2015 Volume 79 Issue 4 Pages 770-777
Background: Patients with heart failure (HF) have abnormal endothelial function. Although use of a continuous flow left ventricular assist device (CF-LVAD) results in significant hemodynamic improvement, the effects on systemic endothelial function are unclear.
Methods and Results: Eighteen HF patients with CF-LVAD implantation were included in this prospective observational study. We measured reactive hyperemia index (RHI) before and after CF-LVAD implantation to evaluate sequential changes in endothelial function. Patients were followed clinically for the occurrence of adverse cardiovascular events, a composite of death, thrombosis, bleeding, HF, renal failure, and arrhythmia. Preoperative RHI was 1.77±0.39. Early in the postoperative period (7–14 days after operation) RHI significantly decreased to 1.19±0.31 (P<0.001, compared with preoperative RHI). At first and second follow-up (4–6 weeks and 3–7 months after operation) RHI remained lower at 1.48±0.50 (P=0.030) and 1.26±0.37 (P=0.002), respectively, compared with preoperative RHI. The decrease in early postoperative RHI relative to preoperative RHI was significantly associated with adverse cardiovascular events after CF-LVAD (age-adjusted risk ratio for 0.25 decrease in RHI, 1.35; 95% confidence interval: 1.13–1.62, P=0.001).
Conclusions: Peripheral endothelial function had a significant and persistent decline up to 5 months following implantation of CF-LVAD, and this decline was associated with adverse cardiovascular events. These findings may provide insight into some of the vascular complications following CF-LVAD in HF patients. (Circ J 2015; 79: 770–777)
Left ventricular assist devices (LVAD) are increasingly implanted in patients with end-stage heart failure (HF) as a bridge to transplantation and as destination treatment.1–3 Over the past few decades, much progress has been made in the development and refinement of LVAD. Newer axial flow and centrifugal flow devices have replaced the older pulsatile devices. The relatively non-pulsatile flow generated by these devices4 is generally well tolerated, but certain complications, including risk of bleeding and thrombosis, persist,5 as reported by a recent study highlighting this issue.6
Vascular endothelium plays an important role in maintaining the function and viability of the vessel and is extremely sensitive to injury. Endothelial dysfunction is prevalent among patients with HF7,8 and, independent of the initial mechanism for HF, it plays an important role in the progression of the syndrome and also has a significant impact on clinical outcomes.9–12 Endothelial cells synthesize and release vasoactive substances to control vascular tone and peripheral circulation.13 Nitric oxide (NO) is the principal endothelium-derived relaxing factor, and its reduced release may be one of the mechanisms responsible for increased systemic vascular resistance in the setting of non-pulsatile blood flow.14–17 In contrast, pulsatile blood flow enhances basal and stimulated release of endothelium-derived NO, resulting in reduced resistance.16,17 In HF patients supported by continuous flow (CF) LVAD, although symptoms are largely alleviated, endothelial function may be altered further by diminished arterial pulsatility. This progressive endothelial dysfunction might contribute to vascular complications after CF-LVAD implantation.
Reactive hyperemia-peripheral arterial tonometry (RH-PAT) is an accepted non-invasive method, with satisfactory reproducibility, to assess vascular endothelial function.11,18–23 RH-PAT correlates well with coronary endothelial dysfunction,24,25 and cardiovascular events.26,27
The aim of this study was to assess endothelial function (as reflected by RH-PAT) after implantation of CF-LVAD in patients with HF and to investigate the relationship between endothelial function and occurrence of adverse cardiovascular events after CF-LVAD implantation.
Consecutive patients considered for durable CF-LVAD treatment at Mayo Clinic, Rochester, MN, USA, were recruited into the study between April 2011 and May 2012. Pre-specified exclusion criteria were as follows: age <18 years or >80 years, acute HF or hemodynamic instability requiring urgent inotropic support or an intra-aortic balloon pump before operation (for whom preoperative RH-PAT was not feasible) and current or recent smokers. All patients underwent implantation of CF-LVAD (HeartMate II; Thoratec, Pleasanton, CA, USA; or HeartWare HVAD; Framingham, MA, USA; Figure 1). Informed consent was obtained from all patients before initiating the study. The study protocol was approved by the Mayo Clinic institutional review board.
Study flow chart. CF-LVAD, continuous flow left ventricular assist device; RH-PAT, reactive hyperemia-peripheral arterial tonometry.
Peripheral endothelial function was assessed in each patient with RH-PAT as previously described.18,24 RH-PAT has been found to have good reproducibility.18–20,28,29 The system (Itamar Medical, Caesarea, Israel) consists of finger probes to assess the digital volume changes accompanying pulse waves. A blood pressure cuff is placed on 1 upper arm (study arm), while the other arm serves as control (control arm). Tonometry probes are placed on 1 finger of each hand for continuous recording of the PAT signal. After a 5-min equilibration period, the blood pressure cuff is inflated to supra-systolic pressures for 5 min. The cuff is then deflated, while PAT recording continues for 5 min. The RH-PAT data are analyzed on computer in an operator-independent manner. As a measure of reactive hyperemia, RH-PAT index (RHI) is calculated as the ratio of average amplitude of PAT signal over a 1-min time interval starting 1.5 min after cuff deflation, divided by average amplitude of PAT signal of a 2.5-min time period before cuff inflation (baseline), using a computer algorithm automatically normalizing for baseline signal and indexed to the contralateral arm. In the present study, RHI >1.67 was defined as normal; this cut-off was calculated based on a previous report and has been submitted to Food and Drug Administration (FDA).24 All patients were divided into 2 groups according to postoperative RHI: low RHI (all postoperative RHI ≤1.67) and enhanced RHI (at least 1 postoperative RHI >1.67).
RH-PAT was performed prospectively at the following intervals: (1) baseline, within 1 month before CF-LVAD surgery; (2) early postoperative, 5–10 days after surgery, after transfer from the intensive care to the regular floor; (3) first follow-up, 4–6 weeks after surgery; and (4) second follow-up, 3–7 months after surgery (Figure 1).
Serum Markers of Endothelial FunctionSerum samples were collected at each RH-PAT evaluation and stored frozen for further evaluation. Using enzyme-linked immunosorbent assay these serum samples were analyzed for endothelin-1, E-selectin, intracellular adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1). Asymmetric dimethylarginine (ADMA), an NO synthase inhibitor, was evaluated using liquid chromatography-tandem mass spectrometry (reference range, 63–137 ng/ml). Laboratory analysis was performed at the Mayo Clinic cardiovascular biochemistry laboratory, using standard institutional procedure.
Follow-upAll the patients were followed clinically until July 2013 and were censored by death or cardiac transplantation. Significant clinical events reported by the patient that required medical attention, and those noted during routine CF-LVAD follow-up were charted. Adverse cardiovascular event was defined as the occurrence of a composite of death, thrombosis, bleeding, HF, renal failure, and arrhythmia as previously defined.30 These events were recorded by co-authors (S.S. and R.G.C.) blinded to the RH-PAT results.
Statistical AnalysisContinuous data with normal distributions are presented as mean±SD and non-normally distributed variables are presented as median (interquartile range [IQR]). Categorical data are represented as frequencies and percentages. Two-sided Wilcoxon signed rank test was used for comparisons of continuous data, and Pearson’s chi-squared test for discrete variables. Spearman’s correlation coefficient (rho) was used for the evaluation of possible association between change in RHI and change in 6-min walk distance. For sequential data analysis, mixed linear regression models using a random effect, to account for correlation of measures on the same patients, were used. Compound symmetry was assumed for the correlation of measures within patients. Mixed linear regression models were also used to model the association between change in RHI and changes in serum markers (changes in each variable were calculated as relative to preoperative values). The association between the incidence of clinical events during follow-up and changes in RHI were modeled using Poisson regression (log-link) with number of events as the dependent variable and log of follow-up duration as an offset variable. A 2-sided P<0.05 was considered statistically significant. Analysis was performed using JMP version 9.0.0 (SAS Institute, Cary, NC, USA) and SAS 9.3(SAS Institute).
Twenty-nine patients met the inclusion criteria and were recruited for the study. Of these 18 had baseline and at least 1 postoperative RH-PAT measurement and were included in the analysis (Figure 1). RH-PAT was performed at prespecified time periods: before operation (median, 5 days; range, 2–29 days before surgery); early after operation (median, 11 days; range, 7–14 days); first follow-up (median, 1.5 months; range, 1–1.9 months); and second follow-up (median, 4.9 months; range, 3.6–7.2 months after surgery). Early postoperative measurements were not available in 4 patients (22.2%), first follow-up measurements in 5 (27.8%) and second follow-up measurements were not available in 9 (50.0%). Reasons for missing results included medical instability, death and missed follow-up appointments. As required by the protocol, all included patients had a baseline RH-PAT measurement and at least 1 postoperative measurement, 6 patients had all scheduled measurements, 6 missed one and 6 missed 2 RH-PAT measurements. Baseline and postoperative patient characteristics are listed in Table 1. Sixteen patients (88.9%) had HeartMate II implantation (Thoratec) and 2 (11.1%) had HeartWare HVAD. Eleven patients were in the low RHI and 7 in the enhanced RHI group. Except for a higher prevalence of diabetes mellitus in the low RHI group, no significant differences were present between the 2 groups with respect to preoperative and operative characteristics. Baseline RHI also did not differ significantly between the groups (low RHI group, 1.75±0.36 vs. enhanced RHI group 1.79±0.46, P=0.85).
All (n=18) | Low RHI (n=11) | Enhanced RHI (n=7) | P-value | |
---|---|---|---|---|
Age (years) | 57±10 | 57±10 | 56±11 | 0.855 |
Male | 16 (89) | 11 (100) | 5 (71) | 0.137 |
Hypertension | 14 (78) | 7 (64) | 7 (100) | 0.119 |
Diabetes | 11 (61) | 9 (82) | 2 (29) | 0.049* |
Chronic kidney disease | 10 (56) | 6 (54) | 4 (57) | >0.99 |
Ischemic etiology | 8 (44) | 5 (45) | 3 (43) | >0.99 |
SBP (mmHg) | 99±15 | 98±17 | 101±11 | 0.786 |
Weight (kg) | 92±16 | 96±15 | 86±18 | 0.277 |
NYHA class | IIIb: 3 (17), IV: 15 (83) |
IIIb: 2 (18), IV: 9 (82) |
IIIb: 1 (14), IV: 6 (86) |
>0.99 |
Medication before operation | ||||
ACEi or ARB | 9 (50) | 5 (45) | 4 (57) | >0.99 |
β-blockers | 12 (67) | 6 (54) | 6 (86) | 0.351 |
Statin | 9 (50) | 7 (64) | 2 (28) | 0.335 |
Hydralazine | 4 (22) | 3 (27) | 1 (14) | >0.99 |
Nitrate | 4 (22) | 3 (27) | 1 (14) | >0.99 |
Preoperation echocardiography and catheterization | ||||
LVEF (%) | 16.9±6 | 18±7 | 15±5 | 0.363 |
Mean pulmonary pressure (mmHg) | 38±9 | 35±9 | 42±7 | 0.084 |
Mean wedge pressure (mmHg) | 24±7 | 22±6 | 26±7 | 0.219 |
Systemic vascular resistance (Wood units) | 17±5 | 16±4 | 19±5 | 0.160 |
Cardiac index (L·min−1·m−2) | 1.9±0.5 | 2.0±0.5 | 1.7±0.5 | 0.221 |
Operation and discharge | ||||
Bypass time (min) | 111±55 | 106±37 | 120±78 | 0.618 |
Device | ||||
HeartMate II | 16 (89) | 10 (91) | 6 (86) | 0.730 |
HeartWare | 2 (11) | 1 (9) | 1 (14) | |
Duration of hospitalization (days) | 16 (11–29) | 14 (11–26) | 18 (12–31) | 0.716 |
CF-LVAD pump speed (revs/min) | 9,300±242 | 9,320±169 | 9,267±350 | 0.729 |
Estimated CF-LVAD flow (L/min) (n=16) | 5.5±0.4 | 5.4±0.4 | 5.7±0.3 | 0.099 |
Pulsatility index (HeartMate II, n=16) | 5.2±0.7 | 5.3±0.8 | 5.1±0.7 | 0.588 |
F/U echocardiography | ||||
Aortic valve opening | 12 (67) | 7 (64) | 5 (71) | >0.99 |
Outflow graft Doppler (m/s) | 1.5±0.5 (n=13) | 1.4±0.6 (n=9) | 1.7±0.3 (n=4) | 0.214 |
Data given as mean±SD, median (IQR), or n (%). *P<0.05 (Wilcoxon test or Fisher’s exact test). ACEi, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; F/U, follow-up; IABP, intra-aortic balloon pump; CF-LVAD, continuous flow left ventricular assist device; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association; RHI, reactive hyperemia-peripheral arterial tonometry index; SBP, systolic blood pressure.
Mean RHI before surgery was 1.77±0.39 (>1.67 in 10 patients, 55.6%). After surgery, RHI further decreased in most patients. The trends in RHI after surgery are shown in Table 2, Figure 2. Early postoperative RHI decreased to 1.19±0.31 (P<0.001; >1.67 in 1 patient, 7.1%). Subsequently, in the first follow-up period a trend towards partial recovery in RHI was observed in 9 patients with overall RHI of 1.48±0.50 (P=0.22 compared with early postoperative; P=0.029 compared with preoperative; >1.67 in 6 patients, 46.2%). This was followed by a later trend for overall decline during the second follow-up period with RHI of 1.26±0.37 (P=0.001 compared with preoperative; >1.67 in 1 patient, 11.1%). Overall, most patients had a decline in postoperative RHI (P<0.001 for overall trend; Figure 3).
Before operation |
Early after operation |
P-value vs. before operation |
1st F/U | P-value vs. before operation |
2nd F/U | P-value vs. before operation |
P-value overall | |
---|---|---|---|---|---|---|---|---|
ADMA (ng/ml) | 145±23 (n=18) |
150±26 (n=14) |
0.239 | 153±24 (n=12) |
0.265 | 158±33 (n=8) |
0.118 | 0.371 |
Endothelin-1 (pg/ml) | 3.1±1.9 (n=16) |
2.7±0.9 (n=4) |
0.222 | 1.9±0.7 (n=4) |
0.023 | 2.1±1.0 (n=8) |
0.098 | 0.086 |
E-selectin (ng/ml) | 49.4±14.8 (n=16) |
53.5±13.8 (n=4) |
0.363 | 56.8±12.4 (n=4) |
0.264 | 48.1±20.6 (n=8) |
0.815 | 0.524 |
ICAM-1 (ng/ml) | 436±189 (n=18) |
552±192 (n=14) |
0.006 | 416±94 (n=13) |
0.541 | 372±117 (n=9) |
0.189 | 0.004 |
VCAM-1 (ng/ml) | 1,532±717 (n=18) |
1,758±682 (n=14) |
0.004 | 1,636±631 (n=13) |
0.267 | 1,723±874 (n=9) |
0.306 | 0.032 |
RHI | 1.77±0.39 (n=18) |
1.19±0.31 (n=14) |
<0.001 | 1.48±0.50 (n=13) |
0.029 | 1.26±0.37 (n=9) |
0.001 | <0.001 |
Data given as mean±SD. P-value calculated using the generalized mixed linear model. ADMA, asymmetric dimethylarginine; ICAM, intracellular adhesion molecule; VCAM, vascular cell adhesion molecule. Other abbreviations as in Table 1.
Sequential changes in reactive hyperemia-peripheral arterial tonometry index (RHI) after continuous flow left ventricular assist device surgery. Data given as mean±SD. P<0.001 for overall postoperative decline on the generalized linear mixed model. *P<0.05 for mean RHI vs. preoperative RHI.
Effect of continuous flow left ventricular assist device implantation on endothelial function. Patient distribution of change in reactive hyperemia-peripheral arterial tonometry index from the preoperative level (negative values indicate a decrease from baseline).
Table 2 lists sequential changes in serum markers of endothelial function and RHI. Increases in ADMA and E-selectin were noted after CF-LVAD implantation, but these changes were not significant compared with preoperative levels (ADMA, P=0.37; E-selectin, P=0.52). Although the number of patients with available measurements was small, overall endothelin-1 tended to decrease after CF-LVAD implantation (P=0.086) and significantly decreased at first follow-up compared with preoperation values (P=0.023). ICAM-1 and VCAM-1 increased in the early postoperative period and overall changes were significant (ICAM-1, P=0.004; VCAM-1, P=0.032). A significant correlation was not detected between changes in serum markers and the change in RHI (endothelin-1, P=0.63; E-selectin, P=0.48; ICAM-1, P=0.53; VCAM-1, P=0.20). The results of both pre- and postoperative 6-min walk tests were available in 10 out of 18 patients, and the exercise capacity tended to improve (from 304 m to 361 m, P=0.12; median duration between CF-LVAD implantation and postoperative 6-min walk test, 113 days; IQR, 70–183 days). The changes in postoperative RHI (at any follow-up time point), relative to preoperative measurements, were positively related to the change in 6-min walk distance (Table 3), indicating that less worsening in endothelial function might be associated with better improvement in exercise capacity.
Change in RHI | n | Spearman ρ | P-value |
---|---|---|---|
Before→Early after operation | 7 | 0.536 | 0.22 |
Before→1st F/U | 9 | 0.209 | 0.59 |
Before→2nd F/U | 6 | 0.486 | 0.33 |
6-MWT, 6-min walk test. Other abbreviations as in Table 1.
Clinical events were evaluated for a median follow-up of 1.48 years (IQR, 0.76–1.72 years). During follow-up, 3 patients (16.7%) underwent cardiac transplantation, and a total of 69 adverse cardiovascular events were observed (death, n=3, thrombotic events, n=5; bleeding, n=21; HF, n=11; renal failure, n=3; and arrhythmia, n=26). The incidence of clinical events during follow-up modeled as a function of postoperative RHI is shown in Table 4. Low RHI (≤1.67), at all postoperative examinations, showed a trend for association with the occurrence of adverse cardiovascular events, without reaching statistical significance (age-adjusted risk ratio [RR], 1.69; 95% confidence interval [95% CI]: 0.99–2.91, P=0.056). The decrease in early postoperative RHI, relative to preoperative measurements, was significantly associated with higher risk of overall adverse cardiovascular events (age-adjusted RR, 1.35; 95% CI: 1.13–1.62, P=0.001) as well as increased risk of arrhythmia (age-adjusted RR, 1.45; 95% CI: 1.07–1.96, P=0.017). Decrease in first follow-up RHI was also significantly associated with higher risk of adverse cardiovascular events (age-adjusted RR, 1.20; 95% CI: 1.05–1.36, P=0.007).
No. events | Low RHI vs. enhanced RHI |
Change in RHI before to early after operation |
Change in RHI before operation to first F/U |
||||
---|---|---|---|---|---|---|---|
RR (95% CI) | P-value | RR† (95% CI) | P-value | RR‡ (95% CI) | P-value | ||
Univariate model | |||||||
Adverse cardiovascular events | 69 | 1.69 (0.99–2.89) | 0.056 | 1.34 (1.12–1.60) | 0.001 | 1.19 (1.05–1.34) | 0.006 |
Death | 3 | 4.18 (0.22–81.0) | 0.34 | 0.98 (0.43–2.26) | 0.97 | 1.38 (0.60–3.15) | 0.45 |
Thrombotic events | 5 | 6.57 (0.36–118) | 0.20 | 2.17 (0.93–5.08) | 0.073 | 5.57 (0.74–42.1) | 0.096 |
Bleeding | 21 | 1.91 (0.70–5.21) | 0.21 | 1.16 (0.87–1.54) | 0.32 | 1.12 (0.91–1.37) | 0.29 |
Heart failure | 11 | 1.59 (0.42–6.00) | 0.49 | 1.33 (0.81–2.18) | 0.25 | 1.32 (0.91–1.90) | 0.14 |
Renal failure | 3 | 1.19 (0.11–13.2) | 0.89 | 1.24 (0.51–3.03) | 0.55 | 1.18 (0.68–2.04) | 0.55 |
Arrhythmia | 26 | 1.13 (0.50–2.53) | 0.77 | 1.45 (1.07–1.95) | 0.015 | 1.11 (0.92–1.34) | 0.28 |
Age-adjusted model | |||||||
Adverse cardiovascular events | 69 | 1.69 (0.99–2.91) | 0.056 | 1.35 (1.13–1.62) | 0.001 | 1.20 (1.05–1.36) | 0.007 |
Death | 3 | 4.55 (0.27–76.1) | 0.29 | 0.90 (0.36–2.26) | 0.82 | 1.60 (0.55–4.64) | 0.39 |
Thrombotic events | 5 | 6.68 (0.40–111) | 0.19 | 2.17 (0.92–5.11) | 0.076 | 5.17 (0.96–27.7) | 0.055 |
Bleeding | 21 | 1.91 (0.69–5.24) | 0.21 | 1.16 (0.86–1.55) | 0.34 | 1.13 (0.91–1.41) | 0.26 |
Heart failure | 11 | 1.59 (0.42–6.03) | 0.50 | 1.46 (0.86–2.46) | 0.16 | 1.34 (0.90–1.99) | 0.15 |
Renal failure | 3 | 1.19 (0.11–13.4) | 0.89 | 1.51 (0.60–3.77) | 0.38 | 1.19 (0.67–2.11) | 0.56 |
Arrhythmia | 26 | 1.12 (0.50–2.53) | 0.78 | 1.45 (1.07–1.96) | 0.017 | 1.10 (0.91–1.33) | 0.31 |
†Calculated for 0.25 decrease in RHI from before to early after operation. Four patients did not have early postoperative RHI measurement. ‡Calculated for 0.25 decrease in RHI from before operation to first F/U. Five patients did not have first F/U RHI measurement. CI, confidence interval; RR, risk ratio. Other abbreviations as in Table 1.
The current study demonstrates that, in end-stage HF patients, implantation of CF-LVAD is associated with persistent decrease in endothelial function, as assessed on RH-PAT, up to 5 months after implantation. Of importance, the increase in serum markers of endothelial function, VCAM-1 and ICAM-1, also indicate deterioration in endothelial function after CF-LVAD implantation. It is worthy of note that decline in RHI was significantly associated with adverse cardiovascular events after CF-LVAD implantation. These findings shed new light on the vascular physiologic responses to newer, second-generation CF pumps.
The vascular endothelium is a monolayer of cells lining the internal lumen of all blood vessels, and plays an important role in modulating vascular tone, inflammation, thrombosis, and platelet adhesion. Endothelial dysfunction is a significant risk factor for future cardiovascular events.31 Most cardiovascular risk factors have been associated with an impairment of the endothelial NO synthase pathway, which is considered one of the earliest events in vascular disease.32 It has been reported that improvement in endothelial function is associated with future reduction in cardiovascular events.18,33 This suggests that endothelial function could serve as a useful surrogate for vascular health. A well-defined distribution of NO is usually maintained in the systemic circulation under normal physiological conditions.34 Laminar and pulsatile patterns of blood flow in the arteries preserve endothelial function.34 Mechanical stimuli, such as pulsatile flow and fluid shear stress, could play a significant role in regulating the synthesis and release of vasoactive substances by the endothelium.35
Comparison With Other StudiesThere are few reports that evaluated the effects of CF-LVAD support on endothelial function. Amir et al reported a cross-sectional study comparing the effects of LVAD type (CF-LVAD vs. pulsatile flow LVAD) on endothelial function, as assessed on flow mediated dilation.36 They found that patients supported with CF-LVAD had significantly lower endothelial function than patients supported with pulsatile flow LVAD, indicating that the pulsatile blood flow characteristics of pulsatile flow LVAD offer superior vascular benefits over the continuous blood flow characteristics of CF-LVAD. Recently, a small cross-sectional study, which compared patients 1–4 months after CF-LVAD support with HF controls, reported similar (low) RHI in the 2 groups, suggesting no improvement in endothelial function with CF-LVAD support.37 It was also reported that endothelial microparticles did not change before and after CF-LVAD implantation.38 Because of the preliminary pilot study, which enrolled only 8 patients, a high inter-individual variability in the measurements of endothelial microparticles was observed. The present results support previous reports that endothelial function in HF patients does not improve after implantation with CF-LVAD. Moreover, the present serial data demonstrated that CF-LVAD treatment was associated not only with lack of improvement but also with significant deterioration in endothelial function. To our knowledge this is the first study to demonstrate deterioration in endothelial function after CF-LVAD implantation, with important clinical consequences on serial RH-PAT examinations.
Patients with diabetes mellitus were reported to be at higher risk of mortality after LVAD implantation compared to those without.39 In the current study, diabetes mellitus was more prevalent in the low RHI group compared with the enhanced RHI group. Moreover, the change in RHI from before to 1 month after CF-LVAD implantation was significantly smaller in patients with diabetes mellitus than in those without (–0.63±029 vs. 0.03±0.68, P=0.04). The present results suggest that the attenuation of endothelial function following CF-LVAD implantation may play some role in poor outcomes in diabetes patients.
Possible MechanismsIt has been reported that pulsatility in blood flow enhances basal and stimulated release of endothelium-derived NO compared with non-pulsatile systemic blood flow.14–17 A recent study by John et al noted activation of endothelial function and coagulation system in CF-LVAD recipients.40 It was reported that elevation in inducible endothelial markers (ICAM, E-selectin, and tissue factor) persisted for up to 6 months after CF-LVAD implantation.40 Past in vitro studies demonstrated that pulsatile flow improved endothelial cell elongation and realignment compared with CF,41,42 and maintained elevated endothelial NO synthase and Cox protein expression and activity.43 Similarly, in vivo studies reported that pulsatile flow enhanced endothelium-derived NO release in peripheral vasculature.16,17 Taken together, attenuation of pulsatility in blood flow could be a reason for the deterioration in endothelial function after CF-LVAD implantation.
Stimulation of inflammatory cascade after CF-LVAD implantation in HF patients has been reported in several previous studies.44,45 Major causes of this activation of inflammation cascade may include exposure of blood to the artificial surface of CF-LVAD, triggering a foreign body reaction and altered rheologic conditions with different velocities of blood flow, and blood stasis in the CF-LVAD recipient heart. Inflammation is associated with reduced basal and stimulated NO release from the arterial endothelial cells through various mechanisms,46 and could possibly be the mechanism responsible for the present results.47
In the present study, although the overall change did not reach statistical significance, endothelin-1 decreased at first follow-up. The mechanism for elevated plasma endothelin-1 concentration in patients with HF may be multifactorial. In a previous in vitro study, it has been demonstrated that an increase in mRNA for endothelin as well as enhanced release of endothelin occur in response to lower fluid mechanical shear stress.48,49 Therefore, it is possible that, similar to other neurohumoral vasoconstrictors, endothelin-1 is released as a compensatory mechanism in chronic circulatory failure to maintain arterial pressure. Thus the decrease in endothelin-1 in the present patients might be secondary to an improvement in systemic circulation by CF-LVAD support.
Future ResearchAlthough improvements in surgical techniques and patient care have improved outcome in patients with end-stage HF, the incidence of hemorrhagic and thromboembolic complications in patients with CF-LVAD support is still high.6,50 We showed that decline in endothelial function was significantly associated with the composite endpoint of adverse cardiovascular events after CF-LVAD implantation. Progressive endothelial dysfunction might partly contribute to adverse events observed after CF-LVAD implantation. Presence of endothelial dysfunction after CF-LVAD support may involve an increased risk for cardiovascular events and RH-PAT assessment may be used to monitor the risk in CF-LVAD-supported patients. Additional research, however, in the form of large-scale, multicenter, prospective studies is needed to further elucidate the importance of endothelial function in adverse cardiovascular outcomes after CF-LVAD implantation. It is likely that the next generation of VAD under development will deliver a greater degree of pulsatility, and thus the decline in peripheral endothelial function may be mainly an observation related to the present generation of devices.51
Study LimitationsFirst, the present single-center study was limited by the small number of patients, and selection bias may affect the results. Thus the clinical implications of deterioration in endothelial function in CF-LVAD recipients remain to be elucidated. Second, this is a longitudinal observational study and we could not assess the causality of endothelial dysfunction. Third, we did not have a control group, such as patients with HF who received non-cardiac surgery and patients with HF who received implantation of a pulsatile flow LVAD. Fourth, although the patients were studied prospectively, there were missing data for RHI and serum markers. Therefore the present results should be considered in the context of the limitation of missing data. Fifth, although some previous studies adopted RH-PAT to assess endothelial function in patients with LVAD,37,52 RH-PAT during CF-LVAD support has not been validated well.
Peripheral endothelial function, as assessed on RHI, declines following implantation of CF-LVAD. This decline in RHI is associated with adverse cardiovascular events after implantation. Dynamics of endothelial function after CF-LVAD implantation may provide an insight into some of the mechanisms of vascular complications in CF-LVAD recipients.
Funding: Y.M. and T.A. are supported by a research fellowship from Banyu Life Science Foundation International. The work was supported by the National Institute of Health (NIH Grants HL-92954 and AG-31750) and the Mayo Foundation.
The authors accepted a non-restricted clinical research grant from Thoratec, to study endothelial function after CF-LVAD.