PURPOSE. To investigate changes in the variability of white-on-white perimetry in patients with glaucoma who are undergoing surgical or medical reduction of intraocular pressure (IOP). METHODS. This retrospective study included 67 eyes of 50 patients who underwent trabeculectomy (25 eyes) or medical IOP reduction (25 eyes) about midway through the follow-up; 17 fellow eyes of the surgical group were the control group. Their pre- and postinterventional visual fields were analyzed (full-threshold, 24-2 program, Humphrey Field Analyzer; Carl Zeiss Meditec, Inc., Oberkochen, Germany) to evaluate changes in short-term fluctuation (SF) and test-retest variability (TRV). The longest follow-up periods possible were considered, provided that glaucoma was stable (as confirmed by intraocular pressure, optic nerve appearance and visual field). For each patient and in each period, TRV was defined as the mean of the square roots of the mean variances in each of the 54 locations. RESULTS. Trabeculectomy reduced mean IOP (57 ± 4.7 mm Hg; P < 0.0001), IOP fluctuations (-2.9 ± 4.4 mm Hg; P = 0.003), and eye drop use (-1.3 ± 1.4; P = 0.0001). Pre- and postsurgical SF was 3.0 ± 1.1 and 2.7 ± 1.0 dB, respectively (P = 0.34), and pre- and postsurgical TRV was 2.90 ± 0.97 and 2.53 ± 0.86 dB, with a decrease of 0.37 dB (P = 0.0006). In the medical group, mean IOP decreased (-4.0 ± 2.0 mm Hg; P < 0.0001) as did IOP fluctuation (-3.7 ± 4.1 mm Hg; P = 0.0008), but eye drop use increased by 1.0 ± 0.6 (P < 0.0001). TRV (2.58 ± 0.53 vs. 2.72 ± 0.67 dB; P = 0.37) and SF (2.3 ± 0.8 vs. 2.3 ± 0.8 dB; P = 0.89) remained stable during the study. There was no change in perimetric variability throughout the study in the control group. The individual reductions in TRV were higher in the surgical group than in the medical (P = 0.004) or control (P = 0.015) groups. No differences were found between the control and medical groups (P = 0.55). CONCLUSIONS. After trabeculectomy, there was a small but statistically significant reduction in long-term perimetric variability compared with the medical and control groups. There were no significant changes in short-term variability in any of the groups during the study.