Shorter birth length and decreased T-cell production and function predict severe infections in children with non-severe combined immunodeficiency cartilage-hair hypoplasia

Abstract

Background: Cartilage-hair hypoplasia (CHH) is a syndromic inborn error of immunity caused by variants in the RMRP gene. Disease manifestations vary, and their ability to predict outcome is uncertain. The optimal management of infants with CHH who do not fulfill classical severe combined immunodeficiency (SCID) criteria is unknown.

Objective: We described longitudinal changes in lymphocyte counts during childhood and explored correlations of early childhood clinical and laboratory features with clinical outcomes on long-term follow-up of CHH patients.

Methods: Immunologic laboratory parameters, birth length, the presence of Hirschsprung disease, and severe anemia correlated to the primary end points of respiratory and severe infections. We implemented traditional statistical methods and machine learning techniques.

Results: Thirty-two children with CHH were followed up for 2.7 to 22.1 years (median, 8.2 years, in total 331.3 patient-years). None of the patients had classical SCID. Median lymphocyte subclass counts, apart from CD16⁺/56⁺ cells, were subnormal throughout childhood, but did not show age-related decline seen in healthy children. Low immunoglobulin levels were uncommon and often transient. Respiratory and/or severe infections developed in 14 children, 8 of whom had low naive T-cell counts, absent T-cell receptor excision circles, and/or partial "leaky" SCID-level lymphopenia. Shorter birth length correlated with lower lymphocyte counts and the occurrence of infections. Of the laboratory parameters, decreased naive T-cell counts and abnormal lymphocyte proliferation responses contributed most to the development of severe infections. In addition, all participants with absent T-cell receptor excision circles developed severe infections. Opportunistic infections occurred only in children with leaky SCID-level lymphopenia.

Conclusions: Shorter birth length and a combination of laboratory abnormalities can predict the development of severe infections in children with CHH.

Keywords: Immunodeficiency; RMRP; TREC; inborn error of immunity; lymphopenia; machine learning.

Fig 1

Flowchart demonstrating inclusion of patients with cartilage hypoplasia to study, as well as 2 cohorts used for machine learning approach.

Fig 2

Clinical course of 32 children with CHH. Patients are arranged according to severity of T-cell abnormalities. Within each laboratory subgroup, patients are arranged in order of decreasing CD3⁺ median counts. Low CD3⁺ T cells, low naive CD4⁺ T cells, and Leaky-SCID were all defined according to PIDTC 2022 criteria. TRECs are shown as copies/μL. Autoimmune disease in a single patient was hemolytic anemia. N/A, Data not available.

Fig 3

Lymphocyte subclass counts in children with CHH (dotted lines) and healthy children (gray-shaded area) are depicted as medians (bold lines) with 10th to 90th percentile range.

Fig 4

Random forest classification demonstrating relative importance of clinical and laboratory features in development of severe (A and C) and respiratory (B and D) infections in all 32 (A and B) study participants (naive cell counts not included) and in 24 study participants (C and D) for whom naive cell counts were available. X-axis indicates relative importance of each feature from 0 to 1. This can be interpreted as percentage scale, where 1 = 100%. Bars are colored according to prevalence of feature: white, black, and gray for features more common, less common, or similar, respectively, in patients with infections.

Supplementary Fig II