- Design
- systematic review and meta-analysis with random-effects pooling, Poisson rate ratios and meta-regression; GRADE certainty low to very low
- Population
- 33 studies, 5,235 glaucoma-naive adult eyes after Descemet membrane or Descemet stripping automated endothelial keratoplasty
- Primary outcome
- de novo glaucoma and ocular hypertension after surgery
- Effect
- glaucoma 3% vs 11%, ocular hypertension 9% vs 34% (P ≤ 0.0006); rate ratio for ocular hypertension 0.31 (95% CI 0.14-0.71)
Post-keratoplasty glaucoma threatens both the optic nerve and the graft, and surgeons have lacked procedure-specific numbers to counsel with. This review pooled 33 studies and 5,235 glaucoma-naive adult eyes undergoing Descemet membrane endothelial keratoplasty or Descemet stripping automated endothelial keratoplasty, with counts re-audited against the source papers.
Event rates were substantial in both arms and consistently lower after the membrane procedure: de novo glaucoma 3% against 11%, ocular hypertension 9% against 34%, and medication started or escalated 6% against 20% (all P ≤ 0.0006). Glaucoma surgery was 1.0% against 2.5% and did not separate statistically. The automated stripping cohorts were followed almost twice as long (68 against 35 months), but converting to rates per eye-year still favoured the membrane procedure for ocular hypertension (rate ratio 0.31, 95% CI 0.14-0.71) and, less securely, for glaucoma (0.45, 0.19-1.07).
Two findings complicate attributing this to technique. Restricting the glaucoma definition to optic nerve or visual field criteria — rather than counting pressure-based diagnoses — narrowed the gap to 5.0% against 8.1% with no significant difference. And greater steroid exposure was independently associated with ocular hypertension, which is a plausible route by which the two procedures differ without the technique itself being responsible. Certainty was low to very low throughout, and every randomised trial used ultrathin or microthin grafts.
The part that should change behaviour is not procedure selection but surveillance. A third of eyes after the automated stripping procedure developed ocular hypertension, in patients who had no glaucoma beforehand and are therefore not on anyone's glaucoma pathway. Pressure monitoring belongs in the routine postoperative schedule for both procedures, with the steroid taper reviewed as an active variable rather than a default.
- Include scheduled pressure monitoring in postoperative follow-up after any endothelial keratoplasty, not only in eyes with known glaucoma
- Review the steroid taper actively — greater steroid exposure was independently associated with ocular hypertension
- Counsel patients before surgery that raised pressure is common and may need treatment
- Do not select the procedure on this evidence alone; the difference narrowed to nothing under stricter glaucoma definitions
- Record optic nerve and visual field status where pressure rises, rather than diagnosing glaucoma on pressure alone
Why it matters
These are eyes with no glaucoma history, so nobody is monitoring the pressure that a third of them will develop a problem with.
Don't overread it
The difference between procedures depended heavily on how glaucoma was defined and on steroid exposure, so it should not by itself drive procedure choice.
The statistics, in plain English
Pooled single-arm incidences from 33 mostly observational studies are sensitive to how each study defined glaucoma, which is exactly what the sensitivity analysis shows: restricting to optic nerve or field criteria moved 3% against 11% to 5.0% against 8.1% and erased the significance. Converting to rates per eye-year is the right correction for the unequal follow-up, and the ocular hypertension rate ratio of 0.31 (0.14-0.71) survives it while the glaucoma ratio of 0.45 (0.19-1.07) crosses 1.0. GRADE certainty of low to very low means these point estimates should be expected to move.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for ophthalmology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free