| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2003;75:S66-S71
© 2003 The Society of Thoracic Surgeons
a Division of Cardiology, University of Minnesota, Minneapolis, Minnesota, USA
* Address reprint requests to Dr Miller, Division of Cardiology, University of Minnesota, 420 Delaware St SE, MMC 508, Minneapolis, MN 55455, USA.
e-mail: mille278{at}umn.edu
Presented at the Heart Failure & Circulatory Support Summit, Cleveland, OH, Aug 22–25, 2002.
Abstract
Patient selection is a critical factor in the outcome associated with the use of mechanical assist devices for the treatment of refractory heart failure/shock. Numerous risk factors impact on the outcome, many of which can be identified and treated before device surgery. This manuscript reviews all the risk factors that have been identified to date and the use of composite risk scores to predict outcome.
The number of patients with advanced heart failure is increasing. Current estimates suggest that more than 5 million people in the U.S. have been diagnosed with heart failure (HF) [1, 2]. Of this total, an estimated 35% are asymptomatic, 50% are stable on oral drug therapy, and approximately 10% to 15% have advanced HF, with half of the latter group having refractory failure. Recent reports have shown that as many as 30% to 40% of all patients diagnosed with heart failure have primarily diastolic dysfunction with preserved systolic function [3]. Collectively, these data suggest that as many as 400,000 patients have advanced (stage D) HF [4] due to primary systolic dysfunction, of which, a significant percentage may be at an age and overall health to be potential candidates for cardiac replacement therapies. This includes an estimated 75,000 to 100,000 patients who are less than 65 years of age, and as many as 150,000 to 200,000 less than 75 years of age. Heart transplantation has the best demonstrated outcome for selected patients with refractory heart failure [5], but is severely limited by less than 2,500 available donors per year, making it an epidemiologically trivial therapy for the larger number of patients with refractory HF.
The prognosis associated with refractory HF is extremely poor[1, 4, 6, 7] and has not improved significantly over the past 20 years. This poor prognosis has been the impetus for the development of mechanical assist devices to help support the circulation in such patients [7–14]. These devices initially were limited to intraaortic balloon pumps and extracorporeal centrifugal pumps-with or without membrane oxygenators, but has expanded to include initially paracorporeal, and most recently, pneumatic or electrically driven internal pulsatile ventricular assist devices (VADs), as well as total artificial heart pumps.
Experience with the use of these pumps has increased significantly, but patient selection remains the major question associated with their use. Data compiled over the past 10 years have consistently shown that nearly one-third of the patients who undergo device insertion as a bridge to transplant do not survive to undergo transplantation [15–18]. (Table 1) Importantly, the 30% to 35% mortality rate has been independent of the device used, or the period of use, including an analysis limited to only patients supported during the past year rather than cumulative data. A few centers with significant experience have reported better outcomes than the cumulative data, shown in Table 1 [19–21]. Whereas the mortality with this support before transplant is fairly high, the outcome of the patients who survive to transplantation is similar to those transplanted without mechanical support [5]. Collectively, when examined by intent to treat analysis, mechanical support for refractory heart failure as a bridge to transplant is associated with survival as low as 50% and as high as 80% from implant to 1 year posttransplant.
|
Patient selection may therefore be the primary determinant of success with left ventricular assist device (LVAD) use. This manuscript will review the factors involved in patient selection that have been shown to have an adverse effect on outcome, and the use of several risk scoring systems designed to provide guidance for patient selection for LVAD use primarily as a bridge to transplant.
Renal function
Nearly all reports addressing outcome and patient selection have identified renal function as one of the most important variables adversely effecting outcome [21–23]. Renal function is likely a continuous variable. The exact cut point where serum creatinine, as the primary surrogate for renal function, becomes a significant adverse risk factor is unclear, as patients with cachexia and decreased muscle mass may not have an elevation of serum creatinine that is reflective of the amount of impairment in renal function. In general, patients with a serum creatinine more than 5 mg/dL, or on chronic dialysis, have been considered to not be good candidates for mechanical support [22]. Farrar found that blood urea nitrogen (BUN) was more predictive of adverse outcome than creatinine in a large registry of 186 patients from 42 centers supported from 1990 to 1994 [23]. The mortality was 46% for patients that had a BUN more than 40, versus 20% for patients with a BUN less than 20 at the time of device implantation. Oz and colleagues [21, 24] found urine output to be the best surrogate of renal function and end organ compromise, with a range between 20 to 30 mL/h for a 6- to 8-hour period associated with the highest increase in relative risk for outcome with mechanical support.
Hemodynamic criteria
The criteria frequently used to define refractory HF or shock [25] are also common potential indications for mechanical support and include: systolic blood pressure less than 80 mm Hg or mean arterial pressure (MAP) less than 65 mm Hg, pulmonary capillary wedge (PCW) (pressure) more than 20 mm Hg, and cardiac index less than 2.0 L/min on maximal drug therapy. The definition of maximal drug support may vary from one or multiple inotropic agents, with or without use of vasopressor agents. Many patients may not have a cardiac index less than 2.0 L/min due to physiologic compensation with severe tachycardia. This has led to use of a stroke volume less than 25 mL/beat as an alternative criteria to cardiac index. Right atrial pressure is the only consistent hemodynamic value reported to be associated with adverse outcome [22, 24], and that again varies with the percentage use of right ventricular assist devices (RVADs). Whereas hemodynamics seem a clear, objective criteria to guide patient selection, Stevenson has shown that presenting hemodynamics in patients with advance HF did not predict 2-year survival without urgent transplantation [26]. Use of strict hemodynamic criteria may lead to delays in the decision to use VADs with adverse outcomes in terms of morbidity and mortality.
Right heart failure
The status of the right ventricle is often critical to the success of using LVADs. Most reports suggest that at least 20% to 30% of patients have severe right ventricular failure at the time of LVAD implant [27, 28]. Oz and others have shown that right atrium (RA) greater than 16 to 20 mm Hg is an adverse risk factor for successful outcome [21, 22]. The introduction of implantable LVADs, with portable drivers that allowed patients to be discharged to home, has led to a concerted attempt to use only left-sided support. This approach has been associated with a significant decrease in use of mechanical right ventricular support, but often with prolonged postoperative recovery due to extended need for inotropic or pressor drug support, renal insufficiency, and mechanical ventilation after isolated LVAD use. Significant primary right heart failure has been defined as a right atrial pressure greater than 20 mm Hg on maximal drug support or an RA more than PCW. Many initial reports on the use of mechanical support for refractory HF employed biventricular support in the majority of patients [27–30], and most found no difference in outcome or morbidity with use of biventricular support. Importantly, Nakatani and colleagues [31, 32] identified three variables that helped to predict right ventricle (RV) dysfunction after LVAD use. These included RV size (end diastolic [200 mL] and end systolic volume [177 mL]); higher preload (right atrial pressure [RAP] > 23 ± 6); and afterload, with transpulmonary gradient more than 15 or pulmonary vascular resistance (PVR) more than 3.8. The identification of ventricular size as an important predictor of outcome is consistent with data from Stevenson showing left ventricular volume and size were more predictive of outcome than any hemodynamic variable [26].
Infection
Most reports on adverse risk factors associated with VAD use have identified infection as an important risk factor [22, 33]. Surprisingly, several investigators have shown that neither preoperative leukocytosis greater than 15,000 or temperasture greater than 101.5°F had significant impact on outcome with mechanical support [24]. It may be that active infection is such an obvious risk that many patients with overt infection have not been supported. However, patients with device endocarditis may only be effectively managed with device removal and transplantation [25]. Infection, and occult sepsis, may be the cause of hemodynamic deterioration of a patient with chronic heart failure, and needs to be carefully considered in all such patients. Nutritional deficiency may also have a large impact on postoperative infection.
Mechanical ventilation
Pulmonary dysfunction has been shown to be one of the most important adverse risk factors for success with mechanical support. Pulmonary dysfunction may be due to underlying pneumonia as well as pulmonary congestion from heart failure. Patients requiring mechanical ventilation or FIO2 greater than 70% have been shown to have a worse outcome that those not on a ventilator or lower oxygen concentration [21, 24].
Hepatic function
Abnormal hepatic function, specifically elevation of serum bilirubin, has been noted to be associated with worse outcome [22, 34, 35] El-Banayosy and associates [34] recently reported a series of 104 patients who had mechanical support as a bridge to transplant, half of whom had biventricular support. They examined a total of 25 variavbles, including multiple biochemical tests of hepatic and pancreatic function, and showed that only an elevated bilirubin was associated with an adverse impact, with a 1.41 increase in odds ratio or relative risk on mortality, especially in patients undergoing biventricular support. Liver enzyme tests had no correlation with outcome.
Bleeding/abnormal coagulation
Many patients undergoing LVAD implantation have had previous sternotomies for coronary or valvular heart disease, and many (most) are on warfarin anticoagulation at the time of surgery. Reoperation and prolonged anticoagulation have been shown to have an adverse impact on outcome [21, 24, 36] and compound the risk of postoperative bleeding, which is in part due to the very high pressure and upstroke of the pulse delivered by the VAD into the aorta. El-Banayosy and associates [34] found that the risk of bleeding was significantly increased in patients undergoing biventricular support. Requirement for multiple transfusions perioperatively may lead to worsening RV function as well as the development of preformed antibodies.
Reoperation/previous sternotomy
Farrar reviewed data from 186 patients who were supported as a bridge to transplant with the Thoratec PVAD between 1990 and 1994 at 42 centers [23]. It should be noted that 74% of patients in this series had biventricular support. The time from previous sternotomy was important in outcome with mechanical support, with a survival of only 39% if operated on more than 30 days before mechanical support, whereas it was 61% if less than 30 days, compared with 67% for patients with no previous sternotomy. Oz and associates [21] also found reoperation to have a 1.8-fold increase in relative risk for mortality, and ranked fifth in all risk factors and was closely related to the risk of prolonged coagulation time at the time of surgery.
Nutrition
One variable that has not been examined in patients undergoing mechanical support is nutrition. Early satiety due to hepatomegaly, and overall decline in appetite and nutrition, often lead to overt cachexia in patients with long-standing HF. Nutrition is likely a very underestimated risk factor in mechanical support, and is closely linked to overall postoperative morbidity, especially infection. Consideration of parenteral intravenous support, begun at the time of selection for mechanical support and confirmed until adequate enteral feeding can be established, may be important in reducing infection and improving overall outcome with VAD support.
Age
The majority of mechanical assist device use has been restricted to patients who are candidates for heart transplantation, or typically younger than 65 years in most programs. There are only a few reports examining age as a risk factor for mechanical support, and they have used specific cut point analyses (eg, >60 years), rather than a continuous variable [21, 22, 34]. It appears that like heart transplantation, there may be an inverse relationship between age and outcome, especially age greater than 60 years. However, use of LVADs as support for shock associated with an acute myocardial infarction may be successful even in older well-selected patients. The experience with mechanical support as destination therapy suggested that the inverse relationship holds for age greater than 60 years as well [37].
Composite risk scores
It seems clear that many factors may influence outcome, and that no one risk factor, of and by itself, constitutes an absolute contraindication to a successful outcome with use of mechanical support for refractory HF. Therefore, the development of composite risk assessment tools may help improve the outcomes associated with their use. Pennington and associates [30] were the first to examine a large number of risk factors to create a composite risk assessment to guide patient selection. They used a weighted assessment of 24 individual variables in a series of 26 patients referred for mechanical support. A variety of devices were used in this series, in which nearly 70% of patients received biventricular support. The risk factors with the greatest correlation with an adverse outcome included age greater than 50 years, preoperative CPR, mechanical ventilation or FIO2 greater than 70%, use of ECMO preoperatively, and BUN greater than 60 or creatinine greater than 2.2 mg/dL. Right atrial pressure greater than 20 did not have a significant adverse impact, but this may have been due to the very high use of RV mechanical support. Use of ECMO preoperatively (n = 5) was associated with 100% mortality in this small series. Using this weighted scoring system, they identified low-, medium-, and high-risk groups, and showed that there was a significant difference between the score in survivors and nonsurvivors, (5.86 ± 3.15 vs 11.14 ± 3.01).
Oz and colleagues reviewed a series of 56 patients supported at the Cleveland Clinic or Columbia Presbyterian Hospitals between 1990 and 1994 [21, 24]. They examined only seven variables and separated them into continuous or binary factors with the continuous variables dichotomized into yes or no categories for specific cut points such as a CVP greater than 16 mm Hg. They then applied linear regression analysis and calculated a relative risk for each variable. The factors shown to have the greatest adverse impact, or relative risk (RR), in rank order were oliguria less than 30 mL/h (RR = 3.9), CVP greater than 16 (RR = 3.1), mechanical ventilation (RR = 3.0), prolonged coagulation time (RR = 2.4), and repeat sternotomy (RR = 1.8). Of note, neither hospital days before VAD use, white blood count greater than 15,000, or temperature greater than 101.5°F conveyed increased relative risk for adverse outcome. Using a receiver operating curve, they determined that a score of 5.0 defined significant difference, with an AOC of 0.81. The mean risk factor survival score was 2.45 ± 1.7 in survivors versus 5.43 ± 2.8 in nonsurvivors (p < 0.001). Eighty percent of the patients with a score less than 5.0 survived, whereas 75% of patients with a score greater than 5.0 did not survive. Bleeding and RV failure were the most frequent postoperative problems in this series.
One unique approach to risk factor assessment has been the Acute Physiology and Chronic Health Evaluation (APACHE) scoring system, which has been employed primarily in patients with acute severe illness, not often due to heart failure [38–40]. The scoring system includes many biochemical variables such as sodium, hematocrit, renal and hepatic function, as well as hemodynamics and mechanical ventilation. Gracin and associates [41] used this scoring system in a retrospective analysis of 50 patients with severe heart failure who were not supported with mechanical devices, and compared them with 31 fairly matched patients who had LVADs implanted. Survival was highly correlated with APACHE II scores, which were significantly reduced (improved) in patients supported by LVADs. The analysis showed that survival time was better with mechanical support by Kaplan-Meier analysis, as well as Cox proportional analysis after adjustment for APACHE II score (p = 0.0219). They also demonstrated that patients with medium scores (11 to 20) in particular benefited from LVAD therapy, and supported its application for patients with severe HF but not in cardiogenic shock.
Timing of insertion (elective implant)
Deng and associates reviewed an experience in Germany with 41 patients who underwent mechanical support as a bridge to transplant [42]. Patients were divided into elective, urgent, and emergency categories, and compared with patients on the waiting list, but not requiring mechanical support. They showed that outcomes were better in the elective group than the other two groups, and as good as in nonsupported patients.
Alternative strategies (inotropes)
The primary alternative to mechanical support in patients with refractory HF is intravenous inotropic agents, vasodilators, and in some cases vasopressors agents [43, 44]. The outcome with inotropic agents versus mechanical support was initially examined by Reedy and colleagues in 34 status 1 patients awaiting transplant, where were no significant differences between the groups except the VAD patients had better cardiac index at time of transplant [45]. The mortality before transplant was not different between the two groups, but the survival at 1 year posttransplant was 70% in the inotrope group versus 100% in the VAD patients (p = 0.03). In addition, there were more complications such as infection, neurologic events, and renal failure in the inotrope group, although none of these comparisons reached statistical significance. The superiority of implantable LVADs (n = 66) over patients supported with inotropic agents (n = 38) in patients with refractory HF on a heart transplant waiting list was recently reconfirmed by Aaronson and associates [19] They showed that the LVAD-supported patients had a better survival after 3 months of support (81% ± 5% vs 64% ± 11%), and a significant survival advantage at 3 years posttransplant (95% ± 4% vs 65% ± 10%). The overall survival from time of initiating bridge support with LVADs versus inotropic agents was 77% ± 6% versus 44% ± 9% at 3 years. These data seem to demonstrate significant superiority of the LVAD strategy, especially in selected patients who might otherwise be supported with inotropic agents. Other investigators have also shown the superiority of mechanical support in patients who develop refractory heart failure while awaiting transplantation [46, 47].
Reversibility of organ dysfunction
Intensification of medical therapy with drugs and use of IABP may help to significantly improve many of the adverse risk factors before device implantation, including renal and hepatic function, as well as RV failure. Strategies such as preoperative stabilization of patients with significant RV failure with therapies such as ultrafiltration, IABP, and even mechanical ventilation with or without use of nitric oxide, may significantly reduce right heart pressures, improve function, and reduce the need for RV support. These strategies may also be very helpful postoperatively to treat RV dysfunction, often unmasked by improved cardiac output and venous return. More refractory failure may require use of temporary mechanical support with centrifugal or pulsatile pumps for 24 to 48 hours for the RV to obtain the maximum benefit of mechanical unloading with LVAD support, and decrease the need for RVAD use without excessive morbidity.
Destination therapy
It seems clear that patient selection remains the most important determinant of improved outcomes associated with the use of mechanical support for refractory HF. This review has focused only the application of this technology as a bridge to transplantation. Expanded use as destination therapy will likely require some refinement of selection criteria for patients in that group, who are often older and have more potential comorbidities. In addition, the degree of end-organ dysfunction (eg, serum creatinine) that will be allowable or not adversely impact survival will need to be defined, as the criteria in REMATCH Trial were somewhat more liberal than used for slection of transplant candidates in most programs. Careful attention to neurovascular disease, previous stroke, and screen for malignancy are certainly warranted, as well as nutritional status. The device should only be implanted electively (other than for acute myoardial infarction/shock) to allow a complete examination of potential comorbidities.
It is incumbent on the field to perform prospective and retrospective analyses of factors that impact on survival and be willing to use risk assessment tools to help guide utilization of this expensive, but often lifesaving technology.
References
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. Chumnanvej, M. J. Wood, T. E. MacGillivray, and M. F. V. Melo Perioperative Echocardiographic Examination for Ventricular Assist Device Implantation Anesth. Analg., September 1, 2007; 105(3): 583 - 601. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Lee, Y.-H. Park, S.-H. Lim, Y.-T. Kwak, H. Kim, and B.-C. Chang Successful Mechanical Circulatory Support as a Bridge to Transplantation Asian Cardiovasc Thorac Ann, June 1, 2007; 15(3): 243 - 245. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Lachapelle Invited commentary Ann. Thorac. Surg., January 1, 2004; 77(1): 149 - 150. [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 |
