Objective: To compare phacoemulsification with extracapsular cataract surgery in patients with diabetes and to identify determinants of postoperative visual acuity.
Design: Prospective, randomized, paired-eye trial.
Participants: Forty-six patients with diabetes and bilateral cataract.
Intervention: Patients were allocated to phacoemulsification surgery with silicone intraocular lens to one randomly determined eye, and extracapsular cataract surgery with 7-mm polymethylmethacrylate intraocular lens to the other.
Main outcome measures: Logarithm of minimum angle of resolution visual acuity (logMAR VA), incidence of clinically significant macular edema (CSME), retinopathy progression, indices of anterior segment inflammation, and incidence of capsulotomy.
Results: Compared with eyes undergoing phacoemulsification, eyes managed with extracapsular surgery had more anterior chamber cells (P = 0.0004) and flare (P = 0.007) 1 week after surgery and a higher incidence of posterior synechiae (P = 0.04) and intraocular lens deposits (P < 0.0005) in the first postoperative year. The need for posterior capsulotomy was greater in eyes undergoing extracapsular surgery (16 of 46 vs. 5 of 46, P = 0.01). No difference in incidence of postoperative CSME, progression of retinopathy, or development of high-risk proliferative retinopathy was identified between techniques (P = 1.0, 0.8, and 0.2). Median 1-year logMAR VA was worse in eyes undergoing extracapsular surgery (0.08 vs. 0.06, P = 0.02), especially in those with retinopathy (0.14 vs. 0.08, respectively; P = 0.01). The presence or absence of CSME at the time of surgery was the most significant determinant of 1-year logMAR VA in regression models for both extracapsular (P = 0.0004, R2 = 0.45) and phacoemulsification groups (P < 0.00005, R2 = 0.46).
Conclusions: Phacoemulsification is associated with better postoperative VA, less postoperative inflammation, and less need for capsulotomy than extracapsular cataract surgery in patients with diabetes. However, with both techniques, the principal determinant of postoperative VA appears to be the presence or absence of CSME at the time of surgery. Early intervention, reducing the risk that unrecognized CSME is present at the time of surgery, may be more critical to outcome than choice of surgical technique.