Development and Evaluation of the Prognostic Nomogram to Predict Refractive Error in Patients With Primary Angle-Closure Glaucoma Who Underwent Cataract Surgery Combined With Goniosynechialysis.

Objective

To evaluate the accuracy of different intraocular lens (IOL) power calculation formulas and develop prognostic nomograms to predict the risk of postoperative refractive error in primary angle-closure glaucoma (PACG) patients.

Methods

A total of 111 eyes with PACG underwent goniosynechialysis combined with phacoemulsification and IOL implantation were included. SRK/T, Barrett II, Hoffer Q, and Kane formulas were used to predict postoperative refraction. Prediction error (PE) and absolute predictive error (APE) produced by the four formulas were calculated and compared. An APE >0.50 D was defined as the event. Binary logistic regression analysis and prognostic nomogram models were conducted to investigate reliable predictors associated with postoperative refraction.

Results

The Kane (-0.06 D) and Barrett II (-0.07 D) formulas had mean prediction error close to zero (p = 0.44, p = 0.41, respectively). The Hoffer Q and SRK/T produced significantly myopic outcomes (p = 0.003, p = 0.013, respectively). The percentage of eyes within ± 0.5 D was 49.5% (55/111), 44.1% (49/111), 43.2% (48/111), and 49.5% (54/111), for the Kane, Barrett II, Hoffer Q, and SRK/T formula, respectively. Nomogram showed that AL had the greatest impact on the refractive outcomes, indicating a shorter preoperative AL is associated with a greater probability of refractive error event. The area under the receiver operator curve (AUC) of the nomogram for the Kane, Barrett II, Hoffer Q, and SRK/T was 0.690, 0.701, 0.708, and 0.676, respectively.

Conclusions

The Kane and Barrett II formulas were comparable, and they outperformed Hoffer Q and SRK/T in the total eyes with PACG receiving cataract surgery combined with goniosynechialysis. The developed nomogram models can effectively predict the occurrence of postoperative refractive error events.