Secondary Logo

Journal Logo

Adult Circulatory Support

Pump Position Impacts HeartMate II Left Ventricular Assist Device Thrombosis

Kilic, Ahmet*; Ransom, John; Maltais, Simon; Sun, Benjamin§; Entwistle, John W. III; Bailey, Stephen; John, Ranjit#; Klodell, Charles T.**; Gregoric, Igor††; Sheridan, Brett‡‡; Chuang, Joyce§§; Farrar, David J.§§; Sundareswaran, Kartik§§; Adamson, Robert¶¶

Author Information
doi: 10.1097/MAT.0000000000000840


Left ventricular assist devices (LVADs) have gained acceptance as a surgical treatment for advanced, end-stage heart failure.1 There are more than 15,000 patients in the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) database from 158 institutions.2 With the proliferation of this life-sustaining technology, there has been increasing scrutiny on defining best care practices3 and reducing adverse events.4 Indeed, these have served as the cornerstone of designing new LVAD-centered clinical trials.5

Although LVAD technology is always moving forward, it is important to understand the principles of current surgical implantation techniques and postoperative care that maximize the chance of success before making a leap to the next device. In recent years, concerns about pump thrombosis (PT) have increased with focused reports centering on the HeartMate II (HMII) device.4 The rise in PT has been attributed to a combination of factors including patient selection and management—namely, lowering of anticoagulation targets and pump speeds.6 Factors such as extreme inflow and outflow angulation during placement, inadequate clearance of left ventricular trabeculae, and pump migration also play a role in the development of PT. The role of surgical implantation technique as one of the multifactorial causes behind PT has also begun to gain increasing attention.7–9

The PREVENtion of HeartMate II pump Thrombosis through clinical management (PREVENT) study was a multicenter, prospective investigation to evaluate the rate of PT with adoption of a uniform set of medical and surgical practices for HMII implantation.10 The surgical recommendations were aimed at preventing any pump migration and minimizing turbulent flow by creating unobstructed blood paths through the inflow cannula and outflow graft.9 In this secondary PREVENT analysis, we sought to 1) quantify the baseline anatomic characteristics of pump positioning and 2) retrospectively evaluate the impact of pump position on the primary outcomes of PT incidence and event-free survival.

Materials and Methods

Study Design

The PREVENtion of HeartMate II pump Thrombosis through clinical management study was a prospective, multicenter, single-arm, nonrandomized study to evaluate the HMII PT rate with prespecified medical and surgical recommendations. Namely, the surgical recommendations were as follows:

  1. Creation of an adequately sized pump pocket—inferiorly deep and lateral.
  2. Positioning of the inflow cannula parallel to the septum and oriented centrally within the left ventricle.
  3. Positioning of the outflow graft to avoid compression of the right ventricle.
  4. Positioning and fixation of the pump below the diaphragm to prevent migration.

The primary end-point was confirmed PT at 3 months post-implant and has recently been published.10 Prespecified secondary end-points included the incidence of PT and predefined adverse events at 6 months, as well as assessment of pump positioning via standard of care chest x-rays (CXRs). The study was conducted at 24 centers in the United States and was supervised by the sponsor (Abbott). The institutional review board at each center approved the study protocol. A total of 300 patients were enrolled from September 2014 to November 2015 and followed to 6 months post-implant or until they reached a predefined outcome (death, transplantation, device explant for recovery, or withdrawal). Data collected included adherence to recommendations, adverse events, laboratory assessments, hemodynamics, anticoagulation/antiplatelet medications, and patient outcomes. Chest x-rays were prospectively collected per study protocol at 1 week post-implant (baseline) and at 6 months post-implant.

Chest X-Ray Measurements

All CXRs were evaluated by an independent, outcomes-blinded radiologist to assess pump position. The measured parameters included pocket depth, inflow cannula (IC) angle relative to pump body, IC angle relative to vertical (parallel to spine), and cardiothoracic ratio (Figure 1). These measurements have previously been validated to assess HMII pump position.11 Patients who did not have measurable pocket depths, entirety of the pump in the CXR filed, or IC angles from the baseline CXR were excluded. Because the outflow graft is not visible in CXRs, outflow graft position was not evaluated in this analysis.

Baseline chest x-ray parameters. A: Pocket depth, (B) inflow cannula (IC) angle relative to pump body, and (C) IC angle relative to vertical (parallel to spine).

Pump Position Definition

Within the PREVENT study, we defined extreme pump position using tail-end cutoffs of the pump pocket depth and IC angle distributions. We first identified the individual ranges of pocket depth and IC angles with the best sensitivity and specificity for identifying patients who expired or had a confirmed PT at 6 months. The fifth and 10th percentiles were evaluated as possible cutoffs for the lower end of the distribution, and the 90th and 95th percentiles were evaluated as possible cutoffs for the upper end of the distribution. All combinations were tested with a lower cutoff alone, an upper cutoff alone, and lower and upper cutoffs together. An “extreme” pump position was defined as the thresholds with the best discriminatory power. Patients not meeting the definition for extreme pump position were classified as having a “normal” pump position, which is representative of the majority of implants in the PREVENT study.

We evaluated the association of extreme pump position with clinical outcomes and compared them with outcomes with normal pump positioning. Patients were divided into subgroups based on their initial pump position at implant (extreme versus normal). Adverse events and survival free of confirmed PT were then compared between the two groups.

Adverse Event Definitions

Standard INTERMACS definitions (v3.0) were utilized2 for adverse events, except for PT, which utilized the most recent INTERMACS definition (see Table 1, Supplemental Digital Content, In addition to adverse events, lactate dehydrogenase (LDH) levels were analyzed with an elevation defined as ≥2.5× upper lab normal that was sustained over at least two consecutive measurements and days after the first month post-implant. This threshold has been previously identified as an important risk factor for PT.12

Statistical Analysis

Continuous variables are presented as mean ± SD, and categorical variables are presented as percentages. Differences in continuous variables were evaluated with the Wilcoxon rank sum test, and differences in categorical variables were evaluated with the Fisher exact test. Time-to-event analysis was performed with Cox proportional hazards modeling, and hazard ratios (HR) are reported with 95% confidence intervals (CIs). A multivariable Cox proportional hazards model was used to analyze the impact of extreme pump position on survival free of confirmed PT while controlling for variability with patient management. Covariates for pump speed, anticoagulation, and blood pressure were included in the model if they met the stepwise selection criteria of p < 0.05. All statistical comparisons were two sided with a significance level of p < 0.05. Statistical analyses were performed with SAS software (SAS Institute, Inc., Cary, NC).


Baseline Characteristics

Baseline CXRs were collected from 292 patients (97%) with 236 patients having a CXR with a measurable pocket depth and IC angles meeting the inclusion criteria. Baseline characteristics of these patients are shown in Table 1. The mean age was 57 ± 13 years with a body mass index (BMI) of 29.3 ± 6.4 kg/m2. These baseline characteristics are similar to other reports of LVAD patient profiles.13 There were 11 patients with a confirmed PT (4.7%) in this cohort, which was not significantly different from the PT rate of 6.3% in the 64 patients who were not included in the analysis because of missing or unquantifiable CXRs (p = 0.53).

Table 1.
Table 1.:
Baseline Characteristics

Assessment of Pump Position

The overall distribution of pump angles and pocket depth in the PREVENT study are shown in Figure 2. Pumps were implanted with an average depth of 13 ± 4 cm below the diaphragm. Inflow cannula angles relative to pump and vertical were 69 ± 14° and 23 ± 13°, respectively. Based on the distributions of the pocket depth and IC angles, extreme pump position was defined as the following:

Distribution of baseline pump position. A: Pump pocket depth, (B) inflow cannula (IC) angle relative to pump, and (C) IC angle relative to vertical. Red dotted lines indicate the upper or lower thresholds in the definition of extreme pump position.
  1. 1. Pocket depth ≤8 cm (10th percentile) or ≥19 cm (95th percentile) or
  2. 2. IC angle relative to pump ≤46° (fifth percentile) (extreme acute angulation of the IC to pump body) or
  3. 3. IC angle relative to vertical ≤0° (fifth percentile) or ≥39° (90th percentile) (extreme angulation toward lateral or septal wall).

Fifty-eight patients (25%) had extreme pump position at baseline. Illustrative examples of a normal and extreme LVAD are shown in Figure 3.

Chest x-rays showing the variability in pump positioning. A: Appropriate “normal” pump positioning. B: Extreme pump positioning.

Supplemental Table 2 (Supplemental Digital Content, shows key differences in characteristics between patients with normal versus extreme pump position at baseline. There was a higher prevalence of ischemic heart failure in patients with extreme pump position (60%) compared with normal position (39%; p = 0.006). Patients with extreme versus normal pump position were less likely to have pump speeds ≥9,000 rpm at 30 days (67 vs. 83%; p = 0.021).

Impact of Pump Position on Adverse Events

Table 2 shows the prevalence of adverse events between patients with extreme versus normal pump position on univariable analyses. The occurrence of bleeding, stroke, infection, and right heart failure were similar between the groups. However, extreme pump position was associated with a higher risk of confirmed PT (HR = 6.2; 95% CI = 1.8–21.1; p = 0.004), suspected PT (HR = 4.1; 95% CI = 1.4–12.2; p = 0.011), or pump exchange for suspected PT (HR = 4.4; 95% CI = 1.2–16.5; p = 0.027). Additionally, patients with extreme pump position were at higher risk of clinically significant hemolysis (HR = 5.8; 95% CI = 2.1–16.0; p < 0.001) or elevated LDH (HR = 3.0; 95% CI = 1.5–5.9; p = 0.002).

Table 2.
Table 2.:
Adverse Events in Patients With Extreme vs. Normal Pump Position

Impact of Pump Position on Event-Free Survival

With univariable analysis, survival free of confirmed PT (Figure 4) was significantly lower with extreme pump positioning (HR = 3.6; 95% CI = 1.7–7.7; p < 0.001). To account for variation in patient management, median international normalization ratio (INR) and pump speed were included as covariates in a stepwise multivariable Cox proportional hazards model (Figure 5). Extreme pump positioning, INR < 2.0, and pump speed <9,000 rpm were identified as significant risk factors.

Survival free of confirmed pump thrombosis in patients with extreme and normal pump position at baseline.
Multivariable analysis of the impact of baseline pump position on survival free of confirmed pump thrombosis. Hazard ratios and 95% confidence intervals (CI) from Cox proportional hazards models controlling for differences in INR and pump speed.


The prevalence of early PT (<3 months) after LVAD implantation has been reported in a range between 3.6% and 8.4 %.4,14 In the PREVENT study, a confirmed PT rate of 2.9% at 3 months and 4.8% at 6 months post implantation was achieved.10 The surgical principles adopted by implanting surgeons were thought to contribute in part to the low PT rate in the study. These guidelines provide general recommendations to obtain an unobstructed blood flow path and have since been incorporated into the HMII Instructions for Use.15 Despite the low PT rate resulting from adoption of PREVENT guidelines, corresponding quantitative measures of appropriate IC angulation and pump pocket depth are still lacking. In the original bench and animal testing of the HMII, there was no analysis of pump depth or angulation and its impact on fluid mechanics and flow obstruction within the device. In addition, pump positioning data from large animal studies may not be directly relevant to the human heart given the large differences in cardiac anatomy. In the HMII clinical experience, prior studies have described malpositioning. However, our article uniquely provides a set of quantitative thresholds for pump pocket depth and IC orientation that are associated with poor clinical outcomes. In this analysis, we first measured pump positioning in the PREVENT study where a standardized protocol was followed. Based on these empirical distributions, we then quantitatively defined a “normal” pump, which is representative of the majority of PREVENT implants, and also a more “extreme” pump to evaluate the impact of pump position on adverse events.

Patients defined as having extreme pump positioning had significantly more PT, elevated levels of LDH, and hemolysis. Even after controlling for variation in medical management, extreme pump position remained as a significant risk factor for PT or death. Our definition of extreme pump position includes acute angulation between the IC and pump body. Previous studies have found that this angle is significantly smaller in patients requiring pump exchange for thrombosis compared with those with normal LVAD function.16 It is possible that a sharper angle between the IC and pump body may result in stagnation of blood flow. The IC angle relative to vertical provides an estimate of IC angulation toward either the septal or lateral wall. When the angle is too high or low, blood flow into the pump may be obstructed by nearby myocardium or trabeculae predisposing to PT. Inflow cannula orientation is also related to pump pocket depth. Our study identified a pump pocket depth ≤8 cm as higher risk, which was consistent with previous literature showing shallower pump pocket depth association with PT.8,16

Variability in pump positioning may also be due in part to patient-specific characteristics. Patients with extreme pump position were more likely to have ischemic heart failure. These patients may have a prior coronary arterial bypass grafting (CABG) with patent left-sided coronary grafts—in these cases of a re-do sternotomy with patent grafts to the left ventricle, the dissection is limited to avoid damaging the grafts, which may also limit optimal IC positioning. There was also a trend for obesity to be associated with supoptimal pump position (see Table 2, Supplemental Digital Content, Obese patients tend to have increased intraabdominal pressure compared with patients with lower BMI. After implant, the intraabdominal pressure can push the pump upward and possibly alter IC angulation more than in a patient with low BMI.

The technique of surgical implantation has a significant impact on the complex relationship between LVAD flow, patient hemodynamics, anticoagulation, and thrombus formation. Previous single-center and multicenter experiences have found that creating a deep pump pocket, appropriately positioning the IC and outflow graft, and anchoring the pump are needed to maximize blood flow through the LVAD.8,9,14 In our study, extreme pump position, low pump speed, and low INR were all associated with lower event-free survival. These results emphasize the importance of following not just the PREVENT surgical guidelines but also the patient management recommendations to reduce the overall risk of PT.

The definition of extreme pump position developed from this study provides quantitative guidelines for evaluating pump position after implant. If radiological evidence of extreme positioning is found, clinicians can take appropriate actions to mitigate the risk of PT. During the early postoperative period, extreme pump position can be fixed with return to the operating room for pump adjustment, internal or external anchoring or fixation of the pump,17 creation of a new core site, adjustments of an improper outflow graft length, and development of a larger, deeper pocket. Additionally, in patients who are out of the immediate operative period, knowledge that the pump is not ideally positioned can lead to earlier intervention in the setting of elevated LDH, hemolysis, suspicion of PT, or increasing power usage of the LVAD.18,19 Indeed, the ability to recognize early on the need for surgical pump exchange is an important predictor of outcomes.20,21

Although the newer generation devices can be implanted entirely intrathoracic or intrapericardial, the clear superiority of these newer pumps is uncertain. There are also some potential advantages of a preperitoneal pocket-based pump—namely, easier pump exchange or explantation, easier management of device pocket infection, avoidance of rubbing against chest wall in patients with extreme cardiomegaly, and possibly better ventricular remodeling and recovery. Data from this study will continue to help improve outcomes in patients undergoing HMII implantation. Results from this study, in particular IC angulation data, may also prove useful when developing or evaluating surgical guidelines in newer intrathoracic devices.

It is important to understand the surgical tenants of LVAD implantation and their influence on postoperative outcomes. It may even be more crucial with the newer devices to adhere to principles of LVAD implantation as device exchange with these pumps is more complex and often requires either a re-do sternotomy or thoracotomy. This is the first, multi-institutional study that illustrates the importance of surgical implantation and its impact on patient outcomes and the surgical variability even in a controlled study.

There were several important limitations in this study. The PREVENtion of HeartMate II pump Thrombosis through clinical management study was a nonrandomized, single-arm study designed to meet a performance goal. The study was designed to address the risk of early PT, and follow-up was limited to only 6 months post-implant. Although CXRs were prospectively collected, the analysis was retrospective and the thresholds of pump positioning identified as higher risk need to be prospectively validated. Additionally, there were some patients whose CXR could not be evaluated because of missing or unquantifiable values.

The use of CXRs in itself also presented an additional limitation. The pump pocket depth is dependent on the dome of the diaphragm, which can be affected by variables such as breath holding and diaphragmatic paralysis. With CXRs, the morphology of the outflow graft (e.g., kinking) also could not be assessed. Another major limitation is that CXRs are only a two-dimensional (2D) imaging modality. Three-dimensional (3D) imaging modalities such as cardiac CT and 3D echocardiography can provide precise quantitative measures of pump position relative to the left ventricle.22,23 In comparison, our 2D analysis only provides information about the relative positions of the patient’s body, pump body, and IC. However, it has been previously shown that there is a correlation between 2D and 3D parameters of pump position, and both are clinically useful in evaluating malposition.24 This analysis was conducted with a standard-of-care imaging modality to efficiently evaluate pump position in a large number of patients. As 3D image analysis becomes automated and more cost-effective, a future study with cardiac CT can be designed to assess 3D pump position.

In conclusion, results of this study emphasize the importance of following the PREVENT recommendations with a focus on initial pump positioning to reduce the risk of PT. Extreme pump positioning at implant was described with quantitative values from radiological assessment. Patients with extreme pump position at implant had lower survival free of confirmed PT and experienced significantly more PT, hemolysis, and elevated LDH at 6 months in comparison to those with normal pump position.


The authors thank Bruce L. Bower, MD, for measuring the chest x-ray parameters presented in this study.


1. Kilic A, Ailawadi G. Left ventricular assist devices in heart failure. Expert Rev Cardiovasc Ther 2012.10: 649–656.
2. Kirklin JK, Naftel DC, Pagani FD, et al. Seventh INTERMACS annual report: 15,000 patients and counting. J Heart Lung Transplant 2015.34: 1495–1504.
3. Kilic A, Kiernan MS, Brisco-Bacik MA, et al. Common clinical dilemmas in left ventricular assist device therapy: A glimpse into current trends. VAD J 2017.3.
4. Starling RC, Moazami N, Silvestry SC, et al. Unexpected abrupt increase in left ventricular assist device thrombosis. N Engl J Med 2014.370: 33–40.
5. Heatley G, Sood P, Goldstein D, et al.; MOMENTUM 3 Investigators: Clinical trial design and rationale of the Multicenter Study of MagLev Technology in Patients Undergoing Mechanical Circulatory Support Therapy With HeartMate 3 (MOMENTUM 3) investigational device exemption clinical study protocol. J Heart Lung Transplant 2016.35: 528–536.
6. Mehra MR, Stewart GC, Uber PA. The vexing problem of thrombosis in long-term mechanical circulatory support. J Heart Lung Transplant 2014.33: 1–11.
7. Uriel N, Han J, Morrison KA, et al. Device thrombosis in HeartMate II continuous-flow left ventricular assist devices: a multifactorial phenomenon. J Heart Lung Transplant 2014.33: 51–59.
8. Taghavi S, Ward C, Jayarajan SN, Gaughan J, Wilson LM, Mangi AA. Surgical technique influences HeartMate II left ventricular assist device thrombosis. Ann Thorac Surg 2013.96: 1259–1265.
9. Adamson RM, Mangi AA, Kormos RL, Farrar DJ, Dembitsky WP. Principles of HeartMate II implantation to avoid pump malposition and migration. J Card Surg 2015.30: 296–299.
10. Maltais S, Kilic A, Nathan S, et al.; PREVENT Study Investigators: PREVENtion of HeartMate II Pump Thrombosis Through Clinical Management: The PREVENT multi-center study. J Heart Lung Transplant 2017.36: 1–12.
11. Adamson RM, Bower BL, Sundareswaran KS, Farrar DJ, Dembitsky WP. Radiologic assessment of HeartMate II position: Minimal pump migration after long-term support. J Heart Lung Transplant 2015.34: 1617–1623.
12. Levin AP, Saeed O, Willey JZ, et al. Watchful waiting in continuous-flow left ventricular assist device patients with ongoing hemolysis is associated with an increased risk for cerebrovascular accident or death. Circ Heart Fail. 2016;9.
13. Jorde UP, Kushwaha SS, Tatooles AJ, et al.; HeartMate II Clinical Investigators: Results of the destination therapy post-food and drug administration approval study with a continuous flow left ventricular assist device: A prospective study using the INTERMACS registry (Interagency Registry for Mechanically Assisted Circulatory Support). J Am Coll Cardiol 2014.63: 1751–1757.
14. Klodell CT, Massey HT, Adamson RM, et al. Factors related to pump thrombosis with the Heartmate II left ventricular assist device. J Card Surg 2015.30: 775–780.
15. HeartMate II Post-Approval Updates: Destination Therapy Study and Overview of Risk of Thrombosis. Addendum to the HeartMate II LVAS Instructions for Use: Document #10000987. 2017. C Thoratec Corporation,
16. Kazui T, Zhang A, Greenberg J, et al. Left ventricular assist device inflow angle and pump positional change over time adverse impact on left ventricular assist device function. Ann Thorac Surg 2016.102: 1933–1940.
17. Milano CA, Simeone AA, Blue LJ, Rogers JG. Presentation and management of left ventricular assist device inflow cannula malposition. J Heart Lung Transplant 2011.30: 838–840.
18. Cowger JA, Romano MA, Shah P, et al. Hemolysis: A harbinger of adverse outcome after left ventricular assist device implant. J Heart Lung Transplant 2014.33: 35–43.
19. Ravichandran AK, Parker J, Novak E, et al. Hemolysis in left ventricular assist device: A retrospective analysis of outcomes. J Heart Lung Transplant 2014.33: 44–50.
20. Stulak JM, Cowger J, Haft JW, Romano MA, Aaronson KD, Pagani FD. Device exchange after primary left ventricular assist device implantation: Indications and outcomes. Ann Thorac Surg. 2013;95:1262–1267; discussion 12671268.
21. Stulak JM, Dunlay SM, Sharma S, et al. Treatment of device thrombus in the HeartWare HVAD: Success and outcomes depend significantly on the initial treatment strategy. J Heart Lung Transplant 2015.34: 1535–1541.
22. Addetia K, Uriel N, Maffessanti F, et al. 3D morphological changes in LV and RV during LVAD ramp studies. JACC Cardiovasc Imaging. 2018.11: 159–169.
23. Anselmi A, Collin S, Haigron P, Verhoye JP, Flecher E. Device-specific evaluation of intraventricular left ventricular assist device position by quantitative coaxiality analysis. J Surg Res 2017.213: 110–114.
24. Lima B, Dur O, Chuang J, et al. Novel cardiac coordinate modeling system for three-dimensional quantification of inflow cannula malposition of HeartMate II LVADs. ASAIO J. 2018.64: 154–158.

HeartMate II; pump position; pump thrombosis

Supplemental Digital Content

Copyright © 2018 by the ASAIO