Patients and Methods
Study Population
The Miami Veterans Affairs Medical Center (VAMC) Institutional Review Board reviewed and approved this retrospective study, which was performed in accordance with the principles of the Declaration of Helsinki. Inclusion criteria were patients who had combined cataract extraction and either trabeculectomy or GDD surgery at the Miami VAMC from March 2008 to April 2011. Patients with postoperative best-corrected visual acuity (BCVA) of 20/200 or worse were excluded from our study, due to inability to measure refraction adequately.
Data Collection
All data were obtained through a retrospective chart review and entered into a standard computerized database for subsequent analysis. The data collected included: patient demographics (age, sex), pretreatment refractive findings [visual acuity, refractive error, keratometry (K) values, AL], type of glaucoma (open angle, closed angle, or neovascular), preoperative IOP, surgical complications, postoperative IOP, and postoperative visual and refractive outcomes. K values were obtained using either Pentacam rotating Scheimpflug imaging device (Oculus Optikgeräte GmbH, Wetzlar, Germany), Auto-Keratometer (Topcon Medical Systems, Oakland, NJ), or IOLMaster partial coherence interferometry biometer (Carl Zeiss Meditec AG, Berlin, Germany). K values were measured with multiple instruments in some eyes. AL values were obtained using the IOLMaster partial coherence interferometry biometer or contact A-scan with the Eye Cubed ultrasound system (Ellex, Minneapolis, MN).
IOL Model Selection
The surgeon selected the IOL based on results of the K and AL measurements with the devices discussed above. The surgeons' discretion was used to decide which device measurements were most appropriate for each given patient.
Surgical Techniques
Baerveldt (Abbott Laboratories Inc., IL) and Ahmed (New World Medical Inc., CA) GDD implantation was performed as previously described. All Baerveldt GDD surgeries were performed with the implant placed in the superotemporal or inferonasal quadrants and all implants were placed beneath the respective rectus muscles. All Baerveldt GDD were ligated with 7–0 Vicryl suture. Trabeculectomy with mitomycin-C (MMC) was performed as previously described. In all cases the concentration of MMC was 0.4 mg/mL and length of time for MMC application was ≤4 minutes, left to the discretion of the surgeon. Cataract surgery was performed by a temporal clear cornea approach, with standard phacoemulsification techniques.
Control Group
Forty patients who underwent cataract extraction only with a single piece acrylic lens implantation were selected as an age-matched control. Surgeries were performed during the same time period at the same institution. Indications for surgery were visually significant cataract affecting activities of daily living. All surgeries were performed by a temporal clear cornea approach with standard phacoemulsification techniques.
Outcome Measures
The primary outcome measure evaluated was whether or not spherical equivalent (SE) of −1.00 to +0.50 D was achieved at 3 to 6 months postoperatively. Because of the retrospective nature of this study, we did not have the preoperative keratometric values and therefore, refractive predictions for all patients. Thus, the target range chosen was used as a surrogate measure of a successful refractive outcome. Secondary outcomes measures included: the change in IOP, amount of cylinder induced by combined surgery, and other factors potentially affecting refractive outcome (age, type of glaucoma, type of glaucoma surgery, type of lens implant used, and preoperative BCVA).
Applied cylinder was calculated by a modified approach outlined by Cravy. Preoperative keratometry was performed with automated devices as previously discussed. Manual refraction was used to estimate K values postoperatively. On the basis of these substitute values, we were able to calculate the applied cylinder which is the theoretical vector necessary to produce the change in astigmatism from the preoperative to the postoperative state.
Statistical Analysis
All statistical analyses were performed using SPSS (SPSS Inc., Chicago, IL) statistical package. Continuous variables were summarized with the use of mean and SD; categorical variables with the use of percentages. Logistic regression analysis was used to evaluate which factors predicted whether a SE of −1.00 to +0.50 D was achieved. Linear regression analysis was used to evaluate factors that predicted change in cylinder. The Student t test and χ analyses were used to compare refractive outcomes between the combined surgery versus cataract surgery alone group.