This project was first presented at the 2021 American Society of Hematology Annual Meeting with the title: “Splenic Complications in Pediatric Sickle Cell Disease:

A Retrospective Cohort Review.” An abstract of the presentation was published in a special supplement of the journal Blood.

Conflicts of Interest:

Dr. Alex George: research funding from Novo Nordisk; membership in a study adjudication committee and a data safety monitoring board for Pfizer; royalties from Walter-Kluwers.

Dr. Shannon Conneely: no conflicts to declare.

Dr. Ross Mangum: no conflicts to disclose.

Dr. Titilope Fasipe: consulting services for Pfizer, Novartis, and Novo Nordisk.

Dr. Philip Lupo: no conflicts to disclose.

Dr. Michael Scheurer: no conflicts to disclose.

ABSTRACT

Objective: To delineate the natural history of splenic complications other than the loss of splenic function in children with sickle cell disease (SCD), we performed a retrospective chart review of patients with SCD treated at the Texas Children’s Hospital.

Methods: We determined the dates of diagnoses of splenic complications, the number of ASSC events, and hydroxyurea treatment in patients with SCD. We also examined the association of hydroxyurea therapy with the onset and severity of ASSC.

Results: The cumulative prevalence of splenic complications was 24.7% for splenomegaly, 24.2% for ASSC, 9.6% for hypersplenism, and 5.6% for splenectomy. The cumulative prevalence of all splenic complications was highest in patients with hemoglobin Sβ⁰ (69.2%), intermediate in hemoglobin SS (33.3%), low in hemoglobin SC (9.0%), and non-existent in hemoglobin Sβ⁺. The overall event-rate of ASSC was 8.3 per hundred patient-years. The event-rate was 28.4 in the hemoglobin Sβ⁰, 10.9 in hemoglobin SS, and 3.5 in hemoglobin SC Patients with hemoglobin SS and hemoglobin Sβ⁰ on hydroxyurea therapy had a significantly higher occurrence of ASSC than those who were not, with event-rates of 14.2 and 3.1, respectively. The event-rate was also higher for children who started hydroxyurea before age 2 years than for those who started after this age (19.8 and 9.2 respectively).

Conclusions: The prevalence and severity of splenic problems vary widely between different sickle cell genotypes, with hemoglobin Sβ⁰ having the most severe complications. Hydroxyurea therapy is strongly associated with incidence of ASSC, particularly when initiated before two years of age.

INTRODUCTION

The spleen is among the earliest organs affected in people with sickle cell disease (SCD) ^1,2. Splenic dysfunction, as measured by the appearance of Howell-Jolly bodies and pitted red blood cells (RBCs) in circulation and decreased radiotracer update, can be detected as early as six months of age and can be seen in up to 86% of children by age one year³. A significant proportion of children with homozygous sickle cell disease (Hb SS) and sickle beta-zero thalassemia (Hb Sβ⁰) undergo complete infarction of the spleen early in life. This progressive loss of splenic function is associated with impaired ability to clear encapsulated bacteria from the blood and an increased risk of bacteremia. Loss of splenic function occurs later in life in people with hemoglobin SC disease (Hb SC) and may not occur at all in those with sickle beta-plus thalassemia disease (Hb Sβ⁺)^4,5.

In addition to impaired filtration and bacterial clearance, splenic dysfunction can manifest as splenomegaly, acute splenic sequestration crises (ASSC) characterized by acute trapping of blood in an enlarged spleen, and hypersplenism resulting in chronic cytopenias². ASSC in particular can be a life-threatening complication due to the risk of sudden hypovolemic shock in affected people. Previous studies have reported a prevalence of 7-30% for ASSC in children with SCD². A retrospective cohort study of ASSC in 190 patients with Hb SS and Hb Sβ⁰ revealed a prevalence of 12.6% and an event rate of 6 per 100 patient-years, with a median age of onset of 1.4 years⁶. A prospective cohort study of 153 children with Hb SS disease in Jamaica reported a prevalence of splenomegaly of 65% by age one year⁷. Splenomegaly has been reported in up to 34% of children with Hb SC disease, with ASSC and hypersplenism occurring in 12% and 10% respectively⁸. One study of children with Hb Sβ⁺ found no evidence of splenic dysfunction by age 18 years⁵. Another study of children from Tunisia, however, found a higher prevalence of ASSC in both Hb Sβ⁰ and Hb Sβ⁺ genotypes than in patients with Hb SS disease⁹. Of note, the pattern of splenic disease in sub-Saharan African countries appears to differ significantly from other regions. One meta-analysis of published studies describes a prevalence of splenomegaly of 10-73% but a prevalence of less than 10% for ASSC and 5% for hypersplenism¹⁰. The greater prevalence of splenomegaly in this region is hypothesized to be linked to malaria exposure, though the impact of this interaction on other aspects of splenic disease in SCD has not been described¹¹.

In this study, we describe a retrospective chart review of splenic complications, including splenomegaly, ASSC, and hypersplenism, in a large cohort of pediatric patients with SCD at a single children’s hospital. We identified the prevalence of these complications, as well as that of splenectomy, by age across different SCD genotypes. We also analyzed the event rate of splenic sequestration across genotypes. Finally, we examined the correlations between hydroxyurea therapy and the incidence and severity of splenic sequestration in patients with Hb SS and Hb Sβ⁰.

METHODS

This project was approved by the Institutional Review Board at Baylor College of Medicine prior to initiation of chart review. We searched our electronic medical records (EMR) system using ICD10 codes to identify all children with sickle cell disease (SCD) born between January 1, 2011, and December 31, 2018, who were seen at least once at our institution. The ICD10 codes used were D57.0 for all SCD diagnoses, D57.1 for Hb SS, D57.2 for Hb SC, D57.4 for Hb Sβ⁺ and Hb Sβ⁰, and D57.8 for other genotypes. The D57.3 code was used to exclude children with sickle trait. We also identified extended cohorts from January 1, 2000, to December 31, 2018, for the Hb SC and Hb Sβ⁺ genotypes at Texas Children’s Hospital.

Within our patient population, we used ICD10 codes to identify patients with splenic dysfunction (splenomegaly, ASSC, and hypersplenism), and splenectomy. The ICD10 codes used were: R16.1, D73.1, and D73.2 for splenomegaly; D73.89, D57.02, D57.212, D57.412, D57.432, D57.452, and D57.812 for splenic sequestration; D73.1 for hypersplenism; and Z90.81 for splenectomy. These diagnoses were verified by direct chart review using previously defined diagnostic criteria¹². Splenomegaly was defined as a spleen palpable by ≥2 cm below the costal margin; ASSC was defined as an acute decrease in hemoglobin of >20% from baseline with concomitant acute splenic enlargement and a rise in reticulocyte count; and hypersplenism was defined as a combination of splenomegaly with one or more cytopenias with no evidence of other etiologies. We also reviewed a random selection of charts that did not have an ICD10 code for splenic dysfunction to confirm the accuracy of diagnostic designations by ICD10 coding. For children with a confirmed diagnosis of splenic dysfunction, we identified the date of diagnosis or surgical procedure and number of ASSC events. Among patients with Hb SS and Hb Sβ⁰, we also identified whether hydroxyurea therapy had been initiated and the date of initiation. Patients who had inadequate information about diagnoses or dates of onset of complications were excluded from the study cohort.

Continuous variables are presented as medians or means with standard deviation and compared between groups using the paired or unpaired Student’s t test as appropriate. The event rate (ER) of ASSC within each cohort was calculated as events per one hundred patient-years and compared by generating incidence rate ratios (IRR) with 95% confidence intervals.

For survival analysis of each diagnosis, the age of diagnosis was determined using the date of diagnosis as documented in the patient’s chart. Patients who did not have a particular diagnosis were censored at the date of the last documented visit. The times to diagnoses were compared between sub-cohorts using the log-rank test. All statistical analyses were performed using GraphPad Prism software version 10.0.3 (GraphPad Software LLC) and MedCalc software version 22.014 (MedCalc Software Ltd).

RESULTS

We identified 688 patients with a diagnosis of SCD who were seen in any setting at our institution between January 1, 2011, and December 31, 2018. On further review of this group, we identified four misdiagnoses, nine patients with sickle trait, and sixty-eight duplicates who had ICD10 codes for both homozygous (SS) SCD and a variant (SC, Sβ⁺, Sβ⁰, or S-Other). An additional thirteen patients had insufficient information in their charts for the planned analysis and were excluded. After accounting for these anomalies, we derived a final cohort of 594 patients for analysis. Of these, 361 (60.7%) had Hb SS, 160 (26.9%) had Hb SC, 42 (7.1%) had Hb Sβ⁺, 16 (2.7%) had Hb Sβ⁰, and 15 (2.5%) had other SCD genotypes (Supplemental Figure 1).

The cumulative prevalence of splenomegaly, ASSC, hypersplenism, and splenectomy in our overall cohort by age 8 years was 24.7%, 24.2%, 9.6%, and 5.6%, respectively (Figure 1 and Supplemental Table 1). Stratification of splenic dysfunction by genotype revealed that all types of dysfunction and splenectomy were highest in patients with Hb Sβ⁰, intermediate in Hb SS, low in Hb SC, and non-existent in Hb Sβ⁺ and Hb S-Other, with statistically significantly differences between the SCD genotypes (Figure 2 and Supplemental Table 1).

Since ASSC is the most severe complication of splenic dysfunction and is often recurrent, we assessed its event rate in our patient cohorts. The overall ER of ASSC for the entire patient cohort was 8.3 per 100 patient years. The ER for patients with Hb Sβ⁰ was 28.4 compared to 10.9 for Hb SS and 3.5 for Hb SC. The IRR for patients with Hb Sβ⁰ and Hb SC relative to Hb SS were 2.6 and 0.3 respectively, indicating statistically significant differences in the incidence of ASSC between the groups (Table 1).

To determine if splenic dysfunction worsened with age in patients with Hb SC and Hb Sβ⁺ genotypes, we extended our analysis of patients with these genotypes to include those born between January 1, 2000, and December 31, 2010. We identified an additional 149 patients with Hb SC and 32 with Hb Sβ⁺ for whom detailed medical records were available, for totals of 309 and 74 patients respectively. The prevalence of splenic dysfunction remained very low in the Hb Sβ⁺ group, with only one patient developing ASSC and requiring splenectomy. In the Hb SC group, by contrast, the prevalence of splenomegaly, ASSC, and hypersplenism continued to increase with age, more than doubling by the end of this extended study interval (Figure 3 and Supplemental Table 1).

Finally, we assessed the association between hydroxyurea therapy and the incidence and prevalence of ASSC in patients with Hb SS and Hb Sβ⁰. Log-rank survival analysis of patients on hydroxyurea therapy compared to those who were not revealed a significantly higher prevalence of ASSC in the former group (Figure 4A). The ER of ASSC was 3.1 in patients who had no prescriptions for hydroxyurea and 14.2 in those prescribed hydroxyurea therapy, with an IRR of 4.6 (Table 2; p < 0.001). We also assessed whether the timing of hydroxyurea initiation correlated with prevalence and severity of ASSC. Among patients on hydroxyurea experiencing ASSC, the mean age of hydroxyurea initiation was 1.2 ± 0.4 years and the mean age of first ASSC was 1.6 ± 1.2 years; the median ages of HU initiation and ASSC onset were 1.13 years and 1.24 years respectively. Patients initiated on hydroxyurea before or at age 2 years had a significantly higher prevalence of ASSC than those initiated after age 2 years (Figure 4B). The ER was of ASSC in the early hydroxyurea group was 19.8 compared to 9.2 in the late hydroxyurea group, with an IRR of 2.1 (Table 3; p < 0.001).

DISCUSSION

In this study, we present a retrospective analysis of splenic dysfunction in a cohort of 594 pediatric patients with SCD who were cared for at a single children’s hospital between January 1, 2011, and December 31, 2018. As in previous reports, we identified a high prevalence of splenomegaly, ASSC, and hypersplenism in children with SCD. By age 2 years, the cumulative prevalence of splenomegaly and ASSC in the entire cohort were 13.5% and 11.2% respectively, plateauing at 24.7% and 24.2% respectively by age 6 years. The event rate for ASSC was 8.3 per 100 patient-years. Strikingly, there were significant differences in the prevalence of splenic dysfunction and the event rates of ASSC between genotypes. Children with Hb Sβ⁰ had the highest prevalence of all splenic complications and event rate for ASSC, followed by those with Hb SS and Hb SC. In the context of other studies that have reported a higher baseline hemoglobin level, a lower rate of acute chest syndrome, and similar cerebral hemodynamics in Hb Sβ⁰ patients compared to Hb SS patients, our results demonstrating a more severe course of splenic disease in patients with Hb Sβ⁰ suggest that Hb SS and Hb Sβ⁰are phenotypically distinct entities^13,14. Additionally, while children with Hb SS and Hb Sβ⁰appeared to have a plateau of disease prevalence by age 6 years, children with Hb SC disease continued to develop splenic disease well into the second decade of life. Conversely, children with Hb Sβ⁺ had virtually no evidence of splenic disease, even on extended analysis into late adolescence. Our data on Hb SC and Hb Sβ⁺ patients correspond with those from earlier studies, which indicated that Hb SC patients have progressive splenic disease throughout life while patients with Hb Sβ⁺mostly remain disease-free^4,5.

Hydroxyurea therapy has revolutionized care for people with SCD, dramatically reducing the rates of pain crises and acute chest syndrome and potentially reducing chronic organ damage¹⁵. The effectiveness of hydroxyurea therapy in preserving splenic function is less clear. A major study of hydroxyurea therapy in very young children examined its potential role in preserving splenic function but found no clear evidence of such benefit¹⁶. Children enrolled in this study had other significant benefits, including fewer pain crises, episodes of acute chest syndrome, RBC transfusion burden, and overall hospitalizations. Another study that examined splenic function in children escalated to their maximum tolerated dose of hydroxyurea did reveal some benefit in preserving splenic function¹⁷. Surprisingly, we noted a significantly higher prevalence and event rate of ASSC in children started on hydroxyurea before age 2 years compared to those who started therapy after age 2 years. The difference was even more striking when children started on hydroxyurea were compared to those who were never started on this medication. While the retrospective nature of our analysis and inherent differences between the sub-cohorts clearly limit conclusions about causality, our results indicate a strong correlation between the early initiation of hydroxyurea therapy and the burden of splenic dysfunction in SCD. In general, hydroxyurea therapy was initiated before the onset of ASSC among children who started hydroxyurea before age 2 years, supporting the possibility of a causal role for hydroxyurea (Table 4). Interestingly, there are other studies that suggest a link between hydroxyurea therapy and splenic disease. In a retrospective study of 28,580 patients in the Pediatrics Health Information System database, the prevalence of splenectomy was 7.2% among patients on hydroxyurea therapy compared to 3.2% in those not on therapy. The age of splenectomy was also lower in the treated group¹⁸. In a cohort of 23 patients with Hb SS and Hb Sβ⁰ treated with hydroxyurea, the onset of ASSC was noted at substantially older ages than has been previously reported; the authors hypothesize that preserved splenic function on hydroxyurea therapy may be responsible for this delayed onset of ASSC¹⁹. Splenomegaly and hypersplenism have been associated with elevated Hb F levels in Indian and Saudi Arabian patients with Hb SS disease^20,21,22. A possible mechanism for our association between hydroxyurea therapy and ASSC, therefore, is that hemoglobin F induction by early initiation of hydroxyurea therapy may counteract splenic infarction sufficiently to preserve some splenic activity but paradoxically results in disordered splenic filtration with progressive splenomegaly, ASSC, and hypersplenism. It is also possible, on the other hand, that the correlation we describe is due to increased hydroxyurea use in a subpopulation of patients with more severe disease phenotype, including splenic dysfunction.

There are some limitations to our study. The most prominent of these is the retrospective nature of our analysis, which of necessity introduces differences in baseline characteristics and treatment decisions between groups as well as in the duration of follow-up. In comparing our group of patients started on hydroxyurea by age 2 years and those started after age 2 years, for example, the former group is predominantly from the latter part of our study window while the latter is from the earlier part. Additionally, clinical practices may have changed during the period of our study, making comparisons over time prone to error. The small sample size of our Hb Sβ⁰ patient cohort also limits our analysis of this SCD subtype. Finally, while we reviewed our patient charts extensively to minimize the likelihood of inaccurate or missed diagnoses, it is possible that there are errors in our stratification of patients into groups by genotype, types of splenic dysfunction, and treatment status.

In summary, this study presents the largest retrospective cohort analysis of pediatric splenic disease excluding loss of splenic function. Our results indicate that splenomegaly and ASSC affect approximately 25% of all children with SCD , while chronic hypersplenism has a prevalence of about 10%. Additionally, children with Hb Sβ⁰ have a significantly higher prevalence of splenic dysfunction and incidence of ASSC than those with SS disease. Children with Hb SC disease have a later onset of splenic dysfunction but remain at risk of complications through the second decade of life, while those with Sβ⁺ disease have very low rates of splenic complications. Finally, we identify a strong correlation between the early initiation of hydroxyurea therapy and ASSC, though the causative nature of this association remains unproven. It will be important to prospectively evaluate the prevalence and severity of splenic disease in children treated with hydroxyurea escalated to maximum tolerated dose.

ACKNOWLEDGEMENTS

We acknowledge the invaluable assistance of Debasis Dash in extracting diagnostic data on our study cohort from our electronic medical record system. We also acknowledge the children living with sickle cell disease who receive care at Texas Children’s Hospital and their families for their continuing support of our research activities.

REFERENCES

1. Brousse V, Buffet P, Rees D. The spleen and sickle cell disease: the sick(led) spleen. Br J Haematol . 2014;166(2):165–176.

2. El Hoss S, Brousse V. Considering the spleen in sickle cell disease.Expert Rev Hematol . 2019;12(7):563–573.

3. Rogers ZR, Wang WC, Luo Z, et al. Biomarkers of splenic function in infants with sickle cell anemia: baseline data from the BABY HUG Trial.Blood . 2011;117(9):2614–2617.

4. Lane PA, O’Connell JL, Lear JL, et al. Functional asplenia in hemoglobin SC disease. Blood . 1995;85(8):2238–2244.

5. Barrios NJ, Kirkpatrick DV, Lohman D, et al. Spleen function in children with sickle β⁺ thalassemia. J Natl Med Assoc . 1991;83(9):819–822.

6. Brousse V, Elie C, Benkerrou M, et al. Acute splenic sequestration crisis in sickle cell disease: cohort study of 190 paediatric patients.Br J Haematol . 2012;156(5):643–648.

7. Rogers DW, Vaidya S, Serjeant GR. Early splenomegaly in homozygous sickle-cell disease: An indicator of susceptibility to infection.Lancet . 1978;2(8097):963–965.

8. Zimmerman SA, Ware RE. Palpable splenomegaly in children with haemoglobin SC disease: haematological and clinical manifestations.

9. Ben Khaled M, Ouederni M, Mankai Y, et al. Prevalence and predictive factors of splenic sequestration crisis among 423 pediatric patients with sickle cell disease in Tunisia. Blood Cells Mol Dis . 2020;80:102374.

10. Ladu AI, Aiyenigba AO, Adekile A, Bates I. The spectrum of splenic complications in patients with sickle cell disease in Africa: a systematic review. Br J Haematol . 2021;193(1):26–42.

11. Tubman VN, Makani J. Turf wars: exploring splenomegaly in sickle cell disease in malaria-endemic regions. Br J Haematol . 2017;177(6):938–946.

12. Ballas SK, Lieff S, Benjamin LJ, et al. Definitions of the phenotypic manifestations of sickle cell disease. Am J Hematol . 2010;85(1):6–13.

13. Day ME, Rodeghier M, DeBaun MR. Children with HbSβ⁰ thalassemia have higher hemoglobin levels and lower incidence rate of acute chest syndrome compared to children with HbSS. Pediatr Blood Cancer . 2018;65(11):e27352.

14. Ikwuanusi I, Jordan LC, Lee CA, et al. Cerebral hemodynamics and metabolism are similar in sickle cell disease patients with hemoglobin SS and Sβ⁰ thalassemia phenotypes. Am J Hematol . 2020;95(3):E66–E68.

15. Rankine-Mullings AE, Nevitt SJ. Hydroxyurea (hydroxycarbamide) for sickle cell disease. Cochrane Database Syst Rev . 2022;9(9):CD002202.

16. Wang WC, Ware RE, Miller ST, et al. Hydroxycarbamide in very young children with sickle-cell anaemia: a multicentre, randomised, controlled trial (BABY HUG). Lancet . 2011;377(9778):1663–1672.

17. Nottage KA, Ware RE, Winter B, et al. Predictors of splenic function preservation in children with sickle cell anemia treated with hydroxyurea. Eur J Haematol . 2014;93(5):377–383.

18. Menchaca AD, Style CC, Villella AD, et al. Pediatric Sickle Cell Disease Patients on Hydroxyurea Have Higher Rates of Surgical Splenectomy. J Surg Res . 2023;283:798–805.

19. de Montalembert M, Galacteros F, Oevermann L, Cannas G, Joseph L. Hydroxyurea Is Associated with Later Onset of Occurrence of Acute Splenic Sequestration Episodes in Sickle Cell Disease: Lessons from the European Sickle Cell Disease Cohort - Hydroxyurea (ESCORT-HU) Study.Blood . 2022;140(Supplement 1):449–450.

20. Kar BC, Satapathy RK, Kulozik AE, et al. Sickle cell disease in Orissa State, India. Lancet . 1986;2(8517):1198–1201.

21. Padmos MA, Roberts GT, Sackey K, et al. Two different forms of homozygous sickle cell disease occur in Saudi Arabia. Br J Haematol . 1991;79(1):93–98.

22. Al-Salem AH, Al-Aithan S, Bhamidipati P, Al-Jam’a A, Al Dabbous I. Sonographic assessment of spleen size in Saudi patients with sickle cell disease. Ann Saudi Med . 1998;18(3):217–220.

FIGURE LEGENDS

FIGURE 1: Prevalence of splenic complications by age: The cumulative prevalence of splenomegaly, acute splenic sequestration crises (ASSC), hypersplenism, and splenectomy in the entire cohort of 594 patients are presented with the number at risk by age. Numeric values for cumulative prevalence are presented in Supplemental Table 1.

FIGURE 2: Prevalence of splenic complications stratified by genotype. The cumulative prevalence of (A) splenomegaly, (B) acute splenic sequestration crises (ASSC), (C) hypersplenism, and (D) splenectomy are presented for the Hb SS, Hb SC, and Hb Sβ⁰ genotypes with the number at risk by age for each genotype. Numeric prevalence values for each panel are presented in Supplemental Table 1.

FIGURE 3 Prevalence of splenic complications in Hb SC disease: extended cohort analysis. The cumulative prevalence of splenomegaly, acute splenic sequestration crises (ASSC), hypersplenism, and splenectomy in the extended Hb SC cohort (0-18 years) of 309 patients, with the number at risk by age. Numeric values for cumulative prevalence are presented in Supplemental Table 1.

FIGURE 4 Prevalence of acute splenic sequestration crises stratified by hydroxyurea exposure. The cumulative prevalence of acute splenic sequestration crises (ASSC) for Hb SS and Hb Sβ⁰patients is presented with the number at risk by age. (A) Cumulative prevalence in patients treated with hydroxyurea (HU) compared to those not treated (No HU). (B) Cumulative prevalence in patients with hydroxyurea therapy initiated by age 2 years (HU ≤2) compared to those with therapy initiated after age 2 years (HU >2).