166 - Neonatal critical illness is associated with the development of pediatric aplastic anemia

Publication Number: 2160.166

Brian M. Dulmovits, Childrens Hospital of Philadelphia, Philadelphia, PA, United States; Nicolas P. Goldstein Novick, Children's Hospital of Philadelphia, Rose Valley, PA, United States; Christopher Thom, CHOP/UPenn, Philadelphia, PA, United States

Background: Perinatal illness and preterm birth can portend lifelong neurologic and respiratory complications, but long-term hematologic abnormalities are not well understood. We hypothesized that adverse perinatal exposures affect bone marrow development and function, resulting in hematologic morbidities. The impacts of perinatal illness and prematurity on hematologic disease risk are understudied.

Objective: We aimed to define associations between preterm birth or critical illness with the development of childhood leukemia, bone marrow failure (BMF), pancytopenia, myelodysplasia, or aplastic anemia (AA).

Design/Methods: We performed a case-control study of a single care network consisting of neonates admitted to a level IV NICU or follow-up clinic with a post-NICU diagnosis of either leukemia, BMF, pancytopenia, myelodysplasia, or AA from 2008-2025. We excluded patients with Trisomy 21. We extracted demographics and indicators of critical illness (Table 1) from electronic medical records to create control (term non-critically ill, gestational age (GA) >37 weeks), critically ill term, and preterm (GA < 37 weeks) groups. We used logistic regression with year fixed effects and multivariate analysis to determine associations of prematurity, critical illness, and individual components of our critical illness definition with childhood hematologic pathology.

Results: We analyzed 13,083 neonates divided into control (mean GA +/- SD: 38.6+/-1.2 weeks), term critically ill (38.9 +/- 1.2 weeks), and preterm (31.2+/-4.3 weeks) groups. The percentage of preterm infants with any indicator of critical illness ranged from 2.1-61.4% (Table 1). Logistic regression revealed that prematurity (OR 2.4, 95%CI 1.3-4.1) and critical illness in term neonates (OR 2.5, 95%CI 1.4-4.3) were associated with hematologic pathology diagnosis. Subgroup analyses showed AA was the primary diagnosis driving this association (Table 2). A diagnosis of sepsis or shock was significantly associated with hematologic pathology based on multivariate logistic regressions stratified on prematurity (Figure 1).

Conclusion(s): We demonstrate a novel association between prematurity and neonatal critical illness with a later diagnosis of pediatric AA. A diagnosis of sepsis or shock drives this association in both preterm and term infants, suggesting that infectious and inflammatory exposures (as opposed to prematurity itself) may drive later hematopoietic dysfunction in premature and/or critically ill infants. These findings compel larger studies to determine if long-term monitoring programs for hematologic dysfunction are warranted for these neonates.

Table 1. Demographics, length of stay, and indicators of critical and non-critical illness


Table 2. Logistic regression predicting hematologic outcomes
Table 2.jpeg

Figure 1. Forest plot of odds ratios with components of critical illness predicting any hematologic pathology, stratified on prematurity.
Figure 1.jpegModel = logistic regression with year fixed effects, adjusted for displayed factors and additionally any non-invasive ventilation, extracorporeal membrane oxygenation, bronchopulmonary dysplasia, and hypoxic ischemic encephalopathy