Retinal optical coherence tomography angiography as a biomarker of acute kidney injury after acute coronary syndrome

Abstract

Contrast-induced nephropathy (CIN) after percutaneous coronary intervention (PCI) for acute coronary syndrome (ACS) is frequent and associated with long-term renal impairment and mortality. Retinal vessel density (RVD) measured by OCT-Angiography could reflect the global cardiovascular burden of ACS patients and thus provide a fast and non-invasive assessment of the systemic microcirculation, that may be involved in CIN occurrence. Between October 2016 and March 2017, 452 ACS patients were admitted in our coronary care unit. Retinal OCT-A was performed within two days after PCI. Patients were divided in two groups, according to Acute kidney injury (AKI) occurrence (KDIGO criteria, increase in creatinine > 26.5 μmol/L within 2 days or > 50% within 7 days after contrast injection). Of the 452 consecutive patients admitted for ACS during the inclusion period, 216 had a PCI and an analyzable OCT-A. Overall, AKI occurred in 21 (10%) patients. AKI was significantly associated with age (69 ± 14y vs. 62 ± 12y, P = 0.009), Mehran score (10.1 vs. 4.8, P < 0.001), GRACE score (169 ± 38 vs. 127 ± 31 P < 0.001) and logNT-proBNP (7.4 ± 1.6 vs. 5.7 ± 1.7 P = 0.001), but not with the volume of iodinated contrast media. AKI patients had lower RVD values than non-AKI patients (17.5 ± 2.1 mm −1 vs. 19.3 ± 2.3 mm −1 , P = 0.001). The addition of low RVD to Mehran score and NT-proBNP level (model 1), and to GRACE score and NT-proBNP level (model 2) significantly improved their predictive values ( P = 0.013 and P = 0.002, respectively). AKI after primary PCI for ACS was strongly and independently associated with low RVD evaluated with OCT-A. Moreover, this new retinal biomarker improved the predictive performance of Mehran and GRACE scores even after adjustment to NT-proBNP levels, suggesting that the microvascular systemic impairment is not currently taken into account neither by the classical risk score nor by the known clinical and biological factors associated to CIN onset.