Table 25: GRADE Quality of Evidence for CXL of Keratoconus*.
| Outcome | Design | Quality | Consistency | Directness Appropriate Range of Patients | Other† Modifying Factors | Overall Quality |
|---|---|---|---|---|---|---|
| Safety | Case reports, observational longitudinal pre-post study, comparative pre-post studies (treated to untreated fellow-eye) and 1 RCT | Moderate | Low complication rates reported | Inception cohorts specified with appropriate range | Moderate | |
| Corneal Topography | Observational longitudinal pre-post study, comparative pre-post studies (treated to untreated fellow-eye) and 1 RCT | Moderate | Clinically relevant and statistically significant improvements in treated eye and regression in untreated fellow-eyes | Inception cohorts specified with appropriate range | Moderate | |
| Visual Acuity | Observational longitudinal pre-post study, comparative pre-post studies (treated to untreated fellow-eye) and 1 RCT | Moderate | Clinically relevant and statistically significant improvements, although less predictable than topography | Inception cohorts specified with appropriate range | Moderate | |
| Durability | Observational longitudinal pre-post study | Low | Limited data beyond 2-year follow-up | Inception cohorts specified with appropriate range | Low |
*
RCT indicates randomized controlled trial
†
Studies involving comparisons with untreated fellow-eyes demonstrated significant improvements in the treated eye compared to worsening of the underlying fellow-eye in 1-year follow-up. The treated eye was also the worst eye. Treated patients had progressive keratoconus and given the disease’s natural history would not be expected to recover or improve without treatment. Outcomes were based on standardized validated measurements. Estimates of natural variability or normal ranges exist. Improvements in topography were large, clinically relevant, and statistically significant.