Anticoagulation management following left ventricular assist device implantation is similar across all provider strategies

Abstract

OBJECTIVES

Thromboembolic and bleeding events are potential complications following left ventricular assist device implantation. A tight control of the international normalized ratio (INR) is believed to be crucial in the reduction of postimplant complications. There is significant variability among institutions as to whether a device implanting centre should be managing the INR. In this study, we evaluated the effect of INR management strategies in maintaining a therapeutic INR.

METHODS

A retrospective review was utilized to identify patients implanted with either the HeartMate II or the HeartWare HVAD between January 2011 and February 2016. Patients were stratified into 4 groups based on the post-discharge INR management strategy: outside hospital system anticoagulation clinic, outside hospital primary care provider, implanting centre anticoagulation clinic or implanting centre ventricular assist device office. The INR data were collected and analysed for both the early (discharge, 7, 14, 21 and 30 days) and late (3, 6, 9 and 12 months) postoperative periods.

RESULTS

There were 163 patients identified during the study period who met the study inclusion criteria: 49 (30%) patients were managed by an outside hospital system anticoagulation clinic, 59 (36.2%) patients by an outside hospital physician/primary care provider, 22 (13.5%) patients by the implanting centre anticoagulation clinic and 33 (20.2%) patients by the implanting centre ventricular assist device office. There were no statistically significant differences found between management strategies across all time points.

CONCLUSIONS

There was no statistically significant difference found between the management strategies examined. Regardless of the chosen INR management strategy, patients have similar INR values and postoperative outcomes.

INTRODUCTION

There are nearly 6 million people living with heart failure in the USA [1]. Of this population, an estimated 150 000–250 000 patients have advanced heart failure where medical management therapies are no longer sufficient [2]. Although heart transplantation remains the gold standard for management of end-stage heart failure, the relative flat line of donor availability does not allow for transplantation in the era of a growing heart failure population. Since their inception, left ventricular assist devices (LVADs) have emerged as viable alternatives for patients with advanced heart failure both as a bridge to transplantation and as destination therapy [3]. Although LVADs improve survival and quality of life for patients with advanced heart failure, they have limitations with a defined risk of certain adverse events (AEs). Most notably, there is a constant balance between prevention of pump thrombosis (PT) and thromboembolic complications versus maintenance of appropriate anticoagulation and antiplatelet therapy with resultant bleeding risks [4–6]. The current mainstay of therapy for anticoagulation with LVADs is the use of warfarin with aspirin serving as the antiplatelet therapy of choice. The dilemma of under- or overanticoagulation is a significant burden on the medical system often with a need for readmission to the hospital for proper management and intravenous heparin drip bridging for control of the international normalized ratio (INR) [7]. With focus on avoidance of hospital readmissions, particularly for heart failure, and governance over LVAD therapy, there is a heightened awareness of the INR management in these patients. Consequently, some centres advocate management of INR and anticoagulation therapy by only the LVAD implanting centre. This is a significant burden for patients and their families who are often several hundred miles away from the implanting institution. As such, we seek to assess the effectiveness of provider management strategies at maintaining INR within goal range. Additionally, we seek to evaluate the distribution of AEs by management strategy postoperatively.

METHODS

An institutional review board-approved retrospective review was utilized to identify patients implanted with either the HeartMate II (HM II; Thoratec Corporation, Pleasanton, CA, USA) or the HVAD (HeartWare International, Framingham, MA, USA) ventricular assist devices between January 2011 and February 2016. Inclusion criteria were age greater than 18 years, survival to discharge from index hospitalization, INR record availability during specified times, INR management strategy on file and primary LVAD implant. Patients were excluded if they did not survive to discharge, were implanted at an outside hospital and later transferred care to our site, received a total artificial heart or had no documented management strategy on file. Censoring occurred at the time of heart transplantation or death.

Patients were stratified into 4 groups based on their INR management strategy at the time of hospital discharge from LVAD implantation: (i) outside hospital system anticoagulation clinic (OSH-CC), (ii) outside hospital system primary care provider (OSH-PCP), (iii) implanting centre anticoagulation clinic (OSU-CC) or (iv) implanting centre ventricular assist device office (OSU-VAD). Baseline demographics were collected including age, gender, race, heart failure aetiology, body mass index, smoking status, LVAD type and device strategy at the time of implantation. Additional collected comorbidities included the presence of diabetes mellitus, peripheral arterial disease, renal insufficiency, transient ischaemic attacks or cerebrovascular accidents, pulmonary hypertension, New York Heart Association classification of heart failure and Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) profile at the time of implant.

All patients were discharged with an INR protocol goal of 2.0–3.0 with instructions to have their INR checked weekly through the first month. Subsequently, the INR data collected were in line with the INTERMACS follow-up guidelines: 1 week ± 3 days, 1 month ± 7 days, 3 months ± 1 month, 6 months ± 2 months and 12 months ± 2 months. The INR values were chosen closest to the actual date for each follow-up visit for comparison. We defined the early INR management period as being less than 1 month post-discharge and late as INR management between 1 month and 1 year post-discharge. Postoperative outcomes, including incidence of stroke, haemorrhagic event (gastrointestinal bleed, epistaxis or other bleeding event) or PT, were categorized using INTERMACS AE definitions. PT was identified if there was either suspected pump thrombus or confirmed pump thrombus. Incidence of sub- and supratherapeutic INR were also recorded per management strategy.

Statistical analysis was carried out using descriptive statistics for patient baseline demographics and clinical outcomes. Each was stratified by the INR management strategy with analysis carried out utilizing either a t-test or the Fisher’s exact test. Analysis of variance methods were used to test for differences in INR values in both the early and the late outcomes with respect to management strategies. The Fisher’s exact test was used for testing categorical outcomes. The Bonferroni correction was used for setting P-value cut-offs for statistical significance. Linear mixed modelling with compound symmetry covariance structure was used for comparing across time points. All statistical methods were performed in accordance to guidelines by a co-authoring statistician.

RESULTS

A total of 197 patients were identified who received either an HM II or an HVAD implant during the specified time frame. Of these patients, 163 were identified to fulfil the inclusion criteria for the study. Of these patients, 49 (30.0%) patients had their INR managed by an OSH-CC, 59 (36.2%) patients were managed by an OSH-PCP, 22 (13.5%) patients were managed by the OSU-CC and 33 (20.2%) patients were managed by the OSU-VAD.

Baseline patient characteristics for each group are summarized in Table 1. The OSU-CC group had the youngest patient population of 48.5 ± 13.1 years of age when compared with the other groups (P = 0.01). More Caucasian patients were followed at an outside hospital facility: 85.7% at OSH-CC and 83.1% at OSH-PCP when compared with the implanting enter of 50.0% at OSU-CC and 75.8% at OSU-VAD (P = 0.01). All other variables, aside from unspecified heart failure aetiology, were similar across all groups of INR management strategies.

Table 1.

Baseline demographics and comorbidities by the INR management strategy

	OSH-CC (n = 49)	OSH-PCP (n = 59)	OSU-CC (n = 22)	OSU-VAD (n = 33)	P-value
Age at implant (years)	54.2 ± 10.8	58.6 ± 12.1	48.5 ± 13.1	54.5 ± 10.4	0.01
Male (%)	83.7	69.5	81.8	63.6	0.13
Race (%)
African American	14.3	15.3	45.5	24.2	0.02
Caucasian	85.7	83.1	50.0	75.8	0.01
Other	0	1.7	4.5	0	0.38
Heart failure aetiology (%)
Ischaemic	30.6	23.7	13.6	12.1	0.21
Non-ischaemic	55.1	74.6	72.7	78.8	0.09
Unspecified/other	14.3	1.7	13.6	9.1	0.05
Body mass index, kg/m2	29.4 ± 6.8	28.3 ± 6.7	28.8 ± 7.9	27.9 ± 5.5	0.76
Smoker (%)	42.5	36.4	36.8	41.4	0.94
Diabetes mellitus (%)	32.5	38.6	31.6	34.5	0.92
Peripheral vascular disease (%)	2.5	2.3	0	3.5	1.00
Renal disease (%)	35.0	27.3	47.4	24.1	0.33
Stroke/TIA (%)	2.5	9.1	0	3.5	0.47
Pulmonary hypertension (%)	50.0	54.6	52.6	44.8	0.87
Device implanted (%)
HeartMate II	89.8	86.4	86.4	91.0	0.87
HeartWare	10.2	13.6	13.6	9.0	0.87
Preimplant device strategy (%)
Bridge to transplant	30.6	31.0	40.9	42.4	0.59
Destination therapy	67.3	69.0	59.1	57.6	0.64
Unspecified/other	2.0	0	0	0	0.64
NYHA Class (%)
III	4.1	8.5	4.5	12.1	0.58
IV	93.9	84.7	90.9	87.9	0.49
Other	2.0	6.8	4.6	0	0.39
INTERMACS profile (%)
1	18.4	15.3	22.7	18.2	0.85
2	40.8	39.0	27.3	39.4	0.75
3	34.7	37.3	40.9	36.4	0.96
4–7	6.1	8.5	9.1	6.1	0.93

A P-value less than or equal to 0.05 was considered significant.

Age and body mass index are represented as mean ± standard deviation. All other values are expressed as a percentage.

INR: international normalized ratio; INTERMACS: Interagency Registry for Mechanically Assisted Circulatory Support; NYHA: New York Heart Association; OSH-CC: outside hospital system anticoagulation clinic; OSH-PCP: outside hospital system primary care provider; OSU-CC: implanting centre anticoagulation clinic; OSU-VAD: implanting centre ventricular assist device office; TIA: transient ischaemic attack.

The INR values across all management strategies during the early post-discharge interval, the first month after discharge from the hospital, are summarized in Table 2. The time interval of INR check at discharge, 7 days, 14 days, 21 days and 30 days post-discharge were similar throughout all management strategies. The INR values across all management strategies during the late post-discharge interval, between 1 month and 1 year after discharge, are summarized in Table 3. The time interval of INR check at 3 months, 6 months, 9 months and 12 months post-discharge were not statistically different across the various INR management strategies. The percentage of patients outside of the target INR range, across all time points, are shown in Fig. 1; at any given time period, 16.9–60.6% of patients were not within their target range across management strategies.

Table 2.

One-month (early) post-discharge INR value by different management strategies

	OSH-CC	OSH-PCP	OSU-CC	OSU-VAD	P-value
Discharge	2.30 ± 0.45 (n = 49)	2.48 ± 0.69 (n = 59)	2.56 ± 0.40 (n = 22)	2.55 ± 0.70 (n = 33)	0.19
7 days	2.11 ± 0.56 (n = 49)	2.33 ± 0.63 (n = 58)	2.48 ± 0.70 (n = 22)	2.30 ± 0.74 (n = 33)	0.28
14 days	2.38 ± 0.69 (n = 49)	2.34 ± 0.80 (n = 57)	2.28 ± 0.58 (n = 22)	2.03 ± 0.70 (n = 33)	0.24
21 days	2.14 ± 0.60 (n = 49)	2.35 ± 0.79 (n = 57)	2.36 ± 0.84 (n = 22)	2.05 ± 0.53 (n = 33)	0.33
30 days	1.98 ± 0.52 (n = 49)	2.07 ± 0.47 (n = 56)	2.02 ± 0.51 (n = 22)	2.29 ± 0.49 (n = 33)	0.11

All values are represented as mean ± standard deviation.

INR International Normalized Ratio; OSH-CC: outside hospital anticoagulation clinic; OSH-PCP: outside hospital system primary care provider; OSU-CC: implanting centre anticoagulation clinic; OSU-VAD: implanting centre ventricular assist device office.

Table 3.

One-year (late) post-discharge INR value by different management strategies

	OSH-CC	OSH-PCP	OSU-CC	OSU-VAD	P-value
3 months	2.07 ± 0.54 (n = 49)	2.04 ± 0.52 (n = 53)	2.32 ± 0.48 (n = 22)	2.28 ± 0.54 (n = 32)	0.09
6 months	2.11 ± 0.58 (n = 46)	2.06 ± 0.53 (n = 50)	2.35 ± 0.50 (n = 22)	1.98 ± 0.57 (n = 30)	0.15
9 months	2.18 ± 0.66 (n = 46)	2.16 ± 0.68 (n = 47)	2.31 ± 0.72 (n = 21)	2.29 ± 0.66 (n = 28)	0.84
12 months	2.01 ± 0.48 (n = 46)	2.25 ± 0.64 (n = 46)	2.49 ± 0.81 (n = 19)	2.20 ± 0.62 (n = 28)	0.19

All values are represented as mean ± standard deviation.

INR: international normalized ratio; OSH-CC: outside hospital anticoagulation clinic; OSH-PCP: outside hospital system primary care provider; OSU-CC: implanting centre anticoagulation clinic; OSU-VAD: implanting centre ventricular assist device office.

Figure 1.

Patients outside of target international normalized ratio range by time point and management strategy. All values are expressed as percentages. OSH-CC: outside hospital anticoagulation clinic; OSH-PCP: outside hospital system primary care provider; OSU-CC: implanting centre anticoagulation clinic; OSU-VAD: implanting centre ventricular assist device office.

There were no statistically significant differences in AEs collected following LVAD implantation as summarized in Table 4. The follow-up for OSH-CC was 23.6 ± 3.4 months, OSH-PCP was 17.2 ± 2.8 months, OSU-CC was 17.6 ± 2.8 months and OSU-VAD was 14.6 ± 2.5 months. There were no differences among the complication rates of stroke, haemorrhagic events (gastrointestinal, epistaxis or other), PT and sub- or supratherapeutic INR.

Table 4.

Postoperative outcomes stratified by INR management strategy as events per patient year

	OSH-CC (n = 49)	OSH-PCP (n = 59)	OSU-CC (n = 22)	OSU-VAD (n = 33)	P-value
Stroke	0.052	0.047	0.062	0.075	0.79
Haemorrhagic event
Gastrointestinal	0.062	0.107	0.062	0.199	0.25
Epistaxis	0.021	0.024	0.031	0.050	0.93
Other	0.042	0.071	0.000	0.075	0.49
Pump thrombosis	0.010	0.012	0.000	0.025	0.89
Sub-therapeutic INR	0.073	0.130	0.186	0.224	0.41
Supratherapeutic INR	0.010	0.012	0.000	0.050	0.73

All values are expressed as events per patient year.

INR: international normalized ratio; OSH-CC: outside hospital anticoagulation clinic; OSH-PCP: outside hospital system primary care provider; OSU-CC: implanting centre anticoagulation clinic; OSU-VAD: implanting centre ventricular assist device office.

DISCUSSION

Despite the improvements in survival and device technology, there remain certain areas in need of improvement with LVAD therapy. Many efforts are centred on improving the quality of life, reducing AEs and defining best care principles. The need for anticoagulation with this therapy touches on all of these themes. With regard to improving the quality of life, it is important to keep patients who are on LVAD support out of the hospital and enjoying a quality of life consistent with the goals of their care. Indeed, many patients live hundreds of miles away from their implanting centre and a significant amount of time can be taken out of their lives for routine laboratory draws, weekly clinic visits and routine follow-up that can be done at outside facilities where they had previously received their care. There has been a call for this shared care model for taking care of these patients, and it is important to understand what effect, if any, on postimplantation outcomes will be there with allowing nonimplanting centres to manage the INR [8]. Our study demonstrated no significant difference between implanting centre (OSU-CC and OSU-VAD) INR management and nonimplanting centre (OSH-CC, OSH-PCP) management. If the trends in increasing LVAD implantation continue, it will become even more imperative to educate primary care physicians and community care providers in the care of these patients supported with LVADs.

Another central concern for anticoagulation has been the effects of anticoagulation differences in the prevalence of PT. This has been called into question with the unexpected rise reported from multiple centres in the rate of PT in recent years [9]. This delicate balance between management of anticoagulation with bleeding risk versus PT has led some centres to abandon the practice of intravenous heparin bridging immediately after implantation; however, a recent multicentre study shows the importance of resuming this practice in the prevention of PT [10].

In addition to whether heparin bridging is necessary, another area of debate is whether antiplatelet therapy is needed in conjunction with warfarin therapy. The influence that aspirin therapy has on haemorrhage versus thrombosis in LVAD patients remains unknown [11]. One prospective controlled study, the US TRACE study, showed that holding aspirin in patients with a bleeding history on an HM II support was feasible; however, a conferred increased risk of PT existed [4]. A new, randomized controlled study is the St. Jude Medical-sponsored PREVENTion of Non-Surgical Bleeding by Management of HeartMate II Patients without Antiplatelet Therapy (PREVENT II) trial designed to test patients on warfarin maintenance with and without aspirin therapy. At our institution, it is a standard practice to bridge patients with heparin postoperatively and discharge on antiplatelet therapy. In this study, we found no significant difference in haemorrhagic events, stroke or PT either at the implanting centre or at outside hospital facilities.

In addition to whether aspirin is needed, there are perceived, or assumed, distinct periods where large variations in INR management may exist. As such, we looked at 2 distinct periods—early (<30 days post-discharge) and late (1 month to 1 year) post-discharge. Theoretically, the importance of early post-discharge INR management is the fact that patients have numerous medications that can interact with their warfarin dose. They also start enjoying an increase in activity level and appetite, making it imperative to keep a close watch on the INR values. In this study, there was no difference between the outside facilities and implanting centre INR management in this early period. It was also encouraging to note that no changes occurred in the late period as patients became increasingly more independent and less tethered to their implanting institution. Regardless of the chosen INR management strategy, patients have similar INR values that are within the therapeutic range. Perhaps, the most important issue then is the INR management strategy utilizing an anticoagulation protocol. At our institution, the INR goal is communicated with outside facilities and our VAD coordinators have continued communication with these providers. Communication of INR value is charted in the patients’ medical record. If communication of the INR value has not been received, our VAD coordinators initiate a phone call to the outside provider for all patients.

Another measure of success that has emerged has been the concept of the amount of time in therapeutic range (TTR) for each patient. Protocols that are pharmacy driven have been shown in other disease processes to improve the TTR [12], but the clinical advantage remains unclear [13]. An interesting revelation was that in a single institutional study of 51 patients, only half the time was spent in TTR, with 48% of values being outside of the goal INR range [14]. This study also showed the correlation with the time spent in supratherapeutic range 30 days prior to a bleeding event correlating with such events. In the INR management of LVADs, there has only been 1 study looking at a pharmacist-driven protocol with resulting increase in TTR [15]. The purpose of this study was to compare different INR management strategies and, as such, we did not utilize TTR—the 4 different strategies had variations in the long-term frequency of INR checks, leading to an inability to compare strategies utilizing the TTR.

Another important topic raised by our findings was the suggestion of differences in race with regard to follow-up care. Certainly, there exist not only racial disparities in LVADs and heart transplantation but also access to shared care sites and socioeconomic barriers for uniform management [8, 16, 17]. This will be important as the technology becomes more mainstream to ensure all patients who can benefit from this therapy are afforded the same therapeutic options.

Limitations

Akin to most retrospective studies, there were some limitations with our current study. As a single-centre study, the results may not be applicable to the general population. Given the 4 different management strategies, there was a relatively small number (n) in each group. Outside providers were given instructions on therapeutic range and communicated with our LVAD coordinators, but there was no mandated protocol to follow and variations in management style were unable to be captured.

CONCLUSION

Indeed, this is an exciting time of improvements in LVAD technology and INR monitoring. The future looks bright, but we must not take the focus off of patients. Complications related to anticoagulation management, such as hospital readmissions and haemorrhagic and thrombotic events, create a burden for LVAD patients and caregivers. It is encouraging to know that there was no increase in these events among the various groups in our study. Although protocols are necessary for standardization and improvements in care, there will always be a need for clinical judgement. There is a growing body of literature describing either reduced INR goals with aspirin or reduction of dose of aspirin alone in the reduction of bleeding risks with the HM II [18, 19]. However, the INR is inversely related to thrombotic events outside of hospital care for patients with LVADs with an INR <2.0 [20]. In addition, the effects of comorbidities, such as immobility, deconditioning, atrial fibrillation and valvular replacements, on the need for optimal anticoagulation is unknown. Identifying the most convenient method of INR management is perhaps just as important as finding the best anticoagulation strategy. There are many benefits to home INR testing, including convenience for patients and a higher percentage of TTR [15]; however, correlation with laboratory testing and the effect on outcomes still needs to be established [22]. Similarly, the possibility of either dual antiplatelet therapy or newer oral anticoagulation agents replacing warfarin altogether exists.

There will always be the next set of anticoagulation agents, safer adverse bleeding profiles and better LVAD technology. It is imperative to continue to question best care practices, guiding principles and continuously evolve and investigate what the safest, most efficient and most convenient way of providing care for these complex patients will prove to be.

Funding

The project described was supported by Award Number Grant UL1TR001070 from the National Center For Advancing Translational Sciences. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center For Advancing Translational Sciences or the National Institutes of Health.

Conflict of interest: Sitaramesh Emani: Abbott, consulting, grant support, advisory board; Medtronic, Inc., consulting. Brent C. Lampert: Abbott, consulting. Ahmet Kilic: Abbott, travel, consulting; Medtronic, Inc., travel. The remaining authors have no conflicts to declare.