Hydroxycarbamide versus chronic transfusion for maintenance of transcranial doppler flow velocities in children with sickle cell anaemia-TCD With Transfusions Changing to Hydroxyurea (TWiTCH): a multicentre, open-label, phase 3, non-inferiority trial

Abstract

Background: For children with sickle cell anaemia and high transcranial doppler (TCD) flow velocities, regular blood transfusions can effectively prevent primary stroke, but must be continued indefinitely. The efficacy of hydroxycarbamide (hydroxyurea) in this setting is unknown; we performed the TWiTCH trial to compare hydroxyurea with standard transfusions.

Methods: TWiTCH was a multicentre, phase 3, randomised, open-label, non-inferiority trial done at 26 paediatric hospitals and health centres in the USA and Canada. We enrolled children with sickle cell anaemia who were aged 4-16 years and had abnormal TCD flow velocities (≥ 200 cm/s) but no severe vasculopathy. After screening, eligible participants were randomly assigned 1:1 to continue standard transfusions (standard group) or hydroxycarbamide (alternative group). Randomisation was done at a central site, stratified by site with a block size of four, and an adaptive randomisation scheme was used to balance the covariates of baseline age and TCD velocity. The study was open-label, but TCD examinations were read centrally by observers masked to treatment assignment and previous TCD results. Participants assigned to standard treatment continued to receive monthly transfusions to maintain 30% sickle haemoglobin or lower, while those assigned to the alternative treatment started oral hydroxycarbamide at 20 mg/kg per day, which was escalated to each participant's maximum tolerated dose. The treatment period lasted 24 months from randomisation. The primary study endpoint was the 24 month TCD velocity calculated from a general linear mixed model, with the non-inferiority margin set at 15 cm/s. The primary analysis was done in the intention-to-treat population and safety was assessed in all patients who received at least one dose of assigned treatment. This study is registered with ClinicalTrials.gov, number NCT01425307.

Findings: Between Sept 20, 2011, and April 17, 2013, 159 patients consented and enrolled in TWiTCH. 121 participants passed screening and were then randomly assigned to treatment (61 to transfusions and 60 to hydroxycarbamide). At the first scheduled interim analysis, non-inferiority was shown and the sponsor terminated the study. Final model-based TCD velocities were 143 cm/s (95% CI 140-146) in children who received standard transfusions and 138 cm/s (135-142) in those who received hydroxycarbamide, with a difference of 4·54 (0·10-8·98). Non-inferiority (p=8·82 × 10(-16)) and post-hoc superiority (p=0·023) were met. Of 29 new neurological events adjudicated centrally by masked reviewers, no strokes were identified, but three transient ischaemic attacks occurred in each group. Magnetic resonance brain imaging and angiography (MRI and MRA) at exit showed no new cerebral infarcts in either treatment group, but worsened vasculopathy in one participant who received standard transfusions. 23 severe adverse events in nine (15%) patients were reported for hydroxycarbamide and ten serious adverse events in six (10%) patients were reported for standard transfusions. The most common serious adverse event in both groups was vaso-occlusive pain (11 events in five [8%] patients with hydroxycarbamide and three events in one [2%] patient for transfusions).

Interpretation: For high-risk children with sickle cell anaemia and abnormal TCD velocities who have received at least 1 year of transfusions, and have no MRA-defined severe vasculopathy, hydroxycarbamide treatment can substitute for chronic transfusions to maintain TCD velocities and help to prevent primary stroke.

Funding: National Heart, Lung, and Blood Institute, National Institutes of Health.

Figure 1

TWiTCH CONSORT flow diagram showing the enrollment, randomisation, and follow-up of the TWiTCH study participants.

Figure 2

Laboratory parameters based on intention-to-treat population: Panel A = hemoglobin concentration; Panel B = mean corpuscular volume; Panel C = %HbS; Panel D = %HbF; Panel E = white blood cell (WBC) count; Panel F = absolute neutrophil count (ANC); Panel G = absolute reticulocyte count (ARC); Panel H = serum ferritin. Complete blood counts and reticulocytes were measured locally, while hemoglobin electrophoresis and ferritin were analysed centrally. Data are illustrated as mean ± 1 standard deviation. The Standard Treatment Arm is portrayed in dashes while the Alternative Treatment Arm is shown by the solid line. All parameters are significantly different at exit (p<0.0001) between treatment groups except Panel A, with p=0.800.

Figure 3

Primary study endpoint analysis of TCD velocities. The Standard (Transfusion) Arm data are portrayed as dashes while the Alternative (Hydroxyurea) Arm data are shown as a solid line. Panel A illustrates the TCD data using the mixed model statistical analysis, with baseline equivalence but divergence over time with lower velocities in the Alternative Arm. The curves are significantly different using the per-protocol non-inferiority comparison (p=8.82 × 10⁻¹⁶) and also by post-hoc analysis for superiority (p=0.023). Panel B illustrates the actual TCD velocity data in each arm as mean ± 1 standard deviation, including the number of participants evaluated at each 12-week time point. Panel C illustrates baseline (enrolment) and final (exit) maximum time-averaged mean TCD velocities for each participant. The lines at 170 and 200 cm/sec denote the normal and abnormal TCD boundaries, respectively. Open circles = Alternative Arm, Closed circles = Standard Arm.

Figure 3

Primary study endpoint analysis of TCD velocities. The Standard (Transfusion) Arm data are portrayed as dashes while the Alternative (Hydroxyurea) Arm data are shown as a solid line. Panel A illustrates the TCD data using the mixed model statistical analysis, with baseline equivalence but divergence over time with lower velocities in the Alternative Arm. The curves are significantly different using the per-protocol non-inferiority comparison (p=8.82 × 10⁻¹⁶) and also by post-hoc analysis for superiority (p=0.023). Panel B illustrates the actual TCD velocity data in each arm as mean ± 1 standard deviation, including the number of participants evaluated at each 12-week time point. Panel C illustrates baseline (enrolment) and final (exit) maximum time-averaged mean TCD velocities for each participant. The lines at 170 and 200 cm/sec denote the normal and abnormal TCD boundaries, respectively. Open circles = Alternative Arm, Closed circles = Standard Arm.