Current understanding of causes of intracerebral haemorrhage (ICH) is based on associations with single risk factors or imaging features in isolation. However, people with ICH usually have multimorbidity and multiple risk factors for the disease. Although hypertension is the strongest risk factor with a population attributable risk of 56% due to its frequency (Lancet 388, 761-775) two-thirds of people with ICH in Lothian have at least two risk factors and each person has a median of three chronic comorbidities (personal data).
Multimorbidity is common in people with stroke (Stroke 50, 1919-1926), but the association between multimorbidity and ICH incidence is unknown. Furthermore, there may be specific combinations of risk factors, drugs associated with bleeding, and comorbidities (hereafter referred to as ‘multimorbidity clusters’) that are associated with incidence of ICH. Multimorbidity is associated with a higher risk of death after ICH and may affect functional outcome(Neurology 82, 340-350) and having multiple cardiovascular conditions increases morbidity and mortality after ischaemic stroke, but we do not know if there is an association between multimorbidity clusters and MACE after ICH (International Journal of Stroke 3, 237-248).
Computational approaches can identify underlying patterns in data, and could provide a deeper understanding of the interactions between multimorbidity clusters and ICH which can be difficult to detect using conventional methods in an uncommon disease. Algorithmic approaches can identify high-risk multimorbidity clusters, which could be associated with outcome and treatment effect. Furthermore, a hypothesis-free approach may provide insight into previously undetected associations (Nature 579, 494-496).
HYPOTHESES
1. Multimorbidity clusters are associated with intracerebral haemorrhage incidence and pathological subtype.
2. Multimorbidity clusters are associated with MACE outcomes after intracerebral haemorrhage.