To investigate whether neutrophil lymphocyte ratios (NLR) can predict stroke subtypes in young strokes and differentiate ischaemic and hemorrhagic strokes.

Inflammatory changes have been associated with cerebrovascular disease and in all stages of acute ischaemic strokes (AIS). Neutrophil lymphocyte ratios have been correlated with infarct volume, short-term mortality in AIS and, length of stay. However data on young strokes is limited.

Patient registry data was used to collect clinical details on all patients under the age of 66 years admitted to a single stroke centre with stroke between September 2019 and January 2021. Patients with inflammatory disorders, concurrent malignancies, renal disease, active infection, haematological disorders, those taking immunosuppressive medications and if sufficient clinical data was unavailable were excluded. Ischaemic strokes were further classified into subgroups in terms of TOAST criteria. Multinomial logistic regression was used to assess whether admission NLR levels can predict stroke subtype. Independent samples T Test was used to assess NLR group differences between haemorrhagic and ischaemic strokes.

80 patients had ischaemic stroke. 51 (64%) were male. The mean age was 55.14 years (+/- 10 years). Stroke subtypes were: small vessel disease (25%), large vessel atherosclerosis (26%), cardioembolic (18%), other defined (12%) and cryptogenic (19%). No relationship was found between NLR and TOAST categories using multinomial regression.

11 patients had haemorrhagic stroke. Mean NLR for AIS was 3.6 and 3.9 for hemorrhagic, however difference was not significant (p= .284). When small and large vessel atherosclerosis related AIS only were considered, mean NLR was 3.77 with no group difference compared with haemorrhagic infarcts.

Although there is evidence to suggest that patients with atherosclerosis related causes for ischaemic strokes have inflammatory components to disease mechanism, this was not reflected in our data using NLR. Larger samples are needed for further sub-typing.