The same pH by two routes: protocol v1.0 with prospectively specified exploratory Amendment 02 v1.0
Prospective exploratory Amendment 02 v1.0 — deposited before its clinical execution. This is a new version of the original pH-project registration (version DOI 10.5281/zenodo.24340426; concept DOI 10.5281/zenodo.24340425). The seven original attachments are retained unchanged. PH-AMENDMENT-02-v1.0.md governs the added analyses. The original v1.0 and E1 numerical results are unchanged. E1 is separately documented as a post-execution deposit at 10.5281/zenodo.24342821. These further analyses were designed after the original and E1 results were known. All included sources have prior programme exposure. Registration fixes the new procedures prospectively; it does not make the cohorts independent confirmation. The exposure ledger states the prior inspections, outcome access and related analyses. No Amendment 02 clinical source has been opened or model fitted during preparation. Clinical execution is authorized only after publication and checksum verification. A: Hsieh and Yu receive the six-parameter pH/PaCO2 model with profile-likelihood intervals, pH bands, supported risk surfaces and the fixed E1 curvature/interaction extension. Eligibility is pH-project specific, without ventilation variables, and preserves Yu's documented PaCO2=467 exclusion. B: Challenge's original 6,770 index stays are compared across prespecified oxygenation strata using frozen PaO2/FiO2 timing and conflict rules; the original gas is never reselected. C: Tuebingen admission base excess at fixed pH is analysed separately in mixed arterial/venous MICU samples, with a same-patients lactate-conditioned model; no PaCO2 reconstruction. D: six measured-gas cohorts are fitted with a shared curved association and projected to cohort-specific core slopes using a dedicated fractional-outcome solver. Paired differences, three supporting heterogeneity tests and the prespecified analysis without Challenge are reported. Similar projected slopes do not establish equal curves. All 101 numerical, source and execution-gate checks passed. Nine full synthetic bootstrap procedures completed with 18,000/18,000 valid draws; a separate orchestration/reporting rehearsal passed. Six fixed scenarios were each run 60 times. The sparse-overlap scenario showed imperfect nominal likelihood-ratio calibration (8/60 rejections at nominal 0.05 under a shared curve); this limitation is disclosed. These tests verify implementation and characterize selected scenarios, not universal power or coverage. All p values are nominal and local intervals pointwise. No causal, equivalence, permissive-hypercapnia safety or pooled clinical-OR claim is enabled. No protected cohort or rate-against-volume model is opened. Patient-level records, identifiers, copyrighted papers and account material are excluded. Source data require independent lawful access. Tiny source-format fixtures in the archive are invented synthetic records. CC BY 4.0; files embargoed until 18 November 2027 under the investigator's continuing instruction. Metadata and provenance hashes are public. Package ZIP SHA-256: d2e9a22ca8e8b7e840a67016bd249146d5f2fca9a2fb791da8262a43c44e5476Local lock SHA-256: afe20409a8348a6b0d28af97a9acb5734577968f3c132c434cf5ac767d7de4e6Amendment SHA-256: 2c57b3bcd83f87cb6c7c22bf02d6246413aef50b66935c12a345c846ebc87f2bPackage manifest SHA-256: 68c90133917a41f7c8dd1d1cba768a38f41e2d7ad1b750c27154d8ed41bc250b
Authors
- Ruggero Amà
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-18
- DOI
- https://doi.org/10.5281/zenodo.24340425
- Primary Topic
- Renal function and acid-base balance
- Type
- article
- Field-Weighted Citation Impact
- 0.00