Achieved Anticoagulation vs Prosthesis Selection for Mitral Mechanical Valve Replacement A Population-Based Outcome Study

Background: Thromboembolic events (TEs) are frequent after mechanical mitral valve replacement (MVR), but their association to anticoagulation quality is unclear and has never been studied in a population-based setting with patients who have a complete anticoagulation record. Methods: We compiled a complete record of all residents of Olmsted County, MN, who underwent mechanical MVR between 1981 and 2004, for all TE, bleeding episodes, and international normalized ratios (INRs) measured from prosthesis implantation. Results: In the 112 residents (mean [+/- SD] age, 57 +/- 16 years; 60% female residents) who underwent mechanical MVR, 19,647 INR samples were obtained. While INR averaged 3.02 +/- 0.57, almost 40% of INRs were < 2 or > 4.5. Thirty-four TEs and 28 bleeding episodes occurred during a mean duration of 8.2 +/- 6.1 years of follow-tip. There was no trend of association of INR (average, SID, growth variance rate, or intensity-specific incidence of events) with TE. Previous cardiac surgery (p = 0.014) and ball prosthesis (hazard ratio [HR], 2.92; 95% CI, 1.43 to 5.94; p = 0.003) independently determined TE. With MVR using a ball prosthesis, despite higher anticoagulation intensity (p = 0.002), the 8-year rate of freedom from TE was considerably lower (50 +/- 9% vs 81 +/- 5%, respectively; p < 0.0001). Compared with expected stroke rates in the population, stroke risk was elevated with non-ball prosthesis MVR (HR 2.6; 95% CI, 1.3 to 5.2; p = 0.007) but was considerable with ball prosthesis MVR (HR 11.7; 95% CI, 7.5 to 18.4; p < 0.0001). INR variability (SD) was higher with a higher mean INR,,line (p < 0.0001). INR variability (HR 2.485; 95% CI, 1.11 to 5.55; p = 0.027) and cancer history (p < 0.0001) independently determined bleeding rates. Conclusion: This population-based comprehensive study of anticoagulation and TE post-MVR shows that, in these closely anticoagulated patients, anticoagulation intensity was highly variable and not associated with TE incidence post-MVR. Higher anticoagulation intensity is linked to higher variability and, thus, to bleeding. The MVR-ball prosthesis design is associated with higher TE rates notwithstanding higher anticoagulation intensity, and its use should be retired worldwide. (CHEST 2009; 136:1503-1513)