Fig. 2. 95% Confidence interval (thin solid line) and SD (thick solid line) for a hypothetical noninferiority trial comparing Custodiol and plegisol (active control), using serum area under the curve troponin I concentration ([cTnI]) as the primary outcome variable (treatment effect). The confidence interval is calculated from the difference between the treatment effect of the two cardioplegia solutions. (A ) Shows noninferiority of the two solutions;i.e. , the upper end of the confidence interval is 0.5 SD or less (δ). (B ) Depicts the converse (see text).