The initial management of penetrating ocular injuries is a major sight-threatening problem for both civilian and military medicine. A novel device (Eye-Aid) temporarily tamponades leakage from such injuries while being easy to remove upon arrival to specialized ophthalmologic care. Eye-Aid consists of a protective eye shield with an adhesive backing that connects to a portable canister containing rapidly deployable thermoresponsive foam. The aim of this study was to compare the use of the novel Eye-Aid device to control in a new live swine ocular injury model. Bilateral penetrating ocular injuries were created on 14 male Yorkshire swine in a standardized manner using a 16-gauge needle device to puncture the central cornea and cause a full-thickness wound. Researchers randomized eye intervention side, with the contralateral eye used as paired control. Two minutes after the injury, the eye shield components of the Eye-Aid system, which has a sticky pad for attachment to the skin and a luer-lock for foam deployment, were placed bilaterally. Eight minutes after the injury, foam was deployed for the intervention eye according to the device instructions for use. For the control eye, no additional procedures were performed. Six hours post-injury, end A-scan and intraocular pressure (IOP) were measured. Primary study outcome was change in axial length of the globe. Secondary outcomes were as follows: (1) Presence of full anterior chamber collapse, defined as a lack of measurable anterior lens capsule-reflex (ALC-reflex) on A-scan and (2) change in IOP. Outcomes were analyzed as paired intra-animal data, with intervention and control data for each animal. A paired t-test was used to analyze the difference in axial length change and IOP change between treatment groups, whereas a conditional logistic regression was used to analyze dichotomous ALC-reflex outcome and estimate the odds ratio associated with the Eye-Aid device. A significant difference (P < .0001) in mean change in axial length between intervention (-210 μm) and control (-1,202 μm) groups was found. There was a significant difference in ALC-reflex presence, with 79% of eyes having an ALC-reflex in the intervention group, compared to 14% in the control (P = .008). IOP remained higher in the intervention group, with a mean change of -1.5 mmHg for the intervention group compared to -4.0 mmHg in the control (P = .0001). This study describes the first development of an in vivo large animal ocular injury model that realistically approximates the emergent time course and pathophysiology of patients with full-thickness corneal open globe injuries. It also gives the first description of using thermoreversible hydrogel foam for such injuries. Eye-Aid was found to be significantly better than control for treatment of such injuries, based on measurements of both structure and pressure. Assuming that the absence of an ALC-reflex demonstrates complete anterior chamber collapse, the Eye-Aid group demonstrated a 79% eye "save" rate compared to only 14% in the control group, as described earlier. This results in a Number Needed to Treat of 3 for this finding. Eye-Aid additionally demonstrated several characteristics that would be beneficial in a device targeted for emergent deployment by non-ophthalmologists.
Read full abstract