Cerebral hemodynamic response to chronic transfusion is associated with cerebral vasculopathy in sickle cell anemia

Abstract Background Diffuse correlation spectroscopy (DCS) is a novel, non-invasive optical technique that measures an index of regional microvascular cerebral blood flow (CBF). Previously, we used DCS to demonstrate that transfusion significantly decreases CBF in SCA patients on chronic transfusion (Lee et al. 2022). This decreased response was expected because it is known that transfusion improves oxygen-carrying capacity, thereby reducing the need for compensatory hyperperfusion. However, while we observed an average decrease in CBF with transfusion, the CBF response was considerably variable across participants. In this secondary analysis, we explore factors contributing to this variability. Given that chronic inflammation, endothelial disfunction, and altered red blood cell biomechanics contribute to both macro- and microvascular complications in SCA, we test the hypothesis that microvascular CBF and its response to transfusion are associated with the presence of large vessel vasculopathy.Methods Details of the original study design can be found in (Lee et al. 2022). In brief, children with HbSS or HbSβ0 on chronic transfusion therapy were enrolled in an IRB-approved study at Children's Healthcare of Atlanta; exclusion criteria included prior history of stroke, Moyamoya, or previous surgical revascularization. An index of regional CBF was assessed in the frontal cortex immediately before and after transfusion with DCS. Vasculopathy was assessed via review of the clinical report from the most recent standard-of-care head MRI/MRA prior to and after the time of CBF measurement. Patients were considered to have vasculopathy if the report cited narrowing, stenosis, or occlusion of any cerebral vessel. Vasculopathy status was categorized as binary (present/absent) at the time of transfusion and categorically at follow-up (present pre-transfusion, never developed, developed at follow-up). To test the hypothesis that CBF and its response to transfusion are associated with vasculopathy, multivariate models were used that included terms for vasculopathy status (either at time of transfusion or at follow-up), age or sex, and their interaction, given that CBF was observed to be significantly correlated with age and CBF transfusion response was correlated with sex in the original analysis. Model fit was assessed using adjusted R2 values, and individual predictors were considered significant if p 0.05. All models were fitted using R (version 4.4.1).ResultsA total of 28 transfusion events were included in this analysis. The cohort was all HbSS and predominantly female (75%), with a median age of 14 (IQR 11,16) years. At the time of transfusion, 11 patients (39%) had evidence of vasculopathy on their most recent MRI/MRA, while 17 (61%) did not. The median time from MRA/MRI to transfusion was -0.76 (IQR -0.50, -1.19) years. Follow-up imaging was available for 26 of the 28 transfusions: 12 (43%) did not have vasculopathy, 3 (12%) developed vasculopathy, and 13 (42%) had persistent evidence of vasculopathy that was also seen on pre-transfusion imaging. The median time to follow-up imaging was 0.75 (IQR 0.34,1.51) years. In multivariate models accounting for the effect of age, lower pre-transfusion CBF was significantly associated with the presence of vasculopathy at the time of transfusion (model R² = 0.76, p0.001) and with vasculopathy status at follow-up (R² = 0.79, p 0.001). Multivariate models incorporating sex revealed that pre-existing vasculopathy (R² = 0.45, p0.001) and vasculopathy status at follow-up (R² = 0.50, p0.001) were significantly associated with an increase in CBF in response to transfusion.Conclusions Our results show that both pre-transfusion CBF and transfusion-induced CBF response measured with DCS were significantly associated with patient vasculopathy status at the time of transfusion and at follow-up, after accounting for the effects of age and sex. Notably, the pattern of lower CBF before transfusion followed by an increase afterwards in patients with vasculopathy is opposite to the expected hemodynamic response to transfusion (Gulliams et al., 2018; Juttukonda et al., 2019; Lee et al., 2022), which suggests altered cerebral hemodynamics in these patients that warrant further investigation. Prospective studies in larger cohorts are needed to determine if DCS-measured CBF dynamics can provide a predictor for vasculopathy development in the context of chronic transfusion.