APF530 for nausea and vomiting prevention following cisplatin: phase 3 MAGIC trial analysis

Eligible men and women were 18-80 years of age with histologically or cytologically confirmed malignancy (cancer type information was not captured) and were entering the first cycle of their single-day HEC treatment (defined by ASCO 2011 emetogenicity criteria).⁷ Patients had Eastern Cooperative Oncology Group Performance Status (ECOG-PS) of 0 or 1, no history or presence of significant cardiac disease or QT interval prolongation, and adequate bone marrow, kidney, and liver function. All patients provided written informed consent.

Procedures

Patients were stratified by planned receipt of the cisplatin regimen ≥50 mg/m² (Yes/No), randomized 1:1 to receive APF530 500 mg SC (granisetron 10 mg) or ondansetron 0.15 mg/kg IV (up to a maximum of 16 mg as a single dose) on day 1 (Figure 1). The APF530 arm received the ondansetron saline placebo, and the ondansetron arm received the APF530 SC placebo containing the TEG-POE vehicle. All patients were scheduled to receive fosaprepitant 150 mg IV and dexamethasone 12 mg IV on day 1, then oral dexamethasone 8 mg once daily on day 2 and 8 mg twice daily on days 3 and 4. Rescue medication was allowed at the investigator’s discretion.

Outcomes

The primary objective of the trial was to demonstrate the superiority of APF530 500 mg SC compared with ondansetron 0.15 mg/kg IV, as part of the current guideline-recommended 3-drug regimen, in preventing delayed-phase CINV after HEC. The primary endpoint was delayed-phase (24-120 h) CR (no emetic episodes [vomit or retch] and no rescue medication use). In addition, a prespecified analysis of delayed-phase CR by randomization strata (planned use of cisplatin) was performed.

,

Secondary and other endpoints included overall-phase CR
(0-120 h); delayed-, overall-, and acute-phase complete control (CC: CR and no more than mild nausea); delayed-, overall-, and acute-phase total response (TR; CR and no nausea); and rescue medication use. A post hoc analysis of nausea severity was also conducted. Safety assessments included treatment-emergent adverse events (TEAEs), injection-site reactions (ISRs), laboratory parameters, and vital signs. TEAEs were assessed by type, duration, severity, and relationship to study drug. ISR timing and severity were captured in patient diaries.

Statistical analysis

All efficacy analyses were conducted using the modified intent-to-treat population (mITT; randomized patients who received study drug and a HEC regimen and had post-baseline efficacy data). Safety assessments were performed on the safety population (randomized patients who received study drug).

This analysis conducted on the subgroup of patients with intent to receive cisplatin (cisplatin randomization stratum, ≥50 mg/m², Yes) was exploratory and was not powered to detect treatment differences. Preplanned analyses compared CR, CC, and TR rates across treatment arms using 95% CIs.

Post hoc analyses of time to first rescue medication use, proportion of patients with rescue medication use, and less frequent nausea were performed. All P values were calculated using the Cochran-Mantel-Haenszel chi square test. Rescue medication use results were based on observed data, without imputation for missing results (ie, calculated from the number of patients with a response). Further analyses of efficacy endpoints CR, CC, and TR in the subset of female patients in the cisplatin randomization stratum were performed. Safety assessments were summarized descriptively.

Results

A total of 942 patients were randomized across 77 US centers during March 31, 2014 and May 15, 2015 (471 APF530, 471 ondansetron). Among those, 264 had intent to receive cisplatin and were included in the cisplatin randomization stratum (≥50 mg/m², Yes) (Figure 2). A total of 256 patients in the cisplatin stratum received study drug and were included in the safety population (126 APF530, 130 ondansetron); 252 patients were included in the mITT population (124 APF530, 128 ondansetron).

Baseline demographics were generally balanced between treatment arms (Table 1). The proportion of female patients was 41.1% (51/124) and 48.4% (62/128) in the APF530 and ondansetron arms, respectively. The majority of patients had an ECOG PS of 0 (57.3% [71/124] APF530; 60.2% [77/128] ondansetron). The most common cisplatin-based chemotherapy regimen in both treatment arms was cisplatin and gemcitabine (25.0% [31/124] APF530; 28.9% [37/128] ondansetron) (Suppl Table 1). Two patients in the APF530 arm and 3 patients in the ondansetron arm either received a lower cisplatin dose (<50 mg/m²) or did not go on to receive cisplatin as intended at randomization (Suppl Table 1). As previously reported, in the cisplatin stratum (Table 2), delayed-phase CR was numerically higher in the APF530 arm versus the ondansetron arm, with a corresponding treatment difference of 10.6% (65.3% [81/124] APF530; 54.7% [70/128] ondansetron; 95% CI [-1.4, 22.7]; P = .085). Although the CI contains 0, the result is consistent with the significant benefit observed in the overall study population (64.7% [291/450] APF530; 56.6% [256/452] ondansetron; 95% CI [1.7, 14.4];
P = .014).¹⁸ This more in-depth analysis found similar trends favoring the APF530 over the ondansetron regimen across overall- and acute-phase CR (Table 2).

CC rates were consistently higher across all phases in the APF530 arm compared with the ondansetron arm, with treatment differences ranging from 10.5% to 8.1%. For TR, the most stringent measure of CINV control, there were trends favoring the APF530 arm over the ondansetron arm in acute, delayed, and overall phases (Table 2). Among female patients in the cisplatin stratum, the numerically higher trends for CR, CC, and TR in the APF530 arm versus the ondansetron arm persisted across acute, delayed, and overall phases (Suppl Table 2).

A significantly greater proportion of patients in the APF530 arm, compared with the ondansetron arm, reported no rescue medication use during the delayed phase (74.4% [90/121] APF530; 62.6% [77/123] ondansetron; P = .048). Trends in favor of APF530 were observed in the overall phase (71.1% [86/121] APF530; 61.8% [76/123] ondansetron; P = .125) and acute phase (86.9% [106/122] APF530; 81.9% [104/127] ondansetron; P = .278). Time to first rescue medication use was consistently longer in the APF530 arm, compared with the ondansetron arm, although not statistically significantly (P = .150) (Figure 3).

In a post hoc analysis of nausea frequency, the APF530 regimen was associated with a trend toward less frequent nausea (0-2 vs ≥3 episodes), although not a statistically significant difference, compared with the ondansetron regimen in the delayed and overall phases (Suppl Table 3).

The APF530 regimen was generally well tolerated in the cisplatin subgroup, and no new safety signals were identified (Table 3). Most patients experienced at least one TEAE. Excluding ISRs, TEAE incidences were 72.2% and 66.9% in the APF530 and ondansetron arms, respectively; most common were constipation, fatigue, nausea, diarrhea, dehydration, and headache. Excluding ISRs, the most common treatment-related TEAEs in the APF530 and ondansetron arms were constipation (2.4% and 2.3%, respectively and headache (3.2% and 4.6%).

ISRs occurred in 49.2% and 54.6% of patients in the APF530 and ondansetron arms, respectively (Table 3); all ISRs were considered by the sponsor to be treatment related. ISRs were generally mild or moderate in severity. No ISRs were considered serious TEAEs, and most resolved by study end. One patient in the APF530 arm and none in the ondansetron arm experienced a treatment-related serious TEAE (atrial fibrillation, APF530). There were 2 TEAEs leading to death, 1 in the APF530 arm and 1 in the ondansetron arm. Both were acute renal failure due to underlying disease, and neither was considered related to study drug.