Outcome comparison between wavefront guided and wavefront optimized photorefractive keratectomy: A systematic review and meta analysis

Author: Khaled M Hamam

Geographical coverage: Not reported

Sector: Biomedical

Sub-sector: Treatment

Equity focus: None

Study population: Patients undergoing photorefractive keratectomy.

Review type: Effectiveness review

Quantitative synthesis method: Meta-analysis

Qualitative synthesis method: Not applicable

Background: Photorefractive keratectomy (PRK) eye surgery is widely used for patients at risk for corneal ectasia to maintain an aspheric corneal shape. Wavefront‑guided (WFG) ablation profile was designed to reduce pre‑existing higher‑order aberrations (HOA).

Objectives: To compare the visual outcomes and corneal aberrations between WFO and WFG PRK and evaluate their effect on myopia for patient‑perceived quality of vision to determine whether one treatment profile leads to more optimal vision than the other.

Main findings:

In summary, authors found no significant difference between WFG and WFO in UDVA, CDVA, mean manifest sphere and spherical equivalent, but the total of MD of manifest cylinder after three and 12 months’ follow‑up was significantly lower with the WFG ablation profile.

A total of eight papers were included for the qualitative analysis and six articles for quantitative analysis. None of the eight studies had the selection, attrition or reporting bias. Two studies had performance bias as the participants and personnel were not fully blinded, and one study had a different type of bias.

Based on the pooled results, authors observed no significance between uncorrected distance visual acuity (UDVA) and corrected distance visual acuity (CDVA) between both groups who underwent WFG and WFO PPR after three months’ follow-up (MD = ‑0.03; 95% CI: [‑0.06, 0.00]; p = 0.07), (MD = ‑0.02; 95% CI: [‑0.04, 0.01]; p = 0.22), respectively. Authors also observed no significant difference between mean manifest cylinder after three and 12 months’ follow-up, but the total MD for mean manifest cylinder difference was significantly lower with the WFG treatment method (MD = ‑0.12, (95% CI: [0.23:‑0.01], p = 0.03). This shows a slight advantage of the WFG over the WFO method. Authors found similarities between the visual performance and excellent refractive outcomes in both WFO and WFG PRK. There were no significant statistical differences between the two approaches. On further comparison, there was a slight advantage of the WFG over the WFO method.

Authors note the following limitations. It is important to note that the review was limited by studies reporting UDVA at different times following the procedure; some studies reported UDVA at three months, others at six, while others at 12, and some studies had strict enough follow-up to report it on all those time periods. It also included a limited number of trials.

Authors note the need of larger sample sizes, and more studies are required to show equivalency or superiority of WFG compared to WFO effectively. Further and more extensive studies are needed with stricter follow-up and more diverse follow‑up checklists to determine the best guidelines for myopia PRK treatment.

Methodology:

Studies were included under the following criteria: (1) population are patients undergoing photorefractive keratectomy; (2) intervention and comparator are Wavefront‑guided versus Wave‑front‑optimised photorefractive keratectomy operations; (3) study design: included randomised clinical trials with no restrictions for languages. Animal trials, conference abstracts, non‑randomised trials and studies without relevant outcomes were excluded.

Authors searched PubMed, Scopus, Web of Science and Cochrane CENTRAL at March 2020, and updated the search in September 2020 for relevant clinical trials comparing optimised and guided photorefractive keratectomy. The references of included trials and relevant reviews were screened to ensure high‑quality searching.

Two authors independently screened the titles and abstracts of all expected included studies followed by full‑text screening and then manually searching the references of the finally included papers eligible to meta‑analysis. Two independent authors performed the extraction step, and any disclosure among authors was resolved through discussion and referred to the study senior. To assess methodological quality of trials, authors used Cochrane’s risk-of-bias tool. Authors did assess publication bias because of the small number of included studies.

Authors used the mean difference (MD) to perform analysis of continuous outcomes, and risk ratio (RR) to analyse dichotomous outcomes. The analysis was performed using Review Manager Software under a fixed‑effect model in all outcome data. Statistical heterogeneity between studies was assessed by observation of the graphs and measured by Chi‑square test and I² test for the degree of the heterogeneity. Subgroup analysis was conducted according to postoperative follow‑up of UDVA to stratify the surgical efficacy on UDVA.

Applicability/external validity: Authors indicate that the included studies used different follow-up durations, which may limit its precise applicability regarding reported results. Authors also indicate that they were only able to include a limited number of trials.

Geographic focus: No information provided on geographic origin of included trials. No discussion of how findings might vary by geographic context.

Summary of quality assessment:

The approaches taken to identifying, including and critically appraising studies were relatively robust, with all key tasks being undertaken by at least two authors. However, there was no attempt to include unpublished material and it is unclear what the time cut-off for the search was (and its rationale). While the analysis of the data was also relatively robust, there was no consideration of unit of analysis errors (which may have been required, if included trials involved both eyes for included individuals). In addition, there was no attempt to analyse the reported results by the quality of included studies – although we appreciate that this would have been challenging, given the small number of studies included. For these reasons, we have medium confidence in the findings of this review.