Right ventricular failure (RVF) after left ventricular assist device (LVAD) implantation appears to be associated with increased mortality. However, the determination of which patients are at greater risk of developing postoperative RVF remains controversial and relatively unknown. We sought to determine the preoperative risk factors for the development of RVF after LVAD implantation. The data were obtained for 175 consecutive patients who had received an LVAD. RVF was defined by the need for inhaled nitric oxide for ≥48 hours or intravenous inotropes for >14 days and/or right ventricular assist device implantation. An RVF risk score was developed from the β coefficients of the independent variables from a multivariate logistic regression model predicting RVF. Destination therapy (DT) was identified as the indication for LVAD implantation in 42% of our patients. RVF after LVAD occurred in 44% of patients (n = 77). The mortality rates for patients with RVF were significantly greater at 30, 180, and 365 days after implantation compared to patients with no RVF. By multivariate logistic regression analysis, 3 preoperative factors were significantly associated with RVF after LVAD implantation: (1) a preoperative need for intra-aortic balloon counterpulsation, (2) increased pulmonary vascular resistance, and (3) DT. The developed RVF risk score effectively stratified the risk of RV failure and death after LVAD implantation. In conclusion, given the progressively growing need for DT, the developed RVF risk score, derived from a population with a large percentage of DT patients, might lead to improved patient selection and help stratify patients who could potentially benefit from early right ventricular assist device implantation.
Implantation of a left ventricular assist device (LVAD) has proved to be a successful treatment option for patients with end-stage heart failure, as either a bridge to transplantation (BTT) or permanent (“destination”) therapy (DT). However, a significant proportion of patients who undergo implantation with an LVAD develop significant right ventricular failure (RVF) that adversely affects the outcome. In medically nonresponsive patients, implantation of a right ventricular assist device (RVAD) might be necessary. It appears, however, that early planned institution of biventricular mechanical circulatory support results in improved outcomes compared to delayed conversion of an LVAD to biventricular mechanical support. However, the determination of which patients under consideration for advanced therapy with an LVAD are at a greater risk of developing intraoperative or early postoperative RVF remains controversial and relatively unknown. Hence, the aim of the present study was to attain a better understanding of the perioperative risk factors that are predictive of postoperative RVF in patients receiving an LVAD as BTT or DT. Moreover, we sought to define a novel risk score model derived from a large, single-center population with a large number of both DT and BTT patients.
Methods
The analysis was performed using a prospectively collected database from a large-volume, single center of 175 LVAD patients who underwent implantation from 1993 to 2008. The patients underwent implantation with the HeartMate XVE (n = 82, 47%), HeartMate VE (n = 42, 24%), HeartMate 1000 IP (n = 17, 10%), HeartMate II (n = 25, 14%; all devices manufactured by Thoratec, Pleasanton, California) or the Novacor device (n = 9, 5%; World Heart, Oakland, California). The LVADs were placed as BTT in 58% of the patients and as DT in 42% of the population. RVF was defined by the need for RVAD implantation, the need for inhaled nitric oxide for ≥48 hours, or the need for intravenous inotrope therapy for >14 consecutive days.
Clinically relevant data were collected from the patient population, ≤24 hours before implantation, within a 2-week period to determine the clinical and hemodynamic predictors for RVF after LVAD implantation. The clinical variables obtained before LVAD implantation included demographics, medications, preimplant inotrope dependency, preoperative intra-aortic balloon pump, intubation, and co-morbidities. The preoperative laboratory data included sodium, hemoglobin, white blood cell count, platelets, uric acid, total cholesterol, albumin, total bilirubin, alanine aminotransferase, aspartate aminotransferase, γ-glutamyl transferase, creatinine, blood urea nitrogen, and so forth. The hemodynamic data included measurements of the right atrial pressure, right ventricular systolic pressure, systolic pulmonary artery pressure, mean pulmonary artery pressure, and diastolic pulmonary artery pressure, pulmonary capillary wedge pressures, cardiac output, cardiac index, pulmonary vascular resistance and systemic vascular resistance. The right ventricular stroke work was calculated using the equation: right ventricular stroke work = cardiac output/[heart rate × (mean pulmonary artery pressure − right atrial pressure) × 0.0136]. Before LVAD implantation, 2-dimensional echocardiography with pulsed, continuous wave, and color flow Doppler was obtained. The measurements obtained included left ventricular ejection fraction and fractional shortening, left and right atrial area, left ventricular internal dimensions at end-diastole and end-systole, intraventricular septum, and left ventricular posterior wall dimensions.
Univariate analysis was performed using the Student t test for continuous variables and the chi-square test for the categorical variables to perform between-group comparisons for those patients with RVF and those without. For all analyses, p ≤0.05 was considered statistically significant and a univariate predictor of RVF after LVAD implantation. Backward and forward step-wise multivariate logistic regression analyses were performed for all variables according to groupings of related variables. Nominal statistical significance was set at p ≤0.05, with other variables included for p <0.10, with an odds ratio >1.5 or <0.67 or for those parameters with substantial confounding effects (change in the regression β coefficient >10%) on variables included using the first 2 criteria. A RVF risk score was created by rounding the exponentiated regression model coefficients to the nearest 0.5. The risk score thresholds were determined using recursive partitioning according to data from the receiver operating characteristic curve of the RVF risk score’s ability to predict patients’ RVF status. The area under the curve was calculated using a receiver operating characteristic curve of the RVF risk score. Kaplan-Meier survival curves were generated, and post-LVAD survival between those patients with and without RVF were analyzed using the log-rank test for linear trend. The analysis was conducted for 30-day, 6-month, and 1-year survival. Post-LVAD survival was defined as current LVAD support at the cutoff date for the study or successful cardiac transplantation. All data were analyzed using Statistical Package for Social Sciences, version 15.0 (SPSS, Chicago, Illinois) and are expressed as the mean ± SD, unless otherwise specified. The institutional review board approved the study, and all patients provided informed consent for collection of data used in the present study.
Results
Of the 175 patients analyzed, 77 (44%) experienced RVF after LVAD implantation. Of the 77 patients who were diagnosed with RVF, 45 (58%) were treated medically with either inotropes or inhaled nitric oxide, and 32 (42%) required the placement of an RVAD after LVAD implantation. The survival rate between the non-RVF and RVF groups was 96% versus 80% at 30 days (p = 0.0012), 90% versus 70% at 180 days (p = 0.0011; Figure 1 ), and 83% versus 62% at 365 days (p = 0.002).
Table 1 displays the preoperative clinical characteristics of those patients whose postoperative care was complicated by RVF compared to those who were free of RVF. The preoperative medications and laboratory values are listed in Table 2 . Tables 3 and 4 summarize the preoperative hemodynamic and echocardiographic parameters. Independent variables or those with significant confounding effects that were retained after multivariate analysis ( Table 5 ) were entered as factors of the RVF risk score. The risk score was calculated as the sum of the points assigned for the existence of each of 8 perioperative variables ( Table 6 ). The risk score was then broken into 4 categories: ≤5.0, 5.5 to 8, 8.5 to 12, and ≥12.5 points ( Table 6 ). Of the 36 patients who did not develop RVF, 32 (89%) had a right ventricular risk score of ≤5.0. Of the 18 patients with a risk score of ≥12.5 points, 15 (83%) developed RVF after LVAD implantation. The area under the curve for the risk score was 0.743 ± 0.037 ( Figure 2 ). A comparison with a recently published risk score was also undertaken. By application of that risk score’s point values for vasopressor requirement, aspartate aminotransferase ≥80 IU/L, creatinine ≥2.3 mg/dl, and bilirubin ≥2.0 mg/dl to risk in this study population, an area under the curve of 0.61 ± 0.04 was found, considerably lower than the 0.73 reported by Matthews et al.
| Variable | RVF | p Value | |
|---|---|---|---|
| Yes (n = 77) | No (n = 98) | ||
| Age (years) | 58.2 ± 12.9 | 56.5 ± 14.4 | 0.417 |
| Therapy | 0.171 | ||
| Bridge to transplantation | 52% | 62% | |
| Destination therapy | 48% | 37% | |
| Men | 79% | 87% | 0.185 |
| Body surface area (m 2 ) | 2.00 ± 0.32 | 2.06 ± 0.27 | 0.249 |
| Body mass index (kg/m 2 ) | 27.4 ± 6.3 | 27.3 ± 5.6 | 0.89 |
| Obesity (body mass index ≥30 kg/m 2 ) | 34% | 23% | 0.132 |
| Etiology | 0.516 | ||
| Ischemic | 42% | 37% | |
| Nonischemic | 58% | 63% | |
| Preoperative intra-aortic balloon counterpulsation | 49% | 31% | 0.012 |
| Preoperative intubation | 43% | 26% | 0.016 |
| Diabetes mellitus | 23% | 28% | 0.531 |
| Previous sternotomy | 42% | 45% | 0.713 |
| Smokers ⁎ | 45% | 41% | 0.464 |
| Variable | RVF | p Value | |
|---|---|---|---|
| Yes (n = 77) | No (n = 98) | ||
| Angiotensin-converting enzyme inhibitor | 22% | 34% | 0.084 |
| β Blocker | 26% | 21% | 0.504 |
| Intravenous inotropes | 87% | 79% | 0.147 |
| Angiotensin receptor blocker | 9% | 9% | 0.502 |
| Aldosterone | 27% | 27% | 0.934 |
| Allopurinol | 9% | 8% | 0.844 |
| Statin | 22% | 24% | 0.681 |
| Platelets (k/mm 3 ) | 178.8 ± 95.9 | 212.2 ± 101.8 | 0.029 |
| Lymphocyte | 11.7 ± 7.4 | 12.8 ± 9.0 | 0.452 |
| Uric acid (mg/dl) | 8.2 ± 3.4 | 8.4 ± 3.2 | 0.811 |
| Cholesterol (mg/dl) | 117.7 ± 43.6 | 149.8 ± 50.2 | 0.019 |
| Albumin (g/dl) | 3.3 ± 0.7 | 3.3 ± 0.7 | 0.663 |
| Total bilirubin (mg/dl) | 1.8 ± 1.8 | 1.3 ± 1.1 | 0.05 |
| Alanine aminotransferase (U/L) | 212.1 ± 570.9 | 130.2 ± 286 | 0.227 |
| Aspartate aminotransferase (U/L) | 258.4 ± 720.2 | 157.5 ± 360.8 | 0.115 |
| Creatinine (mg/dl) | 1.8 ± 1.0 | 1.6 ± 0.8 | 0.176 |
| Blood urea nitrogen (mg/dl) | 34.4 ± 20.5 | 31.3 ± 19.4 | 0.315 |
| Variable | RVF | p Value | |
|---|---|---|---|
| Yes (n = 77) | No (n = 98) | ||
| Systolic blood pressure (mm Hg) | 108.1 ± 17.6 | 107.6 ± 13.3 | 0.834 |
| Diastolic blood pressure (mm Hg) | 62.3 ± 11.4 | 63.0 ± 10.1 | 0.651 |
| Heart rate (beats/min) | 95.2 ± 14.5 | 96.0 ± 14.8 | 0.738 |
| QRS >120 ms | 68% | 59% | 0.297 |
| Right atrial pressure (mm Hg) | 11.6 ± 6.2 | 9.5 ± 5.1 | 0.023 |
| Pulmonary artery systolic pressure (mm Hg) | 48.9 ± 13.1 | 51.0 ± 14.6 | 0.381 |
| Pulmonary artery diastolic pressure (mm Hg) | 25.6 ± 7.2 | 24.4 ± 8.4 | 0.375 |
| Mean pulmonary artery pressure (mm Hg) | 34.5 ± 10.9 | 34.4 ± 10.6 | 0.921 |
| Pulmonary capillary wedge pressure (mm Hg) | 21.9 ± 7.5 | 21.8 ± 8.4 | 0.909 |
| Pulmonary vascular resistance (Wood units) | 3.6 ± 2.0 | 2.9 ± 1.8 | 0.485 |
| Cardiac output (L/min) | 4.5 ± 1.4 | 4.7 ± 1.7 | 0.451 |
| Cardiac index (L/min/m 2 ) | 2.2 ± 0.5 | 2.3 ± 0.9 | 0.415 |
| Right ventricular stroke work (mm × Hg × ml/beat) | 14.9 ± 7.8 | 17.4 ± 8.6 | 0.08 |
| Variable | RVF | p Value | |
|---|---|---|---|
| Yes (n = 77) | No (n = 98) | ||
| Ejection fraction (%) | 21.1 ± 6.8 | 20.3 ± 6.7 | 0.456 |
| Fractional shortening (%) | 10.8 ± 6.8 | 10.1 ± 7.0 | 0.572 |
| Left ventricular end-diastolic diameter (cm) | 6.0 ± 1.1 | 6.5 ± 0.9 | 0.003 |
| Left ventricular end-systolic diameter (cm) | 5.4 ± 1.2 | 5.9 ± 1.1 | 0.007 |
| Intraventricular septum (cm) | 1.2 ± 0.3 | 1.1 ± 0.2 | 0.283 |
| Left ventricular posterior wall (cm) | 1.2 ± 0.3 | 1.1 ± 0.2 | 0.306 |
| Left atrial area (cm 2 ) | 28.4 ± 9.7 | 28.9 ± 6.8 | 0.738 |
| Right atrial area (cm 2 ) | 22.6 ± 6.6 | 22.6 ± 6.4 | 0.962 |
| Variable | Odds Ratio | p Value |
|---|---|---|
| Full model | ||
| Destination therapy | 3.31 | 0.005 |
| Inotrope dependency | 2.47 | 0.08 |
| Obesity (body mass index ≥30 kg/m 2 ) | 1.99 | 0.08 |
| Intra-aortic balloon counterpulsation | 3.88 | 0.002 |
| Pulmonary vascular resistance | ||
| ≤1.7 Wood unit | 1.0 | |
| 1.8–2.7 Wood unit | 1.95 | 0.22 |
| 2.8–4.2 Wood unit | 3.01 | 0.045 |
| ≥4.3 Wood unit | 4.14 | 0.012 |
| Angiotensin-converting enzyme inhibitor/angiotensin receptor blocker | 0.49 | 0.054 |
| β Blocker | 1.60 | 0.258 |
| Model with variables not included in full model ⁎ | ||
| Left ventricular end-diastolic diameter | 0.73 | 0.13 |
| Left ventricular end-systolic diameter | 0.77 | 0.19 |
| Right atrial pressure | 1.03 | 0.442 |
| Right ventricular stroke work † | 0.94 | 0.09 |
| Bilirubin | 1.17 | 0.261 |
| Platelets | 0.998 | 0.221 |
| Intubation | 1.944 | 0.208 |
| Left ventricular ejection fraction | 1.01 | 0.727 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
