The impact of in utero transfusions on perinatal outcomes in patients with alpha thalassemia major: the UCSF registry

Abstract

Alpha thalassemia major (ATM) is a hemoglobinopathy that usually results in perinatal demise if in utero transfusions (IUTs) are not performed. We established an international registry (NCT04872179) to evaluate the impact of IUTs on survival to discharge (primary outcome) as well as perinatal and neurodevelopmental secondary outcomes. Forty-nine patients were diagnosed prenatally, 11 were diagnosed postnatally, and all 11 spontaneous survivor genotypes had preserved embryonic zeta-globin levels. We compared 3 groups of patients; group 1, prenatally diagnosed and alive at hospital discharge (n = 14), group 2, prenatally diagnosed and deceased perinatally (n = 5), and group 3, postnatally diagnosed and alive at hospital discharge (n = 11). Group 1 had better outcomes than groups 2 and 3 in terms of the resolution of hydrops, delivery closer to term, shorter hospitalizations, and more frequent average or greater neurodevelopmental outcomes. Earlier IUT initiation was correlated with higher neurodevelopmental (Vineland-3) scores (r = -0.72, P = .02). Preterm delivery after IUT was seen in 3/16 (19%) patients who continued their pregnancy. When we combined our data with those from 2 published series, patients who received ≥2 IUTs had better outcomes than those with 0 to 1 IUT, including resolution of hydrops, delivery at ≥34 weeks gestation, and 5-minute appearance, pulse, grimace, activity, and respiration scores ≥7. Neurodevelopmental assessments were normal in 17/18 of the ≥2 IUT vs 5/13 of the 0 to 1 IUT group (OR 2.74; P = .01). Thus, fetal transfusions enable the survival of patients with ATM and normal neurodevelopment, even in those patients presenting with hydrops. Nondirective prenatal counseling for expectant parents should include the option of IUTs.

Graphical abstract

Figure 1.

Patient characteristics. (A) Flowchart depicting all patients with ATM in the UCSF registry and the 3 groups included in the statistical analyses (group 1: prenatal diagnosis, alive (n = 14); group 2: prenatal diagnosis, deceased (n = 5), and group 3, postnatal diagnosis, alive (n = 11). Group 1 included one patient with a nondeletion variant (SEA/Hb H Adana), and group 3 included 3 patients with alpha thalassemia nondeletion variants (homozygous Zurich-Albisrieden [n = 1], SEA/Hb H Adana [n = 2]). ∗Previously published patients: 4 patients in group 1^,and 6 patients in group 3.^, (B) Genotypes of patients prenatally or postnatally diagnosed. The prenatally diagnosed group included 49 patients with the following genotypes: homozygous SEA (n = 29), SEA/FIL (n = 16), SEA/THAI (n = 1), homozygous MED I (n = 1), homozygous BRIT (n = 1), and SEA/Hb Adana (n = 1). The postnatally diagnosed group included 11 patients with the following genotypes: SEA/SEA (n = 8), SEA/Hb Adana (n = 2), and homozygous Zurich-Albisrieden (n = 1). The black outline in the prenatal diagnosis group indicates genotypes that have deletions spanning the zeta-globin gene (THAI, FIL), which were not observed in the postnatal diagnosis group (P = .02, Fisher’s exact test). (C) Cardiothoracic ratio observed by ultrasonography at the indicated GAs. Each data point is a unique fetus, the lines indicate the fifth, fiftieth, and ninety-fifth percentiles. The triangle represents a patient with a nondeletion variant.

Figure 2.

The effect of IUTs on perinatal outcomes. We compared outcomes in fetuses who were diagnosed prenatally and survived (group 1, blue fill) with those diagnosed prenatally who had a perinatal demise (group 2, green fill) and those who were postnatally diagnosed and alive (group 3, red fill). (A) Percentage of fetuses in whom hydrops had resolved at the time of birth (group 1: n = 12, group 2: n = 3, group 3: n = 4 with data available). P = .005 across groups, P = .01 between groups 1 and 3. (B) GA at delivery in weeks (group 1: n = 14, group 2: n = 4, and group 3: n = 11 with data available). P = .004 across groups, P = .003 between groups 1 and 3, P = .022 between groups 1 and 2. (C) APGAR scores at 5 minutes (group 1: n = 14, group 2: n = 2, group 3: n = 6 with data available). P = .03 across groups, P = .03 between groups 1 and 3. (D) Percentage of neonates who required invasive mechanical ventilation after birth (group 1: n = 14, group 2: n = 2, group 3, n = 8 with available data) P = .03 across groups, P = .052 between groups 1 and 3). (E) Length of neonatal hospitalization (group 1: n = 14, group 2: n = 1, group 3, n = 9 with data available) P = .04 across groups, P = .049 between groups 1 and 3. (F) Percentage with average or greater neurodevelopment (group 1: n = 11, group 3: n = 11 with data available). P = .04 between groups 1 and 3. In panels B, C, and E, each symbol represents 1 patient; triangles represent patients with nondeletion variants (Zurich-Albisrieden, Hb H Adana); and the horizontal line indicates the median. For comparisons between groups 1 and 3, ∗P ≤ .05, ∗∗P ≤ .01; ∗∗∗P ≤ .005. † Represents a patient who died after a 175-day hospital stay. (Kruskal-Wallis one-way analysis of variance test with pairwise comparisons was used to compare the continuous variables across the 3 groups, and Fisher’s exact test was used for ordinal variables).

Figure 3.

Vineland-3 percentile score and correlation with GA at the first IUT. (A) Composite Vineland percentile score vs GA at first IUT (Pearson r = −0.72, P = .02). (B) Communication subdomain Vineland percentile (Pearson r = −0.82, P = .004). In each graph, the filled circles represent patients in group 1, and the triangle represents a group 1 patient with a nondeletion variant (Hb H Adana). The empty circle represents a patient who was born spontaneously at 28 weeks and diagnosed after birth (this patient was not included in the statistical analysis to evaluate the impact of IUTs on neurodevelopment). The superior dotted line corresponds to 1 standard deviation below the mean (sixteenth percentile); which signifies the risks of developmental delay. The inferior dotted line corresponds to 2 standard deviations below the mean (2.5^th percentile); which signifies the risks for intellectual disability and is not seen in any patient in the transfused cohort.

Figure 4.

Combined analysis of perinatal outcomes in our case series and those reported in 2 previously published case series (Chan et al 2018, Zhang et al 2021). Forest plots with 95% confidence intervals comparing patients who received optimal IUTs (OT, ≥2 IUTs) and those with suboptimal IUTs (SOT, 0-1 IUT). Denominators for each analysis reflect the number of patients for whom the data were available. (A) At the time of birth, hydrops resolved in 17/18 (94.4%) OT and 3/9 (33.3%) SOT fetuses (OR 3.25 [0.90, 5.60]; P = .01). (B) The number of births at ≥34 weeks was 21/27 (77.8%) in OT vs 3/19 (15.8%) in SOT patients (OR 2.92 [1.36, 4.47]; P = .003). (C) OT patients were significantly more likely than SOT patients to have reassuring APGAR scores (≥7) at 5 minutes (OT: 16/18 (88.2%) vs SOT: 5/12 (41.6%) (OR 2.44 [0.59, 4.30]; P = .01). (D) OT patients were significantly more likely than SOT patients to have average or greater neurodevelopment (OT: 17/18 (94.4%) vs SOT 5/13 (38.5%) (OR 2.74 [0.76, 4.71]; P = .01).