Purpose To evaluate the effectiveness and predictability of arcuate keratotomy (AK) for post-keratoplasty astigmatism and to present the complications and rate of repeat procedures.

Methods Sixteen eyes from 14 patients were included. Paired 70–80° arc length AKs centred on the steep axis were carried out 0.5 mm within the graft–host junction. The depth of the AKs was set at approximately 80–90% of the depth of the cornea, based on a topographic pachymeter at the incision location. The outcome measures included preoperative and postoperative topographic astigmatism, uncorrected and corrected visual acuity, surgical complications and repeat procedures.

Results In 12 of the 16 eyes (75%), at least one additional surgical procedure was required to obtain the desired result: suturing for overcorrection or wound gape in six eyes (38%), lengthening of the incisions for undercorrection in four eyes (25%) and additional AKs for marked astigmatic axis displacement in three eyes (19%). The mean preoperative astigmatism was 10.45±3.82 dioptres (D); the postoperative astigmatism at the last visit was 2.99±1.14 D (in a mean follow-up of 17.6±5.55 months). The efficacy index was 0.83 and the safety index was 1.68.

Conclusions In treatment of post-keratoplasty astigmatism, AK does not have a good predictability. Additional procedures such as lengthening of the AK incisions for undercorrection or using compression sutures for overcorrection with significantly gaping wounds are frequently required to improve the final outcome.