Frailty is strongly associated with self-reported symptom burden among patients with cirrhosis

Abstract

Objectives:

Although patients with cirrhosis often experience debilitating symptoms, few are referred for palliative care. Frailty is increasingly incorporated in liver transplantation evaluation and has been associated with symptom burden in other populations. We hypothesized that frail patients with cirrhosis are highly symptomatic and thus are likely to benefit from palliative care.

Methods:

Patients with cirrhosis undergoing outpatient liver transplantation evaluation completed the Liver Frailty Index (LFI; grip strength, chair stands, and balance) and a composite of validated measures including the Edmonton Symptom Assessment Scale (ESAS), distress and quality of life (QOL) measures.

Results:

Of 233 patients (median age 61 years, 43% female), 22% were robust, 59% pre-frail and 19% frail. Overall, 38% of patients reported ≥1 severe symptoms based on preestablished ESAS criteria. Higher frailty categories were associated with increased prevalence of pain, dyspnea, fatigue, nausea, poor appetite, drowsiness, depression and poor wellbeing (test for trend, all p<0.05). Frail patients were also more likely to report psychological distress and poor QOL (all p<0.01). In univariate analysis, each 0.5 increase in LFI was associated with 44% increased odds of experiencing ≥1 severe symptoms (95% CI: 1.2–1.7, p<0.001), which persisted (OR 1.3, 95% CI: 1.0–1.6, p=0.004) even after adjusting for Model for End-Stage Liver Disease-Sodium, ascites, hepatic encephalopathy, and age.

Conclusion:

In patients with cirrhosis, frailty is strongly associated with physical/psychological symptoms including pain and depression, and poor QOL. Frail patients with cirrhosis may benefit from palliative care co-management to address symptoms and improve QOL.

INTRODUCTION

The clinical course of patients with end stage liver disease (ESLD) is characterized by progressive decline and episodic exacerbations that can result in rapid deterioration and death. Patients often experience debilitating symptoms, including pain, anorexia, and depression, as well as poor quality of life (QOL) [1–4]. Unfortunately, many patients with ESLD are not eligible for liver transplantation or die while on the waiting list. Although emerging data suggests that palliative care can improve symptoms and QOL in this population, few patients are currently referred to palliative care [5–8]. Therefore, we need establish systems to routinely screen and identify patients who are appropriate for palliative care.

Frailty is increasingly recognized to have an important impact on patient outcomes and has been associated with increased hospitalizations, waitlist mortality, and physical function after transplant in patients with cirrhosis [9–11]. Although the relationship between frailty and symptoms has not been established in patients with cirrhosis, studies in other populations suggest that frailty is not only associated with increased symptom burden, but that the relationship may be bidirectional [12–14]. We hypothesized that frailty would be associated with patient-reported symptoms and thus may help to identify patients likely to benefit from palliative care co-management.

The objectives of this study were to 1) assess the symptom burden, distress, and QOL in patients with cirrhosis and 2) investigate the correlation of these symptoms with frailty.

PATIENTS AND METHODS

Study population

Adult patients with cirrhosis undergoing outpatient evaluation for liver transplantation from 7/1/19 – 9/30/19 were eligible for inclusion. Patients underwent informed consent, and the study was approved by our institutional ethics review board (IRB #11–07513). Patients were excluded if they had severe hepatic encephalopathy, as defined as a Numbers Connection Test time >120 seconds (given concerns regarding ability to comply with performance-based tests of frailty), and if they did not speak English or Spanish given the lack of availability of the symptom assessment instruments in other languages.

Study measures

Frailty was assessed using the Liver Frailty Index (LFI), a validated measure of physical function in patients with cirrhosis [15]. Patients were instructed by trained study personal to complete three performance-based tests: grip strength, chair stands, and balance. The LFI was calculated using the online calculator: https://liverfrailtyindex.ucsf.edu. LFI scores were then stratified by frailty category using previously established LFI cut-points: robust (<3.2), pre-frail (3.2–4.3), and frail (≥4.4) [15,16].

To characterize symptom burden, patients completed the Palliative Care Quality Network Symptom and Well-being Survey, which is a composite of validated measures including the Edmonton Symptom Assessment Scale (ESAS), National Comprehensive Cancer Network Distress Thermometer, and single-item spiritual distress and QOL measures [17]. The ESAS is a 9-item self-report scale commonly used to assess symptoms in palliative care [18–20]. For each symptom, patients rate how they are feeling now from 0 (no symptom) to 10 (worst possible symptom) on a 11-point numeric rating scale. In the ESAS, a moderate symptom is defined as score 4–6 and a severe symptom as score 7–10 [21,22]. A symptom score ≥4 (moderate to severe) is recommended as a trigger for more comprehensive assessment in the clinical setting [21]. The presence of one or more severe symptoms (ESAS score ≥ 7) has been associated with patient-reported high symptom burden and decreased physical, emotional, and social functioning [23]. ESAS can be further divided into a physical symptom subscale (pain, fatigue, drowsiness, nausea, poor appetite, and dyspnea) and psychological symptom subscale (depression and anxiety) [20].

The National Comprehensive Cancer Network Distress thermometer is a tool frequently used in cancer populations to screen for psychological distress [24]. Patients rate their current level of distress on a numeric rating scale ranging from 0 (no distress) to 10 (extreme distress), with a positive screen defined as score ≥ 4 [25]. Spiritual distress was assessed using a one-item descriptive rating scale: “Are you at peace?” [26] Lack of spiritual distress was defined by selections of “not at all” or “a little bit,” while spiritual distress was considered “a moderate amount,” quite a bit,” or “completely.” Finally, patients completed a single global QOL item [27,28]. Responses were categorized in three tiers: 1) Poor or very poor, 2) Fair, 3) Good or excellent.

Medical records were reviewed to collect demographic information and clinical data, including etiology of liver disease, history of hepatocellular carcinoma, and presence of ascites. Hepatic encephalopathy was categorized as present if ≥60 seconds were needed to complete the Numbers Connection Test A [29]. Model for End-Stage Liver Disease-sodium (MELDNa) was calculated from labs collected within 6 months of study visit.

Statistical Analysis

Data were summarized by medians with interquartile ranges (IQR) for continuous variables or numbers with percentages for categorical variables. Differences in demographic and clinical characteristics were stratified by severe symptoms (one or more ESAS score ≥7) using chi-square tests and Fisher exact tests (<5 events) for categorical variables and Mann Whitney U tests for non-parametric continuous variables. We compared differences in ESAS symptom scores, spiritual and psychological distress, and QOL by frailty categories using test for trend.

A multivariate model was constructed to identify variables that were independently associated with reporting ≥1 severe symptoms. Bivariate analyses were performed on variables identified a priori as potentially associated with severe symptoms: LFI, age, MELDNa, ascites, and hepatic encephalopathy. These variables, regardless of p-value in the bivariate analyses, were then entered in a forward stepwise regression model in order of increasing significance. Subgroup analyses were performed in patients with and without ascites or hepatic encephalopathy, and low and high MELDNa as defined by the 75th percentile score.

An alpha of <0.05 was used to determine statistical significance. All analyses were conducted using SPSS (Version 23.0. Armonk, NY: IBM Corp).

RESULTS

Baseline Patient Characteristics

Of 233 patients enrolled in the study, the median age was 61 (IQR 54–65), 99 (43%) were female, 42 (18%) were non-white, and 191 (82%) were listed for transplant. The median MELDNa was 15 (IQR 11–19). The median LFI was 3.8 (IQR 3.3–4.2); according to LFI cut-offs, 52 (22%) were robust, 138 (59%) pre-frail, and 43 (19%) frail (Table 1).

Table 1.

Baseline characteristics of study sample stratified by symptom burden

	N = 233 Median (IQR) or n (%)	Mild/moderate symptoms	Severe symptom(s)†	p-value
Age, years	61 (54–65)	62 (54–66)	60 (55–65)	0.52
Female	99 (43)	54 (38)	45 (50)	0.09
Race/ethnicity
White	191 (82)	114 (80)	77 (85)
Black	10 (4)	6 (4)	1 (1)
Hispanic	3 (1)	2 (1)	1 (1)
Asian	25 (11)	19 (13)	6 (7)
Native American	3 (1)	1 (1)	2 (2)
Other	1 (0.4)	0 (0)	1 (0.4)	0.41
BMI, kg/m2	29 (25–33)	28 (25–33)	29 (25–33)	0.34
Primary etiology of cirrhosis
Hepatitis C	75 (32)	44 (31)	31 (34)
Alcohol	63 (27)	30 (28)	24 (26)
NASH	50 (22)	27 (19)	23 (25)
Cholestatic	17 (7)	13 (9)	4 (4)
Hepatitis B	7 (3)	6 (4)	1 (1)
Other	21 (9)	13 (9)	8 (9)	0.45
Dialysis	10 (4)	5 (4)	5 (6)	0.47
Ascites	37 (16)	16 (11)	21 (23)	0.02
Hepatic encephalopathy	43 (19)	21 (15)	22 (24)	0.07
HCC	77 (33)	52 (37)	25 (28)	0.15
Listed for transplant	191 (82)	116 (82)	75 (82)	0.89
MELDNa	15 (11–19)	14 (11–19)	18 (14–20)	<0.001
Liver Frailty Index	3.76 (3.26–4.23)	3.61 (3.01–4.05)	4.00 (3.41–4.49)	<0.001
Robust	52 (22)	41 (29)	11 (12)
Pre-frail	138 (59)	86 (61)	52 (57)
Frail	43 (19)	15 (11)	28 (31)	<0.001

^†Severe symptom(s) = 1 or more Edmonton Symptom Assessment Scale score ≥7

BMI=body mass index, HCC = hepatocellular carcinoma, IQR=interquartile range, MELDNa=Model for End-Stage Liver Disease–sodium, NASH=Non-alcoholic steatohepatitis

Symptom Burden

The prevalence of moderate to severe symptoms were as follows: fatigue (47%), poor well-being (42%), pain (35%), poor appetite (35%), drowsiness (30%), anxiety (23%), dyspnea (20%), depression (17%), and nausea (11%). Overall, 91 (39%) of patients reported one or more severe symptoms. Likelihood of having severe symptom(s) did not differ by demographic factors such as age, gender, race or by whether patients were listed for transplant (all p>0.09) (Table 1).

Higher frailty categories were significantly associated with increased prevalence of moderate to severe pain, dyspnea, fatigue, nausea, poor appetite, drowsiness, depression, and poor well-being (Figure 1). Additionally, frailty was associated with higher scores on the ESAS physical symptom subscale (IRR [Incidence rate ratio] 1.5, 95% confidence interval [CI] 1.3–1.9, p<0.001) as well as the psychological symptom subscale (IRR 1.3, 95% CI 1.0–1.7, p<0.03). Overall, frail patients were more likely to report one or more severe symptoms (robust: 21% vs. pre-frail: 38% vs. frail: 65%, p<0.001), as well as increased mean (standard deviation) total number of severe symptoms (robust: 0.5 [1.5] vs. pre-frail 0.8 [1.3] vs. 2.0 [2.2], p<0.001).

Figure 1.

Moderate to severe symptoms by frailty status

In univariate analysis, each 0.5 unit increase in LFI was associated with 44% increased odds of experiencing one or more severe symptom(s) (95% CI: 1.2–1.7, p<0.001). After adjusting for MELDNa, ascites, HE, and age, LFI was the only variable significantly associated with the severe symptom group (odds ratio [OR] 1.3, 95% CI: 1.1–1.6, p=0.004) (Table 2). In subgroup analyses, LFI was no longer associated with severe symptom(s) in patients with ascites (OR 1.4, 95% CI 0.8–2.4, p=0.29) and hepatic encephalopathy (OR 1.0, 95% CI 0.6–1.7), p=0.98), but remained significant in patients without ascites (OR 1.3, 95% CI 1.1–1.6, p=0.006) and hepatic encephalopathy (OR 1.4, 95% CI 1.2–1.8, p=0.001) and in patients with high MELDNa (OR 1.8, 95% CI 1.2–2.7, p=0.009) and low MELDNa (OR 1.3, 95% CI 1.0–1.6, p=0.02) (Supplementary Tables 1–2).

Table 2.

Forward stepwise regression of factors associated with severe symptom(s)

	Step 1	Step 2	Step 3	Step 4	Step 5
	OR (95%CI)	OR (95%CI)	OR (95%CI)	OR (95%CI)	OR (95%CI)
Liver Frailty Index †	1.4 (1.2–1.7)	1.4 (1.1–1.7)	1.3 (1.1–1.6)	1.3 (1.1–1.6)	1.3 (1.1–1.6)
MELDNa		1.1 (1.0–1.1)	1.1 (1.0–1.1)	1.1 (1.0–1.1)	1.1 (1.0–1.1)
Ascites			1.6 (0.7–3.4)	1.5 (0.7–3.2)	1.5 (0.7–3.3)
Hepatic encephalopathy				1.4 (0.7–2.9)	1.4 (0.7–2.9)
Age					1.0 (1.0–1.0)

^†per 0.5-unit increase

MELDNa=Model for End-Stage Liver Disease–sodium, CI=confidence ratio, OR= Odds ratio

Distress and Quality of Life

Psychological distress was present in 98 (42%) of patients, and spiritual distress was present in 29 (12%) of patients. In terms of QOL, 153 (66%) reported good or excellent, 42 (25%) reported fair, and 22 (10%) reported poor or very poor. Frail patients were significantly more likely to report psychological distress and poor QOL, and there was a trend for increased spiritual distress (Figure 2). Patients who were listed for transplant compared with those who were not listed did not differ by self-reported psychological distress, spiritual distress, or QOL (all p>0.07).

Figure 2.

Distress and quality of life by frailty status

DISCUSSION

ESLD causes tremendous suffering, and when cure is only possible for a minority of patients, the goal of medical therapy should focus on aggressively managing symptoms, improving QOL, and optimizing function. There is emerging evidence showing that early palliative care decreases symptom burden, improves mood, and reduces resource utilization in patients with cirrhosis [5,30]. Unfortunately, outpatient palliative care is utilized infrequently and often too late, and no evidence-based guidelines yet exist to guide clinicians (e.g. hepatologists, primary care providers) in identifying patients with cirrhosis who are likely to benefit from palliative care [7,8]. Our study, demonstrating a strong association between physical frailty and symptom burden, raises the possibility that frailty may be considered an indicator for palliative care referral among patients with cirrhosis.

More than one-third of our study sample reported one or more severe symptoms. Similar to past studies, the most commonly reported symptoms were fatigue, poor well-being, pain, and poor appetite [31,32]. The high prevalence of pain in our study parallels prior research showing that rates of pain in patients with ESLD is comparable to that among patients with lung and colon cancer [3]. Notably, experiencing one or more severe symptoms is clinically meaningful as this group of highly symptomatic patients is likely to report significant impairments in physical, emotional, and social functioning [22].

Importantly, we found that symptom burden is strongly associated with frailty. Frail patients were more likely to report both physical and psychological symptoms as well as distress and poor QOL. Symptom burden increased in a “dose-dependent” fashion with worsening frailty. Compared to robust patients, frail patients had 2–3 times the rates of moderate to severe symptoms and increased total number of severe symptoms. Furthermore, frailty was associated with the severe symptom group as determined by ESAS score independent of MELDNa, ascites, hepatic encephalopathy, and age. In fact, MELDNa was not a good predictor of symptom burden, confirming prior reports [33–35].

Given the potential for portal hypertensive complications to drive the relationship between frailty and symptom burden in patients with cirrhosis, we conducted subgroup analyses to further explore our main research question. In subgroup analyses, we found that frailty was not associated with ESAS score among the subgroups of patients with ascites or hepatic encephalopathy. This is, perhaps, not unexpected, as these portal hypertensive complications likely drive the symptomatology experienced by these patients; patients with known ascites or hepatic encephalopathy are easily identified in clinical practice and may benefit from referral to palliative care based on these complications alone. However, we found that frailty remained strongly associated with symptom burden among those without ascites or hepatic encephalopathy. Assessment of frailty may add significant value to clinical practice to help identify patients in greatest need of palliative care who do not have obvious signs of clinical portal hypertension.

Overall, our findings suggest a patient’s global health and functional status is strongly correlated with symptom burden and QOL. Thus, frailty, which is increasingly utilized to predict waitlist mortality and post-transplant outcomes, is a good surrogate for symptom burden and should be considered as a trigger for palliative care referral in patients with cirrhosis [10–12]. Multiple frailty instruments have been established in the geriatric population, but the LFI is currently the only frailty assessment tool developed specifically for patients with cirrhosis. The LFI is a simple, objective metric that can be administered in approximately 90 seconds in the clinic setting [36]. The LFI should not replace asking patients about symptoms. Instead, it efficiently serves dual purposes as a functional assessment tool with prognostic ability and a symptom screening tool to trigger further evaluation in high risk patients.

In addition, studies in other populations suggest that the relationship between frailty and symptom burden may be bidirectional [13,14, 37–40]. While no data currently exists, symptom management theoretically may improve functional status, whereas rehabilitating frail patients may potentially reduce symptoms. As next steps, we plan to test a system of referring patients to outpatient palliative care based on frailty status and evaluate the role of palliative care in addressing symptoms and frailty.

We acknowledge several limitations of our study. The generalizability of findings from this single center study is unknown. Patients who were ineligible for transplant or delisted were not included and may have had even higher symptom burden. These patients should also be referred to palliative care.

In conclusion, our study is the first to demonstrate a strong association between frailty and symptom burden in patients with cirrhosis awaiting liver transplantation. Frailty should be considered an indicator for detailed symptom assessment and consideration of palliative care referral. Our findings are a promising step towards identifying patients with cirrhosis who are most likely to benefit from palliative care co-management to address symptoms and improve quality of life.