Author + information
- Received July 21, 2016
- Revision received September 6, 2016
- Accepted September 11, 2016
- Published online January 30, 2017.
- Farhan Zafar, MDa,
- Chet R. Villa, MDa,
- David L. Morales, MDa,
- Elizabeth D. Blume, MDb,
- David N. Rosenthal, MDc,
- James K. Kirklin, MDd and
- Angela Lorts, MDa,∗ ()
- aCincinnati Children’s Hospital Medical Center, Cincinnati, Ohio
- bBoston Children’s Hospital, Boston, Massachusetts
- cLucile Packard Children’s Hospital, Palo Alto, California
- dUniversity of Alabama at Birmingham, Birmingham, Alabama
- ↵∗Reprint requests and correspondence:
Dr. Angela Lorts, The Heart Institute, MLC 2003, Cincinnati Children’s Hospital Medical Center, 3333 Burnet Avenue, Cincinnati, Ohio 45229.
Objectives This study investigated how small patient size affects clinical outcomes in patients implanted with a continuous flow left ventricular assist device (CFLVAD).
Background The development of smaller CFLVADs has allowed ventricular assist device (VAD) use in anatomically smaller patients; however, limited outcome data exist regarding CFLVAD use in patients with a body surface area (BSA) ≤1.5 m2.
Methods All CFLVAD patients entered in the Interagency Registry for Mechanically Assisted Circulatory Support registry April 2008 to September 2013 and with BSA data were included. Biventricular VAD patients were excluded. Patient characteristics and clinical outcomes were compared between patients with BSA ≤1.5 m2 (small patients) and those >1.5 m2.
Results Of 10,813 CFLVAD recipients, 231 had a BSA ≤1.5 m2. Small patients were more commonly female patients (68% vs. 20%; p < 0.01), Hispanic (10% vs. 6%; p < 0.03), and on intravenous inotropes (88% vs. 80%; p < 0.01). Small patients had higher bleeding (p < 0.01) and driveline infection (p < 0.01) rates, while exhibiting lower rates of right heart failure (p < 0.01) and renal dysfunction (p < 0.01). Device malfunction rate (p > 0.05), overall survival (p > 0.05), and 1-year competing outcomes (p > 0.05) were similar between BSA groups.
Conclusions Patients with a BSA ≤1.5 m2 supported with a CFLVAD have similar survival to larger patients. These data support the use of CFLVAD in anatomically small patients.
Ventricular assist devices (VADs) have revolutionized the treatment of end-stage heart failure. The increased utilization that has been seen in the past 5 years has been accelerated by the transition from historic large, pulsatile VADs to the newer, smaller, continuous flow pumps (1,2). Historically, women and children have been underserved by VAD therapy despite favorable VAD outcomes in these populations (3,4). Multiple reasons exist for the lower VAD utilization rates in these patients; however, small patient size is one of the most frequently cited causes. The size concerns are reflected in the small number of patients with a body surface area (BSA) ≤1.5 m2 in the HeartMate II (Thoratec Inc., Pleasanton, California) and HeartWare (HeartWare Inc., Framingham, Massachusetts) clinical trials and continued access protocols (5–8). In addition, the instructions for use for each device specifically state that the safety and effectiveness of the devices have not been established in patients with a BSA ≤1.5 m2. In spite of the labeling and limited data, the use of continuous flow devices has rapidly expanded to smaller patients (9–13), including children (14) with BSAs as low as 0.7 m2, because there are no ideal alternative therapies.
Given the limited data assessing the safety of these devices in small adults, this study sought to characterize the use of continuous flow left VADs (CFLVADs) in patients with a BSA ≤1.5 m2 and to describe the differences in clinical outcome associated with the use of CFLVADs in these “undersized” patients. On the basis of the positive clinical experience to date (9,11–13), we hypothesized patients with a BSA ≤1.5 m2 would have similar survival to larger patients. In addition, we thought it possible that the morbidity profile would differ given both the expected demographics and chest/device size mismatch for patients with a BSA ≤1.5 m2.
Data for this study were obtained from the INTERMACS (Interagency Registry for Mechanically Assisted Circulatory Support) Registry, funded by the National Heart, Lung, and Blood Institute, National Institutes of Health, Department of Health and Human Services, under Contract No. HHSN268201100025C. The INTERMACS database is a national registry sponsored by National Heart, Lung, and Blood Institute to collect data on patients treated with U.S. Food and Drug Administration (FDA)-approved mechanical circulatory support devices for the treatment of advanced heart failure. Participation is mandatory for Centers for Medicare & Medicaid Services–approved mechanical circulatory support implantation centers. The data are audited, and adverse events are reviewed and validated by a Medical Events Committee. A project-specific research proposal was reviewed and approved by INTERMACS data coordinating center before the release of a completely deidentified dataset for this study.
The INTERMACS registry was used to identify all adults (>18 years) implanted with a CFLVADs from April 2008 to September 2013. Patients (117) with missing BSA and sex information were removed. Patients (329) with biventricular VAD implants were also excluded from the study. The remaining (10,813) patients were divided into 2 cohorts based on BSA: “standard” patients with BSA >1.5 m2 (matching fit criteria for at least 1 of the FDA-approved continuous flow [CF] devices) and “small” patients with BSA ≤1.5 m2 (not matching fit criteria for any FDA-approved CF devices) (9,10).
Patient characteristics and outcomes
Pre-implant characteristics, implantation strategies, and clinical and laboratory data were compared between the 2 groups. Because of the deidentified nature of the dataset, age was available only in groups of 10-year periods. The primary outcome was all cause-morality with data censored at transplantation or device explanation for myocardial recovery. Secondary outcomes were major adverse events measured both as events per 100 patient months and time to first major bleeding, device malfunction, infection, neurological dysfunction, renal dysfunction, and right heart failure (RHF) (15). Bleeding event was further classified by site: mediastinal, gastrointestinal, and others. The “Bleeding: others” category included bleeding from the pump pocket, pleural space, abdominal, pulmonary, retroperitoneal, urinary tract, and ear-nose-throat/dental. Similarly, infection was also further classified by location: pump related (pocket and pump interior), driveline, and others. The “Other infection” category included line sepsis, pulmonary, urinary tract, mediastinum, peripheral wound, and gastrointestinal.
Categorical characteristics were compared using chi-square test or Fischer’s exact test as appropriate. Continuous variables were compared using the Mann-Whitney U test. The primary endpoint was presented as a competing outcome analysis (transplant, recovery, alive on device, or death) using time to transplant, recovery, or death. Favorable outcomes (transplanted, recovered, alive with device in place) were compared at 1 year and overall using the chi-square test. Freedom from first episode of major adverse events was compared using the Kaplan-Meier method. The exact event rate per 100 patient months and 95% confidence intervals for each major adverse event were estimated using the Byar method and compared using z-test statistics between the 2 groups. All the statistics was performed using SPSS, version 21 (IBM Corporation, Armonk, New York).
Baseline characteristics by BSA
A total of 231 (2%) small patients (BSA ≤1.5 m2) received a CFLVAD during the study period compared with 10,582 (98%) standard patients (BSA >1.5 m2). The patient size distribution is shown in Figure 1. Pre-implant characteristics, device strategies, and clinical and laboratory values are compared in Table 1. The small patients were more likely to be women, Hispanic, and on intravenous inotropes at the time of implant. Age, diagnosis, INTERMACS profile, and biventricular VAD use were similar between the 2 groups. Smaller patients were more commonly in the group deemed unlikely to be listed (5% vs. 3%; p = 0.05). There were only 15 patients with BSA ≤1.2 m2 and only 1 patient with a BSA <1.0 m2 (0.76 m2).
Overall survival was similar between the 2 groups (Table 2). One-year competing outcomes were favorable in 81% of the small patients (15% transplanted, 1% recovered, and 65% alive on device) and was similar to standard patients (p = 0.8). However, fewer small patients were transplanted at 1 year compared with standard patients (15% vs. 23%; p < 0.01) (Figure 2). Cause of death was similar between the 2 groups except for major infection (1.8% vs. 9.2%; p = 0.05) (Online Table 1).
The major adverse event rates are presented in Table 3 and time to first event in Figure 3. Overall bleeding rate was higher in small patients compared with standard patients (7.0 events/100 patient-months vs. 5.9 events/100 patient-months; p < 0.01). Small patients had higher rates of gastrointestinal bleeding and other site bleeding; however, there was no difference in mediastinal bleeding. The higher number of bleeding events in the small patient cohort occurred primarily in men rather than women (9.5 events/100 patient-months vs. 6.0 events/100 patient-months; p < 0.01), and number of bleeding events were similar between women with BSA ≤1.5 m2 and >1.5 m2 (5.06 events/100 patient-months vs. 5.92 events/100 patient-months; p = 0.66).
The localized driveline infection rate was significantly higher in the small patients compared with standard patients (1.9 events/100 patient-months vs. 1.4 events/100 patient-months; p = 0.01). There was no difference in pump or other infection rates. There was no difference in infection rates based on sex in the small patient cohort (6.4 events/100 patient-months vs. 7.0 events/100 patient-months; p = 0.48). Infection rates were also similar between women with BSA ≤1.5 m2 and >1.5 m2 (6.4 events/100 patient-months vs. 6.8 events/100 patient months; p = 0.39).
The small patients experienced less renal dysfunction (0.6 events/100 patient-months vs. 1.1 events/100 patient-months; p = 0.01) and RHF (0.2 events/100 patient-months vs. 0.5 events/100 patient-months; p = 0.01). No small patients experienced early RHF (<1 month). The small patients also had lower rates of late RHF (>1 month) compared with standard size patients (0.2 events/100 patient-months vs. 0.4 events/100 patient-months; p = 0.04).
The rate of all other major adverse events, including arterial noncentral nervous system thrombus, device malfunction, hemolysis, hepatic dysfunction, neurological dysfunction, stroke, rehospitalization, and wound dehiscence, were not different between the 2 groups (p > 0.3).
Time-to-event analysis showed that there was no difference in freedom from first bleeding event, renal dysfunction, infection, and RHF (Figure 3), despite differences in the overall event rates noted previously. There was also no difference in freedom from device malfunction (p = 0.62) and neurological dysfunction (p = 0.57).
The current study found there was no significant difference in mortality for patients in the INTERMACS registry with a BSA ≤1.5 m2 compared with patients with a BSA >1.5 m2. The smaller patients demonstrated no early RHF and lower rates of late RHF compared with patients with a BSA >1.5 m2, but did experience higher rates of nonmediastinal bleeding and driveline infection. Despite incremental differences in these outcomes, overall survival was similar between small and standard size patients. These data support the continued use of these continuous flow devices in small patients.
The finding that survival is similar in patients ≤1.5 m2 is notable given the rise of CFLVAD technology and the increasing use of these devices in small adults and children (2). The current data as well as the recent report from the PediMACS (Pediatric Interagency Registry for Mechanical Circulatory Support) (13) provide the most robust data to date that CFLVADs can safely and effectively support small patients. The smaller footprint of current-generation CFLVADs has allowed these devices to be implanted in populations that have traditionally been underserved because of small body size, including women, non-Caucasian ethnicities, and children (11,12,14,16,17). The lower rates of VAD utilization and delay in the use of VAD therapy within these populations have also been linked to higher waitlist mortality (18,19). Although the industry instructions for use for the most common CFLVADS still cite a BSA of 1.5 m2 as their lower size recommendation, the lack of therapeutic alternatives has driven the use of these devices in small patients despite the limited data. The current study provides some reassurance that CFLVADs are safe and effective in patients with a BSA ≤1.5 m2 as outcomes are favorable and similar to those of larger patients. Of note, the small patients included in this study were not significantly older, did not undergo VAD implantation at a less favorable INTERMACS patient profile, and had similar indications for implantation (35% vs. 37% implanted as destination therapy). The smaller patients were more likely to be on inotropes (88% vs. 80%) at the time of implantation, suggesting that they may have waited longer before the decision was made to implant. The patients with a BSA ≤1.5 m2 were also more likely to be women (68%) compared with the larger patient cohort (20%), although post-VAD survival has not been shown to be different between men and women (3,16,20).
Although the lack of a significant difference in mortality is the most notable finding within the current report, there were subtle differences in adverse event rates that should be noted. First, small patients experienced slightly higher rates of driveline infection (1.9 events/100 patient-months vs. 1.4 events/100 patient-months), although small patients were less likely to have a major infection as the cause of death (1.6% vs. 9.2%). Of interesting, small patient size has not consistently been identified as a risk factor for driveline infection. In fact, studies reporting an association between infection and size have generally identified larger size as a risk factor (21–23), although this association is inconsistent (24,25). Assessing the risk of infection in small patients is likely not possible from the existing data given that the patients in the majority of studies have a median BSA in the 1.8 to 2.0 m2 range. We suspect the relationship between patient size and infection risk is likely not linear, and attempts to analyze the data may be confounded by risk of infection in obese patients (22). It may be that the risk of infection is in fact U-shaped (i.e., the risk for infection is higher in small and obese patients). However, even focused examination of studies including larger numbers of lower BSA patients has alternately reported smaller size as a risk factor for infection (26,27) or not (28). It may be that a patient who is “appropriately” small because of age, sex, or ethnicity will likely carry lower infectious risk profile compared with a patient who is small because of cardiac cachexia. Cachectic, or frail, patients could not be readily identified within the current dataset from rest of the cohort given the lack of patient-specific data or trends in the pre-implant data (prealbumin, weight, body mass index, and so on). Alternately, there may be an inherent risk for driveline infection in smaller patients with less central adiposity or abdominal musculature; this will require further study. Small patients may require modified surgical technique (29) and exit site management protocols if the current findings are confirmed in future studies. Previous studies have reported success in lowering driveline infections through a specific driveline care protocol (30); this may be especially relevant in smaller patients. There may also be an inherent risk for infection within the patient cohorts that is not assessed with the current data. The white blood cell count was similar between groups; however, infectious history, chronicity of heart failure, and frailty were not assessed and may modify infectious risk (26,31).
Second, the small patients experienced greater rates of bleeding, especially early in the postoperative course, compared with the larger patients. The bleeding events occurred primarily in small men rather than in women. The increased bleeding rate was not driven by higher rates of mediastinal bleeding within the small patient cohort, which was a potential concern given the small thoracic cavity size and resulting device/chest size mismatch. That women were not at particularly high risk of bleeding is especially notable given previous reports suggesting female sex (20,32) may increase the risk of bleeding complications. It is also notable that although bleeding complications were higher within the smaller group, there was no difference in the rates of thromboembolism, neurological dysfunction, or device malfunction. These were all potential concerns given that the CFLVADs were designed and optimized for patients with a BSA of ∼1.9 to 2.0 m2. Running CFLVADs at lower rates per min (to accommodate for smaller patient size) could theoretically increase the risk of thrombosis given lower flow rates through the device. The observed difference in coagulation-related complications may also be due to inherent differences in hemostasis or the details of anticoagulation management, which are not available in the current dataset.
Third, patients with a BSA ≤1.5 m2 had lower rates of RHF (0.19 events per 100 patient-months vs. 0.47 events per 100 patient-months), despite having a higher proportion of women and more patients supported with inotropes at the time of implantation, both of which have been reported as risk factors for post-VAD RHF (16,33,34). The lower rates of RHF in the small patient cohort are especially notable given the theoretical potential for worsening RHF with inappropriately high LVAD speed (for size) with excessive left ventricular decompression and coincident septal shift (35). However, there is precedent for lower rates of RHF, specifically, late RHF in smaller patients. Takeda et al. (36), showed larger patients had higher rates of late RHF and renal dysfunction. The current study showed similar risks for the larger patient cohort, although, as with the Takeda et al. (36) article, we are unable to discern which insult is primary, RHF or renal dysfunction. It is reasonable to hypothesize that the adverse changes in right ventricular function associated with increasing patient size (37,38) are driving post-VAD renal failure. The effect of device management, especially pump speed, could not be assessed within the current data set.
Finally, although the focus of this study has been the clinical outcomes of VAD implantation in small adults, the data are also relevant to another traditionally underserved population: children. PediMACS recently reported the initial CFLVAD experience in children (13). Children supported with a continuous flow VAD experienced an excellent overall outcome, with a 92% favorable outcome at 6 months as well as favorable rates of bleeding, infection, renal dysfunction, stroke, and neurological dysfunction. Unfortunately, the rate of RHF and specifics of bleeding and infectious complications were not specifically reported in the initial PediMACS study and thus could not be compared with the INTERMACS data. Nonetheless, the combined data from the current study and the PediMACS study provide further reassurance that the use of current-generation VADs in small patients should be considered. Although we believe the data support implantation in smaller patients as a whole, the “safe” lower limit of patient size has not yet been determined.
The INTERMACS registry provides the most comprehensive method for assessing outcomes in specific populations that cannot be performed at a single center. Thus, the INTERMACS dataset provides a unique method to assess outcomes in a relatively small patient population (such as patients with a low BSA); however, it is not without limitations. There is limited granularity regarding patient-specific clinical variables (especially regarding cardiac cachexia and frailty) that may provide further clarity regarding the differences in adverse outcomes that are reported. There is also a possibility of selection bias in that the small patients who received a CFLVAD in the current study may not be reflective of the larger population with a small BSA (i.e., centers implanted the device only in small patients they thought would be most likely to succeed). The preoperative demographic, INTERMACS profile, and end-organ laboratory do not suggest this is the case, but it remains possible. The limited number of small patients and differences in patient characteristics make it difficult to assess the effect of size alone and may also confound the reported differences in adverse event rates while making adjusted analysis difficult. Thus, some of the differences in adverse event rates reported may reflect the differences in patient selection, characteristics, or management rather than size. Specific pump thrombosis data were not collected until 2014 and therefore not analyzed in this study. Although less likely, given that a few statistical tests resulted in rejection of null hypothesis, the possibility of false-positive rate cannot be excluded in multiple comparisons; no corrections were made for multiple comparisons in this study.
In a large, national registry, there was no difference in post-VAD mortality in relation to the size of the patient. Although there were small differences in specific adverse events, which require further study, these differences did not affect survival, and thus small size alone should not be a deterrent to CFLVAD implantation. These data should provide care providers reassurance to implant patients with a BSA of <1.5 m2 if they meet other criteria for VAD support.
COMPETENCY IN MEDICAL KNOWLEDGE: CFLVADs are being placed in patients who are smaller than the current size recommendations in the device labeling. In this analysis of the INTERMACS registry, we demonstrate that patients ≤1.5 m2 supported with CFLVADs have similar survival to larger patients.
TRANSLATIONAL OUTLOOK: Further studies are needed to define the inflection point where small patient size adversely impacts survival in patients implanted with a CFLVAD.
The authors thank the INTERMACS investigators, coordinators, and participating institutions for the data they have provided for this registry.
For a supplemental table, please see the online version of this paper.
Data collection for this work was funded in whole or in part with federal funds from the National Heart, Lung, and Blood Institute, National Institutes of Health, Department of Health and Human Services, under contract no. HHSN268201100025C. Dr. Morales has served as a consultant for Syncardia. Dr. Rosenthal has received research funds from Berlin Heart; and educational support from Heartware.
All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
Drs. Zafar and Villa contributed equally to this work.
- Abbreviations and Acronyms
- body surface area
- continuous flow left ventricular assist device
- Food and Drug Administration
- Interagency for Mechanically Assisted Circulatory Support
- right heart failure
- ventricular assist device
- Received July 21, 2016.
- Revision received September 6, 2016.
- Accepted September 11, 2016.
- American College of Cardiology Foundation
- Khazanie P.,
- Hammill B.G.,
- Patel C.B.,
- et al.
- Hsich E.M.,
- Naftel D.C.,
- Myers S.L.,
- et al.
- Pagani F.D.,
- Miller L.W.,
- Russell S.D.,
- et al.
- Aaronson K.D.,
- Slaughter M.S.,
- Miller L.W.,
- et al.
- ↵Summary of safety and effectiveness data. Premarket approval application number 060040. Thoratec Heartmate II Left Ventricular Assist System. April 21, 2008. Available at: http://www.accessdata.fda.gov/cdrh_docs/pdf6/P060040b.pdf. Accessed January 2017.
- ↵Summary of safety and effectiveness data. Premarket approval application number P100047. Heartware Heartware Ventricular Assist System. November 20, 2012. Available at: http://www.accessdata.fda.gov/cdrh_docs/pdf10/P100047b.pdf. Accessed January 2017.
- Wu L.,
- Weng Y.G.,
- Dong N.G.,
- et al.
- Rossano J.W.,
- Lorts A.,
- VanderPluym C.J.,
- et al.
- Birks E.J.,
- McGee E.C. Jr..,
- Aaronson K.D.,
- et al.
- Hsich E.M.,
- Starling R.C.,
- Blackstone E.H.,
- et al.
- John R.,
- Aaronson K.D.,
- Pae W.E.,
- et al.
- Gordon R.J.,
- Weinberg A.D.,
- Pagani F.D.,
- et al.
- Koval C.E.,
- Thuita L.,
- Moazami N.,
- Blackstone E.
- Imamura T.,
- Kinugawa K.,
- Nitta D.,
- et al.
- Asaki S.Y.,
- Dean McKenzie E.,
- Elias B.,
- Adachi I.
- Cagliostro B.,
- Levin A.P.,
- Fried J.,
- et al.
- Loyaga-Rendon R.Y.,
- Acharya D.,
- Pamboukian S.V.,
- et al.
- Boyle A.J.,
- Jorde U.P.,
- Sun B.,
- et al.
- Moon M.R.,
- Bolger A.F.,
- DeAnda A.,
- et al.
- Takeda K.,
- Takayama H.,
- Colombo P.C.,
- et al.
- Wong C.Y.,
- O'Moore-Sullivan T.,
- Leano R.,
- Hukins C.,
- Jenkins C.,
- Marwick T.H.