30-Second Takeaway
- Oral nonpeptide GLP-1 RA HRS-7535 produced clinically meaningful HbA1c reductions versus placebo in phase 2.
- CGM modestly improves treatment satisfaction versus SMBG but evidence for quality-of-life benefits is inconclusive.
Week ending June 6, 2026
Selective evidence brief: GLP-1 RA advances, CGM patient-reported outcomes, PanNET molecular risk, and AI for guideline surveillance
Oral HRS-7535 added to metformin lowers HbA1c and causes modest weight loss
In a 16-week phase 2 randomized trial (n=194) oral HRS-7535 added to metformin reduced HbA1c by -1.82% (60 mg group) with placebo-adjusted differences up to -1.57% (all P<.001). Between 48.7% and 63.2% of treated patients achieved HbA1c <7.0% versus 15.4% with placebo. Weight fell modestly, greatest at 90 mg (-2.63%), and adverse events were mostly mild-to-moderate gastrointestinal events with rare hypoglycemia. These results are from a short (16-week) phase 2 program in China and require phase 3 confirmation and broader safety data before practice change.
UK survey: high GLP-1RA satisfaction but concerns and misuse reported
A cross-sectional UK survey (n=684; 544 current/former users) found high user satisfaction with GLP-1RAs for weight management. Respondents reported reduced energy, altered dietary habits, and concerns about loss of muscle mass and long-term safety. Many sourced medication from online pharmacies and social media, and a minority reported dose-exceeding behaviors consistent with misuse. These findings signal need for clinician counseling on sourcing, dose adherence, and monitoring muscle mass and functional status.
AI pre-ranking with cutoff improves efficiency and recall for living guideline maintenance
In a simulation using 854 articles across 38 topics, overarching single-database AI-enabled pre-ranking achieved 92–96% recall versus 76–89% for topic-specific searches. Overarching searches reduced screening workload by 63–70%, lowering articles screened per relevant study to 28–48 versus 76–160. Topic-specific searches were more efficient per topic but had significantly lower recall, and AI ranking added modest guideline-level gains. Consider overarching single-database AI pre-ranking to preserve recall while substantially reducing screening burden for living guidelines.
References
Numbered in order of appearance. Click any reference to view details.
Additional Reads
Optional additional studies from this edition.