Purpose
To assess the relationship between graft thickness and best-corrected visual acuity (BCVA) after Descemet stripping endothelial keratoplasty (DSEK).
Design
Systematic review and meta-analysis.
Methods
PubMed, EMBASE, Web of Science, and conference abstracts were searched for studies published up to October 2015 with standard systematic review methodology. Eligibility criteria included studies evaluating graft thickness in primary DSEK and visual outcomes. There were no restrictions to study design, study population, or language. Correlation coefficients were pooled using random-effects models.
Results
Of 480 articles and conference abstracts, 31 met inclusion criteria (2214 eyes) after full-text review. Twenty-three studies assessed correlations between BCVA and graft thickness, and 8 studies used different statistical methods. All associations were reported dimensionless. Studies generally had small sample sizes and were heterogeneous, especially with respect to data and analysis quality ( P = .02). Most studies did not measure BCVA in a standardized manner. The pooled correlation coefficient for graft thickness vs BCVA was 0.20 (95% CI, 0.14–0.26) for 17 studies without data concerns; this did not include 7 studies (815 eyes) that used different statistical methods and did not find significant associations.
Conclusions
There is insufficient evidence that graft thickness is clinically important with respect to BCVA after DSEK, with meta-analysis suggesting a weak relationship. Although well-designed longitudinal studies with standardized measurements of visual acuity and graft thickness are necessary to better characterize this relationship, current evidence suggests that graft thickness is not important for surgical planning.
Endothelial keratoplasty has become the standard method to restore endothelial function and improve vision in diseases such as Fuchs endothelial corneal dystrophy and pseudophakic corneal edema. However, it is well known that many patients do not reach best-corrected visual acuity (BCVA) of 20/20 after Descemet stripping (automated) endothelial keratoplasty (DSEK or DSAEK, both of which will simply be referred to as DSEK in this article). Controversy exists as to whether thicker DSEK grafts are associated with worse visual outcomes compared to thinner DSEK grafts. Based on the hypothesis that thicker grafts impair BCVA after DSEK, some corneal surgeons request thinner grafts from eye banks or prepare “ultrathin” corneal grafts, which, in some cases, can increase the cost of care or result in donor wastage.
This systematic review and meta-analysis assembled all individual studies that measured DSEK graft thickness and postoperative visual acuity to determine if a relationship exists. Current evidence related to the effect of graft thickness on visual outcomes is summarized to improve our understanding of visual limitations after DSEK and to facilitate evidence-based surgical decisions.
Methods
Search Methods for Identifying Studies
This systematic review and meta-analysis was conducted according to the PRISMA-P guidelines. The search strategy was built in conjunction with a reference librarian. The main search keywords were “Descemet stripping endothelial keratoplasty,” “Descemet stripping automated endothelial keratoplasty,” and “endothelial keratoplasty” in combination with “thickness,” “thick,” or “thin” (detailed search strategy is provided in the Appendix , Supplemental Material available at AJO.com ). PubMed, MEDLINE, Embase, Web of Science, Scopus, and Cochrane Database of Systematic Reviews were systematically reviewed to include all results up to October 6, 2015. Gray literature was searched for conference abstracts presented at the Association for Research and Vision in Ophthalmology between 2002 and 2015 ( www.iovs.org ) that were not published elsewhere. Authors of conference abstracts were contacted to provide additional information. Ongoing trials and negative unpublished clinical trials were searched for with ClinicalTrials.gov , projectreporter.nih.gov , and WHO International Clinical Trials Registry. The reference lists of identified articles were explored. We did not have any restrictions for publication status, language, or difficulty with respect to accessing journals. The search strategy and the completeness of the articles were verified by 2 experts in the field (W.M.B. and S.V.P.). The systematic review protocol was registered at the Prospective Register for Systematic Reviews (PROSPERO; http://www.crd.york.ac.uk/prospero ; registration number: CRD42015017310).
Eligibility Criteria for Considering Studies for This Review
Inclusion criteria for studies were: (1) design: observational studies (more than 2 participants) or clinical trials; (2) population: adults without ocular comorbidities other than the indication for primary DSEK or cataract; (3) exposure: graft thickness or corneal thickness in DSEK; (4) primary outcome: postoperative BCVA. Articles were not restricted according to method of measurement of thickness or BCVA, or by methods to minimize bias in the individual studies.
Study Selection
Two independent reviewers (K.W. and W.M.B.) screened all article titles and abstracts according to the eligibility criteria to identify studies for full-text review. Disagreements were resolved by consensus of all 3 authors. EndNote 4 and X7 (Thomson Reuters) were used to manage the identified articles and eligibility status.
Data Collection
The following study design information was recorded from all eligible full-text studies by 2 independent investigators (K.W. and S.V.P.) in a pilot-tested data extraction form: study characteristics and design, study population with inclusion and exclusion criteria, indication for DSEK, proportion of patients receiving concomitant cataract surgery, number of eyes without comorbidities and complications, and quality parameters ( Table 1 ) including industry sponsorship. Quantitative data that were recorded included graft thickness (or corneal thickness) and BCVA, timing and methods of measurement of these variables, length of follow-up, metric to assess association (eg, correlation coefficients and significance levels), and the number of eyes for association analyses ( Table 2 ).
| Study Characteristics | Quality Assessment | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| First Author | Year | Journal | Country | Study Design | Indication for DSEK | Triple DSEK | Standardized Measurement | Comorbidities Excluded | Data Concern | |
| 1. BCVA | 2. Thickness | |||||||||
| Price | 2006 | Ophthalmology | USA | R | 91% FECD, 10% PCE | 11% | N | Y | Y | |
| Chen | 2008 | Cornea | USA | P | 81% FECD, 16% PCE, 3% other | 51% | N | Y | Y | A, B |
| Di Pascuale | 2009 | Am J Ophthalmol | USA | R | 71% FECD, 29% PCE | 42% | N | Y | Y | |
| Nieuwendaal | 2009 | Cornea | NL | P | 100% FECD | 0% | Y | Y | Y | |
| Pogorelov | 2009 | Br J Ophthalmol | GER | P | 100% FECD | 20% | N | Y | Y | C |
| Terry | 2009 | Ophthalmology | USA | P | 83% FECD, 12% PCE | 55% | N | Y | Y | |
| Ahmed | 2010 | Am J Ophthalmol | USA | P | 100% FECD | – | Y | Y | Y | |
| Neff | 2011 | Cornea | USA | R | 91% FECD, 6% PCE, 3% other | 56% | N | Y | Y | |
| Rice | 2011 | Cornea | UK | R | 57% FECD, 29% PCE, 10% both, 4% ACE | 4% | N | Y | N | |
| Van Cleynenbreugel | 2011 | Cornea | NL | P | 54% FECD, 43% PCE, 3% after vitrectomy | 3% | N | Y | Y | |
| Villarrubia | 2011 | Arch Soc Esp Ophthal | ESP | R | FECD, or PCE | – | N | N | N | |
| Heinzelmann a | 2012 | ARVO Abstract | GER | R | 85% FECD, 15% PCE | – | N | Y | Y | |
| Seery | 2012 | Am J Ophthalmol | USA | P | 100% FECD | 83% | Y | Y | Y | |
| Shinton | 2012 | Br J Ophthalmol | UK | R | 53% FECD, 20% PCE, 16% both, 8% failed DSEK, 4% PPMD | 23% | N | Y | Y | |
| Terry | 2012 | Ophthalmology | USA | R | 100% FECD | – | N | Y | Y | |
| Van der Meulen | 2012 | Cornea | NL | P | 100% FECD | 0% | Y | Y | Y | |
| Daoud | 2013 | Am J Ophthalmol | USA | R | 70% FECD | – | N | Y | Y | |
| Dickman | 2013 | JAMA Ophthalmol | NL | R | 67% FECD, 30% PCE | 14% | N | Y | Y | B |
| Hindman | 2013 | Cornea | USA | P | 80 % FECD, 20% PCE | 0% | N | Y | Y | |
| Phillips | 2013 | Cornea | USA | P | 95%, 5% other | – | N | Y | Y | |
| Woodward | 2013 | Cornea | USA | R | 72% FECD, 16% PCE, 12% regrafts | – | N | Y | Y | |
| Acar | 2014 | Int J Ophthalmol | TUR | R | 100% PCE | 0% | N | Y | Y | A |
| Ivarsen | 2014 | Br J Ophthalmol | DK | R | 96% FECD, 4% PCE | – | N | Y | Y | |
| Maier | 2014 | Ophthalmologe | GER | R | 92% FECD, 8% PCE | 40% | N | Y | Y | A |
| Ragunathan | 2014 | Cell Tissue Bank | DK | R | 95% FECD, 5% PCE | 28% | N | Y | Y | |
| Wisse | 2014 | Cornea | USA | R | 92% FECD, 1% PCE, 3% failed grafts | 0% | N | Y | N | |
| Ang | 2015 | Br J Ophthalmol | SG | R | 52% PCE, 48% FECD | 78% | N | Y | Y | |
| Menucci | 2015 | BMC Ophthalmol | IT | P | 100% FECD | 0% | N | Y | Y | |
| Unterlauft | 2015 | J Glaucoma | GER | R | 35% FECD, 65% PCE | – | N | N | Y | |
| Davidson b | 2015 | ARVO Abstract | USA | R | – | 18% | N | Y | Y | B |
| Van Laere c | 2015 | ARVO Abstract | USA | R | – | – | N | Y | Y | |
a Heinzelmann S, et al, Invest Opthalmol Vis Sci 2012;53: ARVO E-Abstract 43. Additional data for table acquired by contacting the authors (personal communication with Sonja Heinzelmann, MD on June 11, 2015).
b Davidson R, et al, Invest Opthalmol Vis Sci 2015;56: ARVO E-Abstract 1572. Additional data for table acquired by contacting the authors (personal communication with Michael J. Taravella, MD on November 27, 2015).
c Van Laere L, et al, Invest Opthalmol Vis Sci 2015;53: ARVO E-Abstract 1563. Additional data for table were requested from authors but not received.
| Study | Graft (Cornea) | Best-Corrected Visual Acuity | Association | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
| First Author | Thickness [μm] | Instrument | Timing | Preoperative (logMAR) | Postoperative (logMAR) | Follow-up (Months) | Eyes for Correlation | Correlation Coefficient | P Value | Method |
| Price | (1) 690 ± 77 (surgeon) (2) 610 ± 62 (precut) | US | Post | – | (1) ≤0.4 (2) ≤0.4 | 6 | 155 | – | .25 | ANOVA, regression |
| Chen | 660 ± 52 (cornea) | US | Post | 0.48 | 0.18 | 6 | 74 | 0.34 | .001 | Pearson |
| Di Pascuale | 147.8 ± 44 | AS-OCT | Post | 3.0–0.4 a | 2.0–0.0 | 3–9 | 11 | 0.42 | .13 | Spearman |
| Nieuwendaal | Median, 128.3 (55–181) | AS-OCT | Post | 0.6 (1.3–0.4) a | 0.18 (0.48–0.0) | 15.5 (6–32) | 13 | 0.251 | .69 | Linear regression |
| Pogorelov | 100 ± 38 | AS-OCT | Post | 0.84 ± 0.63 | 0.37 ± 0.18 | 6 | 10 | 0.59 | .05 | Pearson |
| Terry | 169 ± 36 (88–257) | US | Pre | 0.62 a , b | 0.29 | 12 | 61 | 0.362 | .06 | Pearson |
| Ahmed | 156 ± 51 | Confocal | Post | 0.44 ± 0.21 | 0.16 ± 0.16 | 12 | 29 | 0.06 | .67 c | Pearson/Spearman, GEE for significance |
| Neff | (1) 77–131 (2) 138–182 | AS-OCT | Post | (1) 0.50 (2) 0.40 | (1) 0.10–0.0 (2) 0.30–0.0 | 13 | 33 | – | .01 | Nonparametric comparison of 2 groups |
| Rice | 140 ± 52 (76–273) | AS-OCT | Post | 1.20 ± 0.80 a | 0.40 ± 0.20 | 20 ± 6 | 17 (11) | 0.139 | .56 | Wilcoxon signed rank test |
| Van Cleynenbreugel | 175 (94–304) | US | Intra | 0.58 ± 0.20 (1.0–0.3) | 0.29 ± 0.16 (0.7–0.0) | 6 | 34 | 0.164 | .13 | Pearson |
| Villarrubia | 165 ± 44 (88–263) | AS-OCT | Early post | 0.7 ± 0.7 (1.00–0.0) a | 0.3 ± 0.5 (1.3–0.1) | 15 (1–36) | 51 (22) | 0.40 | n. s. | Pearson/Spearman |
| Heinzelmann d | 259 (178–401) | AS-OCT | Post | – | 0.4 ± 0.3 | 8 | 27 | – | .29 | ANOVA, linear regression |
| Seery | – | Confocal | Post | – | – | 24 | 27 | 0.24 | .20 | Pearson, GEE for significance |
| Shinton | 142 (99–172) | AS-OCT | Post | 0.71 (0.48–1.05) a | 0.24 (0.24–0.54) | 13 | 39 | 0.010 | .95 | Spearman |
| Terry | 163 ± 29 | (1) US (2) AS-OCT | (1) Pre(2) Post | 0.38 ± 0.23 | 0.14 ± 0.11 | 6 | 418 | 0.23 | <.001 | Pearson, Deciles, ANOVA |
| Van der Meulen | 609 ± 56 (494–722) (Cornea) | AS-OCT | Post | 0.6 ± 0.18 (0.0–0.8) | 0.33 ± 0.19 (0.0–0.7) | 8 (6–64) | 34 | – | n. s. | Pearson/Spearman |
| Daoud | (1) 92 ± 12 (47–99) (2) 126 ± 14 (100–150) (3) 159 ± 11 151–196) | US | Intra | (1) 0.99 (2) 0.88 (3) 0.86 | (1) 0.64 (2) 0.56 (3) 0.52 | 6 | (1) 67 (2) 316 (3) 77 | – | .48 | – |
| Dickman | 97 ± 25 | AS-OCT | Post | 0.5 b | 0.2 | 6 | 34 | 0.35 | .02 | Pearson |
| Hindman | – | AS-OCT | Post | 0.64 ± 0.34 | 0.22 ± 0.03 | 12 | 20 | – | n. s. | Spearman |
| Phillips | 145 ± 27 (60–215) | US | Pre | 0.37 | 0.10 | 12 | 65 | 0.047 | .71 | Pearson |
| Woodward | (1) 199 ± 45 (106–303) (2) 165 ± 53 (88–335) | (1) US (2) AS-OCT | (1) Pre (2) Post | – | 0.16 | 27 (4–52) | 32 | (1) 0.11 (2) 0.26 | (1) .57 (2) .16 | Pearson |
| Acar | (1) 73–146 (2) 167–195 (3) 220–317 | AS-OCT | Post | (1) 1.30 ± 1.52 (2) 1.52 ± 1.4 (3) 1.7 ± 1.0 | (1) 0.2 ± 1.2 (2) 0.3 ± 1.1 (3) 0.6 ± 2.0 | 12 | (1) 15 (2) 9 (3) 13 | 0.96 | <.001 | Pearson |
| Ivarsen | 183 ± 49 | AS-OCT | Post | – | 0.30 ± 0.13 | 6–36 | 55 | – | n. s. c | Pearson |
| Maier | (1) 169 ± 64 (2) 153 ± 64 | (1) US (2) AS-OCT | (1) Intra (2) Post | 0.8 (2–0.3) | 0.3 (2.0–0.0) | 23 ± 11 | 65 | (1) 0.241 (2) 0.273 | (1) .1 (2) .03 | Spearman |
| Ragunathan | (1) 236 ± 7 (surgeon) (2) 182 ± 6 (precut) | Scheimpflug | Post | – | (1) 0.25 ± 0.02 (2) 0.24 ± 0.02 | 12 | (1) 55 (2) 49 | (1) 0.015 (2) 0.03 | (1) .91 (2) .84 | Pearson |
| Wisse | (1) 257 ± 47 (350- microkeratome) (2) 222 ± 33 (400- microkeratome) | US | Intra | 0.57 ± 0.36 | 0.17 ± 0.25 | 12 | 37 | 0.18 | .27 | Pearson |
| Ang | 156 (110–180) | US | Pre | 1.3 ± 0.7 (PCE) 0.66 ± 0.6 (FECD) | 0.27 ± 0.1 (PCE) 0.22 ± 0.9 (FECD; at 12 months) | 24 | 128 | 0.228 | .171 c | Spearman, regression |
| Menucci | 132 ± 33 | AS-OCT, confocal | Pre | 0.78 ± 0.35 | 0.13 ± 0.09 | 12 | 39 | 0.28 | >.05 | Structural equation modeling |
| Unterlauft | 621 ± 73 (cornea) | Optical pachymeter | Post | 0.98 ± 0.48 | 0.36 ± 0.17 | 12 | 16 | 0.35 | .0005 | Spearman |
| Davidson e | 56 (28–158) | AS-OCT | Post | 0.19 ± 0.16 | 6 | 25 | 0.13 | n. s. | Pearson | |
| Van Laere f | (1) ≤120 (2) 121–150 (3) 151–180 (4) ≥180 | US | Pre | (1) 0.37 (2) 0.36 (3) 0.42 (4) 0.33 | (1) 0.15 (2) 0.19 (3) 0.20 (4) 0.19 | 3–6 | 64 | – | n. s. | – |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
