Variation in Serial TCD Velocity Measurements in the TCD with Transfusions Changing to Hydroxyurea (TWiTCH) Trial

Introduction: The multicenter phase 3 randomized clinical trial TCD With Transfusions Changing to Hydroxyurea (TWiTCH, NCT01425307) investigated whether hydroxyurea could maintain TCD velocities and prevent stroke in children with sickle cell anemia and abnormal TCD velocities. The trial was stopped early after reaching the primary endpoint, when the average TCD velocity in the alternative (hydroxyurea) arm was found to be non-inferior and even slightly better (138 versus 143 cm/sec) than the standard (transfusion) arm. A planned secondary analysis was to determine the variation in serial TCD velocity measurements for individual study participants, to document the normal TCD velocity fluctuations that occur and to inform the need for frequent TCD evaluations in this clinical setting.

Methods: All TCD examinations collected in TWiTCH were analyzed for the maximum time-averaged mean velocity (TAMV) measured in the main intracranial arteries. TCD studies were performed by certified examiners using identical non-imaging instruments, no more than one week before a scheduled transfusion or phlebotomy procedure. Measurements on the index side, defined as the cerebral hemisphere with the higher mean arterial velocity at baseline assessment, were used for these statistical analyses. TCD values were analyzed according to four phases of the trial: (1) screening with three monthly TCD examinations performed at Weeks -8, -4, and 0 before randomization; (2) initial treatment period (hydroxyurea dose escalation with overlap transfusions or transfusions only) with three quarterly TCD measurements at Weeks 12, 24, and 36; (3) steady-state treatment period (hydroxyurea only or transfusions only) with four quarterly TCD measurements at Weeks 48, 60, 72, and 84; and (4) study exit with three monthly TCD examinations performed at Weeks 96, 100, and 104 or during the accelerated close-out period. The within-patient variance and standard deviation of the TAMV values were found using a linear mixed model, which was calculated using SAS Version 9.3.

Results: TAMV measurements on the index side from 1458 TCD examinations on 121 randomized patients formed the overall dataset. TAMV values ≥170 cm/sec were identified in 140 measurements on 40 children (9.6% of TCD values) and exceeded 200 cm/sec in 1.8% of examinations (26 values, 8 children) during the study treatment phase. The within-patient TCD variation, representing fluctuation of repeated TCD measurements in the same study participant, was 12.0 cm/sec for the entire cohort during the initial screening phase (363 TCD measurements). In the initial treatment phase, TCD variation in the hydroxyurea arm (180 values, 60 children) was 10.5 cm/sec, similar to 10.2 cm/sec on the transfusion arm (183 values, 61 children). In the steady-state treatment phase, the TCD variation in the hydroxyurea arm was 10.2 cm/sec (207 values, 57 children), compared to 12.3 cm/sec on the transfusion arm (212 values, 58 children). In the final exit phase, TCD variation in the hydroxyurea arm was 9.8 cm/sec (155 values, 58 children), similar to 10.2 cm/sec on the standard arm (155 values, 57 children). TCD variation was greater with higher baseline TCD velocity and shorter transfusion duration, but was not affected by age or gender. For example, participants with baseline TCD velocity of <170 cm/sec had a post-randomization measurement variation of 10.8 cm/sec, while those ≥170 cm/sec had higher variation (13.7 cm/sec, p=0.001).

Conclusions: The TWiTCH trial provides prospective data regarding the biological variation in serial TCD values that occurs in children with previous abnormal TCD examinations but no severe vasculopathy. The observed within-patient variation in TCD velocities was consistently 10-14 cm/sec, with higher variation observed in children with higher baseline TCD velocities. This fluctuation can be expected with serial TCD examinations and should be considered before making treatment decisions based on a single TCD measurement. Since comparison of entry and exit TCD values in the TWiTCH trial documented that hydroxyurea at maximum tolerated dose was non-inferior to transfusions, frequent TCD examinations may not be warranted.