Author + information
- Received November 4, 2015
- Revision received January 14, 2016
- Accepted January 19, 2016
- Published online March 29, 2016.
- Mani Arsalan, MDa,b,
- Molly Szerlip, MDa,
- Sreekanth Vemulapalli, MDc,
- Elizabeth M. Holper, MDa,
- Suzanne V. Arnold, MDd,
- Zhuokai Li, PhDc,
- Michael J. DiMaio, MDa,
- John S. Rumsfeld, MDe,
- David L. Brown, MDa and
- Michael J. Mack, MDa,∗ ()
- aThe Heart Hospital Baylor Plano, Plano, Texas
- bKerckhoff Heart-Center, Bad Nauheim, Germany
- cDuke Clinical Research Institute, Durham, North Carolina
- dSaint Luke's Mid America Heart Institute, Kansas City, Missouri
- eDenver VA Medical Center, Denver, Colorado
- ↵∗Reprint requests and correspondence:
Dr. Michael J. Mack, Heart Hospital Baylor Plano, 1100 Allied Drive, Plano, Texas 75093.
Background Data demonstrating the outcome of transcatheter aortic valve replacement (TAVR) in the very elderly patients are limited, as they often represent only a small proportion of the trial populations.
Objectives The purpose of this study was to compare the outcomes of nonagenarians to younger patients undergoing TAVR in current practice.
Methods We analyzed data from the Society of Thoracic Surgeons/American College of Cardiology TVT (Transcatheter Valve Therapy) Registry. Outcomes at 30 days and 1 year were compared between patients ≥90 years versus <90 years of age using cumulative incidence curves. Quality of life was assessed with the 12-item Kansas City Cardiomyopathy Questionnaire.
Results Between November 2011 and September 2014, 24,025 patients underwent TAVR in 329 participating hospitals, of which 3,773 (15.7%) were age ≥90 years. The 30-day and 1-year mortality rates were significantly higher among nonagenarians (age ≥90 years vs. <90 years: 30-day: 8.8% vs. 5.9%; p < 0.001; 1 year: 24.8% vs. 22.0%; p < 0.001, absolute risk: 2.8%, relative risk: 12.7%). However, nonagenarians had a higher mean Society of Thoracic Surgeons Predicted Risk of Operative Mortality score (10.9% vs. 8.1%; p < 0.001) and, therefore, had similar ratios of observed to expected rates of 30-day death (age ≥90 years vs. <90 years: 0.81, 95% confidence interval: 0.70 to 0.92 vs. 0.72, 95% confidence interval: 0.67 to 0.78). There were no differences in the rates of stroke, aortic valve reintervention, or myocardial infarction at 30 days or 1 year. Nonagenarians had lower (worse) median Kansas City Cardiomyopathy Questionnaire scores at 30 days; however, there was no significant difference at 1 year.
Conclusions In current U.S. clinical practice, approximately 16% of patients undergoing TAVR are ≥90 years of age. Although 30-day and 1-year mortality rates were statistically higher compared with younger patients undergoing TAVR, the absolute and relative differences were clinically modest. TAVR also improves quality of life to the same degree in nonagenarians as in younger patients. These data support safety and efficacy of TAVR in select very elderly patients.
It is estimated that the number of people age ≥90 years (nonagenarians) in the United States will quadruple by the year 2050 to reach 8.7 million (1). As such, clinicians are being confronted with an increasing number of nonagenarians with severe aortic stenosis (AS), which significantly reduces quality of life and survival. Due to the morbidity and mortality of surgical aortic valve replacement in patients at advanced age, surgery is often denied to very elderly patients (2,3). Over the last 10 years, transcatheter aortic valve replacement (TAVR) has emerged as a viable treatment option for patients with severe AS who are inoperable or at high surgical risk, prolonging survival and improving quality of life in the majority of patients (4,5). However, the effect of TAVR in nonagenarians is largely unknown, as they represent only a small fraction of patients enrolled in the pivotal clinical trials. A few small, single-center series have reported outcomes of TAVR in the very elderly and showed acceptable results (6–9). Due to this lack of outcomes data for TAVR in the very elderly, decision-making for TAVR in nonagenarians is complicated. As such, the aim of this study was to compare the procedural, 30-day, and 1-year outcomes of nonagenarians with patients age <90 years undergoing TAVR in current clinical practice using comprehensive data from the STS/ACC (Society of Thoracic Surgeons/American College of Cardiology) TVT (Transcatheter Valve Therapy) Registry.
The STS/ACC TVT registry
The TVT Registry collects clinical information including patient demographics, comorbidities, functional status, quality of life, and procedural details in addition to post-operative, 30-day, and 1-year outcomes using standardized definitions on virtually all patients undergoing TAVR with a commercially approved device in the United States (10,11). The Chesapeake Central Institutional Review Board (IRB) and the Duke University School of Medicine IRB approved the registry. Both IRB committees granted a waiver of informed consent and authorization for this study.
Nonagenarians were defined as patients age ≥90 years at the time of the procedure, and a small number of centenarians were included in this study (n = 24). TVT Registry clinical records for procedures performed from November 2011 through September 2014 were linked to Medicare administrative claims using direct patient identifiers (name and social security number) by the Centers for Medicare & Medicaid Services (CMS). Per the CMS National Coverage Determination for reimbursement, all patients were required to have site documentation of echocardiographically defined severe AS and an assessment by 2 cardiothoracic surgeons who independently deemed the patients as at high or prohibitive surgical risk of mortality from surgical aortic valve replacement. Of the 24,025 index TVT procedure records, 8,502 were not linked to Medicare, either because of patient nonparticipation in the Medicare Parts A and B fee-for-service program at the time of the index procedure or an inability to link the index admission to a Medicare inpatient claim. For quality of life, the study cohort was limited to procedures performed on or before July 17, 2014, for 30-day assessment and procedures on or before August 1, 2013, for 1-year assessment to allow for appropriate follow-up.
Primary outcomes studied included death, stroke, rehospitalization due to heart failure, aortic valve reintervention, myocardial infarction (MI), and quality of life (QOL) at 30 days and 1 year. QOL was assessed with the 12-item Kansas City Cardiomyopathy Questionnaire (KCCQ-12), a 12-item condensed psychometrically valid version of the full Kansas City Cardiomyopathy Questionnaire (KCCQ) (12). A disease-specific health status survey originally developed to describe and monitor health status in patients with heart failure, the KCCQ has also been validated in patients with aortic stenosis (13). For this study, we focused on the overall summary score of the 12-item Kansas City Cardiomyopathy Questionnaire (KCCQ-os), which ranges from 0 to 100 with higher scores indicating less symptom burden, less physical and social limitations, and better quality of life.
In-hospital outcomes were collected as part of the TVT Registry. Valve Academic Research Consortium definitions were used for major bleeding (10,11). All site-reported stroke and valve reintervention events were adjudicated by a board-certified cardiologist using Valve Academic Research Consortium definitions. This process involved review of specific site queries and de-identified source records as needed. Following hospital discharge, death was identified using the Medicare Denominator file. Medicare in-hospital administrative claims files were used for detection of re-hospitalization events through October 2014 using the following International Classification of Diseases, Ninth Revision, Clinical Modification codes: for stroke; 433.x1, 434.x1, 997.02, 436, 437.1, 437.9, 430, 431, and 432.x; for heart failure, 398.x, 402.x1, 404.x1, 404.x3, and 428.x; for aortic valve reintervention, 35.11, 35.21, 35.22, 35.01, 35.05, 35.06, and 35.09; and for MI, 410.x0 and 410.x1. For rehospitalization, follow-up was censored at the end of fee-for-service coverage, loss of Part A or B coverage, or end of the follow-up period (October 31, 2014), whichever occurred first.
Baseline characteristics and in-hospital outcomes of the study population were presented according to age. Categorical variables were summarized as percentages, and continuous variables as medians with interquartile ranges. The baseline characteristics and in-hospital outcomes of patients ≥90 and <90 years of age were then compared using the Pearson chi-square test for categorical variables and the Wilcoxon rank sum test for continuous variables.
Cumulative incidences of death and nonfatal outcomes at 30 days and 1 year post-TAVR were estimated for patients ≥90 and <90 years of age. For nonfatal outcomes, including stroke, heart failure readmission, aortic valve reintervention and post-procedural MI, death was considered a competing risk, and therefore the cumulative incidence indicated the probability of a nonfatal outcome occurring given that death may impede its occurrence (14). The 30-day observed-to-expected mortality ratios were calculated on the basis of the baseline Society of Thoracic Surgeons Predicted Risk of Operative Mortality (STS PROM) score, with 95% confidence intervals (CIs) obtained using a normal approximation to the binomial distribution (15).
The Cox proportional hazards models were used to assess the unadjusted and adjusted effects of age on 30-day and 1-year mortality. Nonfatal outcomes were assessed using Fine and Gray proportional subdistribution hazards models in the presence of competing risk of death (14,16). The multivariable models included the covariates in the recently developed TVT model for in-hospital mortality (unpublished data). The risk of adverse outcomes for patients ≥90 years versus <90 years of age was reported using hazard ratios (HRs) with 95% CIs.
Complete case analysis was performed for 30-day and 1-year KCCQ-os scores. KCCQ-os scores were summarized as medians with interquartile ranges and compared between patients ≥90 and <90 years of age using the Wilcoxon rank sum test. The effect of age on follow-up KCCQ-os scores was assessed using linear regression models that included a binary indicator variable for age group. Models were constructed as unadjusted (including age only) and adjusted, including age, baseline STS PROM score, and baseline KCCQ-os score. As the STS PROM score is calculated including age as well as its interactions with several risk factors, this complicates the multivariable adjustment. As such, the baseline STS PROM score was recalculated assuming all patients were 85 years of age, the median age of the study population. All analyses were performed using SAS software, version 9.4 (SAS Institute Inc., Cary, North Carolina), and a p value <0.05 was considered statistically significant.
From November 2011 through September 2014, 24,025 patients, of which 3,773 (15.7%) were nonagenarians, underwent TAVR at 329 participating hospitals. The median age was 92 years in nonagenarians and 82 years in the younger cohort. Compared to patients <90 years of age, nonagenarians were more likely to be female and less likely to have high-risk features including prior nonaortic valve cardiac surgery procedure, diabetes, prior stroke, and prior MI, but overall, they had higher estimated surgical mortality (STS PROM scores, age ≥90 years vs. <90 years: 9.2% vs. 6.3%; p < 0.001). Although nonagenarians had slower 5-m walk tests (8.7 s vs. 8.0 s; p < 0.001), they had better self-reported QOL (KCCQ-os: 41.7 vs. 37.5; p < 0.001) prior to TAVR (Table 1).
Compared with patients <90 years of age, nonagenarians were more likely to experience in-hospital stroke, major vascular assess site complications, and major bleeding events (Table 2). They also had longer intensive care unit stays and increased rates of blood transfusions. In-hospital death was higher among nonagenarians (6.5% vs. 4.5%; p < 0.001), and they were also more likely to be discharged to extended care/transitional care unit/rehabilitation or nursing home.
No differences in stroke rate, aortic valve reintervention, or MI were evident after 30 days or 1 year (Table 3, Figures 1 and 2⇓⇓). There was a higher rate of heart failure readmission after 30 days for nonagenarians (5.3% vs. 4.3%; p = 0.014) (Figure 3) but not after 1 year (14.9% vs. 14.5%; p = 0.377). The 30-day and 1-year mortality rates were significantly higher among nonagenarians (age ≥90 years vs. <90 years: 30-day rate: 8.8% vs. 5.9%; p < 0.001; 1-year rate: 24.8% vs. 22.0%; p < 0.001, absolute risk 2.8%, relative risk 12.7%) (Figure 4). Among the small number of patients who were ≥100 years of age (n = 15), 30-day mortality was 0% and 1-year mortality was 6.7%.
After adjusting for multiple demographic and clinical factors, nonagenarians continued to show an increased risk of mortality at 30 days (adjusted HR: 1.46; 95% CI: 1.25 to 1.71; p < 0.001), heart failure at 30 days (adjusted HR: 1.35; 95% CI: 1.11 to 1.65; p = 0.003), and mortality at 1 year (adjusted HR: 1.20; 95% CI: 1.09 to 1.32; p < 0.001). However, as nonagenarians had a higher STS PROM than the younger cohort, the observed to expected mortality ratios were 0.81 (95% CI: 0.70 to 0.92) and 0.72 (95% CI: 0.67 to 0.78), respectively.
Quality of life
Nonagenarians had a lower median 30-day KCCQ-os score compared with younger patients (70.8 vs. 72.9; p = 0.006) but similar KCCQ-os scores at 1 year after TAVR (79.2 vs. 81.3; p = 0.539) (Table 4). After adjusting for STS PROM risk score and baseline KCCQ, nonagenarians had, on average, 3.57-point lower KCCQ-os scores at 30 days but no significant difference at 1 year.
This is the largest study comparing short-term and midterm outcomes of nonagenarians with younger patients undergoing TAVR in U.S. clinical practice. Patient demographics of nonagenarians in this study highlight that nonagenarians who are currently undergoing TAVR represent a highly selected group. Factors that cause a patient to be at higher risk for poor outcomes after TAVR, such as reduced ejection fraction, prior cardiac surgery, and prior stroke, are less commonly present in nonagenarians compared with the younger population; thus, very advanced age is the primary factor for the nonagenarian to be classified as high risk for surgery. Although these patients are at increased risk for morbidity and mortality after TAVR simply on the basis of their age, our findings show that many nonagenarians have good outcomes after TAVR, with prolonged survival and improved QOL, making TAVR reasonable to consider in select nonagenarians with severe AS.
Most of the existing studies report on single-center experiences with small patient cohorts and thus may only provide limited decision-making information on the outcomes of nonagenarians undergoing TAVR (6–9,17). In the largest previous study on TAVR in nonagenarians (346 nonagenarians of 2,254 enrolled patients; reported from the FRANCE-2 [French National Transcatheter Aortic Valve Implantation Registry]), Yamamoto et al. (18) demonstrated a 30-day mortality of 11.2% in nonagenarians without any significant difference compared with patients age 80 to 84 years or 85 to 89 years (18). We found a lower 30-day mortality rate in nonagenarians (8.8%), which may reflect differences in patient selection and improvements in devices and techniques. Although the mortality rate for nonagenarians in our study remained higher than that observed in younger patients (5.9%), this rate may be acceptable, particularly given the QOL benefits observed among surviving nonagenarians. This is further illustrated by the similar observed-to-expected ratios. Of note, the observed nonagenarian mortality rate is not only lower than that previously reported on nonagenarians, but is also comparable to published mortality rates for octogenarians (18,19).
In regard to long-term survival, the FRANCE-2 data showed a trend toward decreased survival in older patients, but the study did not meet statistical significance due to the limited number of patients at risk by 1 year (18). The current study confirms a higher mortality rate in nonagenarians at 1 year, but due to the high mortality rates in both groups, it is doubtful that the measured difference of 2.8% is clinically significant or simply reflects the shorter underlying life expectancy of nonagenarians. Some clarity on this issue can be achieved by comparing our 1-year outcomes to the life expectancy of an age-matched general population. The younger patient cohort showed a 1-year mortality of 22%, which is significantly higher compared with the age-matched general population (6.1% in males and 4.4% in females) (20). In nonagenarians, however, the observed difference between 1-year mortality in TAVR patients (24.8%) and the same-age general population (male 20.5% and female 16.5%) was much smaller (Central Illustration).
Thus, we confirm the promising results from other reports about TAVR in select nonagenarians, and we additionally show that the difference between outcomes of select nonagenarians and younger patients might not be of major clinical relevance.
We further describe the outcomes of a very small cohort of patients ≥100 years of age undergoing TAVR. Overall, 24 centenarians were treated in the TVT registry; thus, this is the largest report on TAVR in this age group. We are aware that our patient cohort (n = 15 with CMS linkage) is too small to give a reliable statement, but nevertheless, the procedural outcomes are encouraging. There was no mortality at 30 days, which is likely indicative of very careful patient selection. The very low 1-year mortality (6.7%) is excellent, especially considering the high 1-year death probability in a 100-year-old patient (36% in males and 31% in females). Of note, Bridges et al. (21), reviewing the STS national database, reported that only 5 centenarians underwent cardiac surgery in the United States between 1997 and 2000.
Increased age alone may heighten the risk of particular post-procedural complications. As vascular complications were 1 of the most common complications, assessing age-related differences is relatively straightforward. The increasing prevalence of vascular complications in very old patients has been demonstrated by several studies, but, because of varying definitions, rates of vascular complications are difficult to compare between studies. Havakuk et al. (22) reported that the rate of minor vascular injuries is increased in patients age ≥85 years (7.5% vs. 16%; p = 0.02), without any difference in major vascular complications (4.3% vs. 2.5%; p = 0.41). However, our data, similar to those of Yamamoto et al. (9), shows higher rates of major vascular complications in the very elderly, but no significant difference in minor vascular injury rates.
Additional in-hospital complications that differed between nonagenarians and younger patients were major bleeding events, need for blood transfusion, and stroke, all of which were higher in the nonagenarian group. The in-hospital stroke rate was higher for nonagenarians in the present study, but we did not detect any significant effect of age on incidence of stroke following TAVR at 30 days or at 1 year (9,18,22). The observed 30-day stroke rate was similar to prior reports on elderly patients undergoing TAVR (6,9,19,22,23). The increased incidences of these in-hospital complications are comprehensively reflected in the longer intensive care unit stays and higher likelihood of discharge to extended care or rehabilitation facilities experienced by nonagenarians.
Although our in-hospital results might suggest worse short-term outcomes for nonagenarians, 30-day and, more importantly, 1-year outcomes are more suitable to determine the appropriateness of TAVR in the nonagenarian age group. However, it is important to note that survival is not the sole factor defining good outcomes in TAVR, especially in the elderly population where survival with reasonable functional capacity and QOL is what matters most. Improvement in QOL should also be included to evaluate whether a procedure should be reasonably offered to nonagenarians.
In a review of the PARTNER (Placement of Aortic Transcatheter Valve) trial, Thourani et al. (24) have previously demonstrated that QOL improves and stabilizes 6 months after TAVR in nonagenarians. Our data confirm these findings. We found that there was a significant increase in KCCQ scores by 30 days, but scores were significantly lower in nonagenarians compared with younger patients. However, there were no differences in QOL between age groups by 1 year after TAVR. These findings suggest that nonagenarians likely recover more slowly after TAVR and thus need more time until the beneficial effect of the procedure is measurable. However, if given time to recover, older patients are able to achieve similar QOL levels as younger patients. This information may be important for patients to know prior to undergoing TAVR, for post-procedure planning and setting realistic expectations of recovery times.
This analysis should be interpreted in light of several important potential limitations. The TVT Registry only captures information on patients receiving commercially approved devices. As several newer TAVR devices are currently under investigation in the United States, our data thus do not represent an all-comers population or the most recent iterations of transcatheter devices. Furthermore, a large number of patients (35%; 8,502 of 24,025) could not be included in long-term outcome analysis due to inability to link with the CMS. Compared with patients without CMS linkage, patients with CMS linkage were more likely to be females; to have prior transient ischemic attack; and to have higher left ventricular ejection fraction, STS PROM, and KCCQ-12 scores. They were less likely to have prior aortic valve procedure, have diabetes, have New York Heart Association functional class III/IV, and be currently on dialysis. Furthermore, there was a high rate of missing KCCQ data (50% at 30-day and 62% at 1-year follow-up). Compared with patients with nonmissing KCCQ-os scores, patients with missing KCCQ scores at follow-up were more likely to be males, have a slower 5-m walk test, have lower left ventricular ejection fraction, have higher STS PROM, have lower baseline KCCQ score, have insulin dependent diabetes, and be on dialysis. It is unclear how inclusion of these patients with missing follow-up data would have altered our findings. However, our study still represents the largest study of nonagenarians who have undergone TAVR and provides much needed data to describe the outcomes of these high-risk patients.
Given their limited life expectancy, it has been unclear whether to perform TAVR in the very elderly, as existing data on outcomes of these patients have been limited. The primary concerns are that nonagenarians might not survive the procedure as frequently, recover from the procedure as quickly, or experience an improved functional outcome and quality of life. We report in current U.S. clinical practice that approximately 16% of patients undergoing TAVR are nonagenarians or older. Although 30-day and 1-year mortality rates were higher in this age group compared with younger TAVR patients, the absolute and relative differences were clinically modest. Furthermore, although nonagenarians generally take longer to recover their physical function and QOL than younger patients, TAVR improved long-term QOL to a similar degree in both age groups. As such, the reported data support both the safety and the efficacy of TAVR in select elderly patients, suggesting that TAVR should not be denied solely on the basis of patient age.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: Although 30-day and 1-year mortality rates are higher in nonagenarians undergoing TAVR, the absolute and relative differences compared with younger patients are clinically modest.
TRANSLATIONAL OUTLOOK: As the indication for TAVR expands to encompass patients at lower risk, more research is needed to clarify which patients, regardless of age, benefit most from the procedure.
The Society of Thoracic Surgeons/American College of Cardiology Transcatheter Valve Therapy (STS/ACC TVT) Registry is an initiative of the Society of Thoracic Surgeons and the American College of Cardiology. This research was supported by the American College of Cardiology’s National Cardiovascular Data Registry (NCDR). The views expressed in this paper represent those of the author(s) and do not necessarily represent the official views of the NCDR or its associated professional societies identified at CVQuality.ACC.org/NCDR. Dr. Szerlip has served as a speaker for Edwards Lifesciences. Dr. Vemulapalli has received research grants from Abbott Vascular, Boston Scientific, and the American College of Cardiology. Dr. Holper has been a consultant to Boston Scientific and Asahi Intecc. Dr. Rumsfeld is Chief Science Officer of the NCDR. Dr. Mack is an uncompensated member of the executive committee of the PARTNER (Placement of Aortic Transcatheter Valve) trial of Edwards Lifesciences. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- aortic stenosis
- confidence interval
- institutional review board
- 12-item Kansas City Cardiomyopathy Questionnaire
- overall summary score of the 12-item Kansas City Cardiomyopathy Questionnaire
- myocardial infarction
- quality of life
- transcatheter aortic valve replacement
- Received November 4, 2015.
- Revision received January 14, 2016.
- Accepted January 19, 2016.
- 2016 American College of Cardiology Foundation
- Vincent G.K.,
- Velkoff V.A.
- Nishimura R.A.,
- Otto C.M.,
- Bonow R.O.,
- et al.
- Vahanian A.,
- Alfieri O.,
- Andreotti F.,
- et al.
- Noble S.,
- Frangos E.,
- Samaras N.,
- et al.
- Spertus J.A.,
- Jones P.G.
- Arnold S.V.,
- Spertus J.A.,
- Lei Y.,
- et al.
- ↵Social Security Administration. Period life table for the Social Security area population. 2011. Available at: https://www.ssa.gov/oact/STATS/table4c6.html. Accessed February 12, 2016.