Potential conflict of interest: Dr. Kumagi received salary support from an unrestricted educational grant from Axcan Pharma. Dr. Hirschfield is a consultant for and is on the speakers' bureau of Axcan Pharma.
Quality of life is an important concern for patients with chronic liver disease. We sought to describe the frequency, severity, and associations of fatigue, in patients with primary biliary cirrhosis (PBC). We performed association testing between PBC-40 multidomain disease-specific quality of life responses and clinical findings. Three hundred twenty-seven patients from a single clinic with PBC (94% female, 92% AMA-positive) were evaluated. The average age was 57 years and average disease duration 7.2 years. Verbally reported fatigue was noted in 48% but present in the overwhelming majority on PBC-40 completion, with 44% having moderate or severe symptoms. Of those not complaining of fatigue clinically, 25% documented moderate or severe fatigue by questionnaire. Age had an inverse relationship with fatigue (P < 0.01), whereas body mass index (BMI) was positively associated (P < 0.01), as was the presence of pruritus (P < 0.001), sicca symptoms (P < 0.001), depression (P < 0.001), fibromyalgia (P < 0.004), and scleroderma (P < 0.05). For those with varices (P < 0.05) or cirrhosis clinically (P < 0.05), higher fatigue scores were noted, although those who initially presented with noncirrhotic disease had higher scores at the time of testing (P < 0.005). Fatigue was associated with greater use of prescription medication (P < 0.01), in particular for antipruritics (cholestyramine: P < 0.001; rifampin: P < 0.001), proton pump inhibitors (P < 0.002), beta-blockers (P < 0.02), and antidepressants (P < 0.001), whereas those taking calcium and vitamin D appeared less fatigued (P < 0.05). In a multivariate model, calcium and vitamin D use, BMI, stage of disease at diagnosis, as well as symptomatic fatigue or pruritus, were significant. Biochemical response to UDCA was not associated with lower fatigue scores. Conclusion: Attempts at defining the biological basis of fatigue in patients with PBC, and improving its treatment, must account for its multifactoral causes. (HEPATOLOGY 2010)
Primary biliary cirrhosis (PBC) is a chronic autoimmune liver disease commonly seen in middle-aged women, characterized by the presence of cholestasis secondary to nonsuppurative destructive cholangitis.1 In addition to the potential for liver-related morbidity and mortality, it is recognized that patients with PBC frequently suffer from a marked impairment in their quality of life (QOL).2 Fatigue has been identified as one of the principal factors contributing to this functional impairment across most studies of patients with PBC, and this potentially disabling symptom is reported to significantly affect a variable minority of patients.3-8
Given such a high prevalence for fatigue in patients with PBC, some have suggested that this symptom is specific and should be recognized as a component of the disease itself.9 There does not appear to be a relationship between symptom severity and liver disease activity, and others have questioned the direct association between PBC and fatigue.10-12 Notably the symptom complex is also a feature of other cholestatic13 and noncholestatic liver disease.14 The pathophysiology underpinning this symptom remains poorly appreciated, and effective interventions are absent. Although there are studies to suggest both a central nervous system and a peripheral motor system contribution,15-17 these mainly derive from animal models that bare little resemblance to human pathological conditions. Other more recent data also suggest that abnormalities in sleep and autonomic dysfunction may be significant contributors to fatigue,18-23 validation of these findings by independent research groups is needed. As expected for a disease in which the average age at diagnosis is older than 50 years, there is a high prevalence of co-morbidities in patients with PBC.24 To that end, there is justification in clarifying the role played by extrahepatic factors in fatigue severity, because these need to be accounted for in any biological models of disease, or treatment studies. One prior study12 evaluated fatigue in 49 Italian patients with PBC and found co-morbidities (38% of patients) were independently associated with higher fatigue scores. Depressed patients (30%) were also more fatigued than patients without depression.
Existing generic tools used to quantify fatigue are derived from other, often nonhepatic, chronic diseases,25 and have not been validated for use in PBC as such. PBC-40 is a multidomain, quality-of-life (QOL) measure developed and validated specifically for patients with PBC.26 The domains within the questionnaire allow quantification of disease-related factors that contribute collectively to the overall quality of life in patients with PBC. We set out to use this validated questionnaire to describe the frequency, severity, and associations of fatigue in a very large well-defined cohort of Canadian patients with PBC. Furthermore, we evaluated external factors that may relate to the presence of fatigue.
Patients and Methods
Patient Population and Data Collection.
All patients with PBC attending clinic between January 2007 and November 2008 were asked to complete the PBC-40 questionnaire. A retrospective chart review was then performed during which demographic, clinical, and laboratory data were collected and tabulated in a study database. This retrospective chart review was approved by the University Health Network Research Ethics Board.
All patients had a diagnosis of PBC made according to traditional criteria,1 with documentation of prior normal cholangiography, if negative for antimitochondrial antibody (AMA). At questionnaire, history of associated autoimmune diseases (such as rheumatoid arthritis, thyroid-associated disease, Raynaud's syndrome, scleroderma/calcinosis Raynaud esophagus sclerosis teleangiectasiae, Sjogren syndrome), other co-morbidities (diagnosed by patients primary practitioners such as depression, rheumatic fibromyalgia, diabetes, hypertension, reflux), and detailed history of medications were recorded. Verbally reported symptoms, including fatigue, pruritus, sicca symptoms, and right upper quadrant pain, were routinely documented in the electronic clinic chart.
Demographic data collected at the time of questionnaire included age, sex, ethnicity, duration of disease, use of ursodeoxycholic acid (UDCA, 13-15 mg/kg/day), and body mass index (BMI). Standard liver biochemistry (alanine aminotransferase, aspartate aminotransferase, total bilirubin, gamma-glutamyltranspeptidase, and alkaline phosphatase [ALP]) along with other standard laboratory investigations (creatinine, hemoglobin, and thyroid-stimulating hormone levels) were retrieved. Serum immunoglobulin G, immunoglobulin M, and titer of serum AMA (routine immunofluorescence) or AMA-M2 (Pharmacia Diagnostics, Dorval, Quebec) were recorded. Serum biochemical data were available for all subjects at the time of questionnaire and from within the year immediately before symptom assessment. Data from liver biopsy, abdominal ultrasound, as well as upper endoscopy, were also collected.
PBC-40 is a 40-item scale measuring health-related quality of life in PBC, readily applicable to routine clinic practice, as a way of patients evaluating their symptoms.26 It consists of specific symptom domains (Cognition, Itch, Fatigue, Social, Emotional, and Symptoms) and is designed for self-completion. Participants rate items on a five-point scale (1 = ‘never’ to 5 = ‘always’), with high scores denoting greater symptoms impact and poorer quality of life. A previous study defined ranges of severity for the symptom domains contained in the PBC-40.21 By using these clinically meaningful cutoff values applied to the scores from the PBC-40Fatigue domain, no fatigue was a score of 11 or less, mild was a score of 12 to 28, moderate was a score of 29 to 39, and severe was a score of 40 or greater. To test the reliability of the questionnaire in our PBC patient population, the PBC-40 questionnaire was applied twice, at a 1-year interval, to a random sample of 196 patients.
Data were analyzed using SAS. Results are reported as mean ± standard deviation. Categorical variables were analyzed using a series of t tests and chi-squared test (or Fisher's exact test where appropriate). Pearson correlation coefficient (or analysis of variance where appropriate) was performed to analyze correlations between fatigue scores and various biological, demographic, and clinical variables. Finally, variables that were found to be statistically significant were further analyzed with multivariate analysis using a backward selection procedure to determine predictive factors for fatigue. A P value < 0.05 was considered significant.
ALP, alkaline phosphatase; AMA, antimitochondrial antibody; BMI, body mass index; PBC, primary biliary cirrhosis; QOL, quality of life; UDCA, ursodeoxycholic acid.
Demographic and Clinical Data.
Three hundred twenty-seven unselected patients with PBC were included in the review. Clinical, biochemical, and histological stage of disease in the participants are summarized in Table 1. At the time of questionnaire, 94% of the participating cohort were women, and the mean age was 57.3 ± 11.5 (range, 24-90). Two hundred seventy-five patients (84%) had a history of at least one potential co-morbidity.
Table 1. Baseline Characteristics of Patients with PBC at Questionnaire (n = 327)
N (%) or mean ± SD
Age at questionnaire (years)
57.3 ± 11.5
Duration of disease
7.2 ± 5.4
AMA-positive (at questionnaire)
26.9 ± 5.1
Associated autoimmune diseases
Proton pump inhibitors
Clinically documented symptoms
188.1 ± 126.8
40 ± 29
Total bilirubin (μmol/L)
12.7 ± 7.6
0.97 ± 0.12
259 ± 96
2.6 ± 6.1
The laboratory profile was predominantly that of anicteric cholestasis but with normal liver synthetic function and platelet count: alanine aminotransferase 40 ± 29 IU/L (normal range [NR] ≥ 40), aspartate aminotransferase 40 ± 23 IU/L (NR ≤ 35), gamma-glutamyltranspeptidase 142 ± 191 IU/L (NR ≤ 40), ALP 188 ± 126 IU/L (NR ≤ 110), bilirubin 12.7 ± 7.6 μmol/L (NR ≤ 22), INR 0.97 ± 0.12 (NR 0.80-1.20), Platelet 259 ± 96 × 109/L (NR 150-400), and immunoglobulin M 3.4 ± 2 g/L (NR 0.4-2.3). Hemoglobin, thyroid-stimulating hormone, or creatinine levels did not identify a metabolic basis for fatigue in any patient reviewed.
Disease Characteristics in Patients with PBC at Time of Questionnaire.
The average disease duration was 7.2 ± 5.4 years (range, 0-21). At diagnosis, 301 patients (92%) were AMA-positive. At the time of the questionnaire, only 258 patients (79%) remained AMA-positive, as evaluated by AMA-M2 enzyme-linked immunosorbent assay, in keeping with the reported fall of AMA titers on treatment.27 Baseline histological data were available for 291 patients (89%), with most of the patients (78%) diagnosed at an early stage of fibrosis. None of the patients with PBC included had decompensated liver disease, but 31 patients had a history of esophageal varices. The overwhelming majority of patients (n = 315, 96%) were treated with UDCA. Treatment response data28-30 were available across a total of 261 patients.
PBC-40 Domain Scores.
Of the 327 patients, 323 patients (98.8%) successfully completed the PBC-40 in full. Scores for all domains were as follows: Symptoms domain 17.3 ± 4.7 (possible range, 7-3); Itch domain 4.1 ± 3.4 (possible range, 3-15); Fatigue domain 27.4 ± 11.2 (possible range, 11-55); Cognition domain 12.0 ± 5.8 (possible range, 6-30); Social and Emotional (S&E) domains 27.6 ± 11.6 (possible range, 13-65). The distributions of scores of the PBC-40 domains are as shown in Fig. 1. The most frequently seen status was none (54%) for Itch, mild for Symptoms (64%), Cognition (44%), Social and Emotional (72%), and Fatigue (50%). The scores observed in the current study are similar to, albeit numerically lower than, those previously reported from Newcastle, UK (Table 2).
Table 2. Comparison of PBC-40 Domain Scores Across Different Studies
Mean Age (years)
Comparison of responses to PBC-40 domains across various published cohorts demonstrating broad similarity in the symptoms of patients with PBC. S&E, Social and Emotional.
There was as expected inter-domain correlation across the PBC-40 domains (Table 3), with the Fatigue domain scores showing close correlations with Symptoms, Cognition, and Social and Emotional domains score but less correlation with Itch scores. Although each domain of PBC-40 is not equally weighted, the total score similarly correlates strongly with each domain. However, the validity of a total PBC-40 score requires further evaluation.
Table 3. Interdomain Correlations (r Values) for the PBC-40 Domains
Interdomain correlations within the PBC-40 questionnaire are demonstrated with significant correlations shown (*P < 0.001). S&E, Social and Emotional.
Serial Data Confirm Validity of Fatigue Scores.
One hundred ninety-six patients completed the questionnaire for the second time, approximately 1 year later. Within the two PBC-40 domains Fatigue and Itch, there were no significant changes in the score reported, P > 0.05; in other words, the score was reproducible over time.
Subjective Reporting of Fatigue by Patients Underestimates Fatigue Severity.
We found that fatigue was more commonly present when evaluated formally than reported routinely to the physician in clinic. Verbally reported fatigue as a subjective complaint was noted in 156 patients (48%) but found in the majority on PBC-40 completion: mild in 159 (49%), moderate in 92 (28%), and severe in 51 (16%). Of the 167 patients (52%) who did not verbally report fatigue, at questionnaire the symptom was noted as being mild in 63% (n = 105), moderate in 17% (n = 28), and severe in 8% (n = 13) (Fig. 2). Patients who had verbally reported fatigue did, however, have significantly higher scores than those with no verbally reported fatigue (32.4 ± 10.5 versus 22.7 ± 9.8, P < 0.001) (Table 4).
Table 4. Univariate Analysis: Significant Associations with Fatigue in Patients with PBC
Twenty-one patients (6.5%) did not report any fatigue at questionnaire, most of whom were asymptomatic at diagnosis of PBC (n = 18). These patients were not clinically depressed or receiving medications associated with fatigue (such as beta-blockers or antidepressants), and only four patients reported associated autoimmune disease.
Hepatic and Extrahepatic Associations with Fatigue.
Univariate analysis was performed to identify clinical or laboratory markers of fatigue (Table 4). It was noted that a patient's BMI was positively associated with fatigue (r = 0.17; P = 0.002), whereas those patients who were younger at diagnosis had greater fatigue (r = −.16; P = 0.005). The association of fatigue with disease markers was mixed, likely representing varying confounding factors. Sixty-six patients (20%) reported pruritus at the time of questionnaire, and this was associated with higher fatigue scores than those who did not report itch (32.9 ± 11.1 versus 26.0 ± 10.8, P < 0.001). Our average disease duration was just over 7 years, and notably, if patients were fatigued at presentation they were more likely to remain fatigued at the time of questionnaire (P < 0.001). For those diagnosed with noncirrhotic disease, fatigue was more frequent (P = 0.005). However, at the time of questionnaire, the presence of varices (P = 0.034) or cirrhosis on imaging (P = 0.031) was associated with higher fatigue scores, confirming a complex interrelationship between disease severity and fatigue.
Amongst associated autoimmune diseases, scleroderma/calcinosis Raynaud esophagus sclerosis teleangiectasiae was significantly associated with increased fatigue scores (P = 0.022), whereas other autoimmune disorders were not. The presence of fibromyalgia (P = 0.004) and depression (P < 0.001) were similarly associated with fatigue, as was the cumulative number of medical conditions (P = 0.017). Those with two or more co-morbidities had significantly higher fatigue scores (0-1: 26.3 ± 11 versus >2: 29.5 ± 11.5, P = 0.017).
Surrogate markers associated by univariate analysis with a higher fatigue score were use of antipruritics (cholestyramine P < 0.001 and rifampin P < 0.001), proton pump inhibitor prescription (PPI) (P = 0.002), beta-blocker use (P = 0.017), and antidepressant medication (P < 0.001). Patients taking more than three medications were more fatigued than those who were not (29.4 ± 11 versus 25.7 ± 11.2; P = 0.003).
Prior studies have associated the presence of AMA with muscle bioenergetics, thus conceivably providing a link with fatigue.22, 31 In our study, neither the AMA at diagnosis nor that at questionnaire was associated with fatigue (P > 0.05), hence not supporting this hypothesis.
Using a backwards selection procedure to perform multivariate analysis, with significance defined as P < 0.05, we identified calcium and vitamin D use, elevated BMI, stage of disease at diagnosis, presence of varices, clinically reported fatigue at questionnaire, and pruritus as the significant predictors of fatigue when evaluated formally in the PBC-40 questionnaire (Table 5). This broad modeling of our data reinforces the concept that fatigue in PBC is multifactoral. Within each variable it remains highly likely that there are related factors that we are unable to capture or define accurately that contribute to fatigue severity.
Table 5. Multivariate Analysis of Associations with PBC-40 Domain Scores
Reduced multivariate model (backwards selection procedure): A multivariate model was built that simultaneously included scleroderma/CREST, depression, calcium and vitamin D use, cholestyramine use, rifampin use, PPI use, beta-blocker use, BMI, age at questionnaire, stage of disease at diagnosis, varices, presence of cirrhosis by ultrasound, fatigue, pruritus, and sicca symptoms at questionnaire, as well as IgM at questionnaire.
Calcium and vitamin D USE
Stage of disease at diagnosis
Fatigue at questionnaire
Pruritus at questionnaire
Biochemical Treatment Response to UDCA and PBC-40 Domains.
The Toronto criteria for treatment response is derived from this clinic practice and predicts no histological progression at 10 years if patients have ALP values less than 1.67 × upper limit of normal after 2 years of UDCA.30 Comparative criteria were also applied as per Pares (normalization of ALP or >40% reduction of ALP after 1 year of UDCA)28 and Corpechot (ALP <3 × upper limit of normal and aspartate aminotransferase less than 2 × upper limit of normal and bilirubin less than 1 mg/dL after 1 year of UDCA).29 Student t tests were used to compare PBC-40 responses between treatment responders and nonresponders. Complete biochemistries for at least one treatment response were available in 261 patients. As demonstrated in Table 6, there were significantly lower symptom scores in all domains other than Fatigue and Cognition, if patients responded as per the Toronto definition. Applying the alternative definitions of treatment response also demonstrated significantly lower total PBC-40 scores in responders than in nonresponders (range, 8.7-14.9 points lower; P < 0.05). Itch scores were significantly lower, absolute difference 1.1-1.9, according to the Toronto (P = 0.02) and Corpechot (P = 0.001) criteria, but not Pares (P = 0.71). Responders by any criteria scored lower values in the Social and Emotional domains; range 3.5-4.1 points lower; P < 0.05.
Table 6. Biochemical Response to Treatment and Association with PBC-40 Domains
RES (n = 142)
NR (n = 75)
RES (n = 170)
NR (n = 43)
RES (n = 141)
NR (n = 81)
PBC-40 domain responses for each domain were compared between treatment responders and nonresponders, using three biochemical definitions of treatment response (see main text). NR, nonresponder; RES, responder.
80.1 ± 30.4
91.5 ± 30.4
82.2 ± 30.4
97.1 ± 33.0
82.9 ± 31.6
91.6 ± 32.4
16.5 ± 4.3
17.8 ± 5.0
16.9 ± 4.7
18.1 ± 4.7
17.3 ± 4.8
17.2 ± 4.9
3.4 ± 2.7
4.5 ± 3.6
3.7 ± 2.8
5.6 ± 4.1
4.0 ± 3.2
4.2 ± 3.2
25.4 ± 10.7
28.3 ± 11.1
25.3 ± 10.3
30.7 ± 12.1
25.9 ± 10.6
28.2 ± 11.8
11.1 ± 5.6
12.0 ± 5.6
11.6 ± 5.6
12.7 ± 6.4
11.5 ± 5.6
12.6 ± 6.1
25.4 ± 10.8
28.9 ± 11.6
26.2 ± 10.8
30.3 ± 12.7
25.9 ± 10.7
29.4 ± 12.5
Fatigue is a common but complex symptom that is poorly understood and lacks effective treatment. Up to 85% of patients with PBC will complain of fatigue, and it is often a symptom that negatively impacts on the quality of life of patients, as well as having been suggested to be associated with early mortality.32 In this study, we set out to explore and describe the frequency and severity of fatigue in patients with PBC, through the use of a multidomain disease-specific QOL tool, the PBC-40, and to specifically define the role of comorbidities in fatigue. We confirm the importance of this symptom for patients with PBC but clearly show the relevance of comorbidities in determining fatigue severity. Future studies that seek to identify mechanisms and interventions for fatigue need to account for this lack of disease specificity.
Although it is clear that symptoms are an important component of this disease for patients, it was significant to us to note the discrepancy between verbally reported fatigue and self-assessed fatigue, demonstrating an ongoing reluctance of patients to complain about their symptoms with their physicians. The reasons behind this are speculative but may reflect an ability of many patients to cope with, and therefore not acknowledge, mild symptoms. It is also possible that patient perception remains that physicians are too busy to focus on fatigue. If there were clearer data regarding the significance of fatigue as a symptom itself, this might be beneficial to patients. The relative scoring for the individual domains of PBC-40 (Fatigue as well as Social and Emotional domains scoring the highest, and Itch the lowest) mirrored those previously observed21, 23, 26 (Table 2). This in itself is an important observation that supports the PBC-40 as a reproducible and representative QOL measure in PBC, applicable in centers outside of the originating unit. However, a more formal evaluation using matched patients would be valuable, because symptoms of disease are likely influenced to some degree by local factors. In that regard, we were able to demonstrate how fatigue is in fact a variable symptom for those with PBC, with a minority of patients untroubled by energy-related problems. Of the 323 participants in the study, fatigue was present in any severity in 94% (n = 302), that is, fatigue appears to be a normal part of life. As a symptom it was considered the worst, or one of the worst, symptoms in 44% (n = 143).This is in contrast to previous studies conducted in Newcastle upon-Tyne, showing moderate and severe fatigue to be present in 60% to 63% of patients.21, 33 Such distinctions in clinical cohorts may herald from a number of local demographic and referral practice issues, but they do reinforce the concept that what is described by one unit as characteristic for a disease may not be universally applicable. As new medications are developed for PBC, their efficacy will be measured not only by biochemical or histological end points, but in terms of symptoms as well. Because any new treatment study will have to be multi-center and multinational, prior distinctions in symptom severity, and their baseline associations, are important to document.
In our current study, we examined the role of extrahepatic factors related to fatigue and characterized the associations between these factors and severity of fatigue, with a clear focus on patients with PBC. The data presented here confirm that the association of chronic health conditions with fatigue in PBC were significant and cannot be disregarded. Our study adds significantly to the findings of an Italian group who also showed an apparent association between comorbidity in PBC,12 but in a very limited sample size and without recourse to a disease-specific QOL tool. In this uniquely sized North American study, we extend the previous work by using for the first time in this context a more sensitive and disease-specific QOL measure for use in PBC. Our study showed that a higher number of comorbidities (>2) as well as a higher number of medications (>3) was associated with higher fatigue scores. The association with depression is notable. Clearly we are not able to distinguish between direct potential causes for fatigue (for example, beta-blocker use) and surrogate associations (e.g. the possible increased prevalence of functional bowel disorders in those using a proton pump inhibitor). Our study design prevents us from doing this but this does not detract from the concept that fatigue is not specific to PBC, nor does it make less relevant the message to clinicians to think broadly about fatigue and its cause, if they are to successfully intervene. Others before have shown, for example, that 12% of patients with PBC fulfilled criteria for irritable bowel syndrome and had more gastrointestinal symptoms and higher fatigue scores.11 The pruritus reported by our patients was also associated with higher fatigue scores, and at the time of evaluation 33 patients were currently using antipruritic prescription. Despite that, we show that patients who were on these medications scored high on the Fatigue domain of PBC-40 (34.5 ± 10.8 versus 26.5 ± 11.0, P < 0.0001). This could be attributed to the symptom being in need of better control as well as speculatively poor sleep quality because of itching.
The associations we demonstrate with disease severity are complicated by unknown confounders. Therefore, although at the time of survey disease severity does seem to associate with fatigue, it had an inverse relationship when factored backwards with regards to stage at presentation. Confounders regarding modes of presentation may help explain this apparent paradox, as might age at presentation, for example. Our multivariate analysis went on to provide a model for fatigue association that included BMI, disease severity, and clinical symptoms. The association with calcium and vitamin D use is presumably a surrogate for an unknown factor. Of note, AMA status was not related to the presence of fatigue. It is well recognized that AMA titers can fall during the course of treated PBC, and our failure to demonstrate a relationship between AMA titers and fatigue argues against the hypothesis that there is a direct metabolic link between mitochondrial antibodies and changes in mitochondrial function in patients.22, 31 For the symptom of fatigue, we were also not able to show an effect of treatment response, although intriguingly some components of the PBC-40 domains, Symptom and Itch, did seem to improve. This provides further support to the complex interactions that underpin fatigue in patients with PBC.
All such studies are prone to limitations even after accounting for the clear strengths of our work, namely, that of an extensive clinical dataset from a very large single-centre cross-sectional study. Criticism could be raised for the absence of a control arm, but this study set out not to compare fatigue severity across liver disease, but to precisely evaluate fatigue in PBC in the context of the whole patient. Therefore, a control group was not absolutely necessary or appropriate. Future studies will also need to be particular in having uniform criteria in the definition and assessment of comorbidities and assessing for other causes of fatigue, The presence of fatigue in other liver disease supports our overarching findings. Future studies are required to validate and refine our findings, particularly in clinic populations from different parts of the world, and where possible with longitudinal evaluation of the significance of any observations to outcomes, because this remains a point of concern.8 Additionally the methods we applied to define comorbidities likely underreport such associations, because more formal involved evaluations of patients would be helpful. It was not possible in this study, for example, to have an in-depth depression evaluation. This does not detract from our findings, because our definitions of comorbidities were conservative. Having objective numerical evaluations of fatigue is difficult, and as is clearly shown in our study, there is a disparity between physician-reported and objective assessment. The analysis, however, is by its nature based on the numerical scores reported, and this represents a limitation in terms of clinical significance at an individual patient level. However, the purposes of such analyses are to guide future research studies and help define and refine the questions they set out to answer. In this way, future work can come closer than we have been able to, in specifying the factors that account for fatigue as a whole.
In conclusion, we confirm that fatigue is a prevalent concern for patients with PBC that is underreported to physicians routinely. We demonstrate that the symptom complex has a multifactorial cause and is not specific to the disease. Careful appraisal in clinic is therefore relevant when addressing this symptom. Furthermore, when evaluating the biological basis of this symptom, or developing novel interventions, studies must account for these demonstrated extrahepatic associations.
We thank our patients for their ongoing support of our clinic, Jenny Heathcote for her guidance, and Tamara Arenovich for statistical input.