| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2006;81:1745-1751
© 2006 The Society of Thoracic Surgeons
a Cardiology Division, Vanderbilt University, Nashville, Tennessee
b Radiology Department, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts
c WorldHeart, Inc, Oakland, California
d Department of Cardiothoracic Surgery, Stanford University, Stanford, California
e Division of Cardiology, Duke University, Durham, North Carolina
f Division of Cardiology, Columbia University, New York, New York
g Department of Cardiothoracic Surgery, University of Maryland, Baltimore, Maryland
h Baltimore Veterans Administration Medical Center, Baltimore, Maryland
Accepted for publication November 29, 2005.
* Address correspondence to Dr Butler, Cardiology Division, 383-PRB, Vanderbilt University Medical Center, Nashville, TN 37232 (Email: javed.butler{at}vanderbilt.edu).
| Ms Howser and Dr Portner disclose that they have a financial relationship with WorldHeart, Inc.
|
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
METHODS: To assess the relationship between renal function and LVAD use, changes in creatinine clearances (CrCl, in mL/min) were assessed retrospectively in 220 patients who underwent LVAD placement. These patients were also divided into four groups based on CrCl quartiles (< 47, 48–68, 69–95, and > 95) and compared for outcomes post-LVAD placement.
RESULTS: Eighty-four patients died on LVAD support. Survival on LVAD was worse for patients with the worst baseline CrCl (42%, 52%, 63%, and 79% for 6 month and 26%, 34%, 47%, and 66% for 12 month survival for quartiles 1–4; both p < 0.01 for trend). Adjusting for other covariates, patients in the lowest CrCl quartile were at a higher risk of dying postimplant (odds ratio 1.95, 95% confidence interval 1.14–3.63). Paired sample analysis showed the following changes in CrCl: preoperative to week 1, 77.0 ± 46.6 to 92.1 ± 51.1 (p < 0.01; n = 202), week 1 to 2, 89.4 ± 49.2 to 95.2 ± 52.4 (p = 0.01, n = 171), week 2 to 3, 107.5 ± 58.1 to 113.7 ± 66.1 (p = 0.16, n = 74), and week 3 to 4, 111.1 ± 56.6 to 110.5 ± 56.8 (p = 0.87, n = 60). For the 60 patients with baseline CrCl less than 50, CrCl increased from 36.7 ± 9.2 to 60.1 ± 35.5 (p < 0.01; n = 55 pairs) from preimplant to week 1. In 37 of these patients (62%) on intraaortic balloon pump support preimplant, CrCl increased from 38.4 ± 8.2 to 67.9 ± 40.3 mL/minute (p < 0.01) during week 1 postimplant. Recovery of renal function to CrCl greater than 50 was associated with a trend towards better 30-day survival (84% vs 66%, p = 0.09).
CONCLUSIONS: Baseline poor renal function is associated with worse outcomes after LVAD implantation. However, renal function improves substantially and rapidly in post-LVAD survivors and is associated with improved outcomes. These data underscore the importance of careful patient selection for LVAD therapy.
Deteriorating renal function is common in patients with advanced heart failure and portends poor outcomes [1–3]. Depressed renal function is also considered a contraindication to cardiac transplantation; however, this remains controversial [4, 5]. First, there is no uniformly accepted renal function below which transplantation is considered absolutely contraindicated, although many centers require a creatinine clearance (CrCl) of at least 50 mL/minute for transplant eligibility. This is based on studies showing a higher risk of mortality posttransplant with advanced baseline renal insufficiency [6]. For example, Ostermann and colleagues [7] found that cardiac transplant patients with a preoperative CrCl equal to 50 mL/minute or less had twice the risk of mortality at 30 days than those with a CrCl greater than 50 mL/minute. Moreover, there are no definitive tests reliably shown to predict reversibility of abnormal renal function. Inotropic infusion with hemodynamic optimization and intraaortic balloon pump insertion are commonly used to optimize renal blood flow in these patients in order to assess the "true" baseline renal function, but these maneuvers are not always definitive. Left ventricular assist devices (LVADs) offer another option to optimize peripheral tissue perfusion and renal function serving as a "bridge to transplant candidacy" among patients with advanced heart failure who are otherwise not candidates. Although LVADs are currently only approved for destination therapy and bridge to transplantation (BTT), there is a growing interest in the concept of bridge to candidacy; in this respect, our data explore this concept with respect to renal dysfunction [8].
Conversely, abnormal renal function is associated with poor outcomes in patients undergoing cardiac surgery in general, and we expect that this may also be true for patients undergoing LVAD placement [9]. Thus there are potentially countervailing influences on the interaction between preimplant renal function and LVAD outcomes. On one hand, LVAD implantation may improve renal function and therefore may improve prognosis; on the other hand, poor renal function may increase the risk of adverse post-LVAD outcomes. Neither the effect of baseline renal function on outcomes post-LVAD implantation, nor the efficacy of LVAD therapy to reverse renal insufficiency, have previously been quantitatively reported in a large cohort of LVAD patients. Earlier studies on the association between renal dysfunction and LVAD outcomes are inconsistent and contradictory [10–16]. In this study, we therefore sought to assess the relationship between renal function and LVAD use from both perspectives; the impact of renal function on LVAD outcome and the impact of LVAD implantation on renal function.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Renal Function Assessment
Creatinine clearance (CrCl) was calculated using the Cockcroft-Gault formula [17]. To assess the impact of preimplant renal function on post-LVAD outcomes, the patients were divided into four groups based on the CrCl quartile distribution: group-1 (CrCl < 47 mL/minute); group-2 (CrCl 48–68 mL/minute); group-3 (69–95 mL/minute); and group-4 (CrCl > 95 mL/minute). The morbidity and mortality of these groups were compared after LVAD implantation.
To assess changes in renal function over time, renal function was assessed from the preoperative period (n = 220) to week 1 (postimplant day 6–8, n = 203), week 2 (postimplant day 13–15, n = 178), week 3 (postimplant day 20–22, n = 81), and week 4 (postimplant day 27–29, n = 82) postimplantation. In the event of more than one assessment during this three day period each week, the highest serum creatinine value was chosen to calculate CrCl (biasing against a favorable effect of the device on CrCl and assessing the worst CrCl postimplantation). We assumed the highest serum creatinine postimplantation would provide the best estimate of true CrCl because postoperative hemodilution is common. Moreover, this was conservative, reducing the risk of overestimating LVAD impact. In order to account for possible selection bias (patients living longer postimplantation may have had better renal function, and censoring due to death of patients with poor renal function could falsely overestimate LVAD impact on renal function), paired sample analyses were performed among patients who had renal function assessed at each of two consecutive time points. This "paired analysis" was performed comparing renal function over the following intervals: pretransplant to week 1 (202 paired observations), week 1 to week 2 (171 paired observations), week 2 to week 3 (74 paired observations), and week 3 to week 4 (60 paired observations).
Outcomes and Definitions
The outcomes studied included impact of baseline renal function on survival after LVAD placement and time-dependent changes in renal function in survivors. Survival was assessed as the proportion of patients within each CrCl quartile alive at 30, 180, and 365 days postimplantation censoring patients at the time of transplantation. In addition, since the majority of the LVADs are currently placed as a BTT, the composite endpoint of mortality either on LVAD or within 30 days after transplantation was assessed among the BTT patients. Crude mortality in BTT and destination patients was evaluated separately to rule out the possibility of differing results in the two patient cohorts. The results were similar with respect to survival. Therefore, except as specified, all subsequent analyses were performed on the combined database.
Cause of death was studied by dividing deaths into the following categories: cardiovascular, neurologic, infectious, multisystem organ failure, and other (miscellaneous or unknown). The following complications were also compared among the four groups.
Neurologic
Including transient ischemic attack, embolic stroke, hemorrhagic stroke, thrombotic stroke, seizures, and metabolic encephalopathy. Considering the established interaction among thromboembolism, particularly in VAD recipients, the incidence of transient ischemic attack, embolic stroke, and hemorrhagic stroke were specifically studied among the four groups.
Infectious
Any positive blood or tissue culture resulting in antibiotic therapy. We also specifically studied local (pump pocket or driveline) and systemic (sepsis) infections.
Respiratory
Ventilator support for more than five days for the index operation or any other unplanned intubation thereafter, excluding reoperations.
Bleeding
Bleeding in any organ system requiring greater than five units of packed red blood cell transfusions within a 24 hour period.
Reoperation
Any surgery required after LVAD implantation, excluding transplantation. In order to differentiate reoperations for early postoperative complications from other surgical procedures, we studied the proportion of reoperations before and after 7 days of LVAD implantation. Specific indications for reoperation were not available.
Statistical Analysis
Overall changes in renal function for each successive time period were compared using the t test. Paired sample t tests were performed to assess changes in patients in whom renal function was assessed at two successive time intervals. Similar analyses were performed on a restricted cohort of patients with a baseline CrCl equal to 50 mL/minute or less, those with CrCl equal to 50 mL/minute or less who were on intraaortic balloon pump support at the time of implantation, and those with a history of diabetes. Univariate analyses were performed to assess associations between patient characteristics and the various patient groups, based on renal function quartiles using 
2 tests (testing for linear trends) for categorical variables and analysis of variance for continuous variables. Complications after LVAD implantation were then compared among the four groups using the same analytic approach. Among patients with baseline CrCl equal to 50 mL/minute or less, additional analyses were performed to assess the predictors of recovery of CrCl to greater than 50 mL/minute, and 30-day survival difference between those who did and did not recover their renal function.
The Kaplan-Meier survival method was used to assess both overall survival and the composite end point of survival on LVAD and within 30 days of transplantation. Log-rank statistics were performed to assess statistical significance in survival differences among the four groups. Cox regression analyses were preformed to assess the independent predictors of survival and to calculate odds ratios (OR) and 95% confidence intervals (CI) for risk assessment. The following characteristics were studied as potential predictors: age, gender, race, heart failure etiology, prior history of hypertension, diabetes, thoracotomy, abdominal surgery, smoking, transient ischemic attack and stroke, several serum chemistries, and hemodynamics. Variables with missing values for more than 15% of the study sample were excluded from consideration. These included pulmonary capillary wedge pressure, serum albumin, glucose, and hemoglobin.
A 2-tailed p value of 0.05 was used to designate statistical significance. All p values 0.20 or less are shown in the results; the remaining are designated as nonsignificant (NS). Categorical variables are presented as proportions and continuous variables as mean ± standard deviation. All analyses were performed using SPSS for Windows Release 11.5 (SPSS Inc, Chicago, IL).
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
|
|
|
Changes in Renal Function
The overall change in CrCl from the preoperative period to weekly intervals up to week 4 postimplant was as follows: 76.6 ± 46.1 to 91.8 ± 51.2 to 95.7 ± 52.6 to 114.9 ± 63.8 to 107.1 ± 50.7 mL/minute, respectively. Paired sample analysis on patients with renal function measurements at two consecutive time intervals showed the following changes: preoperative to week 1, 77.0 ± 46.6 to 92.1 ± 51.1 mL/minute (p < 0.01); week 1 to 2, 89.4 ± 49.2 to 95.2 ± 52.4 mL/minute (p = 0.01); week 2 to 3, 107.5 ± 58.1 to 113.7 ± 66.1 mL/minute (p = 0.16); and week 3 to 4, 111.1 ± 56.6 to 110.5 ± 56.8 mL/minute (p = 0.87). Similar early improvement was also seen in patients with diabetes (preoperative to week 1, 78.81 ± 36.4 to 90.4 ± 47.3 mL/minute; week 1 to 2, 90.3 ± 41.5 to 96.1 ± 37.8 mL/minute; week 2 to 3, 90.1 ± 39.5 to 92.5 ± 48.3 mL/minute; and week 3 to 4, 94.7 ± 42.2 to 87.6 ± 40.8 mL/minute).
A similar analysis of paired samples was performed in 60 patients (27%) with CrCl equal to 50 mL/minute or less prior to LVAD implantation. The CrCl increased from 36.7 ± 9.2 preimplant to 60.1 ± 35.5 (n = 55; p < 0.01) one week post-LVAD placement, and from 59.7 ± 36.8 mL/minute to 66.8 ± 47.3 mL/minute (n = 47; p = 0.15) from week 1 to week 2. Thirty-seven of the 60 patients with a CrCl equal to 50 mL/minute or less (62%) who were on intraaortic balloon pump support preimplant also showed significant improvement in CrCl after LVAD implantation. In this group, the CrCl increased from 38.4 ± 8.2 preimplant to 67.9 ± 430.3 mL/minute (p < 0.01) during the first week after LVAD surgery. Forty-one of the 60 patients (68%) improved their CrCl to greater than 50 mL/minute after LVAD placement. These findings were similar among patients in the BTT or the destination group.
Predictors of Reversibility and Outcomes of Patients Who Improved Renal Function
Figure 1
shows the survival difference in patients who had baseline CrCl equal to 50 mL/minute or less stratified by whether or not, postimplant, their CrCl improved to greater than 50 mL/minute. Among patients with baseline CrCl equal to 50 mL/minute or less, 30-day survival for patients who improved their CrCl to greater than 50 mL/minute was 84% as compared with 66% for those who did not (p = 0.09, relative risk 2.16, 95% CI 0.88–5.28). Patients who did improve their CrCl to greater than 50 mL/minute tended to have lower cardiac index (1.9 ± 0.7 vs 2.2 ± 0.5 L/min/m2, p = 0.09) and lower body mass index (24.4 ± 3.8 vs 26.7 ± 5.4 kg/m2, p = 0.11) prior to LVAD implantation. Similarly these patients had less prevalent diabetes (18% vs 41%, p = 0.10). There was no difference in the use of intraaortic balloon pump support prior to implantation between those who did and did not improve CrCl (60% vs 64%, p = 0.81).
|
|
|
Comment |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
In this context, our study evaluates the importance of baseline renal function with respect to LVAD use. We demonstrate not only that the patients with the worst renal function preimplant have the highest mortality risk post-LVAD placement, but also that LVAD use is associated with a significant improvement in renal function in survivors in the short-term after implantation. This improvement was seen not only in the overall group, but also in those with preoperative CrCl equal to 50 mL/minute or less even with intraaortic balloon pump support, and in those with baseline diabetes. These results could justify a more liberal LVAD use among patients with renal insufficiency as even the lowest CrCl quartile had improved renal function and prolonged survival. Although significantly reduced renal function clearly portends a relatively poor prognosis after LVAD implantation and should therefore be considered at least a relative contraindication to LVAD use, we conclude that identifying preoperative factors to discriminate between those who will and will not benefit from VAD therapy must be a priority for the field, underscoring the importance of collecting credible data in large registries [19].
Although we had relatively few patients with baseline CrCl equal to 50 mL/minute or less, in this group we found three variables that trended to an association with improved renal function and survival; absence of diabetes, lower cardiac index preimplant, and lower body mass index. Although these risk factors need validation, as a practical matter we infer that severely depressed renal function in the presence of relatively preserved cardiac index despite maximum therapy mandates thorough assessment of intrinsic renal pathology before recommending LVAD therapy, particularly in the setting of diabetes.
If the contraindications to LVAD therapy come to mirror those for cardiac transplantation, then the role of LVAD therapy could become severely constrained, and patients still be left without an alternative. Thus it is critical to identify not only the risk factors affecting LVAD outcomes, but also why such relationships exist and whether these risks are alterable to improve outcomes. The specific reasons for higher mortality risk among patients in the lowest CrCl quartile cannot be elucidated by this retrospective analysis. Surprisingly, complications commonly associated with renal insufficiency (pneumonia, infection) do not seem to explain the survival differences. Patients with worse renal function were least likely to undergo repeat operations postimplantation, both perioperatively and long-term. In routine cardiac surgery patients, renal dysfunction is associated with an increased incidence of reoperation and therefore we speculate that the lowest rate for reoperations in our patients is unlikely to be related to diminished need [9]. This unexpected result may reflect a higher threshold to provide aggressive care for patients with severe renal insufficiency and it needs further study. As newer devices that are implanted with less surgical trauma become available, and other risk factors like diabetes are better controlled, the adverse relationship between impaired renal function and LVAD outcomes may improve over time.
The improvement in renal function elucidated in our results is encouraging, especially in the subgroup analysis for those with CrCl equal to 50 mL/minute or less. Of those patients who would not be considered appropriate cardiac transplant candidates on the day of LVAD implantation when utilizing renal function as the only indicator (CrCl 
50 mL/min) 68% had a CrCl greater than 50 mL/minute within four weeks after LVAD implantation, reversing this relative contraindication to heart transplant therapy (patient population heavily weighted toward BTT, but implanted with this disqualifier, thus supporting notion of bridge to candidacy). Patients with CrCl equal to 50 mL/minute or less who were supported on an intraaortic balloon pump on average also experienced a significant increase in CrCl after one week of LVAD support. These data underscore the fact that current measures of assessment for renal dysfunction reversibility short of mechanical circulatory support are inadequate to discriminate between those who will or will not recover renal function after transplant. Moreover, a strong case is made for using this technology as bridge to candidacy in patients with potentially reversible contraindications to transplant. Most of the benefit in renal function was seen within two weeks of implantation, and thus it appears that the decision to list or not list for transplantation can be made fairly quick after LVAD implantation.
Our study has several limitations, the most important being the relatively small number of patients with CrCl equal to 50 or less at baseline. Although we had too few patients to carry out a meaningful comparison between patients with insulin dependent versus noninsulin dependent diabetes within those with baseline CrCl equal to 50 mL/minute or less, we have recently shown similar post-LVAD survival in these two groups [20]. Some data from the central registry used for this analysis (such as cause of death) may be inaccurate due to documentation and ascertainment bias. This problem is typical of multicenter device registries: in the second annual Mechanical Circulatory Support Device Database report, the reported cause of death was "multisystem organ failure" in 27%, and "other" or unspecified in an additional 31% [21]. These deficiencies are partially compensated by completeness of data capture and auditing in these two FDA registration studies with a single LVAD device. Also, our data represent the experience with one device; whether or not similar outcomes can be expected with other devices is not known.
In conclusion, in this study we demonstrate not only a significant improvement in renal function with LVAD implantation, but we also demonstrate that poor baseline renal function is associated with an increased risk for mortality. For patients with a severe reduction in baseline renal function, improvement in renal function is associated with an improved prognosis. Larger studies are needed to assess, identify, and validate the predictors of reversibility of renal dysfunction in LVAD candidates. Our findings support the notion that LVAD therapy can be used in a select group of patients with severely impaired renal function as a bridge to candidacy, in addition to its accepted role in bridge to transplant and destination therapy.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. R. Wilson, G. H. Mudge Jr, G. C. Stewart, and M. M. Givertz Evaluation for a Ventricular Assist Device: Selecting the Appropriate Candidate Circulation, April 28, 2009; 119(16): 2225 - 2232. [Full Text] [PDF]
|
|
|
|
 |
|
 S. E. Sandner, D. Zimpfer, P. Zrunek, A. Rajek, H. Schima, D. Dunkler, M. Grimm, E. Wolner, and G. M. Wieselthaler Renal Function and Outcome After Continuous Flow Left Ventricular Assist Device Implantation Ann. Thorac. Surg., April 1, 2009; 87(4): 1072 - 1078. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
= CrCl > 50 mL/minute; 
= overall; 
=