Abstract
Purpose :
Using mediation analysis, we investigate the interrelationships between ocriplasmin and vision outcomes to elucidate the role of resolution of vitreomacular adhesion (VMA). Presuming VMA resolution (VMAR) is surrogate marker for treatment success in vitreomacular traction (VMT), we aim to understand how this anatomical surrogate is associated with improvement in functional and patient-reported outcomes.
Methods :
We included 218 participants from the OASIS study (clinicaltrials.gov identifier: NCT01429441) with a diagnosis of VMT defined as the presence of VMA related to decreased visual function. Eyes had VMT only, or VMT with a macular hole. Logistic regression analyses were used to estimate the total treatment effect (TTE) on the binary vision outcomes. Outcomes, assessed at month 24 month, included visual acuity improvement (VA-I): ≥ 2-lines increase in VA; visual function questionnaire improvement (VFQ-I): ≥ 5-points increase in the National Eye Institute VFQ-25 composite score (CS); visual function improvement (VF-I): a composite endpoint defined as either a VA-I ‘OR’ a clinically meaningful improvement in the VFQ-25 CS. In mediation, a quantity of interest is the extent to which the TTE on outcomes is transmitted through a mediating variable vs. other pathways. Using a simulation-based approach, the TTE is decomposed into an indirect (through VMAR) and direct effect. Causal quantities are expressed in mean risk difference.
Results :
Indirect effects for VFQ-I, VA-I and VF-I vary between 5.2 and 11.8%, representing what would have happened to vision outcomes if VMAR status were changed for each participant to the extent that it is affected by ocriplasmin. Direct effects range from 8.3% to 24.1% and capture the effect of ocriplasmin on outcomes while holding VMAR at a value that would have been observed naturally had the participant been assigned to sham. For example, the effect of ocriplasmin on VA-I induced by its effect on VMAR is 11.8%, while the effect on VA-I through any pathway other than VMA resolution is 12.2%. The mediation proportion shows that 48.9% of the TTE on VA-I (23.9%) is mediated by the effect of ocriplasmin on VMAR.
Conclusions :
The effect of ocriplasmin on VA-I was to a large extent the result of VMA resolution, while improvement in patient-reported outcomes was partially explained by this intermediary variable.
This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.