TREC Based Newborn Screening for Severe Combined Immunodeficiency Disease: A Systematic Review

Abstract

Background: Newborn screening (NBS) by quantifying T cell receptor excision circles (TRECs) in neonatal dried blood spots (DBS) enables early diagnosis of severe combined immunodeficiency disease (SCID). In recent years, different screening algorithms for TREC based SCID screening were reported.

Purpose: To systematically review the diagnostic performance of published algorithms for TREC based NBS for SCID.

Methods: PubMed, EMBASE and the Cochrane Library were systematically searched for case series and prospective cohort studies describing TREC based NBS for SCID. We extracted TREC content and cut-off values, number of retests, repeat DBS and referrals, and type and number of typical SCID and other T cell lymphopenia (TCL) cases. We calculated positive predictive value (PPV), test sensitivity and SCID incidence.

Results: Thirteen studies were included, re-confirming 89 known SCID cases in case series and reporting 53 new SCID cases in 3.15 million newborns. In case series, the sensitivity for typical SCID was 100%. In the prospective cohort studies, SCID incidence was ~1.7:100,000, re-test rate was 0.20-3.26%, repeat DBS rate 0.0-0.41% and referral rate 0.01-1.35%. PPV within the five largest cohorts was 0.8-11.2% for SCID and 18.3-81.0% for TCL. Individual TREC contents in all SCID patients was <25 TRECs/μl (except in those evaluated with the New York State assay).

Conclusions: The sensitivity of TREC based NBS for typical SCID was 100 %. The TREC cut-off score determines the percentage of non-SCID TCL cases detected in newborn screening for TCL. Adapting the screening algorithm for pre-term/ill infants reduces the amount of false positive test results.

Fig. 1

Flowchart of article selection for systematic review of TREC based NBS for SCID. Exclusion criteria were applied hierarchically. Studies were included as case series if stored Guthrie cards of known patients were tested. Studies were included as cohort studies if newborns were tested prospectively and sufficient details on algorithm outcome were reported

Fig. 2

Neonatal TREC content of individual patients with typical SCID (a) and other TCL (b) diagnoses at different screening sites (colors). TREC content reported in TRECs/reaction was converted to TRECs/μl. Partial DiGeorge syndrome and 22q11 deletion syndrome are displayed together as DiGeorge (b). Thirty three of 74 typical SCID patients displayed had 0 TRECs, compared to 22 out of 182 other TCL patients. Anonimyzed SCID patients had ADA-SCID (n = 4), γ chain-SCID (n = 2), Omenn-SCID (n = 2), RAG-SCID (n = 2), PNP-SCID (n = 1), undefined SCID (n = 7)

Fig. 3

Neonatal TREC content of individual patients with typical SCID (a) and other TCL (b) diagnoses divided by screening site-specific TREC cut-off value. Colors correspond to different data sources

Fig. 4

The percentage of typical SCID and other TCL cases of which the individual TREC score was below a TREC cut-off score from 0 to 25 TRECs/μl. Patients included in this figure are all patients for which an individual TREC score was reported in the included articles. Two typical SCID cases reported by Vogel et al. [33] had a TREC value above 25, as is consistent with the New York cut-off score of 200 TRECs/μl