Keywords
frailty - primary pulmonary hypertension - inpatient mortality
Introduction
Primary pulmonary hypertension (PPH), a rare and progressive condition marked by elevated
blood pressure in pulmonary arteries, leads to significant morbidity and mortality.[1] Inactivating mutations in the BMPR2 gene are the most common genetic cause of PPH, leading to dysregulation of pulmonary
vascular remodeling.[2]
Frailty, a clinical syndrome characterized by decreased reserve and resistance to
stressors, is assessed by the hospital frailty risk score (HFRS), a tool designed
to quantify the risk of frailty in hospitalized patients based on the International
Classification of Diseases, Tenth Revision (ICD-10) codes, which helps in identifying
patients at higher risk of poor outcomes.[3]
[4]
[5]
[6] Despite the significance of frailty in PPH outcomes, there is a notable lack of
comprehensive data evaluating the influence of HFRS on inpatient outcomes among PPH
patients.
The association between frailty and adverse outcomes in PPH patients can be attributed
to various factors.[7]
[8]
[9] Frailty is a complex syndrome involving decreased physiological reserve and resistance
to stressors, and in the context of PPH, it may reflect the cumulative impact of cardiovascular
and respiratory impairments, as well as the burden of comorbidities.[10]
[11]
[12]
[13] Additionally, frailty is often accompanied by systemic inflammation, hormone resistance,
and increased muscle protein degradation, which further exacerbate the vulnerability
of PPH patients to adverse events.[4]
[14]
[15]
Methods
ICD-10 is a system to code all medical diagnoses, symptoms, and procedures designed
by the World Health Organization. The system is modified by Centers for Medicare and
Medicaid Services and the National Center for Health Statistics to better fit the
U.S. health system.[6] Hospital billing data include patient demographics and ICD-10 codes. The Agency
for Healthcare Research and Quality (AHRQ) is a federal agency under the Department
of Health and Human Services that standardizes states billing data obtained from different
hospitals in each state to create uniform databases such as Nationwide Readmission
Database (NRD). NRD is a publicly available database designed to generate national
estimates of all-cause and condition-specific inpatient readmissions. Quality measures
are applied to ensure accuracy. The study was deemed exempt by the Institutional Review
Board as the NRD contains deidentifying patients' information.
For this study, we used NRDs from 2016 as we are interested in utilizing ICD-10 codes
only. We did not include data after 2019 to avoid possible bias related to the coronavirus
disease 2019 pandemic. We included all patients aged 18 years and older with a primary
diagnosis (I10_Dx1) of PPH who were admitted from January to November in each year
studied. We calculated HFRS for each case. HFRS is an ICD-10-based score validated
to provide a low-cost systematic way to screen for frailty and predict hospital adverse
outcomes.[14] To simplify the analysis, we characterized patients into two categories based on
HFRS: low frailty score (HFRS <5) and high frailty score (HFRS ≥5), as shown in [Fig. 1]. The latter can be categorized into three classes: HFRS 5 to 10, HFRS 10 to 15,
and HFRS above 15, as shown in [Supplementary Table S1] (online only).
Table 1
Baseline characteristics
|
Variable
|
Total
|
With HFRS >5
|
With HFRS <5
|
p-Value
|
|
Age, mean (SD)
|
56.8 (23.2)
|
60.6 (21.3)
|
53.8 (23.9)
|
<0.0001
|
|
Female gender
|
3,498 (76.8%)
|
1,499 (75.0%)
|
1,999 (78.2%)
|
0.0936
|
|
Chronic lung disease
|
1,629 (35.8%)
|
809 (40.5%)
|
820 (32.1%)
|
<0.0001
|
|
Dementia
|
65 (1.4%)
|
52 (2.6%)
|
13 (0.5%)
|
<0.0001
|
|
Depression
|
684 (15.0%)
|
352 (17.6%)
|
332 (13.0%)
|
0.0019
|
|
Diabetes mellitus
|
1,301 (28.6%)
|
693 (34.7%)
|
608 (23.8%)
|
<0.0001
|
|
Hypertension
|
1,940 (42.6%)
|
1,039 (52.0%)
|
901 (35.3%)
|
<0.0001
|
|
CHF
|
2,995 (65.8%)
|
1,569 (78.5%)
|
1,426 (55.8%)
|
<0.0001
|
|
Malignancy
|
187 (4.1%)
|
125 (6.3%)
|
62 (2.4%)
|
<0.0001
|
|
Obesity
|
1,221 (26.8%)
|
579 (29.0%)
|
642 (25.1%)
|
0.032
|
|
PVD
|
238 (5.2%)
|
135 (6.8%)
|
102 (4.0%)
|
0.0016
|
|
Deficiency anemia
|
1,029 (22.6%)
|
593 (29.7%)
|
435 (17.0%)
|
<0.0001
|
|
Blood loss
|
36 (0.8%)
|
22 (1.1%)
|
13 (0.5%)
|
0.0764
|
|
Coagulopathy
|
814 (17.9%)
|
511 (25.6%)
|
303 (11.9%)
|
<0.0001
|
|
Chronic liver disease
|
633 (13.9%)
|
364 (18.2%)
|
269 (10.5%)
|
<0.0001
|
|
Movement disorder
|
101 (2.2%)
|
58 (2.9%)
|
43 (1.7%)
|
0.0646
|
|
Seizure
|
122 (2.7%)
|
84 (4.2%)
|
38 (1.5%)
|
0.0006
|
|
Encephalopathies
|
143 (3.1%)
|
123 (6.2%)
|
20 (0.8%)
|
<0.0001
|
|
Paralysis
|
48 (1.0%)
|
38 (1.9%)
|
9 (0.4%)
|
0.0009
|
|
Psychosis
|
161 (3.5%)
|
87 (4.4%)
|
73 (2.9%)
|
0.0794
|
|
CKD
|
1,102 (24.2%)
|
806 (40.3%)
|
296 (11.6%)
|
<0.0001
|
|
PUD
|
31 (0.7%)
|
16 (0.8%)
|
15 (0.6%)
|
0.5265
|
|
CBVD (POA)
|
58 (1.3%)
|
39 (2.0%)
|
19 (0.7%)
|
0.0026
|
|
CBVD sequelae
|
31 (0.7%)
|
28 (1.4%)
|
3 (0.1%)
|
<0.0001
|
|
Hospital location
|
|
Central metropolitan
|
1,219 (26.8%)
|
520 (26.0%)
|
698 (27.3%)
|
0.0647
|
|
Fringe metropolitan
|
1,217 (26.7%)
|
596 (29.8%)
|
621 (24.3%)
|
|
Medium metropolitan
|
831 (18.3%)
|
330 (16.5%)
|
501 (19.6%)
|
|
Small metropolitan
|
514 (11.3%)
|
226 (11.3%)
|
289 (11.3%)
|
|
Micropolitan counties
|
435 (9.5%)
|
190 (9.5%)
|
245 (9.6%)
|
|
Other
|
321 (7.0%)
|
129 (6.5%)
|
192 (7.5%)
|
|
Socioeconomic status
|
|
1
|
1,223 (26.9%)
|
505 (25.3%)
|
718 (28.1%)
|
0.1294
|
|
2
|
1,184 (26.0%)
|
505 (25.3%)
|
678 (26.6%)
|
|
3
|
1,142 (25.1%)
|
510 (25.5%)
|
632 (24.7%)
|
|
4
|
958 (21.0%)
|
457 (22.9%)
|
501 (19.6%)
|
Abbreviations: CBVD, cerebrovascular disease; CHF, chronic heart failure; CKD, chronic
kidney disease; HFRS, hospital frailty risk score; POA, present on admission; PUD,
peptic ulcer disease; PVD, peripheral vascular disease; SD, standard deviation.
Fig. 1 Study design. HFRS, hospital frailty risk score; LOS, length of stay.
We examined comorbidities that are related to the risk of inpatient mortality and
readmission, which are included in the AHRQ Elixhauser Comorbidity Score and have
been validated to predict both inpatient mortality and 30-day readmissions.[16] To maintain privacy, ethnicity is not available in NRD datasets.
Complex survey design, weights, and clustering were considered during the analysis
using “Survey” procedures in Statistical Analysis Software (SAS), producing a nationwide
analysis for discharge estimates from almost all hospitals in the United States. Categorical
and continuous variables were reported as percentages and mean ± standard deviation
(SD), respectively. Differences in mean and percentage were assessed using least-squares
means and chi-squared tests, respectively. Logistic regression was used to analyze
the independent impact of frailty on categorical outcomes. The final parsimonious
model included age, gender, encephalopathy, chronic heart failure, chronic liver disease,
and frailty. Statistical significance was considered as p-value < 0.05. All analyses were performed using SAS version 9.4 (SAS Institute Inc.,
SAS 9.4, Cary, North Carolina, United States).
Results
We classified patients included with PPH into two categories to simplify the study:
those at low risk for frailty (HFRS < 5) and those likely to have frailty (HFRS ≥
5). Almost half of the cases (44%) were found to be at risk for frailty (HFRS > 5).
[Table 1] shows that patients at risk for frailty were older, with a mean age of 60 years
(compared with 53 years for patients at low risk for frailty, p-value 0.0001) and had a similar distribution of female sex. The majority of patients
were female (77%).
Patients at risk for frailty were significantly (p-value < 0.05) associated with a higher frequency of chronic comorbidities, including
chronic lung disease, chronic kidney disease, chronic liver disease, chronic heart
failure, hypertension, diabetes mellitus, obesity, acute encephalopathies, dementia,
depression, cerebrovascular disease, malignancies, and malnutrition.
Univariate analysis showed that the odds ratio for inpatient mortality is 7.6 times
higher in patients who are frail (5.1–11.3). The higher the score, the higher the
odds for inpatient mortality ([Fig. 2]). Multivariate analysis ([Fig. 3]) showed that even after adjustment for significant comorbidities, age, and gender,
HFRSs more than 5 were the most important single factor associated with higher inpatient
mortality (adjusted odds ratio: 6 [95% CI: 4–9], p-value 0.0001). [Table 2] shows that cases at risk for frailty (HFRS >5) also had a prolonged hospital stay
(13 vs. 6 days, p-value < 0.0001) and subsequently total hospitalization charges ($212,609 vs. $77,775,
p-value < 0.001). They also had a higher rate of 30-day daily admission (19 vs. 13%,
p-value < 0.001) and a slight increase in inpatient mortality during first readmission
after the index hospitalization (12 vs. 8%, p-value 0.095).
Table 2
Outcomes
|
Variable
|
Total
|
With HFRS >5
|
With HFRS <5
|
p-Value
|
|
Index mortality
|
329 (7.2%)
|
277 (13.9%)
|
52 (2.0%)
|
<.0001
|
|
Index malnutrition
|
459 (10.1%)
|
337 (16.9%)
|
122 (4.8%)
|
<0.0001
|
|
Index LOS, mean (SD)
|
9.0 (16.8)
|
13.3 (21.7)
|
5.6 (9.5)
|
<0.0001
|
|
Total charges mean (SD)
|
136,936 (516,428)
|
212,609 (680,305)
|
77,775 (311,036)
|
<0.0001
|
|
30-d readmission
|
692 (15.2%)
|
372 (18.6%)
|
320 (12.5%)
|
<0.0001
|
|
Readmission mortality
|
70 (10%)
|
45 (12%)
|
25 (8%)
|
0.0953
|
Abbreviations: HFRS, hospital frailty risk score; LOS, length od stay; SD, standard
deviation.
Fig. 2 HFRS and mortality. HFRS, hospital frailty risk score.
Fig. 3 Multivariate analysis. CI, confidence interval; OR, odds ratio.
Discussion
This study aimed to investigate the association between frailty, as measured by the
HFRS, and inpatient outcomes among patients diagnosed with PPH at a national level.
The HFRS has emerged as a valuable tool for identifying frailty in hospitalized patients
using routinely collected administrative data to quantify frailty and predict adverse
outcomes such as prolonged hospital stays, readmissions, and mortality in a large
cohort of older adults.[10]
[14]
In examining the impact of frailty on increased mortality and disease progression,
our findings align with and expand upon the existing literature. Boyd et al highlighted
that frailty significantly increased hospitalization rates and disability progression
among older women, suggesting a similar trajectory of adverse outcomes in frail populations,
which parallels our findings in PPH patients, where higher frailty scores were associated
with increased inpatient mortality and prolonged hospital stays.[5] Wang et al and Guan and Niu further reinforced the notion that frailty exacerbates
disease progression and adverse events in chronic respiratory conditions, supporting
our results that frailty significantly worsens the prognosis for PPH patients.[12]
[17] Furthermore, Dinesh et al explored the relationship between rehabilitation and frailty
in patients with advanced heart or lung disease, demonstrating that frailty is a significant
predictor of poor surgical outcomes, underscoring the necessity for targeted rehabilitation
strategies to mitigate these risks.[18]
Resource utilization in the hospital setting is markedly influenced by frailty, as
demonstrated in our study and corroborated by the existing literature. Bernabeu-Mora
et al showed that frailty was a predictive factor for 90-day readmissions following
hospitalization for chronic obstructive pulmonary disease exacerbations, a finding
that aligns with our observation of higher readmission rates among frail PPH patients.[4] Hadaya et al examined the impact of frailty on clinical outcomes and resource use
following emergency general surgery, reporting increased hospital costs, longer stays,
and higher readmission rates for frail patients, which parallels our findings of significantly
higher hospitalization charges and longer stays for frail PPH patients.[9]
Our findings are consistent with other studies, such as Makary et al, which showed
that frailty predicts increased health care resource use, including longer hospital
stays and higher costs.[8]
Recent studies in older populations indicate that frailty may be reversible with targeted
exercise and nutritional interventions, although evidence supporting nutritional strategies
remains limited.[19]
[20] Overall, these studies collectively highlight the substantial health care burden
imposed by frailty, reinforcing the necessity for tailored management strategies to
mitigate resource utilization and improve outcomes in frail PPH patients.
The strengths of our study include the use of a large national database, which allowed
for a comprehensive analysis and broad applicability of our findings. However, the
study also has several limitations. First, the retrospective nature of the study limits
our ability to establish causality between frailty and inpatient outcomes. Additionally,
the lack of detailed clinical data (e.g., laboratory results and imaging) may have
influenced our findings. Another limitation of the study is that the HFRS was not
developed for the NRD (or any other U.S. database), which may underestimate frailty
due to limited outpatient data integration. The score also was developed for patients
above the age of 75 years, raising questions about its reliability in younger populations,
such as those with PPH. Additionally, PPH patients often undergo comprehensive evaluations,
which may reduce the utility of the HFRS in this context. Prospective studies are
needed to validate our findings, establish causative relationships, and explore the
efficacy of various interventions, including nutritional support and exercise programs,
in improving outcomes for all frail PPH patients.
Conclusion
In conclusion, our study demonstrates that frailty assessment, using HFRS, can be
easily incorporated into clinical practice. It is a significant predictor of adverse
inpatient outcomes in patients with PPH. Frailty is associated with higher hospital
mortality, increased 30-day readmission rates, and prolonged hospital stays. Further
research is necessary to confirm our findings and to develop and test interventions
aimed at reducing the impact of frailty in PPH patients.
