Clinical implications of frailty in hospitalized patients with pulmonary arterial hypertension

Abstract

Objectives

Frailty is associated with increased morbidity and mortality in chronic lung disease, but its prognosis has not been evaluated in pulmonary arterial hypertension (PAH). This study aimed to assess: (1) impact of frailty on hospital length of stay (LOS) and health-care utilization in PAH; (2) association of frailty with 1-year post-discharge outcomes.

Methods

Retrospective, single-centered cohort study of consecutive PAH patients admitted non-electively (January 2009-December 2018), predominantly for right heart failure (57%). Frailty was defined as ≥ 0.25 using a cumulative deficits frailty index. Disease characteristics, hospital factors, and mortality were compared using univariate analysis and multivariable regression, adjusting for age and sex.

Results

44/96 (46%) PAH patients were frail. Frailty was associated with older age, greater comorbidities, and lower six-minute walk distance pre-admission (p < 0.05). Frail patients had a longer hospital LOS (4 days 95% (0.4–6.3), p = 0.04) and were more likely to receive social work consultation (36% vs 13%, p = 0.01), independent of age and sex. There were no adjusted differences (frail vs non-frail) in hospital mortality (OR:1.01 95% (0.28–3.72) or 12-months mortality post-discharge (HR:1.26 95% (0.48–3.29).

Conclusion

Frailty was associated with greater hospital LOS and interdisciplinary support, but not 1-year mortality. Future studies should explore whether alternative frailty models may be more informative of longer-term PAH outcomes.

Introduction

Pulmonary arterial hypertension (PAH) is an under-recognized condition resulting in significant morbidity and economic burden. ¹ PAH has a progressive course associated with impairments in many aspects of daily life, including physical and psychosocial function. ¹ PAH carries a poor prognosis with the median survival estimated at 3 years after diagnosis in the absence of pharmacotherapy or lung transplantation. ²

Given the progressive nature of PAH, these patients are at high risk for hospitalizations. The United States REVEAL Registry characterized hospitalizations of 862 PAH patients with greater than half of patients admitted during their first year after PAH diagnosis with a wide range of factors (i.e. pneumonias, gastrointestinal bleeding, arrhythmias, etc.). ³ Given the high incidence of hospitalizations, evaluation of prognostic factors may help with hospital management and discharge planning. ⁴ Presently, prognosis relies on a multifaceted approach based on clinical presentation, pulmonary vascular hemodynamics, and exercise capacity.^5,6 However, a prognostic multidimensional measure such as frailty, a syndrome characterized by an accumulation of physiologic deficits across multiple organ systems, has not been evaluated in PAH.

Several tools have been developed and applied in both clinical practice and research to assess frailty. For example, the Short Physical Performance Battery uses functional tests including gait speed, sit-to-stand and balance tests to categorize frailty. ⁷ The Fried frailty phenotype has also been commonly used, where it evaluates five criteria: unintentional weight loss, self-reported exhaustion, muscle weakness, slow walking speed, and reduced physical activity levels.^8,9 Although the frailty phenotype has been shown to be prognostic of adverse clinical outcomes, it relies on physical assessments and does not integrate psychosocial aspects, which are important elements of daily function, hospitalization, and recovery post-hospital discharge. ¹⁰ The cumulative deficits frailty index (FI) assesses the proportion of health deficits present over the total number of deficits considered.^11,12 The FI encompasses a multi-dimensional assessment that includes physical, psychological, and social parameters. ¹¹ A number of frailty indices have been applied in several studies of chronic lung disease (CLD) ¹³ but to our knowledge, a FI has not been previously utilized in PAH patients.

Frailty is important in the inpatient setting as hospitalizations are a significant source of physiologic stress. Frailty is often associated with a lower adaptive response to physiological stressors and increased risk of adverse health outcomes post-hospital discharge. ¹⁴ It has been associated with longer hospital length of stay (LOS), greater disability, and mortality in non-PAH populations with CLD.^13,15 However, PAH patients are unique as they often have an increased risk of bleeding, predisposition to right-sided heart failure, and a higher risk of infections due to vasodilatory therapy.^16,17 Thus, understanding the contribution of frailty in PAH patients may provide unique insights related to hospital outcomes and functional recovery post hospital discharge.

Given the burden of hospitalizations and unique clinical characteristics of PAH, we aimed to assess the effect of frailty using the FI on: (1) hospital LOS, palliative symptom burden, resource utilization, and discharge disposition, and (2) 1-year outcomes post-hospital discharge including exercise capacity, hospital re-admissions and mortality. We hypothesized that frailty would be associated with a longer hospital LOS, increased consultation of services during hospital, and greater mortality 1-year post-hospital discharge.

Methods

This was a retrospective cohort of consecutive adult patients with PAH admitted to Toronto General Hospital with an index admission between January 1, 2009-December 31, 2018, with 1-year follow-up period post hospital discharge until December 31, 2019. The study was approved by the Research Ethics Board at the University Health Network (REB 20-6345).

PAH patients were included if they were admitted for an acute indication such as right-sided heart failure, infection (i.e. lower respiratory tract infection, sepsis, etc.), or bleeding (i.e. hemoptysis, gastrointestinal, etc.). Patients were excluded if they: (1) had no PAH diagnosis, ¹⁸ (2) had adult congenital heart disease, (3) actively listed for lung transplant (LTx) on hospital admission, or (4) admitted electively for PAH pharmacotherapy management.

Frailty assessment

Frailty was defined using a 39-point FI comprised of comorbidities (n = 26), laboratory measures (n = 11), and functional parameters (n = 2) (Supplemental Table 1). The FI scale has been previously utilized in LTx and CLD populations at our center, and other studies.^12,19–21 It was adapted for this study based on relevant and available PAH parameters. Specifically, items in our FI were included if they represented health-related deficits spanning diverse physiologic systems (e.g., cardiac, metabolic, hematologic), and if they achieved a minimum of 80% data completeness. This was in line with previously established procedures used to develop the FI. ¹²

Deficits included in the FI were chosen to encompass a range of clinical domains, with each deficit scored as 0 (absent/normal) or 1 (present/abnormal).^12,19 As per convention, each patient’s FI was calculated as a ratio of present deficits to the total number of deficits.^11,12 A FI ≥0.25 was used to define frailty, consistent with CLD literature.^19,20,22 The frailty scoring was performed by one research team member (AL) with approximately 10% of the patients verified by a secondary reviewer (DR), and any discrepancies or uncertainties discussed for all participants.

Clinical characteristics

Clinical characteristics such as age, sex, body mass index, and previous hospital admissions (within 12 months of index admission) were abstracted from chart review. Assessment of daily function, comorbidities and disease severity were evaluated using the New York Heart Association (NYHA) functional class, the Charlson Comorbidity Index (CCI), and COMPERA-2.0 score, respectively.^23–25 Comorbid conditions were quantified using the CCI, a weighted index used to predict the burden of comorbidities. ²³ Disease severity was also characterized using the COMPERA-2.0 score, a risk stratification tool used in patients with PAH to estimate their 1-year mortality risk. ²⁵ Parameters captured in this tool include NYHA functional class, 6MWD, and BNP. The low and intermediate-low risk scores were grouped together, as previous findings demonstrated no difference in survival between these categories. ²⁵ Performance of ADLs and IADLs was categorized as either independent or dependent based on clinical history found in progress or multidisciplinary assessment notes. Additionally, supplemental oxygen use at rest and with activity, home care supports and PAH-specific therapies were extracted.

Pulmonary function and exercise capacity

Baseline pulmonary function and six-minute walk test (6MWT) data were abstracted using the closest measurement preceding hospital admission within 12 months. Both tests were performed in the Pulmonary Function Lab at the University Health Network as per American Thoracic Society standards. ²⁶ The parameters abstracted included forced expiratory volume in 1 s (FEV₁), forced vital capacity (FVC), and the diffusion capacity for carbon monoxide (DLCO). Data collected from the 6MWT included the distance walked and percent predicted, and the use of any gait aids.

Pulmonary vascular hemodynamics

Pulmonary vascular hemodynamics were characterized using transthoracic echocardiography (TTE) and right heart catheterization measurements. These measures were obtained using the closest test preceding the index hospital admission. TTE parameters included right ventricular systolic pressure (RVSP) and the tricuspid annular plane systolic excursion (TAPSE). RVSP was categorized from reports as normal (<40 mmHg), mild, moderate, severe, or hypokinetic (severity not specified). ²⁷ TAPSE ≥16 mm was considered normal. ²⁸

Hemodynamic parameters abstracted from right heart catheterization included mean pulmonary arterial pressure (mPAP), cardiac output, cardiac index, and pulmonary vascular resistance (PVR). ²⁹ Cardiac index and PVR were calculated, if not reported in the chart, using the formulas: cardiac index = cardiac output/body surface area; PVR = (mPAP-left atrial pressure)/cardiac output.^30,31 Serum levels of B-type natriuretic peptide (BNP) were also abstracted.

Clinical factors during hospitalization

For patients with multiple hospital admissions, the first eligible admission during the study time period was the index admission. Cause of the admission was ascertained based on clinical documentation. During the index admission, medical specialties consulted, multidisciplinary supports such as social work, rehabilitative services (physiotherapy/occupational therapy) and referrals for home and community care services were collected.

For patients that had palliative care (PC) consultation, the Edmonton Symptom Assessment System (ESAS) scores were obtained, as previously applied in CLD and LTx patients.^32,33 ESAS is a patient-rated scale assessing severity of 11 symptoms on a Likert scale from 0 to 10 (e.g. dyspnea, sleeping problems, etc). ³⁴

Complications during the hospitalization were defined as any bleeding, infections, falls, and/or syncopal events. If patients were admitted to the intensive care unit (ICU), the number of patients requiring mechanical ventilation and ICU LOS were ascertained. Other clinical outcomes captured included hospital LOS (days), discharge disposition (e.g. home, rehabilitation facility, etc.), and hospital mortality.

Post-hospitalization parameters

Variables abstracted at 3, 6 and 12-months post-hospitalization included BNP and 6MWT distances. Survival was assessed up to 12-months post-discharge. For patients with no clinical records 1-year post-discharge, online obituaries were searched to determine if patients had died during this time period. Patients were categorized as lost to follow-up, if there was no documentation of death or follow-up at our center.

Statistical analysis

The statistical analysis was performed using GraphPad Prism 9.0 (GraphPad Inc, San Diego, CA) and R-Studio (Version 4.04). Continuous variables were assessed using medians (interquartile ranges (IQR), 25–75%) or mean ± standard deviation. Categorical variables were expressed using counts and percentages. The Kruskal-Wallis non-parametric test was used to evaluate the difference in hospital LOS between frail and non-frail patients. To assess differences in participation with multidisciplinary support, PC, discharge disposition, and hospital mortality, tests of proportions (chi-squared and Fishers exact) were utilized. Post-discharge outcomes were grouped based on time frame post-discharge; 3–6 months or 12 months post-discharge. Multivariable logistic and Cox regression models were applied to evaluate the differences in hospital and post-discharge outcomes, adjusted for age and sex. A p-value of <0.05 was considered statistically significant across all analyses.

Results

Study population

96 PAH patients were included out of 371 admitted to hospital during the study period (see Figure 1). Two main reasons for study exclusion were: (1) majority did not have a diagnosis of PAH (n = 51) or had other forms of pulmonary hypertension (n = 70); (2) admitted for elective reasons, such as PAH workup, optimization of PAH pharmacotherapy, or LTx assessments.

Figure 1.

Consort diagram of study participants. * Patients listed for lung transplantation at time of non-elective admission. **Non-acute admissions include being admitted for PH workup or assessment, electively admitted to manage pharmacotherapy, to optimize PAH medication, to monitor PH therapy, to reinitiate PH therapy, or expedite lung transplant assessment. Abbreviations: PAH: Pulmonary Arterial Hypertension; PH: Pulmonary Hypertension.

Frailty characteristics

Frailty was observed in 44 (46%) of the PAH population during hospital admission, with the frail group having a median FI score of 0.31 IQR [0.28–0.33] and the non-frail group 0.17 IQR [0.11–0.22]. The distribution of frailty index scores is displayed in Figure 2.

Figure 2.

Distribution of cumulative deficit scores in the study cohort. FI ≥0.25 indicated cut-off for frailty with vertical dotted line.

Baseline characteristics of patients prior to hospitalization is outlined in Table 1. Frail patients were older, had a higher BMI and a greater CCI, with p < 0.05 for all comparisons. The FI was correlated with CCI (r = 0.55, p-value <0.0001). A greater CCI ≥5 was associated with older age, greater dependence on ADLs and IADLs, and lower 6MWD, with p < 0.05 for all comparisons. (Supplemental Table 2). When stratified based on the COMPERA-2.0 risk scores (Table 1), frail patients were more likely to be in the intermediate-high or high risk categories (p = 0.006) than lower risk categories. Frail patients also had a lower six-minute walk distance (6MWD) and higher BNP levels prior to hospitalization.

Table 1.

Baseline characteristics of study cohort.

Baseline characteristics	Total (n = 96)	Frail (n = 44)	Non-frail (n = 52)	P-value
Age (median)	63 IQR [49–75]	68 IQR [58–77]	58 IQR [38–70]	0.01
Females (n, %)	69 (72%)	31 (70%)	38 (73%)	0.82
Body mass index (kg/m2)	27 ± 6	29 ± 7	25 ± 5	0.01
NYHA functional class (n = 80; 32; 48)
Class I-II	28 (35%)	8 (25%)	20 (42%)	0.13
Class III-IV	52 (65%)	24 (75%)	28 (58%)
Charlson comorbidity index	4 IQR [2–5]	5 IQR [4–6]	2 IQR [1–4]	<0.0001
COMPERA 2.0 risk (n = 69; 28; 41)
Low – Intermediate-low risk	29 (42%)	6 (21%)	23 (56%)	0.006
Intermediate-high – High risk	40 (58%)	22 (79%)	18 (44%)
Oxygen use (at rest)	34 (35%)	18 (41%)	16 (31%)	0.30
Oxygen use (on exertion)	36 (38%)	19 (43%)	17 (33%)	0.29
Gait aid	26 (27%)	17 (39%)	9 (17%)	0.02
Independence in ADLs (n = 82; 36; 46)	56 (68%)	19 (54%)	37 (82%)	0.01
Independence in IADLs (n = 82; 36; 46)	47 (57%)	14 (37%)	33 (73%)	0.004
Home care supports prior to admission (n = 60; 29; 31)	15 (25%)	9 (20%)	6 (12%)	0.38
PAH therapy				0.19
No PAH therapy a	14 (15%)	8 (18%)	6 (12%)
Monotherapy	47 (49%)	24 (55%)	23 (44%)
Dual therapy	30 (31%)	9 (20%)	21 (40%)
Triple therapy	5 (5%)	3 (7%)	2 (4%)
PAH-specific medications				0.73
Phosphodiesterase-5 inhibitors	42 (44%)	19 (43%)	23 (44%)
Endothelin receptor antagonists	45 (47%)	18 (41%)	27 (52%)
Soluble guanylate cyclase stimulator	4 (4%)	2 (5%)	2 (4%)
Prostanoid therapy	10 (10%)	6 (14%)	4 (8%)
Prostanoid therapy				0.71
Intravenous infusion	5 (5%)	3 (7%)	2 (4%)
Subcutaneous infusion	3 (3%)	1 (2%)	2 (4%)
Oral therapy	2 (2%)	2 (5%)	0
Right heart catheterization
mPAP (mmHg) (n = 86; 38; 48)	45 IQR [36–55]	44 IQR [36–49]	45 IQR [36–57]	0.23
PVR (dynes·s·cm-5) (n = 74; 34; 40)	692 IQR [501–1044]	634 IQR [457–868]	792 [578 – 1140]	0.07
Wedge pressure (mmHg) (n = 76; 36; 40)	9 IQR [6–12]	10 IQR [6–13]	8 IQR [5–12]	0.36
Cardiac output (L/min) (n = 78; 35; 43)	5 ± 5	4 ± 1	5 ± 6	0.16
Cardiac index (L/min/m2) (n = 81; 36; 45)	2.2 ± 0.7	2.3 ± 0.6	2.2 ± 0.7	0.94
Six-minute walk test
Distance (m) (n = 82; 38; 44)	298 IQR [182–382]	211 IQR [171–305]	324 [262–513]	<0.0001
Distance (% predicted) (n = 82; 38; 44)	60 ± 24	51 ± 22	67 ± 23	0.0004
Pulmonary function testing
FEV1 (% predicted) (n = 64; 28; 36)	65 ± 17	61 ± 18	67 ± 16	0.05
FVC (% predicted) (n = 64; 29; 35)	79 ± 26	74 ± 32	80 ± 23	0.14
DLCO (% predicted) (n = 49; 21; 28)	53 ± 17	45 ± 17	54 ± 21	0.36
Echocardiogram
RVSP (mmHg) (n = 89; 41; 48)	79 IQR [69–92]	75 IQR [67–87]	81 IQR [70–98]	0.14
TAPSE (mm) (n = 53; 28; 25)	14 IQR [11–18]	14 IQR [11–16]	15 IQR [12.5–18]	0.28
Right ventricle systolic function (n = 91; 42; 49)				0.69
Normal	13 (14%)	5 (12%)	8 (16%)
Mild	23 (25%)	10 (24%)	13 (27%)
Moderate	33 (36%)	17 (40%)	16 (33%)
Severe	16 (18%)	6 (14%)	10 (20%)
Hypokinesis (severity not specified)	6 (7%)	4 (10%)	2 (4%)
BNP (pg/mL) prior to admission (n = 83; 38; 45)	257 IQR [83–861]	351 IQR [158–941]	242 IQR [35–587]	0.04
Previous hospital admissions (within 12 months of index admission) b	16 (17%)	12 (27%)	4 (8%)	0.01

Data presented as mean ± SD, median [IQR] and proportions (%).

Abbreviations: ADLs: Activities of Daily Living; BNP: B-type Natriuretic Peptide; DLCO: Diffusing Capacity for Carbon Dioxide; FEV1: Forced Expiratory Volume in the first second; FVC: Forced Vital Capacity; IADLS: Instrumental Activities of Daily Living; mPAP: Mean Pulmonary Arterial Pressure; NYHA: New York Heart Association; PVR: Pulmonary Vascular Resistance; RVSP: Right Ventricular Systolic Pressure; TAPSE: Tricuspid Annular Plane Systolic Excursion.

^aPatients not on any PAH therapy due to medication intolerance (n = 6), patient declined, (n = 1), due to contraindications/other clinical decisions (n = 5) or started on PAH-specific therapy post-index admission (n = 2).

^bPrevious hospital admissions were defined as admissions that occurred within 12 months of the index hospital admission.

Hospitalization characteristics and outcomes

The most common causes for hospitalization were right-sided heart failure (57%), infection (10%), and worsening dyspnea (6%), with others listed in Supplemental Table 3. There was no difference in frailty between patients admitted for right-sided heart failure compared to other causes (p = 0.38). Frail PAH patients had a longer hospital LOS compared to non-frail patients (median 11 IQR [6–16] vs 7 [4–11] days, p = 0.004), which remained significant when adjusted for age and sex (median difference in LOS 3.7 days 95% CI [0.37–6.31], p = 0.04). There was no difference in the prevalence of hospital complications between frail and non-frail groups (Supplemental Table 4).

The most common medical consultations included LTx evaluation (14%), nephrology (13%), and rheumatology (11%). Frail patients were more likely to receive consultations from social work (16 (36%) versus 7 (13%), p = 0.01), PC (15 (34%) versus 7 (13%), p = 0.02), and physiotherapy or occupational therapy (22 (50%) versus 15 (29%), p = 0.02). When adjusted for age and sex, frailty was an important determinant of needing social work consultation, whereas age was a more important contributor of requiring rehabilitative services or PC support (Table 2). For patients assessed by PC (20 (21%)), there were no statistical differences in the median ESAS scores between those who were frail and non-frail (Supplemental Figure 1). Patients with frailty were more likely to require initiation or continuation of home supports or community care services at time of hospital discharge (13 (30%) versus 4 (8%), p = 0.01).

Table 2.

Multivariable analysis of hospital outcomes.

Parameter	Social work OR (95%CI)	Rehabilitative services OR (95% CI)	Palliative care OR (95%CI)	Hospital mortality OR (95% CI)
Frailty	3.13 (1.10–8.86)*	1.97 (0.80–4.82)	2.52 (0.87–7.26)	1.01 (0.28–3.72)
Age (per 10 years)	1.29 (0.94–1.78)	1.38 (1.05–1.82)*	1.55 (1.07–2.23)*	1.16 (0.83–1.63)
Sex	0.59 (0.18–1.90)	0.54 (0.20–1.51)	0.94 (0.30–2.95)	1.21 (0.30–4.88)
C-stat	0.72	0.70	0.72	0.54

Data presented as odds ratios with 95% confidence interval (95% CI).

*p < 0.05.

There were no differences in ICU admissions between frail 7 (16%) and non-frail patients 9 (17%), p = 0.85. Of the 16 (17%) PAH admitted to the ICU, 5 (31%) required mechanical ventilation and 9 (56%) required vasopressors/inotropic support, with no differences between the frail and non-frail groups (p > 0.05 for both comparisons). Out of the 12 (13%) hospital deaths, 7/12 (58%) patients required an ICU admission during their hospital stay. There was no difference in hospital mortality between the frail and non-frail group, (5 (11%) versus 7 (13%), p = 0.76), even after adjusting for age and sex (p = 0.98, Table 2).

Post-discharge outcomes

Of the 84 PAH patients surviving to hospital discharge, 10 patients were discharged from PH team care and 2 patients were lost to follow-up after the hospital admission (Supplemental Figure 2). As such, there were 72 patients initially followed after hospital discharge. 46 (64%) of the 72 patients were readmitted to hospital within 12 months of discharge at our center. Out of those readmitted to hospital, 21/31 (68%) were frail and 25/41 (61%) were non-frail (p = 0.55), with no difference observed when adjusted for age and sex (OR: 0.87 95% CI (0.33–2.31, p = 0.78).

Frail patients that were readmitted were more likely to have a greater CCI (5 IQR [3-6] vs 2 IQR [0-4], p < 0.001) from their baseline admission, a lower walk distance at baseline (219 IQR [187-308] vs 324 IQR [240-447] meters, p = 0.01), and greater NYHA class (p = 0.03) (Supplemental Table 5). At 12-months post-discharge, data was available for 48 individuals in our study cohort (Supplemental Figure 2).

Frail patients had a significantly lower 6MWD at 3–6 months after hospital discharge (203 IQR [173–283] vs 420 IQR [334–520] m; 47 ± 5 vs 73 ± 3 % predicted; for frail and non-frail groups respectively, p < 0.0001, Figure 3). By 12-months post discharge, frail patients compared to non-frail demonstrated some recovery in their 6MWD (350 IQR [238–398] vs 433 IQR [352–552] m; 68 ± 16 vs 75 ± 17% predicted, p = 0.12), respectively.

Figure 3.

(a)Six-minute walk distance (6MWD) measurements at baseline prior to hospitalization and post-discharge for frail and non-frail patients. Data presented as median [IQR]. (b) Percent (%) predicted for the six-minute walk distance (6MWD) at baseline prior to hospitalization and post-discharge for frail and non-frail patients. Data presented in mean ± standard error.

There were 18/72 deaths (25%) over the 12-months post-discharge, with 13/18 (72%) occurring in the first 6-months. There was no difference in all-cause mortality post-hospital discharge within 12-months between frail and non-frail groups (9/34 (26%) versus 9/45 (20%) respectively, p = 0.50) (Figure 4), with no differences after adjusting for age and sex (HR: 1.26 95% (0.48–3.29), p = 0.64). Patients lost to follow-up were more likely to be older, have a higher frailty index, and a lower 6MWD, with p < 0.05 for all comparisons (Supplemental Table 6).

Figure 4.

Kaplan-meier plot of 12-months survival post hospital discharge. 72 patients followed post-discharge. 5 patients censored as lost to follow-up prior to 12 months.

Discussion

To our knowledge, this is the first study to characterize the clinical implications of frailty in PAH patients during hospitalization and up to 1-year post-discharge. We observed that approximately half of the hospitalized PAH patients were identified as frail by the FI, with frail patients more likely to be older, have a higher CCI, a lower 6MWD, and a higher risk COMPERA-2.0 risk strata pre-admission. During hospitalization, frailty was associated with a longer LOS and greater reliance on social work consultations independent of age and sex. There was a high prevalence of hospital readmissions in both frail and non-frail PAH patients. Frail patients had a slower recovery in 6MWD in their first 6-months post-hospital discharge.

Frailty was observed in 46% of patients at the time of hospital admission using a 39-point FI, similar to other CLD populations. A study by Wilson et al. observed a prevalence of 45% using a 32-point frailty deficit index in LTx candidates, with prevalence independent of lung disease. ²⁰ In patients with fibrotic ILD, frailty was prevalent in 38-50%.^21,35 Several potential factors contributing to the high prevalence of frailty in CLD include age, psychosocial factors and physical function limitations.^36,37 As in other CLD,^35,38 another factor for frailty in the PAH population may be increased burden of comorbidities, given the moderate correlation between FI and CCI in the present cohort. PAH patients with a greater CCI were more likely to be older and have greater functional limitation with increased dependence on their ADLs, IADLS, and lower 6MWD (Supplemental Table 2).

During the index hospitalization, patients identified as frail by the FI had a longer LOS by a median of 4 days, along with greater requirements for inpatient consultations and home care services. Although frailty has been shown to be associated with longer hospitalizations, ³⁶ PAH may also be associated with a longer stay compared to other CLD admissions given increased complexity of managing several medications such as vasodilators, anticoagulants, and diuretics.^1,39 This may also be compounded in the PAH population by older age, a greater number of comorbidities, and lower functional capacity pre-hospitalization. Thus, a prolonged hospitalization may result in a greater need for interdisciplinary support, longer recovery period, and increased healthcare costs. ⁴⁰

We did not observe any significant differences in hospital readmissions or mortality within 12-months of hospital discharge between frail and non-frail PAH patients measured with FI. This was in contrast with other studies, which have demonstrated higher readmission rates and mortality rates in frail patients with pulmonary hypertension and other chronic lung diseases.^13,41,42 The discrepancy may arise from differences in frailty definitions, cohort characteristics, or follow-up duration. For example, Rauf et al. utilized the Hospital Frailty Risk Score to characterize frailty in PAH patients. ⁴² Compared to patients with other CLDs, PAH patients are often younger, and survival may be more influenced by disease-specific hemodynamic factors. In addition, PAH management advances over the 9-year period (2009–2018) may have also impacted prognosis.

Although frailty was not associated with a greater number of hospital re-admissions within our cohort, there was a high prevalence of readmissions with about two-thirds of PAH patients re-hospitalized within 1-year of discharge. The most common cause of hospital readmission was right-sided heart failure,^3,4,17 an established risk factor for clinical worsening after hospital discharge and often associated with subsequent readmissions. ⁴ Unfortunately, PAH patients are vulnerable to deterioration in the early post-discharge period, often attributed to medication non-adherence and challenges related with fluid management at home. ¹⁷ Patients with PAH may also have increased risk of re-admissions given their vulnerability to thrombosis, arrhythmias, and infections, which may increase risk of subsequent re-admissions.^3,43

Frail patients had a lower exercise capacity pre-admission and at 3–6 months post-discharge compared to non-frail patients. The 6MWD at 12-months post-discharge was not statistically different despite the mean difference of greater than 80 m between the frail and non-frail groups, which is more than the clinically important difference of 30 m^44,45 The impact of frailty on functional capacity may be influenced by factors such as reduced cardiopulmonary reserve, impaired musculoskeletal fitness, and increased comorbidities. ¹⁵ Further, the recovery in exercise capacity post-discharge in PAH can be influenced by other factors such as physical lifestyle, PAH medications, and rehabilitation practices. ⁴⁶ Thus, frailty may further contribute to lower exercise capacity post-hospitalization given the effects of multi-systemic involvement. ³⁶

There were several limitations in the present study. Firstly, our study was conducted in a single transplant and PAH center and had a modest sample size, which may limit the generalizability of our findings to other settings. The FI utilized in this study has not been previously validated in PAH populations, although a similar index has been applied to LTx candidates and ILD patients at our center, as well as other CLD states.^19,20,35,47 Frailty was assessed at time of admission therefore some scores were influenced by the acute severity of illness. However, capturing frailty at this timepoint allowed for the identification patients that were most vulnerable to adverse outcomes during hospitalization and recovery. Further, 26 out of 39 (67%) items in our FI were based on comorbidities and non-laboratory parameters, which would not necessarily be influenced by the acute hospital admission. Given the retrospective study design, some patients did not have data up to 1-year post hospital discharge and thus we were under-powered for assessment of secondary outcomes post-hospital discharge. Patients lost-to-follow up or discharged from PH team care were found to be older and more frail (Supplemental Table 6), which may have impacted post-discharge outcomes. As such, discharge to a long-term care facility may represent an important outcome that should be considered in future studies as a clinically relevant endpoint. Further, the follow-up duration of this study was relatively short, which may not fully capture the impact of frailty on long-term outcomes.

In summary, this study highlights that frail PAH patients had a longer hospital stay and an increased requirement for multidisciplinary healthcare support. Patients with frailty also had lower exercise capacity pre- and post-hospital discharge. Characterizing frailty at the time of hospital admission in PAH patients may help identify a subgroup of patients who may benefit from enhanced multidisciplinary care and additional supports following discharge. The FI is composed of parameters that can be abstracted from routine clinical records, which allows for the concept of a FI to be applied across centers. Emerging studies also demonstrate the use of automated frailty indices derived from electronic health records, which could allow for the frailty assessment to be more feasibly integrated into routine practice. ⁴⁸ Although frailty was not associated with 1-year readmissions or mortality, further study is needed to explore the impact of frailty on longer term outcomes and functional recovery post hospitalization.

ORCID iDs

Nadia Sharif https://orcid.org/0000-0001-9961-515X

Dmitry Rozenberg https://orcid.org/0000-0001-8786-9152