Table 3.
Mean differences in the main outcomes and adverse events in non-randomized studies of interventions after corneal collagen cross-linking (epi-on vs. epi-off)
| Study or subgroup | Epi-off CXL | Epi-on CXL | ||||
|---|---|---|---|---|---|---|
| Mean ± SD | n (%) | n | Mean ± SD | n (%) | n | |
| Kmax (D) | ||||||
| Vinciguerra et al. 2019a [47] | − 0.1 ± 4.6 | – | 20 | − 0.4 ± 4.4 | – | 20 |
| Jouve et al. 2017b [52] | − 1.1 ± 4.2 | – | 40 | 0.2 ± 5.2 | – | 40 |
| Henriquez et al. 2017a [53] | − 0.9 ± 5.2 | – | 25 | 0.1 ± 5.3 | – | 36 |
| Eraslan et al. 2017a [55] | − 1.4 ± 2.6 | – | 18 | − 0.6 ± 3.1 | – | 18 |
| Vinciguerra et al. 2016a [57] | − 1.0 ± 1.5 | – | 20 | − 0.3 ± 1.9 | – | 20 |
| UDVA (logMAR) | ||||||
| Rossi et al. 2018 [51] | − 0.1 ± 0.2 | – | 10 | − 0.2 ± 0.2* | – | 10* |
| − 0.1 ± 0.2Þ | – | 10Þ | ||||
| Henriquez et al. 2017 [53] | − 0.1 ± 0.4 | – | 25 | − 0.1 ± 0.2 | – | 36 |
| Eraslan et al. 2017 [55] | − 0.0 ± 0.1 | – | 18 | − 0.1 ± 0.2 | – | 18 |
| CDVA (logMAR) | ||||||
| Vinciguerra et al. 2019 [47] | − 0.1 ± 0.1 | – | 20 | − 0.1 ± 0.2 | – | 20 |
| Rossi et al. 2018 [51] | − 0.1 ± 0.1 | – | 10 | − 0.1 ± 0.0* | – | 10* |
| − 0.1 ± 0.1 Þ | – | 10Þ | ||||
| Jouve et al. 2017 [52] | − 0.1 ± 0.1 | – | 40 | − 0.1 ± 0.2 | – | 40 |
| Henriquez et al. 2017 [53] | − 0.1 ± 0.1 | – | 25 | − 0.1 ± 0.1 | – | 36 |
| Eraslan et al. 2017 [55] | − 0.1 ± 0.1 | – | 18 | − 0.1 ± 0.1 | – | 18 |
| Vinciguerra et al. 2016 [57] | − 0.0 ± 0.1 | – | 20 | − 0.1 ± 0.1 | – | 20 |
| Higher Order Aberrations (µm) | ||||||
| Vinciguerra et al. 2019 [47] | − 0.1 ± 0.3 | – | 20 | − 0.7 ± 0.2 | – | 20 |
| Vinciguerra et al. 2016 [57] | − 0.0 ± 0.2 | – | 20 | − 0.3 ± 0.8 | – | 20 |
| Comatic Aberrations | ||||||
| Vinciguerra et al. 2019 [47] | − 0.3 ± 1.0 | – | 20 | 1.7 ± 0.3 | – | 20 |
| Rossi et al. 2018 [51] | − 0.4 ± 1.2 | – | 10 | − 0.7 ± 1.4* | – | 10* |
| − 0.7 ± 1.3 Þ | – | 10Þ | ||||
| Eraslan et al. 2017 [55] | − 0.2 ± 0.3 | – | 18 | − 0.1 ± 0.5 | – | 18 |
| Vinciguerra et al. 2016 [57] | − 0.2 ± 0.3 | – | 20 | − 1.2 ± 1.6 | – | 20 |
| Demarcation Line (depth in µm at 1 month) | ||||||
| Jouve et al. 2017 [52] | 291.0 ± 61.0 | – | 40 | 216.0 ± 49.0 | – | 40 |
| Eraslan et al. 2017 [55] | 272.3 ± 28.6 | – | 18 | 136.6 ± 17.9 | – | 18 |
| Central corneal thickness (µm) | ||||||
| Rossi et al. 2018 [51] | − 2.9 ± 18.9 | – | 10 | − 1.6 ± 35.5* | – | 10* |
| − 4.7 ± 30.2 Þ | – | 10Þ | ||||
| Corneal Thinnest Point (µm) | ||||||
| Vinciguerra et al. 2019 [47] | − 57.0 ± 103.0 | – | 20 | 5.0 ± 38.0 | – | 20 |
| Henriquez et al. 2017 [53] | − 12.5 ± 40.2 | – | 25 | 1.5 ± 51.5 | – | 36 |
| Eraslan et al. 2017 [55] | − 11.3 ± 14.9 | – | 18 | − 8.8 ± 14.8 | – | 18 |
| Vinciguerra et al. 2016 [57] | − 41.1 ± 35.3 | – | 20 | 1.0 ± 7.2 | – | 20 |
| ECD (cells/mm2) | ||||||
| Rossi et al. 2018 [51] | − 32.7 ± 99.3 | – | 10 | − 46.1 ± 197.9* | – | 10* |
| − 27.0 ± 62.5Þ | – | 10Þ | ||||
| KC progression | ||||||
| Jouve et al. 2017 [52] | – | 3 (7.5%) | 40 | – | 8 (20.0%) | 40 |
| Henriquez et al. 2017 [53] | – | 3 (12.0%) | 25 | – | 2 (5.6%) | 36 |
| Lost ≥ 2 lines CDVA | ||||||
| Eraslan et al. 2017 [55] | – | 0 (0.0%) | 18 | – | 1 (5.6%) | 18 |
| Persistent stromal haze | ||||||
| Henriquez et al. 2017 [53] | – | 1 (0.1%) | 25 | – | 0 (0.0%) | 36 |
| Eraslan et al. 2017 [55] | – | 5 (27.8%) | 18 | – | 0 (0.0%) | 18 |
| Sterile infiltrates | ||||||
| Henriquez et al. 2017 [53] | – | 1 (0.1%) | 25 | – | 0 (0.0%) | 36 |
CXL = corneal collagen cross-linking; CDVA = corrected distance visual acuity; D = diopter; ECD endothelial cell density; KC = keratoconus; Kmax = maximum keratometry; logMAR = logarithm of the minimum angle of resolution; n (%)= absolute frequency (relative frequency); SD = standard deviation; UDVA = uncorrected distance visual acuity
*Transepithelial CXL treatment study group
ÞIontophoresis CXL treatment study group
aScheimpflug imaging analysis (Oculus Pentacam GmbH, Wetzlar, Germany)
bScanning slit technique (Orbscan IIz; Bausch & LombSurgical, Rochester, NY)