Two allogeneic spacers lowered intraocular pressure in crude analyses, but the site effect changed the story
Bio-interventional cyclodialysis is built around a simple surgical premise: create a controlled cleft and keep it open long enough to improve uveoscleral outflow. In this multicenter CREST analysis, surgeons used an acellular scleral allograft spacer to maintain that channel, and some eyes received one spacer while others received two. The question was straightforward and clinically relevant: does the second spacer add meaningful pressure lowering, or does it mainly reflect how and where the procedure was performed?
At 12 months, the unadjusted data favored the two-spacer approach. Eyes treated with two spacers showed a greater reduction in intraocular pressure than eyes treated with one spacer, with an estimated difference of 2.52 mmHg. Patient-level adjustment did not erase the signal right away. Propensity-score matching still showed a 2.36 mmHg advantage, and doubly robust overlap weighting showed a 1.99 mmHg difference. For surgeons looking only at the crude numbers, the second allograft looked like the more potent mechanical support.
Then the analysis turned to the part of the dataset that mattered most: enrolling site.
Site imbalance dominated the comparison between one spacer and two
The clearest asymmetry in the cohort was not baseline pressure or medication burden. It was geography. One center contributed 43.6% of the two-spacer eyes but only 7.1% of the one-spacer eyes, making site the largest measured imbalance in the study. Baseline IOP was close between groups, averaging 17.9 mmHg in the one-spacer arm and 18.7 mmHg in the two-spacer arm. Medication burden was also only modestly different, at 1.19 versus 1.46 agents.
Once models accounted for site, the apparent benefit of the second spacer weakened substantially. The treatment difference fell to 0.86 mmHg and lost conventional statistical significance. Within-site analyses pushed the estimate even lower, to 0.42 mmHg with both analytical strategies. Those site-adjusted results make the surgical environment hard to ignore. Technique, surgeon selection, patient mix, and perioperative management can all influence whether a second spacer appears to add benefit when the underlying biology may not be changing as much as the crude comparison suggests.
A second spacer may still matter mechanically, but this dataset cannot separate device effect from practice pattern.
Why the second spacer may look helpful before adjustment
The biologic rationale for placing more material is easy to understand. A larger or more stable cleft should, in theory, preserve aqueous outflow and resist early closure. In glaucoma surgery, small differences in how long a surgically created pathway stays patent can translate into measurable pressure changes. A second spacer could provide additional scaffolding, reduce local collapse, or improve the durability of the cleft if the first implant is insufficient.
Yet the same logic also explains why observational comparisons can mislead. Surgeons may preferentially use two spacers in eyes they expect to be harder to control, or in centers with technique patterns that differ from those elsewhere in the network. Site-level effects can reflect case selection, operative nuance, postoperative care, and thresholds for supplemental intervention. In a procedure as operator-dependent as cyclodialysis, those factors can mimic a device effect unless the comparison is randomized.
The study design tried to limit bias in several ways. Investigators used unadjusted comparison, baseline-adjusted regression, 1:1 propensity-score matching, overlap weighting, and models that explicitly accounted for enrolling site. IOP values after a secondary IOP-lowering procedure were censored, which helps isolate the effect of the index surgery, but also means the analysis remains focused on a selected postoperative course rather than the full complexity of long-term glaucoma care. Even with these safeguards, the central message stayed the same: the apparent advantage of two spacers shrank once the center effect entered the model.
Medication use and retreatment stayed similar across arms
Despite the pressure differences seen in some analytic models, medication reduction did not separate the groups. The change in medication burden was −0.69 in the one-spacer group and −0.62 in the two-spacer group, a nonsignificant difference. Secondary IOP-lowering procedures also occurred at the same rate in both arms, 2.9% versus 2.9%. Those parallel safety-and-rescue outcomes are important because they suggest the pressure signal did not translate into a clearly different downstream treatment burden over the first year.
In practical terms, that means the second spacer did not demonstrate a broader clinical footprint across the measures reported here. If an additional implant truly produced a meaningful incremental effect, one might expect that to be mirrored by lower medication reliance, fewer rescue procedures, or both. Instead, the benefit was concentrated in some analytical views of IOP alone and became fragile when site was controlled.
What this means for glaucoma surgery research
The trial registration, multicenter structure, and prospective CREST framework give the analysis real clinical value, especially in a field where procedure-specific evidence is often sparse. The findings also land in a familiar glaucoma reality: pressure control depends not only on the device or surgical concept, but on how that procedure is deployed across different hands and centers. A spacer that seems superior in pooled data may simply be aligned with one center’s technique or patient selection pattern.
Because this was an observational study rather than a randomized comparison, the results can support hypothesis generation but not causal attribution. The most defensible conclusion is narrow. Two-spacer treatment was associated with greater 12-month IOP reduction in crude and patient-level adjusted analyses, yet that association was substantially attenuated after accounting for where the surgery was performed. The study therefore leaves unresolved whether the second spacer independently improves outflow or whether the apparent advantage is largely a function of site-specific practice.
A randomized head-to-head trial would be the cleanest way to answer the question. Until then, the incremental value of adding a second allogeneic reinforcement remains plausible, but unproven.