Randomized controlled trials in children’s heart surgery in the 21st century: a systematic review

Objectives: Randomised controlled trials are the gold standard for evaluating healthcare interventions yet are uncommon in children’s heart surgery. We conducted a systematic review of clinical trials in paediatric cardiac surgery to evaluate the scope and quality of the current international literature.
Methods: We searched MEDLINE, CENTRAL, LILACS and manual screening of retrieved references and systematic reviews, to identify all randomised controlled trials reporting the effect of any intervention on the conduct or outcomes of heart surgery in children published in any language since January 2000; secondary publications and those reporting inseparable adult data were excluded. Two reviewers independently screened studies for eligibility and extracted data; the Cochrane Risk of Bias tool was used to assess for potential biases.
Results: We identified 333 trials from 34 countries randomising 23,902 children. Most were early phase (313, 94.0%), recruiting few patients (median 45, IQR 28-82), and only 11 (3.3%) directly evaluated a surgical intervention. 109 (32.7%) trials calculated a sample size, 52 (15.6%) reported a CONSORT diagram, 51 (15.3%) were publicly registered and 25 (7.5%) had a Data Monitoring Committee. Overall risk of bias was low in 22 (6.6%), high in 69 (20.7%) and unclear in 242 (72.7%).
Conclusions: The recent literature in children’s heart surgery contains few late phase clinical trials. Most trials did not conform to accepted standards of reporting and the overall risk of bias was low in few studies. There is a need for high quality, multi-centre clinical trials to provide a robust evidence-base for contemporary paediatric cardiac surgical practice.