Hematological Abnormalities and Malnutrition Mediate Pathway between Cancer and Outcomes in Ischemic Stroke Patients

2020 
Abstract Background The present study aimed to examine whether variables including D-dimer, high-sensitivity C-reactive protein (hsCRP), hemoglobin, platelet count, and nutritional status mediate the pathway between cancer and ischemic stroke outcomes. Methods We reviewed data from consecutive patients with ischemic stroke admitted to Osaka University Hospital between January 1, 2006, and December 31, 2016. Patients with ischemic stroke were grouped according to the presence of cancer. Nutritional status was assessed using Controlling Nutritional Status (CONUT) scores. Mediation analyses were utilized to address the study aims. Results Among 1,570 patients with ischemic stroke, 185 (12%) had active cancer. Relative to patients with ischemic stroke in the non-cancer group, those in the cancer group exhibited higher National Institutes of Health Stroke Scale scores on admission, higher D-dimer and hsCRP levels, lower hemoglobin levels and platelet counts, higher CONUT scores, and poorer modified Rankin Scale scores at discharge. Mediation analysis revealed that D-dimer, hsCRP, hemoglobin, platelet count, and CONUT scores acted as mediators of poor prognosis in the cancer group. The association between the exposure and outcome variables was no longer significant in the models containing D-dimer and CONUT scores as mediator variables, suggesting that they were strong mediators. Regarding the association between the mediator and outcome variables, hemoglobin, platelet count, and CONUT exhibited non-linearity (p for non-linearity Conclusions D-dimer, hsCRP, hemoglobin, platelet count, and CONUT score act as mediators of poor prognosis in patients with ischemic stroke with comorbid cancer. Such abnormalities can help to predict ischemic stroke outcomes.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    22
    References
    1
    Citations
    NaN
    KQI
    []