What is typical "poolability criteria" for pooling control (sham in this case) arms? Standard errors are provided but not confidence intervals. What can we learn, if anything, from data that has been provided by APLS with respect to the poolability of sham arms that would not result in any misinterpretation of efficacy? 

What are the consequences, if any, to changing the primary endpoint from 12 months to 18 months? While masking may remain in tact, could bias be introduced in the analysis of post-hoc later time points (ie - 2:1 randomization, AEs)? Are any adjustments to the "standard" p-value hurdle of 0.05 likely required? 

Please describe MMRM analysis in layman's terms. What information regarding dropouts or missed injections would be required to determine the appropriateness of MMRM analysis? APLS has noted an increasing effect of pegcetacoplan over time (at 18 and 24 months relative to 0-12 months). Could this be the result of MMRM bias and not truly an unexpected and unexplained biological effect? How is such bias typically corrected? What sensitivity analyses would you expect from the FDA to ensure the result is not being misinterpreted? 

It's my understanding that by requiring a slope analysis as opposed to looking at lesion change from baseline at a given timepoint, the FDA is in essence suggesting that the slope (or rate of change) is expected to be constant. I struggle with the idea then that the slope would effectively change at some point after 12 months and that not be attributed to variability but rather an increasing biological effect over time. Can you help make sense of this? Have you observed this in any other clinical trials?