Long-acting growth hormones in East Asia: comparative efficacy and safety
Liu Q, Wu C, Peng D, et al. · Endocrine Practice. 32(8):1303–1310. · Online 1 April; issue August 2026.
This five-study network analysis included Pegpesen, Jintrolong, and somapacitan. For height velocity, the Pegpesen estimate versus daily GH was −0.10 cm/year (95% CI −0.53 to 0.32). The authors reported more favorable growth estimates for Jintrolong. Differences in adverse-event and serious-adverse-event rates were not statistically significant.
Interpretation: the figures above are the original authors’ reported results, not independently reproduced estimates. Alongside the limits of indirect comparisons, a forthcoming letter raises specific questions about study identification and safety-data handling. Read the methodological correspondence below before interpreting comparative rankings.
Data provenance and safety analysis: three questions requiring reassessment
Robin Wan. Request for Editorial Review of Data Provenance and Outcome-Altering Analytical Issues. Letter to the Editor, Endocrine Practice.
Acceptance reported by the author; journal publication details and DOI are pending. This summary is based on the author-supplied revised manuscript.
The letter requests an editorial reassessment of the populations and safety outcomes used in Liu et al.’s analysis. It identifies the following issues and calls for access to the extraction data and executable model code.
- Trial identity and the analyzed population. Wan reports that the study-characteristics table labels Luo 2017 as Phase II (108 participants), while the extracted Jintrolong efficacy and adverse-event values match Phase III (343 participants). If that Phase III cohort supplied the modeled data, the corresponding total would be 1,087 rather than 852 (852 − 108 + 343). The original Luo report confirms the two distinct trial populations; the network’s actual input file is needed to resolve which cohort was analyzed.
- Exclusion of the 391-participant Pegpesen Phase III safety dataset. According to the letter, this trial was included for efficacy but excluded for safety on the basis that adverse events were insufficiently differentiated. Its published Table 4 does report treatment-emergent adverse events, drug-related events, discontinuations, and serious adverse events over 52 weeks. Wan argues that relying on the smaller, 12-week Phase II dataset warrants reassessment with aligned event definitions and follow-up periods.
- Severe versus serious events, and a double-zero comparison. The letter questions whether the source statement about no severe events was recoded as no serious adverse events (0/228 versus 0/115), then used to produce a relative risk and SUCRA ranking. Severity and seriousness are different concepts. As the letter emphasizes, a comparison with no observed events in either arm does not, by itself, establish a direction of relative safety; any rank requires scrutiny of coding and model assumptions.
What the Pegpesen Phase III source actually reports
Participants with at least one event, n/N (%), over 52 weeks. These are participant counts, not the total number of event episodes.
| Outcome | Pegpesen (N = 261) | Daily rhGH (N = 130) |
|---|---|---|
| Any treatment-emergent adverse event | 229/261 (87.7%) | 123/130 (94.6%) |
| Drug-related treatment-emergent adverse event | 52/261 (19.9%) | 26/130 (20.0%) |
| Any serious adverse event | 17/261 (6.5%) | 8/130 (6.2%) |
This single-trial table establishes that differentiated safety data are available. It is not a reanalysis of the network and does not, by itself, establish a comparative safety advantage.
Requested next steps: verify study-by-outcome extraction, reproduce the model independently, reconcile study identities and denominators, reassess inclusion of Phase III safety data, and remove rare-event rankings that lack support.
The concerns are attributed to Wan’s letter. They should not be represented as an issued journal correction, a retraction, or a finding of misconduct. The letter itself states that the discrepancies do not establish misconduct.