Global Longitudinal Strain for Detection of Cardiac Iron Overload in Patients with Thalassemia: A Meta-analysis of Observational Studies with Individual-Level Participant Data

Abstract

Background Although cardiac magnetic resonance (CMR) is the most reliable tool for assessment of CIO in patients with thalassemia, it is not always readily available. Recent studies have explored the potential of GLS as an alternative for diagnosis of CIO. We aimed to investigate the efficacy of global longitudinal strain (GLS) for detection of cardiac iron level (CIO). Methods We searched SCOPUS, MEDLINE, and Embase to identify the studies which used GLS for assessment of CIO. We searched for individual participant data (IPD) in eligible studies to perform ROC curve analysis. CMR with a T2* cut-off value of 20ms was considered as the gold standard. A meta-analysis was performed and the risk of bias was assessed using the JBI Checklist. Results A total of 14 studies with 789 thalassemia patients (310 and 430 with and without CIO respectively and 49 with undetermined condition) were considered eligible for meta-analysis. IPDs of 405 participants were available. GLS was significantly lower in patients with CIO (-17.5 ± 2.7%) compared to those without CIO (-19.9 ± 2.3%; WMD = 1.6%, 95% CI= [0.76–2.4], p = 0.001, I2 = 77.1%) and to normal population (-20.61 ± 2.26%; WMD = 2.2%, 95% CI= [0.91–3.5], p = 0.001, I2 = 83.9%). A GLS<-19.5% could predict CIO with 92.8% sensitivity and 34.63% specificity (AUC = 0.659, 95% CI= [0.6–0.72], p-value < 0.0001). A GLS value≤-5.74% has 100% positive predictive and ≥-24.5% has 100% negative predictive values for detection of CIO. Conclusions According to our study, GLS is a strong predictor of CIO and when CMR is not available, it may be a useful screening method for identification of CIO in thalassemia patients.