Edinburgh Research Explorer

The Edinburgh CT and genetic diagnostic criteria for lobar intracerebral haemorrhage with cerebral amyloid angiopathy: model development and diagnostic test accuracy study

Research output: Contribution to journalArticle

Related Edinburgh Organisations

Open Access permissions



  • Download as Adobe PDF

    Rights statement: © The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license.

    Final published version, 485 KB, PDF document

Original languageEnglish
JournalLancet Neurology
Early online date10 Jan 2018
Publication statusE-pub ahead of print - 10 Jan 2018


Background: Identifying lobar spontaneous intracerebral haemorrhage (ICH) associated with cerebral amyloid angiopathy (CAA) is important because it has a higher risk of recurrent ICH compared with arteriolosclerosis-associated ICH. We aimed to develop a prediction model for identifying CAA-associated lobar ICH using CT features and genotype.
Methods: We identified adults with first-ever ICH diagnosed by CT, who died and underwent research post-mortem in a prospective, population-based inception cohort study. We determined apolipoprotein E (APOE) genotype and radiologists rated CT imaging appearances masked to clinical, genetic and histopathological features. A neuropathologist rated brain tissue for small vessel diseases including CAA masked to clinical, radiographic and genetic features. We used CT and APOE data in a logistic regression model, which we internally validated using bootstrapping, to predict the risk of CAA-associated lobar ICH, derive diagnostic criteria, and estimate diagnostic accuracy.
Findings: Among 110 adults (median age 83 years [IQR 76-87], 49 [45%] male), ICH was lobar in 62 (56%), deep in 41 (37%) and infratentorial in seven (6%). Among 62 lobar ICH, 36 (58%) were CAA-associated and were independently associated with subarachnoid haemorrhage (32/36 [89%] versus 11/26 [42%]; p=0·01), ICH with finger-like projections (14/36 [39%] versus 0/26 [0%]; p=0·04) and APOE ε4+ (18/36 [50%] versus 2/26 [8%]; p=0·002). A diagnostic model for CAA-associated lobar ICH using these three variables had excellent discrimination (c-statistic 0·92, 95%CI 0·86-0·98), confirmed by internal validation. The rule-out criteria neither subarachnoid haemorrhage nor APOE ε4+ had 100% sensitivity (95%CI 88-100%). The rule-in criteria subarachnoid haemorrhage and either APOE ε4+ or finger-like projections had 96% specificity (95%CI 78-100%).
Interpretation: The CT and APOE genotype prediction model for CAA-associated lobar ICH shows excellent discrimination in this cohort, but requires external validation. The Edinburgh ‘rule-in’ and ‘rule-out’ diagnostic criteria might inform prognostic and therapeutic decisions that depend on identifying CAA-associated lobar ICH.

Download statistics

No data available

ID: 47405204