- Email: [email protected]
In vivo tissue reaction within the outflow conduit in patients supported by HeartWare HVAD
Journal Pre-proof In vivo tissue reaction within the outflow conduit in patients supported by HeartWare HVAD. Jain P, Robson D, Shehab S, Muthiah K, Jansz P, Qiu MR, Barrett W, Sivasubramaniam V, Kumaradevan N, Macdonald PS, Hayward CS PII:
S1054-8807(19)30321-7
DOI:
https://doi.org/10.1016/j.carpath.2019.107156
Reference:
CVP 107156
To appear in:
Cardiovascular Pathology
Received Date: 1 August 2019 Revised Date:
23 September 2019
Accepted Date: 24 September 2019
Please cite this article as: P J, D R, S S, K M, P J, MR Q, W B, V S, N K, PS M, CS H, In vivo tissue reaction within the outflow conduit in patients supported by HeartWare HVAD., Cardiovascular Pathology, https://doi.org/10.1016/j.carpath.2019.107156. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Inc.
In vivo tissue reaction within the outflow conduit in patients supported by HeartWare HVAD.
Running title: Outflow conduit tissue reaction in LVAD patients. Authors: Jain P1,2, Robson D1, Shehab S3, Muthiah K1,3, Jansz P1, Qiu MR4, Barrett W4, Sivasubramaniam V4, Kumaradevan N4, Macdonald PS1,2,3, Hayward CS1,2,3. Author affiliations: 1. 2. 3. 4.
Cardiology Department, St Vincent’s Hospital, Sydney, Australia University of New South Wales, Sydney Australia Victor Chang Cardiac Research Institute, Sydney Australia Anatomical Pathology Department, St Vincent’s Hospital, Sydney, Australia
Corresponding author: Prof Christopher Hayward Heart Failure and Transplant Unit St Vincent’s Hospital Darlinghurst NSW 2010 Australia Email: [email protected] Phone: +61 2 8382 6880 Fax: +61 2 8382 6881
Word count: 1988 Key words: Continuous-flow left ventricular assist device, outflow conduit, tissue reaction, foreign body, pump flow
Abstract: Aim: The frequency, extent and nature of tissue ingrowth within the continuous-flow left ventricular assist device outflow conduit has not been systematically assessed. We sought to characterise conduit histopathology at explant in a cohort of patients with HeartWare HVAD and assess the effect on pump performance. Methods: Patients undergoing routine histopathological assessment of a HeartWare HVAD removed at transplant or autopsy were assessed. Outflow conduits were examined macroscopically, and visible tissue was sectioned for microscopic evaluation. In patients who had undergone prior contrast enhanced computerised tomography (CT) with HVAD in situ, the outflow conduit was measured at the aortic anastomosis and 5cm proximal to the anastomosis, in the axial and sagittal planes. All patients had their pump flow, flow pulsatility, current and speed determined from log files examined at 1, 3, 6, 9 and 12 months post LVAD implantation. Results: 25 consecutive patients were assessed (24 LVAD, 1 BiVAD). Of the 26 outflow grafts assessed, there was evidence of tissue ingrowth reaction in 24 (92%). The most common site was the distal anastomosis (18/24, 75%), with the graft body involved in 14/24 (58%). Microscopic evaluation revealed acute inflammatory infiltrate in 4/24 grafts (17%), chronic inflammatory infiltrate in 14/24 (58%), neointima formation in 18/24 (75%) and fibrosis in 18/24 (75%). The median depth of tissue was 1mm (range 0-2mm). The mean conduit diameter was 9.5±0.6mm at the aortic anastomosis compared to 11.1±0.5mm 5cm proximal to the anastomosis (p<0.0001). In patients with unchanged pump speed one month post implant, analysis of log files revealed a significant (5.8±8.6%) decrease in pump flow (4.65±0.86 vs 4.38±0.92L/min, p=0.01) as well as flow pulsatility (5.00±1.10 vs 4.16±1.05L/min, p=0.006). Conclusions: There is evidence of tissue formation within the HVAD outflow conduit in the vast majority of patients, most commonly located at the aortic anastomosis. This is associated with significantly decreased pump flow over time.
Background and aims It is well established that continuous-flow left ventricular assist devices (cf-LVADs) are susceptible to acute deterioration in pump performance as a result of outflow graft complications such as thrombus formation, extrinsic compression, kinking or twisting[1-3]. Increasingly, however, it is recognized that chronic changes such as stenosis at the aortic anastomosis can gradually affect pump flow and ventricular unloading over time[4]. Additionally, it has been demonstrated that changes in outflow graft diameter can have dramatic effects on conduit gradient[5] as well as blood flow within the great vessels[6]. Despite these advances in our understanding, a number of questions remain. These include the prevalence, anatomic distribution, extent and histopathological nature of tissue ingrowth within the outflow graft in clinically stable patients; the underlying cause of accelerated tissue ingrowth; and the effects of such changes, if present, on long-term pump performance. In the era of destination therapy with cf-LVADs, understanding these changes is key to optimizing patient outcomes and quality of life through better management of existing pumps in the short term, and better pump design in the long-term. We therefore sought to characterize the histopathological changes that occur within the outflow graft, as well as the changes in pump flow over time, in a cohort of consecutive patients undergoing removal of the HeartWare HVAD device. Methods: Stable patients with HVAD (HeartWare, Medtronic) in situ who were undergoing device removal at the time of transplantation or autopsy were included in the study. Ethics approval was obtained from our Hospital’s Ethics Review Committee (Reference 2019/STE00072). Outflow grafts were examined macroscopically for the presence, extent and location of tissue ingrowth. If present, this tissue was removed from the outflow graft and sectioned for microscopic analysis. Inflammatory reactions were defined as ‘acute’ by the presence of neutrophils; ‘giant cell’ by the presence of multinucleated giant cells and/or granulomas; and ‘chronic’ by the presence of other chronic
inflammatory cells. Architectural tissue changes were classified as ‘fibrosis’, ‘neointima’ or ‘atherosclerosis’ at the discretion of the examining anatomical pathologist. All patients had their pump flow, flow pulsatility, current and speed retrospectively assessed at 1, 3, 6, 9 and 12 months post-implant from pump controller log files. Pump flow and flow pulsatility were corrected offline for changes in hematocrit[7]. Patients who had undergone contrast-enhanced computerized tomography (CT) had their outflow graft diameter measured at the aortic anastomosis and 5cm proximal to the anastomosis in the axial and sagittal planes. Pump flow changes over time and CT diameter were statistically assessed using the paired t-test or Wilcoxon signed-rank test for normally distributed and non-normally distributed data respectively. A p-value of <0.05 was considered significant. Results 25 consecutive patients were assessed (24 LVAD, 1 BiVAD). Baseline patient details are outlined in Table 1. One device was assessed at autopsy following a fatal intracerebral hemorrhage. The remainder were assessed following heart transplantation. There were no device explants performed for myocardial recovery or pump malfunction. Twenty one out of 26 (81%) had undergone device implantation via median sternotomy, while the remainder (5 patients, 19%) had undergone a minimally invasive approach via left lateral thoracotomy. Of the 26 outflow grafts assessed, there was evidence of tissue ingrowth in 24 (92%). An example is shown in figure 1. The most common site was the distal anastomosis (18/24, 75%), with the graft body involved in 14/24 (58%). Microscopic evaluation revealed acute inflammatory infiltrate in 4/24 grafts (17%), granulomatous infiltrate in 11/24 (46%), and other chronic inflammatory infiltrate in 14/24 (58%). The majority of grafts had evidence of either neointima formation (18/24, 75%) and/or fibrosis (18/24, 75%), with atherosclerosis present in 2/24 (8%). The median depth of tissue was 1mm (range 0-2mm). Typical microscopic findings are demonstrated in figure 2.
The median time on-pump for those with acute inflammatory infiltrate was 158 days (range 92-327 days), granulomatous infiltrate 238 days (range 92-482 days) other chronic inflammatory infiltrate 306 days (range 92-482 days), fibrotic changes 294 days (range 92-1032 days) and neo-intima formation 307 days (range 119-1032 days). These differences were not statistically significant. Twelve out of 25 patients (48%) had microbiological evidence of bloodstream (3/25, 12%) or driveline (9/25, 36%) infection at some stage whilst supported on-pump. There was no association between presence of device-related infection and either inflammatory (p=0.70) or neointimal/fibrotic changes (p=0.61) within the conduit. Contrast enhanced CT was performed in 19 patients at a median of 87 days post implant (range 10296 days). All outflow grafts were widely patent. The mean conduit diameter was 9.5±0.6mm at the aortic anastomosis compared to 11.1±0.5mm 5cm proximal to the anastomosis (p<0.0001). Analysis of patient log files revealed a trend towards decreased pump flow over time (4.61±0.76 vs 4.43±0.80L/min, p=0.07), despite a trend towards increased pump speed (2488±109 vs 2532±125 RPM, p=0.07). Where pump speed remained unchanged compared to baseline, there was a significant (5.8±8.6%) decrease in both pump flow (4.65±0.86 vs 4.38±0.92L/min, p=0.01) and flow pulsatility (5.00±1.10 vs 4.16±1.05L/min, p=0.006) over a median time period of 5 months (range 211 months). Discussion With approval of cf-LVAD devices for destination therapy in severe heart failure, a greater understanding the long-term pump-patient interaction – comprising not only the effects of the pump on the body, but also the effects of the body on the pump – is required. It is well recognized in the non-LVAD setting that synthetic vascular graft insertion can cause a foreign body reaction and neointima formation, with a demonstrated association between these processes and graft failure in one analysis of arteriovenous fistulae[8]. In the setting of cf-LVADs, meanwhile, case reports have
identified stenosis at the aortic anastomosis as a rare but important cause of pump dysfunction[1, 4]. Our study is the first to build on this anecdotal experience by systematically assessing the human tissue response to the presence of the cf-LVAD outflow graft, and long-term changes in pump flow, in a consecutive cohort of patients without evidence of pump dysfunction or clinical outflow obstruction. The key findings of our study are: firstly, that some degree of tissue ingrowth into the outflow graft is a near-universal phenomenon, and most commonly occurs at the aortic anastomosis; second, that this reaction may variably be comprised of an acute inflammatory, chronic inflammatory, fibrotic or neointimal reaction; and third, that this is associated with a small but measurable decrease in pump flow and flow pulsatility over time. It is increasingly recognized that outflow graft characteristics have a significant impact on the function of rotary pumps through afterload effects. It has been demonstrated that hemodynamically significant gradients are generated across the HVAD outflow graft in a pulsatile circulation; furthermore, minimum luminal area has a significant, non-linear effect on the magnitude of these gradients[5]. From the baseline HVAD conduit cross-sectional area of 0.79cm2 – smaller than the widely accepted threshold for severe aortic stenosis[9] – a circumferential tissue ingrowth depth of 1mm would result in a 36% reduction to 0.50cm2. This would be expected to be sufficient to cause a significant reduction in pump performance due to increased pressure head across the cf-LVAD impeller. This hypothesis is reflected in the results of our study, in which there were statistically significant reductions in pump flow and flow pulsatility over time where pump speed was kept constant. Although the absolute magnitude of pump flow reduction across the whole cohort was small, the upper range limit – 0.95L/min – certainly represents a clinically significant decrease. Our observed trend towards increased pump speed over time lends further weight to the hypothesis that, in some patients at least, the reduction in pump flow over time may be significant enough to warrant clinician intervention.
The precise characteristics of the tissue response was highly variable in terms of extent, location and histopathological features; moreover, the factors that influence these characteristics are unclear. There was no association between time on-pump and depth of neointima assessed histopathologically. This finding is not unexpected, given the results of a previous large animal study assessing neointima formation within synthetic coronary grafts, which showed that following an initial period of endothelial and smooth muscle cell hyperplasia lasting 3 months, a ‘steady-state’ is achieved in which there is turnover but not further accumulation of neointima[10]. Additionally, neither time on-pump nor the presence of driveline or bloodstream infection significantly affected the location of tissue ingrowth, or whether or not this tissue was inflammatory in nature. Finally, the anastomosis angle with the aorta and flow characteristics within the outflow graft were not specifically assessed in this study; given the demonstrated association in the setting of arteriovenous fistulae between neointimal growth and both low wall shear stress and turbulent flow [11-13], these factors warrant dedicated study. Contrast-enhanced CT demonstrated a significantly smaller graft diameter at the anastomosis than within the body of the graft, with an example shown in figure 3. There are a number of possible explanations for this discrepancy. Certainly, the aortic anastomosis was the most common site of tissue ingrowth in our patients, and so the CT findings may reflect relative stenosis at this point. Additionally, suturing of the graft is likely to prevent expansion at the anastomotic site, both on a beat-to-beat basis during systole, and due to stretching of graft material in the long-term. Regardless of whether the underlying cause is pathological or a result of outflow graft design and implant technique, the absolute and relative reductions in graft diameter at the anastomosis may have significant effects on both flow rate and turbulence. The long-term effects of tissue ingrowth on pump performance could be mitigated using either a pre-emptive approach – through modifications in outflow graft design – or alternatively through selective treatment of clinically significant lesions. Pre-emptively, the graft diameter could be
increased, either along its entire length or at its distal end only. An increased diameter would, given the non-linear effect of graft diameter on pressure losses[5] mean that the same degree of diameter reduction would result in a smaller increase in pressure gradient and therefore a smaller reduction in pump performance. It should be noted that the HeartMate 3 device employs a 13mm outflow graft as opposed to the 10mm HeartWare HVAD graft assessed in this study and is therefore already less susceptible to the effects of tissue ingrowth. A second potential modification would involve the use of drug-eluting grafts; both Paclitaxel and Sirolimus-eluting grafts have demonstrated efficacy in preventing synthetic arteriovenous fistula restenosis in animal studies, although they are not yet marketed for human use[14, 15]. Finally, drug-eluting balloon angioplasty has demonstrated efficacy in the treatment of arteriovenous fistula restenosis, and may similarly represent a potential treatment modality in cases of clinically relevant cf-LVAD outflow graft stenosis[16]. This approach may ultimately be the most appropriate, given that while tissue ingrowth is almost universally present, it is not associated with clinically significant flow reduction in the majority of patients. The limitations of our study warrant discussion. This was a retrospective analysis, and as such is hypothesis-generating only. While we established that pump flow decreases over the same time period that tissue ingrowth occurs, we were not able to definitively establish a causal relationship between these phenomena. There was no correlation on an individual patient level between depth of tissue ingrowth histologically and the magnitude of pump flow reduction, although it should be noted that histologic assessment of tissue depth may be prone to significant measurement error. There was similarly no relationship between conduit diameter on contrast-enhanced CT and magnitude of pump flow reduction, a finding that is not surprising given that CT was performed according to clinical need and therefore at variable time post-LVAD implant. The observed reduction in pump flow may therefore have been due to other changes in loading conditions; the specific contribution of outflow graft diameter reduction to decreased pump flow cannot be accurately delineated from our study.
Conclusions There is evidence of tissue formation within the HVAD outflow graft in the vast majority of patients, most commonly located at the aortic anastomosis and variably consisting of an inflammatory infiltrate, neointima or fibrosis. This is associated with significantly decreased pump flow over time and may have implications for the design of future outflow grafts.
References 1. 2.
3.
4. 5. 6.
7. 8. 9.
10.
11.
12. 13.
Scandroglio, A.M., et al., Diagnosis and Treatment Algorithm for Blood Flow Obstructions in Patients With Left Ventricular Assist Device. J Am Coll Cardiol, 2016. 67(23): p. 2758-2768. Gruger, T., et al., Late post-pump blood flow obstruction in a novel left ventricular assist device: The unusual case of a twisted outflow graft. J Thorac Cardiovasc Surg, 2018. 155(1): p. e33-e35. Trankle, C.R., et al., Internal Versus External Compression of a Left Ventricular Assist Device Outflow Graft: Diagnosis With Intravascular Ultrasound and Treatment With Stenting. Circ Heart Fail, 2018. 11(5): p. e004959. Pieri, M., et al., Heart Failure After 5 Years on LVAD: Diagnosis and Treatment of Outflow Graft Obstruction. ASAIO J, 2017. 63(1): p. e1-e2. Jain, P., et al., Pulsatile Conduit Pressure Gradients in the HeartWare HVAD. ASAIO J, 2019. 65(5): p. 489-494. Bhat, S., et al., Effect of Outflow Graft Size on Flow in the Aortic Arch and Cerebral Blood Flow in Continuous Flow Pumps: Possible Relevance to Strokes. ASAIO J, 2017. 63(2): p. 144149. Granegger, M., et al., Development of a pump flow estimator for rotary blood pumps to enhance monitoring of ventricular function. Artif Organs, 2012. 36(8): p. 691-9. Mehta, R.I., et al., Pathology of explanted polytetrafluoroethylene vascular grafts. Cardiovasc Pathol, 2011. 20(4): p. 213-21. Nishimura, R.A., et al., 2017 AHA/ACC Focused Update of the 2014 AHA/ACC Guideline for the Management of Patients With Valvular Heart Disease: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. J Am Coll Cardiol, 2017. 70(2): p. 252-289. Clowes, A.W., T.R. Kirkman, and M.M. Clowes, Mechanisms of arterial graft failure. II. Chronic endothelial and smooth muscle cell proliferation in healing polytetrafluoroethylene prostheses. J Vasc Surg, 1986. 3(6): p. 877-84. Berceli, S.A., et al., Flow-induced neointimal regression in baboon polytetrafluoroethylene grafts is associated with decreased cell proliferation and increased apoptosis. J Vasc Surg, 2002. 36(6): p. 1248-55. Binns, R.L., et al., Optimal graft diameter: effect of wall shear stress on vascular healing. J Vasc Surg, 1989. 10(3): p. 326-37. Mattsson, E.J., et al., Increased blood flow induces regression of intimal hyperplasia. Arterioscler Thromb Vasc Biol, 1997. 17(10): p. 2245-9.
14.
15. 16.
Baek, I., et al., Suppression of neointimal hyperplasia by sirolimus-eluting expanded polytetrafluoroethylene (ePTFE) haemodialysis grafts in comparison with paclitaxel-coated grafts. Nephrol Dial Transplant, 2012. 27(5): p. 1997-2004. Glickman, M., Drug eluting grafts for hemodialysis access. J Vasc Access, 2017. 18(Suppl. 1): p. 53-55. Khawaja, A.Z., et al., Systematic review of drug eluting balloon angioplasty for arteriovenous haemodialysis access stenosis. J Vasc Access, 2016. 17(2): p. 103-10.
Table 1: Baseline characteristics Baseline Characteristics (n=25) Age % male Median INTERMACS class Etiology (%) Ischemic Non-ischemic Comorbidities (%) Type 2 Diabetes Mellitus Chronic renal impairment Stage 3-4 Stage 5 Previous stroke Configuration (%) LVAD BiVAD % bridge to transplant Implant technique (%) Minimally invasive Median sternotomy Medications (%) Beta blocker RAAS blockade Reason for explant (%) Transplantation Autopsy Myocardial recovery Days on-pump (median, IQR) Clinical outcome (%) Alive Died on-pump Died following transplantation LVAD-related adverse events (%) Ischemic stroke or TIA Hemorrhagic stroke Pump thrombus Gastrointestinal bleeding Right heart failure Driveline infection
55.7 (9.4) 80 2 44 56 24 32 0 12 96 4 100 20 80 68 48 96 4 0 298 (223-352) 76 4 20 28 12 12 8 16 36
Figure 1: Macroscopic specimen demonstrating the anastomosis between HVAD outflow conduit and aorta following LVAD explantation at the time of cardiac transplant. 1A: Anastomosis and
outflow conduit intact. 1B: The outflow conduit has been transected to expose a layer of new tissue formation within the conduit at the anastomosis.
Figure 2: Typical histopathological findings (hematoxylin and eosin-stained sections). 2A: Focally prominent inflammation, including moderate numbers of plasma cells and neutrophils (inset: highpower field). 2B: Prominent chronic inflammation including large numbers of plasma cells and some lymphocytes. 2C: Fibrotic tissue with focal mild atheromatous change in addition to a multinucleated giant cell reaction. 2D: Fibrotic tissue consistent with large vessel intima and occasional chronic inflammatory cells.
Figure 3: Contrast-enhanced computerized tomography demonstrating subtle stenosis at the outflow graft-aorta anastomosis. 2A-B: axial plane. 2C-D: sagittal plane.
Supplementary appendix: Summary table outlining patient and tissue characteristics at HVAD explant. Patient
Age at explant
Days on pump
Pump location
Tissue ingrowth?
Location
Depth
Microscopic Description
1
52
92
LVAD
Yes
Aortic anastomosis
1mm
Foreign body giant cell reaction
2
68
93
LVAD
No Aortic anastomosis and body of graft
2mm
Plasma cells, giant cells and neutrophils
Yes
Aortic anastomosis
2mm
Fibromyxoid stromal tissue, foreign body giant cell reaction with histocytes, plasma cells
2mm
Fibromyxoid stroma with fibroblasts, histiocytes, occasional giant cells
3
61
119
LVAD
Yes
4
69
188
LVAD
No
5
37
198
LVAD
5
37
198
RVAD
Yes
Pulmonary anastomosis and body of graft
6
42
220
LVAD
Yes
Aortic anastomosis
11.5mm
7
54
233
LVAD
Yes
Aortic anastomosis
NS
8
70
238
LVAD
Yes
Body of graft
<1mm
9
53
259
LVAD
Yes
Body of graft
2mm
10
66
265
LVAD
Yes
Aortic anastomosis
<1mm
Aortic anastomosis and body of graft
1mm
11
40
284
LVAD
Yes
12
62
290
LVAD
Yes
13
54
305
LVAD
Yes
14
41
306
LVAD
Yes
15
60
308
LVAD
16
63
312
LVAD
Aortic anastomosis Aortic anastomosis
1mm 1mm
Body of graft
<1mm
Yes
Body of graft
<1mm
Yes
Aortic anastomosis
<1mm
Myxoid fibrosis, giant cells, plasma cells, epithelioid cells Intima-like myxoid fibrotic tissue with scattered plasma cells and lymphocytes Giant cell reaction, myxoid fibrosis, neointima Fibrotic tissue, adipocytes, cardiomyocytes, peripheral nerve bundles Foreign body, giant cells, plasma cells, foamy macrophages Prominent chronic inflammation - plasma cells, lymphoid aggregates, eosinophils, neovascularisation Reactive changes Myxoid stromal tissue. No inflammation Neointima, focal atheromatous change, foamy macrophages, giant cell foreign body reaction Myxoid fibrotic tissue Foreign body giant cell reaction
17
66
327
LVAD
Yes
Aortic anastomosis
<1mm
18
56
350
LVAD
Yes
Body of graft
<1mm
Yes
Aortic anastomosis and body of graft
1mm
Yes
Aortic anastomosis and body of graft
<1mm
19
20
59
57
353
416
LVAD
LVAD
21
51
435
LVAD
Yes
Body of graft
1mm
22
59
449
LVAD
Yes
Aortic anastomosis
1mm
23
60
463
LVAD
Yes
Body of graft
<1mm
Yes
Aortic anastomosis and body of graft
1mm
Yes
Aortic anastomosis and Body of graft
1mm
24
25
55
52
482
1032
LVAD
LVAD
Scattered neutrophils Neointima with minimal revascularisation Fibrotic tissue consistent with large vessel intima, scattered chronic inflammatory cells Chronic inflammatory changes with plasma cells Mild myxoid changes, mild atheromatous change, chronic inflammatory changes (lymphocytes) Intima-like myxoid fibrotic tissue with scattered plasma cells, lymphocytes and MNG cells Neo-intima. No significant inflammation Foreign body giant cell reaction, mixed inflammation and granulomata Fibroblast proliferation, endothelial lining
Highlights • • •
Tissue ingrowth into the HeartWare HVAD outflow graft is a near-universal finding It may comprise an acute or chronic inflammatory, fibrotic or neointimal reaction This is associated with a decrease in pump flow and flow pulsatility over time
Conflicts of Interest There are no relevant conflicts of interest.
S1054-8807(19)30321-7
DOI:
https://doi.org/10.1016/j.carpath.2019.107156
Reference:
CVP 107156
To appear in:
Cardiovascular Pathology
Received Date: 1 August 2019 Revised Date:
23 September 2019
Accepted Date: 24 September 2019
Please cite this article as: P J, D R, S S, K M, P J, MR Q, W B, V S, N K, PS M, CS H, In vivo tissue reaction within the outflow conduit in patients supported by HeartWare HVAD., Cardiovascular Pathology, https://doi.org/10.1016/j.carpath.2019.107156. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Inc.
In vivo tissue reaction within the outflow conduit in patients supported by HeartWare HVAD.
Running title: Outflow conduit tissue reaction in LVAD patients. Authors: Jain P1,2, Robson D1, Shehab S3, Muthiah K1,3, Jansz P1, Qiu MR4, Barrett W4, Sivasubramaniam V4, Kumaradevan N4, Macdonald PS1,2,3, Hayward CS1,2,3. Author affiliations: 1. 2. 3. 4.
Cardiology Department, St Vincent’s Hospital, Sydney, Australia University of New South Wales, Sydney Australia Victor Chang Cardiac Research Institute, Sydney Australia Anatomical Pathology Department, St Vincent’s Hospital, Sydney, Australia
Corresponding author: Prof Christopher Hayward Heart Failure and Transplant Unit St Vincent’s Hospital Darlinghurst NSW 2010 Australia Email: [email protected] Phone: +61 2 8382 6880 Fax: +61 2 8382 6881
Word count: 1988 Key words: Continuous-flow left ventricular assist device, outflow conduit, tissue reaction, foreign body, pump flow
Abstract: Aim: The frequency, extent and nature of tissue ingrowth within the continuous-flow left ventricular assist device outflow conduit has not been systematically assessed. We sought to characterise conduit histopathology at explant in a cohort of patients with HeartWare HVAD and assess the effect on pump performance. Methods: Patients undergoing routine histopathological assessment of a HeartWare HVAD removed at transplant or autopsy were assessed. Outflow conduits were examined macroscopically, and visible tissue was sectioned for microscopic evaluation. In patients who had undergone prior contrast enhanced computerised tomography (CT) with HVAD in situ, the outflow conduit was measured at the aortic anastomosis and 5cm proximal to the anastomosis, in the axial and sagittal planes. All patients had their pump flow, flow pulsatility, current and speed determined from log files examined at 1, 3, 6, 9 and 12 months post LVAD implantation. Results: 25 consecutive patients were assessed (24 LVAD, 1 BiVAD). Of the 26 outflow grafts assessed, there was evidence of tissue ingrowth reaction in 24 (92%). The most common site was the distal anastomosis (18/24, 75%), with the graft body involved in 14/24 (58%). Microscopic evaluation revealed acute inflammatory infiltrate in 4/24 grafts (17%), chronic inflammatory infiltrate in 14/24 (58%), neointima formation in 18/24 (75%) and fibrosis in 18/24 (75%). The median depth of tissue was 1mm (range 0-2mm). The mean conduit diameter was 9.5±0.6mm at the aortic anastomosis compared to 11.1±0.5mm 5cm proximal to the anastomosis (p<0.0001). In patients with unchanged pump speed one month post implant, analysis of log files revealed a significant (5.8±8.6%) decrease in pump flow (4.65±0.86 vs 4.38±0.92L/min, p=0.01) as well as flow pulsatility (5.00±1.10 vs 4.16±1.05L/min, p=0.006). Conclusions: There is evidence of tissue formation within the HVAD outflow conduit in the vast majority of patients, most commonly located at the aortic anastomosis. This is associated with significantly decreased pump flow over time.
Background and aims It is well established that continuous-flow left ventricular assist devices (cf-LVADs) are susceptible to acute deterioration in pump performance as a result of outflow graft complications such as thrombus formation, extrinsic compression, kinking or twisting[1-3]. Increasingly, however, it is recognized that chronic changes such as stenosis at the aortic anastomosis can gradually affect pump flow and ventricular unloading over time[4]. Additionally, it has been demonstrated that changes in outflow graft diameter can have dramatic effects on conduit gradient[5] as well as blood flow within the great vessels[6]. Despite these advances in our understanding, a number of questions remain. These include the prevalence, anatomic distribution, extent and histopathological nature of tissue ingrowth within the outflow graft in clinically stable patients; the underlying cause of accelerated tissue ingrowth; and the effects of such changes, if present, on long-term pump performance. In the era of destination therapy with cf-LVADs, understanding these changes is key to optimizing patient outcomes and quality of life through better management of existing pumps in the short term, and better pump design in the long-term. We therefore sought to characterize the histopathological changes that occur within the outflow graft, as well as the changes in pump flow over time, in a cohort of consecutive patients undergoing removal of the HeartWare HVAD device. Methods: Stable patients with HVAD (HeartWare, Medtronic) in situ who were undergoing device removal at the time of transplantation or autopsy were included in the study. Ethics approval was obtained from our Hospital’s Ethics Review Committee (Reference 2019/STE00072). Outflow grafts were examined macroscopically for the presence, extent and location of tissue ingrowth. If present, this tissue was removed from the outflow graft and sectioned for microscopic analysis. Inflammatory reactions were defined as ‘acute’ by the presence of neutrophils; ‘giant cell’ by the presence of multinucleated giant cells and/or granulomas; and ‘chronic’ by the presence of other chronic
inflammatory cells. Architectural tissue changes were classified as ‘fibrosis’, ‘neointima’ or ‘atherosclerosis’ at the discretion of the examining anatomical pathologist. All patients had their pump flow, flow pulsatility, current and speed retrospectively assessed at 1, 3, 6, 9 and 12 months post-implant from pump controller log files. Pump flow and flow pulsatility were corrected offline for changes in hematocrit[7]. Patients who had undergone contrast-enhanced computerized tomography (CT) had their outflow graft diameter measured at the aortic anastomosis and 5cm proximal to the anastomosis in the axial and sagittal planes. Pump flow changes over time and CT diameter were statistically assessed using the paired t-test or Wilcoxon signed-rank test for normally distributed and non-normally distributed data respectively. A p-value of <0.05 was considered significant. Results 25 consecutive patients were assessed (24 LVAD, 1 BiVAD). Baseline patient details are outlined in Table 1. One device was assessed at autopsy following a fatal intracerebral hemorrhage. The remainder were assessed following heart transplantation. There were no device explants performed for myocardial recovery or pump malfunction. Twenty one out of 26 (81%) had undergone device implantation via median sternotomy, while the remainder (5 patients, 19%) had undergone a minimally invasive approach via left lateral thoracotomy. Of the 26 outflow grafts assessed, there was evidence of tissue ingrowth in 24 (92%). An example is shown in figure 1. The most common site was the distal anastomosis (18/24, 75%), with the graft body involved in 14/24 (58%). Microscopic evaluation revealed acute inflammatory infiltrate in 4/24 grafts (17%), granulomatous infiltrate in 11/24 (46%), and other chronic inflammatory infiltrate in 14/24 (58%). The majority of grafts had evidence of either neointima formation (18/24, 75%) and/or fibrosis (18/24, 75%), with atherosclerosis present in 2/24 (8%). The median depth of tissue was 1mm (range 0-2mm). Typical microscopic findings are demonstrated in figure 2.
The median time on-pump for those with acute inflammatory infiltrate was 158 days (range 92-327 days), granulomatous infiltrate 238 days (range 92-482 days) other chronic inflammatory infiltrate 306 days (range 92-482 days), fibrotic changes 294 days (range 92-1032 days) and neo-intima formation 307 days (range 119-1032 days). These differences were not statistically significant. Twelve out of 25 patients (48%) had microbiological evidence of bloodstream (3/25, 12%) or driveline (9/25, 36%) infection at some stage whilst supported on-pump. There was no association between presence of device-related infection and either inflammatory (p=0.70) or neointimal/fibrotic changes (p=0.61) within the conduit. Contrast enhanced CT was performed in 19 patients at a median of 87 days post implant (range 10296 days). All outflow grafts were widely patent. The mean conduit diameter was 9.5±0.6mm at the aortic anastomosis compared to 11.1±0.5mm 5cm proximal to the anastomosis (p<0.0001). Analysis of patient log files revealed a trend towards decreased pump flow over time (4.61±0.76 vs 4.43±0.80L/min, p=0.07), despite a trend towards increased pump speed (2488±109 vs 2532±125 RPM, p=0.07). Where pump speed remained unchanged compared to baseline, there was a significant (5.8±8.6%) decrease in both pump flow (4.65±0.86 vs 4.38±0.92L/min, p=0.01) and flow pulsatility (5.00±1.10 vs 4.16±1.05L/min, p=0.006) over a median time period of 5 months (range 211 months). Discussion With approval of cf-LVAD devices for destination therapy in severe heart failure, a greater understanding the long-term pump-patient interaction – comprising not only the effects of the pump on the body, but also the effects of the body on the pump – is required. It is well recognized in the non-LVAD setting that synthetic vascular graft insertion can cause a foreign body reaction and neointima formation, with a demonstrated association between these processes and graft failure in one analysis of arteriovenous fistulae[8]. In the setting of cf-LVADs, meanwhile, case reports have
identified stenosis at the aortic anastomosis as a rare but important cause of pump dysfunction[1, 4]. Our study is the first to build on this anecdotal experience by systematically assessing the human tissue response to the presence of the cf-LVAD outflow graft, and long-term changes in pump flow, in a consecutive cohort of patients without evidence of pump dysfunction or clinical outflow obstruction. The key findings of our study are: firstly, that some degree of tissue ingrowth into the outflow graft is a near-universal phenomenon, and most commonly occurs at the aortic anastomosis; second, that this reaction may variably be comprised of an acute inflammatory, chronic inflammatory, fibrotic or neointimal reaction; and third, that this is associated with a small but measurable decrease in pump flow and flow pulsatility over time. It is increasingly recognized that outflow graft characteristics have a significant impact on the function of rotary pumps through afterload effects. It has been demonstrated that hemodynamically significant gradients are generated across the HVAD outflow graft in a pulsatile circulation; furthermore, minimum luminal area has a significant, non-linear effect on the magnitude of these gradients[5]. From the baseline HVAD conduit cross-sectional area of 0.79cm2 – smaller than the widely accepted threshold for severe aortic stenosis[9] – a circumferential tissue ingrowth depth of 1mm would result in a 36% reduction to 0.50cm2. This would be expected to be sufficient to cause a significant reduction in pump performance due to increased pressure head across the cf-LVAD impeller. This hypothesis is reflected in the results of our study, in which there were statistically significant reductions in pump flow and flow pulsatility over time where pump speed was kept constant. Although the absolute magnitude of pump flow reduction across the whole cohort was small, the upper range limit – 0.95L/min – certainly represents a clinically significant decrease. Our observed trend towards increased pump speed over time lends further weight to the hypothesis that, in some patients at least, the reduction in pump flow over time may be significant enough to warrant clinician intervention.
The precise characteristics of the tissue response was highly variable in terms of extent, location and histopathological features; moreover, the factors that influence these characteristics are unclear. There was no association between time on-pump and depth of neointima assessed histopathologically. This finding is not unexpected, given the results of a previous large animal study assessing neointima formation within synthetic coronary grafts, which showed that following an initial period of endothelial and smooth muscle cell hyperplasia lasting 3 months, a ‘steady-state’ is achieved in which there is turnover but not further accumulation of neointima[10]. Additionally, neither time on-pump nor the presence of driveline or bloodstream infection significantly affected the location of tissue ingrowth, or whether or not this tissue was inflammatory in nature. Finally, the anastomosis angle with the aorta and flow characteristics within the outflow graft were not specifically assessed in this study; given the demonstrated association in the setting of arteriovenous fistulae between neointimal growth and both low wall shear stress and turbulent flow [11-13], these factors warrant dedicated study. Contrast-enhanced CT demonstrated a significantly smaller graft diameter at the anastomosis than within the body of the graft, with an example shown in figure 3. There are a number of possible explanations for this discrepancy. Certainly, the aortic anastomosis was the most common site of tissue ingrowth in our patients, and so the CT findings may reflect relative stenosis at this point. Additionally, suturing of the graft is likely to prevent expansion at the anastomotic site, both on a beat-to-beat basis during systole, and due to stretching of graft material in the long-term. Regardless of whether the underlying cause is pathological or a result of outflow graft design and implant technique, the absolute and relative reductions in graft diameter at the anastomosis may have significant effects on both flow rate and turbulence. The long-term effects of tissue ingrowth on pump performance could be mitigated using either a pre-emptive approach – through modifications in outflow graft design – or alternatively through selective treatment of clinically significant lesions. Pre-emptively, the graft diameter could be
increased, either along its entire length or at its distal end only. An increased diameter would, given the non-linear effect of graft diameter on pressure losses[5] mean that the same degree of diameter reduction would result in a smaller increase in pressure gradient and therefore a smaller reduction in pump performance. It should be noted that the HeartMate 3 device employs a 13mm outflow graft as opposed to the 10mm HeartWare HVAD graft assessed in this study and is therefore already less susceptible to the effects of tissue ingrowth. A second potential modification would involve the use of drug-eluting grafts; both Paclitaxel and Sirolimus-eluting grafts have demonstrated efficacy in preventing synthetic arteriovenous fistula restenosis in animal studies, although they are not yet marketed for human use[14, 15]. Finally, drug-eluting balloon angioplasty has demonstrated efficacy in the treatment of arteriovenous fistula restenosis, and may similarly represent a potential treatment modality in cases of clinically relevant cf-LVAD outflow graft stenosis[16]. This approach may ultimately be the most appropriate, given that while tissue ingrowth is almost universally present, it is not associated with clinically significant flow reduction in the majority of patients. The limitations of our study warrant discussion. This was a retrospective analysis, and as such is hypothesis-generating only. While we established that pump flow decreases over the same time period that tissue ingrowth occurs, we were not able to definitively establish a causal relationship between these phenomena. There was no correlation on an individual patient level between depth of tissue ingrowth histologically and the magnitude of pump flow reduction, although it should be noted that histologic assessment of tissue depth may be prone to significant measurement error. There was similarly no relationship between conduit diameter on contrast-enhanced CT and magnitude of pump flow reduction, a finding that is not surprising given that CT was performed according to clinical need and therefore at variable time post-LVAD implant. The observed reduction in pump flow may therefore have been due to other changes in loading conditions; the specific contribution of outflow graft diameter reduction to decreased pump flow cannot be accurately delineated from our study.
Conclusions There is evidence of tissue formation within the HVAD outflow graft in the vast majority of patients, most commonly located at the aortic anastomosis and variably consisting of an inflammatory infiltrate, neointima or fibrosis. This is associated with significantly decreased pump flow over time and may have implications for the design of future outflow grafts.
References 1. 2.
3.
4. 5. 6.
7. 8. 9.
10.
11.
12. 13.
Scandroglio, A.M., et al., Diagnosis and Treatment Algorithm for Blood Flow Obstructions in Patients With Left Ventricular Assist Device. J Am Coll Cardiol, 2016. 67(23): p. 2758-2768. Gruger, T., et al., Late post-pump blood flow obstruction in a novel left ventricular assist device: The unusual case of a twisted outflow graft. J Thorac Cardiovasc Surg, 2018. 155(1): p. e33-e35. Trankle, C.R., et al., Internal Versus External Compression of a Left Ventricular Assist Device Outflow Graft: Diagnosis With Intravascular Ultrasound and Treatment With Stenting. Circ Heart Fail, 2018. 11(5): p. e004959. Pieri, M., et al., Heart Failure After 5 Years on LVAD: Diagnosis and Treatment of Outflow Graft Obstruction. ASAIO J, 2017. 63(1): p. e1-e2. Jain, P., et al., Pulsatile Conduit Pressure Gradients in the HeartWare HVAD. ASAIO J, 2019. 65(5): p. 489-494. Bhat, S., et al., Effect of Outflow Graft Size on Flow in the Aortic Arch and Cerebral Blood Flow in Continuous Flow Pumps: Possible Relevance to Strokes. ASAIO J, 2017. 63(2): p. 144149. Granegger, M., et al., Development of a pump flow estimator for rotary blood pumps to enhance monitoring of ventricular function. Artif Organs, 2012. 36(8): p. 691-9. Mehta, R.I., et al., Pathology of explanted polytetrafluoroethylene vascular grafts. Cardiovasc Pathol, 2011. 20(4): p. 213-21. Nishimura, R.A., et al., 2017 AHA/ACC Focused Update of the 2014 AHA/ACC Guideline for the Management of Patients With Valvular Heart Disease: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. J Am Coll Cardiol, 2017. 70(2): p. 252-289. Clowes, A.W., T.R. Kirkman, and M.M. Clowes, Mechanisms of arterial graft failure. II. Chronic endothelial and smooth muscle cell proliferation in healing polytetrafluoroethylene prostheses. J Vasc Surg, 1986. 3(6): p. 877-84. Berceli, S.A., et al., Flow-induced neointimal regression in baboon polytetrafluoroethylene grafts is associated with decreased cell proliferation and increased apoptosis. J Vasc Surg, 2002. 36(6): p. 1248-55. Binns, R.L., et al., Optimal graft diameter: effect of wall shear stress on vascular healing. J Vasc Surg, 1989. 10(3): p. 326-37. Mattsson, E.J., et al., Increased blood flow induces regression of intimal hyperplasia. Arterioscler Thromb Vasc Biol, 1997. 17(10): p. 2245-9.
14.
15. 16.
Baek, I., et al., Suppression of neointimal hyperplasia by sirolimus-eluting expanded polytetrafluoroethylene (ePTFE) haemodialysis grafts in comparison with paclitaxel-coated grafts. Nephrol Dial Transplant, 2012. 27(5): p. 1997-2004. Glickman, M., Drug eluting grafts for hemodialysis access. J Vasc Access, 2017. 18(Suppl. 1): p. 53-55. Khawaja, A.Z., et al., Systematic review of drug eluting balloon angioplasty for arteriovenous haemodialysis access stenosis. J Vasc Access, 2016. 17(2): p. 103-10.
Table 1: Baseline characteristics Baseline Characteristics (n=25) Age % male Median INTERMACS class Etiology (%) Ischemic Non-ischemic Comorbidities (%) Type 2 Diabetes Mellitus Chronic renal impairment Stage 3-4 Stage 5 Previous stroke Configuration (%) LVAD BiVAD % bridge to transplant Implant technique (%) Minimally invasive Median sternotomy Medications (%) Beta blocker RAAS blockade Reason for explant (%) Transplantation Autopsy Myocardial recovery Days on-pump (median, IQR) Clinical outcome (%) Alive Died on-pump Died following transplantation LVAD-related adverse events (%) Ischemic stroke or TIA Hemorrhagic stroke Pump thrombus Gastrointestinal bleeding Right heart failure Driveline infection
55.7 (9.4) 80 2 44 56 24 32 0 12 96 4 100 20 80 68 48 96 4 0 298 (223-352) 76 4 20 28 12 12 8 16 36
Figure 1: Macroscopic specimen demonstrating the anastomosis between HVAD outflow conduit and aorta following LVAD explantation at the time of cardiac transplant. 1A: Anastomosis and
outflow conduit intact. 1B: The outflow conduit has been transected to expose a layer of new tissue formation within the conduit at the anastomosis.
Figure 2: Typical histopathological findings (hematoxylin and eosin-stained sections). 2A: Focally prominent inflammation, including moderate numbers of plasma cells and neutrophils (inset: highpower field). 2B: Prominent chronic inflammation including large numbers of plasma cells and some lymphocytes. 2C: Fibrotic tissue with focal mild atheromatous change in addition to a multinucleated giant cell reaction. 2D: Fibrotic tissue consistent with large vessel intima and occasional chronic inflammatory cells.
Figure 3: Contrast-enhanced computerized tomography demonstrating subtle stenosis at the outflow graft-aorta anastomosis. 2A-B: axial plane. 2C-D: sagittal plane.
Supplementary appendix: Summary table outlining patient and tissue characteristics at HVAD explant. Patient
Age at explant
Days on pump
Pump location
Tissue ingrowth?
Location
Depth
Microscopic Description
1
52
92
LVAD
Yes
Aortic anastomosis
1mm
Foreign body giant cell reaction
2
68
93
LVAD
No Aortic anastomosis and body of graft
2mm
Plasma cells, giant cells and neutrophils
Yes
Aortic anastomosis
2mm
Fibromyxoid stromal tissue, foreign body giant cell reaction with histocytes, plasma cells
2mm
Fibromyxoid stroma with fibroblasts, histiocytes, occasional giant cells
3
61
119
LVAD
Yes
4
69
188
LVAD
No
5
37
198
LVAD
5
37
198
RVAD
Yes
Pulmonary anastomosis and body of graft
6
42
220
LVAD
Yes
Aortic anastomosis
11.5mm
7
54
233
LVAD
Yes
Aortic anastomosis
NS
8
70
238
LVAD
Yes
Body of graft
<1mm
9
53
259
LVAD
Yes
Body of graft
2mm
10
66
265
LVAD
Yes
Aortic anastomosis
<1mm
Aortic anastomosis and body of graft
1mm
11
40
284
LVAD
Yes
12
62
290
LVAD
Yes
13
54
305
LVAD
Yes
14
41
306
LVAD
Yes
15
60
308
LVAD
16
63
312
LVAD
Aortic anastomosis Aortic anastomosis
1mm 1mm
Body of graft
<1mm
Yes
Body of graft
<1mm
Yes
Aortic anastomosis
<1mm
Myxoid fibrosis, giant cells, plasma cells, epithelioid cells Intima-like myxoid fibrotic tissue with scattered plasma cells and lymphocytes Giant cell reaction, myxoid fibrosis, neointima Fibrotic tissue, adipocytes, cardiomyocytes, peripheral nerve bundles Foreign body, giant cells, plasma cells, foamy macrophages Prominent chronic inflammation - plasma cells, lymphoid aggregates, eosinophils, neovascularisation Reactive changes Myxoid stromal tissue. No inflammation Neointima, focal atheromatous change, foamy macrophages, giant cell foreign body reaction Myxoid fibrotic tissue Foreign body giant cell reaction
17
66
327
LVAD
Yes
Aortic anastomosis
<1mm
18
56
350
LVAD
Yes
Body of graft
<1mm
Yes
Aortic anastomosis and body of graft
1mm
Yes
Aortic anastomosis and body of graft
<1mm
19
20
59
57
353
416
LVAD
LVAD
21
51
435
LVAD
Yes
Body of graft
1mm
22
59
449
LVAD
Yes
Aortic anastomosis
1mm
23
60
463
LVAD
Yes
Body of graft
<1mm
Yes
Aortic anastomosis and body of graft
1mm
Yes
Aortic anastomosis and Body of graft
1mm
24
25
55
52
482
1032
LVAD
LVAD
Scattered neutrophils Neointima with minimal revascularisation Fibrotic tissue consistent with large vessel intima, scattered chronic inflammatory cells Chronic inflammatory changes with plasma cells Mild myxoid changes, mild atheromatous change, chronic inflammatory changes (lymphocytes) Intima-like myxoid fibrotic tissue with scattered plasma cells, lymphocytes and MNG cells Neo-intima. No significant inflammation Foreign body giant cell reaction, mixed inflammation and granulomata Fibroblast proliferation, endothelial lining
Highlights • • •
Tissue ingrowth into the HeartWare HVAD outflow graft is a near-universal finding It may comprise an acute or chronic inflammatory, fibrotic or neointimal reaction This is associated with a decrease in pump flow and flow pulsatility over time
Conflicts of Interest There are no relevant conflicts of interest.