Results

Five eligible published studies were identified, three that compared bare sclera resection without and with mitomycin C application,15 23 24 one that compared bare sclera resection with conjunctival autograft placement,14 and one comparing the three techniques.11

An initial objective of the study was to perform meta-analysis calculations comparing bare sclera resection with mitomycin C use and conjunctival autograft placement; however, only one eligible study was identified.11 Also, we originally intended to evaluate the role of β irradiation and three studies evaluating different treatment groups were retrieved; one comparing bare sclera resection with mitomycin C and β irradiation applications,25 one comparing bare sclera resection with or without β irradiation,26 and one comparing bare sclera resection with β irradiation and conjunctival autograft placement.27Since no two studies compared similar treatment groups no analysis attempt was performed.

Table 1 provides a description of the population characteristics of studies included, follow up periods, and pterygium recurrence definitions. Table 2 gives a general outline of the surgical technique and the postoperative treatment utilised in the studies included. Table3 shows the individual quality scoring of the studies included, along with that of three studies excluded for non-randomisation and scores below 0.5.

Ten studies were totally and one partially excluded from this analysis. Studies by Hayasaka et al,4 Riordan-Evaet al,28 Schrage et al,29 and a part of Singh et al 23 partially because of no patient randomisation; Mahar et al 5 because of a quality score below 0.5; Güler et al,30 and Sebban and Hirst13 because of combined treatment results from primary and recurrent pterygia; Dowlut and LaFlamme,27and Chayakul25 because of the concomitant use of β irradiation; Singh et al 3 because of results included in a subsequent study²³; and Vaniscotte et al 31because of insufficient information.

Table 4 shows the 2 × 2 tables data constructed based on information provided by the study. Homogeneity testing determined a χ² of 1.23 for the bare sclera versus conjunctival autograft and 3.06 for bare sclera with or without mitomycin C use. For a χ² distribution with 1 and 3 degrees of freedom respectively, and at an alpha level of 0.05, the obtained χ² determined that it was acceptable to combine information from the studies using the Mantel–Haenszel method.

The pooled odds ratio for postoperative pterygium recurrence in patients who only had bare sclera resection was 6.1 (95% confidence intervals, 1.82 to 18.75) compared with those who had conjunctival autograft placement, and 25.4 (9.02 to 66.69) compared with those who received mitomycin C. Individual and pooled odds ratios and confidence intervals are depicted in Figures 1 and2.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Individual study and pooled odds ratios and 95% confidence intervals calculated from data provided by original studies comparing bare sclera resection with postoperative mitomycin C versus placebo. Singh et al 23 OR: 109.33 (12.18–980.34), Frucht-Pery et al 24 OR: 25.38 (4.68–136.4), Cano-Parra et al 15 OR: 18.45 (2.24–150.27), Chen et al 11 OR: 12.5 (0.81–192.92). (OR = individual study odds ratio, 95% confidence intervals for individual studies are in parenthesis.) Cumulative odds ratio: 25.4 (95% cumulative confidence intervals 9.02 to 66.69)

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Individual study and pooled odds ratios and 95% confidence intervals calculated from data provided by original studies comparing bare sclera resection versus conjunctival autograft placement. Lewallen et al 14 OR: 3.11 (0.17–55.21), Chen et al 11 OR:11.67 (2.14–64.04). (OR = individual study odds ratio, 95% confidence intervals for individual studies are in parenthesis.) Cumulative odds ratio: 6.1 (95% cumulative confidence intervals 1.82–18.75).

Discussion

This study provides evidence of the inherent risk for pterygium recurrence of bare sclera resection alone when compared with other currently available surgical techniques. Our results show with a 95% confidence interval that the odds for a pterygium recurrence is at least nine times higher if mitomycin C is not applied intraoperatively or postoperatively. Despite the fact that when comparing bare sclera resection with conjunctival autograft placement only two eligible studies were retrieved, we provide evidence with a 95% confidence interval with the fact that there is an increased risk of pterygium recurrence with bare sclera resection of at least 1.8 times greater, a figure that could be in excess of 18. These results should strongly discourage surgeons and clinical researchers in the use of bare sclera resection, if recurrences are to be prevented, bearing in mind that the Declaration of Helsinki states that in clinical trials no suboptimal treatment can be offered to a control group.

Conjunctival autograft placement and bare sclera resection with mitomycin C use have both advantages and limitations. Conjunctival autograft placement following pterygium resection offers the theoretical advantages of reconstructing the architecture of the corneoscleral limbus and transplanting limbal stem cells which may facilitate corneal epithelial healing.32 It is also a time consuming procedure. Mitomycin C is an antibiotic-antineoplastic agent that selectively inhibits DNA, cellular RNA and protein synthesis³ and can have serious complications following administration.33 Identification of an ideal dose and concentration of mitomycin C is beyond the scope of this study.

Our study has several limitations inherent in meta-analysis. A low number of studies was retrieved by our search as a result of several factors. Medline searches may yield only two thirds of relevant published papers.34 35 Among the identified studies, only a few randomised controlled clinical trials provided the quality standards for eligibility. This underscores the need for the development and publication of adequately planned and conducted clinical research in this area. Publication bias (the lack of publication of trials with negative results) is also a limitation to any meta-analysis. Unidentified eligible studies, as well as future publications should increase the power of this survey. As for any meta-analysis evaluating a surgical technique, the study populations, the individual surgeon expertise, as well as the postoperative treatment are different in every study. The test of homogeneity performed provides confidence in establishing that, statistically speaking, the samples can be combined because they could come from a same theoretical population from which each sample from each study is a representation. An evaluation of the individual population risk factors and recurrence rates was out of the scope of this meta-analysis. Also, according to the authors’ description of the surgical technique and postoperative management, all studies followed standard surgical and postoperative procedures (Table 2), although some differences were noticeable. Mitomycin C concentration, administration, and dosing were different between the studies evaluated. Follow up periods and definitions of pterygium recurrence are also slightly different among studies evaluated. Recurrences are usually evident 1–2 months after surgery,35 a period covered by all eligible studies. Included for analysis are studies by Singh et al,23 Frucht-Pery et al 24reporting a total of four recurrent cases. Bearing in mind all the study’s limitations, we consider that our study had eligibility criteria adequate for meta-analysis. We believe that the differences between the eligible studies should only minimally affect the general results and conclusions of this study.

At present we are unable, through meta-analysis, to determine if there is a significant difference in the recurrence rate between bare sclera resection with mitomycin C and conjunctival autograft placement. In this regard, Singh et al 23 reported that to detect 2% and 5% recurrence rate differences, adequately designed clinical trials require sample sizes of 736 and 211 patients respectively. Only one study with a small sample size was detected in our search.11 New research in this area should focus on answering, through randomised controlled clinical trials with adequate sample sizes, if significant differences exist with regard to safety and efficacy issues of techniques which, like bare sclera resection with mitomycin C and conjunctival autograft placement, have a lower risk for postoperative pterygium recurrence.