Prognostic significance of early acute splenic sequestration in children with severe sickle cell genotypes: A comprehensive longitudinal neonatal cohort study

Acute splenic sequestration crisis (ASSC) is one of the earliest acute clinical manifestations of sickle cell anemia (SCA), with a median age at first episode of 1.8 years [range: 0.4–12.9] as reported for our recently published regional longitudinal newborn cohort, beginning with the introduction of newborn screening (1986) and ending just before the introduction of preventive intensification with hydroxyurea (HU) in 2015.¹ Early predictive biomarkers have been identified for ASSC, but little is known about the impact of early ASSC on disease severity.² Unlike early dactylitis, early ASSC was not found to be associated with an increase in the risk of adverse outcomes, including death, stroke, frequent vaso-occlusive crisis (VOC), and recurrent acute chest syndrome (ACS), in a cohort of newborns with SCA.³

Our main objective here was to determine, from our SCA birth cohort, whether children experiencing early ASSC have a higher disease burden. In addition, we aimed to update clinical information on ASSC and confirm the prognostic factors identified in previous studies. Consistent with the French standards of care, for the whole cohort, disease-modifying therapies (DMT) were started only after the occurrence of complications: transfusion program (TP) was mainly implemented for stroke prevention, and HU was prescribed only to children over the age of 3 years for low hemoglobin (Hb) levels and/or recurrence of VOC/ACS. Specifically at our center, TP was offered for frequent VOC/ACS or anemia despite HU, or in children younger than 3 years, and hematopoietic stem cell transplantation (HSCT) to patients with cerebral vasculopathy or frequent VOC/ACS with a human leukocyte antigen-identical sibling. In our cohort-study, the use of DMT was thus considered a surrogate for disease severity.

ASSC was defined as splenic enlargement (increase of at least 2 cm from baseline) measured below the costal margin and associated with acute anemia (decrease in Hb concentration >2 g/dL relative to the previous measurement). Early and late ASSC were defined as a first episode of ASSC occurring before or after the age of 2 years respectively. During ASSC, standard management was prompt transfusion to restore effective circulating volume. After the resolution of a first ASSC, local guidelines recommended watchful waiting, unless children had another reason for receiving TP or HU. After the second or third episode, then either splenectomy or a temporary prophylactic TP were considered, to prevent ASSC recurrence. The age at which splenectomy was considered (usually after 3 years of age) and the indication for splenectomy after TP (only if persistent splenomegaly during TP or systematic) varied over time.

Children were classified into two groups on the basis of the timing of the first ASSC: before 2 years (early ASSC group), or after 2 years or no ASSC (other group). Descriptive statistics were used to summarize the data and groups were compared using Student's t-tests and Fisher's exact tests. Multivariate logistic regression analyses with stepwise selection identified significant factors predictive of early ASSC. Kaplan–Meier (KM) estimates and Hazard Ratios (HRs) were provided for SCA complications and initiation of DMT. Incidence rates were compared between ASSC groups using a Poisson regression.

Our severe sickle cell genotype cohort consisted of 292 subjects: 280 with HbSS, nine with HbSβ⁰, and three with HbS-Dpunjab (HbSD) genotypes. During the study period, 105 children, 56 (53%) boys and 49 girls experienced a first episode of ASSC, resulting in two- and 5-year probabilities of 21% [95% CI:16–25%] and 31% [95% CI:25–36%], respectively.¹ The first ASSC occurred before and after the age of 2 years in 61 and 44 children, respectively. Hundred children were not receiving any DMT when first ASSC occurred whereas four were taking HU, and one child had initiated TP. They all experienced first ASSC episode after 2 years of age.

During the first ASSC episode, mean Hb and platelet levels dropped to 6.0 ± 1 g/dL and 148 ± 77G/L, respectively, with no significant difference between the two age groups (6.0 ± 1 vs. 6 ± 0.1 g/dL; and 143 ± 67 vs. 155 ± 87G/L respectively, p = 0.91 and 0.53) suggesting similar severity in the two age groups. Approximately 104 children received a blood transfusion during their first ASSC. After the resolution of the first ASSC, watchful waiting was applied for 95 (90%) children, 43 (41%) of whom experienced no further episode. A temporary TP was used for four children. All had additional reasons for receiving blood transfusions. Only one child was started on HU because of recurrent VOCs. Five children underwent splenectomy.

Sixty-two of the 105 children (59%) experienced more than one episode: 40 of the 61 with a first ASSC before the age of 2 years and 22 of the 44 children with a first ASSC after the age of 2 years. There were 238 episodes in total, with a global incidence of 6.9/100 patient-years (PY). None of the episodes were fatal or accompanied by stroke. The median interval between the first and second episodes was 116 days [range 66–293 days] overall, 112 days [range: 63–288] and 131 days [range: 71–291] for children experiencing a first ASSC before and after the age of 2 years, respectively.

Overall, a temporary TP was used in 31/105 (30%) children, mostly in those with recurrent ASSC (28/62 (45%)). The median duration of the TP was 25 months [range: 16–35 months]. Sequestration recurred during the TP in 12 cases (39%), a median of 14.5 months after TP initiation [range: 7–27]. Splenectomy was performed in 39/105 (37%) patients overall, and in 15/31 (48%) children initially placed on a TP. Median age at splenectomy was 4.5 years [range: 3.6–6.8] overall, and 4.5 years [3.3–5.5] and 4.8 years [3.8–7.5] in the subgroups of children with and without prior TP, respectively.

We then focused on early ASSC, to identify prognostic markers of early ASSC occurrence among baseline blood parameters and before any DMT, and to determine whether early ASSC occurrence was predictive of adverse outcomes throughout childhood. We restricted analysis to the 287 children with more than 2 years' follow-up (FU). We thus compared the group of 60 children with early ASSC and a group of 227 children with first ASSC after 2 years (late ASSC) or no ASSC.

Mean FU was 13.7 ± 4.6 years in the early ASSC group and 13.4 ± 4.9 years in the other patients. In univariate analysis, early ASSC was significantly associated with being male, β-globin Bantou/Bantou haplotype, alpha-thalassemia (including one or two α-chain gene deletion), lower baseline Hb and HbF levels and higher baseline reticulocyte, leukocyte and neutrophil counts (Table S1). In the multivariable analysis, low HbF level (OR 0.88, 95% CI 0.83–0.94; p < .0001), high leukocyte count (OR 1.11, 95% CI 1.03–1.19; p < .007), and alpha-thalassemia (OR 2.11, 95% CI 1.02–4.38; p < .004) were independently associated with early ASSC.

We next investigated whether children with early ASSC experienced more severe disease by performing log-rank tests to compare KM estimates for sickle complications. KM curves showed that early ASSC was associated with a significantly higher cumulative risk of conditional transcranial Doppler (TCD) velocities, but did not increase the risk of abnormal TCD velocities, overt stroke, silent cerebral infarcts (SCI), or extracerebral cumulative organ damage, including a high tricuspid regurgitation velocity (TRV≥2 m/s), sickle nephropathy, and retinopathy (Table 1A). As expected, early ASSC was significantly associated with a higher cumulative risk of requiring transfusion, acute exacerbation of anemia (AEA), splenectomy, initiation of a TP, and HSCT. Interestingly, early ASSC was not associated with a higher cumulative risk of hydroxyurea use.

TABLE 1. Adverse outcomes according to age at the first acute splenic sequestration crisis in children with severe sickle cell genotypes.

1A
	Total (N = 287)	Early ASSC group (N = 60)	Late ASSC or no ASSC group (N = 227)	HR [95% CI]	p value
Neurological complications (n (%))
Conditional TCD velocities		30 (50)	82 (36.3)	1.52 [1.0–2.31]	.0491
Abnormal TCD velocities		21 (35)	60 (26.7)	1.34 [0.81–2.20]	.2526
Overt stroke		0	6 (2.6)	0	.2031
Silent cerebral infarcts		9 (15.3)	38 (18.4)	0.82 [0.4–1.7]	.5979
Extracerebral chronic organ complications
TRV ≥2.5 M/S		15 (25.4)	45 (21.1)	1.25 [0.7–2.24]	.4564
Nephropathy		3 (5.2)	21 (9.5)	0.51 [0.15–1.70]	.2628
Retinopathy		13 (21.7)	34 (15.9)	1.40 [0.74–2.65]	.3017
Osteonecrosis		1 (1.7)	15 (6.6)	0.24 [0.03–1.84]	.1364
Cholelithiasis		24 (40)	89 (39.2)	0.89 [0.57–1.40]	.6177
Cholecystectomy		20 (33.3)	83 (36.6)	0.77 [0.47–1.25]	.2901
Splenectomy		23 (38.3)	28 (12.3)	3.64 [2.09–6.32]	<.0001
Acute sickle cell disease-realated events
Vaso-occlusive crisis		54 (90)	186 (81.9)	1.46 [1.07–1.97]	.0147
Acute chest syndrome		44 (73.3)	157 (69.2)	1.18 [0.84–1.65]	.3310
Acute exacerbation of anemia		52 (86.7)	106 (46.7)	3.90 [2.78–5.47]	<.0001
First transfusion		60 (100)	210 (92.9)	6.10 [4.33–8.60]	<.0001
Initiation of disease-modifying therapy
Transfusion program		49 (81.7)	130 (57.3)	2.36 [1.56–2.83]	<.0001
Hydroxyurea		42 (70)	152 (67)	1.12 [0.79–1.57]	.5208
Hematopoîetic stem cell transplantation		17 (28.3)	39 (17.2)	1.9 [1.07–3.37]	.0251

1B
	Early ASSC group (N = 60)	Late ASSC or no ASSC group (N = 227)	RR [95% CI]	p value
Total VOC episodes/100PY	74	61	1.21 [1.1–1.34]	.0002
Total ACS episodes/100PY	20	15	1.29 [1.05–1.56]	.0140
Total acute pneumonia EPISODES /100PY	8	8	0.96 [0.71–1.30]	.8045
Total ACS /100PY (excluding pNEUmonia episodes)	12	7	1.63 [1.26–2.12]	.0002
Total acute exacerbations of anemia /100PY	25	7	3.38 [2.75–4.16]	<.0001
Total transfusion episodes /100PY	446	265	1.68 [1.61–1.76]	<.0001
Total occasional transfusion episodes /100PY	78	48	1.64 [1.48–1.81]	<.0001

• Note: High tricuspid regurgitation velocity (TRV) (≥2.5 m/s); sickle nephropathy (microalbuminuria defined as a urine albumin-to-creatinine ratio ≥30 mg/g); sickle retinopathy (proliferative or non-proliferative); avascular bone necrosis was recorded only if symptomatic and accompanied by necrosis on X-ray or MRI; VOC, vaso-occlusive crisis requiring hospitalization; We differentiated between ACS episodes developing during the clinical course of a VOC and acute pneumonia (restricted to episodes combining fever, cough, and/or wheezing at admission and isolated lobe involvement, with no concomitant VOC). Total transfusion episodes include transfusions as part of a transfusion program. Total occasional transfusion episodes exclude transfusions delivered as part of a transfusion program. Analysis was restricted to children with more than 2 years of follow-up. Children were classified into two groups on the basis of the timing of the first ASSC episode: (1) before 2 years (early ASSC group n = 60) or (2) no ASSC or first ASSC episode after the age of 2 years (n = 227). Kaplan–Meier survival estimates were calculated and compared in log-rank tests between the two groups. Hazard ratios (HRs) are given for each variable, together with the associated 95% confidence intervals (95% CI). Incidence rate was calculated as the total number of events divided by total patient-years (PY) at risk. Data were censored at the last clinical visit, or at the time of hematopoietic stem cell transplantation. We used a Poisson regression model with log follow-up as the offset and acute splenic sequestration crisis (ASSC) group as the main factor to compare the two groups. Risk ratios (RRs) are given for each variable, together with the associated 95% confidence intervals (95% CI). Significant differences between the two groups (p < .05) are indicated in bold. 1A: Cumulative risks of neurological complications, extracerebral chronic organ damage, acute sickle cell disease-related events, initiation of disease-modifying therapy, according to age at first acute splenic sequestration crisis in children with severe sickle cell disease (HbSS/Sβ⁰/SD genotype group). 1B: Incidence rate throughout childhood, according to age at first acute splenic sequestration crisis in children with severe sickle cell disease (HbSS/Sβ⁰/SD genotype group).
• Abbreviations: ASSC, acute splenic sequestration crisis; ACS, acute chest syndrome; TCD, transcranial doppler.

We then compared incidence rates to determine whether children with early ASSC had a higher vaso-occlusive burden. Early ASSC predicted a modest but significant increase in the numbers of both VOC (74.4 vs. 61.4/100 PY) and ACS episodes (19.6 vs. 15.3/100 PY) during FU. Interestingly, ACS rates increased only for ACS developing during the clinical course of a VOC (11.9 vs. 7.3/100 PY), not for ACS in the form of acute pneumonia episodes (7.8 vs. 8.1/100 PY) (Table 1B).

In conclusion, this single-center cohort study of 238 episodes of ASSC in 105 children with SCA, with no concomitant strokes or deaths, confirms the benefits of a well-developed healthcare system providing newborn screening, parental education, and early access to comprehensive sickle cell disease referral centers. As previously reported, high baseline HbF levels provided the strongest protection against early ASSC in multivariate analysis (p < .0001). We also show, for the first time, that alpha-thalassemia is independently predictive of early ASSC.

This is the first study to evaluate the prognostic significance of early ASSC for disease severity throughout childhood. We found that children with early ASSC did not have higher rates of neurological complications other than conditional cerebral velocities. This finding was unexpected because early ASSC was significantly associated with a higher cumulative risk of AEA in our cohort, and AEA was identified as a significant independent risk factor for SCI in other cohorts.⁴ These findings may reflect the protective effect of alpha-thalassemia against cerebral vasculopathy,⁵ as the prevalence of alpha-thalassemia was significantly higher in the group of children with early ASSC. Alternatively, our data may attest to the protective effect of the DMT widely used in our SCA genotype group, particularly for HSCT (probability of 83.1% [95% CI: 78.3–87.4%] overall and of 16.7% [95% CI: 12.4–21.5%] for HSCT by the age of 10 years).¹ Interestingly, no abnormal TCD velocities or SCI developed after transplantation in any of the 56 children undergoing HSCT median age at HSCT: 4.8 years (range: [2.6–17.3] for the 17 children with early ASSC, and 7.3 years [3.2–19.7] for the other 39 children).

Interestingly, early ASSC increased the odds of more frequent VOCs and more frequent ACSs throughout childhood. Our data argue for a broader use of HU, even in children with ASSC complications, as HU did not influence ASSC occurrence rates in the Baby-hug randomized trial,⁶ because many of these children experience recurrent vaso-occlusive complications. Some of the strongest predictors of splenic function preservation in children treated with HU included younger age at initiation of HU therapy, shorter time to reach the maximum tolerated dose (MTD), and a larger increase in HbF levels relative to baseline.⁷ Additional studies in real-life conditions are warranted because HU may delay the development of functional asplenia, and it is important to investigate the effects of introducing HU earlier and of the dose escalation to the MTD, on ASSC risk over time.