Evaluation of hemoglobin S percent threshold to prevent severe pain events: a secondary analysis of the SIT trial

TO THE EDITOR:

The 2020 American Society of Hematology guidelines suggested the use of blood transfusion therapy after all other measures to decrease the incidence rates of vaso-occlusive pain events have failed. However, unlike regular blood transfusion therapy for primary and secondary stroke prevention, for which the goal is hemoglobin (Hb) S level ≤30%, no optimal Hb S threshold has been established to attenuate the severe acute vaso-occlusive pain event incidence rate. The goal of achieving a Hb S <30% is based on the strategy for primary and secondary stroke prevention in children with sickle cell anemia via blood transfusion therapy. The American Society of Hematology 2020 guidelines for sickle cell disease¹ suggested the use of blood transfusion therapy for decreasing acute vaso-occlusive pain events only when other methods such as hydroxyurea, L-glutamine, and crizanlizumab have proven unsuccessful. Despite these recommendations, no Hb S percentage goal threshold was established.

In a secondary data analysis of the silent cerebral infarct transfusion (SIT) trial,² we tested the hypothesis that among children receiving regular blood transfusion therapy to keep the mean Hb S level <30%, participants who achieved a mean Hb S level <30% would have lower incidence rates of severe acute vaso-occlusive pain events requiring hospitalization than children regularly receiving transfusion with mean Hb S level ≥30%. Data to support our hypothesis would provide, to our knowledge, the first evidence for the optimal Hb S threshold levels to decrease the incidence rate of acute vaso-occlusive pain events in children with sickle cell anemia (HbSS).

In the multicenter, single-blind clinical trial for children with sickle cell anemia, each participant was randomly allocated to receive regular blood transfusion therapy (transfusion group) or standard care (observation group) in 29 centers in the United States, Canada, France, and the United Kingdom. The eligibility criteria include participants aged between 5 and 15 years, with no history of stroke, ≥1 silent cerebral infarcts on magnetic resonance imaging of the brain, and a neurologic examination showing no focal deficits associated with the magnetic resonance imaging. Blood transfusion therapy in the SIT trial was leukocyte-poor, Hb S–negative, packed red blood cells (RBCs), administered at 10 to 15 mL/kg per transfusion, at a rate not exceeding 5 mL/kg per hour. The blood was matched for ABO and Rh (C, D, E, and Kell) antigens. The transfusion therapy goal was to keep the maximum Hb S level <30%. The optimal strategy was RBC apheresis; however, manual exchange and simple transfusion were acceptable because of the decreased net iron balance.³ 

For the secondary analysis, participants were included only from the trial's transfusion arm. The primary end point was the incidence rate of hospitalization for severe vaso-occlusive pain episodes among those receiving only regular blood transfusion therapy. We estimated a zero-inflated negative binomial regression model (Table 1) to determine the association between the incidence rate for severe acute vaso-occlusive pain events requiring hospitalization and independent covariates. The covariates included Hb S level at baseline and mean Hb S level (<30% and ≥30%) throughout the transfusion period.

A total of 90 participants received regular blood transfusion therapy in the SIT trial and were included in this analysis. Participants were followed up and received transfusion for a median of 2.9 years. Among participants who received transfusion, 58.9% (53/90) and 41.1% (37/90) had a mean Hb S level during transfusion of ≥30% and <30%, respectively. The mean age at the time of randomization was 10.1 years in the Hb S ≥30% group and 9.8 years in the Hb S <30% group (P = .55). In the Hb S ≥30% group, 35.8% (19/53) were females, and in the Hb S <30% group, 40.5% (15/37) were females (P = .65). The median Hb S percent at baseline or before transfusion was similar in both groups; 83.0 (interquartile range, 74.0-89.5) for the Hb S ≥30% group and 86.5 (interquartile range, 71.8-92.2) for the Hb <30% group (P = .31). The baseline mean total Hb was statistically different, with 8.0 g/dL in the Hb S ≥30% group and 7.5 g/dL in the Hb S <30% group (P = .01).

Among the participants, 10.8% (4/37) who had a mean Hb S <30% and 20.8% (11/53) of those with a mean Hb S ≥30% experienced at least 1 hospitalization for a severe acute vaso-occlusive pain event. The mean Hb S levels in the <30% and ≥30% groups were 25.4% and 37.5%, respectively. The incidence rates for hospitalization for severe acute vaso-occlusive pain events in the <30% and ≥30% groups were 9.4 (95% confidence interval, 4.5-17.2) and 29.2 per 100 person-years (95% confidence interval, 21.0-39.5), respectively (P = .078). In the multivariable analysis (Table 1), the incidence rate ratio for severe acute vaso-occlusive pain events was not associated with a mean Hb S level <30% (incidence rate ratio, 0.65; P = .43).

As a secondary analysis of the SIT trial with a sample size of 90 participants, our results need confirming in a prospective trial to determine the optimal maximum Hb S level to prevent recurrent severe vaso-occlusive pain. However, our results provide preliminary evidence that the maximum Hb S percent threshold for primary and secondary prevention of strokes may not be the same as for preventing acute pain episodes requiring hospitalization. A significant limitation is that 83% of the participants did not have any serious vaso-occlusive pain events requiring hospitalization, indicating that pain events occur at relatively low frequency in children receiving regular transfusions.

In this secondary analysis of the SIT trial, we demonstrate, to our knowledge, for the first time that regular RBC transfusions to achieve a mean Hb S level <30% were as effective as a mean Hb S level ≥30% in decreasing the incidence rate of severe acute vaso-occlusive pain events requiring hospitalization. The 95% confidence intervals are wide for the pain incidence, suggesting that the lower threshold of <30% may be beneficial or the higher threshold <50% may be of no benefit (supplemental Material). Given the broad confidence intervals, the Hb S <30% may be beneficial in decreasing severe acute vaso-occlusive pain event incidence rate. Thus, for patients started on regular blood transfusions to decrease their severe vaso-occlusive pain incidence rate, we would start with a goal of a maximum Hb S level of <50%. If the participant has a beneficial effect with a concomitant decrease in the severe acute vaso-occlusive pain event incidence rate, we will not lower the therapeutic goal of the Hb S to <30%; however, if the patient had no significant change in the severe acute vaso-occlusive pain event incidence rate after 3 to 6 months, we would decrease the threshold to a maximum level of <30%. Our results only apply to secondary prevention of acute vaso-occlusive pain events in children with sickle cell disease.

Acknowledgments: The authors thank the Burroughs Welcome Fund (J.M.) and the American Society of Hematology (J.M.) 2024 ASH Medical Student Physician-Scientist Award for supporting this project.

Contribution: M.R.D. and J.M. designed research and provided data; A.R.G.-B. helped write the manuscript; M.R. performed statistical analysis and helped write multiple versions of the manuscript; and M.R.D. and J.M. wrote the initial draft of the manuscript, and analyzed and interpreted data.

Conflict-of-interest disclosure: M.R.D. reports being the chair of the steering committee for a secondary prevention of priapism trial in sickle cell disease supported by Novartis Pharmaceutical Corporation. The remaining authors declare no competing financial interests.

Correspondence: Michael R. DeBaun, Vanderbilt-Meharry, Center of Excellence in Sickle Cell Disease, Vanderbilt University School of Medicine, 2200 Children’s Way, 11206DOT, Nashville, TN 37232-9000; email: m.debaun@vumc.org.