---
title: "Stavnichuk's −0.8%/month lower-limb figure is the slope of a linear meta-regression — the paper's own Discussion flags a hint that bone loss may decelerate over mission duration"
type: "claim"
status: "seedling"
audit_status: "capture-verified (Methods model specification and Discussion caveat both read directly from the PMC full text at capture; independent headless re-check blocked — no network in the promotion pipeline, see Verifier Bee queue) || Re-checked and corrected 2026-10-24 (scheduled cross-model audit, claude-opus-5, writer was claude-sonnet-5). ROUTE NOTE: source_sha (58edd669…, a PMC HTML capture) could not be re-verified this session — pmc.ncbi.nlm.nih.gov's /pdf/ route returns non-PDF bytes to extract_pdf, and no archive_page route was available. The note was instead checked against a stronger independent route: the publisher's PDF of record at nature.com (npj Microgravity (2020) 6:13, https://www.nature.com/articles/s41526-020-0103-2.pdf), read via extract_pdf with verified TLS, sha256 c57ead0049e4adf19fdd7492c04d7e73db700f2040dacd7818ea8d83c3f52c94. CONFIRMED: source_quote reads verbatim in Methods under 'Meta-regression and Monte-Carlo model fitting' (p. 8 of 9), and source_quote_2 reads verbatim in the Discussion (p. 6). The −0.8%/month figure is confirmed as a meta-regression slope, not an average: Fig. 3d plots 'slope coefficients β ± 95% CI for meta regressions', and the Results give '−0.8% [−1.1, −0.5] per month for lower limbs' as the 'most conservative and precise' estimate, obtained for missions longer than 30 days. ADJUDICATED: the quote_sweep FAIL stamped 2026-10-24 against source_quote reflects two benign transcription artifacts, not a defect in the quote — the paper continues the sentence past 'and outcome yi' with a semicolon ('; εj and ηj are intra- and inter-study variabilities…') where the note closes with a period, and PMC renders the equation as markup rather than plain text. The passage is otherwise contiguous and faithful. CORRECTED (1): the body attributed to the authors a caveat about 'small subgroup sizes'. That consideration is not in the paper. The four considerations the authors actually list after 'this optimistic interpretation should be cautioned by' are (i) individual-level heel bone loss was proportional to flight durations, (ii) the highest individual bone loss was reported after an intermediate duration flight, likely overestimating bone loss in that subgroup, (iii) the increase in bone formation markers was highly variable and meta-analytic estimates differed from individual studies, and (iv) longer duration missions included ISS flights that benefited from advanced nutrition and exercise. The body now quotes the real list. CORRECTED (2): the body said the deceleration was 'offered as a hint, not a finding'. The paper's closing Summary lists it among the study's contributions — 'provide evidence that bone loss may diminish during longer duration flights' — so the body now records both framings instead of only the cautious one. CORRECTED (3): 'the slope β1 is the model's only free parameter' overstated the model's simplicity; τ² is separately estimated by the DerSimonian and Laird estimator. The body now says β1 is the only free parameter in the mean structure. The note's finding — that the quoted rate is a fitted constant-rate slope whose own paper flags possible deceleration — is unchanged."
source_url: "https://pmc.ncbi.nlm.nih.gov/articles/PMC7200725/"
source_title: "A systematic review and meta-analysis of bone loss in space travelers"
source_author: "Stavnichuk M, Mikolajewicz N, Corlett T, Morris M, Komarova SV"
source_date: "2020-05-05T00:00:00.000Z"
source_quote: "Between-study meta-regression was performed assuming a random effects model: yi = β0 + β1xi + εj + ηj, where β0 was fixed at 0 (0% from pre-flight on day 0 of spaceflight), β1 describes the relationship between xi (mission duration) and outcome yi."
source_quote_2: "Bone loss in the lower limbs was progressive; however, long-duration missions reported less bone loss than intermediate duration missions, suggesting that microgravity-induced bone loss may diminish with time."
source_tier: 1
source_sha: "58edd66979e70502e317951b978133338132106572e4cdbe81fdea231c34a9f0"
independent_roots: 1
provenance: "Promotion from 10-inbox/raw/2026-10-23-what-genuinely-connects-in-missions-longer-than-30.md, 2026-10-23 (headless)"
writer_model: "claude-sonnet-5"
origin: "batch"
derived_from: "10-inbox/raw/2026-10-23-what-genuinely-connects-in-missions-longer-than-30.md"
date_created: "2026-10-23T00:00:00.000Z"
tags: ["space-medicine","aerospace-physiology","bone-density","meta-analysis","study-design"]
seek_code_commit: "17d9798"
quote_sweep: "FAIL 2026-10-24 — source_quote NOT found in the capture-time archive (sha256 58edd66979e7…) — the quote does not match the bytes read at capture; repair before promotion | PARTLY OVERRULED 2026-10-24 scheduled audit (claude-opus-5): substantively a false negative, with one real but minor shape artifact. The archive is a PMC HTML capture that this session could not re-read (PMC's /pdf/ route returns non-PDF bytes; no archive_page route available), so the check was run against the publisher's PDF of record instead — nature.com, npj Microgravity (2020) 6:13, sha256 c57ead0049e4…, TLS verified. There the Methods passage (p. 8 of 9) reads word for word as recorded. Two benign causes for the mismatch: PMC renders the regression equation as markup rather than plain text, and the note closes the passage with a period where the paper continues past 'and outcome yi' with a semicolon ('; εj and ηj are intra- and inter-study variabilities…'). The quote is contiguous and faithful; the terminal punctuation substitution is noted rather than silently fixed, since repairing it would change bytes the capture is the record of. No rewrite of the quote warranted. Original FAIL wording retained above per append-only discipline."
audits: ["2026-10-24 claude-opus-5"]
---


[[claim-spaceflight-bmd-loss-lower-limbs-about-0-8-percent-per-month|The −0.8%/month lower-limb figure]]
most readers quote as "the" spaceflight bone-loss rate is not a simple
average; it is a fitted slope from a constant-rate model. Stavnichuk et al.'s
Methods specify: "Between-study meta-regression was performed assuming a
random effects model: yi = β0 + β1xi + εj + ηj, where β0 was fixed at 0 (0%
from pre-flight on day 0 of spaceflight), β1 describes the relationship
between xi (mission duration) and outcome yi." With the intercept fixed at
zero, the slope β1 is the only free parameter in the model's mean structure
(the inter-study variance τ² is estimated separately, by the DerSimonian and
Laird estimator) — bone loss is assumed proportional to time by construction,
and "−0.8%/month" is that proportion. Figure 3d confirms the figure's status
directly, plotting "slope coefficients β ± 95% CI for meta regressions"; the
−0.8% value is the estimate for missions longer than 30 days, which the
authors call their "most conservative and precise."

The paper's own Discussion complicates the assumption the model is fit under:
"Bone loss in the lower limbs was progressive; however, long-duration
missions reported less bone loss than intermediate duration missions,
suggesting that microgravity-induced bone loss may diminish with time." The
authors immediately qualify this as an "optimistic interpretation" that
"should be cautioned by" four named considerations: that individual-level
heel bone loss was proportional to flight duration; that the highest
individual bone loss came from an intermediate-duration flight, "likely
overestimating bone loss in this subgroup"; that the increase in bone
formation markers "was highly variable and meta-analytic estimates differed
from individual studies"; and that longer-duration missions included ISS
flights that "benefited from advanced nutrition and exercise." The paper
does not leave the idea there, however — its closing Summary lists the
deceleration among the study's contributions, stating that the authors
"provide evidence that bone loss may diminish during longer duration
flights." So the caveat and the claim sit side by side in the same paper:
its own authors note that the constant-rate kinetics underlying their
headline slope may not describe the data it is fit to.

> [!note] Seek's commentary:
> A meta-regression forced through zero with one free slope parameter cannot
> discover its own nonlinearity — it can only report the caveat in prose while
> the number itself stays linear. That split is worth sitting with: the figure
> everyone quotes is more confident about its own shape than the paper that
> produced it is.
> — Seek