Acute retinal pigment epithelium detachments after photocoagulation.

PURPOSE: To characterize the morphology of patterned scanning laser (PASCAL) panretinal photocoagulation.

METHODS

In this prospective cohort study, patients with proliferative diabetic retinopathy or severe nonproliferative diabetic retinopathy with high-risk characteristics, who were treated with PASCAL panretinal photocoagulation as part of their indicated clinical course, were serially imaged with spectral domain optical coherence tomography. Thirty eyes of 25 patients were studied from 1 hour to 21 weeks after laser treatment.

RESULTS

Over a quarter (26.1%) of all treatment spots were imaged by spectral domain optical coherence tomography 1 hour after PASCAL panretinal photocoagulation. At 1 hour (±30 minutes) after PASCAL treatment, spectral domain optical coherence tomography demonstrated retinal pigment epithelium detachment in 23 of 27 eyes (85.2%) and in 36.1% of all imaged laser spots. Detachments occurred preferentially at the photocoagulation edges in 48.4% of pigment epithelium detachments (PEDs). Linear regression analysis revealed that average laser power (Pearson's r = 0.671, P < 0.001) and average laser energy (Pearson's r = 0.588, P = 0.001) were significantly associated with PEDs observed 1 hour after treatment. Pigment epithelium detachments occurred at a rate of 9.2% ± 4.9% at an average power of 0 mW to 400 mW, 37.8% ± 9.5% at 401 mW to 800 mW, 42.1% ± 5.6% at 801 mW to 1,200 mW, and 53.6% ± 5.7% at >1,200 mW. By a 1-week follow-up, no PEDs were observed, and the retinal pigment epithelium appeared morphologically similar to its preoperative structure by 3 weeks. Patient characteristics (study eye, sex, race, diagnosis, age, preoperative blood glucose, hemoglobin A1C, duration of diabetes, and body mass index) and other PASCAL parameters (number of laser applications, spot size, pulse duration, and average laser fluence) were not significantly associated with PEDs.

CONCLUSION

Retinal pigment epithelium detachment occurs 1 hour after PASCAL treatment over a wide range of laser settings. Laser power and energy are positively correlated with the occurrence of PEDs, which are no longer observed by 1-week follow-up. Future studies might examine various acute posttreatment time points and directly compare the morphology of PASCAL burns with that of longer pulse-duration laser modalities.