Research paper
Nørrelund H. The metabolic role of growth hormone in humans with particular reference to fasting. Growth Horm IGF Res. 2005 Apr;15(2):95-122. doi:10.1016/j.ghir.2005.02.005. PMID: 15809014.
Why we cite it
A 28-page review (pp. 95–122, per the PubMed record) of growth hormone's metabolic role in humans during fasting. **Nothing of it is held here** — PubMed carries no abstract and the full text is paywalled — so what this record says of it arrives through secondary literature citing it. As that literature reports it, the review frames GH as a master fuel switch from carbohydrate and protein to lipid utilisation during fasting, with lipolysis as the route to protein sparing. **That is the review's proposed account, not this library's finding.** `cgh05`, the library's one direct test (held in full, with Nørrelund its second author), blocked the GH receptor over a 36-hour fast and found lipid mobilisation suppressed but protein metabolism unaffected. (Corrected 2026-09-10, card E1AWST, from *"Provides the mechanistic framework … that explains why fasting preserves muscle"* and an undated count, *"Referenced by 3 k-files (11 references) in this corpus, making it the most heavily referenced citation for GH's protein-sparing mechanism"*. The count matched the corpus on 2026-09-10 — three knowledge files, eleven anchors — but *most heavily referenced* was never surveyed, and a count with no date goes stale without anyone noticing.)
**FLAGGED 2026-09-05 (card MERNGY) AND REPAIRED 2026-09-10 (card E1AWST, on the Auditor's ruling that the repair is this desk's) — A CONFIRMATION VERB ON A REVIEW NOBODY HERE HAS READ.**
The bullet below that read *"Inhibiting lipolysis eliminates GH's ability to restrict protein loss during fasting, **confirming the causal mechanism**"* stated a confirmation **in this record's own voice**, for a 28-page review that is **paywalled and has never been read here** — its claims arriving, by this record's own Verification depth, *"through secondary sources citing this review."*
**Compare `cif01`, which blankets its extracted knowledge with *"These are Fung's assertions, transcribed — not this library's findings."*** This record carried no such framing, so a two-removes report read as the library's own finding. **It now does** (see Extracted knowledge).
**And `cgh05` ran that exact chain forwards and it came out empty** — see the back-link above. **The wording was left as found so that the repair would be made once, deliberately, by the desk that owns it; this is that repair.**
DOI: 10.1016/j.ghir.2005.02.005 · PubMed 15809014 · Publisher