Dear Editor, Cardiac allograft vasculopathy (CAV) remains a major cause of long-term mortality after heart transplant (HT).1 Current invasive diagnostic methods have limited sensitivity in the early stage of CAV and underestimate the severity of the disease.2-4 To address that, we conducted a non-targeted transcriptomic study focused on identifying circulating miRNAs as potential non-invasive biomarkers for CAV detection at 1-year surveillance. In the discovery phase, a total of 70 plasma samples from patients who underwent HT were analysed using RNA-seq technology (57 without CAV and 13 with CAV). Subsequently, the cohort was expanded to 103 patients for validation using RT-qPCR. Patients with CAV were classified into low-grade CAV (CAV1, n = 6) and high-grade CAV (CAV2-3, n = 9). The clinical characteristics of the patients included at the time of sample collection are summarised in Table 1. Both groups were similar, except for donor age in the validation phase (p < 0.05). Discovery Phase (RNA-seq) Validation Phase (RT-qPCR) No CAV (N = 57) CAV (N = 13) No CAV (N = 88) CAV (N = 15) Large-scale screening identified nine miRNAs significantly upregulated in patients with CAV, all of which demonstrated good diagnostic performance with area under the curve (AUC) values greater than 0.700 (Figure S1A–C). Cubic spline plots revealed a near-linear association between these miRNAs and CAV risk, with all microRNAs showing a significant overall effect (p < 0.05) except miR-374b-5p (Figure S1D). The upregulation of six miRNAs in patients with CAV was validated using RT-qPCR: miR-21-5p (FC = 2.27, p < 0.01) (Figure 1A), miR-223-3p (FC = 3.34, p < 0.001) (Figure 1B), miR-23a-3p (FC = 1.43, p < 0.05) (Figure 1C), miR-23b-3p (FC = 1.53, p < 0.05) (Figure 1D), miR-374b-5p (FC = 1.49, p < 0.05) (Figure 1E), and miR-98-5p (FC = 1.87, p < 0.05) (Figure 1F). Receiver operating characteristic (ROC) curves were plotted to analyse the ability of altered miRNAs to detect CAV, and all of them showed a significant AUC (Figure S2A). Based on evidence from our previous study indicating that a combination of miRNAs provides greater diagnostic accuracy than individual miRNAs,5 we developed a three-miRNA signature combining miR-223-3p, miR-23a-3p, and miR-23b-3p. This signature showed higher diagnostic efficacy than individual miRNAs (AUC = 0.850; p < 0.001) (Figure S2B). Moreover, considering that donor age reportedly has a strong association with the disorder6 and that it was a clinical variable altered in our cohort with an AUC of 0.665 (p < 0.05) (Figure S2C), this variable was incorporated into the model together with the three-miRNA signature, obtaining an AUC of 0.869 (p < 0.001) (Figure S2D). Sensitivity, specificity, positive (PPV) and negative (NPV) predictive values and positive (LR+) and negative (LR-) likelihood ratios for the diagnosis of CAV for individual miRNAs and their combinations are summarised in Table 2. The combination of the three-miRNA signature with donor age achieved a sensitivity of 87% and specificity of 81%, and it showed the highest PPV and NPV (45% and 97%, respectively) with superior LR+ (4.7) and lower LR- (0.16). After internal validation with bootstrapping, the optimism-corrected AUC remained stable (corrected AUC = 0.824), supporting model robustness. Overall, this model demonstrated the highest diagnostic accuracy for discriminating between patients with and without CAV. We observed that our model could detect CAV independently of disease severity. The combination of the three-miRNA signature with donor age achieved an AUC of 0.783 (p < 0.05) for detecting CAV1 and an AUC of 0.917 (p < 0.001) for detecting CAV2-3 (Figure 2A,B). In addition, the model demonstrated the ability to distinguish patients with high-grade CAV from those with low-grade CAV, with an AUC of 0.944 (p < 0.01; Figure 2C). Sensitivity, specificity, PPV, NPV, LR+, and LR- for distinguishing different severity grades of CAV are detailed in Table S1. Furthermore, to assess whether the combination of the three-miRNA signature and donor age is related to low- and high-grade CAV, binary logistic regression analysis was performed, adjusting for age and sex. Therefore, our model was independently associated with the presence of CAV1 with odds ratio of 41.5 (95% CI 5.7–305.0, p < 0.001) and a C statistic of 0.741 (95% CI 0.455–1.000, p < 0.05) and for the presence of CAV2-3 with odds ratio of 45.2 (95% CI 5.2–395.1, p < 0.001) and a C statistic of 0.852 (95% CI 0.663–1.000, p < 0.001). Our study has several limitations. Although the number of cases identified in our cohort aligns with the expected incidence of CAV at 1-year post-heart transplantation (≈8%),7 the absolute number of affected patients is small, a limitation inherent to the low frequency of this complication in early follow-up and the single-centre nature of the study. Nevertheless, the number of events is comparable to that of previously published research in this field.7 Therefore, while the results obtained are consistent and methodologically robust, the model should be interpreted with caution. Overall, our findings provide a solid foundation and constitute a necessary first step to encourage future studies that overcome these limitations, particularly through validation in larger cohorts. In conclusion, this study identifies a circulating miRNA signature associated with CAV and demonstrates that combining miR-223-3p, miR-23a-3p, and miR-23b-3p levels with donor age provides significant discriminatory ability to detect CAV. The proposed approach may represent a promising non-invasive strategy to complement invasive coronary angiography, enhancing diagnostic capabilities for detecting CAV and severity stratification. Estefanía Tarazón and Esther Roselló-Lletí designed and supervised the study; Irene González-Torrent and Carlota Benedicto conducted experiments; Lorena Pérez-Carrillo, Marta Delgado-Arija and Isaac Giménez-Escamilla acquired and analysed data; Estefanía Tarazón, Esther Roselló-Lletí and Lorena Pérez-Carrillo contributed to data interpretation and/or discussion; Irene González-Torrent wrote the manuscript. All authors reviewed the manuscript. The authors are grateful to the Hemodynamic Unit (University and Polytechnic La Fe Hospital) for their help with obtaining samples. This work was supported by the National Institute of Health “Fondo de Investigaciones Sanitarias del Instituto de Salud Carlos III” (Projects: PI20/00071, PI24/00088 and PI24/01310) co-funded by European Union; Miguel Servet contract: CP21/00041 co-funded by European Union; contracts FI21/00034 and FI21/00186]; Ministry of Science and Innovation (MCIN, 10.13039/501100011033) and State Investigation Agency (AEI) (Project CNS2022-135769 and CNS2024-1544396) co-funded by European Union “Next Generation EU” and the European Recovery, Transformation and Resilience Plan (PRTR); Conselleria de Educación, Universidades y Empleo (Project CIAICO/2022/246, contract CIACIF/2022/429). The authors declare no conflict of interest. The study was approved by the Ethics Committee (Biomedical Investigation Ethics Committee of University and Polytechnic Hospital La Fe of Valencia, Spain) and was conducted in accordance with the principles outlined in the Declaration of Helsinki8, and all subjects gave written informed consent to participate in the study. The data that support the findings of this study are available from the corresponding author upon reasonable request. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.
González‐Torrent et al. (Sun,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: