Moderate fixed-dose hydroxyurea for primary prevention of strokes in Nigerian children with sickle cell disease: Final results of the SPIN trial

To the Editor:

In children with sickle cell anemia (SCA) living in high-income set- tings, the routine use of transcranial Doppler (TCD) measurements, coupled with monthly blood transfusion therapy for children with abnormal velocities (≥200 cm/s, non-imaging), resulted in a 92% relative risk reduction in strokes when compared to no treatment.¹ More recently, the TWITCH Trial demonstrated that children with abnormal TCD measurements and no evidence of magnetic reso- nance angiography-defined cerebral vasculopathy, after 1 year of initial regular blood transfusion therapy, could be transitioned to hydroxyurea therapy at the maximum tolerated dose.² However, initial regular blood transfusion therapy for primary stroke preven- tion is not feasible for the vast majority of the children with SCA living in low- and middle-income settings.³ Furthermore, limited data are available to determine whether initial treatment with hydroxyurea rather than blood transfusion therapy is a durable therapy for primary stroke prevention.

To address this gap in knowledge for preventing strokes in chil- dren with SCA living in Africa, we tested the hypothesis that moder- ate fixed-dose hydroxyurea (~20 mg/kg/d) for primary stroke prevention was feasible in a low-income setting (Kano, Nigeria). We previously demonstrated in the Stroke Prevention in Nigeria (SPIN) Trial (NCT01801423) that hydroxyurea was acceptable and safe for children with SCA and abnormal TCD measurements.⁴ We have extended the feasibility trial for approximately 5 years to test the hypotheses that moderate fixed-dose hydroxyurea will: (a) not result in an excess incidence rate of serious adverse events (death or stroke) when compared to a group of children with SCA and TCD measure- ments <200 cm/s not receiving hydroxyurea; (b) be comparable to the stroke incidence rate in children with abnormal TCD measurements in the STOP Trial that were initially and only treated with regular blood transfusion therapy.¹ We report the final results of the SPIN Trial.

At trial entry, eligible participants were screened with the TCD non-imaging technique to determine increased stroke risk, defined as two independent measurements of time-averaged mean maximum velocity (TAMMV) of ≥200 cm/s or one measurement of ≥220 cm/s in the middle cerebral artery (MCA). Families of children with abnor- mal TCDs were offered regular blood transfusion therapy as standard care. If families refused regular blood transfusion, moderate fixed- dose hydroxyurea (~20 mg/kg/d) was offered via the SPIN trial; chil- dren were evaluated monthly with surveillance complete blood counts (CBCs). To primarily address whether hydroxyurea was associated with an increased incidence of death when compared to children with SCA in the same age group not treated with hydroxyurea, we included a com- parison group of children with SCA who were screened and had a TCD measurement of <200 cm/s and who agreed to be followed as part of routine care. Malaria prophylaxis and penicillin prophylaxis were pre- scribed to all participants as standard care. Serious adverse events includ- ing death or stroke in the treatment and comparison groups, based on the World Health Organization criteria,⁵ were recorded and compared.

Twenty-nine children with abnormal TCD measurements were identified and treated with moderate fixed-dose hydroxyurea for pri- mary stroke prevention, and 206 children were included in the com- parison group. No caregiver of a child with abnormal TCD measurements elected to have their child treated with regular blood transfusion therapy. Among the comparison group, standard care TCD screening was performed on average every 12 months. Four children developed abnormal TCD measurements and crossed over to the treatment group. Baseline features and clinical outcomes of the treat- ment and comparison groups are shown in Table 1.

TABLE 1.

Clinical and laboratory features of participants. Baseline characteristics and clinical outcomes in children with SCA and abnormal transcranial Doppler (TCD) measurement (non-imaging ≥200 cm/s) and comparison group (non-imaging <200 cm/s) followed prospectively for 4.8 and 4.9 years, respectively.

Participant characteristics	All Participants (n = 235)	Comparison group (n = 206)	Treatment group (n = 29)	P value
Age, median (IQR)	8.1 (6.1–10.3)	8.2 (6.2–10.4)	7.0 (5.6–9.3)	.067
Sex, male, percent (n)	46.4 (109)	46.1 (95)	48.3 (14)	.827
Time followed, median (IQR) years	4.9 (4.6–5.2)	4.9 (4.6–5.2)	4.8 (3.7–5.6)	.408
Ethnicity, Hausa-Fulani, percent (n)	89.8 (211)	91.7 (189)	75.9 (22)	.016b
Head of household university/professional education, percent (n), (n = 230)	58.3 (134)	60.9 (123)	39.3 (11)	.030
Height, (cm), median (IQR)	124.0 (115.0–132.0)	126.0 (115.8–132.0)	120.0 (114.0–131.0)	.227
Weight (kg), median (IQR)	20.0 (17.0–24.0)	20.0 (17.0–24.0)	20.0 (16.5–23.5)	.576
BMI (kg/m2), mean (std. dev)	13.4 (1.9)	13.4 (2.0)	13.7 (1.7))	.534c
TCD at baseline, (cm/s), median (IQR)	140.0 (120.0–159.0)	134.0 (118.0–150.0)	208.0 (205.0–226.0)	<.001a
All hospitalizations per 100 person-y, mean (std. dev.)	33.3 (40.7)	34.4 (41.4)	26.0 (36.6)	.113d
Acute chest syndrome requiring hospitalization, per 100 person-y, mean (std. dev.)	1.42 (6.0)	1.53 (6.2)	0.76 (3.7)	.566d
Acute vaso-occlusive pain requiring hospitalization, no associated fever, per 100 person-y, mean (std. dev.)	13.7 (24.2)	14.9 (24.9)	5.3 (17.2)	.002d
Acute vaso-occlusive pain requiring hospitalization with associated fever, per 100 patient-y, mean (std. dev.)	12.6 (19.8)	12.6 (20.0)	12.2 (18.4)	.924d
Fever requiring hospitalization, no associated pain, per 100 patient-y, mean (std. dev.)	3.0 (8.5)	3.0 (7.4)	3.1 (13.9)	.936d
Stroke, percent (n)	1.7 (4)	1.5 (3)	3.4 (1)	.412b
Stroke, events per 100 patients-y, mean (std. dev.)	0.47 (2.8)	0.43 (3.0)	0.76 (3.9)	.603d
Death, percent (n)	8.1 (19)	9.2 (19)	0.0 (0)	.141b
Death, rate per 100 patients-y, mean (std. dev.)	1.8 (18.9)	2.0 (19.6)	0.0 (0.0)	.081d

^aChi-squared test for categorical variables or Mann-Whitney U-test for continuous variables, unless otherwise noted.

^bFisher’s exact test.

^cT-test.

^dMid-p exact test.

The median time on hydroxyurea therapy (follow-up time) was 4.8 years (IQR: 3.7–5.6). The stroke incidence rate among participants on hydroxyurea was 0.76 per 100 person-years (95% CI: 0.11–5.24), which was not significantly different from participants in the comparison group who did not have an abnormal TCD velocity (incidence rate 0.43 per 100 person-years; 95% CI: 0.16–1.15) (P = .603). Importantly, the inci- dence rate of strokes was significantly lower than the stroke incidence rate reported in the standard care group for the STOP trial (10.7 per 100 person-years, with a total of 102 person-years), where all children had abnormal TCD velocity.¹ The stroke incidence rate among partici- pants treated with hydroxyurea was also similar to the stroke incidence rate in the STOP trial transfusion group (0.9 per 100 person-years with a total of 110 patient-years).¹

In 29 participants monitored with monthly CBCs (n = 1560), no participant had hydroxyurea withheld due to myelosuppression (abso- lute neutrophil count <1000 × 10⁹/L) or low platelet count (<80 × 10⁹/L) on two consecutive evaluations. These data suggest that, in low- and middle-income settings, monthly or even quarterly routine CBC assessment for myelosuppression associated with mod- erate fixed-dose hydroxyurea has limited clinical utility.

A total of 19 deaths occurred in the comparison group, with no deaths in the treatment group. One death occurred in a child that was originally in the treatment group, but the death occurred after the par- ticipant was withdrawn from the trial because of progressive renal dis- ease unrelated to study treatment. The death rate in the comparison group was 2.0 per 100 person-years. There was no statistically signifi- cant difference in the death rate between treatment and comparison groups (P = .081). The leading cause of death was suspected or confirmed malaria (defined as fever ≥37.5°C in the presence of Plas- modium falciparum in peripheral blood smear or a positive rapid diag- nostic antigen test) occurring in 79% of the participants (15 of 19). As expected, we found a statistically significant lower incidence rate of uncomplicated, afebrile acute vaso-occlusive pain events requiring hospitalization in the hydroxyurea treatment group than in the com- parison group (5.3 per 100 person-years and 14.9 per 100 person- years, respectively; P = .002). There was no difference in the incidence rates of febrile acute vaso-occlusive pain events requiring hospitaliza- tions. Incidence rates for other indications for hospitalization are described in Table 1. An early (3 months) and sustained (2 years) decrease in the initial abnormal TCD measurements was unexpected, but an important finding in our clinical trial (Figure 1). The results from this trial bolster pre-existing pooled data from 10 studies demonstrat- ing that hydroxyurea therapy significantly lowers TCD velocity.⁶

FIGURE 1.

Boxplots of serial TCD measurements in children treated with hydroxyurea. Screening, 3-month and 2-year TCD measurements in children receiving moderate fixed-dose hydroxyurea therapy (~20 mg/kg/d) for primary prevention of strokes. The highest time-averaged mean maximum velocity in the right and the left middle cerebral artery was obtained in 25 individuals assessed at baseline, 3 months, and 2 years after starting therapy, paired Wilcoxon signed rank test, baseline to 3 months (P < .001); baseline to 2 years (P < .001)

In Nigeria and other low- and middle-income settings, expenses for hydroxyurea therapy and laboratory monitoring are self-pay for the majority of families of children with SCA. To ensure the sustain- ability of stroke prevention efforts in low- and middle-income set- tings, minimizing costs and inconvenience of both treatment and laboratory surveillance is critical. In our trial, hydroxyurea was pro- duced locally (Bond Chemical, Awe, Oyo State of Nigeria), and at a rel- atively inexpensive cost of approximately $4.00 for 1 package (30 tablets) of Oxyurea 500mg per month for a 20 mg/kg/d dose per child that weighs 25 kg. The cost of a CBC is approximately $3.00 USD; therefore, minimizing CBC surveillance would decrease family medical expenses. Our findings are similar to the BABY HUG Trial, where moderate fixed-dose hydroxyurea therapy was not associ- ated with absolute neutrophil count <500/mm3 or platelet count <80 × 103/mm³.⁷ Our study results did not reveal either laboratory or clinical evidence of myelosuppression. Taken together, the results in this clinical trial and the BABY HUG Trial strongly suggest that monthly CBCs are not needed to identify myelosuppression at a moderate fixed-dose of hydroxyurea. Based on these data, the optimal CBC interval is unknown. The best CBC interval in this set- ting may be determined in the Primary Prevention of Stroke in Chil- dren With SCD in sub-Saharan Africa II (SPRING) Trial (NCT02560935) when children with SCA and abnormal TCD mea- surements (n=220), will be randomly allocated to moderate fixed- dose vs. low fixed-dose hydroxyurea therapy, 20 vs. 10 mg/kg/day, respectively) for 3 years.

In summary, the results of our NIH/NINDS-funded feasibility SPIN Trial provide additional evidence supporting the American Soci- ety of Hematology recommendation that for children with SCA and with abnormal TCD measurements, living in a low- and middle-income settings, initial treatment with moderate fixed-dose hydroxyurea (~20 mg/kg/d) is comparable to initial treatment with regular blood transfusion in the STOP trial (0.9 per 100 person-years with a total of 110 patient-years) and superior to no treatment.⁶ In addition, monthly CBC assessments for myelosuppression associated with moderate fixed-dose hydroxyurea have limited clinical utility in low- and middle- income settings.