Figure 2.
Three-month best spectacle-corrected visual acuity (BSCVA) vs baseline BSCVA for patients receiving voriconazole and natamycin, with Fusarium species (A) and non-Fusarium species (B) as the causative organism. The curve is a nonparametric locally weighted scatterplot smoothing regression fit, with the shaded bands indicating ±1 estimated SD. Patients who experienced perforation or corneal transplantation prior to the 3-month observation may have excellent visual acuity despite this adverse outcome and were assigned a low-vision score of 1.7 logMAR (or the 3-week BSCVA, whichever was worse). Observations over 1.5 logMAR were jittered for plotting.