been systematically investigated.

Therefore, this study aimed to systematically investigate the associations between inflammatory markers and MIA syndrome in patients undergoing maintenance dialysis, with a focus on longitudinal changes and their clinical implications. Specifically, by collecting data on inflammatory marker levels (CRP, IL-6, and TNF-α) at multiple time points over 24 months, we sought to (1) examine the correlation between dynamic changes in inflammatory markers and MIA syndrome progression; (2) evaluate the independent predictive value of these markers for cardiovascular events; and (3) compare inflammatory profiles between hemodialysis patients and peritoneal dialysis patients and explore whether inflammation mediates the relationship between dialysis modality and clinical outcomes. These findings may provide insights for identifying high-risk dialysis patients and developing targeted anti-inflammatory interventions, with potential implications for improving cardiovascular outcomes in this vulnerable population.

Methods

Study design and population

This retrospective cohort study was conducted at the dialysis center of Dongyang Hospital of Traditional Chinese Medicine, utilizing data from the “Long-term Prognosis Assessment Project for Dialysis Patients,” which has systematically collected clinical indicators and research parameters since 2017.

We enrolled adult patients with end-stage renal disease receiving maintenance dialysis (hemodialysis or peritoneal dialysis) between January 2017 and January 2022. The inclusion criteria were as follows: (1) aged 18–80 years; (2) maintenance dialysis for ≥ 6 months; (3) a stable dialysis regimen without modality change in the preceding 3 months; and (4) complete baseline inflammatory marker measurements. For peritoneal dialysis patients, only those on prevalent continuous ambulatory peritoneal dialysis (CAPD) or automated peritoneal dialysis (APD) for ≥ 6 months were included. Patients who experienced PD technique failure and who switched to hemodialysis were excluded from the PD group to avoid confounding by indications.

The exclusion criteria included acute infection or inflammatory conditions at baseline, major surgery within 3 months, active malignancy, severe liver dysfunction (Child‒Pugh class C), kidney transplantation during follow-up, or incomplete data preventing MIA syndrome assessment.

Data were extracted from electronic medical records by two independent researchers with quality control and consensus resolution of discrepancies. This retrospective study employed convenience sampling without prospective sample size calculations. Post hoc power analysis indicated 80% power to detect a hazard ratio ≥ 1.5 for cardiovascular events (α = 0.05, two-sided), given our sample of 197 patients and an observed event rate of 42.6%.

Data collection

Clinical characteristics

Demographic data (age, sex, BMI), dialysis modality and vintage, comorbidities, and medication use were extracted from medical records. Comorbidities were systematically documented according to standard definitions: hypertension (blood pressure ≥ 140/90 mmHg or antihypertensive use), diabetes mellitus (fasting glucose ≥ 126 mg/dL, HbA1c ≥ 6.5%, or glucose-lowering medication), and cardiovascular disease (coronary artery disease, heart failure, arrhythmias, or valvular disease). Residual kidney function was assessed by 24-hour urine output (≥ 200 mL/day considered significant). Dialysis adequacy was measured by Kt/V (target ≥ 1.2 for hemodialysis, ≥ 1.7 for peritoneal dialysis).

Laboratory data

Baseline data were collected at study enrollment (January 2017 to January 2022), with subsequent measurements at 6, 12, and 24 months (± 2 weeks). Blood samples were collected at four time points: baseline, 6 months, 12 months, and 24 months (± 2 weeks). To minimize variability, standardized sampling protocols were employed. For hemodialysis patients, blood was drawn predialysis (before heparin) during mid-week sessions (typically Wednesday or Thursday) to represent steady-state conditions. For PD patients, samples were obtained during morning clinic visits (8:00–10:00 AM) after an overnight dwell with complete peritoneal drainage prior to venipuncture. The samples were processed within 2 h, centrifuged, aliquoted, and stored at -80 °C until batch analysis.

Inflammatory markers

(CRP, IL-6, and TNF-α) were measured via standardized immunoassays: CRP was measured via high-sensitivity turbidimetric immunoassay (normal range < 3 mg/L), and IL-6 and TNF-α were measured via enzyme-linked immunosorbent assay (ELISA). Nutritional indicators included serum albumin, prealbumin, total cholesterol, and transferrin at all time points; leptin was measured only at baseline owing to resource constraints. Atherosclerosis-related parameters included lipid profiles (LDL-C, HDL-C, and triglycerides), which are used to define dyslipidemia (LDL-C > 2.6 mmol/L or HDL-C < 1.0 mmol/L) as a component of MIA syndrome scoring.

Imaging assessment

Carotid intima‒media thickness (CIMT) was measured via B-mode ultrasound at the far wall of the common carotid artery (1–2 cm proximal to the carotid bulb) following standardized protocols. Bilateral measurements were performed by experienced sonographers blinded to the inflammatory marker status, and the mean left and right CIMT values were used for analysis. A CIMT > 1.0 mm was considered indicative of subclinical atherosclerosis. Measurements were performed at baseline and at 24 months.

Nutritional assessment

The subjective global assessment (SGA) [18, 19] was performed by trained nephrologists or dietitians, who categorized patients as A (well-nourished), B (mild-to-moderate malnutrition), or C (severe malnutrition) on the basis of medical history and physical examination. The objective nutritional parameters included serum ALB (< 35 g/L defined as hypoalbuminemia), prealbumin (< 200 mg/L considered deficient), and BMI (< 23 kg/m² considered underweight for Asian dialysis patients). A composite nutritional score (0–4) was calculated by assigning one point for each abnormal parameter; this composite score was used in stratified analyses.

Classification and assessment systems

Inflammatory status classification

Patients were classified as having high inflammation if any marker exceeded established thresholds associated with adverse outcomes in dialysis populations: CRP > 8 mg/L (approximately twice the upper normal limit, consistently linked to cardiovascular mortality) [20–22], IL-6 > 20 pg/mL (median value in large dialysis cohorts, independently predictive of mortality) [23], or TNF-α > 25 pg/mL (upper tertile in previous studies, associated with protein-energy wasting). The “any criterion” approach captures the activation of multiple inflammatory pathways [24, 25]. Sensitivity analyses used alternative classifications: (1) quartile-based grouping; (2) median splits for individual markers; and (3) ≥ 2 elevated markers for high inflammation status.

MIA syndrome severity

MIA syndrome severity was quantified via a composite score (range 0–10) based on established criteria [7–9] and expert consensus, which incorporates three components: (1) Malnutrition (0–4 points): albumin < 35 g/L, prealbumin < 200 mg/L, SGA B/C, and BMI < 23 kg/m²; (2) Inflammation (0–3 points): CRP > 8 mg/L, IL-6 > 20 pg/mL, and TNF-α > 25 pg/mL; (3) Atherosclerosis (0–3 points): CIMT > 1.0 mm, vascular calcification on imaging, and dyslipidemia (LDL-C > 2.6 mmol/L or HDL-C < 1.0 mmol/L). One point was assigned for each criterion met. The scoring system was calculated at baseline and at 24 months. Internal validation in a 30-patient subset demonstrated acceptable internal consistency (Cronbach’s α = 0.81) and test‒retest reliability (r = 0.79), although external validation in independent cohorts is needed.

Follow-up and outcome assessment

Patients were followed from enrollment through January 2023 or until death, kidney transplantation, transfer, or loss to follow-up (median: 24 months; IQR: 18–24 months). Data, including dialysis records, hospitalizations, and clinic visits, were obtained from electronic medical records.

The primary outcome was cardiovascular events, which were defined as newly diagnosed events after enrollment (for patients with baseline cardiovascular disease, only new events or documented recurrences were counted). Events included myocardial infarction, heart failure, stroke, and angina pectoris, which were diagnosed according to established clinical guidelines. Myocardial infarction required ≥ 2 of the following: typical chest pain, ECG changes, elevated troponin, or imaging evidence [26]. Heart failure requires clinical manifestations, physical signs, and objective evidence (echocardiography: LVEF < 50% or diastolic dysfunction; and/or BNP > 400 pg/mL or NT-proBNP > 900 pg/mL) [27]; importantly, diagnosis requires assessment after achieving euvolemia (by clinical examination, stable weight, and bioimpedance when available) to distinguish cardiac dysfunction from volume overload in dialysis patients. Stroke was defined as a neurological deficit > 24 h confirmed by neuroimaging. Angina required typical symptoms confirmed by stress testing or coronary stenosis > 70% [28].

All events were documented with supporting evidence and verified by attending physicians. Major endpoints (myocardial infarction, stroke, death) were adjudicated by an independent committee (two cardiologists, one nephrologist) blinded to the patients’ inflammatory status, with disagreements resolved by consensus.

Statistical analysis

All analyses were performed via SPSS 26.0 and R 4.2.0, with two-sided P < 0.05 considered significant. Continuous variables are reported as the means ± SDs or medians (IQRs) and were compared via t tests/ANOVAs or Mann‒Whitney/Kruskal‒Wallis tests; categorical variables are reported as frequencies and were compared via the chi‒square test or Fisher’s exact test. Correlations between inflammatory markers and nutritional/atherosclerosis indicators were assessed via Pearson or Spearman correlation coefficients.

The inflammatory marker growth rate over 24 months was calculated as [(24-month value − baseline)/baseline] × 100, with the last observation carried forward for incomplete follow-up (n = 15, 7.6%). Cox proportional hazards models were used to examine time-to-cardiovascular events, with inflammatory markers analyzed as continuous (per unit increase) and categorical (high vs. low) variables, adjusting for age, sex, dialysis characteristics, comorbidities, hemoglobin, albumin, and residual kidney function. The proportional hazards assumption was verified; model discrimination was assessed by the C-index. Multivariate linear regression was used to examine inflammatory-MIA associations with identical covariates. Formal mediation analysis using the PROCESS macro [29] was employed to test whether inflammation mediates dialysis modality‒outcome relationships via bootstrap methods (5000 iterations). Sensitivity analyses excluded patients with diabetes/cardiovascular disease, outliers, and alternative inflammatory thresholds. Subgroup analyses stratified by sex and age were used to assess effect modification.

Results

Baseline characteristics

A total of 197 maintenance dialysis patients were enrolled, comprising 117 males (59.4%) and 80 females (40.6%), with a mean age of 62.7 ± 12.3 years and a mean duration of dialysis of 4.2 ± 2.3 years. The most common comorbidities were hypertension (83.8%), diabetes mellitus (54.3%), and cardiovascular disease (39.6%). Among the participants, 149 (75.6%) received hemodialysis, and 48 (24.4%) received peritoneal dialysis(Table 1).

Table 1.

Baseline characteristics of the study participants

Characteristics	All patients (n = 197)	Hemodialysis (n = 149)	Peritoneal dialysis (n = 48)	P value
Demographics
Age (years)	62.7 ± 12.3	63.1 ± 11.8	61.3 ± 13.1	0.314
Male (%)	59.4	58.4	62.5	0.592
BMI (kg/m²)	24.3 ± 3.8	24.5 ± 3.6	23.8 ± 4.1	0.273
Dialysis vintage (years)	4.2 ± 2.3	4.3 ± 2.1	4.0 ± 2.7	0.406
Residual urine output (mL/day)	185 ± 132	172 ± 128	221 ± 139	0.031
Kt/V	1.43 ± 0.21	1.39 ± 0.18	1.84 ± 0.24	< 0.001
Comorbidities
Hypertension (%)	83.8	84.6	81.3	0.635
Diabetes mellitus (%)	54.3	55.7	50	0.517
Cardiovascular disease (%)	39.6	40.9	35.4	0.497
Laboratory Parameters
Hemoglobin (g/dL)	10.2 ± 1.8	10.1 ± 1.7	10.4 ± 1.9	0.342
Serum creatinine (mg/dL)	9.8 ± 3.2	10.1 ± 3.3	9.0 ± 2.9	0.048
Serum potassium (mmol/L)	4.8 ± 0.7	4.9 ± 0.7	4.6 ± 0.6	0.015
Calcium (mmol/L)	2.21 ± 0.23	2.22 ± 0.25	2.19 ± 0.21	0.453
Phosphate (mmol/L)	1.83 ± 0.54	1.89 ± 0.52	1.67 ± 0.47	0.012
Inflammatory Markers
CRP (mg/L)	7.7 ± 4.3	8.2 ± 4.4	6.1 ± 3.7	0.011
IL-6 (pg/mL)	18.3 ± 10.1	18.7 ± 9.9	16.5 ± 11.1	0.121
TNF-α (pg/mL)	21.8 ± 14.7	22.5 ± 15.1	19.4 ± 13.2	0.174
Nutritional Parameters
Serum albumin (g/L)	37.2 ± 4.8	36.8 ± 5.1	38.4 ± 3.9	0.065
Prealbumin (mg/L)	278 ± 82	271 ± 85	296 ± 71	0.089
SGA B/C (%)	48.2	51	39.6	0.172
Atherosclerosis Indicators
CIMT (mm)	1.08 ± 0.31	1.10 ± 0.33	1.03 ± 0.26	0.187
CIMT > 1.0 mm (%)	56.9	59.1	50	0.281
LDL-C (mmol/L)	2.87 ± 0.94	2.91 ± 0.97	2.76 ± 0.85	0.342
Dyslipidemia (%)	64.5	66.4	58.3	0.305
MIA Syndrome Score	5.8 ± 2.3	6.1 ± 2.4	5.1 ± 2.0	0.014

Note: Values are the means ± SDs or percentages. P < 0.05 indicated statistical significance. BMI = body mass index; Kt/V = dialysis adequacy; CRP = C-reactive protein; IL-6 = interleukin-6; TNF-α = tumor necrosis factor-alpha; SGA = Subjective Global Assessment; CIMT = carotid intima‒media thickness; LDL-C = low-density lipoprotein cholesterol; MIA = malnutrition‒inflammation‒atherosclerosis

Baseline characteristics were generally similar between dialysis modalities, although hemodialysis patients had significantly lower residual urine output (172 ± 128 vs. 221 ± 139 mL/day, P = 0.031), lower dialysis adequacy (Kt/V: 1.39 ± 0.18 vs. 1.84 ± 0.24, P < 0.001), higher serum phosphate (1.89 ± 0.52 vs. 1.67 ± 0.47 mmol/L, P = 0.012), and higher CRP levels (8.2 ± 4.4 vs. 6.1 ± 3.7 mg/L, P = 0.011). However, the IL-6 and TNF-α levels did not differ significantly between modalities (P = 0.121 and P = 0.174, respectively).

Longitudinal changes in MIA syndrome components

Over the 24-month follow-up, all three MIA syndrome components demonstrated progressive deterioration (Table 2). The levels of the following inflammatory markers significantly increased: CRP increased from 7.7 ± 4.3 to 9.2 ± 5.1 mg/L (P < 0.001 for trend), IL-6 from 18.3 ± 10.1 to 21.8 ± 12.3 pg/mL (P < 0.001), and TNF-α from 21.8 ± 14.7 to 24.9 ± 16.2 pg/mL (P < 0.001). Nutritional parameters deteriorated progressively, with the serum ALB concentration decreasing from 37.2 ± 4.8 to 35.8 ± 5.2 g/L (P = 0.002), the prealbumin concentration increasing from 278 ± 82 to 254 ± 89 mg/L (P = 0.004), and the proportion of patients with malnutrition increasing from 48.2% to 61.9% (P < 0.001). Atherosclerosis indicators also worsened, with the mean CIMT increasing from 1.08 ± 0.31 to 1.18 ± 0.35 mm (P < 0.001) and the CIMT > 1.0 mm increasing from 56.9% to 68.5% (P = 0.003). Reflecting cumulative deterioration across all the components, the MIA syndrome score increased significantly from 5.8 ± 2.3 to 7.2 ± 2.6 points (P < 0.001).

Table 2.

Longitudinal changes in MIA syndrome components: overall and stratified by inflammatory status and dialysis modality

Parameters	Baseline	6 Months	12 Months	24 Months	P for Trend
A. Overall Cohort ( n  = 197)
Inflammatory Markers
CRP (mg/L)	7.7 ± 4.3	8.3 ± 4.6	8.8 ± 4.9	9.2 ± 5.1	< 0.001
IL-6 (pg/mL)	18.3 ± 10.1	19.5 ± 10.8	20.8 ± 11.6	21.8 ± 12.3	< 0.001
TNF-α (pg/mL)	21.8 ± 14.7	22.9 ± 15.3	24.1 ± 15.8	24.9 ± 16.2	< 0.001
Nutritional Parameters
Serum albumin (g/L)	37.2 ± 4.8	36.8 ± 5.0	36.2 ± 5.1	35.8 ± 5.2	0.002
Prealbumin (mg/L)	278 ± 82	271 ± 85	263 ± 87	254 ± 89	0.004
Malnutrition (SGA B/C, %)	48.2	52.8	57.9	61.9	< 0.001
Atherosclerosis Indicators
CIMT (mm)	1.08 ± 0.31	1.12 ± 0.32	1.15 ± 0.34	1.18 ± 0.35	< 0.001
CIMT > 1.0 mm (%)	56.9	60.4	64.5	68.5	0.003
MIA Syndrome Score (0–10)	5.8 ± 2.3	6.2 ± 2.4	6.8 ± 2.5	7.2 ± 2.6	< 0.001
B. Stratified by Inflammatory Status
High-Inflammation Group ( n  = 105)
CRP (mg/L)	11.2 ± 5.8	11.9 ± 6.2	12.5 ± 6.4	12.9 ± 6.5	< 0.001
Serum albumin (g/L)	35.1 ± 5.3	34.6 ± 5.5	34.3 ± 5.6	34.2 ± 5.6	0.012
CIMT (mm)	1.06 ± 0.29	1.12 ± 0.33	1.21 ± 0.36	1.28 ± 0.38	< 0.001
MIA Score	7.8 ± 2.2	8.1 ± 2.3	8.5 ± 2.3	8.7 ± 2.4	< 0.001
Low-Inflammation Group ( n  = 92)
CRP (mg/L)	3.8 ± 1.9	4.1 ± 2.1	4.2 ± 2.1	4.1 ± 2.0	0.187
Serum albumin (g/L)	39.5 ± 3.5	39.2 ± 3.7	38.5 ± 4.0	37.8 ± 4.3	0.003
CIMT (mm)	1.09 ± 0.32	1.12 ± 0.31	1.14 ± 0.32	1.16 ± 0.31	0.021
MIA Score	3.6 ± 1.7	4.2 ± 1.8	5.1 ± 1.9	5.9 ± 2.1	< 0.001
P value (between groups)	< 0.001	< 0.001	< 0.001	< 0.001	-
C. Stratified by Dialysis Modality
Hemodialysis ( n  = 149)
CRP (mg/L)	8.2 ± 4.4	8.9 ± 4.7	9.6 ± 5.0	10.1 ± 5.3	< 0.001
CRP growth rate (%)	-	-	-	22.8 ± 8.4	-
Serum albumin (g/L)	36.8 ± 5.1	36.2 ± 5.2	35.6 ± 5.4	35.0 ± 5.5	< 0.001
Albumin decline (g/L)	-	-	-	1.8 ± 0.9	-
MIA Score	6.1 ± 2.4	6.6 ± 2.5	7.2 ± 2.6	7.6 ± 2.5	< 0.001
Peritoneal Dialysis ( n  = 48)
CRP (mg/L)	6.1 ± 3.7	6.5 ± 3.9	6.8 ± 4.0	7.0 ± 4.1	0.008
CRP growth rate (%)	-	-	-	14.3 ± 6.2	-
Serum albumin (g/L)	38.4 ± 3.9	38.1 ± 4.0	37.8 ± 4.2	37.5 ± 4.3	0.032
Albumin decline (g/L)	-	-	-	0.9 ± 0.7	-
MIA Score	5.1 ± 2.0	5.4 ± 2.1	5.9 ± 2.2	6.2 ± 2.3	< 0.001
P value (between modalities)	0.011	0.008	0.002	0.001	-

Note: Values are the means ± SDs or percentages. P for trend was assessed by repeated-measures ANOVA or the Cochran‒Armitage test. High inflammation was defined as CRP > 8 mg/L, IL-6 > 20 pg/mL, or TNF-α > 25 pg/mL. CRP = C-reactive protein; IL-6 = interleukin-6; TNF-α = tumor necrosis factor-alpha; SGA = Subjective Global Assessment; CIMT = carotid intima-media thickness; MIA = malnutrition-inflammation-atherosclerosis

Stratification by inflammatory status revealed marked differences in progression trajectories. Compared with the low-inflammation group (n = 92), the high-inflammation group (n = 105) experienced more rapid deterioration in all MIA components. At 24 months, the high-inflammation group had significantly lower albumin levels (34.2 ± 5.6 vs. 37.8 ± 4.3 g/L, P < 0.001), more pronounced CIMT increases (1.06 ± 0.29 to 1.28 ± 0.38 mm vs. 1.09 ± 0.32 to 1.16 ± 0.31 mm, P < 0.001 for interaction), and substantially higher MIA scores (8.7 ± 2.4 vs. 5.9 ± 2.1, P < 0.001). Similarly, hemodialysis patients exhibited greater inflammatory marker increases (CRP growth rate: 22.8% vs. 14.3% in peritoneal dialysis, P = 0.007) and more marked nutritional deterioration (albumin decline: 1.8 ± 0.9 vs. 0.9 ± 0.7 g/L, P = 0.002), resulting in higher 24-month MIA scores (7.6 ± 2.5 vs. 6.2 ± 2.3, P = 0.001).

Associations between inflammatory markers and MIA syndrome components

At baseline, inflammatory markers were significantly correlated with both nutritional and atherosclerosis parameters (Table 3). CRP was strongly negatively correlated with serum ALB (r=-0.41, P < 0.001) and prealbumin (r=-0.27, P = 0.003) and positively correlated with CIMT (r = 0.36, P < 0.001) and LDL-C (r = 0.28, P = 0.003). Similarly, IL-6 and TNF-α correlated significantly with albumin (r=-0.35 and r=-0.32, respectively, both P < 0.001), prealbumin, and CIMT. These associations indicate that a greater inflammatory burden is linked to worse nutritional status and more advanced subclinical atherosclerosis. Notably, 56.9% of patients had a baseline CIMT > 1.0 mm, indicating a high prevalence of subclinical atherosclerosis in this population.

Table 3.

Correlations between inflammatory markers and Nutritional/Atherosclerosis indicators at baseline

Inflammatory markers	Serum albumin	Prealbumin	Total cholesterol	CIMT	LDL-C	HDL-C
CRP (mg/L)	-0.41 (< 0.001)	-0.27 (0.003)	-0.19 (0.021)	0.36 (< 0.001)	0.28 (0.003)	-0.22 (0.014)
IL-6 (pg/mL)	-0.35 (< 0.001)	-0.29 (0.002)	-0.21 (0.017)	0.31 (< 0.001)	0.33 (< 0.001)	-0.19 (0.032)
TNF-α (pg/mL)	-0.32 (< 0.001)	-0.24 (0.008)	-0.17 (0.039)	0.29 (0.001)	0.26 (0.004)	-0.18 (0.042)

Note: Values represent correlation coefficients (r) with P values in parentheses. P < 0.05 indicated a statistically significant correlation. CIMT = carotid intima‒media thickness; LDL-C = low-density lipoprotein cholesterol; HDL-C = high-density lipoprotein cholesterol

The strength of the inflammation‒nutrition coupling varied by dialysis modality (Table 4). Stratified analyses revealed more pronounced negative correlations between inflammatory markers and nutritional status in hemodialysis patients than in peritoneal dialysis patients (CRP-nutrition: r=-0.38 vs. r=-0.32, P < 0.001 and P = 0.009, respectively; similar patterns for IL-6 and TNF-α). Scatter plot analyses (Fig. 1) confirmed that the relationship between inflammatory markers and MIA syndrome severity was consistently stronger in hemodialysis patients, as evidenced by steeper regression slopes, suggesting that inflammatory markers correlate more strongly with overall MIA manifestation in the hemodialysis population.

Table 4.

Correlations between inflammatory markers and composite nutritional Scores, stratified by Dialysis modality

Inflammatory Markers	All Patients	Hemodialysis	Peritoneal Dialysis	P for Interaction
Composite Nutritional Score*
CRP (mg/L)	r = -0.38	r = -0.38	r = -0.32	0.041
	P < 0.001	P < 0.001	P = 0.009
IL-6 (pg/mL)	r = -0.33	r = -0.31	r = -0.24	0.038
	P < 0.001	P < 0.001	P = 0.021
TNF-α (pg/mL)	r = -0.28	r = -0.25	r = -0.20	0.045
	P < 0.001	P = 0.003	P = 0.045
MIA Syndrome Score
CRP and MIA score	r = 0.44***	r = 0.47***	r = 0.38**	0.028
IL-6 and MIA score	r = 0.39***	r = 0.42***	r = 0.34**	0.041
TNF-α and MIA score	r = 0.36***	r = 0.38***	r = 0.31*	0.052

Note: Composite nutritional score (0–4 points): one point each for albumin < 35 g/L, prealbumin < 200 mg/L, SGA B/C, and BMI < 23 kg/m². Higher scores indicate worse nutrition. For individual parameter correlations (e.g., CRP-albumin), see Table 3. Significance: *P < 0.05, **P < 0.01, ***P < 0.001

Fig. 1.

Correlations between inflammatory markers and MIA syndrome scores according to dialysis modality. Inflammatory markers (CRP, IL-6, and TNF-α) correlate more strongly with MIA syndrome severity in hemodialysis patients (solid circles/lines) than in peritoneal dialysis patients (open circles/dashed lines), as evidenced by steeper regression slopes. HD: hemodialysis; PD: peritoneal dialysis; MIA: malnutrition-inflammation-atherosclerosis

Inflammatory markers as predictors of MIA syndrome severity and cardiovascular events

Patients were stratified into high-inflammation (n = 105, 53.3%) and low-inflammation groups (n = 92, 46.7%) on the basis of prespecified thresholds. The high-inflammation group had significantly higher baseline MIA scores (7.8 ± 2.2 vs. 5.6 ± 1.9 points, P < 0.001) and substantially greater cardiovascular event incidence during follow-up (53.3% vs. 30.4%, P < 0.001). Overall, 84 patients (42.6%) experienced cardiovascular events, including myocardial infarction (n = 30, 35.7%), heart failure (n = 25, 29.8%), stroke (n = 19, 22.6%), and angina (n = 10, 11.9%).

Cox proportional hazards regression demonstrated that inflammatory markers independently predicted cardiovascular events after comprehensive adjustment for age, sex, dialysis characteristics, comorbidities, hemoglobin, albumin, and residual kidney function (Table 5). Each 1 mg/L increase in CRP was associated with a 17% increase in cardiovascular risk (adjusted HR = 1.17, 95% CI: 1.08–1.26, P < 0.001); similarly, each 10 pg/mL increase in IL-6 (HR = 1.12, 95% CI: 1.04–1.20, P = 0.002) and TNF-α (HR = 1.14, 95% CI: 1.05–1.23, P = 0.001) independently predicted events. When analyzed categorically, a high inflammation status was associated with an 89% increased cardiovascular risk (HR = 1.89, 95% CI: 1.34–2.67, P < 0.001). The fully adjusted model demonstrated good discrimination (C-index = 0.73, 95% CI: 0.68–0.78).

Table 5.

Cox proportional hazards regression analysis: inflammatory markers as predictors of cardiovascular events

Variables	Events/Total (%)	Univariable		Multivariable*
		HR (95% CI)	P value	HR (95% CI)	P value
Inflammatory Markers (Continuous)
CRP (per 1 mg/L)	-	1.21 (1.14–1.29)	< 0.001	1.17 (1.08–1.26)	< 0.001
IL-6 (per 10 pg/mL)	-	1.18 (1.11–1.26)	< 0.001	1.12 (1.04–1.20)	0.002
TNF-α (per 10 pg/mL)	-	1.19 (1.12–1.27)	< 0.001	1.14 (1.05–1.23)	0.001
Inflammatory Status (Categorical)
High-inflammation†	56/105 (53.3%)	2.34 (1.68–3.26)	< 0.001	1.89 (1.34–2.67)	< 0.001
Low-inflammation	28/92 (30.4%)	1.00 (reference)	-	1.00 (reference)	-
Dialysis Modality
Hemodialysis	69/149 (46.3%)	1.73 (1.21–2.48)	0.003	1.26 (0.87–1.83)	0.218
Peritoneal dialysis	15/48 (31.3%)	1.00 (reference)	-	1.00 (reference)	-
Other Clinical Predictors
Age (per 10 years)	-	1.34 (1.18–1.52)	< 0.001	1.28 (1.12–1.47)	< 0.001
Diabetes mellitus	52/107 (48.6%)	1.89 (1.41–2.54)	< 0.001	1.52 (1.11–2.09)	0.009
Baseline CVD	48/78 (61.5%)	2.12 (1.58–2.84)	< 0.001	1.73 (1.27–2.36)	< 0.001
Serum albumin (per g/L)	-	0.94 (0.91–0.97)	< 0.001	0.96 (0.93–0.99)	0.012

Note: *Multivariable models adjusted for age, sex, dialysis duration, dialysis modality, diabetes mellitus, cardiovascular disease, hemoglobin, serum albumin, and residual kidney function. †High inflammation was defined as CRP > 8 mg/L, IL-6 > 20 pg/mL, or TNF-α > 25 pg/mL. The fully adjusted model (high-inflammation categorical) demonstrated good discrimination (C-index = 0.73, 95% CI: 0.68–0.78). Total cardiovascular events: 84/197 (42.6%). HR = hazard ratio; CI = confidence interval; CVD = cardiovascular disease

K‒M analysis confirmed significantly lower cardiovascular event-free survival in the high-inflammation group, with estimated 24-month event-free survival rates of 46.7% versus 69.6% in the low-inflammation group (log-rank P < 0.001) (Fig. 2).

Fig. 2.

Kaplan‒Meier survival curves of cardiovascular event incidence in patients with different inflammation levels. This figure confirmed significantly lower cardiovascular event-free survival in the high-inflammation group, with estimated 24-month event-free survival rates of 46.7% versus 69.6% in the low-inflammation group (log-rank P < 0.001)

Dynamic inflammatory changes and MIA syndrome progression

Longitudinal analysis revealed that inflammatory marker trajectories were significantly associated with MIA syndrome progression. Compared with the low-inflammation group, the high-inflammation group presented increased inflammatory marker levels over 24 months (CRP growth rate: 15.3 ± 6.7% vs. 7.2 ± 3.9%, P < 0.001; IL-6: 13.8 ± 5.9% vs. 6.5 ± 3.7%, P < 0.001; TNF-α: 12.6 ± 5.4% vs. 5.8 ± 3.4%, P < 0.001). Importantly, the percentage changes in inflammatory markers from baseline to 24 months correlated significantly with concurrent changes in MIA scores (composite inflammatory change: r = 0.37, P < 0.001; individual markers: CRP r = 0.34, IL-6 r = 0.31, TNF-α r = 0.29, all P ≤ 0.001). In the multivariable linear regression adjusted for baseline factors, the inflammatory marker growth rate remained an independent predictor of MIA syndrome progression (standardized β = 0.32, P < 0.001), explaining an additional 8.4% of the variance beyond the baseline characteristics.

Dialysis modality, inflammatory burden, and cardiovascular outcomes

Compared with peritoneal dialysis patients, hemodialysis patients presented a greater incidence of cardiovascular events (46.3% vs. 29.2%, P = 0.027). Mediation analysis revealed that this association was substantially attenuated after adjusting for inflammatory markers (Fig. 3). According to the unadjusted models, hemodialysis was associated with 73% increased cardiovascular risk (HR = 1.73, 95% CI: 1.21–2.48; P = 0.003). However, after adjusting for CRP, IL-6, and TNF-α, the hazard ratio decreased to 1.26 (95% CI: 0.87–1.83, P = 0.218), with the confidence interval now including 1.0. Formal mediation analysis demonstrated that inflammatory markers collectively mediated approximately 54% of the total effect of dialysis modality on cardiovascular events (indirect effect: HR = 1.37, 95% CI: 1.12–1.67, P = 0.002), with CRP showing the strongest mediating effect (β = 0.31, P < 0.001), followed by IL-6 (β = 0.28, P = 0.002) and TNF-α (β = 0.25, P = 0.004). These findings suggest that a substantial portion of the increased cardiovascular risk in hemodialysis patients may be attributable to their increased inflammatory burden.

Fig. 3.

Mediation analysis: Inflammatory markers mediate the relationship between dialysis modality and cardiovascular outcome. Hazard ratios comparing hemodialysis (HD) to peritoneal dialysis (PD) for cardiovascular events across sequential adjustment models. The HR decreased from 1.73 (P = 0.003) to 1.26 (P = 0.218) after adjusting for inflammatory markers, which collectively accounted for 54% of the total effect. *Demographics: age, sex, dialysis vintage. †Comorbidities: diabetes, hypertension, and baseline CVD

Gender differences and sensitivity analyses

Subgroup analysis by sex revealed stronger associations between inflammatory markers and both MIA syndrome severity and cardiovascular risk in male patients than in female patients (inflammatory-MIA correlation: male r = 0.43 vs. female r = 0.35, P for interaction = 0.041; cardiovascular risk per SD increase in inflammation: male HR = 1.22 vs. female HR = 1.14, P for interaction = 0.038), suggesting potential sex-specific differences in inflammatory pathogenesis.

To verify the robustness of the results, sensitivity analyses were performed in patients without baseline diabetes or cardiovascular disease (n = 68). The associations between inflammatory markers and MIA scores remained significant (CRP: r = 0.29, P < 0.001; IL-6: r = 0.27, P = 0.003; TNF-α: r = 0.31, P < 0.001), and inflammatory markers independently predicted cardiovascular events (CRP: HR = 1.15, 95% CI: 1.04–1.27, P = 0.006) (Table 6). Additional sensitivity analyses using alternative inflammatory classification methods (quartile-based, median-based, requiring ≥ 2 elevated markers) and excluding outliers yielded consistent results, confirming the reliability of our findings.

Table 6.

Sensitivity analysis: associations between inflammatory markers and clinical outcomes in patients without baseline diabetes or cardiovascular disease

Analysis	All Patients	Sensitivity Cohort*	Comparison
Inflammatory Markers and MIA Score
CRP (r, P value)	r = 0.44, P < 0.001	r = 0.29, P < 0.001	Consistent
IL-6 (r, P value)	r = 0.39, P < 0.001	r = 0.27, P = 0.003	Consistent
TNF-α (r, P value)	r = 0.36, P < 0.001	r = 0.31, P < 0.001	Consistent
Inflammatory Markers and Serum Albumin
CRP (r, P value)	r = -0.41, P < 0.001	r = -0.34, P = 0.005	Consistent
IL-6 (r, P value)	r = -0.35, P < 0.001	r = -0.29, P = 0.017	Consistent
TNF-α (r, P value)	r = -0.32, P < 0.001	r = -0.26, P = 0.032	Consistent
Predictive Value for Cardiovascular Events
Cardiovascular event rate:
High-inflammation group	56/105 (53.3%)	14/34 (41.2%)	Lower but significant
Low-inflammation group	28/92 (30.4%)	8/34 (23.5%)	Consistent pattern
Cox regression (adjusted HR†):
High-inflammation status	1.89 (1.34–2.67)	2.14 (1.21–3.78)	Consistent
P value	P < 0.001	P = 0.009
CRP (per 1 mg/L)	1.17 (1.08–1.26)	1.15 (1.04–1.27)	Consistent
P value	P < 0.001	P = 0.006

Note: *The sensitivity cohort excludes patients with baseline diabetes or cardiovascular disease (n = 68 of 197). †Adjusted for age, sex, dialysis characteristics, hemoglobin, albumin, and residual kidney function. “Consistent” indicates that associations remained significant with similar magnitudes. Additional analyses excluding outliers and using alternative thresholds yielded consistent results. MIA = malnutrition-inflammation-atherosclerosis; HR = hazard ratio; CI = confidence interval

Discussion

This 24-month longitudinal study demonstrated that elevated inflammatory markers (CRP, IL-6, and TNF-α) are significantly associated with MIA syndrome severity, progression, and cardiovascular events in maintenance dialysis patients. Our key findings include (1) strong associations between the inflammatory burden and both nutritional deterioration and atherosclerosis progression; (2) temporal relationships between longitudinal inflammatory changes and accelerated MIA syndrome progression; (3) higher inflammatory levels and cardiovascular risk in hemodialysis patients than in peritoneal dialysis patients; and (4) substantial mediation of the dialysis modality‒outcome relationship by the inflammatory burden. These findings underscore the clinical importance of inflammatory monitoring in dialysis patients and suggest that inflammation is a potential therapeutic target, although interventional studies are needed to establish causality.

Inflammatory markers and MIA syndrome: clinical significance and biological plausibility

Our results demonstrate strong associations between inflammatory markers and MIA syndrome components, with important clinical implications. The observed correlations between CRP and nutritional parameters (r=-0.41 with albumin) and atherosclerosis indicators (r = 0.36 with CIMT) align with findings from earlier studies [30–32]. While previous investigations established the prognostic value of CRP and IL-6 for all-cause mortality in dialysis populations [23], our study extends these findings by demonstrating through multivariable Cox regression that these markers independently predict cardiovascular events—the leading cause of death in this population [1–6]. The strength of the association (hazard ratios of 1.12–1.17 per unit increase) is consistent with meta-analyses in general CKD populations [22, 33], supporting the robustness of the results. However, the moderate correlation coefficients (r = 0.29–0.41) indicate that inflammation explains only a portion of the MIA syndrome variance, highlighting its multifactorial nature.

From a mechanistic perspective, inflammation may be linked to MIA syndrome via interconnected pathways, although our study did not directly assess these mechanisms. The directionality of these relationships remains to be established. Oxidative stress has been associated with inflammatory signaling via NF-κB activation [34, 35], while inflammatory cytokines (IL-6, TNF-α) correlate with reactive oxygen species production, suggesting potential bidirectional relationships. These interconnected processes may be associated with endothelial dysfunction, including reduced nitric oxide bioavailability and vascular injury, which is consistent with our observed CRP–CIMT correlation [36].

Our longitudinal data showing a correlation between inflammatory changes and MIA score progression (r = 0.37, P < 0.001) demonstrate temporal associations between inflammatory markers and syndrome deterioration, although whether inflammation reflects, precedes, or results from disease progression cannot be determined from our observational design. However, our observational design cannot establish directionality. Bidirectional relationships are likely–malnutrition and vascular dysfunction may be associated with increased inflammation. Only interventional studies targeting inflammatory pathways can definitively establish causality and determine whether reducing inflammation improves outcomes.

Dialysis Modality, inflammatory Burden, and clinical outcomes

An important finding is the differential inflammatory burden between dialysis modalities, which appears to contribute substantially to outcome differences. Our observation of higher CRP in hemodialysis patients than in peritoneal dialysis patients (8.2 ± 4.4 vs. 6.1 ± 3.7 mg/L, P = 0.011) aligns with previous reports [37, 38]. Notably, mediation analysis revealed that after adjusting for inflammatory markers, the association between dialysis modality and cardiovascular events was substantially attenuated (HR decreased from 1.73 to 1.26, with loss of significance), with inflammation mediating approximately 54% of the total effect. These findings suggest that a considerable portion of hemodialysis patients’ increased cardiovascular risk may be attributable to a higher inflammatory burden rather than to the dialysis modality per se.

Several mechanisms have been hypothesized to explain this differential inflammatory response. Hemodialysis has been associated with more robust inflammatory activation, possibly related to complement activation from blood–membrane contact, exposure to dialysis membranes with varying biocompatibility, recurrent ischemia–reperfusion events, and hemodynamic stress observed in extracorporeal circulation [39–41]. Additionally, the intermittent nature of hemodialysis, with larger fluctuations in volume status and uremic toxin levels, has been associated with sustained inflammation. Conversely, peritoneal dialysis provides continuous ultrafiltration and solute removal without direct blood–membrane contact, which may be associated with more stable hemodynamics and less complement activation, although it has its own inflammatory considerations, including glucose degradation products and peritonitis risk [17]. Our finding that IL-6 and TNF-α levels did not differ significantly between modalities (P = 0.121 and P = 0.174) suggests that different inflammatory pathways may be preferentially activated by different techniques.

The stronger inflammation–malnutrition coupling in hemodialysis patients (CRP-nutrition: r=-0.38 vs. r=-0.32 in peritoneal dialysis) suggests potential modality-specific differences. The more pronounced inflammatory response during hemodialysis correlates more strongly with protein-energy metabolism abnormalities, which have been attributed to cytokine-mediated anorexia, increased catabolism, and impaired synthesis [42], potentially influenced by intermittent uremic toxin exposure and intradialytic amino acid losses. These findings suggest that hemodialysis patients may benefit from more aggressive inflammatory monitoring and nutritional support, particularly those with elevated inflammatory markers. While our observational data cannot guide modality prescription, they suggest that strategies to reduce the inflammatory burden—including more biocompatible membranes, optimized dialysis prescriptions, or adjunctive anti-inflammatory therapies—might improve hemodialysis outcomes, although this requires prospective validation.

Temporal dynamics and gender considerations

The divergent inflammatory marker trajectories between the high- and low-inflammation groups (CRP growth rate: 15.3% vs. 7.2%, P < 0.001) suggest that chronic inflammation is associated with progressive disease deterioration. The correlation between inflammatory changes and MIA score progression (r = 0.37, P < 0.001), which remained significant after adjusting for baseline factors, suggests that ongoing inflammatory activity is temporally associated with syndrome progression, although the causal direction remains uncertain. However, alternative explanations include shared upstream drivers (residual function decline, dialysis inadequacy, infections), reverse causation (worsening malnutrition promoting inflammation), or bidirectional relationships forming vicious cycles. Previous studies have predominantly used single-time point measurements [10–12]; our multitime point data provide stronger evidence for active associations but cannot definitively establish directionality. Importantly, regular inflammatory marker monitoring in high-risk patients may facilitate identification of earlier intervention opportunities, although whether interventions reducing inflammation (through therapies, improved dialysis, or management of modifiable triggers such as infections and inadequate dialysis) are associated with slowed MIA progression requires testing in interventional trials.

Our subgroup analysis revealed stronger inflammation–MIA associations and cardiovascular risk prediction in male patients than in female patients (males: r = 0.43, HR = 1.22; females: r = 0.35, HR = 1.14; P for interaction < 0.05), consistent with potential hormonal modulation of immune responses. Estrogen has been associated with anti-inflammatory properties related to NF-κB inhibition and cytokine modulation [43], whereas androgens have been linked to certain inflammatory pathways. Additionally, sex differences in body composition may influence the effects of inflammation on metabolism. These findings suggest potential value in sex-specific risk stratification and management strategies, with male patients potentially benefiting from more aggressive inflammatory monitoring, although these exploratory findings require validation in larger cohorts.

Clinical implications

Our findings have several practical implications. First, inflammatory markers (particularly CRP, IL-6, and TNF-α) may serve as useful biomarkers for identifying high-risk patients, with our prediction model demonstrating good discrimination (C-index = 0.73). Regular monitoring could complement traditional cardiovascular risk assessment. Second, while observational data cannot guide modality prescription, the finding that inflammation statistically accounts for a substantial portion of modality–outcome relationships suggests that the inflammatory burden should be considered when evaluating comparative risks and benefits for individual patients. Third, our findings suggest potential therapeutic avenues, including dialysis optimization (biocompatible membranes, high-flux dialysis), the management of inflammatory triggers (vascular access infections, catheter use), nutritional interventions with anti-inflammatory components, and pharmacological therapies, although such approaches must be balanced against infection risks in immunocompromised patients. Fourth, our MIA scoring system, pending external validation, provides a framework for holistic assessment, facilitating longitudinal monitoring and individualized care.

Strengths, limitations, and future directions

This study has several methodological strengths. The longitudinal design with systematic four-time point data collection enabled temporal relationship assessment and comprehensive evaluation of all three MIA syndrome components. Formal mediation analysis was used to examine the effects of dialysis modality, and multiple inflammatory markers captured different pathway activations. Rigorous outcome adjudication by an independent blinded committee minimized misclassification, and extensive sensitivity analyses confirmed the robustness of the results across various analytical approaches.

Several limitations warrant consideration. The single-center design and potential unmeasured confounding factors necessitate multicenter validation with broader covariate assessment. More fundamentally, our observational design precludes causal inference—while temporal associations suggest that inflammation is associated with MIA progression, directionality remains uncertain, and bidirectional relationships are plausible. Only randomized trials targeting inflammation can establish causality and determine therapeutic efficacy.

Our MIA scoring system requires external validation with outcome-based component weighting. Future priorities include mechanistic studies elucidating inflammatory pathways, extended follow-up capturing late outcomes, genetic profiling enabling risk stratification, and large-scale biomarker validation supporting precision medicine approaches in dialysis populations.

Conclusion

This study demonstrated that inflammatory markers are independently associated with MIA syndrome severity, progression, and cardiovascular events in dialysis patients. Hemodialysis patients presented a greater inflammatory burden than did peritoneal dialysis patients, with inflammation substantially mediating this difference. While causality cannot be established from our observational design, these findings highlight the clinical importance of inflammatory monitoring as a potential therapeutic target, warranting interventional validation.

Author contributions

Sha Chen conceptualized the study, drafted the manuscript, and performed the statistical analysis. Ping Yang and Hongyun Lv were responsible for data acquisition and the preparation of tables and figures. Weiwei Wang, Qingxia Zhang, and Hong Pan handled manuscript typesetting and graphic enhancements. Shuhan Yu provided critical review and revision of the manuscript. All authors participated in the review process and have read and approved the final submitted manuscript.

Funding

This study was not funded.

Data availability

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Ethics Committee of Dongyang Hospital of Traditional Chinese Medicine. All methods were in accordance with the Helsinki Declaration and its contemporary amendments. All patient-related information was approved by the patient or guardian published.

Consent for publication

Written informed consent for publication was obtained from all patients included in this study.

Competing interests

The authors declare no competing interests.