The study included 10 consecutive eyes from 10 patients treated at Chongqing Nanping Aier Eye Hospital in China between April 2024 and July 2025. All had active polypoidal choroidal vasculopathy, or PCV, and fovea-involving submacular hemorrhage larger than five disc areas. Eleven eyes underwent the procedure, but one was excluded because it did not complete the prespecified 12-month follow-up.

The analyzed patients had a mean age of 65.0 ± 11.7 years, and six were male. Their hemorrhage covered an average of 30.8 ± 22.7 disc areas, with a range of 5.03 to 71.92 disc areas, and extended beyond the macular arcade in eight eyes. Before surgery, median best-corrected visual acuity was 1.25 logMAR, central macular thickness averaged 1,006.0 ± 183.0 micrometers, and every eye had a hemorrhagic pigment epithelial detachment.

All operations used a standard three-port, 25-gauge pars plana vitrectomy. Through a 41-gauge cannula, the surgeon injected tenecteplase into the subretinal space at the thickest part of the hemorrhage while avoiding the foveal center. Controlled fluid movement washed across and liquefied the clot under direct observation.

Seven eyes received one injection of 50 micrograms in 0.05 milliliters. Three eyes with extensive hemorrhage received the drug at two or three extrafoveal sites, with no more than 50 micrograms at any one site and a cumulative dose no greater than 100 micrograms. The surgeon then performed a fluid–air exchange and introduced 0.4 milliliters of pure C3F8 gas.

No active subretinal drainage was performed; after clot liquefaction, the gas tamponade displaced the liquefied hemorrhage away from the fovea. Complete displacement of the subfoveal hemorrhage was observed by postoperative day 1 in all 10 eyes, corresponding to 100% of the cohort and an exact 95% confidence interval of 69.2% to 100%. Patients then maintained face-down positioning for two weeks, for at least 16 hours per day, to support displacement of the remaining blood away from the central macula.

The investigators observed no submacular scar formation during follow-up. The result concerns displacement of the subfoveal clot rather than confirmed disappearance of all hemorrhage, and the study did not test gas tamponade or tenecteplase in isolation. Visual and anatomical recovery continued over the following year.

Median best-corrected visual acuity improved from 1.25 logMAR before surgery to 0.30 logMAR at 12 months, a median improvement of 0.80 logMAR; the Hodges–Lehmann estimate had a bootstrap 95% confidence interval of 0.40 to 1.20. The Wilcoxon signed-rank test gave a Z value of −2.617 and a P value of 0.010. Central macular thickness fell from 1,006.0 ± 183.0 micrometers to 205.9 ± 48.2 micrometers, a mean reduction of 800.1 micrometers with a 95% confidence interval of 654.6 to 945.6 micrometers.

Nine eyes, or 90.0%, had complete resolution of the pigment epithelial detachment by month 2, with an exact 95% confidence interval of 55.5% to 99.7%. Outer-retinal recovery was less complete. The external limiting membrane and ellipsoid zone had been disrupted across a median 96% of the macular scan before surgery, with ranges from 60% to 100%.

At 12 months, these lines were intact in two eyes, while the other eight retained disruption involving a median 57% of the scan. Eyes treated within 10 days of symptom onset generally had less residual disruption, including the two eyes with complete restoration, whereas the three eyes treated more than 10 days after onset had disruption across 76% to 100% of the scan. These observations describe an association in a very small observational cohort and do not establish that earlier surgery prevented structural damage.

Patients received a mean of 1.8 postoperative anti-VEGF injections during the first year, with a median of two and an interquartile range of 1.3 to 2.0. Intravitreal ranibizumab was started two weeks after surgery and administered as needed for persistent or recurrent intraretinal or subretinal fluid, recurrent hemorrhage, or a vision-threatening pigment epithelial detachment involving the foveal center. Because the study had no comparison group, the observed injection count cannot show that the operation reduced treatment requirements relative to anti-VEGF monotherapy.

The authors noted that PCV activity can diminish after rupture of a major polyp, potentially contributing to the limited treatment burden. Likewise, clot evacuation, gas tamponade, surgery and anti-VEGF therapy cannot be separated when considering the high observed rate of pigment epithelial detachment resolution. No intraoperative retinal tear, vitreous rehemorrhage, uncontrolled bleeding, postoperative endophthalmitis, or clinically apparent retinal toxicity attributed to tenecteplase was observed.

For the absence of a drug-related adverse event, the exact 95% confidence interval was 0% to 30.8%, showing that the small sample cannot exclude uncommon toxicity. One 81-year-old patient with a pre-existing nuclear lens opacity experienced progression of the cataract and could not maintain face-down positioning for physical reasons. Preclinical evidence cited by the researchers found no outer-retinal toxicity from a single 50-microgram subretinal tenecteplase dose in pigs, while direct clinical evidence for divided cumulative doses above 50 micrograms remained anecdotal.

The present results therefore provide preliminary safety information rather than a precise estimate of rare complication rates. Two clinical measurements were associated with final vision, but their relationships remain exploratory. Twelve-month logMAR acuity correlated positively with the interval from symptom onset to surgery, with a Spearman correlation of 0.84 and a P value of 0.002, meaning longer delays were associated with worse acuity.

Acuity also correlated with hemorrhage area, at r = 0.87 and P = 0.001. Preoperative central macular thickness had a moderate correlation with final acuity, at r = 0.67 and P = 0.035, although this relationship was not adequately captured by a simple linear model. Correlations with baseline acuity, sex and age were not statistically significant, with all P values above 0.08.

The study also identified one high-leverage observation: the eye treated 60 days after symptom onset had a leverage of 0.95 and Cook’s distance of 21.7 in the univariable time-to-surgery model. Leave-one-out analysis showed that the direction and statistical significance of the two principal rank associations persisted whenever any single patient was removed. Across those iterations, the correlation for treatment delay ranged from 0.78 to 0.90, with all P values no greater than 0.013, while the hemorrhage-area correlation ranged from 0.81 to 0.93, with all P values no greater than 0.008.

Even so, multivariable estimates were unstable. In an exploratory model containing baseline acuity and treatment delay, each additional day was associated with an adjusted increase of 0.024 logMAR in final acuity; when hemorrhage area and delay were modeled together, the delay coefficient weakened and was no longer statistically significant. The strong correlation between hemorrhage size and delay, r = 0.918, made their independent effects difficult to estimate, and no adjustment was made for multiple comparisons.

PCV is characterized by branching vascular networks and terminal polypoidal dilations arising from the inner choroid, and it is increasingly recognized within the pachychoroid disease spectrum. Ruptured polyps can release large amounts of blood beneath the retina, where blood-related effects may include a diffusion barrier, iron toxicity and mechanical shearing from fibrin contraction. Because the cohort was restricted to severe PCV-related hemorrhage requiring surgery, its findings apply to a selected population rather than all patients with PCV.

The study nevertheless documents a detailed 12-month snapshot of a technically demanding approach: rapid displacement of the subfoveal clot, sustained visual improvement in the treated eyes, a high observed rate of pigment epithelial detachment resolution, and no observed tenecteplase-related adverse events. Prospective, multicenter controlled studies are needed to determine comparative effectiveness, define the safety of higher cumulative doses, and separate the contributions of surgery, tenecteplase, gas tamponade and anti-VEGF treatment.

Subject of Research: A 12-month retrospective case series of subretinal tenecteplase-assisted vitrectomy for large submacular hemorrhage caused by polypoidal choroidal vasculopathy.

Article Title: Twelve-month outcomes of subretinal tenecteplase-assisted vitrectomy for large submacular hemorrhage secondary to polypoidal choroidal vasculopathy: a retrospective consecutive case series

Article References: Ai, L., Meng, X., He, S., Yuan, S., Zhang, Y., & Wang, Y. (2026). Twelve-month outcomes of subretinal tenecteplase-assisted vitrectomy for large submacular hemorrhage secondary to polypoidal choroidal vasculopathy: a retrospective consecutive case series. Frontiers in Medicine, 13, Article 1981363. https://doi.org/10.3389/fmed.2026.1981363

Image Credits: AI Generated

DOI: 10.3389/fmed.2026.1981363

Keywords: tenecteplase, submacular hemorrhage, polypoidal choroidal vasculopathy, vitrectomy, gas tamponade, anti-VEGF therapy, visual acuity, retinal imaging, case series, surgical outcomes, Twelve-month, outcomes