Advertisement

Isolated aortic valve replacement in North America comprising 108,687 patients in 10 years: Changes in risks, valve types, and outcomes in the Society of Thoracic Surgeons National Database

      Objective

      More than 200,000 aortic valve replacements are performed annually worldwide. We describe changes in the aortic valve replacement population during 10 years in a large registry and analyze outcomes.

      Methods

      The Society of Thoracic Surgeons National Database was queried for all isolated aortic valve replacements between January 1, 1997, and December 31, 2006. After exclusion for endocarditis and missing age or sex data, 108,687 isolated aortic valve replacements were analyzed. Time-related trends were assessed by comparing distributions of risk factors, valve types, and outcomes in 1997 versus 2006. Differences in case mix were summarized by comparing average predicted mortality risks with a logistic regression model. Differences across subgroups and time were assessed.

      Results

      There was a dramatic shift toward use of bioprosthetic valves. Aortic valve replacement recipients in 2006 were older (mean age 65.9 vs 67.9 years, P < .001) with higher predicted operative mortality risk (2.75 vs 3.25, P < .001); however, observed mortality and permanent stroke rate fell (by 24% and 27%, respectively). Female sex, age older than 70 years, and ejection fraction less than 30% were all related to higher mortality, higher stroke rate and longer postoperative stay. There was a 39% reduction in mortality with preoperative renal failure.

      Conclusions

      Morbidity and mortality of isolated aortic valve replacement have fallen, despite gradual increases in patient age and overall risk profile. There has been a shift toward bioprostheses. Women, patients older than 70 years, and patients with ejection fraction less than 30% have worse outcomes for mortality, stroke, and postoperative stay.

      Abbreviations and Acronyms:

      AVR (aortic valve replacement), STS (Society of Thoracic Surgeons)

      CTSNet classification

      The first aortic valve replacement (AVR) was performed by Harken 48 years ago.
      • Reardon M.J.
      • Reardon P.R.
      • Baldwin J.C.
      The evolution of aortic valve surgery.
      In the setting of aortic valve disease, patients are predominantly seen with aortic stenosis, which necessitates valve replacement. Ferguson and colleagues
      • Ferguson Jr., T.B.
      • Hammill B.G.
      • Peterson E.D.
      • DeLong E.R.
      • Grover F.L.
      A decade of change—risk profiles and outcomes for isolated coronary artery bypass grafting procedures, 1990-1999: a report from the STS National Database Committee and the Duke Clinical Research Institute. Society of Thoracic Surgeons.
      reviewed the Society of Thoracic Surgeons (STS) database with regard to coronary artery bypass grafting. They found that the population undergoing this procedure had aged and was at higher risk yet had a lower mortality. Edwards and associates
      • Edwards F.H.
      • Peterson E.D.
      • Coombs L.P.
      • DeLong E.R.
      • Jamieson W.R.
      • Shroyer A.L.
      • et al.
      Prediction of operative mortality after valve replacement surgery.
      developed a model for risk prediction in the setting of valve replacement surgery. This model was validated and proved accurate in predicting outcomes. Two percent of the population have bicuspid aortic valves, which are at risk for stenosis.
      • Goland S.
      • Czer L.S.
      • De Robertis M.A.
      • Mirocha J.
      • Kass R.M.
      • Fontana G.P.
      • et al.
      Risk factors associated with reoperation and mortality in 252 patients after aortic valve replacement for congenitally bicuspid aortic valve disease.
      • Roberts W.C.
      The congenitally bicuspid aortic valve. A study of 85 autopsy cases.
      Four percent of the elderly population have significant aortic stenosis.
      • Freeman R.V.
      • Otto C.M.
      Spectrum of calcific aortic valve disease: pathogenesis, disease progression, and treatment strategies.
      • Nkomo V.T.
      • Gardin J.M.
      • Skelton T.N.
      • Gottdiener J.S.
      • Scott C.G.
      • Enriquez-Sarano M.
      Burden of valvular heart disease: a population-based study.
      The size of the population older than 65 years will grow 50% between 2000 and 2030.

      Day CJ. Population projections of the united states by age, sex, race and Hispanic origin: 1995 to 2050. Curr Popul Rep Popul Estim Proj. 25–1130.

      • Chiappini B C.N.
      • Loforte A.
      • Di Marco L.
      • Di Bartolomeo R.
      • Marinelli G.
      Outcome after aortic valve replacement in octogenarians.
      Devices for transcatheter AVR are in development and clinical investigation. Early results suggest the feasibility of transcatheter valve therapy
      • Walther T.
      • Mohr F.W.
      Aortic valve surgery: time to be open-minded and to rethink.
      • Vassiliades Jr., T.A.
      • Block P.C.
      • Cohn L.H.
      • Adams D.H.
      • Borer J.S.
      • Feldman T.
      • et al.
      The clinical development of percutaneous heart valve technology.
      ; however, standard outcome measures must continue to drive clinical behavior. To this end, we explored the last 10 years of isolated AVR in the STS database with regard to the characteristics and outcomes of this group of patients as a whole and with time.

      Materials and Methods

      Patient Population

      The study population consisted of patients 20 years old or older who underwent isolated AVR at STS-participating hospitals between January 1, 1997, and December 31, 2006. From an initial population of 115,163 isolated AVR cases, we identified a subset of 108,791 patients (94.5%) without a history of endocarditis. From these, we excluded 104 patients (0.1%) with missing data on two key study variables, age and sex. The final study population consisted of 108,687 patients from 928 participating hospitals and surgeon groups. Although 928 participants contributed data during the study period, the number of participants in any single calendar year ranged from 365 to 756.

      End Points

      Outcome measures consisted of in-hospital mortality, permanent stroke, and postoperative stay. Postdischarge 30-day mortality was not analyzed, because this end point was not captured consistently by many participants during the study period.

      Analysis

      The distributions of patient characteristics and outcomes were summarized with percentages for categorical variables and means and medians for continuous variables. Differences in the prevalence of risk factors and outcomes in 1997 versus 2006 were assessed with stratified Mantel–Haenszel χ2 statistics, with STS participant identity serving as the stratification variable. Confidence intervals for the relative change in risk factor prevalence in 1997 versus 2006 were calculated by fitting generalized linear models with a log link function. SEs were calculated with an empirical sandwich estimator to account for correlation of observations within the same participant.
      To create a patient-level summary measure of case severity, we used logistic regression to estimate the probability of mortality for each patient in the study sample. Explanatory variables consisted of age, sex, ejection fraction, congestive heart failure, diabetes, renal failure, cerebrovascular accident, peripheral vascular disease, myocardial infarction, and surgical status. The patient's estimated probability of death is a simple summary measure that combines several individual risk factors into a single number. The observed to expected ratio statistic was then used to compare temporal trends in actual mortality with the average predicted probability from the logistic regression model. For each calendar year, the risk-adjusted mortality was calculated by multiplying the observed to expected ratio times the overall mortality during the study period. Finally, a test of trend was calculated by adding surgery year to the logistic regression model described previously and testing whether the coefficient was zero.

      Missing Data

      Patients with missing data were included in the denominator when reporting the prevalence of binary (yes/no) risk factors. We report the percentage of patients for whom each risk factor was coded as present and the percentage of patients for whom the data were unavailable. We included missing data in the denominator, because patient records frequently list risk factors that are present without enumerating the risk factors that are absent. For four variables (chronic obstructive pulmonary disease, New York Heart Association functional class, ejection fraction, and aortic insufficiency), there were large differences in the frequencies of missing data in 1997 and 2006. For these variables, we repeated the analysis with the subset of participants with at least 90% complete data for the variable. Records missing data for outcomes (mortality and stroke) were imputed to the no category.

      Results

      Both the number of AVRs and the number of participating programs increased during the 10-year study period (Table 1). Table 2 comprehensively outlines patient population characteristics in the overall population, in 1997, and in 2006. Selected patient characteristics are presented in Table 3. The incidences of age older than 70 years, obesity, diabetes, hypertension, chronic obstructive pulmonary disease, cerebrovascular disease, previous stroke, and renal failure all increased during the 10-year study period. The use of bioprosthetic valves increased to 78.4% of total valves used, whereas the mechanical valve use declined to 20.5% (P < .000001; Figure 1).
      Table 1Numbers of patients and participants by year of surgery
      Isolated atrial valve replacement recipientsParticipants
      YearPatients in databaseTotalExcluding endocarditisExcluding missing age or sexTotalIn study population
      1997252,6889,9329,4219,407466447
      1998240,59010,1089,5709,551460431
      1999210,2959,1088,6128,597380367
      2000200,8049,0398,5848,562365359
      2001205,4189,6569,1459,135422410
      2002227,78311,34210,71910,705499485
      2003232,04412,07011,35811,355549537
      2004235,33512,67111,94811,947616594
      2005250,02814,90714,03414,031709688
      2006258,41716,33015,40015,397756742
      Table 2Distribution of patient characteristics in study population: Overall, in 1997, and in 2006
      Overall (n = 108,687)In 1997 (n = 9407)In 2006 (n = 15,397)P value
      Demographic characteristics
       Age
      There were no missing data regarding age.
      (y)
      <.0001
        Median69.0069.0070
        25th59.0058.0060
        75th77.0076.0078
        Mean66.9065.9167.91
       Dichotomous age (No.)
        <70 y54,363 (50.02%)4946 (52.58%)7368 (47.85%)<.0001
        ≥70 y54,324 (49.98%)4461 (47.42%)8029 (52.15%)
       Age ranges (No.)<.0001
        <55 y19,659 (18.09%)1857 (19.74%)2470 (16.04%)
        55 to <60 y9371 (8.62%)719 (7.64%)1344 (8.73%)
        60 to <65 y11,120 (10.23%)1000 (10.63%)1595 (10.36%)
        65 to <70 y14,213 (13.08%)1370 (14.56%)1959 (12.72%)
        70 to <75 y17,349 (15.96%)1654 (17.58%)2283 (14.83%)
        75 to <80 y18,610 (17.12%)1547 (16.45%)2707 (17.58%)
        80 to <85 y13,127 (12.08%)911 (9.68%)2180 (14.16%)
        85 to <90 y4556 (4.19%)294 (3.13%)765 (4.97%)
        ≥90 y682 (0.63%)55 (0.58%)94 (0.61%)
       Sex (No.).7097
        Male62,013 (57.06%)5382 (57.21%)8827 (57.33%)
        Female46,674 (42.94%)4025 (42.79%)6570 (42.67%)
       Ethnicity (No.)
        Missing1951 (1.80%)288 (3.06%)99 (0.64%).0010
        White95,140 (87.54%)8164 (86.79%)13453 (87.37%)
        Nonwhite11,596 (10.67%)955 (10.15%)1845 (11.98%)
      Risk factors
       Body mass index (No.)<.0001
        Missing1964 (1.81%)445 (4.73%)85 (0.55%)
        <30 kg/m269,810 (64.23%)6387 (67.90%)9502 (61.71%)
        ≥30 kg/m236,913 (33.96%)2575 (27.37%)5810 (37.73%)
       Diabetes (No.)<.0001
        Missing1314 (1.21%)448 (4.76%)21 (0.14%)
        No85405 (78.58%)7505 (79.78%)11460 (74.43%)
        Yes21968 (20.21%)1454 (15.46%)3916 (25.43%)
       Type 1 diabetes (No.)
        Missing1703 (1.57%)520 (5.53%)32 (0.21%).0008
        No101,280 (93.19%)8451 (89.84%)14,375 (93.36%)
        Yes5704 (5.25%)436 (4.63%)990 (6.43%)
       Hypertension (No.)<.0001
        Missing900 (0.83%)303 (3.22%)18 (0.12%)
        No39713 (36.54%)4241 (45.08%)4292 (27.88%)
        Yes68074 (62.63%)4863 (51.7%)11087 (72.01%)
       Surgical status (No.)<.0001
        Missing373 (0.34%)63 (0.67%)17 (0.11%)
        Elective88,016 (80.98%)7896 (83.94%)12,042 (78.21%)
        Urgent19,361 (17.81%)1318 (14.01%)3252 (21.12%)
        Emergency808 (0.74%)101 (1.07%)80 (0.52%)
        Emergency salvage129 (0.12%)29 (0.31%)6 (0.04%)
       Nonelective status (No.)<.0001
        Missing373 (0.34%)63 (0.67%)17 (0.11%)
        No88,016 (80.98%)7896 (83.94%)12,042 (78.21%)
        Yes20,298 (18.68%)1448 (15.39%)3338 (21.68%)
       Chronic obstructive pulmonary disease (No.).0005
        Missing17,641 (16.23%)5575 (59.26%)70 (0.45%)
        No73,908 (68.00%)3243 (34.47%)12,262 (79.64%)
        Yes17,138 (15.77%)589 (6.26%)3065 (19.91%)
       Chronic obstructive pulmonary disease in subset (No.)<.0001
        Missing257 (0.29%)17 (0.48%)15 (0.10%)
        No71,299 (80.89%)2992 (85.12%)12,130 (79.92%)
        Yes16,582 (18.81%)506 (14.40%)3033 (19.98%)
       Peripheral vascular disease (No.).6522
        Missing1568 (1.44%)562 (5.97%)29 (0.19%)
        No97,874 (90.05%)8074 (85.83%)13,969 (90.73%)
        Yes9245 (8.51%)771 (8.20%)1399 (9.09%)
       Cerebrovascular disease (No.)<.0001
        Missing1696 (1.56%)636 (6.76%)29 (0.19%
        No95,659 (88.01%)8059 (85.67%)13,459 (87.41%)
        Yes11332 (10.43%)712 (7.57%)1909 (12.40%)
       Cerebrovascular accident (No.).7272
        Missing1513 (1.39%)536 (5.70%)29 (0.19%)
        No100,661 (92.62%)8357 (88.84%)14,397 (93.51%)
        Yes6513 (5.99%)514 (5.46%)971 (6.31%)
       Renal failure (No.).0135
        Missing1576 (1.45%)542 (5.76%)27 (0.18%)
        No101,719 (93.59%)8484 (90.19%)14,521 (94.31%)
        Yes5392 (4.96%)381 (4.05%)849 (5.51%)
       Preoperative dialysis (No.).3389
        Missing1705 (1.57%)578 (6.14%)29 (0.19%)
        No105,344 (96.92%)8701 (92.49%)15,104 (98.10%)
        Yes1638 (1.51%)128 (1.36%)264 (1.71%)
       Immunosuppressive treatment (No.).0016
        Missing1785 (1.64%)605 (6.43%)64 (0.42%)
        No103,899 (95.59%)8616 (91.59%)14,885 (96.67%)
        Yes3003 (2.76%)186 (1.98%)448 (2.91%)
      Previous cardiovascular interventions
       Coronary artery bypass grafting (No.)<.0001
        Missing937 (0.86%)203 (2.16%)62 (0.40%)
        No97,956 (90.13%)8582 (91.23%)13,702 (88.99%)
        Yes9794 (9.01%)622 (6.61%)1633 (10.61%)
       Valve surgery (No.).0058
        Missing967 (0.89%)257 (2.73%)40 (0.26%)
        No101,518 (93.4%)8490 (90.25%)14,541 (94.44%)
        Yes6202 (5.71%)660 (7.02%)816 (5.3%)
      Preoperative cardiac status
       Myocardial infarction (No.).0004
        Missing1594 (1.47%)569 (6.05%)30 (0.19%)
        No97,466 (89.68%)8079 (85.88%)13,809 (89.69%)
        Yes9627 (8.86%)759 (8.07%)1558 (10.12%)
       Myocardial infarction within 21 d (No.).0126
        Missing1917 (1.76%)642 (6.82%)36 (0.23%)
        No104,728 (96.36%)8609 (91.52%)15,057 (97.79%)
        Yes2042 (1.88%)156 (1.66%)304 (1.97%)
       Congestive heart failure (No.)<.0001
        Missing1652 (1.52%)510 (5.42%)32 (0.21%)
        No65,837 (60.57%)5016 (53.32%)10,169 (66.05%)
        Yes41,198 (37.91%)3881 (41.26%)5196 (33.75%)
       Angina (No.).6543
        Missing1508 (1.39%)503 (5.35%)33 (0.21%)
        No77,586 (71.38%)6302 (66.99%)10,971 (71.25%)
        Yes29593 (27.23%)2602 (27.66%)4393 (28.53%)
       Arrhythmia (No.)<.0001
        Missing1695 (1.56%)592 (6.29%)41 (0.27%)
        No89,475 (82.32%)7047 (74.91%)13,203 (85.75%)
        Yes17517 (16.12%)1768 (18.79%)2153 (13.98%)
       New York Heart Association functional class (No.)<.0001
        Missing8877 (8.17%)1819 (19.34%)636 (4.13%)
        I14,690 (13.52%)859 (9.13%)2225 (14.45%)
        II28,866 (26.56%)1613 (17.15%)5348 (34.73%)
        III42,452 (39.06%)3621 (38.49%)5672 (36.84%)
        IV13,802 (12.70%)1495 (15.89%)1516 (9.85%)
       New York Heart Association functional class in subset (No.)<.0001
        Missing416 (0.45%)52 (0.86%)25 (0.17%)
        I13,756 (14.97%)777 (12.81%)2177 (15.09%)
        II26,861 (29.23%)1277 (21.05%)5233 (36.26%)
        III38,539 (41.93%)2797 (46.10%)5532 (38.34%)
        IV12,339 (13.42%)1164 (19.19%)1463 (10.14%)
       Ejection fraction (%)<.0001
        N (% missing)94,207 (13.32%)7100 (24.52%)14,169 (7.98%)
        Median55%52%56%
        25th percentile45%41%50%
        75th percentile61%60%62%
        Mean53.39%51.86%54.38%
       Ejection fraction ranges (No.).0003
        Missing14,480 (13.32%)2307 (24.52%)1228 (7.98%)
        <30%5488 (5.05%)447 (4.75%)715 (4.64%)
        ≥30%88,719 (81.63%)6653 (70.72%)13,454 (87.38%)
       Ejection fraction in subset (%)<.0001
        N (% missing)62,360 (2.52%)2868 (3.01%)10,925 (2.43%)
        Median55%50%56%
        25th45%45%50%
        75th60%60%62%
        Mean53.36%51.34%54.43%
       Ejection fraction ranges in subset (No.).0183
        Missing1614 (2.52%)89 (3.01%)272 (2.43%)
        <30%3460 (5.41%)157 (5.31%)539 (4.81%)
        ≥30%58,900 (92.07%)2711 (91.68%)10,386 (92.76%)
       Aortic stenosis (No.)<.0001
        Missing2273 (2.09%)712 (7.57%)61 (0.40%)
        No18,467 (16.99%)1550 (16.48%)2327 (15.11%)
        Yes87,947 (80.92%)7145 (75.95%)13,009 (84.49%)
       Aortic insufficiency (No.)<.0001
        Missing10,916 (10.04%)3233 (34.37%)277 (1.80%)
        No46,062 (42.38%)2472 (26.28%)7831 (50.86%)
        Yes51,709 (47.58%)3702 (39.35%)7289 (47.34%)
       Aortic insufficiency in subset (No.).1996
        Missing442 (0.52%)41 (1.52%)31 (0.21%)
        No41,736 (49.15%)1319 (48.76%)7615 (52.38%)
        Yes42,744 (50.33%)1345 (49.72%)6891 (47.40%)
      Operative data
       Valve type (No.)<.0001
        Missing1947 (1.79%)335 (3.56%)72 (0.47%)
        None280 (0.26%)24 (0.26%)16 (0.10%)
        M35,284 (32.46%)4695 (49.91%)3164 (20.55%)
        B69,448 (63.90%)4104 (43.63%)12,069 (78.39%)
        H1192 (1.10%)169 (1.80%)20 (0.13%)
        A324 (0.30%)72 (0.77%)10 (0.06%)
        R147 (0.14%)8 (0.09%)12 (0.08%)
        BA65 (0.06%)0 (0%)34 (0.22%)
      There were no missing data regarding age.
      Table 3Relative changes in frequency of selected patient characteristics between 1997 and 2006
      Relative change95% Confidence intervalχ2P value
      Age ≥70 y10.0%6.4% to 13.6%31.997<.0001
      Female−0.3%−3.6% to 3.2%0.025.87513
      Nonwhite18.0%2.8% to 35.6%5.491.01911
      Body mass index ≥30 kg/m237.9%31.7% to 44.3%191.399<.0001
      Diabetes64.6%55.5% to 74.1%299.396<.0001
      Type 1 diabetes38.7%22.7% to 56.8%27.376<.0001
      Hypertension39.3%35.7% to 43.0%603.955<.0001
      Nonelective operation40.8%20.9% to 64.1%19.274.00001
      Emergency operation−51.6%−67.6% to −27.8%12.648.00038
      Chronic obstructive pulmonary disease217.9%165.0% to 281.5%154.655<.0001
      Chronic obstructive pulmonary disease in subset38.8%23.9% to 55.5%32.11<.0001
      Peripheral vascular disease10.9%−3.9% to 27.9%1.994.15796
      Cerebrovascular disease63.8%44.9% to 85.2%62.422<.0001
      Cerebrovascular accident15.4%3.6% to 28.6%6.729.00949
      Renal failure36.1%19.5% to 55.1%21.581<.0001
      Dialysis26.0%0.9% to 57.3%4.168.04121
      Immunosuppressant medication47.2%19.2% to 81.7%12.913.00033
      Previous coronary artery bypass grafting60.4%45.0% to 77.5%83.535<.0001
      Previous valve surgery−24.5%−32.6% to −15.3%23.248<.0001
      Previous myocardial infarction25.4%10.0% to 43.0%11.433.00072
      Myocardial infarction within 21 d19.1%−18.8% to 74.6%0.799.37146
      Congestive heart failure−18.2%−23.9% to −12.1%29.7<.0001
      Angina3.2%−5.6% to 12.7%0.476.49024
      Arrhythmia−25.6%−31.3% to −19.5%53.366<.0001
      New York Heart Association functional class IV−38.1%−47.1% to −27.4%35.059<.0001
      New York Heart Association functional class IV in subset−47.2%−55.5% to −37.2%52.569<.0001
      Ejection fraction <30%−2.3%−15.9% to 13.6%0.09.76396
      Ejection fraction <30% in subset−9.3%−24.1% to 8.2%1.176.27826
      Aortic stenosis11.2%8.2% to 14.4%55.839<.0001
      Aortic insufficiency20.3%11.9% to 29.3%25.148<.0001
      Aortic insufficiency in subset−4.7%−14.6% to 6.4%0.723.39503
      Mechanical valve−58.8%−61.8% to −55.7%549.909<.0001
      Bioprosthetic valve79.7%67.6% to 92.6%273.527<.0001
      Figure thumbnail gr1
      Figure 1Percentage use of bioprosthetic valves relative to mechanical valves from 1997 through 2006. Bioprosthetic valve use increased progressively during 10 years. Asterisk indicates P < .000001.
      Despite increases in comorbidity and predicted risk of death after AVR, the overall observed mortality and risk-adjusted mortality decreased (P < .01; Table 4 and Figure 2). Similarly, the incidence of stroke declined during the last 10 years (P < .05; Table 4 and Figure 3). Subgroup analysis (Table 5) demonstrated that mortality was higher for women than for men overall (not shown) and in the 1997 and 2006 populations (P < .01; Figure 4). Nonwhite patients had a higher mortality from AVR, which was not improved during the study period (Table 5). Age greater than 70 years, diabetes, peripheral vascular disease, ejection fraction less than 30%, and body mass index were all associated with higher mortality (Table 5). Mortality decreased more than 30% between 1997 and 2006 in the presence of diabetes, nonurgent cases, and renal failure and in the younger than 55 years, 55 to 60 years, 60 to 65 years, and older than 85 years age groups (Figure 5). In 2006 there were 2,431 patients who had undergone previous cardiac surgery and their associated mortality was 6.17%. By comparison, there were 12,966 patients with no prior heart surgery and the mortality was 2.35% (not shown). Stroke was reduced in the 55 to 60 years and the 65 to 70 years age groups (Figure 6). Overall stay was unchanged during the study period (not shown). Age older than 75 years, female sex, and ejection fraction less than 30% were associated with longer median stay (9.23 days vs 7.06 days, 7.80 days vs 7.40 days, and 9.53 days vs 7.67 days, respectively, P < .01 for all).
      Table 4Patient characteristics and outcomes by year
      1997199819992000200120022003200420052006
      Expected mortality (%)2.8%2.9%2.9%2.9%3.0%3.1%3.1%3.1%3.2%3.2%
      Observed mortality (%)3.4%3.3%3.5%3.3%3.2%2.9%2.8%2.8%2.3%2.6%
      Expected/observed mortality ratio1.201.161.211.141.070.950.900.910.730.80
      Risk-adjusted mortality (%)3.53%3.42%3.58%3.35%3.15%2.79%2.64%2.66%2.16%2.36%
      Observed stroke (%)1.7%1.9%1.6%1.6%1.6%1.7%1.4%1.4%1.5%1.3%
      Mean postoperative stay (d)87.97.87.87.87.87.87.87.77.8
      Mean age (y)65.966.266.267.066.966.966.967.067.367.9
      Mechanical valve (%)49.9%43.9%40.7%37.9%34.3%31.9%28.8%27.0%24.4%20.5%
      Bioprosthetic valve (%)43.6%48.3%50.7%56.4%62.1%65.8%69.3%70.8%74.3%78.4%
      Mechanical and bioprosthetic valves (%)93.5%92.2%91.4%94.3%96.4%97.8%98.1%97.8%98.8%98.9%
      Ratio of bioprosthetic to mechanical valves0.4660.5240.5550.5980.6440.6730.7060.7240.7530.792
      Figure thumbnail gr2
      Figure 2Risk-adjusted mortality for aortic valve replacement during 10 years in Society of Thoracic Surgeons database. Mortality for aortic valve replacement decreased with time. Asterisk indicates P < .01.
      Figure thumbnail gr3
      Figure 3Stroke rate after aortic valve replacement (AVR) in Society of Thoracic Surgeons database from 1997 through 2006. Stroke rate decreased during 10-year study period. Asterisk indicates P < .01.
      Table 5Changes in outcome between 1997 and 2006 by subgroup
      MortalityStroke
      NRate (%)Relative change (%)Rate (%)Relative change (%)
      1997200619972006Value95% Confidence intervalP value19972006Value95% Confidence intervalP value
      All940715,3973.4%2.6%−24.3%−35.1% to −11.6%<.011.7%1.3%−21.1%−35.9% to −2.9%.03
      Male
       No40256,5704.1%3.2%−20.8%−35.3% to −3.0%.022.0%1.4%−28.1%−45.6% to −5.1%.02
       Yes53828,8272.8%2.1%−27.9%−42.6% to −9.4%<.011.4%1.2%−13.7%−36.1% to 16.5%.34
      Nonwhite
       No845213,5523.3%2.5%−24.5%−36.1% to −10.7%<.011.6%1.3%−15.3%−32.3% to 5.9%.15
       Yes9551,8454.4%3.3%−26.1%−49.8% to 8.9%.132.5%1.2%−52.6%−73.1% to −16.5%<.01
      Age ≤70 y
       No49467,3682.2%1.3%−43.8%−57.4% to −25.7%<.011.2%0.7%−43.3%−62.7% to −13.9%<.01
       Yes44618,0294.6%3.7%−19.1%−32.8% to −2.6%.022.2%1.9%−13.9%−32.5% to 9.8%.23
      Diabetes
       No795311,4812.8%2.2%−24.0%−36.9% to −8.4%<.011.4%1.3%−8.2%−27.6% to 16.3%.48
       Yes14543,9166.3%3.7%−41.1%−53.8% to −24.8%<.013.0%1.4%−55.3%−69.7% to −34.0%<.01
      Peripheral vascular disease
       No863613,9983.0%2.3%−24.6%−36.3% to −10.9%<.011.6%1.3%−19.0%−36.0% to 2.6%.08
       Yes7711,3997.1%5.2%−26.9%−48.9% to 4.6%.093.1%2.0%−35.7%−62.8% to 11.0%.11
      Previous valve surgery
       No874714,5813.3%2.4%−27.0%−38.0% to −14.0%<.011.7%1.3%−21.6%−36.8% to −2.8%.03
       Yes6608164.8%5.6%16.3%−25.6% to 81.6%.511.7%1.5%−11.8%−60.4% to 96.7%.76
      Myocardial infarction
       No864813,8393.1%2.2%−28.3%−39.4% to −15.1%<.011.6%1.1%−29.5%−43.6% to −11.9%<.01
       Yes7591,5586.1%5.3%−13.2%−39.6% to 24.9%.452.2%2.9%29.0%−23.9% to 118.5%.34
      Congestive heart failure
       No552610,2012.1%1.6%−23.6%−40.0% to −2.7%.031.4%1.2%−17.4%−37.2% to 8.8%.17
       Yes38815,1965.2%4.4%−14.9%−29.9% to 3.4%.102.1%1.6%−20.6%−41.9% to 8.4%.15
      Urgent operation
       No795912,0592.8%1.9%−31.9%−43.4% to −18.1%<.011.6%1.2%−23.8%−39.7% to −3.6%.02
       Yes14483,3386.5%4.9%−24.8%−41.8% to −2.7%.032.0%1.6%−17.7%−47.5% to 28.9%.39
      Ejection fraction <30%
       No896014,6823.2%2.4%−14.3%−46.2% to 36.5%.521.7%1.3%−9.7%−61.3% to 110.5%.81
       Yes4477156.0%5.2%−25.1%−36.2% to −12.0%<.012.0%1.8%−21.8%−36.6% to −3.5%.02
      Body mass index ≥30 kg/m2
       No68329,5873.5%2.7%−21.9%−35.0% to −6.1%<.011.8%1.5%−17.4%−35.0% to 4.8%.12
       Yes25755,8103.0%2.3%−25.6%−42.8% to −3.1%.031.2%1.0%−19.7%−47.1% to 21.9%.30
      Renal failure
       No902614,5482.9%2.2%−24.6%−36.2% to −10.8%<.011.6%1.3%−17.9%−33.6% to 1.6%.07
       Yes38184913.9%8.5%−39.0%−55.6% to −16.2%<.014.2%1.9%−55.1%−76.5% to −14.4%.01
      Age (y)
       <5518572,4701.5%0.9%−38.7%−64.8% to 6.6%.080.6%0.5%−18.0%−63.5% to 84.4%.63
       55 to <607191,3442.2%0.6%−73.3%−88.5% to −37.9%<.011.1%0.2%−79.9%−94.5% to −26.2%.01
       60 to <6510001,5953.2%1.6%−51.0%−70.7% to −18.0%<.011.1%0.9%−14.5%−61.3% to 89.1%.70
       65 to <7013701,9592.6%1.9%−26.2%−52.9% to 15.7%.192.0%1.0%−52.5%−74.5% to −11.8%.02
       70 to <7516542,2833.2%2.9%−9.8%−37.4% to 30.0%.581.9%1.5%−23.0%−51.9% to 23.2%.28
       75 to <8015472,7074.6%3.3%−29.2%−47.6% to −4.2%.032.3%2.1%−5.3%−37.4% to 43.3%.80
       80 to <859112,1806.3%4.9%−22.3%−42.9% to 5.8%.112.2%2.0%−10.2%−45.0% to 46.8%.67
       85 to <902947657.8%4.1%−48.2%−69.0% to −13.5%.014.1%2.4%−42.4%−71.9% to 18.4%.13
       ≥9055943.6%9.6%163.3%−40.2% to 1059.2%.201.8%2.1%17.0%−89.2% to 1168.6%.90
      Missing data were imputed to the no category.
      Figure thumbnail gr4
      Figure 4Mortality from aortic valve replacement among male and female patients. Female patients had greater mortality in 1997, in 2006, and in overall population. Asterisk indicates P < .01.
      Figure thumbnail gr5
      Figure 5Mortality versus age in aortic valve replacement study population. Mortality was age dependent in 1997 and in 2006. For age groups as shown, mortality was less in 2006 than in 1997. Asterisk indicates P < .05.
      Figure thumbnail gr6
      Figure 6Stroke versus age in aortic valve replacement population between 1997 and 2006. Stroke rate was age dependent but also reduced as shown between 1997 and 2006. Asterisk indicates P < .05.

      Discussion

      We defined a population of patients undergoing AVR in STS-participating North American centers from 1997 to 2006. Methods used in this study allow for completeness and accuracy of database information, as delineated in Table 1, Table 2, Table 3, Table 4, Table 5. Despite increases in morbidity and predicted mortality, this study demonstrated improved mortality and stroke rates during the 10-year study period. For patients younger than 60 years, mortality from AVR in 2006 was less than 1.0%; for those younger than 70 years, 1.3%; for those younger than 80 years, below 3.5%; and for those under 85 years, below 5% (Table 5). Overall mortality fell 24%, and risk-adjusted mortality fell 33%. Assuming that the risk models that have been developed and validated
      • Ambler G Omar R.Z.
      • Royston P.
      • Kinsman R.
      • Keogh B.E.
      • Taylor K.M.
      Generic, simple risk stratification model for heart valve surgery.
      • Jin R.
      • Grunkemeier G.L.
      • Starr A.
      Validation and refinement of mortality risk models for heart valve surgery.
      • Nowicki E.R.
      • Birkmeyer N.J.
      • Weintraub R.W.
      • Leavitt B.J.
      • Sanders J.H.
      • Dacey L.J.
      • et al.
      Multivariable prediction of in-hospital mortality associated with aortic and mitral valve surgery in Northern New England.
      • Kuduvalli M.
      • Grayson A.D.
      • Au J.
      • Grotte G.
      • Bridgewater B.
      • Fabri B.M.
      A multi-centre additive and logistic risk model for in-hospital mortality following aortic valve replacement.
      • Hannan E.L.
      • Wu C.
      • Bennett E.V.
      • Carlson R.E.
      • Culliford A.T.
      • Gold J.P.
      • et al.
      Risk index for predicting in-hospital mortality for cardiac valve surgery.
      for heart valve surgery and the model developed by the STS and Duke Clinical Research Institute
      • Edwards F.H.
      • Peterson E.D.
      • Coombs L.P.
      • DeLong E.R.
      • Jamieson W.R.
      • Shroyer A.L.
      • et al.
      Prediction of operative mortality after valve replacement surgery.
      are accurate, we conclude that surgical teams participating in the STS database have dramatically improved performance for isolated AVR.
      With regard to specific subgroups, the population of AVR recipients became older and more obese and had increased incidences of diabetes, hypertension, pulmonary disease, cerebrovascular disease, and renal failure during the 10 years. Despite these changes, overall mortality fell for each subgroup. It also fell for most patient subsets outlined in Table 5. Subgroup stroke rate also decreased during the 10-year period despite increasing age and risk in this AVR population (Table 5 and Figure 6). To a degree, stroke and mortality are dependent, because stroke leads to higher mortality. Nonetheless, for patients younger than 70 years, risk of stroke after AVR was 0.7% in 2006. Between the ages of 70 and 80 years, stroke rate in 2006 was less than 2.0%, and even for octogenarians, stroke was less than 2.5% (Table 5 and Figure 6). Stroke rate in this study was time dependent as well as age dependent. Female patients had higher mortality, higher stroke rate, and longer postoperative stay relative to male patients. This was true for the overall population, the 1997 group, and the 2006 group. Bridges and coworkers
      • Bridges C.R.
      • O'Brien S.M.
      • Cleveland J.C.
      • Savage E.B.
      • Gammie J.S.
      • Edwards F.H.
      • et al.
      Association between indices of prosthesis internal orifice size and operative mortality after isolated aortic valve replacement.
      previously demonstrated a relationship between size and outcome in the STS database in the setting of AVR. Because female patients have a smaller body size on average than do male patients, the increased mortality among female patients is consistent with reports linking body size to outcome. Factors that cause this effect of higher female adverse outcome rate and could possibly be manipulated to ameliorate it are unclear, however, and will require further study. Increased adverse outcomes in the nonwhite patients were also observed in this study. This observation in the setting of heart surgery has also been made in previous reports.
      • Bridges C.R.
      Cardiac surgery in African Americans.
      Again, the study design of this review was not sufficient to explain this finding.
      The dramatic shift away from mechanical heart valves toward bioprosthetic heart valves is difficult to explain because of the relatively short time frame in which it occurred. Nonetheless, many young patients refuse long-term anticoagulation, and elderly patients are at high risk when receiving anticoagulation. There has been evidence that reoperation to replace a failed bioprosthetic valve can be accomplished with good outcomes driven by factors other than simple replacement of the valve, such as age, degree of heart failure, and coronary disease.
      • Byrne J.G.
      • Aranki S.F.
      • Couper G.S.
      • Adams D.H.
      • Allred E.N.
      • Cohn L.H.
      Reoperative aortic valve replacement: partial upper hemisternotomy versus conventional full sternotomy.
      • Jamieson W.R.
      • Burr L.H.
      • Miyagishima R.T.
      • Janusz M.T.
      • Fradet G.J.
      • Ling H.
      • et al.
      Re-operation for bioprosthetic aortic structural failure—risk assessment.
      Newer generation tissue valves are expected to provide longer reoperation-free survivals. Finally, the population of patients has aged during the study period, and it is expected that the elderly segment of the population will continue to grow dramatically. Multiplying and adding risk through the patient's lifetime to derive a predicted total lifetime risk for valve implantation at the time of the index operation favors a bioprosthetic valve over mechanical valve and may explain the finding in this study of a nationwide shift toward bioprosthetic valves.
      • El Oakley R.
      • Klein P.
      • Bach D.S.
      Choice of prosthetic heart valve in today's practice.
      All these factors taken together have influenced surgeon and patient valve choices.
      In conclusion, predicted risk and comorbidities of patients undergoing AVR have increased during the last 10 years in this country. Despite these changes, outcomes, including rates of death and stroke, not only have improved but are quite low for isolated AVR. There has been a dramatic shift toward the use of bioprosthetic valves during the 10-year study period. Female sex is associated with higher rates of death and adverse outcomes in the setting of isolated AVR, a finding that requires a search for cause.

      Study Limitations

      This study was based on the STS database and therefore by definition was a retrospective review of patient data submitted by participating centers. Furthermore, the cases studied were nonconsecutive and based on voluntary participation in the STS database. In addition, this was a study of AVR only. New pioneering therapies, such as aggressive and effective repair techniques for aortic insufficiency, will change the focus to short- and long-term outcomes from treatment of a disease rather than outcomes from a particular procedure.
      • Lausberg H.F.
      • Aicher D.
      • Langer F.
      • Schafers H.J.
      Aortic valve repair with autologous pericardial patch.
      Long-term data cannot yet be linked to the in-hospital and 30-day outcome measures provided by the STS database. Further investigation will require inclusion of long-term outcomes and health-related quality of life in any assessment of surgical therapy of valve disease.

      References

        • Reardon M.J.
        • Reardon P.R.
        • Baldwin J.C.
        The evolution of aortic valve surgery.
        Contemp Surg. 1997; 50: 161-168
        • Ferguson Jr., T.B.
        • Hammill B.G.
        • Peterson E.D.
        • DeLong E.R.
        • Grover F.L.
        A decade of change—risk profiles and outcomes for isolated coronary artery bypass grafting procedures, 1990-1999: a report from the STS National Database Committee and the Duke Clinical Research Institute. Society of Thoracic Surgeons.
        Ann Thorac Surg. 2002; 73: 480-490
        • Edwards F.H.
        • Peterson E.D.
        • Coombs L.P.
        • DeLong E.R.
        • Jamieson W.R.
        • Shroyer A.L.
        • et al.
        Prediction of operative mortality after valve replacement surgery.
        J Am Coll Cardiol. 2001; 37: 885-892
        • Goland S.
        • Czer L.S.
        • De Robertis M.A.
        • Mirocha J.
        • Kass R.M.
        • Fontana G.P.
        • et al.
        Risk factors associated with reoperation and mortality in 252 patients after aortic valve replacement for congenitally bicuspid aortic valve disease.
        Ann Thorac Surg. 2007; 83: 931-937
        • Roberts W.C.
        The congenitally bicuspid aortic valve. A study of 85 autopsy cases.
        Am J Cardiol. 1970; 26: 72-83
        • Freeman R.V.
        • Otto C.M.
        Spectrum of calcific aortic valve disease: pathogenesis, disease progression, and treatment strategies.
        Circulation. 2005; 111: 3316-3326
        • Nkomo V.T.
        • Gardin J.M.
        • Skelton T.N.
        • Gottdiener J.S.
        • Scott C.G.
        • Enriquez-Sarano M.
        Burden of valvular heart disease: a population-based study.
        Lancet. 2006; 368: 1005-1011
      1. Day CJ. Population projections of the united states by age, sex, race and Hispanic origin: 1995 to 2050. Curr Popul Rep Popul Estim Proj. 25–1130.

        • Chiappini B C.N.
        • Loforte A.
        • Di Marco L.
        • Di Bartolomeo R.
        • Marinelli G.
        Outcome after aortic valve replacement in octogenarians.
        Ann Thorac Surg. 2004; 78: 85-89
        • Walther T.
        • Mohr F.W.
        Aortic valve surgery: time to be open-minded and to rethink.
        Eur J Cardiothorac Surg. 2007; 31: 4-6
        • Vassiliades Jr., T.A.
        • Block P.C.
        • Cohn L.H.
        • Adams D.H.
        • Borer J.S.
        • Feldman T.
        • et al.
        The clinical development of percutaneous heart valve technology.
        J Am Coll Cardiol. 2005; 45: 1554-1560
        • Ambler G Omar R.Z.
        • Royston P.
        • Kinsman R.
        • Keogh B.E.
        • Taylor K.M.
        Generic, simple risk stratification model for heart valve surgery.
        Circulation. 2005; 112: 224-231
        • Jin R.
        • Grunkemeier G.L.
        • Starr A.
        Validation and refinement of mortality risk models for heart valve surgery.
        Ann Thorac Surg. 2005; 80: 471-479
        • Nowicki E.R.
        • Birkmeyer N.J.
        • Weintraub R.W.
        • Leavitt B.J.
        • Sanders J.H.
        • Dacey L.J.
        • et al.
        Multivariable prediction of in-hospital mortality associated with aortic and mitral valve surgery in Northern New England.
        Ann Thorac Surg. 2004; 77: 1966-1977
        • Kuduvalli M.
        • Grayson A.D.
        • Au J.
        • Grotte G.
        • Bridgewater B.
        • Fabri B.M.
        A multi-centre additive and logistic risk model for in-hospital mortality following aortic valve replacement.
        Eur J Cardiothorac Surg. 2007; 31: 607-613
        • Hannan E.L.
        • Wu C.
        • Bennett E.V.
        • Carlson R.E.
        • Culliford A.T.
        • Gold J.P.
        • et al.
        Risk index for predicting in-hospital mortality for cardiac valve surgery.
        Ann Thorac Surg. 2007; 83: 921-929
        • Bridges C.R.
        • O'Brien S.M.
        • Cleveland J.C.
        • Savage E.B.
        • Gammie J.S.
        • Edwards F.H.
        • et al.
        Association between indices of prosthesis internal orifice size and operative mortality after isolated aortic valve replacement.
        J Thorac Cardiovasc Surg. 2007; 133: 1012-1021
        • Bridges C.R.
        Cardiac surgery in African Americans.
        Ann Thorac Surg. 2003; 76: S1356-S1362
        • Byrne J.G.
        • Aranki S.F.
        • Couper G.S.
        • Adams D.H.
        • Allred E.N.
        • Cohn L.H.
        Reoperative aortic valve replacement: partial upper hemisternotomy versus conventional full sternotomy.
        J Thorac Cardiovasc Surg. 1999; 118: 991-997
        • Jamieson W.R.
        • Burr L.H.
        • Miyagishima R.T.
        • Janusz M.T.
        • Fradet G.J.
        • Ling H.
        • et al.
        Re-operation for bioprosthetic aortic structural failure—risk assessment.
        Eur J Cardiothorac Surg. 2003; 24: 873-878
        • El Oakley R.
        • Klein P.
        • Bach D.S.
        Choice of prosthetic heart valve in today's practice.
        Circulation. 2008; 117: 253-256
        • Lausberg H.F.
        • Aicher D.
        • Langer F.
        • Schafers H.J.
        Aortic valve repair with autologous pericardial patch.
        Eur J Cardiothorac Surg. 2006; 30: 244-249