Objective- Stroke is a heterogeneous disease with diverse causes, which affect the risk of recurrence. This study aimed to identify novel biomarkers that are clinically relevant to the diagnosis of cardioembolic stroke (CE) and the prediction of stroke recurrence using metabolomics. Approach and Results- We obtained blood samples and clinical data from a consecutively registered, hospital-based acute stroke registry and from healthy controls. Mass-spectrometry-based profiling was performed, and several metabolomic signatures were selected for the discrimination of CE and stroke recurrence, coupled with multivariate statistical analysis. Finally, 190 acute ischemic stroke participants (43 CE patients and 147 non-CE patients) and 30 control participants were included. We obtained 29 metabolomics signatures, and of these, 2 medium-chain acylcarnitines (decanoylcarnitine and octanoylcarnitine) were selected as independent discriminants for CE (odds ratio, 2.839; 95% CI, 1.241-6.493 for decanoylcarnitine; odds ratio, 2.839; 95% CI, 1.241-6.493 for octanoylcarnitine). Elevated medium-chain acylcarnitines were also associated with a higher risk of stroke recurrence (hazard ratio, 3.767; 95% CI, 1.276-11.117 for decanoylcarnitine; hazard ratio, 5.519; 95% CI, 1.22-18.781 for octanoylcarnitine). The levels of decanoylcarnitine and octanoylcarnitine were correlated as known surrogate markers of CE. The levels of decanoylcarnitine and octanoylcarnitine were significantly higher in stroke patients with a high-risk potential of cardioembolism than in those with low or intermediate risk. Conclusions- Metabolomics provided an improved understanding of CE pathogenesis and stroke recurrence. We have identified decanoylcarnitine and octanoylcarnitine as novel biomarkers for CE and stroke recurrence.
Keywords: carnitine; metabolomics; patients; recurrence; stroke.