Finding. In DBA/2J glaucoma, it is not the highest pressure spike that best predicts retinal ganglion cell loss. It is how long pressure sits above 19 mmHg. Cox and Andersen–Gill models, fed monthly IOP and terminal RGC counts, agree. Peak and same-day IOP add little once that cumulative burden is in the regression. A mitochondria-targeted peptide, HDAP2, still cuts the progression hazard after that IOP history is adjusted away. Whatever HDAP2 is doing, it is not only lowering pressure.
Why this paper matters
Experimentalists often log a physiological time series and a single terminal cell count, then argue by endpoint. Lesniewski and MacNeil turn that awkward design into a latent time-to-event problem so you can ask which part of the IOP history does the damage and whether a drug still works after you subtract that history.
The mitochondrial reason this brief exists is HDAP2. A mitochondria-targeted neuroprotective peptide that remains significant after longitudinal IOP adjustment is a pressure-independent claim. This paper does not show how the peptide touches organelles. It shows that its benefit is not an IOP artifact.
What they actually measured
Monthly IOP, terminal RGC counts, Cox models, and a time-dependent Andersen–Gill extension that lets risk depend on current and cumulative pressure. Reconstructed survival curves under real and hypothetical IOP paths. Across disease thresholds, cumulative burden above 19 mmHg dominates. HDAP2 reduces hazard after that adjustment.
There is no Seahorse trace, no mtDNA data, no mitophagy stain. Readers who came for mitochondrial mechanism will be disappointed and should still keep the hazard result.
How to read the score
Low seventies. Useful methods paper, one mitochondrial drug residual. Confidence is medium for the IOP-burden claim in DBA/2J and low for any organelle mechanism. Score 72.
What to do with it
If you model experimental glaucoma or any terminal-endpoint degeneration with a longitudinal stressor, steal the framework. If you follow mitochondria-targeted neuroprotectants, file HDAP2 as pressure-independent in this model. Do not convert 19 mmHg into a human target pressure. The directional implication is that sustained IOP integral beats peaks, and a mitochondrial peptide can still move the hazard after that integral is known.