Newborn Screening for Primary Congenital Hypothyroidism: Estimating Test Performance at Different TSH Thresholds

Context Active surveillance of primary congenital hypothyroidism (CH) in a multiethnic population with established newborn bloodspot screening. Objective To estimate performance of newborn screening for CH at different test thresholds and calculate incidence of primary CH. Design Prospective surveillance from June 2011 to June 2012 with 3-year follow-up of outcomes. Relative likelihood ratios (rLRs) estimated to compare bloodspot TSH test thresholds of 6 mU/L and 8 mU/L, with the nationally recommended standard of 10 mU/L for a presumptive positive result. Setting UK National Health Service. Patients Clinician notification of children aged <5 years investigated following clinical presentation or presumptive positive screening result. Main Outcome Measure(s) Permanent primary CH status determined by clinician report of continuing T4 requirement at 3-year follow-up. Results A total of 629 newborns (58.3% girls; 58.7% white ethnicity) were investigated following presumptive positive screening result and 21 children (52.4% girls; 52.4% white) after clinical presentation; 432 remained on treatment at 3-year follow-up. Permanent CH incidence was 5.3 (95% CI, 4.8 to 5.8) per 10,000 infants. With use of locally applied thresholds, sensitivity, specificity, and positive predictive value were 96.76%, 99.97%, and 66.88%, respectively. Compared with a TSH threshold of 10 mU/L, positive rLRs for 8 mU/L and 6 mU/L were 1.20 (95% CI, 0.82 to 1.75) and 0.52 (95% CI, 0.38 to 0.72), and negative rLRs were 0.11 (95% CI, 0.03 to 0.36) and 0.11 (95% CI, 0.06 to 0.20), respectively. Conclusions Screening program performance is good, but a TSH threshold of 8 mU/L appears superior to the current national standard (10 mU/L) and requires further evaluation. Further research should explore the implications of transient CH for screening policy.