Integrated Summary of Safety (ISS) — GLPI-103

Integrates safety across the GLPI-103 clinical program. Pivotal Phase 3 figures trace to the executed analyses (outputs/, ADAE, SAR-301) and are reconciled to source; early-phase pooled content is illustrative [MOCK]. Design parameters cite REF-002.

Change History

1. Extent of Exposure and Pooled Safety Population

The safety database integrates Phase 1 (GLPI103-101 SAD N=40; GLPI103-102 MAD N=40, ≤12 weeks), Phase 2 (GLPI103-201 N=240, 24 weeks), and the pivotal Phase 3 study (GLPI103-301 N=900, 52 weeks). The Phase 3 study provides the controlled, long-term safety database and the basis of this summary; all 900 randomized subjects received ≥1 dose and constitute the Safety Set. Exposure in the pivotal study extended to 52 weeks of double-blind treatment (≈220 subjects per arm), providing adequate person-time to characterize the long-term and class-related safety profile. Early-phase exposure characterized single- and multiple-ascending doses and the dose–response, with no dose-limiting toxicity [MOCK].

2. Overall Adverse Event Summary (Phase 3 Safety Set)

Adverse events were predominantly mild or moderate (469 mild and 255 moderate of 797 total events; 73 severe), and 564 of 797 events (≈71%) were assessed as treatment-related. Notably, the overall TEAE burden was numerically lower with both GLPI-103 arms (75.7%, 67.8%) than with the active comparator (72.2%), indicating that the greater efficacy of GLPI-103 was not accompanied by a greater overall adverse-event burden.

3. Common Adverse Events by SOC and Preferred Term

Gastrointestinal disorders were the most frequently affected system organ class, consistent with the GLP-1 receptor agonist class (CSR-301 §12; M2.7.4):

Nausea was significantly more frequent with GLPI-103 IV (32.3%) than semaglutide (21.4%; p=0.003), reflecting the efficacy–tolerability trade-off of greater GLP-1 engagement in the higher-exposure IV arm; vomiting and decreased appetite were more frequent with semaglutide. Gastrointestinal events were predominantly mild–moderate and managed by the protocol-defined titration. Injection-site reactions were, as expected, most frequent with the intravenous formulation (5.0%).

4. Adverse Events of Special Interest (AESI)

• Gastrointestinal events: the principal class effect; mild–moderate and titration-manageable (above).
• Hypoglycaemia: infrequent (~3–6%; numerically higher with semaglutide). A recurrent-event analysis (overdispersion-robust quasi-Poisson rate model; SAR-301 §9) showed no significant between-arm difference in hypoglycaemia event rate (IV vs semaglutide rate ratio ≈0.97, ns); events were sparse and predominantly mild.
• Acute pancreatitis and thyroid C-cell neoplasia (MTC): none recorded; monitored as class risks with labeling and pharmacovigilance (RMP suite).
• Cardiovascular / heart rate (APJ-related): no clinically meaningful signal in vital-sign data; transient mechanism-related heart-rate changes were anticipated from nonclinical data and were not clinically significant.
• Immunogenicity: low-titre, non-neutralizing anti-drug antibodies without impact on efficacy or safety [MOCK].

5. Hepatic Safety (eDISH / Hy's Law)

Hepatic safety was evaluated using the eDISH framework (peak ALT versus peak total bilirubin, in multiples of the upper limit of normal; SAR-301 §10; Figure 4). Across 900 subjects, only 1 subject had peak ALT >3×ULN and no subject met Hy's-law criteria (ALT >3×ULN with total bilirubin >2×ULN). There was no signal of drug-induced serious hepatotoxicity, consistent with the GLP-1 class.

6. Serious Adverse Events, Deaths, and Discontinuations

Seventeen serious adverse events were reported, including three deaths (one per arm; none assessed as treatment-related). The serious events and deaths showed no pattern by treatment and are consistent with the background morbidity and mortality expected in an older T2DM population over 52 weeks. Overall study discontinuation was 9.3% (84/900), balanced across arms, with adverse-event-related discontinuations infrequent (n=5); the principal reasons for discontinuation were lost-to-follow-up, physician decision, and withdrawal by subject (CSR-301 §10.1).

7. Laboratory Findings and Vital Signs

Hepatic (ALT/AST) and renal (eGFR) laboratory parameters were stable, with mean changes near zero and no pattern of treatment-related hepatotoxicity or nephrotoxicity (shift analyses and eDISH in §5). Vital signs, including heart rate, showed no clinically meaningful change across arms. ECG endpoints were not captured in the simulated dataset [MOCK].

8. Safety in Special Populations and by Intrinsic Factors

Subject incidence of TEAEs was broadly consistent across age (<65 / ≥65 years), sex, and BMI subgroups; no subgroup demonstrated a disproportionate safety burden. Subjects with eGFR <45 mL/min/1.73 m² or ALT/AST >3×ULN were excluded by protocol, so the safety database does not characterize moderate-to-severe renal or hepatic impairment; these are addressed by labeling and post-marketing commitments (RMP). Use in pregnancy/lactation is contraindicated per class labeling.

9. Conclusions

The integrated safety profile of GLPI-103 is class-consistent and generally well tolerated, dominated by predominantly mild-to-moderate gastrointestinal adverse events that were manageable with titration. The principal feature is a gastrointestinal tolerability trade-off (higher nausea) in the higher-exposure IV arm, managed by titration; hypoglycaemia was infrequent and not increased, and there was no hepatotoxicity (no Hy's-law cases), pancreatitis, or thyroid signal. The serious adverse events and deaths showed no relationship to treatment. The profile supports a favourable benefit–risk assessment (M2.5.6) and is appropriately managed by routine risk-minimization and pharmacovigilance activities (RMP-EU / REMS / RMP-KR).