A formal amendment to the original protocol was to change the primary method for statistical analysis from an analysis of covariance (ANCOVA) model based on the last observation carried forward (LOCF) to MMRM and to add additional clarity to some trial procedures. Clarify that every effort was made to complete efficacy scales prior to administering rescue medication at the last trial (or early termination) visit and that efficacy scales were not done if a new antipsychotic was given before the scales were completed. Add change from Baseline in PANSS Excited Component and PANSS Marder Factor scores as secondary efficacy variables. Clarify that participants who were sterile (ie, women who have had an oophorectomy and/or hysterectomy or have been postmenopausal for at least 12 consecutive months; or men who have had orchidectomy) were not required to use two different methods of birth control and add “other approved birth control device” to the list of acceptable birth control methods. Clarify instructions for preparation of whole blood sample for metabolic profiling. Clarify that participants may remain on stable doses of propranolol during the trial if the propranolol is being taken for an indication other than akathisia. Revise Appendix 9 to include the correct version of the PANSS (ie, 2006). The 2006 version of the PANSS was the version that was distributed to the sites at the start of the trial and was used by raters throughout the trial. In addition, administrative changes were made and typographical errors identified during review of the protocol amendment were corrected.

A second formal amendment to the original protocol was done to change the primary and secondary efficacy analysis used the Hochberg procedure to control multiplicity. This testing scheme was replaced by the Average Method to be carried out in 2 steps to control the family-wise error rate. In addition, administrative changes were made.