The Association of Provider-Assessed Psychosocial Risk With Outcomes in Destination Therapy Left Ventricular Assist Device Patients: An Intermacs Registry Analysis

Society guidelines recommend assessment of psychosocial risk, including treatment adherence, mental health, substance use, cognitive status, and social support, in candidates for long-term left ventricular assist device (LVAD) support (both bridge to transplant [BTT], and destination therapy [DT] candidates).1 However, validated and objective tools developed specifically for the assessment of psychosocial risk in this population are lacking. This may contribute to underutilization, or delay in allocation of lifesaving LVAD therapy. Indeed, psychosocial risk factors (particularly inadequate social support) contribute to about a quarter of rejections or deferrals for LVAD therapy, and 1 year survival is only 64% in such patients.2 The importance of psychosocial risk factors while assessing patients who are candidates for heart transplantation is irrefutable given the impact of these factors on post-transplant outcomes, contextualized against the ethical considerations associated with allocation of a limited resource. On the other hand, there are sparse data on the impact of psychosocial factors on outcomes in patients who are not transplant candidates (DT LVAD), who warrant distinct risk-benefit considerations. These patients represent a higher-risk population, typically older in age with more comorbidity. However, current-generation LVADs are associated with significant survival benefit (median survival >5 years) and improvements in quality of life (QOL) in these patients3; and although an expensive therapy, LVADs are not a limited resource. The primary aim of this study is to explore the association of psychosocial factors with post-DT LVAD outcomes. Additionally, because prior research suggests that vulnerable populations such as Black patients and women are likely to be more harshly evaluated for psychosocial risk; we report on differences in psychosocial risk by sex and race.4 The study population included DT patients from the Intermacs registry who underwent primary LVAD implant between 2009 and 2017, and excluded Right Ventricular Assist Device (RVAD) recipients. Psychosocial variables captured in the registry include limited social support, limited cognitive understanding, repeated noncompliance, alcohol abuse, drug use history, severe depression, and other major psychiatric illnesses. For this study, we considered a psychosocial risk factor present if it was thought significant enough to be a contraindication to transplant listing by the medical and surgical team. We compared prevalence of psychosocial risk by sex and race using chi-squared tests. Multivariable Cox proportional hazard models were used to examine the association between psychosocial risk with all-cause mortality, and LVAD-related adverse events (AE) including driveline infection and hemocompatibility-related AEs (HRAE: stroke, GI bleed, or pump thrombosis). Poisson regression models were used to examine the association between psychosocial risk and rehospitalization count within 1 year of implant. Models were adjusted for age, sex, race, body surface area, history of sternotomy, LVAD implant year, heart failure etiology, pump type, INTERMACS profile, creatinine, blood urea nitrogen, bilirubin, albumin levels, and needing inotropic support for 2 or more weeks after implantation. Each variable had less than 10% missing data. Additionally, we performed a sensitivity analysis examining the association of psychosocial risk with improvements in QOL in the subgroup of patients that had this data captured (n = 1,429). Outcomes included improvement in Kansas City Cardiomyopathy Questionnaire-12 (KCCQ-12) summary scores, stratified by the cohort’s median change (+33.3) over 1 year follow-up, and persistently low health-related QOL (HRQOL) defined as KCCQ-12 score less than 45 at baseline and at 1 year follow up.5 This study was exempt from institutional IRB approval. Of 8,072 DT patients (mean age: 62 ± 12 years, 20% female, 24% Black), 1,089 (13.5%) had at least 1 psychosocial risk factor. The proportion of patients with psychosocial risk factors implanted in the later part of the study period was slightly higher than those implanted in the earlier era (2009–2012: 12.1%, 2013–2017: 14.0%, p = 0.03). Compared to White patients, Black patients were more likely to be assessed as having psychosocial risk factors, including limited social support, limited cognitive understanding, repeated noncompliance, alcohol abuse, and history of drug use. Compared to men, women were less likely to have documented alcohol abuse (Table 1). Table 1. - Distribution of Psychosocial Risk by Sex and Race Subgroup (n = 8,072) Contraindication FemaleN = 1,626 (20.1%) MaleN = 6,446 (79.9%) p Value BlackN = 1,930 (23.9%) WhiteN = 5,666 (70.2%) OtherN = 476 (5.9%) p Value ≥1 risk factors 203 (12.5) 886 (13.8) 0.32 404 (20.9) 609 (10.8) 76 (16.0) <0.001 Limited social support 84 (5.2) 270 (4.2) 0.21 157 (8.1) 173 (3.1) 24 (5.0) <0.001 Limited cognitive understanding 17 (1.1) 86 (1.3) 0.64 36 (1.9) 56 (1.0) 11 (2.3) 0.001 Noncompliance 57 (3.5) 221 (3.4) 0.94 124 (6.4) 129 (2.3) 25 (5.3) <0.001 Alcohol abuse 28 (1.7) 268 (4.2) <0.001 93 (4.8) 181 (3.2) 22 (4.6) 0.002 History of drug use 52 (3.2) 263 (4.1) 0.24 133 (6.9) 169 (3.0) 13 (2.7) <0.001 Depression 17 (1.1) 44 (0.7) 0.32 17 (0.9) 43 (0.8) 1 (0.2) 0.32 Other major psychiatric diagnosis 12 (0.7) 58 (0.9) 0.81 14 (0.7) 51 (0.9) 5 (1.1) 0.70 On multivariable analysis, none of the assessed psychosocial risk factors associated with mortality after LVAD. Patients with limited social support (hazard ratio [HR]: 1.39, 95% confidence interval [CI]: 1.10–1.76) and a major psychiatric diagnosis other than depression (HR: 1.89, 95% CI: 1.24–2.88) had a higher hazard of driveline infection, and those with a history of noncompliance had a higher risk of HRAE (HR: 1.20, 95% CI: 1.01–1.43). Patients with limited social support (rate ratio [RR]: 1.13, 95% CI: 1.06–1.20), noncompliance (RR: 1.09, 95% CI: 1.02–1.18), and a major psychiatric diagnosis other than depression (RR: 1.55, 95% CI: 1.39–1.74) also had a higher 1 year rehospitalization rate. There were no associations between these risk factors and improvement in KCCQ-12 scores and HRQOL in the sensitivity analyses (Table 2). Table 2. - Associations Between Psychosocial Risk Factors and Clinical Outcomes Cox Models (HR, 95% CI) Poisson Model (RR, 95% CI) Logistic Regression (OR, 95% CI) Contraindication All-Cause Mortality (38.1%) Driveline Infection (19.0%) HRAE (55.4%) Rehospitalization Rate at 1 Year (85.8%)* KCCQ-12 Score Change >+33.3 at 1 Year (50.5%) Persistently Poor HRQOL 1 Year After Implantation (14.3%) >1 risk factors 1.02 (0.89–1.17) 1.22 (1.05–1.43)† 1.10 (0.99–1.22) 1.08 (1.03–1.12)† 0.90 (0.65–1.26) 1.30 (0.81–2.08) Limited social support 1.08 (0.87–1.34) 1.39 (1.10–1.76)† 1.12 (0.96–1.32) 1.13 (1.06–1.20)† 0.75 (0.44–1.27) 1.19 (0.58–2.44) Limited cognitive understanding 1.14 (0.79–1.64) 1.32 (0.86–2.04) 1.21 (0.91–1.61) 1.11 (0.99–1.24) 1.35 (0.43–4.29) 0.64 (0.08–5.11) Noncompliance 1.17 (0.92–1.48) 1.27 (0.98–1.65) 1.20 (1.01–1.43)† 1.09 (1.02–1.18)† 1.19 (0.65–2.16) 2.04 (0.99–4.18) Alcohol abuse 0.99 (0.76–1.31) 1.12 (0.83–1.50) 1.11 (0.92–1.33) 0.99 (0.91–1.07) 1.12 (0.60–2.09) 0.95 (0.36–2.52) History of drug use 0.91 (0.69–1.19) 1.06 (0.82–1.38) 0.85 (0.70–1.03) 1.07 (0.99–1.15) 1.15 (0.63–2.10) 0.82 (0.33–2.07) Depression 0.74 (0.41–1.34) 1.39 (0.85–2.29) 1.21 (0.84–1.74) 1.02 (0.88–1.19) 1.06 (0.33–3.41) 2.81 (0.72–11.01) Other major psychiatric diagnosis 1.46 (0.96–2.21) 1.89 (1.24–2.88)† 1.12 (0.80–1.59) 1.55 (1.39–1.74)† 0.63 (0.21–1.97) 1.80 (0.48–6.76) All models adjusted for age, sex, race, body surface area, prior sternotomy, implant year, heart failure etiology, pump type (centrifugal vs. axial), INTERMACS profile, creatinine, blood urea nitrogen, bilirubin, albumin levels, and needing inotropic support for 2 or more weeks after implantation.*Amongst those patients on pump support at 1 year after implant (N = 4,215).†A statistically significant association.CI, confidence interval; HR, hazard ratio; HRAE, hemocompatibility-related adverse events; HRQOL, health-related quality of life; KCCQ, Kansas City Cardiomyopathy Questionnaire; OR, odds ratio; RR, rate ratio. The main findings of this study are 1) baseline psychosocial risk does not predict survival benefit or QOL improvement from DT LVAD implant, and 2) a few psychosocial risk factors (specifically limited social support, noncompliance, and a major psychiatric diagnosis other than depression) associate with LVAD-related AE, however the majority of routinely assessed psychosocial risk factors do not. Although this suggests that programs are able to successfully identify patients who are likely to do well after implantation despite a higher-risk psychosocial profile at baseline, the current approach may still be too conservative. It is concerning that the proportion of patients with psychosocial risk factors being implanted only slightly increased during the course of the study period despite data suggesting an increasing prevalence of these risk factors.6 There are limited data on the association of psychosocial risk with outcomes in DT LVAD patients. DeFilippis et al.7 found that selected psychosocial risk factors predicted survival on pump support and several LVAD-related AE in the Intermacs registry. However, they included both DT (<50% of their study population) and BTT patients in their study, and did not report a subgroup analysis by device strategy.7 Furthermore, our findings are robust to specifically examining psychosocial risk factors that were considered significant enough to be a contraindication to transplant listing. Tools for assessment of individual psychosocial risk factors, particularly adequacy of social support, are lacking.1 Since inadequate social support is a heavily weighted relative contraindication to LVAD, closer attention to nuances in this psychosocial risk factor is important.8 Cultural psychological studies reveal great heterogeneity in cultural preferences for social support, with some cultures tending to rely more heavily on explicit support from family and friends, whereas others may view such attentiveness or support requirements as burdensome.9 In the interim, a few centers have relaxed social support requirements, devising “loner protocols” for individuals who live alone but are otherwise at lower risk for adverse outcomes after LVAD. Although there are clear limits to the granularity of Intermacs data for psychosocial assessment as well as healthcare resource utilization and costs (such as urgent clinic visits, emergency department visits and LVAD coordinator time), and the data herein does not include HeartMate 3 LVAD (Abbott, Abbott Park, IL) implants; the lack of strong data supporting restriction of LVAD therapy based on perceived psychosocial risk should prompt heart failure societies and LVAD programs to reconsider programmatic psychosocial requirements. There is a need to develop reliable and reproducible psychosocial risk assessments specifically validated for use in this population (Figure 1).Figure 1.: Central figure: Association of provider assessed psychosocial risk with outcomes in Destination Therapy Left Ventricular Assist Device patients.Acknowledgment Biorender.com was used to create Figure 1 for this manuscript.