This study examines risk factors for corneal endothelial cell density (ECD) loss after Baerveldt glaucoma implant insertion.
Study design
This prospective cohort study included 72 patients (72 eyes) with glaucoma. Central corneal ECD, peripheral corneal ECD, anterior chamber flare, tube insertion entry site position, anterior segment OCT tube parameters and clinical parameters were evaluated after Baerveldt glaucoma implant insertion. Primary outcomes were central and peripheral ECD loss at 5 years. Secondary outcomes included prognostic factors for ECD loss: tube insertion entry site position relative to Schwalbe’s line (SL); anterior segment OCT tube parameters; anterior chamber flare and clinical factors including IOP. Tube insertion entry site positions were defined as all of the entry site behind SL (tube insertion position 1), less than 50% of entry site anterior to SL (tube insertion position 2), more than 50% but less than 100% of entry site anterior to SL (tube insertion position 3; TIP3), and all of the entry site in front of SL (tube insertion position 4; TIP4).
Outcomes
Sixty-four patients completed the study. The mean percentages of central and peripheral ECD loss at 5 years were 36.8% and 50.1%, respectively. On univariate analysis at 5 years, a lower central ECD was associated with TIP3. Tube angle-to-anterior iris distance, tube length and TIP3 were associated with lower peripheral ECD at 5 years.
Multiple regression analysis revealed TIP3 to be associated with both lower central ECD (P=0.015) and peripheral ECD (P=0.028). A short tube length was associated with lower peripheral ECD (P=0.016). The fastest rate of cell loss was associated with TIP3 for central ECD and TIP4 for peripheral ECD. Anterior chamber implant insertion was associated with ECD loss greatest close to the tube. Tube insertion in the vicinity of or anterior to SL, and short tube length were associated with significant ECD loss with time.
Limitations
This was a well-designed study without major limitations. While a control nonsurgical eye group to compare corneal endothelial cell loss would have been relevant, this can be extracted from the broader literature on age-related loss for comparison.
Clinical significance
Placement of glaucoma drainage devices is common and resulting corneal decompensation has been reported. While some ECD loss occurs with all tube shunts over time, this study provides guidance on risk factors for ECD loss as well as tube insertion landmarks, position and length to minimize ECD loss.