From Pages to Practice

By Balimkiz Senman; Michael Mi, M.D.

Published November 8, 2017


Valve replacement is a life-saving intervention for patients with severe valvular heart disease. Currently, there are two main classes of prostheses — mechanical or biologic — each with its own risks and benefits. Mechanical prosthetic valves necessitate lifelong anticoagulation due to their prothrombotic nature, and increase the risk for stroke and bleeding. The decreased durability of biologic prostheses leaves patients at risk for reoperations, with the attendant surgical risks. Current guidelines recommend biologic prostheses for older patients who typically don’t tolerate anticoagulation well and may not live long enough to require a second operation, and mechanical prostheses for younger patients. Recent observational studies have shown no significant mortality difference between the two valve types in patients in the gray zone (aged 50– 69 years), leading to increased use of biologic prostheses. In this week’s issue of NEJM, Goldstone et al. argue that prior studies may have failed to detect important mortality differences between the two types of prostheses due to low statistical power.

In a retrospective cohort study using administrative databases, the authors examined outcomes for 9942 patients who underwent aortic-valve replacement (AVR) alone and 15,503 patients who underwent mitral-valve replacement (MVR) with or without concomitant procedures. All patients had surgery in one of 142 nonfederal hospitals in California from 1996 through 2013. Patients were stratified by age to assess current practice guidelines: 45–54 years and 55–64 years for AVR and 40–49 years, 50–69 years, and 70–79 years for MVR. The primary endpoint was mortality. Secondary end points included perioperative mortality (≤30 days after surgery) and cumulative incidence of stroke, bleeding, or reoperation.

From 1996 through 2013, the use of biologic prostheses increased from 11.5% to 51.6% for AVR and from 16.8% to 53.7% for MVR. Perioperative mortality was similar between the two prosthesis types and valve positions for all age groups, except for higher 30-day mortality associated with biologic prostheses in MVR patients aged 40 to 49 years. For AVR, long-term 15-year mortality associated with biologic prostheses was significantly higher in patients aged 45–54 years (30.6% vs. 26.4%) but not for those aged 55–64 years (36.1% vs. 32.1%). For MVR, 15-year mortality associated with biologic prostheses was significantly higher for patients aged 40–49 years (44.1% vs. 27.1%) and 50–69 years (50.0% vs. 45.3%), but not for those aged 70–79 years (78.3% vs. 77.3%). These findings persisted after multivariate adjustment for baseline patient characteristics and hospital. Biologic prosthetic valves were associated with higher risk of reoperation and perioperative mortality following reoperation (7.1% for AVR and 14.0% for MVR) whereas mechanical prostheses were associated with higher risk of bleeding and, in some groups, stroke.

Overall, the authors’ findings and conclusion challenge the current practice that increasingly favors biologic prostheses for younger patients. Mechanical prostheses appear to offer long-term survival benefits in patients up to age 55 for AVR and up to age 70 for MVR. Study limitations include the observational design and reliance on administrative databases that lack important clinical detail. Although it’s impossible to fully control for confounding by indication, this study does provide a contemporary and well-powered look at the risks associated with the two types of valves, an excellent starting point to begin rethinking the current guidelines, and motivation for more definitive studies to assess the best valve for specific clinical situations.

Browse more From Pages to Practice »

Balimkiz Senman is a medical student at UMass Medical Center.


Michael Mi, MD is a NEJM Editorial Fellow and a BIDMC Hospitalist.