Biventricular pacing (CRT) shows clear benefits in heart failure with wide QRS, but results in patients with narrow QRS have been conflicting. In this meta-analysis we tested the hypothesis that trial design might have influenced findings.Method and results
We identified all reports of CRT-P/D therapy in subjects with narrow QRS. 12 studies (2074 patients) met the inclusion criteria.Studies were assessed for the presence of bias resistance steps, i.e. a randomised control arm, and blinded outcome measurement. The effect on each endpoint was quantified using a standardised z score (Figure 1). The weighted-mean effect size for CRT studies with no bias resistance features was 0.75 (95% CI 0.45 to 1.05). For those with one feature, randomization but not blinding, it was 0.84 (95% CI -0.14 to 1.81). For those with two features (i.e. blinded, randomised controlled trials), it was -0.02 (95% CI -0.10 to 0.06). Study design had a strong effect on results (p = 0.0007 for contrast between the three groups).Conclusions
Only in studies which do not have randomization and full blinding is CRT found to be effective in patients with narrow-QRS heart failure. When randomization and full blinding are implemented, there appears to be no benefit. This is driven by study design, not choice of endpoint. This means that either randomization or blinding inhibit positive physiological effects or that lack of blinding and randomization introduce positive bias unintentionally.