Cerebral Amyloid Angiopathy-Related Inflammation and Biopsy-Positive Primary Angiitis of the CNS: A Comparative Study
Lou Grangeon,Grégoire Boulouis,Jean Capron,Fouzi Bala,Dimitri Renard,Nicolas Raposo,Ozlem Ozkul-Wermester,Aude Triquenot-Bagan,Xavier Ayrignac,David Wallon,Emmanuel Gerardin,Philippe Kerschen,Denis Sablot,Maïté Formaglio,Fernando Pico,Guillaume Turc,Marc Verny,Lisa Humbertjean,Marie Gaudron,Stéphane Vannier,Nelly Dequatre,Benoit Guillon,Clothilde Isabel,Caroline Arquizan,Olivier Detante,Sophie Godard,Barbara Casolla,Michael Levraut,Cédric Gollion,Mathieu Gerfaud-Valentin,Laurent Kremer,Laure Daelman,Nicolas Lambert,Sylvain Lanthier,Alexandre Poppe,Alexis Régent,David Weisenburger-Lile,Pierre Verdure,Gérald Quesney,Mathieu Vautier,Anne Wacongne,Eric Thouvenot,Jérémie Pariente,Sarah Coulette,Pierre M Labauge,Nadège Olivier,Thibaut Allou,Hélène Zephir,Antoine Néel,Saskia Bresch,Benjamin Terrier,Olivier Martinaud,Romain Schneckenburger,Thomas Papo,Chloé Comarmond-Ortoli,Eric Jouvent,Marie Subréville,Louis Poncet-Megemont,Muhammad Amer Khatib,François Lun,Carole Henry,Eloi Magnin,Quentin Thomas,Mathilde Graber,Yassine Boukriche,Geneviève Blanchet-Fourcade,Diana Ratiu,Christian Pagnoux,Emmanuel Touzé,Hubert de Boysson,Sonia Alamowitch,Ahmad Nehme
DOI: https://doi.org/10.1212/WNL.0000000000209548
IF: 9.9
2024-07-23
Neurology
Abstract:Background and objectives: Cerebral amyloid angiopathy-related inflammation (CAA-RI) and biopsy-positive primary angiitis of the CNS (BP-PACNS) have overlapping clinicoradiologic presentations. It is unknown whether clinical and radiologic features can differentiate CAA-RI from BP-PACNS and whether both diseases have different relapse rates. The objectives of this study were to compare clinicoradiologic presentations and relapse rates in patients with CAA-RI vs BP-PACNS. Methods: Patients with CAA-RI and BP-PACNS were enrolled from 2 retrospective multicenter cohorts. Patients with CAA-RI were biopsy-positive or met probable clinicoradiologic criteria. Patients with BP-PACNS had histopathologic confirmation of CNS angiitis, with no secondary etiology. A neuroradiologist read brain MRIs, blinded to the diagnosis of CAA-RI or BP-PACNS. Clinicoradiologic features were compared using univariable logistic regression models. Relapse rates were compared using a univariable Fine-Gray subdistribution hazard model, with death as a competing risk. Results: This study enrolled 104 patients with CAA-RI (mean age 73 years, 48% female sex) and 52 patients with BP-PACNS (mean age 45 years, 48% female sex). Patients with CAA-RI more often had white matter hyperintense lesions meeting the probable CAA-RI criteria (93% vs 51%, p < 0.001), acute subarachnoid hemorrhage (15% vs 2%, p = 0.02), cortical superficial siderosis (27% vs 4%, p < 0.001), ≥1 lobar microbleed (94% vs 26%, p < 0.001), past intracerebral hemorrhage (17% vs 4%, p = 0.04), ≥21 visible centrum semiovale perivascular spaces (34% vs 4%, p < 0.01), and leptomeningeal enhancement (70% vs 27%, p < 0.001). Patients with BP-PACNS more often had headaches (56% vs 31%, p < 0.01), motor deficits (56% vs 36%, p = 0.02), and nonischemic parenchymal gadolinium enhancement (82% vs 16%, p < 0.001). The prevalence of acute ischemic lesions was 18% in CAA-RI and 22% in BP-PACNS (p = 0.57). The features with the highest specificity for CAA-RI were acute subarachnoid hemorrhage (98%), cortical superficial siderosis (96%), past intracerebral hemorrhage (96%), and ≥21 visible centrum semiovale perivascular spaces (96%). The probable CAA-RI criteria had a 71% sensitivity (95% CI 44%-90%) and 91% specificity (95% CI 79%-98%) in differentiating biopsy-positive CAA-RI from BP-PACNS. The rate of relapse in the first 2 years after remission was lower in CAA-RI than in BP-PACNS (hazard ratio 0.46, 95% CI 0.22-0.96, p = 0.04). Conclusion: Clinicoradiologic features differed between patients with CAA-RI and those with BP-PACNS. Specific markers for CAA-RI were hemorrhagic signs of subarachnoid involvement, past intracerebral hemorrhage, ≥21 visible centrum semiovale perivascular spaces, and the probable CAA-RI criteria. A biopsy remains necessary for diagnosis in some cases of CAA-RI. The rate of relapse in the first 2 years after disease remission was lower in CAA-RI than in BP-PACNS.