Heart failure (HF) is estimated to affect 26 million people worldwide currently and is associated with 20–50% mortality within 1 year of hospitalization.1 Advanced heart failure (American Heart Association/American College of Cardiology Stage D) comprises approximately 1% of all HF patients, and portends exceptionally poor survival estimated at 25% at 1 year with medical therapy alone.2 Left ventricular assist devices (LVADs) have become standard of care for patients with end-stage HF either as a bridge-to-transplantation or destination therapy and improve survival and quality of life in such patients.3–5 However, LVADs are associated with high cost and complication rates, making patient selection essential.3,6
Diabetes mellitus (DM) is common in patients with HF and portends poor prognosis.7 Similarly, it is associated with worse outcomes in patients undergoing cardiac surgery.8 Left ventricular assist device implantation is contraindicated in patients with complicated or uncontrolled DM.3 However, the impact of DM on long-term outcomes after LVAD implantation in eligible patients is not clear. Most prior studies were performed in the era of pulsatile LVADs, which have been replaced by continuous-flow LVADs. We sought to investigate the impact of DM on long-term outcomes of patients bridged-to-heart transplantation with LVADs in the contemporary era.
Materials and Methods
Dataset and Study Design
We used the United Network for Organ Sharing (UNOS) Registry of the Organ Procurement and Transplantation Network. This registry maintains data on all patients listed for solid organ transplantations in the United States from 1987 to present. The data are provided by transplant centers at the time of listing, transplantation, and at follow-up. These data are collected through an electronic system, deidentified, and provided for administrative and research purposes. For this analysis, we identified all adults (≥18 years) listed for heart transplantation with a CF-LVAD (Heartmate II or Heartware) between 2000 and 2015 with DM and compared them with those without DM (controls). Patients were divided by the presence or absence of DM at listing.
Primary outcome was a composite of wait-list mortality or delisting for clinical deterioration. Secondary outcomes included cause-specific wait-list mortality, transplantation, transplantation with status 1A, delisting for improvement, delisting for clinical deterioration, post-transplantation mortality, and organ survival.
Baseline variables were compared using χ2 and analysis of variance (ANOVA) tests as appropriate. Time to event analyses were done using Kaplan-Meier method and Cox proportional hazard models. The models were adjusted for variables that were statistically different (p < 0.05) between patients with DM and controls per the baseline characteristics of both groups (i.e., age, sex, race, smoking, body mass index [BMI], ischemic etiology, serum creatinine, and mean pulmonary artery pressure [PAP]). No assumptions were made for missing variables. All tests were two sided, and p < 0.05 was considered statistically significant.
A total of 4,978 patients were included in this analysis. Characteristics of patients are shown in Table 1. At a median of 5.8 months on the wait-list, 259 (5%) died and 295 (6%) were delisted for deterioration. The primary composite outcome was higher in patients with DM (1 year: 14%, 2 years: 25%, 3 years: 31%) compared with those without DM (1 year: 10%, 2 years: 19%, 3 years: 26%); p < 0.001 (Figure 1A), so was wait-list mortality (DM: 1 year: 7%, 2 years: 12%, 3 years: 16% vs. no DM: 1 year: 5%, 2 years: 9%, 3 years: 12%; p < 0.001; Figure 1B).
Other outcomes were not different between patients without and with DM (3 year cumulative rate, log-rank test): infection-related mortality (0.8% vs. 0.9%; p = 0.74), pulmonary disease–related mortality (0.2% vs. 0.6%; p = 0.10), mortality from cerebrovascular disease (1.6% vs. 2.6%; p = 0.07), improvement and delisting (1.9% vs. 1.4%; p = 0.31), transplantation (76% vs. 73%; p = 0.22), or transplantation with status 1A (61% vs. 58%; p = 0.08). After multiple adjustments, DM was not associated with an increased risk of the composite wait-list mortality/delisting (hazard ratio [HR]: 1.17 [0.97–1.41]; p = 0.11) or any of the secondary outcomes: mortality (HR: 1.16 [0.88–1.53]; p = 0.30), or transplantation (HR: 0.99 [0.92–1.08]; p = 0.89), Table 2. Further, there was no difference in wait-list mortality or delisting between types of DM (type I, versus type II, versus unspecified; p = 0.51).
A total of 3,058 patients underwent transplantation, of whom 32% had DM. One, 3, and 9 year survival were higher in patients without DM (91%, 84%, 64%) compared with those with DM (87%, 79%, 60%, respectively, p (log rank) = 0.001), Figure 2A; as was graft survival (DM 90%, 83%, 65% vs. DM: 87%, 80%, 61%; p = 0.007), Figure 2B. After adjusting for age, sex, ischemic etiology, UNOS status, time on the wait-list, transplant year, total bilirubin, mean PAP, history of malignancy pretransplant use of extracorporeal membrane oxygenation (ECMO) and mechanical ventilation, BMI, ischemic time, creatinine, and donor age, DM was still associated with increased post-transplant mortality (HR: 1.23 [1.002–1.52]; p = 0.048). With respect to post-transplantation mortality, there was no interaction between DM and serum creatinine (pinteraction = 0.08), DM and ischemic cardiomyopathy (pinteraction = 0.81), or DM and cerebrovascular disease (pinteraction = 0.08).
Compared with 25,016 patients without LVAD and without DM who were listed for transplantation during the same period, patients with CF-LVAD and DM had increased risk of post-transplant mortality (HR: 1.28 [1.08–1.51]; p = 0.004; see Figure 1, Supplemental Digital Content, http://links.lww.com/ASAIO/A351).
Among patients with DM who underwent LVAD, multivariable predictors of wait-list mortality included white race (HR: 1.89 [1.10–2.38]; p = 0.022) and preoperative creatinine (HR: 1.28 [1.06–1.56] per 1 mg/dl; p = 0.013). Among patients with DM and LVAD who underwent transplantation, multivariable predictors of post-transplantation mortality include creatinine (HR: 1.39 [1.20–1.60] per 1 mg/dl; p < 0.001) and pretransplantation mechanical ventilation (HR: 4.01 [1.88–8.57]; p < 0.001). Univariable and multivariable predictors of wait-list and post-transplantation mortality among patients with DM are shown in Table 3. There was a nonstatistically significant U- shaped relationship between BMI and wait-list/post-transplant mortality (Figure 3). There was no relationship between time on wait-list and post-transplant survival (Figure 4).
Comparison with Pulsatile Left Ventricular Assist Device
Among 571 patients bridged with pulsatile LVAD (Heartmate XVE) (196 with DM), there was no statistically significant association between DM and adjusted wait-list mortality or delisting (HR: 0.74 [0.43–1.27]; p = 0.27), wait-list mortality (HR: 0.61 [0.31–1.21]; p = 0.16), or post-transplant mortality (adjusted HR: 0.89 [0.58–1.35]; p = 0.58). There was no significant interaction between type of LVAD (continuous-flow versus pulsatile flow) and DM with respect to post-transplant mortality (pinteraction = 0.10). Additionally, in a multivariable model that also adjusts for DM, there was no statistically significant difference between CF-LVAD and PF-LVAD with respect to post-transplantation mortality (p = 0.10).
In this large contemporaneous study of patients with DM bridged-to-transplant with continuous-flow LVADs, we show that DM is prevalent and associated with higher post-transplantation mortality and worse organ survival.
Prior studies on the impact of DM on post-LVAD outcomes have shown conflicting results. Among 201 patients implanted with pulsatile LVAD at a single center, DM was present in 24% of the patients and was not associated with worse survival (7 year survival in DM 43% versus non-DM 61%; p = 0.13).9 There were also no differences in infections, bleeding, stroke, renal failure, or length of stay.1 In another study of 222 patients implanted with pulsatile LVADs, DM was present in 26% of the participants and associated with increased risk of death (odds ratio [OR]: 1.76 [1.05–2.94]).10 These studies were done in the era of pulsatile flow LVADs and do not reflect current practice.
Our results are concordant with two recent single-center studies that investigated the impact of DM on continuous-flow LVAD outcomes. Vest et al.11 reviewed outcomes of 300 consecutive patients who underwent implantation of continuous-flow LVADs (62% bridge to transplantation (BTT)), of whom 43% had DM. In accordance with our findings, there were no differences in overall mortality (adjusted HR: 0.88 [0.57–1.37]; p = 0.58), stroke, intracerebral hemorrhage, pump thrombosis, or device-related infections between those with or without DM. Mohamedali et al.12 reported on the outcomes of 288 CF-LVAD patients (42% had DM) between 2006 and 2013. Diabetes mellitus was not associated with mortality (HR: 0.99; 95% CI: 0.65–1.6; p = 0.97), gastrointestinal or intracerebral bleeding, stroke, infection, or early right ventricle (RV) failure. The authors, however, found an increased rate of hemolysis among patients with DM compared with controls (10% vs. 3%; p = 0.02). These findings are in contrast to two single-center retrospective studies. In one analysis of 341 patients who received LVAD (63% destination therapy (DT)) between 2007 and 2016 (38% have DM), DM was associated with all-cause mortality (adjusted HR: 1.73 [1.18–25.3]; p = 0.005) and risk of LVAD complications (adjusted HR: 2.1 [1.35–3.18]; p = 0.001).13 In another single-center retrospective study14 of 96 patients with DM and 95 patients with DM who received Heartmate II, DM was associated with higher risk of death at 3 years (42% vs. 21%; p = 0.013), but there were no significant differences in risk infection, neurologic dysfunction, and rehospitalization with no differences by baseline hemoglobin A1c. These differences could be explained by the differences in the proportion of BTT patients which is usually associated with lower rates of comorbidities than destination therapy patients.
We additionally show that DM was associated with worse post-transplantation outcomes in patients bridged with LVADs. To our knowledge, only one previous study investigated the impact of DM on post-transplantation outcomes in patients bridged with pulsatile, and not continuous-flow LVADs. In a single-center study of patients bridged with pulsatile LVAD, DM was associated with worse post-transplantation outcomes (7 year survival: DM 57% versus non-DM 81%; p = 0.02).9 The reasons for this finding remain unknown, but could be related to increased LVAD complications which has been shown to impact post-transplantation outcomes.15
Our data are in line with the 2013 International Society for Heart and Lung Transplantation (ISHLT) guidelines for mechanical circulatory support that do not consider uncomplicated DM as a contraindication LVAD implantation.3 However, our data contradict the 2016 ISHLT guidelines that do not consider uncomplicated DM as a contraindication for heart transplantation3 as our data suggest that DM impacts post-transplantation in patients bridged with CF-LVADs.
Our study suggests that carefully selected patients with DM can be bridged safely to transplantation with contemporary LVADs. It is unclear if poor glycemic control, however, is associated with worse LVAD outcomes, but two recent studies showed preimplantation glycated hemoglobin did not impact LVAD outcomes.11,13 Unfortunately, our dataset lacks information on DM control and complications.
We acknowledge important limitations in this study. This is a retrospective, registry-derived study with inherent bias and limitations. For example, the registry lacks granular information on DM control, duration, complications, and medical therapy. We also lack device-related outcomes, such as pump thrombosis, drive-line infections, or strokes which may have been different in patients with and without DM. However, we tried to identify whether DM was associated with more frequent status 1A upgrade, which could have reflected device-related complications. Because the UNOS database only includes patients bridged-to-transplantation, the wait-list time is short (median 6 months) which is reflective of the average waiting time for transplantation of LVAD patients in the United States. Our data may not be generalizable to patients who receive LVAD as destination therapy, who typically have longer LVAD support time. Nevertheless, the large number of patients (more than 10-fold larger compared with prior studies) and the availability of hard clinical outcomes (death) make this analysis informative.
One-third of patients bridged-to-transplantation with LVAD in the United States have DM. While it does not increase wait-list mortality or delisting, DM is associated with decreased post-transplantation survival.
1. Benjamin EJ, Blaha MJ, Chiuve SE, et al.; American Heart Association Statistics Committee and Stroke Statistics Subcommittee: Heart Disease and Stroke Statistics-2017 Update: A Report From the American Heart Association. Circulation 2017.135: e146–e603.
2. Fang JC, Ewald GA, Allen LA, et al.; Heart Failure Society of America Guidelines Committee: Advanced (stage D) heart failure: a statement from the Heart Failure Society of America Guidelines Committee. J Card Fail 2015.21: 519–534.
3. Feldman D, Pamboukian SV, Teuteberg JJ, et al.; International Society for Heart and Lung Transplantation
: The 2013 International Society for Heart and Lung Transplantation
Guidelines for mechanical circulatory support: Executive summary. J Heart Lung Transplant 2013.32: 157–187.
4. Rogers JG, Pagani FD, Tatooles AJ, et al. Intrapericardial left ventricular assist device for advanced heart failure. N Engl J Med 2017.376: 451–460.
5. Slaughter MS, Rogers JG, Milano CA, et al. Advanced heart failure treated with continuous-flow left ventricular assist device. N Engl J Med 2009.361: 2241–2251.
6. Miller LW, Guglin M. Patient selection for ventricular assist devices: A moving target. J Am Coll Cardiol 2013.61: 1209–1221.
7. From AM, Leibson CL, Bursi F, et al. Diabetes
in heart failure: Prevalence and impact on outcome in the population. Am J Med 2006.119: 591–599.
8. Bucerius J, Gummert JF, Walther T, et al. Impact of diabetes
mellitus on cardiac surgery outcome. Thorac Cardiovasc Surg 2003.51: 11–16.
9. Topkara VK, Dang NC, Martens TP, et al. Effect of diabetes
on short- and long-term outcomes after left ventricular assist device implantation. J Heart Lung Transplant 2005.24: 2048–2053.
10. Butler J, Howser R, Portner PM, Pierson RN III. Diabetes
and outcomes after left ventricular assist device placement. J Card Fail 2005.11: 510–515.
11. Vest AR, Mistak SM, Hachamovitch R, Mountis MM, Moazami N, Young JB. Outcomes for patients with diabetes
after continuous-flow left ventricular assist device implantation. J Card Fail 2016.22: 789–796.
12. Mohamedali B, Yost G, Bhat G:. Is diabetes
mellitus a risk factor for poor outcomes after left ventricular assist device placement? Texas Hear Inst J 2017.44: 115–119.
13. Asleh R, Briasoulis A, Schettle SD, et al. Impact of diabetes
mellitus on outcomes in patients supported with left ventricular assist devices
: A single institutional 9-year experience. Circ Heart Fail 2017.10: e004213.
14. Usoh CO, Sherazi S, Szepietowska B, et al. Influence of diabetes
mellitus on outcomes in patients after left ventricular assist device implantation. Ann Thorac Surg 2018.106: 555–560.
15. Quader MA, Wolfe LG, Kasirajan V:. Heart transplantation
outcomes in patients with continuous-flow left ventricular assist device-related complications. J Heart Lung Transplant 2015.34: 75–81.