Pain and Functional Status in Patients with Ventricular Assist Devices

Abstract

Context

Ventricular assist devices (VADs) have been shown to improve survival and overall quality of life, but there are limited data on pain control and functional status in this patient population.

Objectives

This study examined changes in pain, functional status and quality of life over time in VAD patients.

Methods

Patients were enrolled in this prospective cohort study before or as early after VAD implant as possible and then followed for up to forty-eight weeks. The Brief Pain Inventory (BPI) was used to assess pain. The Katz Independent Activities of Daily Living (IADL) questionnaire was used to assess functional status. The Kansas City Cardiomyopathy Questionnaire (KCCQ), a 23-item questionnaire covering five domains (physical function, symptoms, social function, self-efficacy, and quality of life) was used to assess quality of life and health status.

Results

Eighty-seven patients were enrolled at four medical centers. The median BPI severity score was 2.8 (interquartile range [IQR] 0.5, 5.0) before implantation, and 0.0 (IQR 0.0, 5.3) 48 weeks after implantation (P=0.0009). Katz IADL summary scores also demonstrated significant improvement over time (p<0.0001). KCCQ summary scales demonstrated significant improvement with time (P<0.0016).

Conclusion

This study demonstrated that patients with VADs experienced improved pain, functional status, and quality of life, over time. These data may be useful to help patients make decisions when they are considering undergoing VAD implantation.

Introduction

Ventricular assist devices (VADs) are increasingly used to treat heart failure (HF) patients with advanced disease.¹ These mechanical devices are surgically implanted to augment the pumping function of the heart’s ventricles. Although originally Food and Drug Administration-approved as a temporary therapy to “bridge” patients as they waited for a donor organ for cardiac transplantation, VADs are now also implanted in advanced HF patients who are ineligible for transplantation (destination therapy). Data from 2014 demonstrated that nearly 12,500 patients have received a device since 2006² and the number of patients with VADs is expected to continue to grow.

Given the rapidly increasing prevalence of patients with VADs,^{3, 4} there is a corresponding increased interest in integrating palliative care into the treatment offered to patients with advanced HF. This growing interest, combined with Medicare requirements for palliative care involvement in the care of patients who are undergoing VAD implantation as destination therapy,⁵ together highlight the importance of better understanding the post-implant time points when the expertise of palliative care may most benefit patients and the primary team caring for them. By understanding the natural history of symptoms in these patients, it would be easier to target interventions to coincide with periods when patients are in most need.

While VADs have been shown to improve survival in patients with advanced HF, and there are some data exploring quality of life in these patients,^6–8 there have been no studies examining pain, activities of daily living (ADLs), and the natural history of these symptoms over time in this patient population. These outcomes are of interest because although dyspnea and fatigue are the classic symptoms of HF, many of these patients also suffer from pain.^9–12 Studies have demonstrated that as many as 70% of patients with advanced heart failure experience pain.¹³ In addition, patients with advanced HF often have difficulty completing their ADLs because of shortness of breath, fatigue, or generalized debility from the underlying disease process. This constellation of symptoms can negatively impact quality of life, which has been tied to increased risk of hospital admissions and increased mortality.^{14, 15}

Furthermore, although VADs have decreased in size as technology has improved, they are still partially external devices that require a power source connected to a controller unit that must be attached to the patient at all times. Although the survival benefit is clear, dependency on such a device has the potential to adversely affect patients’ quality of life, particularly in terms of social functioning. Additionally, VADs are associated with their own set of adverse events such as gastrointestinal bleeding and an increased risk of stroke that in itself may cause patients to worry, also adversely affecting quality of life. A more granular analysis is necessary to assess the impact of a VAD on patients’ quality of life in order to better inform tailored interventions that address specific aspects of quality of life. We enrolled a prospective cohort of patients with advanced HF at the time of VAD implant to examine changes in pain, functional status, and quality of life over time.

Methods

Design Overview

This prospective cohort study was designed to determine baseline symptoms of patients peri-VAD implantation, and to evaluate how these symptoms and characteristics changed over time following implantation. Patients were enrolled during the hospitalization when the device was implanted at four high-volume VAD implanting medical centers across the United States immediately before or as early after VAD implant as possible, and were then followed for up to 48 weeks. Patients were eligible for this study if they were age 21 years and older, fluent in English, had a caregiver or family member who was willing to be enrolled in a related cohort study, and had consistent and reliable access to a telephone. We limited eligibility to English speakers because most of the study instruments have only been validated in English. Institutional review boards of participating centers and the coordinating center approved the protocol, and all patients provided written informed consent.

Instruments

Outcomes were obtained through structured patient questionnaires. These questionnaires were administered at the bedside peri-implant. Post-implant, surveys were conducted via phone calls by a research assistant. Patients were called during the day; however, if they did not answer, there was at least one attempt at night and one attempt during the weekend to reach them.

Pain was assessed using the Brief Pain Inventory (BPI) at the time periods noted in Appendix A (available at jpsmjournal.com). The BPI is a survey used in clinical practice and research.^{16, 17} While this tool was originally developed to assess pain in cancer patients, it has been demonstrated to be valid in non-cancer patients,¹⁷ and has been used in other studies examining pain in heart failure.⁹ Pain severity is evaluated by asking patients four questions to rate their current, worst, least and average pain in the prior week. Seven questions assess pain interference by asking patients to rate how much their pain interfered with various activities such as mood and walking ability in the prior week. The scale ranges from 0 to 10 and higher numbers indicate a higher level of severity. The average score across domains provides overall severity and interference scores.

Functional status was evaluated with the Katz Independent Activities of Daily Living (IADL) questionnaire¹⁸ at the time periods noted in Appendix A. This scale assesses functional independence and has been shown to be valid in studies on several diseases including HF.^19–21 It explores independence in dressing, bathing, walking across a room, eating, transferring from bed and toileting. Patients rate their ability to perform these tasks on a scale of 1–5 with 1 meaning they can perform the task with no difficulty at all, and 5 meaning they cannot perform the task. Thus, a lower overall score is indicative of higher functioning.

The Kansas City Cardiomyopathy Questionnaire (KCCQ) was used to assess quality of life and health status covering five domains: physical function, symptoms, social function, self-efficacy and quality of life; it was administered at the time periods illustrated in Appendix A. The KCCQ is a valid and reliable measure developed for patients with HF^{22, 23} and it is highly sensitive to clinical change in HF patients over a 6–12 week period.²⁴ The overall score ranges from 0–100. Higher scores indicate better HF-related health status and can predict future cardiovascular mortality and re-hospitalization.²⁵ Poor health status is defined as a KCCQ score below 50 points.

Statistical Analysis

Data were analyzed using SAS v. 9.3 (SAS Institute, Inc., Cary, NC). We used descriptive analyses to summarize the characteristics of our study population. Dichotomous variables were summarized as proportions and continuous variables with a mean and standard deviation if they were normally distributed; otherwise, they were summarized using a median and interquartile range.

We then determined the severity (as applicable) of the physical symptoms over time. The linear mixed model regressions (PROC MIXED for normally distributed data and PROC GLIMMIX for non-normally distributed data) accounted for the structure of the data where assessments were nested within participants. These procedures implement two likelihood-based methods: maximum likelihood (ML) and restricted/residual maximum likelihood (REML).²⁶ In our analysis of the data, we found that missing data were missing at random, meaning there was no specific mechanism to or underlying systematic reason for the missing data.²⁷ A favorable theoretical property of ML and REML is that they accommodate data that are missing at random.²⁸ In addition, linear mixed model regressions are able to model repeated measures data. They can utilize repeated subjects with missing data points as the time points with data still contribute to the maximum likelihood.

Results

Patient Characteristics

One hundred and forty-two patients were screened and 111 (78%) were deemed to be eligible for this study (Fig. 1 Of this group, 95 patients (86% of eligible group) were enrolled, and ultimately 87 patients were analyzed. Forty patients were enrolled before implant. Thirty-four were enrolled immediately after implant. Five patients were enrolled two weeks after implant. Eight patients were enrolled four weeks or later after implant. The mean±SD age of the cohort was 58.0± 11.7 years, 23% were female, 29% identified as black, and 59% as white. A majority of the patients (57%) reported a history of myocardial infarction, 42% reported a history of diabetes mellitus, and few had cancer or other serious medical problems besides their heart disease (Table 1). The indication for VAD placement was bridge to transplant for 39% of this group, and 52% received the VAD as destination therapy. Indication was unknown for 9% of patients. Over the course of the study period, 8% of the patients died and 16% received a heart transplant.

Fig. 1.

CONSORT diagram of participant flow.

Table 1.

Baseline and Implant Characteristics^* Among a Cohort of Patients Undergoing VAD Implant at Four Academic Medical Centers. The number of responses varies due to the timing of when patients completed the questionnaire; in many cases patients were enrolled either immediately before the VAD was implanted (and thus there was not time to complete the demographic form) or immediately after and they may not have been well enough to complete the questionnaires.

Characteristic
Indication for VAD placement (n=87)
Destination Therapy		51.7 % (45)
Bridge to Transplant		39.1% (34)
Unknown Indication		9.2% (8)
Device Type (n=87)
BiVAD		1.1% (1)
LVAD		95.4% (83)
Unknown		3.4% (3)
Manufacturer (n=87)
HeartMate II		82.8%% (71)
Heartware		13.8% (12)
Unknown		4 (4.6%)
Age (n=87)		58.0 +/− 11.7
Gender (n=87)
Women		23% (20)
Men		77% (67)
Marital Status (n=45)
Married		75.6% (34)
Separated		11.1% (5)
Widowed		2.2% (1)
Single/Never Married		11.1% (5)
Divorced		0
Partnered/Living with Significant Other		0
Medical Center (n=87)
Mount Sinai		34.5% (30)
Columbia-Presbyterian		19.5% (17)
University of Pennsylvania		11.5% (10)
Jewish Hospital		34.5% (30)
Number of years of school (n=41)
<12 years		9.8% (4)
12–16 years		75.6% (31)
>16 years		14.6% (6)
Highest Degree Obtained (n=41)
None		4.9% (2)
Technical Degree/Diploma		2.4% (1)
High School Diploma/GED		48.8% (20)
Associates Degree		14.6% (6)
Bachelor's degree		12.2% (5)
Masters Degree		12.2% (5)
Doctorate		2.4% (1)
Other non-US Degree		2.4% (1)
Ethnicity (n=41)
Not of Hispanic/Latino/Spanish Origin		90.2% (37)
Puerto Rican		4.9% (2)
Other Hispanic/Latino/Spanish Origin		4.9% (2)
Race (n=41)
White		58.6% (24)
Black/African American		29.3% (12)
Asian		2.4% (1)
Other		7.3% (3)
White & Native American/Alaska Native		2.4% (1)
Income(n=41)
$10,000–$29,999		12.2% (5)
$30,000–$59,999		9.8% (4)
$60,000–$99,999		17.1% (7)
$100,000–$199,999		19.5% (8)
Don't Know		22.0% (9)
Refused		17.1% (7)
Not Applicable		2.4% (1)
How much money do you have left over at the end of the month? (n=41)
Some money left over		43.9% (18)
Just enough money to make ends meet		34.2% (14)
Not enough money to make ends meet		9.8% (3)
Don't Know		7.32% (3)
Refused		2.4% (1)
Not Applicable		2.4% (1)
Comorbidities (n=50)
History of Heart Attack		57.1%
History of Surgery for Peripheral Vascular Disease		19.5%
History of Stroke/TIA		12.2%
Asthma		7.3%
COPD		2.4%
Stomach problems/Peptic Ulcer Disease		17.1%
Alzheimer's disease/Other Dementia		0.0%
Cirrhosis/Liver damage		4.9%
Diabetes		41.5%
Kidney problems/Dialysis/Transplant		4.9%
RA/Lupus/PMR		0.0%
Lymphoma		0.0%
Leukemia/Polycythemia Vera		0.0%
Other Cancer		7.3%
HIV/AIDS		0.0%

^*Plus-minus values are means ± SD, categorical values are n (%)

Instrument Outcomes

For each instrument, there was a 26%–88% (mean 49%, median 46%) response rate at each included time period. The lowest response rates were immediately after implant. There were no differences in response rates between the instruments at any particular time period (Appendix A).

Pain Outcomes

Median pain severity score was 2.8 (interquartile range [IQR] 0.5, 5.0) of 10 (with 10 representing worst pain) pre-implant, 3.8 (IQR 0.0, 6.5) immediately after implant, 0.0 (IQR 0.0, 5.3) at forty-eight weeks and decreased over time. A generalized linear mixed model using time to predict pain score demonstrated that the score decreased over time (P<0.0001).

Median score representing how much pain interferes with activity was 7.2 (IQR 2.8, 7.5) of 10 (with 10 representing worst pain) pre-implant, 2.0 (IQR 0.0, 5.2) immediately after implant, 0.0 (IQR 0.0, 0.7) at forty-eight weeks and decreased over time. A generalized linear mixed model using time to predict score demonstrated decrease over time (P<0.0001).

Functional Status Outcomes

Katz IADL summary scores demonstrated significant correlation with time. Median summary score was 14 (IQR 6, 14), before implantation, 21 (IQR 11, 26) immediately after implantation, 10 (IQR 6, 10) at 48 weeks and decreased over time. A generalized linear mixed model using time to predict score demonstrated that the summary score decreased over time (P<0.0001).

Quality of Life Outcomes

KCCQ summary scales demonstrated significant correlation with time. Mean±SD overall summary score was 66.4±16.3 before implantation, 58.0±26.1 immediately after implantation, and 81.0± 14.2 at 48 weeks. A generalized linear mixed model using time to predict score demonstrated that the overall summary score increased over time (P=0.0001). Mean clinical summary score was 73.1±18.3 before implantation, 60.5±29.1 immediately after implantation, and 80.9±18.4 at 48 weeks. A generalized linear mixed model using time to predict score demonstrated that the clinical summary score increased over time (P=0.0015).

Discussion

We evaluated pain, functional status, and quality of life over time in a group of patients who underwent clinically indicated VAD implant. To our knowledge, this is the first study to fully explore pain and ADLs in patients who have been implanted with VADs. Our data demonstrate that there are discrete time periods during which patients with recently implanted VADs need extra attention to their physical symptoms and quality of life. Palliative care may play a key role in the weeks immediately following implant in managing pain and directing support for ADLs, while also providing reassurance to these patients that their symptoms will likely improve over time.

Pain Outcomes

We found that pain does not worsen in this group, and in fact may improve compared to baseline pain before implantation. Four weeks after implant, median scores dropped to zero and remained there until the end of the observation period at 48 weeks.

Although shortness of breath is the classic symptom of worsening HF, pain is common in patients with HF. Chronic pain can affect patients’ quality of life and prior studies have demonstrated that 38%–70% of HF patients experience moderate to severe pain.^{9, 29} Per these studies, most of this pain is located in the chest, low back, arms, legs and joints. In our study, patients with pain reported its location to be mostly in the chest, shoulders, stomach, and both upper and lower back. However, to our knowledge, no prior studies have fully explored pain in patients with VADs. Assessing the burden of pain in this group is important as this may be a potential target for palliation in this patient population.

Improvement in pain may be the result of a number of factors directly related to the VAD. Heart failure has many downstream consequences, and affects different organ systems including the renal and neuroendocrine systems. Cytokines implicated in the progression of HF include tumor necrosis factor α (TNFα), interleukin (IL) 1, and IL-6, which are all pro-inflammatory³⁰ and have been implicated in other pain syndromes.^31–33 Elevated levels of these cytokines may be the result of under-perfusion of peripheral tissues. The VAD improves forward flow in patients, which results in improved perfusion of peripheral tissues and may result in decrease in these inflammatory cytokines,^{34, 35} potentially decreasing pain in these patients.

Functional Outcomes

We found that VADs do not worsen ability to perform ADLs over time. To our knowledge, no prior studies have specifically examined the ability to perform ADLs in VAD patients over time.

As HF progresses, shortness of breath, fatigue, and generalized debility increase, which limits patients’ ability to complete ADLs. Indeed, patients with stage D symptoms (American College of Cardiology [ACC] staging system) are defined by their marked limitation of physical activity and symptoms at rest. A partially external device such as a VAD could potentially negatively impact patients’ ability to perform ADLs, which may have other downstream effects such as increasing caregiver burden. However, our study demonstrated that the VAD did not appear to have a negative effect on ADLs. It is possible that improved cardiac output after VAD placement may improve ability to perform ADLs compared to baseline.

Quality of Life Outcomes

In this study, we observed an 8.6 point improvement at four weeks, compared to prior to implantation and a 25 point improvement at 36 weeks, compared to prior to implantation. To put these changes into perspective, a five-point change in the KCCQ has been previously determined to be clinically meaningful.²²

There was a slight decrease in scores at 48 weeks; however, there were only 16 patients remaining in the cohort at this time and 63% of this group were destination therapy patients versus 51% of patients in the total cohort. This decrease may be the result of a “remainder effect” where healthier patients were transplanted and were thus censored, leaving behind the sickest patients. It is also possible that there may be a component of “VAD-fatigue” where patients were weary of being dependent on the VAD at one year after placement of this device.

Our findings of the improvement of the KCCQ Overall Summary Score (which is derived from the physical function, symptom, social function, and quality of life domains) over time are consistent with prior studies. Previous studies examining these issues have demonstrated significant improvements in KCCQ overall summary scores, Minnesota Living With Heart Failure (MLWHF) scores, Quality of Life Index, Heart Failure Symptom Checklist, Sickness Impact Profile, LVAD Stressor Scale, the Jalowiec Coping Scale, the EuroQol questionnaire, and the vertical visual analogue scale.^{7, 36–39}

We found similar results for KCCQ Clinical Summary Score, which is derived from the symptom and physical limitation domain. Together, these data reinforce the findings from other investigators that VADs have the potential to significantly improve the overall quality of life for patients with severe HF.

Limitations

Our study is limited by high non-response rates at certain time points, particularly peri-implant. This may be because the window of time between the decision to proceed with a VAD and the implant surgery can be as short as a few hours. This made the task of completing surveys before the VAD implantation challenging. In addition, the surveys done immediately after implant were performed by cardiology nurse-practitioners in the cardio-thoracic intensive care unit, who also had other primary responsibilities such as teaching the patient and his/her family about VAD management. We chose to analyze our data with linear mixed models in order to accommodate missing data. Linear mixed models can utilize subjects with missing data points as the time points with data still contribute to the overall maximum likelihood of the regression model and provide us information with the overall pattern of symptoms over time.

Non-response at later times may have been because patients were too ill to complete an in-depth survey. Patients may have suffered from complications or been readmitted, potentially resulting in ascertainment bias. We did not collect information on readmissions or complications in this study, and thus are limited in our ability to speculate why patients did not respond. However, this may be offset by patients who were healthier and were thus able to undergo transplantation. As noted above, a larger proportion (63%) of this group remaining at 48 weeks were destination therapy patients.

Conclusions

As VADs are increasingly adopted for the treatment of advanced HF, it is vitally important to not only investigate their impact on survival, but also on overall quality of life, symptoms, and functional status. These key indicators are important to patients and their families. We found that patients with VADs experienced improved pain, functional status, and quality of life, over time. The natural history after VAD implantation may have clinical implications. For example, one or two follow-ups with the palliative care team in the weeks post-implant where pain and ADLs are assessed may be sufficient to meet the needs of the vast majority of these patients. While the series of measures that we used in this study may be too cumbersome for routine practice, shorter versions of these instruments might be adapted to be used as screening tools or automatic triggers for palliative care consultation. This will help palliative care teams better allocate personnel to the patients most in need, assuring better quality of care for patients with advanced cardiac disease and their families.