Optimal markers of treatment response to vasodilatory drugs in small vessel disease: An OxHARP trial analysis.

Abstract

Background and aims: Vasodilating drugs targeting the endothelium could reduce long-term harms due to cerebral small vessel disease (cSVD) but there are no commonly accepted methods to measure short-term disease activity or drug response. In the OxHARP clinical trial, we determined the most sensitive physiological markers of treatment response to sildenafil versus placebo on either transcranial ultrasound (TCD) or magnetic resonance imaging (MRI), and their validity compared to disease severity and other measures of other physiological mechanisms.

Methods: In the OxHARP double-blind, randomized, placebo-controlled crossover trial we measured aortic blood pressure, mean flow velocity (MFV), cerebral pulsatility, cerebrovascular conductance index (CVCi = MFV/aortic mean BP), cerebral perfusion (pcASL-MRI) and cerebrovascular reactivity to inhaled CO2 on TCD (CVR-TCD) and MRI in white (CVR-WM), gray (CVR-GM) and white matter hyperintensities (CVR-WMH). Effects of 3 weeks of sildenafil were compared to placebo. Validity of markers were determined by between-visit repeatability (intraclass correlation coefficient (ICC)); associations with CVR-TCD, CVR-WMH and CVR-GM; associations with other markers; the magnitude of response, and sensitivity, to sildenafil.

Results: In 69 participants, repeatability was greatest for MFV, pulsatility, CVCi and CVR-WMH (ICC > 0.8), very good for CVR-TCD and GM-perfusion (ICC > 0.7), and good for CVR-GM (ICC > 0.6). CVR-TCD was associated with CVR on MRI (CVR-WMH: r² = 0.12, p = 0.02; CVR-GM: r² = 0.22, p = 0.001), while blood flow measures on TCD (MFV, CVCi) were associated with CVR-TCD and perfusion-MRI (all p < 0.05). All markers were associated with WMH volume and improved by sildenafil, but CVCi was most sensitive, requiring only 20 patients for a crossover trial at 80% power, compared to 26 for GM-perfusion or 84 for CVR-GM.

Conclusions: Multiple markers were associated with cSVD, but no single marker reflected all physiological drug effects. CVCi and gray matter perfusion on MRI were the most sensitive markers of disease activity and drug response, although CVR indices may be more specific for endothelial dysfunction.