Unruptured Paraclinoid Aneurysm Treatment Effects on Visual Function: Systematic Review and Meta-analysis.

Abstract

Postoperative visual outcomes following repair of unruptured paraclinoid aneurysms (UPAs) are not well defined. We aim to investigate the influence of treatment modality on visual function. A systematic literature analysis using the Ovid Medline and EMBASE databases was performed, encompassing English language studies (published between 1996 and 2016) reporting treatment outcomes for UPAs. Rates of visual morbidity (new, permanent postoperative deficit, worsening preoperative deficit); angiographic (occlusion, recurrence, retreatment) and clinical outcomes (death, disability, post-treatment subarachnoid hemorrhage) were recorded. Random effects meta-analysis was performed. Twenty-eight studies reported visual outcomes, with data for 1013 endovascular and 691 microsurgical patients. In patients with normal vision undergoing elective repair of UPAs, rates of postoperative visual morbidity were higher following microsurgical (10.8%; 95% confidence interval [CI] 8.5-13.7) than endovascular (2.0%; 95% CI 1.2-3.2) interventions, P < 0.001. In those presenting with preoperative visual impairment, surgery was associated with a modest advantage in visual recovery compared with endovascular therapies (65.2% vs. 48.9%, P < 0.03). There were no differences in visual morbidity following treatment with any of the endovascular modalities. Meta-analysis of comparative studies suggested a trend toward poor visual (ES = 0.42; 95% CI 0.08-2.09) and clinical outcomes (ES = 0.57; 95% CI 0.07-4.44) following microsurgery and a trend toward angiographic recurrence (ES = 2.52; 95% CI 0.80-7.90) and retreatment (ES = 1.62; 95% CI 0.46-5.67) after endovascular interventions. In patients with normal vision undergoing repairs for UPAs, there is a positive correlation between visual outcomes and endovascular treatments. When visual compromise is present, surgery provided modest advantage in visual recovery. However, definitive conclusions were not possible due to data heterogeneity.