Continuous-flow left ventricular assist device (CF-LVAD) implantation is a widely accepted therapy for end-stage heart failure with excellent long-term survival.1–5 However, increased time on mechanical support has also heralded the challenge of managing its mechanical and nonmechanical complications, such as infection and anticoagulation.6,7 These complications not only significantly impair the patients’ freedom from admission and overall quality of life on device support but also increase the resource burden on the healthcare system. Notably, pump thrombosis (PT) is a common and potentially life-threatening complication among HeartMate II (HMII; Thoratec, Pleasanton, CA) patients that has garnered increased attention in the light of recent literature observing a cryptogenic spike in its incidence.8,9 The standard of care in mitigating the risk of PT in LVAD patients continues to be an appropriate anticoagulation therapy (international normalized ratio [INR] 1.5–2.5 and aspirin 81 mg daily), but consistently maintaining patients within therapeutic range has been difficult in practice because of poor compliance or individual variance in response to anticoagulation.10,11 Furthermore, even among properly anticoagulated patients, bleeding or thrombotic risk has stagnated because of the multifactorial etiology of PT including preceding infections, hemolysis, body habitus, and gastrointestinal bleeding that requires cessation of anticoagulation or inherently hypercoagulable state.12–14
Recently, there have been suggestions that HMII geometry, notably acute inflow cannula angles (ICAs < 55) in the immediate postoperative period via chest radiograph, may correlate with poor left ventricular unloading and ultimately higher risk of PT.15 However, studies have yet to corroborate this claim in larger cohorts. In this study, we attempt to validate this correlation between device geometry and PT incidence at our center. Furthermore, prior studies have not found any particular patterns of change in device geometry over time. We hypothesize that device geometry is not a static measure and attempt to characterize the long-term changes in HMII geometry as the heart device hybrid remodels.
Materials and Methods
This study was approved by the institutional review board at the Hospital of the University of Pennsylvania. We performed a retrospective analysis of patients implanted with HMII LVAD from January 2011 to March 2014 at our institution. The final diagnosis of PT required visual confirmation by the operating surgeon during device explantation. For patients who developed multiple PTs, recurrent incidents were analyzed as a separate subset.
Chest radiographs (anterior-posterior [AP] and poster-anterior [PA]) were reviewed by a blinded reviewer to assess the ICA and pump pocket depth (PPD). Measurements were confirmed by two blinded reviewers who independently analyzed a randomly selected subset (25%). Inflow cannula angle and PPD were recorded at two time points: 1) during the immediate postoperative period before discharge, and 2) at the most recent follow-up date before PT, transplantation, device explant, or death. Chest radiographs with >2 mm degree of rotation, defined as displacement between sternal wires and thoracic vertebral bodies, were excluded as poor imaging technique. Chest radiograph that had inconsistent orientation (AP versus PA) at the two time points or that failed to completely visualize the ICA or PPD was excluded. Inflow cannula angle was measured from the inner margin of the IC to the inner margin of the rotor. Pump pocket depth was defined as the vertical distance from the highest point of the diaphragm to the inner margin of the rotor (Figure 1).
All data analyses were performed using GraphPad Prism software, version XML6 (GraphPad Software, Inc., La Jolla, CA). Continuous variables are reported as mean ± SD or median with interquartile ranges, depending on normality. Categorical variables are reported as percentages of the total. Student’s t-test and Fisher’s exact test were used to analyze continuous and categorical variables, respectively. Paired t-test was used to assess the degree of change in patients’ ICA over time. Analysis of variance (ANOVA) was used to compare changes in ICA within the cohort. Receiver operating characteristic curves were used to evaluate ICA discrimination in predicting PT incidence. Statistical significance was defined at p < 0.05.
A total of 90 patients were implanted with HMII CF-LVADs during the study period. Of them, 16 patients (17.8%) developed PT (0.58 event/patient-year). Table 1 reports the preoperative characteristics of the 90 CF-LVAD patients enrolled in this study, categorized into PT and non-PT groups. The PT cohort had higher hemoglobin levels and lower blood urea nitrogen levels. Difference in clinical history or anticoagulation regimen did not reach statistical significance.
HeartMate II Geometry and Pump Thrombosis Incidence
Of the 90 HMII patients enrolled, 87 (96.7%) had chest radiographs in the immediate postoperative period, which met our inclusion criteria. All the radiographs had AP orientation. The three chest radiographs that were excluded were all from the non-PT group for either inadequate visualization of the device or not meeting the time requirement. Table 2 shows the immediate postoperative ICA and PPD values of the total CF-LVAD cohort, which were 54.90° ± 10.61° and 82.20 ± 29.89 mm, respectively, and did not differ between PT and non-PT groups (56.03° ± 10.14° vs. 54.64° ± 10.79°, p = 0.63, and 86.66 ± 30.93 mm vs. 81.12 ± 30.93 mm, p = 0.46). Receiver operating characteristic curve did not report any postoperative ICA discrimination in predicting PT incidence (area under curve (AUC) 0.55) (Figure 1). Inflow cannula angle measured at the most recent time point also was not a predictor (AUC 0.57). There was no correlation between the geometric acuity of ICA and the temporal acuity of PT (days postimplantation) in the PT cohort (R < 0.01).
HeartMate II Geometry Evolution with Heart-Device Remodeling
Sixty-seven (74.4%) patients had chest radiographs with consistent orientations at both time points. The majority (55 patients; 82%) had AP to AP comparison. A higher percentage of patients were excluded from the PT group than from the non-PT group (11 patients [16.4%] from non-PT vs. 5 patients [31%] from PT). Both initial and final ICA and PPD followed a normal distribution.
Changes in HMII geometry were measured over 112.5 (interquartile range = 34.3–337.3) days. During this period, PPD decreased in all groups (p < 0.01), but ICA did not shift (Figure 2). At an individual level, there was a significant variance among patients’ change in ICA although paired t-test did not reveal statistical significance (p 0.21) (Figures 3 and 4A). When patients were stratified by their initial ICA into 1) more than 1 SD below the mean, 2) within 1 SD from the mean, and 3) more than 1 SD above the mean, initial ICA was a significant predictor of future angle change (Δ IC angle +5.32° ± 8.02°, −0.78° ± 8.97°, −7.99° ± 8.13°, respectively; ANOVA p = 0.002). These values suggested a convergence toward the mean (55.4°) but did not correlate with PT (Figure 4B).
Redevelopment of Pump Thrombosis
Of the 16 patients in the PT group, 4 patients (25%) experienced a second episode of PT after device exchange compared with 17.8% incidence in the original cohort (p = 0.49). Their postoperative ICA on the second device was 49.4° ± 7.2° and did not reach statistical significance when compared with the non-PT group. One patient had a third occurrence.
In this single-center retrospective analysis, we tested the correlation between HMII geometry and the risk of PT. We further aimed to characterize the change in device geometry in the process of device-organ remodeling. Our principal findings are as follows:
- Device geometry was not valid in predicting the future risk of PT.
- Long-term changes in ICA varied depending on initial measurements and suggested a convergence toward the mean.
Pump thrombosis is a widely recognized, life-threatening complication for patients receiving mechanical circulatory support with a prevalence that ranges from 2% to 13%.3,8,16,17 It is possible that this may be an underestimation because of inherent limitations in detecting subclinical thrombus formation within the device. Current guidelines recommend a strict anticoagulation regimen—maintaining an INR between 1.8 and 2.5 with a daily aspirin dose of at least 81 mg—studies have suggested that PT may be a multifactorial phenomenon that can occur even in the setting of optimal anticoagulation.17 Infection underlying hypercoagulable state, need for noncardiac surgery, and withholding anticoagulation therapy in the setting of bleeding (gastrointestinal, neurologic, or epistaxis) has been proposed as risk factor that would alter medical management if correctly identified. More recently, Taghavi et al.15 have suggested that surgical technique during HMII implantation may influence device function and ultimately the likelihood of developing PT. Inflow cannula angle below 55° at the time of implantation was reported as having good discrimination in predicting future PT risk. In our study, initial device geometry was not valid in predicting the risk of PT. We also hypothesized that if acute ICA were a mechanical risk factor for PT, patients with acute ICAs should accordingly experience temporally acute PT. A risk factor related to surgical technique would expose patients to highest risk of PT in the relatively early postoperative period. Uriel et al.13 have also suggested that PT with mechanical etiologies (e.g., abnormal inflow cannula position, bend relief disconnect, or deformed outflow graft) presents more rapidly than those with organic or cryptogenic causes. However, in our study, having putatively thrombogenic HMII geometry did not result in earlier or more rapid PT development.
Of note, our cohort’s average IC angle was more acute than that reported in prior studies. This is attributable to a difference in our IC angle placement when analyzing the radiographs rather than a difference in surgical technique at our institution. The angle was drawn from the inner margin of the inflow cannula to the inner margin of the pump body not the rotor. We reasoned that this would be a more visible landmark for surgeons during operations. Surgically, similar to our peer institutions, our HMII placement in the recent years has been generally more conservative with increased IC angle and adequate pocket size in an effort to curb the rising incidence of PT.
We also analyzed the possibility that the initial IC angle measurement may have been a poor predictor of PT risk in our study because the HMII geometry was evolving over time. Although, similar to prior studies on HMII geometry, the average change in IC angle remained stable at the level of the entire cohort in our study, we also noted significant variance among individual patients’ delta angles. Especially, patients with initial IC angle values that deviated from the mean tended to regress toward the mean over time. Accounting for these changes in ICA, we questioned whether the patients’ most recent ICA values before PT would be a more accurate predictor than the initial ICA; however, this was also untrue. Ultimately, acuity of ICA does not seem to be related to PT risk regardless of time point.
More broadly, our findings suggest that device geometry as noted on chest radiographs is not static and that there may be a geometrically optimal destination that is best-suited for the HMII device. To our knowledge, this is the first study attempting to characterize these changes. Although our study provides a valuable foray into analyzing changes in device geometry, factors that influence geometric change are still not clearly understood and may include multifactorial etiologies, such as myocardial remodeling because of ventricular unloading, wound healing, or variations in surgical technique. Future studies that analyze changes in device geometry in the context of changes in device flow-profiles, cardiac function as measured with echocardiography and Swan Ganz hemodynamic monitoring are warranted. Our study was not powered to find any correlations between the change in device geometry and survival or other complication rates, but these outcome measures will be important considerations in future analyses.
Last, there was a small increase in the incidence of recurrent PT compared with that in the general CF-LVAD cohort although our study was not powered to reach statistical significance. Future efforts should evaluate whether certain patients possess intrinsically higher PT risks regardless of surgical technique, which should be considered a potential relative contraindication when implanting an LVAD.
This study is limited in its retrospective approach. Although the human aspects of error were mitigated by having a single observer, two confirming observers, and strict inclusion criteria, the study is also limited by the inherent variations in technique, position, and availability among chest radiographs. Also, ICA and PPD were chosen in our study because chest radiographs are routinely taken post-LVAD implantation at our institution. However, evaluating other measurements that approximate device-organ remodeling may be beneficial for future studies.
1. 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
2. Starling RC, Naka Y, Boyle AJ, et al. Results of the post-U.S. Food and Drug Administration-approval study with a continuous flow left ventricular assist device as a bridge to heart transplantation: A prospective study using the INTERMACS (Interagency Registry for Mechanically Assisted Circulatory Support). J Am Coll Cardiol. 2011;57:1890–1898
3. Kirklin JK, Naftel DC, Pagani FD, et al. Sixth INTERMACS annual report: A 10,000-patient database. J Heart Lung Transplant. 2014;33:555–564
4. Rogers JG, Aaronson KD, Boyle AJ, et al.HeartMate II
Investigators. Continuous flow left ventricular assist device improves functional capacity and quality of life of advanced heart failure patients. J Am Coll Cardiol. 2010;55:1826–1834
5. Park SJ, Milano CA, Tatooles AJ, et al.HeartMate II
Clinical Investigators. Outcomes in advanced heart failure patients with left ventricular assist devices for destination therapy. Circ Heart Fail. 2012;5:241–248
6. Hasin T, Marmor Y, Kremers W, et al. Readmissions after implantation of axial flow left ventricular assist device. J Am Coll Cardiol. 2013;61:153–163
7. Forest SJ, Bello R, Friedmann P, et al. Readmissions after ventricular assist device: Etiologies, patterns, and days out of hospital. Ann Thorac Surg. 2013;95:1276–1281
8. 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
9. Rame JE, Atluri P, Acker MA. Unexpected abrupt increase in left ventricular assist device thrombosis. N Engl J Med. 2014;370:1466–1467
10. Boyle AJ, Russell SD, Teuteberg JJ, et al. Low thromboembolism and pump thrombosis
with the HeartMate II
left ventricular assist device: Analysis of outpatient anti-coagulation. J Heart Lung Transplant. 2009;28:881–887
11. John R, Kamdar F, Liao K, et al. Low thromboembolic risk for patients with the Heartmate II
left ventricular assist device. J Thorac Cardiovasc Surg. 2008;136:1318–1323
12. Bishop MA, Streiff MB, Ensor CR, Tedford RJ, Russell SD, Ross PA. Pharmacist-managed international normalized ratio patient self-testing is associated with increased time in therapeutic range in patients with left ventricular assist devices at an academic medical center. ASAIO J. 2014;60:193–198
13. 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
14. Katz JN, Jensen BC, Chang PP, Myers SL, Pagani FD, Kirklin JK. A multicenter analysis of clinical hemolysis in patients supported with durable, long-term left ventricular assist device therapy. J Heart Lung Transplant. 2015;34:701–709
15. 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
16. Bartoli CR, Ailawadi G, Kern JA. Diagnosis, nonsurgical management, and prevention of LVAD thrombosis. J Card Surg. 2014;29:83–94
17. Eckman PM, John R. Bleeding and thrombosis in patients with continuous-flow ventricular assist devices. Circulation. 2012;125:3038–3047