To develop a scale of domains associated with the health-related quality-of-life (HRQOL) in patients with cirrhosis-related ascites.

We initially undertook literature searches and a qualitative study in order to design a cirrhosis-associated ascites symptom (CAS) scale describing symptoms with a potential detrimental impact on health related quality of life (HRQL) (the higher the score, the worse the symptoms). Discriminatory validity was assessed in a validation cohort including cirrhotic patients with (1) tense/severe; (2) moderate/mild; or (3) no ascites (controls). Patients also completed chronic liver disease questionnaire (CLDQ) and the EuroQoL 5-Dimensions 5-Level (EQ-5D-5L) questionnaire evaluating HRQL. The relation between scale scores was analysed using Spearman correlations.

The final CAS scale included 14 items. The equivalent reliability was high (Chronbach’s alpha 0.88). The validation cohort included 103 patients (72% men, mean age 62.4 years). The mean scores for each question in the CAS scale were higher for patients with severe/tense ascites than for mild/moderate ascites and controls. Compared with controls (mean = 9.9 points), the total CAS scale score was higher for severe/tense ascites (mean = 23.8 points) as well as moderate/mild ascites (mean = 18.6 points) (P < 0.001 both groups). We found a strong correlation between the total CAS and CLDQ score (rho = 0.82, P < 0.001) and a moderate correlation between the CAS and the EQ-5D-5L score (0.67, P < 0.001).

The CAS is a valid tool, which reflects HRQOL in patients with ascites.

Ascites is a serious complication to cirrhosis[1,2]. Without liver transplantation, only about 55% of patients with cirrhosis and ascites are alive, five years after the initial diagnosis. In addition, the severity of symptoms associated with ascites may have a detrimental impact on health-related quality of life (HRQL). Five previous studies have evaluated the HRQL in patients with cirrhosis using generic questionnaires including the Medical Outcome Study Short Form 36 (SF-36), the Nottingham Health Profile and the Sickness Impact Profile[3-7]. All studies found that cirrhosis is associated with a detrimental effect on HRQL. None of the studies were designed to evaluate the association between ascites severity and the quality of life. However, one study found that disease progression, including development of ascites, is associated with an impaired HRQL[7]. Disease-specific questionnaires developed for the assessment of HRQL in patients with liver disease and cirrhosis have reached similar conclusions[8,9]. The questionnaires include aspects that are of particular relevance to patients with cirrhosis, such as concerns about complications and liver transplantation. The 29 item chronic liver disease questionnaire (CLDQ) was used to evaluate HRQL in 204 patients representing all stages of various liver diseases[10]. The scale correlates with the severity of the underlying liver disease and is also associated with active medical and psychiatric comorbidity. Liver failure, development of minimal hepatic encephalopathy, malnutrition, abdominal pain, fatigue, and anxiety may also affect HRQL[11,12].

None of previous studies specifically evaluated the association between the severity of ascites related symptoms and HRQL. However, it is likely that the severity of symptoms is important.

In clinical trials of medical interventions, HRQL as well as symptom management are important parameters in the assessment of intervention benefits and harms[4,5,13,14]. A study including 212 outpatients with cirrhosis found that the management of cirrhosis-related complications is associated with improved HRQOL assessed using the Medical Outcomes Study Form, SF-36 and CLDQ questionnaires. The study found that ascites as well as a decrease in haemoglobin and previous hepatic encephalopathy predicted the HRQOL[15].

We therefore developed and evaluated the predictive ability of a scale specifically made to assess symptoms related to cirhosis-associated ascites. We subsequently undertook a multicenter cohort study to evaluate the association between our scale and a disease specific scale (CLDQ) as well as a generic HRQL questionnaire (EQ-5D-5L).

The present study was conducted as a multi-center study, with participation of three university hospitals in Denmark: Copenhagen University Hospital Hvidovre, Odense University Hospital and Aarhus University Hospital. The study was approved by the Danish Data protection Agency, journal no: 04054, ID: AHH-2015-075.

The development of the cirrhosis-associated ascites symptom (CAS) scale included a literature review and a qualitative study to evaluate face and content validity as well as discriminatory ability.

We initially searched for previously validated scales in MEDLINE and EMBASE (‘ascites’ AND (‘quality of life [Mesh]’ OR ‘life quality’ OR ‘health-related quality of life’ OR ‘health related quality of life’ OR ‘symptoms’) AND (‘cirrhosis’ OR ‘end-stage liver disease’). The searches were combined with manual searches of reference lists in potentially relevant articles and conference proceedings. None of our searches identified questionnaires evaluating symptoms associated with ascites.

We then proceeded with a qualitative study. Initially, the study group and ten independent experts identified and listed what they believed were the most important symptoms with an expected negative impact on HRQL (six hepatologists and four nurses with clinical experience in the management of patients with ascites from Gastro Unit, medical division, Copenhagen University Hospital Hvidovre, were interviewed). In our selection of symptoms (domains), we decided that they should occur frequently or be severe enough to have substantial impact on daily function in order to be considered for our scale.

After the selection of domains, we first conducted open interviews followed by structured interviews of ten patients with cirrhosis and moderate (n = 4) or severe ascites (n = 6). The interviews took place in a quiet room in the hospital ward. The selected domains were initially presented and patients were asked to rate them (Supplementary Table 1) according to the patient’s perceived ‘importance’, on the extent to which the symptom had bothered them. Each patient was then asked to complete the preliminary questionnaire themselves and to answer a debriefing form with the questions: (1) How long did it take to complete the questionnaire? (2) Did you need assistance from others to fill out the questionnaire? (3) Did you find any items confusing or difficult to respond to (and if so, list the items)? (4) Did you find any items upsetting? (5) Were any important items missing? and (6) Do you have any additional comments?

We revised the questionnaire instrument based on the answers. One question, which patients found upsetting and two questions, which patients found to be irrelevant or unimportant were removed. Revisions were decided by consensus among investigators.

The final scale included 14 items (Table 1). Chronbach’s alpha was 0.88 for the total score, which we considered as acceptable. The validation cohort included 103 patients.

Demographic data are stated in Table 2. Seventy-four were male (72%) and mean age was 62.4 years (range 38-85). The proportion of patients with Child Pugh A was 24%, Child Pugh B: 58% and Child Pugh C: 19%. Forty-four percent had severe/tense ascites and 27% had mild/moderate ascites. A control group of 31 patients (30%) had no ascites.

Discriminant validity

The mean scores for each question in the CAS scale suggested that symptoms were worse for patients with severe/tense ascites than for controls (Figure 1).

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Mean cirrhosis-associated ascites symptom scores by question number and ascites severity.

Symptoms also appeared worse for patients with mild/moderate ascites. Accordingly, the CAS scale score found that patients with severe/tense ascites or moderate/mild ascites had significantly worse scores compared with controls (P < 0.001 and P < 0.001; Figure 2).

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Median test scores with SEM divided by control, mild/moderate and tense ascites respectively. A: CAS scale. Control vs mild/moderate, P < 0.001; control vs tense, P < 0.001; B: CLDQ. Control vs mild/moderate P = 0.002; Control vs tense, P < 0.001; C: EQ-5D-5L. Control vs mild/moderate, P = 0.038; Control vs tense, P = 0.002. CAS: Cirrhosis-associated ascites symptom scale; CLDQ: Chronic liver disease questionnaire; EQ-5D-5L: EuroQoL 5-Dimensions 5-Level scale questionnaire.

As expected, patients with severe/tense ascites had the worst symptoms scores. Based on the CLDQ questionnaire, tense/severe ascites had a detrimental impact on HRQOL compared with controls (P < 0.001) as did moderate/mild ascites (P = 0.002). The EQ-5D-5L also found a lower HRQOL in patients with severe/tense ascites (P = 0.002) or moderate/mild ascites (P = 0.038).

We found that the developed CAS scale is effective in discriminating between various severities of ascites. It takes five minutes to perform and is easily implemented in daily practice. Moreover the CAS scale correlates strongly with the well-developed liver disease-specific HRQL instrument, CLDQ[9]. This suggests that our symptom assessment scale reflects the perceived HRQL of patients with cirrhosis and ascites. The CLDQ subparameters: fatigue, activity and systemic symptoms were likewise strongly correlated with the CAS scale suggesting that these domains are influential symptoms or functional limitations for patient with ascites. Surprisingly, the subparameter abdominal symptoms was not associated with the CAS scale. This may be due to use of words or inaccuracy of the questions, and will need further investigations for clarification. The total CLDQ was better correlated with CAS than each subparameter. This may be due to the impact of low quality of life affecting all aspects of life, allowing a total low score higher statistical significance. The CAS versus EQ-5D-5L correlation was moderate. The EQ-5D-5L is a generic scale which aims at examining non-disease specific HRQL[16,18,19], and specific symptoms related to cirrhosis and ascites may be undetected by this questionnaire. On the contrary, the CLDQ is created for patients with chronic liver disease, and it includes questions concerning ascites related symptoms and gains a stronger correlation with the CAS score.

To our knowledge, no previous study has developed and validated a questionnaire specifically aimed at evaluating HRQL in patients with ascites. However, previous literature describes correlations between disease severity in chronic liver disease and HRQL[7,12,20]. One study found that ascites had significant effects on self-related disease progression in cirrhosis[7]. Another showed that ascites was not a predictor for HRQL[12] and two studies found that management of ascites may improve HRQL[15,21]. In a study evaluating the impact of the Medical Outcomes Study Short Form (SF-36) on HRQL in 523 patients with cirrhosis and ascites, a strong association between the physical component score (PCS) and leg edema, history of previous hepatic encephalopathy, severe ascites and low serum sodium levels were found[22]. The PCS includes questions on physical role, physical functioning, and body pain and general health. Also, a lower mental component score in the SF-36 was associated with low levels of serum sodium and treatment with lactulose. Hence, several factors influence HRQL when measured by SF36 and the impact of ascites was not clearly established by this study.

The CAS score comprises only 14 questions with high compliance to answering, takes only five minutes to complete and is readily implemented in daily clinic. The present study supports the results implying correlation between ascites severity and HRQL assuming that the CAS scale is valid and reliable. Progression in liver disease (e.g., measured by Child Pugh score which includes presence of ascites) is associated with ascites severity grade. This argues that our results support previous literature finding a correlation between liver disease severity and HRQL.

Our study has some considerable weaknesses. The relatively small sample size is one of the main limitations. Moreover, the majority of patients had cirrhosis due to alcohol. Dependency of alcohol, concomitant diseases related to alcohol use or cirrhosis that may impact physical health was not assessed in our cohort. Malnutrition is a serious complication to cirrhosis with impact on HRQL[23]. Also, impact of socioeconomic factors may influence quality of life substantially[24,25]. Future studies may address the synergy between various factors affecting quality of life in cirrhosis and ascites. The CAS scale is a specific scale evaluating the impact of symptoms related to ascites. When assessing overall quality of life on chronic liver disease or the impact of multi-morbidity in cirrhosis; validated and recommended scores such as SF36 or CLDQ should be used, with the CAS scale as a supplement[9]. Consequently, CAS scale validation also requires further research on larger sample size and applied on standardized groups of subjects. An interesting study setup would be to test the CAS scale on patients with cirrhosis presented with ascites before and after treatment. Further, the CAS scale ought to be tested as a monitoring tool for efficacy in intervention trials, compared with CLDQ and SF-36 questionnaires.

In conclusion, the present construction and validation of the CAS scale seems promising for future research in the area of ascites symptom management and HRQL. Initially the scale requires testing in other study setups in order to clarify the quality and practicability of the questionnaire. In future perspectives an optimized and/or further validated version of the CAS scale can be used as a monitoring tool in interventional trials and to assess the effects of standardized treatment. Ultimately the CAS scale can contribute to gaining further knowledge on how to treat ascites effectively and consequently improve patients’ quality of life.

The authors would like to thank the health care professionals at the Gastro Unit, Copenhagen University Hospital Hvidovre, who participated in the qualitative study, thereby contributing to the design of the CAS scale. We also wish to thank medical professor Per Bech and senior doctor Finn Zierau for valuable feedback on statistical issues.