Background: Cerebral small vessel disease (SVD) is a common cause of stroke and a major contributor to vascular dementia, but there are few treatments to delay disease progression. Neuroinflammation and increased blood-brain barrier (BBB) permeability have been suggested to play a role in pathogenesis, and both can be visualised in man using advanced imaging. However, whether they drive disease progression and can be targeted therapeutically is uncertain. In a rodent model of SVD, minocycline treatment reduced white matter damage and markers of inflammation and BBB permeability, and improved behavioural outcomes. We determined whether minocycline had a similar effect in patients with moderate to severe SVD. Methods: MINERVA (ISRCTN15483452) was a phase II randomised, double blind, placebo-controlled trial performed in a single UK centre. The study included 44 participants with moderate to severe SVD defined as a clinical lacunar stroke and confluent MRI white matter hyperintensities, who were randomly assigned 1:1 to either minocycline 100mg twice daily or placebo for three months. The co-primary outcome measures were volume of focal areas of increased microglial signal (determined using 11C-PK11195 positron emission tomography) and increased blood-brain barrier permeability (measured using dynamic contrast-enhanced [DCE] MRI). Secondary outcome measures included change in inflammatory biomarkers in serum. Results: 44 participants were recruited between September 2019 and June 2022. All 44 were included in a modified intention-to-treat analysis: 23 in the minocycline group and 21 in the placebo group. Mean age was 73.0 ± 8.9 years in the treatment group and 66.5 ± 11.9 in the placebo group (p = 0.04), but other baseline demographics were balanced. Participants were recruited at a mean of 23.1 ± 24.7 months post stroke. Hotspots of increased 11C-PK11195 binding and increased BBB permeability were present in the participants. There was no treatment effect of minocycline on the change in percentage of 11C-PK11195 binding hotspots (RR 1.01, 95% CI 0.98 – 1.04), or on BBB permeability hotspots (RR 0.97, 95% CI 0.91 – 1.03). All secondary analyses, as well as the per-protocol analysis, showed no treatment effect. There was also no significant treatment effect on C-reactive protein levels or on any of the individual biomarkers in a proteomics panel targeting vascular inflammation and endothelial activation.
Interpretation: Although increased 11C-PK11195 binding consistent with microglial activation, and increased BBB permeability on DCE-MRI, are both present in SVD, treatment with minocycline did not reduce measurements of either process. Our findings do not support a phase 3 trial of minocycline therapy to delay SVD progression. Whether these pathophysiological mechanisms are causal, or secondary to tissue damage, remains to be determined.