Share this article on:

De Novo Aortic Insufficiency During Long-Term Support on a Left Ventricular Assist Device: A Systematic Review and Meta-Analysis

Deo, Salil V.*; Sharma, Vikas; Cho, Yang Hyun,‡; Shah, Ishan K.§; Park, Soon J.

doi: 10.1097/MAT.0000000000000042
Adult Circulatory Support

Aortic insufficiency (AI) may occur while supported on a left ventricular assist device (LVAD). We conducted a systematic review to determine the incidence, predictors, and consequences of AI during LVAD support. MEDLINE was searched for original studies presenting clinical data regarding patients who developed AI during LVAD implant. Seven observational studies (657 patients) were selected for review; 65% of patients underwent implantation with a continuous-flow device (Cf-LVAD). The incidence of AI was 25% (11–42%) (Support period: 412 ± 281 days). AI increased by 4% (1–6%) per month of support (p < 0.01). AI-positive patients were older at implant (weighted mean difference, 7.7 [4.3; 11.1]; p < 0.01). Female sex (0.002 ± 0.001; p = 0.01) and smaller body surface area (−0.003 ± 0.001 per m2; p < 0.01) correlated with progressive AI. Destination therapy patients (odds ratio [OR], 5.3 [1.2, 24]; p = 0.02) and those with Cf-LVAD pumps were likely to develop AI (hazard ratio [HR], 2.2 [1.2, 3.8]; p < 0.01). A closed aortic valve was associated with AI (OR, 4.7 [1.9, 11.8]; p < 0.01). Survival was comparable in both cohorts (HR, 1.5 [0.81, 2.8]; p = 0.2). A significant number of patients develop de novo AI during LVAD support. Advanced age, longer support duration, continuous-flow pumps, and a closed aortic valve are associated with AI. Large cohort studies would improve our understanding of this condition.

From the *Adventist Wockhardt Heart Institute, Surat, Gujarat, India; Division of Cardiovascular Surgery, Mayo Clinic, Rochester, Minnesota; Division of Cardiovascular Surgery, Samsung Medical Center, Sungkyunkwan School of Medicine, Seoul, South Korea; and §Department of Surgery, University of Minnesota, Minneapolis, Minnesota.

Submitted for consideration September 2013; accepted for publication in revised form November 2013.

Disclosure: Dr. SJP is on the advisory committee for Thoratec Corporation. Other authors do not have any disclosures to declare.

Reprint Requests: Soon J. Park, MD, Division of Cardiovascular Surgery, Case Medical Center, Case Western Reserve University, Cleveland OH 44106. Email:

Left ventricular assist device (LVAD) therapy is now a standard surgical therapy for patients with end-stage congestive heart failure.1 Initially introduced as pulsatile devices, the present new-generation pumps are based on a continuous-flow physiology. Initially, the device was being implanted as bridge to transplant (BTT). However, the recent approval of LVAD as destination therapy (DT) has resulted in an almost exponential increase in the LVAD recipients over the past years. For these DT patients, LVAD-related adverse events gain even more importance as a bailout strategy is unavailable.

Progressive aortic insufficiency (AI) has been reported to occur in LVAD-supported patients, especially as the duration of support increases. However, available data regarding natural history and predictors of AI are limited to single institutional experiences. Opinion regarding the impact of de novo AI on long-term survival is divided. Hence, we conducted a systematic review of studies reporting the occurrence of de novo AI in patients on LVAD support, with the aim of analyzing patient and pump factors related to this clinical sequela of LVAD support.

Back to Top | Article Outline


A systematic review of published English-language literature was conducted in MEDLINE using the terms “left ventricular assist device,” “de novo aortic insufficiency,” “continuous-flow pumps,” “pulsatile pumps,” “aortic regurgitation,” and “natural history” with Boolean operators to identify original articles fulfilling the following inclusion criteria:

1. Human subjects having de novo AI while on intracorporeal or paracorporeal LVAD support, either pulsatile or continuous flow.

2. Study will present the demographic parameters of patients with and without AI or it will present adverse clinical events occurring in both cohorts and predictors or it will present a survival analysis of these cohorts.

Articles that reported their experience with concomitant aortic valve surgery during LVAD implant were excluded as our main focus was to investigate the natural history of progressive AI after LVAD implant and its impact on survival. Case reports, small case series (less than 10 patients), review articles, and letters to the editor were excluded.

Two authors (VS, SVD) independently screened abstracts; full-text articles were analyzed to determine fitness for inclusion. Disagreements were resolved by consensus.

Back to Top | Article Outline

Statistical Analysis

Statistical analysis was performed using R Individual proportions were pooled using the Freeman Tukey double arcsine method available in the “meta” package for R.3 The Peto method was implemented for obtaining the pooled odds ratio (OR) as it is the preferred method in the case of low event rates.4 If any one cell of the contingency table reported zero events, a small constant was added to all the cells to ensure inclusion of the study in the pooled estimate.5 However, studies where both arms had no events were not included in the meta-analysis. Hazard ratios (HRs) were obtained from individual studies; if not available, they were calculated by implementing methods described by Parmar et al.6 and simplified by Tierney et al.7 The pooled estimate was obtained using the inverse variance weighting method.4 Results are reported using a 95% confidence interval and p-value <0.05 is considered statistically significant. Continuous data were compared using the weighted mean difference (WMD). Given the small number of studies (less than 10), publication bias was not assessed.

Back to Top | Article Outline


From the initial 5,257 search results generated, further refinement in search criteria determined 112 abstracts, which were then screened. The detailed flowchart of the search protocol is presented in Figure 1. Seven articles8–14 (657 patients) discussed the development of de novo AI after LVAD implant. All articles were retrospective single-center studies prepared by analysis of their respective hospital LVAD databases. From the pooled 657 patients, 428 (65%) underwent implantation with a continuous-flow device while the remaining 35% were implanted the pulsatile device. Four studies implemented both continuous-flow LVAD (Cf-LVAD) and pulsatile-flow LVAD (Pf-LVAD),9–12 two reported results with Cf-LVAD,8,14 and one study from Japan discussed their results with the TOYOBO device.13 HeartMate II (Thoratec Corp, Pleasanton, CA) was the most common Cf-LVAD implanted. Two studies from Japan10,13 implemented the TOYOBO device, a commercially approved paracorporeal pulsatile LVAD. Table 1 presents the pooled results of the demographic parameters of patients included in these studies. Some studies failed to present differences in demographic parameters between AI-positive and AI-negative patients.

The pooled incidence of significant AI was 25% (11–42%) during the support period of 412 ± 281 days (Figure 2). This result demonstrated significant heterogeneity. A meta-regression analysis demonstrated significant association with duration of support (p = 0.0006) but not with age (p = 0.92). Increase in the LVAD support for a month increased the proportion of AI by 4% (1–6%).

Back to Top | Article Outline

Preoperative Variables and Progressive AI

Patients who underwent LVAD implantation at a later age appeared to develop AI over the study period. Patients developing AI (57 ± 17 years) were much older than those who did not develop AI (49 ± 12 years). A pooled analysis of four studies8,10,13,14 demonstrated significant difference in the age of the two cohorts (WMD = 7.7 [4.35, 11.08]; p < 0.0001) (Figure 3). Female sex (0.002 ± 0.001, p = 0.01) and smaller body surface area (−0.003 ± 0.001 per m2, p = 0.0017) also correlated with higher AI during follow-up.9 Mitral regurgitation (OR, 7.6 [1–43.9]; p = 0.04) was also associated with progressive AI.13

Back to Top | Article Outline

Factors Related to LVAD Support

Implant as an LVAD as DT8,14 was associated with a higher odds (5.32 [1.18, 23.95]; p = 0.02) of developing AI (Figure 4). A pooled analysis of four studies9–12 demonstrated that patients with continuous-flow pumps were more likely to develop AI compared to the patients with pulsatile pumps (OR, 2.15 [1.21, 3.82]; p = 0.0083; I2 = 38%) (Figure 5). Aggarwal et al.8 demonstrated that the mean LVAD speed of both (AI+ and AI−) cohorts was comparable (9,327 ± 1,299 vs. 9,099 ± 1,279; p = 0.34).

Pak et al.11 demonstrated that patients who developed AI had slightly larger aortic root diameters before LVAD implant (3.43 ± 0.43 vs. 3.15 ± 0.40; p = 0.06). The authors also reported a positive correlation between aortic sinus diameter (0.04 ± 0.01/mm; p = 0.001) and progressive AI.9 On a pooled analysis of 322 patients,8,11,13,14 we found that a predominantly closed aortic valve was strongly associated with AI (OR, 4.77 [1.94, 11.73]; p = 0.0007) (Figure 6).

Back to Top | Article Outline

Impact of AI on Long-Term Survival

A pooled analysis of HRs from three studies8,12,13 demonstrated that overall survival was comparable in both cohorts (HR, 1.52 [0.81, 2.85]; p = 0.19; I2 = 77%) (Figure 7). Four studies8,10–12 presented data regarding disposition of patients with significant AI. Surgery for AI was performed in 6 of 74 (8.1%) while 17 of 74 (23%) underwent heart transplantation.

Back to Top | Article Outline


As an established therapy for end-stage heart failure, LVAD is being increasingly implemented as DT in older patients.

Our systematic review of seven single-center institutional studies (657 patients) demonstrates that age, aortic valve closure, implant as DT, and duration of implant were important factors associated with the production of de novo AI. Continuous flow devices were more prone to develop AI rather than pulsatile devices.

The important relationship between age and AI has been demonstrated since the Framingham Study.15,16 The LVAD physiology, especially with continuous flow, hastens the age-related degeneration of the valve. The relative inelasticity of the aged aorta, increased sheer stress, and high diastolic luminal pressures adversely affect the valve leaflets as well as the support mechanism. AI is associated with increased medial disruption and degeneration,17 which would naturally lead to poor valve support and progressive AI. Both position and angle of the outflow graft and native aorta are important in determining hemodynamic alterations in the ascending aorta.18,19 Commissural fusion of the aortic valve occurs in patients supported on a Cf-LVAD.20 Martina et al.21 demonstrated that AI was present in all patients with commissural fusion and 67% patients without this abnormality. Importantly, development of this pathology was not related to the age of the patient or the duration of LVAD support.21 This demonstrates the importance of maintaining a pulsatile circuit with at least partial valve opening with each systole.

Our pooled analysis demonstrates that continuous-flow physiology is an important factor in the development of AI. This is highlighted by two important findings: 1) patients who had relatively closed aortic valves were more prone to develop AI and 2) AI was more often seen in patients with Cf-LVAD compared to pulsatile pumps. Partial support was associated with lower amount of commissural pathology.21 Thus, while we are increasingly using Cf-LVADs for their easier implant abilities and improved engineering, we need to take a close look at optimal device settings. Maintaining adequate cardiac output with some native ventricular ejection should be our goal. The Jarvik device does transiently reduce speed to promote valve opening. Reports regarding the long-term clinical benefit of this unique setting are still awaited.

A pooled analysis of three studies failed to demonstrate any difference in long-term survival in both cohorts. While that may be true, a significant number of AI-positive patients (24%) underwent transplantation as a bailout procedure. Only one study in this analysis consisted of patients with a large DT cohort, while other studies predominantly contained a BTT population.8 With a 70% survival on Cf-LVAD therapy at the end of 2 years, the number of LVAD supported with progressive AI is going to increase.22 A multi-institutional prospective study is needed to help us to define the incidence, risk factors, and clinical importance of condition.

Park et al.23 first reported a simple technique of aortic valve repair at the time of LVAD implant, which has been adopted by others with minor modifications.24 A recent retrospective review (with a follow-up > 2 years) demonstrates the Park stitch to be an effective, simple, and convenient solution for moderate or more AI at the time of LVAD.25 In our institution, we routinely perform aortic valve repair for patients with moderate or more AI at the time of LVAD implant.

A large multi-institutional prospective study will help improve our understanding of this clinical condition and guide us regarding the appropriate use of concomitant aortic valve repair during LVAD implant.

Back to Top | Article Outline


We have conducted a meta-analysis of observational data. All included articles are from single-center institutional data; some variation in patient selection, operative technique and, most importantly, LVAD settings are bound to be present. To account for some bias, we have adopted a random effects model and conducted a meta-regression in the presence of high heterogeneity. Publication bias was not assessed due to the small number of studies included in the meta-analysis.

Back to Top | Article Outline


Our study is the first pooled analysis discussing de novo AI after LVAD implant. The assimilation and analysis of studies with small individual event rates empower us to attain statistical significance. In spite of certain limitations, we feel that our study addresses a very important aspect of LVAD therapy, especially as we enter the era of DT.

Back to Top | Article Outline


A significant number of patients develop de novo AI during LVAD support. Advanced age, longer support duration, continuous-flow pumps, and a closed aortic valve were associated with AI. While AI did not adversely impact long-term survival, bailout heart transplantation was often needed. Further large cohort studies are needed to improve our understanding of this clinical condition.

Back to Top | Article Outline


1. Holman WL, Naftel DC, Eckert CE, et al. Durability of left ventricular assist devices: Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) 2006 to 2011. J Thorac Cardiovasc Surg. 2013;146:437–441
2. Team RC. R: A Language and Environment for Statistical Computing. 2013 Vienna, Austria R Foundation for Statistical Computing
3. Schwarzer G Meta-analysis with R version 3.0.2. Available at: Accessed February 4, 2014.
4. Higgins JPT, Green S The Cochrane Handbook for Systematic Reviews of Interventions, version 5.1.0. 2011 The Cochrane Collaboration Available at: Accessed February 4, 2014.
5. Friedrich JO, Adhikari NK, Beyene J. Inclusion of zero total event trials in meta-analyses maintains analytic consistency and incorporates all available data. BMC Med Res Methodol. 2007;7:5
6. Parmar MK, Torri V, Stewart L. Extracting summary statistics to perform meta-analyses of the published literature for survival endpoints. Stat Med. 1998;17:2815–2834
7. Tierney JF, Stewart LA, Ghersi D, Burdett S, Sydes MR. Practical methods for incorporating summary time-to-event data into meta-analysis. Trials. 2007;8:16
8. Aggarwal A, Raghuvir R, Eryazici P, et al. The development of aortic insufficiency in continuous-flow left ventricular assist device-supported patients. Ann Thorac Surg. 2013;95:493–498
9. Cowger J, Pagani FD, Haft JW, Romano MA, Aaronson KD, Kolias TJ. The development of aortic insufficiency in left ventricular assist device-supported patients. Circ Heart Fail. 2010;3:668–674
10. Hatano M, Kinugawa K, Shiga T, et al. Less frequent opening of the aortic valve and a continuous flow pump are risk factors for postoperative onset of aortic insufficiency in patients with a left ventricular assist device. Circ J. 2011;75:1147–1155
11. Pak SW, Uriel N, Takayama H, et al. Prevalence of de novo aortic insufficiency during long-term support with left ventricular assist devices. J Heart Lung Transplant. 2010;29:1172–1176
12. Rajagopal K, Daneshmand MA, Patel CB, et al. Natural history and clinical effect of aortic valve regurgitation after left ventricular assist device implantation. J Thorac Cardiovasc Surg. 2013;145:1373–1379
13. Toda K, Fujita T, Domae K, Shimahara Y, Kobayashi J, Nakatani T. Late aortic insufficiency related to poor prognosis during left ventricular assist device support. Ann Thorac Surg. 2011;92:929–934
14. Soleimani B, Haouzi A, Manoskey A, Stephenson ER, El-Banayosy A, Pae WE. Development of aortic insufficiency in patients supported with continuous flow left ventricular assist devices. ASAIO J. 2012;58:326–329
15. Akasaka T, Yoshikawa J, Yoshida K, et al. Age-related valvular regurgitation: A study by pulsed Doppler echocardiography. Circulation. 1987;76:262–265
16. Singh JP, Evans JC, Levy D, et al. Prevalence and clinical determinants of mitral, tricuspid, and aortic regurgitation (the Framingham Heart Study) Am J Cardiol. 1999;83:897–902
17. Benedik J, Pilarzcyk K, Wendt D, et al. Is there any difference in aortic wall quality between patients with aortic stenosis and those with regurgitation? Eur J Cardiothorac Surg. 2013;44:754–759
18. May-Newman K, Hillen B, Dembitsky W. Effect of left ventricular assist device outflow conduit anastomosis location on flow patterns in the native aorta. ASAIO J. 2006;52:132–139
19. May-Newman KD, Hillen BK, Sironda CS, Dembitsky W. Effect of LVAD outflow conduit insertion angle on flow through the native aorta. J Med Eng Technol. 2004;28:105–109
20. Letsou GV, Connelly JH, Delgado RM III, et al. Is native aortic valve commissural fusion in patients with long-term left ventricular assist devices associated with clinically important aortic insufficiency? J Heart Lung Transplant. 2006;25:395–399
21. Martina JR, Schipper ME, de Jonge N, et al. Analysis of aortic valve commissural fusion after support with continuous-flow left ventricular assist device. Interact Cardiovasc Thorac Surg. 2013;17:616–624
22. Kirklin JK, Naftel DC, Kormos RL, et al. Fifth INTERMACS annual report: Risk factor analysis from more than 6,000 mechanical circulatory support patients. J Heart Lung Transplant. 2013;32:141–156
23. Park SJ, Liao KK, Segurola R, Madhu KP, Miller LW. Management of aortic insufficiency in patients with left ventricular assist devices: A simple coaptation stitch method (Park’s stitch). J Thorac Cardiovasc Surg. 2004;127:264–266
24. Morgan JA, Brewer RJ, Nemeh HW, et al. Management of aortic valve insufficiency in patients supported by long-term continuous flow left ventricular assist devices. Ann Thorac Surg. 2012;94:1710–1712
25. McKellar SH, Deo S, Daly RC, et al. Durability of central aortic valve closure in patients with continuous flow left ventricular assist devices. J Thorac Cardiovasc Surg. 2014;147:344–348

congestive heart failure; left ventricular assist device; aortic insufficiency

Copyright © 2014 by the American Society for Artificial Internal Organs