BACKGROUND: Little is known about the optimal management of impending paradoxical embolism (IPDE), a biatrial thromboembolus caught in transit across a patent foramen ovale. Our aim was to review observational studies on this subject to identify prognostic factors and to compare mortality and systemic embolism between treatments. METHODS: Systematic literature searches in Medline, Embase, and Cochrane Library identified 154 studies (174 patients). The primary end point was 30-day mortality. The secondary end point was systemic embolism during treatment. RESULTS: Thirty-day mortality was 18.4%. On univariate analysis, age (64+/-13.9 vs 56.7+/-16.5; P = .01), coma (12.9% vs 2.2%; P = .02), and systemic embolism (71.9% vs 51.4%; P = .048) at presentation were significantly increased among nonsurvivors. Surgical thromboembolectomy had lower mortality than other treatment groups (10.6%; P = .04). In multivariable models, no prognostic factor was a significant independent predictor of mortality. Surgically treated patients had nonsignificantly reduced mortality (odds ratio [OR], 0.65 [0.24-1.72]; P = .65) and thrombolysis-treated patients had increased mortality (OR, 1.62 [0.43-5.97]; P = .47). However, systemic embolism during treatment and combined mortality and systemic embolism was decreased in the surgery group (OR, 0.13 [0.03-0.67]; P = .02 and OR, 0.26 [0.11-0.60]; P = .001). CONCLUSIONS: This review attempts to help guide what to do in IPDE, despite severe limitations of the methods. Surgical thromboembolectomy showed a nonsignificant trend toward improved survival, significantly reduced systemic embolism, and composite of mortality and systemic embolism, compared with anticoagulation alone. Thrombolysis, on the other hand, had the opposite effect, although not significantly.