The purpose of this global protocol amendment (protocol Version 2.0) was to change the categories for assessing the relationship of any clinical adverse experiences to study medication from Related and Unrelated to Probable, Possible, Unlikely, and Not Related. It was also clarified that: • Suspected Unexpected Serious Adverse Reactions are serious adverse events (SAEs) that were judged by the Investigator or the Sponsor to be possibly or probably related to the study treatment administered and assessed by the Sponsor and/or its delegate as unexpected. • Any SAE that occurred outside of the reporting period, for which there was a reasonable possibility of a study drug relationship in the opinion of the Investigator, had to be reported as soon as the site became aware of the event.

The purpose of this global protocol amendment (protocol Version 3.0) was to facilitate patient recruitment into the study by: A) Removing the upper limit of agitation severity from the inclusion criterion: Therefore, any patients judged to be clinically agitated at Baseline with a value of ≥ 4 out of the 7 items on the CGI-S could be eligible for inclusion in the study. Previous pivotal clinical studies assessing the safety and efficacy of inhaled loxapine or aripiprazole in acutely agitated patients with schizophrenia or bipolar disorder did not specify an upper limit on agitation severity in the eligibility criteria. The eligibility criteria current at the time of the amendment already addressed certain considerations associated with the removal of the upper limit of agitation severity from the inclusion criterion and so limited the recruitment of patients who were most severely affected in a clinically meaningful manner: 1) patients recruited under the protocol amendment, regardless of the severity of their agitation, were required to have documented adequate consent capacity per the Investigator’s judgement and to provide written informed consent (except in Spain); 2) the study was designed to ensure that any patients who were likely to be unable to use the inhalation device were excluded from the study. B) Removing the restriction of the use of injectable depot neuroleptics or long-acting second generation antipsychotics within 1 dose interval prior to study drug administration: This change was feasible as the efficacy and safety of the study drug to treat the acute agitation episode could be assessed in this pragmatic clinical study, regardless of the medication used to treat the underlying condition. Removing this eligibility criterion better reflected the clinical practice and standard of care (SOC) of acutely agitated patients with schizophrenia or bipolar disorder.

The purpose of this global protocol amendment (protocol Version 4.0) was to facilitate patient recruitment into the study on one hand, and also to better reflect the usual clinical practice by changing the Exclusion Criterion 4. Patients were eligible for the study if they were not treated with benzodiazepines or other hypnotics or oral or short-acting IM antipsychotics within 1 hour prior to study drug administration, whereas the previous protocol version stated a 4-hour time interval. The effect of benzodiazepines or other hypnotics or oral or short-acting IM antipsychotics can be seen quite early after administration and, if patients were still judged to be clinically agitated 1 hour later, they were allowed to be enrolled into the study. This change was feasible as the efficacy and safety of the study drug to treat the acute agitation episode could be assessed in this pragmatic clinical study. Changing this eligibility criterion better reflected the clinical practice and SOC of acutely agitated patients with schizophrenia or bipolar disorder and enabled the enrolment of a subject population that is closer to the patient population in which inhaled loxapine and aripiprazole is being used in clinical practice. Other changes made in this protocol amendment included the extension of the planned duration of the clinical study from up to approximately 11 months to up to approximately 36 months.

The purpose of this global protocol amendment was to update the statistical analysis methodology. The current trial characteristics (i.e., short time of follow-up and condition under study) would not lead to early study withdrawals. For this reason, the analysis planned initially (i.e., the Mantel-Haenszel log-rank test) would not represent any advantage over other standard methods. Therefore, the purpose of this amendment was to document the use of the Wilcoxon rank-sum test, instead of the Mantel-Haenszel log-rank test, to assess the primary efficacy analysis. Additionally, the original analysis (i.e., the Mantel-Haenszel log-rank test) would also be conducted for sensitivity purposes. The inferential analysis was also conducted using the stratified non-parametric van Elteren test, using modified ridit scores, which is as a direct extension of the Wilcoxon rank-sum test for two samples. Other changes made in this global protocol amendment were to: • Confirm that the assessment of the initial sample size estimation would be based on the interim analysis, Stage 1 data of the primary end point of time to response of CGI-I and would be used as estimates in a calculation of the sample size at one-sided 0.025 alpha level and approximately 90% power • Clarify that there would be no adjustment for multiplicity upon the stated alpha level for this study • Clarify that meaningful time (difference between median aripiprazole time to response and median loxapine time to response) was approximately 30 minutes (interim analysis) • Update some administrative information (change in Medical Monitor, Monitoring Committee constitution, Biostatistician)