Perioperative visual loss (POVL) is a devastating complication, with no known treatment or prevention, most commonly due to ischemic optic neuropathy (ION), or retinal arterial occlusion (RAO). The highest inci- dence is associated with two of the most commonly performed inpatient surgical procedures in the United States, spinal fusion, and cardiac surgery. The medical and economic burden of sudden, unexpected visual loss is considerable; the emotional toll is immeasurable. It is imperative to understand the risk factors for POVL in order to develop means to prevent these blinding complications. There also exists here a unique opportunity to study the natural history and risk factors for perioperative and spontaneous eye diseases that are among the most common causes of sudden adult vision loss. A most pressing problem in POVL is predicting the high risk individual. Using the National Inpatient Sample, the pro- ject will examine postulated risk factors for POVL, and construct a predictive risk model based on these factors (Aim 1). The risk model will be externally validated using a large commercial insurance database (Aim 2). The central hypothesis is that the presence of specific patient factors, and specific surgical factors, can predict the risk of developing POVL. Aim 1 will study risk factors and develop a predictive model for perioperative visual loss (POVL) in spinal fusion and cardiac surgery using the NIS, with the following specific hypotheses based on the literature and preliminary data: 1) Perioperative RAO and ION are associated with pre-existing abnormalities of the retina and/or optic nerve; 2) Carotid stenosis and specific forms of stroke increase risk of POVL; 3) Specific surgical factors are associ- ated with POVL, including valvular heart surgery, and extent of fusions in spine surgery; 4) Female sex signifi- cantly decreases risk of POVL. Aim 2 will externally validate the predictive model from Aim 1 using Optum. We will test the hypothesis that combinations of patient specific and surgical factors are synergistic in defining risk level, such that the ma- jority of POVL incidents are predictable a priori. The research fits the specific goals of this RFA for 2o analysis of existing data to generate new hypotheses. Innovation includes: 1) the new concept relating pre-existing ophthalmic disease to POVL, 2) using diagnostic and procedural data, and statistical modeling to build a risk profile index, and 3) cross-validation with another database. In addition to direct clinical relevance of both the risk factor determination and the risk model, the translational significance of both is the high likelihood of resulting changes in the care of surgical patients, and development of prevention strategies. The underlying research significance is that the studies will set the groundwork to enable vertical advancement of the field into mechanisms of injury.

R21EY027447

2017-09-30

2020-08-31