Can peak systolic velocities be used for prediction of stroke in sickle cell anemia?

Background and purpose: Ischemic stroke occurs in at least 11% of patients with homozygous sickle cell anemia (SCD) by the time they turn 20 years old. High risk associated with distal intracranial internal carotid (ICA) and proximal middle cerebral artery (MCA) stenosis can be detected by transcranial Doppler (TCD). TCD screening offers the possibility of reducing the risk of first stroke significantly based on a paradigm tested and proven to be effective in a stroke prevention trial in sickle cell anemia (STOP). Children with high flow velocity in the ICA and MCA of 200 cm/s time average mean of the maximum (TAMM) or higher had a 10% per year risk of first stroke that was reduced to <1% with regular red cell transfusion (reduction of hemoglobin S <30%). The clinical application of the STOP results could be enhanced if criteria for treatment could be found that are based on peak systolic velocity (PSV), the measure more commonly used in vascular ultrasound practice. Objective: To compare PSV and end diastolic velocity (EDV) with TAMM for prediction of stroke and to derive PSV cutpoints for STOP protocol definitions of conditional and abnormal TCD. Using the STOP TCD and stroke outcome data to compare PSV and TAMM in terms of stroke prediction, PSV cutpoints comparable to those based on TAMM and used in STOP were derived. Because of their familiarity to the vascular ultrasound community, PSV cutpoints should be an important alternative to TAMM and may increase availability of screening and risk stratification for children with this disease. Materials and methods: Data from 1,937 baseline TCD studies from STOP were correlated with stroke outcome in those not treated with transfusion. Stroke prediction was assessed with survival analysis using TAMM, PSV and EDV as continuous variables individually and then pair-wise in the same model, which contained 53 stroke events. Results: PSV and EDV were highly correlated to the TAMM velocity (r = 0.94). The multivariate model for prediction indicated that TAMM velocity was a better predictor than EDV, and PSV and TAMM were approximately equivalent. PSV cutpoints defining the two relevant STOP risk categories - "conditional," which should lead to increased TCD surveillance, and "abnormal," which should lead to strong consideration for treatment according to STOP - were derived taking into consideration known differences in measurements between the dedicated Doppler systems (TCD) used in STOP and the transcranial Doppler imaging (TCDI) systems commonly used in clinical practice. The recommended PSV cutpoint for conditional TCD is 200 cm/s, and for abnormal TCD triggering consideration for treatment is 250 cm/s. Conclusion: Assuming TCDI equipment is used and the STOP protocol is applied, a PSV cutpoint of 200 cm/s is recommended as the threshold for increased TCD surveillance (comparable to a TCD TAMM of 170 cm/s in STOP); a PSV of 250 cm/s is recommended as the cutpoint at which, if confirmed in a second examination, chronic transfusion should be considered. Assuming the STOP scanning protocol is used, PSV is at least as good as TAMM and can be used to select children with SCD for treatment or increased surveillance to prevent first stroke.