Author + information
- Received August 7, 2013
- Revision received October 11, 2013
- Accepted October 17, 2013
- Published online March 25, 2014.
- Brian R. Lindman, MD, MSCI∗∗ (, )
- Philippe Pibarot, DVM, PhD†,
- Suzanne V. Arnold, MD, MHA‡,
- Rakesh M. Suri, MD, DPhil§,
- Thomas C. McAndrew, MS‖,
- Hersh S. Maniar, MD∗,
- Alan Zajarias, MD∗,
- Susheel Kodali, MD‖,¶,
- Ajay J. Kirtane, MD, SM‖,¶,
- Vinod H. Thourani, MD#,
- E. Murat Tuzcu, MD∗∗,
- Lars G. Svensson, MD, PhD∗∗,
- Ron Waksman, MD††,
- Craig R. Smith, MD¶ and
- Martin B. Leon, MD‖,¶
- ∗Washington University School of Medicine, St. Louis, Missouri
- †Quebec Heart and Lung Institute, Laval University, Quebec City, Quebec, Canada
- ‡Saint Luke's Mid-America Heart Institute, Kansas City, Missouri
- §Mayo Clinic, Rochester, Minnesota
- ‖Cardiovascular Research Foundation, New York, New York
- ¶Columbia University Medical Center/New York Presbyterian Hospital, New York, New York
- #Emory University School of Medicine, Atlanta, Georgia
- ∗∗Cleveland Clinic Foundation, Cleveland, Ohio
- ††MedStar Washington Hospital Center, Washington, DC
- ↵∗Reprint requests and correspondence:
Dr. Brian R. Lindman, Cardiovascular Division, Washington University School of Medicine, Campus Box 8086, 660 South Euclid Avenue, St. Louis, Missouri 63110.
Objectives The goal of this study was to determine whether a less-invasive approach to aortic valve replacement (AVR) improves clinical outcomes in diabetic patients with aortic stenosis (AS).
Background Diabetes is associated with increased morbidity and mortality after surgical AVR for AS.
Methods Among treated patients with severe symptomatic AS at high risk for surgery in the PARTNER (Placement of Aortic Transcatheter Valve) trial, we examined outcomes stratified according to diabetes status of patients randomly assigned to receive transcatheter or surgical AVR. The primary outcome was all-cause mortality at 1 year.
Results Among 657 patients enrolled in PARTNER who underwent treatment, there were 275 patients with diabetes (145 transcatheter, 130 surgical). There was a significant interaction between diabetes and treatment group for 1-year all-cause mortality (p = 0.048). Among diabetic patients, all-cause mortality at 1 year was 18.0% in the transcatheter group and 27.4% in the surgical group (hazard ratio: 0.60 [95% confidence interval: 0.36 to 0.99]; p = 0.04). Results were consistent among patients treated via transfemoral or transapical routes. In contrast, among nondiabetic patients, there was no significant difference in all-cause mortality at 1 year (p = 0.48). Among diabetic patients, the 1-year rates of stroke were similar between treatment groups (3.5% transcatheter vs. 3.5% surgery; p = 0.88), but the rate of renal failure requiring dialysis >30 days was lower in the transcatheter group (0% vs. 6.1%; p = 0.003).
Conclusions Among patients with diabetes and severe symptomatic AS at high risk for surgery, this post-hoc stratified analysis of the PARTNER trial suggests there is a survival benefit, no increase in stroke, and less renal failure from treatment with transcatheter AVR compared with surgical AVR. (The PARTNER Trial: Placement of AoRTic TraNscathetER Valve Trial; NCT00530894)
Diabetes mellitus adversely affects morbidity and mortality for all types of cardiovascular diseases (1,2). In patients with aortic stenosis (AS), diabetes is associated with increased hypertrophic remodeling, decreased left ventricular function, and worse heart failure symptoms (3,4). Diabetes has also been associated with increased morbidity and mortality after surgical aortic valve replacement, even after adjustment for comorbidities such as vascular disease and renal dysfunction (5,6). The mechanisms for this additional surgical risk are not completely known, although it is hypothesized that the inflammation, oxidative stress, and reperfusion injury induced by cardioplegia and cardiopulmonary bypass are particularly harmful in the setting of diabetes and hypertrophic ventricular remodeling from chronic pressure overload due to AS, thereby causing adverse short- and long-term consequences (7–13). As such, a less-invasive method of valve replacement that avoids the injurious effects of cardiopulmonary bypass may lead to improved clinical outcomes among these high-risk patients with diabetes. Accordingly, we examined the clinical outcomes of patients at high risk for surgery enrolled in the PARTNER (Placement of Aortic Transcatheter Valve) trial to evaluate whether outcomes varied according to diabetes status after treatment with transcatheter versus surgical aortic valve replacement (14).
The design, inclusion and exclusion criteria, and primary results of the high-risk cohort (cohort A) of the PARTNER trial have been reported (14). These patients were at high surgical risk as defined by a predicted risk of death ≥15% by 30 days after surgery. After evaluation of vascular anatomy, patients were included in either the transfemoral placement cohort or the transapical placement cohort and randomized to undergo transcatheter therapy with the Edwards SAPIEN heart valve system (Edwards Lifesciences, Irvine, California) or surgical aortic valve replacement. Some patients did not undergo their assigned procedure due to death, refusal, study withdrawal, and/or pre-treatment clinical deterioration. For the current analysis, we included only patients who were randomized to and received the assigned treatment (as-treated population). The diagnosis of diabetes and other clinical characteristics were determined by the enrolling sites. The study protocol was approved by the institutional review board at each enrolling site, and all patients provided written informed consent.
Clinical events, including death (all-cause and cardiac), repeat hospitalizations, stroke, renal failure, major bleeding, myocardial infarction, and vascular complications, were adjudicated by a clinical events committee. The primary endpoint of the PARTNER trial and our analysis was all-cause death at 1 year. A detailed report of the classification of deaths among the diabetic and nondiabetic patients treated with transcatheter or surgical aortic valve replacement in the transfemoral and transapical placement cohorts is provided in Online Table 1. Repeat hospitalizations were defined as hospitalization resulting from symptoms of AS (valve-related deterioration, including heart failure, angina, or syncope) or complications of the valve procedure. Stroke was defined as a focal neurological deficit lasting ≥24 h or a focal neurological deficit lasting <24 h with imaging findings of acute infarction or hemorrhage. Renal failure events were defined as the need for dialysis of any sort (hemodialysis, continuous venovenous hemodialysis, peritoneal). Further details on clinical events definitions are provided in Online Table 2. Many of these clinical event definitions are consistent with the VARC-2 (Valve Academic Research Consortium–2) definitions (e.g., cardiac death, stroke, myocardial infarction), but others differ substantially (e.g., renal failure, major bleeding) (15). An independent core laboratory analyzed all echocardiograms (16). The presence and severity of post-procedural prosthesis–patient mismatch and aortic regurgitation were determined according to VARC-2 criteria (15). The Kansas City Cardiomyopathy Questionnaire (KCCQ), a disease-specific health status measure of heart failure, was used to assess health status (17,18).
Continuous variables are summarized as mean ± SD or median (quartile), and they were compared by using the Student t test or Mann-Whitney rank sum test as appropriate. Categorical variables were compared by using the chi-square or Fisher exact test. Survival curves for time-to-event variables, based on all available follow-up data, were performed with the use of Kaplan-Meier estimates and were compared between groups with the use of the log-rank test. Cox proportional hazards models were used to calculate hazard ratios (HRs) and to test for interactions. KCCQ overall summary scores were compared by using analysis of covariance to adjust for baseline differences in KCCQ scores between groups. All statistical analyses were performed by using SAS version 9.2 (SAS Institute, Inc., Cary, North Carolina).
Among the 699 patients enrolled in the PARTNER trial cohort A, 657 patients were randomized to and received transcatheter or surgical therapy; 313 patients were treated with surgery and 344 patients were treated with transcatheter therapy. Among the as-treated population, 275 (42%) patients had diabetes, 145 in the transcatheter group (103 transfemoral, 42 transapical) and 130 in the surgery group (88 transfemoral, 42 transapical cohort). Among the 382 patients without diabetes, 199 were treated with transcatheter valve replacement (137 transfemoral, 62 transapical) and 183 were treated with surgery (133 transfemoral, 50 transapical).
The clinical characteristics and medication usage of patients in the trial with and without diabetes differed in ways that would be expected based on diabetes status (Online Table 3). Within each subgroup of patients (diabetic and nondiabetic), the clinical characteristics were generally well matched between those who received transcatheter versus surgical valve replacement (Table 1, Online Table 4).
Stratified analyses based on diabetes status were performed for several important clinical outcomes at 1 year. There was a significant interaction between diabetes status and all-cause mortality (interaction p = 0.048) (Fig. 1). Among the patients with diabetes, 1-year all-cause mortality was 18.0% in transcatheter-treated patients versus 27.4% in the surgically treated patients (HR: 0.60 [95% confidence interval (CI): 0.36 to 0.99]; p = 0.044) (Figs. 1 and 2A). The Kaplan-Meier survival curves for the transfemoral placement cohort (Fig. 2B) and the transapical placement cohort (Fig. 2C) demonstrate a consistent relationship of lower all-cause mortality for transcatheter-treated versus surgically treated diabetic patients compared with the overall population of diabetic patients (Fig. 2A).
At 6 months, all-cause mortality was lower in transcatheter-treated diabetic patients compared with surgically treated diabetic patients (10.3% vs. 23.4%; HR: 0.41 [95% CI: 0.22 to 0.76]; p = 0.003) (Fig. 2A). At 2 years, the survival benefit observed at 6 months and 1 year from transcatheter treatment compared with surgical treatment in diabetic patients was no longer significant (HR: 0.76 [95% CI: 0.49 to 1.19]; p = 0.23) (Online Fig. 1A).
The rates of stroke were similar between transcatheter-treated and surgically treated diabetic patients at 30 days (3.5% vs. 2.4%; p = 0.58) and 1 year (3.5% vs. 3.5%; p = 0.88) (Table 2). At 1 year, there was a decreased rate of renal failure requiring dialysis with transcatheter therapy compared with surgical therapy (4.2% vs. 10.6%; p = 0.05), particularly dialysis lasting >30 days (0.0% vs. 6.1%; p = 0.003). Similar to the main trial results, there was an increased risk of major bleeding with surgery among diabetic subjects but an increased risk of major vascular complications with transcatheter therapy at 30 days and 1 year (p < 0.05 for all relationships).
In terms of echocardiographic symptoms and laboratory findings, the incidence of post-operative mild and moderate or severe total aortic regurgitation was higher in diabetic patients treated with transcatheter therapy compared with surgery (Table 3). There was a trend toward a lower incidence of moderate or severe prosthesis–patient mismatch at 30 days with transcatheter therapy, whereas left ventricular mass was lower at 30 days in surgically treated patients. A lower incidence of New York Heart Association functional class III or IV heart failure symptoms, better quality of life, and longer 6-min walk distance were observed at 30 days in diabetic patients treated with transcatheter therapy compared with those receiving surgery; there were no significant between-group differences at 6 months or 1 year (Table 4). Post-procedural troponin levels and white blood cell counts were higher in diabetic subjects treated with surgery compared with those receiving transcatheter therapy (Table 3).
There was no difference in 1-year all-cause mortality in nondiabetic subjects treated with transcatheter versus surgical therapy (Figs. 3A and 3B); however, there was a trend toward increased mortality in the transapical placement cohort from transcatheter therapy compared with surgery (Fig. 3C). A trend toward a higher risk of stroke was observed in nondiabetic patients treated with transcatheter therapy compared with those receiving surgery at 1 year (7.6% vs. 2.8%; HR: 2.60 [95% CI: 0.94 to 7.22]; p = 0.056) (Fig. 1). The rates of repeat hospitalization and renal failure among nondiabetic patients were similar in the 2 treatment groups. At 2 years, there was also no difference in all-cause mortality in nondiabetic subjects treated with transcatheter versus surgical therapy (Online Fig. 2).
We report for the first time, in a post-hoc stratified analysis of the high-risk patients enrolled in the PARTNER trial, a differential response to transcatheter versus surgical treatment based on diabetes status. Although the PARTNER trial demonstrated similar rates of death at 1 year in those treated with transcatheter or surgical therapy for the overall population, we found that diabetic patients who were treated with transcatheter aortic valve replacement had a 9% lower absolute risk of 1-year all-cause mortality and a 40% lower hazard of death over the first year after the procedure compared with diabetic patients treated with surgical valve replacement. Furthermore, diabetic patients treated with transcatheter therapy had a similar rate of stroke and lower incidence of renal failure compared with those treated with surgery. These findings have important clinical implications for the treatment of patients with severe AS and diabetes who are at high risk for surgery.
Both transcatheter and surgical valve replacement relieve left ventricular pressure overload from AS by treating the mechanical obstruction of the valve. Among the overall population, the PARTNER trial demonstrated that survival at 1 year was similar with transcatheter and surgical valve replacement for patients with severe symptomatic AS at high risk for surgery. However, there may be subgroups of patients who will do better with 1 approach than the other. As we gain more experience with these 2 treatment options, we will learn how to individualize treatment strategies based on a variety of potential factors to obtain the best clinical results. Our study raises the intriguing possibility that transcatheter valve replacement may be the preferred approach for diabetic patients with severe symptomatic AS who are at high surgical risk.
There is considerable interest in comparing less-invasive transcatheter or percutaneous therapies with surgery for a variety of cardiovascular problems, including valve disease and coronary, aortic, carotid, and peripheral vascular disease, particularly in diabetic patients (19–23). These comparisons involve differences both in which therapy is provided (e.g., stent vs. bypass graft) and how it is provided (e.g., catheter-based vs. open surgery). When comparing transcatheter versus surgical aortic valve replacement, there is relatively little difference in which therapy is provided. In both cases, the mechanical valve obstruction is treated by the placement of a new valve that relieves the pressure overload on the ventricle. Nonetheless, differences in how well the implanted valve opens the previously restricted orifice (effective orifice area) and how much it leaks could affect outcomes. In contrast, there are more obvious differences in how the therapy is provided, which we suspect underlies the difference in survival among diabetic patients between the 2 treatment groups. In the case of a transcatheter approach, there is rapid ventricular pacing with large sheaths introduced into the major vessels and/or heart, whereas with surgery, there are the injurious effects of cardiopulmonary bypass, cardioplegia, and reperfusion.
Among diabetic patients, the survival curves between the transcatheter and surgical treatment groups separated soon after valve replacement and continued to diverge until approximately 6 months, after which the curves moved modestly toward each other; by 2 years, there was no significant difference in survival between the 2 treatment groups. We hypothesize that this relationship is due to the short-term benefit of a less-invasive approach to replace the valve that avoids cardiopulmonary bypass, which is mitigated over time by nonprocedure-related factors and the known deleterious effects of increased aortic regurgitation after transcatheter valve replacement. In the PARTNER trial, both in the whole population and the subgroup with diabetes, there was a much greater incidence of mild, moderate, and severe aortic regurgitation in the transcatheter treatment group compared with surgery, which is associated with increased all-cause mortality (24). A potential implication is that if the incidence of aortic regurgitation after transcatheter aortic valve replacement is reduced, the early substantial survival benefit of transcatheter valve replacement in diabetic patients may be sustained beyond the first year.
Other observations from this analysis merit further study. Although not the focus of our analysis, the rate of all-cause mortality at 1 year was lower in diabetic patients compared with nondiabetic patients treated with transcatheter therapy. Diabetes is known to adversely affect morbidity and mortality for all types of cardiovascular disease and adversely influence post-procedural outcomes after percutaneous and surgical procedures (1,2,25,26). As such, this result was somewhat surprising. However, it should be noted that there were numerous baseline clinical differences between the diabetic and nondiabetic patients (Online Table 3), which could confound this comparison. In particular, diabetic patients had a much larger body mass index than nondiabetic patients, which was expected. In the PARTNER trial, higher body mass index had an independent protective effect in the transcatheter group but not in the surgical group. This finding may explain, at least in part, the unexpected observation of lower mortality in diabetic patients compared with nondiabetic patients in the transcatheter group. This hypothesis-generating observation of an apparent “diabetes paradox” requires further study and careful adjustment for confounders.
Diabetes is characterized by a milieu of hyperglycemia, insulin resistance, and increased nonesterified fatty acids, which contribute to oxidative stress, lipotoxicity, advanced glycation end products, and altered calcium handling and substrate metabolism (11). Surgical valve replacement involves cardioplegia, cardiopulmonary bypass, and reperfusion injury, which may cause more inflammation, oxidative stress, and myocardial ischemia/injury than with the rapid ventricular pacing performed during transcatheter therapy (7–10,12,13). In patients with diabetes, this can intensify an already existing deleterious myocardial and systemic environment, which may have important short- and long-term adverse consequences for cardiac performance and clinical outcomes after valve replacement. Recently, Sinning et al. (27) demonstrated that the development of systemic inflammatory response syndrome during the first 48 h after transcatheter aortic valve replacement is associated with increased 30-day and 1-year mortality. We speculate that surgical valve replacement may be associated with an increased incidence of systemic inflammatory response syndrome compared with transcatheter replacement. Consistent with this possibility, the 24-h post-procedure blood analyses drawn in the PARTNER trial showed higher levels of white blood cells in patients with diabetes after surgical aortic valve replacement compared with transcatheter valve replacement. The 24-h post-procedure cardiac enzyme levels were also higher in the diabetic patients treated with surgery, suggesting increased ischemic injury compared with a transcatheter approach. Other mechanisms whereby transcatheter therapy may confer a survival benefit in diabetic patients include less prosthesis–patient mismatch and less post-procedural renal failure requiring dialysis, both of which have a known adverse impact on clinical outcomes (16,24). However, ultimately the mechanisms underlying the survival benefit from a transcatheter valve replacement in diabetic patients require further investigation. These mechanisms include the impact of insulin and/or oral diabetic medical treatments and how the metabolic syndrome and diabetes separately and in combination influence outcomes in diabetic patients undergoing transcatheter or surgical valve replacement.
One limitation of our study was that diabetes status was not a pre-specified subgroup analysis. As such, these results should be considered hypothesis-generating and need to be confirmed in future research. However, given the relatively low power to demonstrate superiority of transcatheter replacement over surgical replacement in a subgroup analysis, the statistically significant survival benefit is noteworthy and should encourage further evaluation. Second, the diagnosis of diabetes was determined by enrolling sites and was not verified by using other mechanisms. However, the differences observed between diabetic and nondiabetic patients in the PARTNER trial with respect to baseline clinical characteristics and medication usage are consistent with those that would be expected based on the presence or absence of diabetes. Furthermore, we did not have reliable information on diabetic medication usage (insulin and/or oral medications) or access to data on the severity or duration of diabetes, microvascular complications, or glucose control. How each of these factors contributes to the treatment effect of transcatheter versus surgical aortic valve replacement will require further study. However, by including patients with mild (recent onset, diet controlled, or oral medications only) as well as severe (long-standing, requiring insulin) diabetes, we were less likely to disprove the null hypothesis that survival would be similar between the transcatheter and surgical treatment groups.
Diabetes is associated with increased morbidity and mortality in patients with AS undergoing surgical valve replacement. In a post-hoc stratified analysis of the PARTNER trial in which high-risk patients were randomized to undergo transcatheter or surgical aortic valve replacement, we found that diabetic patients had a survival benefit at 1 year with no increased risk of stroke and less renal failure when treated with transcatheter valve replacement compared with surgery. These results suggest that transcatheter aortic valve replacement may be the preferred treatment approach for patients with AS and diabetes who are at high risk for surgery. Confirmation of these findings, particularly in lower-risk populations, is needed, as well as insights into the underlying mechanisms for the observed survival benefit.
For supplemental tables and figures on the study protocol, please see the online version of this article.
The PARTNER trial was funded by Edwards Lifesciences, and the protocol was developed jointly by the sponsor and study steering committee. The current analysis was conducted by academic investigators with no additional funding from Edwards Lifesciences. Dr. Lindman is supported by K23 HL116660 and the Washington University Institute of Clinical and Translational Sciences grant (UL1 TR000448, KL2 TR000450) from the National Center for Advancing Translational Sciences of the National Institutes of Health; and is a site coinvestigator for the PARTNER Trial. Dr. Suri's institution (Mayo Clinic) receives randomized aortic valve replacement trial funding to the Division of Cardiovascular Surgery from Edwards Lifesciences, St. Jude Medical, and Sorin Medical; and he is a national Principal Investigator for the PERCEVAL Trial (Sorin Medical), on the steering committee for the Portico Trial (St. Jude Medical), and coinvestigator for the PARTNER II (Edwards Lifesciences) and COAPT (Abbott) trials. Dr. Zajarias is a member of the PARTNER trial steering committee; is site Principal Investigator for the PARTNER trial; and is a consultant for Edwards Lifesciences. Dr. Kodali is a member of the PARTNER trial steering committee and consultant for Edwards Lifesciences, a member of the steering committee for the Portico trial (St. Jude Medical), and a member of the scientific advisory board of Thubrikar Aortic Valve. Dr. Thourani is a member of the PARTNER trial steering committee; and is a consultant for Edwards Lifesciences, Sorin Medical, St. Jude Medical, and DirectFlow. Dr. Waksman is a member of the Speakers' Bureau of Boston Scientific, Medtronic, AstraZeneca, Biotronik, and Abbott Vascular. Drs. Tuzcu, Svensson, Smith, and Leon are unpaid members of the PARTNER Executive Committee; and has received travel reimbursements from Edwards Lifesciences for activities related to these positions. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- aortic stenosis
- aortic valve replacement
- confidence interval
- hazard ratio
- Kansas City Cardiomyopathy Questionnaire
- Received August 7, 2013.
- Revision received October 11, 2013.
- Accepted October 17, 2013.
- American College of Cardiology Foundation
- Mak K.H.,
- Topol E.J.
- Lindman B.R.,
- Arnold S.V.,
- Madrazo J.A.,
- et al.
- Falcao-Pires I.,
- Hamdani N.,
- Borbely A.,
- et al.
- Ascione R.,
- Caputo M.,
- Gomes W.J.,
- et al.
- Dybdahl B.,
- Wahba A.,
- Lien E.,
- et al.
- Poornima I.G.,
- Parikh P.,
- Shannon R.P.
- Suleiman M.S.,
- Hancock M.,
- Shukla R.,
- Rajakaruna C.,
- Angelini G.D.
- Kappetein A.P.,
- Head S.J.,
- Généreux P.,
- et al.
- Hahn R.T.,
- Pibarot P.,
- Stewart W.J.,
- et al.
- Reynolds M.R.,
- Magnuson E.A.,
- Wang K.,
- et al.
- Arnold S.V.,
- Spertus J.A.,
- Lei Y.,
- et al.
- Chaitman B.R.,
- Rosen A.D.,
- Williams D.O.,
- et al.
- Kip K.E.,
- Faxon D.P.,
- Detre K.M.,
- Yeh W.,
- Kelsey S.F.,
- Currier J.W.
- Sinning J.M.,
- Scheer A.C.,
- Adenauer V.,
- et al.