OBJECTIVE Ninety eyes were retreated to correct myopic regression, with or without corneal haze, after primary photorefractive keratectomy (PRK); astigmatism ranging from -0.50 to -3 diopters (D) was present in 43 eyes. RESULTS The sphere (mean +/- standard deviation) was -2.82 +/- 1.74 D before repeat surgery. At 6 months, the mean was +0.30 D; at 1 year, it was -0.17 D. Patients were divided into two groups: group 1 included 56 eyes with little or no haze (< or = 1). The mean sphere value for this group was -2.13 D before retreatment; at 1 year, it was -0.20 +/- 0.76 D. At 6 months, 84% of sphere values were within +/- 1 D of the intended correction, with a mean haze value of less than 0.5, similar to that before repeat surgery; there was a mean gain of 0.25 Snellen line of best-corrected visual acuity. Astigmatism equal or greater than -0.5 D was present in 34 eyes (mean, -1.17 +/- 0.55 D). At 6 months, astigmatism was reduced to -0.45 +/- 0.48 D. The second group included 34 eyes with corneal haze greater than 1 (mean, 2.7). The mean sphere value was -3.95 D before photorefractive keratectomy and -0.12 +/- 1.48 D 1 year after treatment. Fifty percent of the mean sphere values were within +/- 1 D at 6 months, with a mean haze value of 1. The gain in mean best-corrected visual acuity was 1.3 Snellen lines. The mean astigmatism in nine eyes was -1.75 +/- 0.75 D before photorefractive keratectomy and -0.72 +/- 0.78 D 6 months after treatment. CONCLUSIONS There is a significant difference in the outcome predictability between the two groups. A second photorefractive keratectomy can be done 6 months after the primary treatment in patients with regression with or without trace haze. This group has a High predictability in achieving a good correction, with a low complication rate. When haze is present, retreatments are less predictable with 40% of patients overcorrected. Generally, however, these eyes have a statistically significant decrease in haze (Student's t test; P < 0.01) and an improvement in best-corrected visual acuity.