Written by: Jeffrey WangAlexis K OkohYuxuan ChenRebecca S SteinbergApoorva GangavelliKrishan J PatelYi-An KoJeffrey D AlexisShivani A PatelDavid J VegaMani DaneshmandErsilia M DefilippisKhadijah BreathettAlanna A Morris
The INTERMACS database includes adult patients (≥ 19 years old) who underwent implantation of continuous-flow durable mechanical circulatory support from 2008–2017 (n = 20,931). The data are publicly available through the National Heart, Lung, and Blood Institute Biologic Specimen and Data Repository Information Coordinating Center repository. Patients who did not have their first LVAD implantation recorded in the INTERMACS registry (n = 16), or who received a durable right ventricular assist device only or a total artificial heart were excluded from this analysis (n = 492). This study was approved by the Emory Institutional Review Board.The primary exposure of interest was the presence of psychosocial risk factors, defined as the presence of any 1 of the following1: substance abuse (alcohol and/or drug)2; limited social support3; limited cognitive understanding4; repeated nonadherence5; and psychiatric disease (including any major psychiatric diagnosis or severe depression).4,8 These data are collected and documented in INTERMACS at the time of LVAD implantation, based on patient self-report and medical chart reviews. The presence of narcotic dependence was not incorporated into the history of substance abuse, given the degree of missing data. Patients were excluded from the analytic cohort if data about presence of psychosocial risk factors were missing (n = 5539).The Kansas City Cardiomyopathy Questionnaire (KCCQ) assesses HRQOL in patients with HF and has been validated in multiple clinical settings.28–30 The 12-item KCCQ (KCCQ-12) quantifies HRQOL by assigning a score ranging from 0–100 after assessment of 4 domains: physical limitation, symptom frequency, social limitation, and quality of life. The KCCQ-12 was administered either by the participant (ie, self-reported), the research coordinator or the family at baseline (preimplantation) and at 1-year post-implant. In this analysis, the overall summary score was used as reported in the INTERMACS database. Hereafter, we refer to the KCCQ-12 overall summary score as the KCCQ-12 score. To assess changes in HRQOL, we calculated the change in the KCCQ-12 score from baseline to 1 year post-implant, and we defined persistently low HRQOL based on an overall KCCQ-12 score of < 45 at baseline and at the 1-year follow-up.31 Only patients whose KCCQ-12 scores were available at baseline and at 1-year follow-up (n = 2936) were included in the analyses, so as to determine persistently low HRQOL and change in KCCQ-12 scores at 1 year, whereas patients missing KCCQ-12 scores either at baseline (n = 1765) or at 1-year follow-up (n = 5131) were included in analyses where appropriate. Patients missing both KCCQ-12 scores at baseline and at 1-year follow-up were excluded from analysis (n = 5565).
Data are presented as mean ± standard deviation (SD), median (interquartile range [IQR]), and N (proportion), as appropriate. Differences in baseline characteristics between patients with and without psychosocial risk factors were compared by using the independent t test for continuous variables and the χ2 test for categorical variables. Multivariable logistic and linear regression were used to determine whether the presence of psychosocial risk factors was associated with persistently low HRQOL and change in KCCQ at 1 year, respectively. Kaplan-Meier analysis was used to compare the cumulative incidence of all-cause hospital readmission at 1 year between those with and without psychosocial risk factors. Univariate and multivariable Cox proportional hazards regression was used to determine the association of psychosocial risk factors with hospital readmission at 1 year. All multivariable models were adjusted for age, gender, self-identified race and ethnicity, marital status, education level, working for income, device strategy, INTERMACS profile, serum creatinine (pre-implant), serum albumin (pre-implant), concomitant implantation of temporary right ventricular assist device, implant year, and history of previous cardiac surgery. All P values are 2-sided tests, with P < 0.05 considered statistically significant. All analyses were conducted in R (version 4.2.1, R Foundation for Statistical Computing, Vienna, Austria).
Results
There were 20,931 patients in the INTERMACS database who received continuous-flow durable LVADs between 2008 and 2017 (Fig. 1). Of these patients, 5534 were excluded due to missing data about psychosocial risk factors, and an additional 5565 patients were excluded due to missing KCCQ-12 scores, at both baseline and 1 year post-implant. This resulted in a final analytic cohort of 9832 patients with complete or partial KCCQ-12 score data. (Table 1) and (Supplementary Table 2) demonstrate that the baseline characteristics of those patients excluded due to missing data are similar to the baseline characteristics of the patients who compose the final analytic cohort.
In the final analytic cohort (n = 9832), 2024 (20.6%) patients had 1 or more psychosocial risk factors. Substance abuse (n = 1253; 12.8%) was the most prevalent psychosocial risk factor, specifically alcohol abuse (n = 772; 7.9%), followed by limited social support (n = 481; 4.9%), major psychiatric disorder (n = 394; 4.0%), repeated nonadherence (n = 336; 3.4%), and limited cognitive understanding (n = 168; 1.7%). Compared to those without psychosocial risk factors, patients with psychosocial risk factors were younger, more likely to be male, and more likely to self-identify as Black (Table 1). Patients with psychosocial risk factors were also less likely to be married, to be working for income and to have lower education attainment. Patients with psychosocial risk factors were more likely to be in New York Heart Association (NYHA) class IV, had higher severities of illness based on INTERMACS profiles at the time of implant, and were more likely to have LVADs implanted as destination therapy.
At baseline, there were no significant differences in KCCQ-12 scores between patients with and without psychosocial risk factors (Table 1). Moreover, there were no significant differences in KCCQ-12 scores when comparing patients with scores present both at baseline and at 1 year, at baseline only, or at 1 year only (Supplementary Table 2). At 1 year post LVAD implant, patients both with and without psychosocial risk factors demonstrated improvements in KCCQ-12 scores from baseline (Table 2). After adjustment for differences in baseline characteristics, the improvement in KCCQ-12 scores was 4.44 to −3.35) points lower in patients with psychosocial risk factors compared to those without such factors (Table 3).
Patients with psychosocial risk factors were also more likely to have persistently poor HRQOL from baseline to 1 year. Even after adjustment for differences in baseline characteristics, the presence of psychosocial risk factors pre-implant remained associated with persistently poor HRQOL (adjusted odds ratio [aOR] 1.41, 95% CI: 1.06–1.88) (Table 4). When examining each psychosocial risk factor individually, limited social support (aOR: 1.66 95% CI:1.04–2.66) and major psychiatric disorder (aOR: 1.91 95% CI: 1.20–3.03) were associated with persistently poor HRQOL, whereas substance abuse, limited cognitive understanding and repeated nonadherence were not.
Compared to patients without psychosocial risk factors, those with psychosocial risk factors had higher risks of all-cause hospitalization at 1 year (Fig. 2). Even after adjustment for differences in baseline characteristics, the presence of psychosocial risk factors remained associated with higher risk of rehospitalization at 1 year (adjusted hazard ratio 95% CI: 1.09–1.25; P < 0.001 (Table 5). However, when examining each psychosocial risk factor individually, substance abuse (aHR: 1.12, 95% CI: 1.03–1.22), limited cognitive understanding (aHR 1.39, 95% CI: 1.12–1.71) and major psychiatric disorder (aHR: 1.14, 95% CI: 1.00–1.23) were associated with a higher risk of 1-year all-cause rehospitalization. There was no association between limited social support and repeated nonadherence to rehospitalization at 1-year.
Discussion
In this analysis of the INTERMACS registry, we examined the association of psychosocial risk factors with HRQOL after LVAD implantation. We found that ~20% of patients undergoing LVAD implantation have psychosocial risk factors. There were no differences in KCCQ scores between those with and without psychosocial risk factors preimplant, but patients with psychosocial risk factors had smaller improvements in KCCQ scores 1 year post-implant compared to those without psychosocial risk factors. However, this difference was lower than the traditionally accepted threshold of 5 points that is typically considered to be a clinically significant difference in KCCQ scores.32 The presence of psychosocial risk factors was also associated with persistently poor HRQOL and 1-year all-cause hospital readmissions. However, when examining individual psychosocial risk factors, major psychiatric disorders and limited social support were associated with persistently poor HRQOL, whereas major psychiatric disorders, substance abuse and limited cognitive understanding were associated with 1-year all-cause hospital readmission. These findings suggested that patients with psychosocial risk factors derive meaningful clinical benefit post-implant and that some psychosocial risk factors portend a more meaningful prognosis than others.
The 2018 International Society of Heart and Lung Transplant guidelines recommend assessment of psychosocial risk factors as part of the comprehensive assessment for advanced therapies.5 Our findings support prior literature demonstrating a high prevalence of psychosocial comorbid conditions among patients with advanced HF. In our analysis, 20% of patients in the INTERMACS registry had 1 or more psychosocial risk factors present at the time of LVAD implantation. Substance abuse was the most prevalent risk factor, followed by limited social support, major psychiatric disorder, repeated nonadherence and, finally, limited cognitive understanding. A recent analysis of the National Inpatient Sample demonstrated that substance abuse is also common among patients with HF, with a tobacco-use disorder prevalence of 10.7%, and an alcohol-use disorder prevalence of 2.7%.33 Similarly, in a meta-analysis of 27 studies, the prevalence of severe mental health disorders was found to be 42% among patients with NYHA class IV HF, compared to 11% among patients with NYHA class I HF.34
The use of validated tools to assess HRQOL is supported by the most recent HF Clinical Practice Guidelines and by governmental agencies, including the Food and Drug Administration.24,35 Previous analyses investigating HRQOL (as measured by the KCCQ) among LVAD recipients have found that in addition to clinical factors and adverse events, psychosocial metrics, including health-related stress, inability to cope with stress and lack of confidence in VAD-related care, are associated with lower KCCQ scores. A recent analysis of the INTERMACS registry also found that patients who reported substance abuse (including alcohol and illicit drug use) prior to LVAD implantation had lower KCCQ scores at 6, 12 and 18 months compared to those without reported substance abuse.36 Our analysis is novel in that we assessed the association of multiple psychosocial risk factors (including substance abuse, social support, major psychiatric disorder, and non-adherence) with HRQOL among LVAD recipients. Our findings confirm and extend prior work documenting that multiple psychosocial risk factors, including substance abuse and others, may have adverse impacts on improvement in HRQOL after LVAD implant. In our multivariable models examining the association between psychosocial risk factors and change in KCCQ-12 scores at 1 year, the presence of psychosocial risk factors was associated with a 3.5-point lower improvement in KCCQ-12 scores. While this difference was statistically significant, it is below the generally accepted threshold of 5 points that is considered to be clinically meaningful,32 indicating that patients with psychosocial risk factors still derive a meaningful improvement in HRQOL after LVAD implantation.
Management of and support for psychosocial risk factors are being recognized increasingly as a target area to improve outcomes in LVAD recipients. Prior analyses document that the presence of psychosocial risk factors is associated with device-related morbidity, including infections, gastrointestinal bleeding, pump thrombosis, and all-cause readmissions.4,36 Our findings confirm that the risk of readmissions post-LVAD implant is high among all LVAD recipients, albeit slightly higher among those with psychosocial risk factors. Given the small absolute difference, our findings suggest that most of the risk for hospitalization in the 1st year after LVAD implantation is attributable to factors other than psychosocial risk factors, including severity of illness, device complications, bleeding, etc. The association with LVAD-related mortality rates is less well described. A recent analysis from INTERMACS found no association of psychosocial risk factors with mortality among LVAD recipients.4 However, a single-center, matched-cohort study of 60 patients found that those with active substance abuse at the time of LVAD implant had higher risks for death and driveline infection. Among the patients with substance abuse, the majority (13/20, [65%]) abused alcohol; however, other substance abuse was reported, including methamphetamines (3 [15%]), crack/cocaine (1 [5%]), narcotics (1 [5%]), and multiple substances (2 [10%]).
Understanding which psychosocial risk factors have the most robust association with adverse outcomes in LVAD recipients is critical as we consider methods to improve health equity. A recent multicenter analysis of patients referred for advanced HF therapies evaluation across the United States demonstrated that ~ 20% of candidates for LVAD are considered ineligible due to psychosocial reasons.3 While several objective tools (eg, SIPAT, PACT) are available to standardize psychosocial evaluations for advanced HF therapies, studies suggest that implicit bias may still influence which patients are more likely to be perceived as adequate candidates. A prior single-center analysis demonstrated that female candidates were less likely to be eligible for advanced therapies, even though male candidates had higher SIPAT scores than females, due primarily to historical or current substance use.37 National survey studies confirm that female candidates are more likely to have perceived barriers to eligibility for advanced therapies, including the assumed burden of childcare and lack of social support.38 It is currently unclear how well perceived risk pre-implant correlates with actual risk post-implant. Psychosocial risk factors are an important component in evaluating candidacy for advanced therapies, but decisions should be holistic, not grounded in the presence of psychosocial risk factors alone. It is important to note that these data are representative only of patients whose psychosocial risk factors were perceived to be low enough that they were still considered eligible for LVADs.
There are several important limitations of our analysis that must be considered. INTERMACS collects data only on LVAD recipients, who have already been deemed eligible for advanced therapies by their respective centers. Thus, our findings cannot be extrapolated to all patients undergoing LVAD evaluation, some of whom may, indeed, have psychosocial risk profiles that are likely to portend worse prognoses with LVADs than is captured in the current analysis. Second, INTERMACS does not provide detailed information about substance abuse, including severity of abuse or whether use is current or past. Although INTERMACS gathers data about substance use and alcohol abuse, a significant portion of the data concerning narcotic dependence is missing. Finally, many patients in INTERMACS had missing data in both psychosocial risk factors as well as KCCQ-12 scores. It is reasonable to conclude that these data may not have been missing at random, particularly among patients who are critically ill (ie, on extracorporeal membrane oxygenation or mechanical ventilation pre-implant). The comparisons documented in the Supplementary Tables suggest that the patients who compose our analytic cohort are generally representative of patients in the INTERMACS database, but there were some statistically significant differences in those missing and not missing psychosocial risk factors, such as INTERMACS profile, mechanical circulatory support, mechanical ventilation, and the 6-minute walk test. Finally, the time period for this analysis was prior to the 2018 heart transplant allocation change and, therefore, contains a higher proportion of patients implanted for bridge to transplant.
Conclusion
In conclusion, this analysis of the INTERMACS registry suggests that among patients with advanced HF who receive LVADs, the presence of psychosocial risk factors is associated with smaller improvements in HRQOL as well as higher risks for readmissions. While these differences were enough to reach statistical significance, it is unclear how clinically meaningful these differences are.32 Further, major psychiatric disorder was the only psychosocial risk factor associated with both endpoints. Psychosocial risk factors are a common reason for patients to be excluded from eligibility for LVADs, but our findings highlight the need for more research to further characterize psychosocial risk in both the pre- and the post-implant phase and to identify effective methods to modify and manage psychosocial risk factors.
AG
Dr. Apoorva Gangavelli
Cardiology · General Non-Invasive CardiologyEmory Universityapoorva