Background: Keratoconus is an ectatic (weakening) disease of the cornea, which is the clear surface at the front of the eye. Approximately 10% to 15% of patients diagnosed with keratoconus require corneal transplantation. This may be full-thickness (penetrating) or partial-thickness (lamellar). Objectives: To compare visual outcomes after deep anterior lamellar keratoplasty (DALK) and penetrating keratoplasty for keratoconus, and to compare additional outcomes relating to factors which may contribute to poor visual outcomes (e.g. astigmatism, graft rejection and failure). Search methods: We searched a number of electronic databases including CENTRAL, PubMed and EMBASE without using any date or language restrictions. We last searched the electronic databases on 31 October 2013. We also handsearched the proceedings of several international ophthalmic conferences. Selection criteria: We included all randomised controlled trials (RCTs) comparing the outcomes of DALK and penetrating keratoplasty in the treatment of keratoconus. Data collection and analysis: Two authors assessed trial quality and extracted data independently. For dichotomous data (graft failure, rejection, achievement of functional vision) results were expressed as odds ratios (ORs) and 95% confidence intervals (CIs). For continuous data (postoperative best corrected visual acuity (BCVA), uncorrected visual acuity (UCVA), keratometric astigmatism and spherical equivalent) results were expressed as mean differences (MDs) and 95% CIs. Main results: We identified two completed studies, with a total of 111 participants (n = 30 and n = 81), both conducted in Iran, that met our inclusion criteria. Participants had moderate to severe keratoconus pre-operatively and were randomly allocated to receive either DALK or penetrating keratoplasty. Only one eye of each participant was treated as part of the trials. The smaller study had 12 month follow-up data for all participants. For the larger study, four DALK surgeries had to be abandoned due to technical failure and visual and refractive outcomes were not measured in these participants. Follow-up length for the remaining 77 participants ranged from 6.8 to 36.4 months, with all 77 followed for at least three months post-suture removal. Details of the randomisation procedure were unavailable for the smaller study and so sensitivity analyses were conducted to determine if the results from this study had affected the overall results of the review. Neither of the included studies reported a difference between groups on any of the measures of post-graft visual achievement, keratometric astigmatism or spherical equivalent. A single case of graft failure in a penetrating keratoplasty was reported. No postoperative graft failures were reported in the DALK group of either study. Instances of graft rejection were reported in both groups, in both studies. The majority of these cases were successfully treated with steroids. The data, which related to all cases in each study - given that the four cases that did not go ahead as planned had already technically failed without presence of rejection - showed that rejection was less likely to occur in DALK (odds ratio (OR): 0.33, 95% confidence interval (CI) 0.14 to 0.81, GRADE rating: moderate). Results of the sensitivity analysis indicated that inclusion of the Razmju 2011 study did not bias the results with regards to rejection episodes. While sensitivity analysis showed altered results with regards to failure rates, the data available from the Javadi 2010 study alone had a very wide 95% CI, suggesting an imprecise estimate. Therefore, even after removal of the Razmju 2011 data, it is still difficult to draw conclusions regarding superiority of one technique over another with regards to graft failure. DALK was unable to be completed as planned in four cases and in a further three cases, complications during dissection required further intervention. Other adverse events, of varying severity, were reported in both intervention groups with similar frequency. For both types of surgery, these included postoperative astigmatism, steroid induced ocular hypertension and persistent epithelial defects. In recipients of DALK, one participant had interface neovascularisation (a proliferation of blood vessels where the host and donor cornea come together) and one had wrinkling of Descemet's membrane, the basement membrane separating the corneal stroma from the corneal endothelium. In the penetrating keratoplasty groups, one participant required graft resuturing and one had an atonic pupil, a condition in which the pupil dilates and is non-reactive. Overall, the quality of the evidence was rated as very low to moderate, with methodological limitations, incomplete data analysis and imprecision of findings, as well as high risk of bias in several areas for both studies. Authors' conclusions: We found no evidence to support a difference in outcomes with regards to BCVA at three months post-graft or at any of the other time points analysed (GRADE rating: very low). We also found no evidence of a difference in outcomes with regards to graft survival, final UCVA or keratometric outcomes. We found some evidence that rejection is more likely to occur following penetrating keratoplasty than DALK (GRADE rating: moderate). The small number of studies included in the review and methodological issues relating to the two, mean that the overall quality of the evidence in this review is low. There is currently insufficient evidence to determine which technique may offer better overall outcomes - final visual acuity and time to attain this, keratometric stabilisation, risk of rejection or failure, or both, and risk of other adverse events - for patients with keratoconus. Large randomised trials comparing the outcomes of penetrating keratoplasty and DALK in the treatment of keratoconus are needed.