Purpose: A2310 IVUS study showed that everolimus (EVR) provides cardiac allograft vasculopathy (CAV) benefit compared to mycophenolate mofetil (MMF) by assessing the mean change in average maximum intimal thickness (MIT) between baseline and Month 12 (M12). However, only half of the enrolled patients had evaluable intravascular ultrasound (IVUS). Imputations for missing data were performed under statistical and medical perspectives and results are presented here.

Methods: The IVUS population consisted of patients from prospectively selected sites with IVUS capability. The primary IVUS efficacy endpoint was the change in average MIT (surrogate for CAV) from baseline to M12. Clinical condition precluding performance of IVUS was the main reason for missing values which were then imputed primarily using three methods [Table 1]. Method A was based on missing at random assumption; methods B and C were based on missing values not at random [e.g. probability of missing values depended on missing data itself due to a medical reason (renal dysfunction, death)].

Results: Mean change in average MIT from baseline to M12 was significantly smaller in the EVR group vs. MMF group regardless of the type of imputations used [Table 2].

Conclusion: Benefit of EVR on CAV compared to MMF is supported by different imputation methods for missing IVUS data.

« Back to 2015 American Transplant Congress