ABSTRACT
Background
Chronic hemodialysis (HD) patients exhibit severe morpho-functional cardiac alterations, putting them at a high risk of death and adverse cardiovascular (CV) outcomes. Despite the fact that an unbalanced expression of various microRNAs (miRNAs) has been related to pathological cardiac remodeling and worse CV outcomes, scarce evidence exists on their role in this setting.
Methods
We evaluated circulating levels of a selected miRNAs panel (30a-5p, 23a-3p, 451a and let7d-5p) in 74 chronic HD patients together with a thorough clinical and echocardiography assessment. Individuals were then prospectively followed (median 22 months). The primary endpoint was a composite of all-cause and CV mortality and non-fatal CV events.
Results
Circulating levels of all miRNAs were lower in HD patients as compared with healthy controls and independently correlated to the severity of cardiac dysfunction. miRNA 30a-5p, 23a-3p and 451a expression was even lower in 30 subjects (40.5%) reaching the composite endpoint (P < .001), while no differences were reported for let7d-5p. The predictive value of these miRNAs was supported by univariate followed by multivariate Cox regression analyses [hazard ratio (HR) ranging from 0.943 to 0.995; P = .05 to .02] while Kaplan–Meier analyses confirmed a faster progression to the endpoint in individuals displaying miRNA levels below an optimal receiver operating characteristic–derived cut-off value (P ranging from .001 to <.0001; crude HRs 7.95 to 8.61).
Conclusions
Lower circulating levels of miRNA 30-5p, 23a-3p and 451a in HD patients may reflect cardiac abnormalities and predict a higher risk of worse clinical outcomes in the short mid-term. Future studies on larger HD populations are needed to generalize these findings.
Keywords: cardiovascular events, hemodialysis, miRNA, mortality
Graphical Abstract
Graphical Abstract.
INTRODUCTION
End-stage kidney disease (ESKD) patients on chronic hemodialysis (HD) treatment exhibit a remarkable risk of worse cardiovascular (CV) outcomes [1, 2]. In these subjects, CV mortality accounts for 40%–50% of all deaths, compared with ∼25% in the general population [3], and roughly half of HD individuals are expected to experience at least one CV event during their lifetime [4]. Cardiac alterations in ESKD largely impact upon this exceeding risk. Up to 70%–80% of chronic HD patients display abnormal concentric or eccentric left ventricular hypertrophy (LVH) which reflects a maladaptive response to increased vascular resistances, chronic volume overload and sustained uremic toxicity [5]. Cardiac fibrosis due to collagen deposition between capillaries and cardiomyocytes often overlays LVH, leading to ventricular wall dilatation and diastolic dysfunction, a threatening condition known as “uremic cardiomyopathy” whose prevalence among individuals on maintenance HD could be as high as 75% [6].
In recent years, experimental and clinical research has flourished at an ever-increasing pace on investigating the biological roles of microRNAs (miRNAs).
miRNAs represent an evolutionarily conserved family of small, non-coding RNAs which regulate gene expression at the post-transcriptional level [7]. miRNAs abound both in the nucleus and in the cytoplasm and regulate a variety of essential cellular processes including development, differentiation, proliferation, metabolism and stress responses. In addition, being released in the bloodstream and in other body fluids, circulating miRNAs can also target the function of remote organs which, in turn, can modulate their expression [7]. Hence, not surprisingly, an aberrant regulation of various miRNAs has been implicated in the pathogenesis of several neoplastic, inflammatory, neurodegenerative and CV disorders, also representing a potential target for therapeutic interventions [8].
At the cardiac level, an unbalanced expression of various miRNAs may trigger asymmetric ventricular hypertrophy, cardiac fibrosis and myocardial apoptosis after acute heart damage or in the setting of systemic diseases [9]. Furthermore, in patients with acute coronary syndromes or heart failure, measurement of circulating miRNAs levels may reflect incipient or more severe complications, thereby allowing early disease detection and outcome prediction [10].
Yet, despite the fact that abnormal morpho-functional changes at the cardiac level are largely pervasive among ESKD patients, very scant evidence is available elucidating possible associations between a deranged miRNAs expression, the severity of cardiac dysfunction and worse clinical outcomes in this setting.
In this pilot study, we investigated in a multicentric cohort of ESKD individuals on maintenance HD the clinical significance and the diagnostic and prognostic capacity of a small panel of circulating miRNAs with respect to adverse clinical outcomes, namely death and CV events.
To the best of our knowledge, this is the first report in the literature documenting that an aberrant expression profile of miRNA 30a-5p, 23a-3p and 451a in HD patients may impart significant prognostic information to refine outcome prediction in such a very high-risk setting.
MATERIALS AND METHODS
Study cohort
We conducted an observational, prospective, multicenter cohort study on 74 chronic prevalent HD patients undergoing regular dialysis at the “Mater-Domini” University hospital of Catanzaro, Italy (n = 28), the University Hospital of Ioannina, Greece (n = 23) and the General Hospital of Filiates, Greece (n = 23). All patients followed a standard 3 times/week dialysis schedule, had a stable dry weight, an unchanged therapeutic scheme for at least 3 months before entering the study and had achieved a normotensive edema-free state. Exclusion criteria were the following: dialysis vintage <6 months, recent hospitalization (<1 month) for any CV event, acute or non-intermittent HD, recent switch from peritoneal dialysis or renal transplantation, presence or a recent history of bleeding, morbid obesity (body mass index >40 kg/m2), malignancy, liver, thyroid or infectious diseases, alterations in leucocyte count or differential, and/or treatment with steroids or immunosuppressors. Fourteen healthy matched subjects served as controls for miRNAs evaluation. The local Ethic Committees (Comitato Etico Regione Calabria-Area Centro n.397/2020, Ethical Committee of the University Hospital of Ioannina, 25/5/21-12-2020) approved the study and a fully informed consent was obtained from all participants.
Clinical and laboratory assessment
Anthropometric, clinical and dialysis parameters were recorded at baseline before starting a mid-week dialysis session. Adequacy of dialysis was assessed using Kt/V, calculated as the natural logarithm of the ratio between initial and final urea concentration. Common biochemical data were measured, according to standard methods used in the routine clinical laboratory. Blood pressure was assessed using an automated sphygmomanometer integrated in the dialysis machine and measured three times consecutively before dialysis start, with the average values being considered for data analysis. Echocardiographic measurements were carried by a skilled operator unaware of patients’ clinical and biochemical characteristics. Key parameters for estimating ventricular function and morphology and for cardiac chamber quantification were recorded as recommended [11].
Selection and measurement of candidate miRNAs
A thorough systematic literature review was performed in search of experimental and/or clinical evidence of miRNAs (i) being involved in pathological cardiac remodeling in response to heart diseases or systemic cardio-metabolic alterations, (ii) holding a documented prognostic value for CV outcomes in at risk populations and (iii) possibly being involved in stress responses to acute or chronic kidney damage or in the cardio-renal axis. The results yielded pointed at miRNA 30a-5p [12–16], miRNA 23a-3p [17–20], miRNA 451a [21–26] and miRNA let7d-5p [16, 27–31] as candidates for answering the opening research question. Under physiological conditions, all the four candidate miRNAs have previously been implicated in the development of various organs and tissues, including the heart, the kidney, the brain and the vascular system. In addition, these miRNAs play also a relevant role in regulating various immune and metabolic processes, such as bone mineralization and lipid control [8]. Laboratory procedures for miRNAs extraction and evaluation are detailed in Supplementary Appendix 1.
Prospective follow-up and study endpoint
After the baseline assessment, patients were followed until the end of the established observation period or the study endpoint was reached. This latter was defined as a combined outcome of all-cause mortality, CV mortality and non-fatal CV events including coronary, cerebrovascular or peripheral artery disease events, acute heart decompensation or severe cardiac arrythmia episodes requiring hospitalization.
Statistical analysis
The statistical analysis was performed using the SPSS package (version 24.0.0.0; IBM corporation), the MedCalc Statistical Software (version 14.8.1; MedCalc Software bvba) and the GraphPad Prism software (version 9.0.0; GraphPad Software LLC). Differences between groups were determined by the unpaired t-test for normally distributed values, the Mann–Whitney U test for non-parametric values and the chi-square followed by a Fisher's exact test for frequency distributions. Correlation analyses were made by assessing the Pearson (R) or the Spearman (Rho) coefficients, as appropriate. Before testing correlations, all variables showing a skewed distribution were log-transformed to approximate Gaussian distribution. Multiple regression analyses to assess independent relationships were performed by building separate models with each miRNA as the dependent variable and including the respective univariate correlates at baseline. Data were expressed as partial correlation coefficients (β) and P-value. Univariate followed by multivariate Cox proportional hazard regression analyses were employed to evaluate time-dependent associations with the outcome for variables which differed at baseline among study subgroups. Receiver operating characteristics (ROC) analyses were performed for miRNAs eventually associated with the composite endpoint, computing the areas under the curve (AUCs) and the best cut-off value (Youden Index) for identifying patients experiencing the outcome of interest. In addition, a cumulative AUC was elaborated from the individual predicted probabilities obtained from a multivariate logistic regression model including single predictive miRNAs as covariates. Kaplan–Meier curves were generated for patients with miRNAs expression above or below the optimal, ROC-derived thresholds and compared by a Log-rank test. All endpoint analyses were conducted on a time-to-first event basis. Results were considered significant if the P-value was ≤.05.
RESULTS
Characteristics of the study population
Tables 1 and 2 summarize the main characteristics of the study cohort. Mean ± standard deviation age was 72.5 ± 12.5 years and the majority of them were male (75.7%). Median dialysis vintage was 35.5 months (interquartile range 17–68.2) with a satisfactory dialysis adequacy on average (mean Kt/V 1.44 ± 0.27). More than a half of patients (55.4%) were on standard bicarbonate dialysis while the remainder were on hemodiafiltration. Twenty-seven percent of patients (n = 20) had diabetes and roughly half (48.6%, n = 36) had previous history of any CV disease.
Table 1:
Main clinical characteristics in all HD patients and in individuals categorized according to the occurrence of the composite endpoint.
| All HD patients N = 74 | Endpoint—no N = 44 | Endpoint—yes N = 30 | P | |
|---|---|---|---|---|
| Age (years) | 68.6 ± 12.3 | 66.6 ± 11.9 | 72.5 ± 12.5 | .04 |
| Males, n (%) | 56 (75.7) | 33 (75) | 23 (76.7) | .90 |
| Dry weight (kg) | 71 ± 15.3 | 72.1 ± 16.8 | 71.7 ± 13.2 | .92 |
| BMI (kg/m2) | 25.4 ± 4.9 | 26.2 ± 5.2 | 25.1 ± 3.8 | .76 |
| Waist–hip ratio (cm) | 0.94 ± 0.12 | 0.92 ± 0.23 | 0.95 ± 0.19 | .69 |
| Type of dialysis | ||||
| HD, n (%) | 41 (55.4) | 20 (45.4) | 21 (70) | .64 |
| Hemodiafiltration, n (%) | 33 (44.6) | 24 (54.5) | 9 (30) | .06 |
| Kt/V | 1.44 ± 0.27 | 1.46 ± 0.27 | 1.40 ± 0.28 | .29 |
| History of any CV disease | 36 (48.6) | 12 (27.3) | 18 (60) | .08 |
| AMI/angina | 26 (35.1) | 14 (31.8) | 12 (40) | .59 |
| Stroke/TIA | 10 (13.5) | 3 (6.8) | 7 (23.3) | .08 |
| Peripheral vasculopathy | 25 (33.8) | 14 (31.8) | 11 (36.6) | .80 |
| Dialysis vintage (months) | 35.5 (17–68.2) | 35.5 (21.2–66.7) | 40 (15.5–71.2) | .76 |
| Diabetes, n (%) | 20 (27) | 11 (25) | 9 (30) | .79 |
| Glycemia (mg/dL) | 113 ± 49.4 | 125 ± 42 | 140.5 ± 58.2 | .18 |
| Hemoglobin (g/dL) | 10.9 ± 1.1 | 11.1 ± 1.15 | 10.7 ± 1.1 | .11 |
| SBP (mmHg) | 139.6 ± 23.8 | 141.4 ± 19.4 | 136.9 ± 29.3 | .42 |
| DBP (mmHg) | 71.4 ± 12.2 | 73.7 ± 10.2 | 58 ± 14.2 | .04 |
| Pulse pressure (mmHg) | 68.1 ± 26.8 | 57.5 ± 23.2 | 76.8 ± 25.6 | .03 |
| Serum creatinine (mg/dL) | 8.2 (6.9–9.2) | 8.9 (7.5–9.9) | 8.60 (7–10.64) | .11 |
| Urea (mg/dL) | 133.1 ± 34 | 132.2 ± 32 | 134.9 ± 32.8 | .58 |
| Sodium (mg/dL) | 136.3 ± 15.2 | 135.6 ± 19.6 | 137.4 ± 3.1 | .60 |
| Potassium (mg/dL) | 4.97 ± 0.68 | 5.01 ± 0.65 | 4.91 ± 0.73 | .52 |
| Phosphate (mg/dL) | 4.7 ± 1.29 | 4.75 ± 1.24 | 4.65 ± 1.38 | .74 |
| Calcium (mg/dL) | 9.23 ± 0.64 | 9.30 ± 0.69 | 9.12 ± 0.57 | .26 |
| iPTH (pg/mL) | 251.9 (145.9–386) | 207 (129–255) | 239.8 (124.5–358.9) | .59 |
| Uric acid (mg/dL) | 5.8 ± 1.1 | 5.8 ± 1.04 | 5.82 ± 1.15 | .92 |
| Albumin (g/dL) | 3.97 ± 0.24 | 3.99 ± 0.19 | 3.94 ± 0.31 | .41 |
| CK-MB (UI/L) | 21 ± 8.1 | 18.8 ± 6.6 | 24.2 ± 9.2 | .02 |
| hs-cTnI (pg/mL) | 21.4 (11.6–61.7) | 19.3 (7.7–22.2) | 23.1 (20.6–53.6) | .04 |
| ALP (U/L) | 79 (62.7–88.5) | 79 (62–87) | 85.5 (67.5–92) | .23 |
| Total cholesterol (mg/dL) | 147.5 ± 45.1 | 141.8 ± 28.9 | 155.8 ± 61.3 | .19 |
| HDL (mg/dL) | 41.6 ± 10.1 | 40.9 ± 9.3 | 42.5 ± 11.3 | .50 |
| LDL (mg/dL) | 83.6 ± 33 | 77.9 ± 25.8 | 91.9 ± 30.5 | .05 |
| Triglycerides (mg/dL) | 126 (85.5–194.5) | 127 (99–173) | 140 (90.5–265.7) | .61 |
| ESR (mm/h) | 29 (20–44.7) | 33 (25–47) | 29 (25.2–45.2) | .36 |
| C-reactive protein (mg/L) | 0.81 (0.23–3.23) | 0.30 (0.12–0.52) | 0.61 (0.11–0.93) | .23 |
| Fibrinogen (mg/dL) | 421.6 ± 100.9 | 414.8 ± 60.2 | 431.6 ± 54 | .58 |
| RBC (n × 103) | 3.70 ± 0.73 | 3.71 ± 0.50 | 3.69 ± 0.99 | .88 |
| WBC (n × 103) | 7.01 ± 3.30 | 7.17 ± 2.9 | 6.77 ± 1.94 | .60 |
| PLT (n × 103) | 219.2 ± 72.1 | 216.7 ± 75.6 | 222.8 ± 67.6 | .72 |
| Serum iron (mg/dL) | 70.9 ± 25.2 | 72.2 ± 25.9 | 69 ± 24.2 | .59 |
| TSAT (%) | 28 (20.2–34.3) | 28.2 (22.1–32.1) | 28.9 (23.1–34.5) | .88 |
| Ferritin (µg/L) | 197.9 (103–318) | 178.3 (92–251) | 258.1 (129.7–341.8) | .32 |
Statistically significant differences between subgroups are highlighted in bold.
Data are presented as mean ± standard deviation, median (interquartile range) or n (%).
ALP, alkaline phosphatase; AMI: acute myocardial infarction; BMI, body mass index; DBP, diastolic blood pressure; ESR, erythrocyte sedimentation rate; hs-cTnI: high-sensitivity c-Troponin I; PLT, platelet count; iPTH, intact parathormone; RBC, red blood cell count; SBP, systolic blood pressure; TIA: transient ischemic attack; TSAT, transferrin saturation; WBC, white blood cell count.
Table 2:
Echocardiography: parameters in in all HD patients and in individuals categorized according to the occurrence of the composite endpoint.
| All HD patients N = 74 | Endpoint—no N = 44 | Endpoint—yes N = 30 | P | |
|---|---|---|---|---|
| LAVi (mL/m2) | 28.3 (21.9–42.8) | 26.3 (18.3–42.9) | 29.2 (23.6–41.9) | .05 |
| LAD (cm) | 3.7 (3.3–4.1) | 3.8 (3.4–4.5) | 3.7 (3.2–3.9) | .37 |
| LVEDVi (mL/m2) | 51.3 ± 21.9 | 50.2 ± 20 | 53.1 ± 25.2 | .17 |
| LVMi (g/m2) | 131.1 ± 36.8 | 128.8 ± 33.2 | 134.8 ± 41.3 | .05 |
| Ejection fraction (%) | 57.5 ± 9.1 | 58.9 ± 7.6 | 55.3 ± 10.9 | .11 |
| Vmax (m/s) | 2.5 (2.2–2.9) | 2.44 (2–2.78) | 2.49 (1.92–2.89) | .23 |
| TAPSE (mm) | 20 ± 4.8 | 2.1 (1.8–2.5) | 2.1 (1.9–2.6) | .97 |
| E/e′ | 13.2 ± 4.2 | 10.5 ± 3.6 | 16.2 ± 4.8 | .01 |
| Fractional shortening (%) | 37 ± 8.7 | 37.5 ± 7.9 | 36.2 ± 10.2 | .60 |
| RAVi (mL/m2) | 29.5 (14–25.3) | 18.1 (11.8–24.3) | 20.3 (15.9–24.9) | .23 |
Statistically significant differences between the two subgroups are highlighted in bold.
Data are presented as mean ± standard deviation or median (interquartile range).
LAVi, left atrial volume index; LAD: left atrial diameter; LVEDVi, left-ventricular end diastolic volume index; LVMi, left ventricular mass index; E/e′, early diastolic peak left ventricular inflow velocity (E)/early diastolic peak lateral mitral annular velocity (e′) ratio; RAVi, right atrial volume index.
Circulating miRNAs expression in HD patients
Collectively, HD patients displayed a lower normalized relative expression of miRNA 23a-3p (P < .0001), 451a (P = .001), 30a-5p (P = .003) and let7d-5p (P < .0001) as compared with healthy controls (Table 3 and Fig. 1) with an average fold regulation (FR) of, respectively, −1.72 for miRNA 23a-3p, −2.03 for miRNA 451a, −3.90 for miRNA 23a-3p and −3.05 for miRNA let7d-5p. No differences in circulating miRNAs were noticed among ESKD individuals treated with bicarbonate dialysis or hemodiafiltration (P ranging from .38 to .66).
Table 3:
Circulating levels of miRNAs [median (interquartile range) 2−ΔCT relative expression] in all HD patients and in individuals categorized according to the occurrence of the composite endpoint.
| All HD patients N = 74 | Endpoint—no N = 44 | Endpoint—yes N = 30 | P | |
|---|---|---|---|---|
| miRNA 30a-5p | 0.023 (0.012–0.048) | 0.027 (0.019–0.059) | 0.013 (0.008–0.033) | .03 |
| miRNA 23a-3p | 0.401 (0.219–0.651) | 0.507 (0.302–0.858) | 0.223 (0.155–0.432) | <.001 |
| miRNA 451a | 20 (16.8–37.6) | 29.3 (18.3–39.2) | 17.9 (14.7–19.7) | <.001 |
| miRNA let7D-5p | 0.033 (0.014–0.065) | 0.034 (0.014–0.072) | 0.030 (0.014–0.058) | .66 |
Statistical differences between subgroups are highlighted in bold.
Figure 1:
Differences in median circulating miRNAs relative expression between HD patients and healthy controls. *P < .0001; **P = .001; ***P = .003.
At univariate followed by multivariate regression analyses (Table 4) significant independent correlates of miRNAs expression were: peak aortic valve velocity (Vmax) (β = 0.739; P = .005) and tricuspid annular plane excursion (TAPSE) (β = −0.627; P = .04) for miRNA 30a-5p; uric acid (β = 0.254; P = .03) and E/e′ (β = −0.395; P = .01) for miRNA 23a-3p; E/e′ (β = −0.268; P = .01), serum potassium (β = 0.233; P = .03) and high-density lipoprotein (HDL) (β = 0.236; P = .03) for miRNA 451a; and serum sodium (β = −0.556; P < .001) and C-reactive protein (β = 0.224; P = .03) for miRNA let7d-5p.
Table 4:
Univariate and multivariate correlates of circulating miRNAs expression in HD patients.
| Univariate R/Rho | P | Multivariate R2 | P | Multivariate β | P | |
|---|---|---|---|---|---|---|
| miRNA 30a-5p | 0.24 | .004 | ||||
| Vmax | 0.434 | .001 | 0.739 | .005 | ||
| TAPSE | –0.331 | .001 | –0.627 | .04 | ||
| Serum phosphate | 0.231 | .05 | 0.044 | .713 | ||
| Hemoglobin | –0.395 | .01 | 0.041 | .736 | ||
| miRNA 23a-3p | 0.31 | <.001 | ||||
| Uric acid | 0.254 | .03 | 0.261 | .03 | ||
| E/e′ | –0.395 | .01 | –0.262 | .04 | ||
| LDL | –0.236 | .04 | –0.213 | .07 | ||
| LVMi | –0.275 | .03 | –0.154 | .15 | ||
| Vmax | 0.315 | .01 | 0.011 | .97 | ||
| TAPSE | 0.332 | .007 | 0.311 | .42 | ||
| miRNA 451a | 0.26 | <.001 | ||||
| E/e′ | –0.371 | .001 | –0.268 | .01 | ||
| Serum potassium | 0.324 | .005 | 0.233 | .03 | ||
| HDL | 0.229 | .05 | 0.236 | .03 | ||
| Pulse pressure | –0.283 | .01 | –0.210 | .06 | ||
| Albumin | 0.236 | .04 | 0.125 | .23 | ||
| DBPa | 0.357 | .002 | ||||
| miRNA let7d-5p | 0.39 | <.001 | ||||
| Serum sodium | –0.504 | .001 | –0.556 | <.001 | ||
| C-reactive protein | 0.205 | .03 | 0.224 | .03 | ||
| SBP | –0.257 | .05 | –0.159 | .14 | ||
| TAPSE | –0.227 | .05 | –0.068 | .53 |
aNot introduced in the model in order to avoid co-linearity with pulse pressure.
Statistically significant associations are highlighted in bold.
DBP, diastolic blood pressure; SBP, systolic blood pressure; LVMi, left ventricular mass index; E/e′, early diastolic peak left ventricular inflow velocity (E)/early diastolic peak lateral mitral annular velocity (e′) ratio.
Composite endpoint during the follow-up period
During a median follow-up of 22 months (range 1–24), 30 subjects (40.5%) reached the composite endpoint. In more detail, 13 subjects died due to a fatal CV event (n = 6) or due to other causes (n = 7) while 17 subjects experienced a non-fatal CV event (4 cerebrovascular disease, 8 coronary events, 5 severe arrhythmia episodes).
At baseline, individuals experiencing the composite endpoint were older (72.5 ± 12.5 vs 66.6 ± 11.9 years; P = .04), had lower diastolic (58 ± 14.2 vs 73.7 ± 10.2 mmHg; P = .04) and pulse pressure (76.8 ± 25.6 vs 57.5 ± 23.2 mmHg; P = .03) but higher myocardium-specific creatine kinase (CK-MB) (24.2 ± 9.2 vs 18.8 ± 6.6 UI/L; P = .02), high-sensitive troponin I [23.1 (206–53.6) vs 19.3 (7.7–22.2) pg/mL; P = .04] and low-density lipoprotein (LDL) cholesterol (91.9 ± 30.5 vs 77.9 ± 25.8 mg/dL; P = .05). Echocardiography displayed an increased left atrial volume [29.2 (23.6–41.9) vs 26.3 (18.3–42.9); P = .05], left ventricular mass (134.8 ± 41.3 vs 128.8 ± 33.2; P = .05) and a more severely impaired diastolic function (E/e′ 16.2 ± 4.8 vs 10.5 ± 3.6; P = .01). There was also a barely prevalent history of any CV disease (P = .08), particularly stroke (P = .08), and a marginally lower probability of being on hemodiafiltration (P = .06). No significant differences were noticed for any other clinical, instrumental or laboratory variable (P ≥ .10).
Separate clinical data in patients experiencing the composite outcome as compared with others are summarized in Tables 1 and 2.
Patients reaching the endpoint displayed a significantly reduced normalized relative expression of miRNA 30a-5p (P = .03; FR = −1.77), miRNA 23a-3p (P < .001; FR = −1.83) and 451a (P < .001; FR = −1.81), while no differences between study subgroups were reported for miRNA let7d-5p (Table 3 and Fig. 2).
Figure 2:
Differences in median circulating miRNAs relative expression in patients experiencing the composite outcome as compared with others. *P = .003; **P < .0001; ns = not significant.
Cox regression analyses for the composite endpoint
Variables which were different at baseline between study subgroups were evaluated by Cox regression analysis to confirm associations with the combined outcome (Table 5).
Table 5:
Univariate: and multivariate Cox regression analyses for variables significantly associated to the composite endpoint.
| Univariate analysis | Units of increase | HR | 95% CI | X2 | P |
|---|---|---|---|---|---|
| Pulse pressure | 1 mmHg | 1.026 | 1.008–1.044 | 8.003 | .005 |
| E/e′ | 1 U | 1.155 | 1.081–1.234 | 18.343 | <.001 |
| CK-MB | 1 UI/L | 1.062 | 1.014–1.112 | 6.618 | .01 |
| miRNA 30a-5p | 2−ΔCT (×103) | 0.965 | 0.938–0.992 | 6.198 | .01 |
| miRNA 23a-3p | 2−ΔCT (×103) | 0.997 | 0.995–0.999 | 10.047 | .002 |
| miRNA 451a | 2−ΔCT | 0.928 | 0.889–0.969 | 11.422 | .001 |
| Age | years | 1.036 | 0.994–1.081 | 2.754 | .09 |
| LAVi | mL/m2 | 1.016 | 0.983–1.049 | 1.137 | .34 |
| LVMi | g/m2 | 1.011 | 0.996–1.025 | 2.232 | .14 |
| LDL | mg/dL | 1.013 | 0.999–1.028 | 1.904 | .11 |
| hs-CTNi | pg/mL | 1.006 | 0.997–1.015 | 1.301 | .22 |
| Multivariate model 1: including miRNA 30–5p | |||||
| E/e′ | 1 U | 1.153 | 1.044–1.274 | 7.912 | .005 |
| miRNA 30a-5p | 2−ΔCT (×103) | 0.995 | 0.985–0.998 | 5.537 | .02 |
| Pulse pressure | 1 mmHg | 1.010 | 0.987–1.033 | 0.711 | .39 |
| CK-MB | 1 UI/L | 1.005 | 0.982–1.029 | 0.420 | .51 |
| Multivariate model 2: including miRNA 23a-3p | |||||
| E/e′ | 1 U | 1.074 | 1.017–1.135 | 6.634 | .01 |
| miRNA 23a-3p | 2−ΔCT (×103) | 0.998 | 0.996–0.999 | 4.562 | .05 |
| CK-MB | 1 UI/L | 1.111 | 0.985–1.253 | 2.963 | .08 |
| Pulse pressure | 1 mmHg | 1.013 | 0.990–1.037 | 1.248 | .26 |
| Multivariate model 3: including miRNA 451a | |||||
| CK-MB | 1 UI/L | 1.083 | 1.028–1.140 | 9.008 | .003 |
| E/e′ | 1U | 1.124 | 1.016–1.244 | 5.117 | .02 |
| miRNA 451a | 2−ΔCT | 0.943 | 0.892–0.997 | 4.337 | .03 |
| Pulse pressure | 1 mmHg | 1.001 | 0.976–1.028 | 0.010 | .92 |
Statistically significant predictors are highlighted in bold.
E/e′, early diastolic peak left ventricular inflow velocity (E)/early diastolic peak lateral mitral annular velocity (e′) ratio; hs-cTnI: high-sensitivity c-Troponin I; LAVi, left atrial volume index; LVMi, left ventricular mass index.
At univariate models, pulse pressure [hazard ratio (HR) 1.026; 95% confidence interval (CI) 1.008–1.044; P = .005], E/e′ (HR 1.155; 95% CI 1.081–1.234; P < .001), CK-MB (HR 1.062; 95% CI 1.014–1.112; P = .01), miRNA 30a-5p (HR 0.965; 95% CI 0.938–0.992; P = .01), miRNA 23a-3p (HR 0.997; 95% CI 0.995–0.999; P = .002) and miRNA 451a (HR 0.928; 95% CI 0.889–0.969; P = .001) were confirmed as significantly correlated with the study endpoint. Conversely, such an association was not found for age, left atrial volume index, left ventricular mass index, LDL cholesterol and high-sensitive troponin I (P ranging from .09 to .34).
In order to prevent model overfitting, we built three separate multivariable Cox regression models to test independent associations, each one encompassing a different miRNA. The first model revealed a 5% decreasing risk of the combined outcome per each 2−ΔCT (×103) increase in the normalized relative miRNA 30a-5p expression. In the same model, an increased E/e′ was the sole other independent predictor of the combined outcome, while the association with pulse pressure and CK-MB were lost. Similarly, in the second model only E/e′ and miRNA 23a-3p [with a 2% decrease in the risk of the combined outcome per each 2−ΔCT (×103) increase in its normalized relative expression] were independently associated with the study endpoint. Finally, in the third model E/e′, CK-MB and miRNA 451a (displaying a remarkable 37% decrease in the overall risk for each 2−ΔCT increase in the normalized relative expression) resulted as independent outcome predictors while pulse pressure did not.
ROC analyses of miRNAs
The diagnostic capacity of miRNAs, alone or in combination, to identify HD patients experiencing the composite outcome was assessed by separate ROC analyses (Fig. 3 and Table 6).
Figure 3:
ROC curves testing the individual (A, B and C) and combined (D) discriminatory capacity of circulating miRNAs expression in identifying patients experiencing the composite outcome.
Table 6:
Diagnostic: profile (area under the curve-AUC) of miRNAs, alone or in combination, in identifying patients experiencing the composite endpoint.
| AUC (95% CI) | P | SE | |
|---|---|---|---|
| miRNA let7d-5p | 0.549 (0.424–0.669) | .489 | 0.0706 |
| miRNA 30a-5p | 0.831 (0.726 to 0.908) | <.001 | 0.0512 |
| miRNA 451a | 0.778 (0.666 to 0.866) | <.001 | 0.0539 |
| miRNA 23a-3p | 0.797 (0.687 to 0.882) | <.001 | 0.0548 |
| Combined miRNAs a | 0.890 (0.796 to 0.951) | <.001 | 0.0363 |
Statistically significant AUCs are highlighted in bold.
aAUC calculated from the individual predicted probabilities obtained from a multivariate logistic regression model including miRNA 30a-5p, miRNA 451a and miRNA 23a-3p as covariates.
miRNA let7d-5p was not discriminant in such regard, displaying an AUC of 0.549 (95% CI 0.424–0.669; P = .489). Conversely, miRNA 30a-5p showed a remarkable diagnostic capacity with an AUC of 0.831 (95% CI 0.726–0.908; P < .001), while the AUCs of miRNAs 451a and 23a-3p were 0.778 (95% CI 0.666–0.866; P < .001) and 0.797 (95% CI 0.687–0.882; P < .001), respectively. There was no statistical difference between the three AUCs (P ranging from .46 to .79).
The combination of miRNA 30a-5p, miRNA 451a and miRNA 23a-3p improved the individual discriminatory capacity, increasing the AUC to 0.890 (95% CI 0.796–0.951; P < .001; P vs single miRNAs AUCs ranging from .01 to .05).
Survival analyses in patients categorized by circulating miRNAs expression
Kaplan–Meier survival curves of event-free patients were generated for individuals with a normalized expression of miRNAs 30a-5p, 451a and 23a-3p below or above the optimal, ROC-derived cut-off (Fig. 4). Patients with miRNA 30a-5p below threshold experienced a significantly faster evolution to the combined endpoint (crude HR 7.95; 95% CI 3.78–17.16; P < .0001; Log-rank test, χ2 27.958), as well as those showing lower expression of miRNA 23a-3p (crude HR 8.61; 95% CI 3.62–20.47; P < .0001; Log-rank test, χ2 23.758) and miRNA 451a (crude HR 7.95; 95% CI 2.14–9.23; P = .0001; Log-rank test, χ2 16.055). A combined survival analysis was then performed, comparing the outcome risk of HD patients displaying all the three miRNAs below cut-off vs others. Such a combined approach (crude HR 7.07; 95% CI 3.39–14.74; P < .0001; Log-rank test, χ2 27.308) outperformed the risk stratification based on the single analysis of miRNA 451a or 23a-3p while it did not apparently improve the predictive capacity of miRNA 30a-5p.
Figure 4:
Kaplan–Meier curves of endpoint-free patients with circulating miRNA 30a-5p (A), 23a-3p (B) and 451a (C) expression below or above the optimal ROC-derived cut-offs. Survival probabilities in individuals with expression of the three miRNAs below cut-off are showed in (D).
DISCUSSION
Findings from our study raise two main points for discussion. First, in chronic HD patients, we found lower circulating levels of a small panel of miRNAs putatively involved in CV disease, as compared with healthy individuals. Second, a clear relationship emerged between a more decreased expression of some miRNAs and worse clinical outcomes in such a high risk population.
Overall, reduced levels of various circulating miRNAs have previously been described in uremic populations [32], being mostly attributed to the presence of impaired inter-cellular communications [33] and nuclear reprogramming induced by uremic toxins, inflammation and sustained oxidative stress [34]. Additionally, potential clearance by the HD treatment itself has also been called into question [35].
In our study cohort, we found a close and independent relationship between low miRNAs 30-5p, 23a-3p and 451a levels and parameters of cardiac (dys)function including Vmax, TAPSE and E/e′. On top of this, miRNAs levels were also influenced by other factors such as uric acid, serum potassium and HDL levels, although a substantial percentage of their variability in this cohort remained unexplained (multivariate R2 ranging from 0.24 to 0.39).
Previous evidence has pointed out the critical involvement of miRNA 30-5p, 23a-3p and miRNA 451a in pathological cardiac remodeling following acute myocardial infarction, chronic cardiomyopathies or in heart failure [12, 15, 18, 22, 25]. Under physiological conditions, these miRNAs down-regulate autophagy and apoptosis of cardiomyocytes and modulate collagen balance. A reduced expression is therefore considered detrimental as it may elicit tissue fibrosis due to deranged collagen deposition and asymmetric ventricular hypertrophy with following wall dilatation [36–38]. Interestingly, similar roles have been described also for miRNA let7d-5p [16, 28, 29]. The low expression of this miRNA, however, was apparently not related to cardiac abnormalities in our study cohort, being rather influenced by inflammation and circulating sodium levels.
The main aim of our study, however, was to investigate the prognostic value of these miRNAs in predicting worse CV outcomes in HD individuals. In this regard, our results corroborate the hypothesis that miRNA 30-5p, 23a-3p and 451a may hold a clinical value for risk stratification while, again, such a capacity was not proved for miRNA let7d-5p.
Our findings pair well with previous evidence reported in other CV disorders. For instance, low to very low circulating miRNA 30a-5p levels predict sudden cardiac or arrhythmic death in patients with coronary heart disease [14]; an aberrant miRNA 23a-3p expression reflects the severity of ST-elevated myocardial infarction [17], predicts heart failure onset [39], and may help risk stratification for CV mortality and worsening New York Heart Association status in the long-term [40]. Finally, early deranged miRNA 451a levels anticipate acute coronary plaque rupture after percutaneous coronary intervention [26] and has prognostic value on the risk of transient or acute stroke in critically ill patients [41].
In our study, patients reaching the composite endpoint showed at baseline significantly lower levels of these three miRNAs as compared with others. By the same token, miRNA 30-5p, 23a-3p and 451a displayed a remarkable discriminatory capacity at ROC analyses in identifying such individuals within the whole study cohort; of note, this capacity resulted even improved by combining the diagnostic information of all the three miRNAs (combined AUC 0.890, P vs single AUCs ranging from .01 to .05). Unadjusted Kaplan–Meier survival analyses confirmed a faster progression to the endpoint in individuals displaying circulating miRNA 30-5p, 23a-3p and 451a below the optimal ROC-derived cut-off values, with crude HRs ranging from 7.95 to 8.61. On the contrary, rather surprisingly, the survival function combining the predictive value of the three miRNAs simultaneously was apparently not superior (crude HR 7.07; 95% CI 3.39–14.74) to estimations based on single-miRNA evaluation.
Finally, and more importantly, a time-dependent association between baseline circulating miRNA 30-5p, 23a-3p and 451a and the composite outcome emerged also at univariate Cox regression analyses and was further confirmed as independent from potential confounders, including the severity of diastolic dysfunction, in three different multivariate proportional-hazards regression models.
This latter observation is, to our opinion, of foremost importance as it would suggest that miRNAs assessment may impart significant prognostic information beyond those traditionally portended by the presence of an overt cardiac dysfunction, thereby bringing an additive value for improving risk prediction.
Our study has strengths and limitations which deserve mentioning. Strengths include the prospective design, a multicentric homogeneous cohort in terms of baseline risk profile and clinical characteristics, an adequate follow-up duration, an inclusive and validated combined endpoint and a robust statistical approach which allowed testing the discriminatory and predictive value of miRNAs from different viewpoints. Key limitations are the small sample size and the relatively low number of events recorded; this latter was indeed sufficient to perform reliable multivariate analyses without model overfitting but prevented stratified approaches based on different types of events.
Finally, the observational nature of the study cannot exclude, in principle, selection bias or residual confounding and may leave unanswered the question as to whether the associations found between circulating miRNAs expression, clinical variables and long-term outcomes rely on causal pathways or represent just a biological epiphenomenon.
In conclusion, we have demonstrated that uremic patients in chronic HD treatment display an aberrantly low expression of a small panel of miRNAs previously acknowledged to be involved in normal and pathological cardiac remodeling. The remarkable prognostic value of miRNA 30-5p, 23a-3p and 451a in predicting a combined CV endpoint may set the stage for larger studies on more heterogeneous HD cohorts to confirm the potential usefulness of this miRNAs signature for improving CV risk assessment in such a particular disease setting.
Supplementary Material
ACKNOWLEDGEMENTS
The authors wish to thank Ms Silvia Fasiello and Ms Giulia Longo for the precious contribution in data collection.
Contributor Information
Davide Bolignano, Magna Graecia University, Nephrology and Dialysis Unit, Catanzaro, Italy.
Marta Greco, Magna Graecia University, Department of Health Sciences, Catanzaro, Italy.
Pierangela Presta, Magna Graecia University, Nephrology and Dialysis Unit, Catanzaro, Italy.
Anila Duni, Department of Nephrology, School of Medicine, University of Ioannina, Ioannina, Greece.
Caterina Vita, Magna Graecia University, Nephrology and Dialysis Unit, Catanzaro, Italy.
Ethymios Pappas, Hemodialysis Unit, General Hospital of Filiates, Filiates, Greece.
Maria Mirabelli, Magna Graecia University, Department of Health Sciences, Catanzaro, Italy.
Lampros Lakkas, Second Department of Cardiology, University Hospital of Ioannina, Ioannina, Greece.
Katerina K Naka, Second Department of Cardiology, University Hospital of Ioannina, Ioannina, Greece.
Antonio Brunetti, Magna Graecia University, Department of Health Sciences, Catanzaro, Italy.
Daniela Patrizia Foti, Magna Graecia University, Experimental and Clinical Medicine, Catanzaro, Italy.
Michele Andreucci, Magna Graecia University, Nephrology and Dialysis Unit, Catanzaro, Italy.
Giuseppe Coppolino, Magna Graecia University, Nephrology and Dialysis Unit, Catanzaro, Italy.
Evangelia Dounousi, Department of Nephrology, School of Medicine, University of Ioannina, Ioannina, Greece.
FUNDING
None declared.
AUTHORS’ CONTRIBUTIONS
Research idea: D.B., M.G. and G.C. Data collection and analysis: P.P., A.D., C.V., E.P., L.L. and K.K.N. Laboratory measurement: M.G., D.P.F. and M.M. Manuscript preparation and revision: D.B., G.C., M.A., D.P.F., E.D. and A.B.
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.
CONFLICT OF INTEREST STATEMENT
The authors declare no conflict of interest with respect to the present work. The results presented in this paper have not been published previously in whole or part, except in abstract format.
REFERENCES
- 1. Provenzano M, Coppolino G, De Nicola L et al. Unraveling cardiovascular risk in renal patients: a new take on old tale. Front Cell Dev Biol 2019;7:314. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Russo D, Corrao S, Battaglia Y et al. Progression of coronary artery calcification and cardiac events in patients with chronic renal disease not receiving dialysis. Kidney Int 2011;80:112–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Thompson S, James M, Wiebe N et al. Cause of death in patients with reduced kidney function. J Am Soc Nephrol 2015;26:2504–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Webster AC, Nagler EV, Morton RL et al. Chronic kidney disease. Lancet 2017;389:1238–52. [DOI] [PubMed] [Google Scholar]
- 5. Glassock RJ, Pecoits-Filho R, Barberato SH. Left ventricular mass in chronic kidney disease and ESRD. Clin J Am Soc Nephrol 2009;4 Suppl 1:S79–91. [DOI] [PubMed] [Google Scholar]
- 6. Alhaj E, Alhaj N, Rahman I et al. Uremic cardiomyopathy: an underdiagnosed disease. Congest Heart Fail 2013;19:E40–5. [DOI] [PubMed] [Google Scholar]
- 7. Kozomara A, Birgaoanu M, Griffiths-Jones S et al. From microRNA sequences to function. Nucleic Acids Res 2019;47:D155–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Ullah S, John P, Bhatti A. MicroRNAs with a role in gene regulation and in human diseases. Mol Biol Rep 2014;41:225–32. [DOI] [PubMed] [Google Scholar]
- 9. Yan B, Wang H, Tan Y et al. microRNAs in cardiovascular disease: small molecules but big roles. Curr Top Med Chem 2019;19:1918–47. [DOI] [PubMed] [Google Scholar]
- 10. Kong AS, Lai KS, Lim SE et al. miRNA in ischemic heart disease and its potential as biomarkers: a comprehensive review. Int J Mol Sci 2022;23:9001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Lang RM, Badano LP, Mor-Avi V et al. Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. J Am Soc Echocardiogr 2015;28:1–39.e14. [DOI] [PubMed] [Google Scholar]
- 12. Yang XX, Zhao ZY. miR-30a-5p inhibits the proliferation and collagen formation of cardiac fibroblasts in diabetic cardiomyopathy. Can J Physiol Pharmacol 2022;100:167–75. [DOI] [PubMed] [Google Scholar]
- 13. Haddad G, Kolling M, Wegmann UA et al. Renal AAV2-mediated overexpression of long non-coding RNA H19 attenuates ischemic acute kidney injury through sponging of microRNA-30a-5p. J Am Soc Nephrol 2021;32:323–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Silverman MG, Yeri A, Moorthy MV et al. Circulating miRNAs and risk of sudden death in patients with coronary heart disease. JACC Clin Electrophysiol 2020;6:70–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Maciejak A, Kostarska-Srokosz E, Gierlak W et al. Circulating miR-30a-5p as a prognostic biomarker of left ventricular dysfunction after acute myocardial infarction. Sci Rep 2018;8:9883. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Long G, Wang F, Duan Q et al. Circulating miR-30a, miR-195 and let-7b associated with acute myocardial infarction. PLoS One 2012;7:e50926. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Bukauskas T, Mickus R, Cereskevicius D et al. Value of serum miR-23a, miR-30d, and miR-146a biomarkers in ST-elevation myocardial infarction. Med Sci Monit 2019;25:3925–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Wu F, Wang F, Yang Q et al. Upregulation of miRNA-23a-3p rescues high glucose-induced cell apoptosis and proliferation inhibition in cardiomyocytes. In Vitro Cell Dev Biol Anim 2020;56:866–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Sheng S, Zou M, Yang Y et al. miR-23a-3p regulates the inflammatory response and fibrosis in diabetic kidney disease by targeting early growth response 1. In Vitro Cell Dev Biol Anim 2021;57:763–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Fan PC, Chen CC, Peng CC et al. A circulating miRNA signature for early diagnosis of acute kidney injury following acute myocardial infarction. J Transl Med 2019;17:139. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Deng HY, He ZY, Dong ZC et al. MicroRNA-451a attenuates angiotensin II-induced cardiac fibrosis and inflammation by directly targeting T-box1. J Physiol Biochem 2022;78:257–69. [DOI] [PubMed] [Google Scholar]
- 22. Sonsoz MR, Yilmaz M, Cevik E et al. Circulating levels of MicroRNAs in hypertrophic cardiomyopathy: the relationship with left ventricular hypertrophy, left atrial dilatation and ventricular depolarisation-repolarisation parameters. Heart Lung Circ 2022;31:199–206. [DOI] [PubMed] [Google Scholar]
- 23. Taverner D, Llop D, Rosales R et al. Plasma expression of microRNA-425-5p and microRNA-451a as biomarkers of cardiovascular disease in rheumatoid arthritis patients. Sci Rep 2021;11:15670. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Muendlein A, Geiger K, Leiherer A et al. Evaluation of the associations between circulating microRNAs and kidney function in coronary angiography patients. Am J Physiol Renal Physiol 2020;318:F315–21. [DOI] [PubMed] [Google Scholar]
- 25. Scrimgeour NR, Wrobel A, Pinho MJ et al. microRNA-451a prevents activation of matrix metalloproteinases 2 and 9 in human cardiomyocytes during pathological stress stimulation. Am J Physiol Cell Physiol 2020;318:C94–102. [DOI] [PubMed] [Google Scholar]
- 26. Li S, Lee C, Song J et al. Circulating microRNAs as potential biomarkers for coronary plaque rupture. Oncotarget 2017;8:48145–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Ou M, Li X, Cui S et al. Emerging roles of let7d in attenuating pulmonary arterial hypertension via suppression of pulmonary artery endothelial cell autophagy and endothelin synthesis through ATG16L1 downregulation. Int J Mol Med 2020;46:83–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Wong LL, Saw EL, Lim JY et al. MicroRNA let-7d-3p contributes to cardiac protection via targeting HMGA2. Int J Mol Sci 2019;20:1522. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Liang H, Pan Z, Zhao X et al. LncRNA PFL contributes to cardiac fibrosis by acting as a competing endogenous RNA of let-7d. Theranostics 2018;8:1180–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Yu ML, Wang JF, Wang GK et al. Vascular smooth muscle cell proliferation is influenced by let-7d microRNA and its interaction with KRAS. Circ J 2011;75:703–9. [DOI] [PubMed] [Google Scholar]
- 31. Gooch JL, King C, Francis CE et al. Cyclosporine A alters expression of renal microRNAs: new insights into calcineurin inhibitor nephrotoxicity. PLoS One 2017;12:e0175242. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Neal CS, Michael MZ, Pimlott LK et al. Circulating microRNA expression is reduced in chronic kidney disease. Nephrol Dial Transplant 2011;26:3794–802. [DOI] [PubMed] [Google Scholar]
- 33. Zietzer A, Steffen E, Niepmann S et al. MicroRNA-mediated vascular intercellular communication is altered in chronic kidney disease. Cardiovasc Res 2022;118:316–33. [DOI] [PubMed] [Google Scholar]
- 34. Franczyk B, Gluba-Brzozka A, Olszewski R et al. miRNA biomarkers in renal disease. Int Urol Nephrol 2022;54:575–88. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Fan W, Li X, Xu X et al. Effects of high-flux dialysis combined with hemoperfusion on serum GRP78 and miR-495-3p in renal failure patients. Biomed Res Int 2021;2021:9591177. [DOI] [PMC free article] [PubMed] [Google Scholar] [Retracted]
- 36. Zhang X, Dong S, Jia Q et al. The microRNA in ventricular remodeling: the miR-30 family. Biosci Rep 2019;39:BSR20190788. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Zeng Z, Wang K, Li Y et al. Down-regulation of microRNA-451a facilitates the activation and proliferation of CD4(+) T cells by targeting Myc in patients with dilated cardiomyopathy. J Biol Chem 2017;292:6004–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Danielson KM, Shah R, Yeri A et al. Plasma circulating extracellular rnas in left ventricular remodeling post-myocardial infarction. EBioMedicine 2018;32:172–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Charrier H, Cuvelliez M, Dubois-Deruy E et al. Integrative system biology analyses identify seven MicroRNAs to predict heart failure. Noncoding RNA 2019;5:22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Rincon LM, Rodriguez-Serrano M, Conde E et al. Serum microRNAs are key predictors of long-term heart failure and cardiovascular death after myocardial infarction. ESC Heart Fail 2022;9:3367–79. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Giordano M, Ciarambino T, D'Amico M et al. Circulating MiRNA-195-5p and -451a in transient and acute ischemic stroke patients in an emergency department. J Clin Med 2019;8:130. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.