Regulation and dysregulation of immunity in growing premature infants remain poorly understood. Premature infants have immature immune systems, including diminished neutrophil numbers and function, T cell lymphopenia, NK cell dysfunction, and relative hypogammaglobulinemia (1–4). Many of the organ-specific disorders seen in prematurity, such as necrotizing enterocolitis (NEC) and bronchopulmonary dysplasia (BPD), are associated with immune dysregulation including altered cytokine production and skewing of T cell populations (5, 6). Clinical manifestations of such conditions are non-specific and may be indistinguishable from organ-specific manifestations of inborn errors of immunity (IEI), encompassing an array of genetic immunodeficiencies and hyper-inflammatory disorders (7).

Consistent with dynamic, rapid changes in composition and function of the immune system in newborns (8, 9), multiple reports have outlined potential pitfalls of interpreting standard state newborn screening for severe combined immunodeficiency (SCID) in those born preterm (10, 11). No standards of care have been established for the clinical immunologist or neonatologist to guide comprehensive workup for IEI in the extremely premature population. Enhanced awareness of the potential to encounter IEI in the neonatal intensive care unit (NICU), as well as formal collaborations between neonatologists and clinical experts in identifying and treating neonatal IEI, will benefit this often-overlooked population. The goal of this review is to educate the clinical immunologist on common clinical features of prematurity which may suggest IEI in certain circumstances. For a review on features of IEI that can present in the neonatal and premature period, please see our previous review (12). Prompt recognition of these disorders in critically ill, preterm neonates will not only enhance our basic understanding of neonatal immunity but will also accelerate development of novel management strategies and therapies for affected infants. IEIs individually are rare, but collectively are not. Reports of IEI in premature infants are uncommon, suggesting that IEI may be underdiagnosed in premature infants and highlighting the need for better detection.

The following cases illustrate challenges in diagnosis of immunodeficiency and genetic inflammatory disorders in premature infants. They are not meant to suggest that each neonate presented has an IEI, but to illustrate the clinical complexity that can occur when trying to determine whether a premature infant has abnormal immunity.

An extremely premature male infant born at 24 weeks gestation developed severe NEC with poor wound healing and required a wound vacuum for several months in order for the abdomen to fully heal. The patient also had recurrent Candida parapsilosis infection with bloodstream and urinary tract infection, as well as severe thrush. Initial immune testing at 30 weeks gestation showed severe T cell lymphopenia but was attributed to the patient’s immaturity and clinical status at that time. The initial SCID screen was reported as normal at 7 days of age by the referring hospital but was persistently abnormal (<252 copies/uL) at our hospital from 1 month on. Repeat immunologic testing at 41 weeks postmenstrual age (neonatal nomenclature for the corrected gestational age accounting for age at birth plus weeks since birth) demonstrated T cell lymphopenia and abnormal T cell proliferation, but the patient became abruptly ill the following week and died from Klebsiella sepsis. Whole exome testing and protein analysis revealed heterozygous null mutations in OSBPL5 in trans, a gene not before implicated in human disease that has been shown to be important for CD4 T cell proliferation in humans (13). Phenotypic and mechanistic evaluation is ongoing in the laboratory. In this situation, the patient’s prematurity and severity of illness clouded the ability to evaluate for IEI. Earlier use of rapid exome testing may have been clinically beneficial.

An infant born at 25 weeks’ gestation developed expected complications of prematurity, including chronic lung disease and surgical NEC. At 4 months of age, she abruptly decompensated with cardiorespiratory failure and required extracorporeal membrane oxygenation (ECMO). She was found to have an abdominal abscess and Pseudomonas sepsis. Standard immune testing showed mild T cell lymphopenia but was otherwise unremarkable. CH50 and AH50 levels were normal. Genetic testing, however, showed a mutation in MBL2 (mannose binding lectin) that was predicted to be pathogenic. MBL2 defects have been shown to increase susceptibility to infection in infants (14), and this genetic disorder may particularly enhance infectious susceptibility in the time period before the baby begins to produce its own IgG in significant amounts. Premature infants with MBL2 polymorphisms have an increased risk for adverse neurological outcomes (odds ratio of 8.67) compared to infants without variants (15). The patient recovered from the initial infection, but she later developed recurrent Pseudomonas peritonitis several months later. This case illustrates how apparent complications of BPD or NEC can be difficult to discern from abnormal infections that signal genetic variants in complement and humoral immunity.

An infant born at 32 weeks gestation developed early-onset Escherichia coli meningitis and sepsis on day of life 1. Despite rapid recognition and treatment with appropriate antibiotics, the patient developed multiple intracranial empyemas that were not amenable to direct drainage. The patient completed seven weeks of antibiotics but exhibited severe loss of brain parenchymal tissue with development of pervasive seizure disorder. The patient again had recurrence of E. coli sepsis and meningitis at 2 and a half months of age, despite negative cultures on multiple occasions in the interval between the first and second infections. Immune testing was normal including a T/B/NK cells and T cell memory panel except for decreased absolute NK cell number (94 cells/uL, ref range 160-1100), and abnormal T cell proliferation to anti-CD3 stimulation (3723 CPM, ref >62,927; background was normal). Trio whole exome sequencing was sent and revealed no convincing candidate genes. E. coli is a well characterized cause of neonatal meningitis, and certain virulence factors are associated with increased risk for neonatal meningitis, such as the presence of the K1 capsule. Gram negative meningitis in premature infants is often devastating neurologically and can cause significant brain volume loss. Severe outcomes have been tied to virulence factors such as the E. Coli K antigen.

A 33-week premature infant was born in the setting of preterm labor. The infant had mild respiratory distress but also had multilineage cytopenias, significant liver dysfunction with coagulopathy, conjugated hyperbilirubinemia, transaminitis, and a ferritin level that peaked over 30,000. Soluble IL-2R level was normal. Her liver MRI was inconsistent with neonatal hemochromatosis (also referred to as gestational autoimmune liver disease/GALD”). Mean fluorescence staining for perforin in NK cells was low. The team was unable to obtain a suitable bone marrow biopsy specimen. Given continued disease progression and suspicion of primary hemophagocytic lymphohistiocytosis (HLH) based on low perforin expression, she was treated with IVIG (with little impact) and then transitioned to steroids. Her ferritin began to decrease with initiation of steroids. Unable to wean steroids without relapse, she eventually was transitioned to etoposide and cyclosporine. She was discharged from the NICU around 2 months of age. Specific gene-based testing for common genetic causes of HLH and whole exome sequencing was non-diagnostic. She went on to have a bone marrow transplant within the first year of life. This case illustrates the clinical overlap between hemophagocytic autoinflammatory disorders and neonatal disorders caused by maternal immune activation.

All of these patients were cared for in a single tertiary NICU within the past three years. Several of these infants did have IEI or clinically important polymorphisms in immune genes. If our unit is similar to other regional referral NICUs, it suggests that IEI may be underappreciated in the premature population. The immune interface is critically important during gestation, and perturbation of the maternal-fetal immune milieu by IEI might actually lead these individuals to be overrepresented in the NICU compared to term infants – it is simply not known.

These cases reinforce how difficult it can be to discern prematurity-related immune immaturity from genetic causes of primary immunodeficiency. Because something is challenging, however, does not mean it is impossible. In the age of cost-effective and rapid genetic testing, especially whole exome sequencing, this presumption of diagnostic helplessness must be reconsidered. Early detection of IEI saves lives, and in some cases, routine enhanced precautions and antibiotics may allow for a bridge to more definitive treatment once the patients are mature and stable enough for bone marrow transplantation or gene therapy. In addition, many of these families will go on to attempt to have additional children, in which case identifying genetic risk can help with family counseling and preimplantation diagnostic efforts.

Among those who contract infections in the NICU, premature infants cannot be distinguished from infants with IEI on the basis of clinical presentation or microbiology. The preterm neonate may exhibit a developmentally appropriate, but still abnormal, response to serious bacterial or fungal infection. The data discussed above support the notion that, in certain circumstances, clinical findings that are common in the premature neonate should prompt the clinician to consider screening for IEI. Given the substantial overlap in clinical symptomatology between premature infants and those with IEI, we advocate a conservative approach. Many other reviews have covered appropriate testing pathways for a range of IEIs (82–84). We will summarize briefly here but encourage clinicians to follow standard diagnostic approaches for symptoms that might suggest an IEI. For symptoms that overlap with autoinflammatory disorders, testing should include the standard acute phase reactant panels and cytokines including IL-18 levels and soluble IL-2R levels (53). Autoinflammatory gene panels can be sent if the first testing step is abnormal or inconclusive, and in fact given the clinical overlap of these disorders is important to send early (53, 85). For innate immune disorders the testing ordered should reflect the clinical scenario, for instance for a patient with a an intraperitoneal fungal ball would raise concern for chronic granulomatous disease and a dihydrorhodamine (DHR) test should be done and followed with CGD genetic testing if abnormal.

When there is concern for SCID or a combined immunodeficiency, which tends to be the most common scenario in premature infants, the first diagnostic steps should still be newborn screening with T cell receptor excision circles (TRECs) followed by flow cytometry and mitogen testing ( Table 1 ). Furthermore, given the difficulty in interpreting flow in this population, where T cell lymphopenia is prominent, clinicians should adopt a more aggressive approach when evaluating for IEI if the results are equivocal. While a suggested reference range for flow cytometry in premature infants has been published (86), most commonly the challenge is interpreting whether abnormal results indicate an IEI or are explained by clinical circumstances, as these premature babies are often very ill and frequently on steroid therapy. If first line screening is abnormal or equivocal and substantial clinical evidence of an IEI is present (very severe infection, multiple infections, unusual organism, or unusual healing), we advocate that genetic testing be undertaken. A “wait and see” approach in this population could have devastating, sometimes fatal, consequences. Historically, there has been hesitancy to send extended testing in premature infants due to their small relative blood volume and the amount of blood needed for testing, but presently most children’s hospitals can do flow cytometry reliably on 1cc of blood or less. Genetic testing can often be done on oral saliva swabs, making blood draws even less of a concern.

Newborn screening for SCID has been of great benefit in the neonatal population. In premature infants, however, the false positive rate is high. In some states, such as Massachusetts, a conservative T cell receptor excision circle (TREC) threshold leads to an even higher rate of false positives in premature infants. As many more infants are referred for consultation and flow cytometry (87, 88), this leads to “TREC fatigue”, where low TRECs in premature infants are typically assumed to be due to immaturity or to critical illness, and the recommendation is to repeat testing every two weeks until either TRECs are normal, or the patients is more mature. In the original report of the first two years of California’s newborn screening for SCID, of the premature infants who tested positive initially, only 11% remained with abnormal TRECs by 3-4 weeks of life (89). If a state algorithm is to repeat NBS in premature infants, this should be time limited and not continue indefinitely. In the end, it is a disservice to premature infants to be continuously administering a screening test that will not be promptly acted upon. However, no clear alternatives to the TREC assay exist for premature infants. In fact, the scope of this problem is recognized in the community and the Association of Public Health Laboratories has been enrolling newborn screening programs in a quality improvement initiative to improve TREC standard ranges for premature infants (https://www.aphl.org/rfp/Pages/NBS-SCID.aspx). While it would neither be cost effective nor clinically effective to perform genetic testing on every premature infant with low TRECs, we recommend that patients who have had a history of abnormal TRECs and any serious or unusual infection have flow cytometry testing. If flow cytometry is inconclusive, even in the setting of illness or steroid use, the baby should be considered for whole exome sequencing (WES). A large number of tertiary children’s hospitals offer rapid WES for scenarios where testing will impact clinical management, with results available in as little as 7 days (90–92). Immunologists have already begun to advocate for more liberal use of genetic testing, even to the point of including genetic testing for IEI in newborn screening panels (93).

In sum, IEI and disorders of prematurity can present with similar features and are not mutually exclusive. Existing screening tests, especially for T cell-based disorders, have low specificity in the premature population. When a patient has a concerning clinical history, definitive genetic testing should be undertaken if immune testing is abnormal, unclear or equivocal, rather than waiting for the patient to be more mature. Clinicians need to have a high index of suspicion for IEI in premature infants, as we do not know the true incidence of IEI in this population. Finally, new diagnostic strategies or better data about normal test values in this population could be especially helpful for diagnosing IEI in the premature population.

Both authors contributed equally to the manuscript, including development of the proposal, research and writing. Both AO and SG edited the final version and approve of submission of the final version.

This work was supported by the National Institutes of Health K08AI151265 (SG) and K08DK120871 (AO); Children’s Hospital of Philadelphia Research Institute (SG); the Manton Center for Orphan Disease Research (AO) and Children’s Hospital Boston Division of Newborn Medicine (AO).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.