The True Impact of Fatigue in Primary Biliary Cirrhosis: A Population Study

Transcription

1 GASTROENTEROLOGY 2002;122: The True Impact of Fatigue in Primary Biliary Cirrhosis: A Population Study JENNIFER GOLDBLATT,* PHILIP J. S. TAYLOR, TOBY LIPMAN, MARTIN I. PRINCE,* ANNA BARAGIOTTA,* MARGARET F. BASSENDINE,* OLIVER F. W. JAMES,* and DAVID E. J. JONES* *Centre for Liver Research, University of Newcastle, and Primary Care Service, Newcastle-upon-Tyne, United Kingdom Background & Aims: Patient surveys suggest that fatigue is a common problem in primary biliary cirrhosis (PBC). The actual extent of the problems caused by fatigue in PBC has yet to be determined as previous studies addressing this question have tended to use selected patient subgroups and subjective or non-quantitative fatigue assessment tools. Here, we have attempted to more accurately assess the extent of fatigue in PBC, and the specificity of the symptom for this disease, by the application of an objective measure of fatigue impact (the fatigue impact score [FIS])to a geographically based patient cohort, age- and sexmatched normal controls, and chronic liver disease controls. Methods: Postal completion of the FIS and linked symptom assessment tools. Results: Median FIS was significantly higher in patients (n 136)than community controls (40 [0 138] vs. 28 [0 156]; P < ) and chronic liver disease controls (n 38)(20.5 [0 145]; P < 0.05). Fatigue scores in the 11 patients who had undergone liver transplantation (median 3.5 years previously)were the same as those in non-transplanted patients with advanced disease. Conclusions: Fatigue is a significant and specific problem in PBC. It is not, however, universal and affects fewer patients than has previously been thought to be the case based on data from selected patient cohorts. This definition of the normal range for fatigue in PBC will assist in future studies of etiology and therapy. The importance of quality of life issues, and the value of formal assessment of health-related quality of life as a disease outcome measure, is increasingly being appreciated in liver as in other diseases. 1 The autoimmune liver disease primary biliary cirrhosis (PBC) 2,3 is particularly characterized, in patient reports, by symptoms such as fatigue that often dramatically impair quality of life. 4 Early descriptive studies suggested that fatigue affects up to 80% of PBC patients, 5 resulting, if true, and given current disease prevalence rates of up to 1 in 700 women over the age of 40 in the United Kingdom 6,7 in a significant population symptom burden. Two problems have beset the study of fatigue in PBC. The first is a lack of appreciation that it is a real symptom resulting from, or associated with, a disease state. The second is the difficulty inherent in quantifying what is a deeply subjective symptom. The first reference to fatigue as a symptom in PBC (affecting 18 of 236 [8%] of a patient cohort) was made by Christensen et al. in Malaise was described as a component of a symptom complex present in 12 of 93 (13%) of the Newcastle-upon-Tyne patient series in By 1990, however, fatigue as a clinical problem was being reported by 141 out of a series of 175 (81%) Canadian patients. 5 The sharply contrasting findings of these studies led us and others to explore the use of more objective measures of fatigue and its impact on daily living. Two separate groups have now validated the Fatigue Impact Score (FIS), an assessment tool quantifying the impact of fatigue on the activities and quality of daily life developed for use in chronic fatigue syndrome and multiple sclerosis, 10 for self-completion in PBC. 11,12 These studies suggested a significant impact of fatigue on the lives of PBC patients. Both were, however, limited by the methods used for selection of the study group. The use of patients referred to tertiary centers, 12 or who belonged to patient support groups, 11 could result in skewing of study populations towards more advanced and/or more symptomatic disease, with a resulting over-estimation of fatigue severity. In this study, we sought to address more accurately the impact of fatigue in the general PBC population by applying the FIS to an epidemiologic cohort of PBC patients (defined by geographical area of residence). A case control approach, in which all patients were paired with age- and sex-matched community controls, was Abbreviations used in this paper: AIH, autoimmune hepatitis; FIS, fatigue impact score; PBC, primary biliary cirrhosis by the American Gastroenterological Association /02/$35.00 doi: /gast

2 1236 GOLDBLATT ET AL. GASTROENTEROLOGY Vol. 122, No. 5 used. This allowed us to address the important question (as yet not studied) of the extent to which the fatigue experienced by PBC patients is different from that experienced by a normal population of the same age and sex. Study of a second control group of patients with another form of chronic liver disease (autoimmune hepatitis [AIH]) allowed us to examine the extent to which fatigue is a problem specific to PBC. Subjects and Methods Study Design and Setting This was a geographically based, controlled study of the impact of fatigue in PBC performed in the North East of England. In addition to comparison with normal community case-controls, fatigue impact in PBC was also compared with that seen in a non-pbc chronic liver disease control group (a comprehensive cohort of patients from the same geographical area previously identified as having AIH 13 ). Previously described exhaustive case finding approaches have allowed the identification of extant cases of PBC within the geographical area NE1-NE25 (encompassing Newcastle upon Tyne and Gateshead). 6,14,15 PBC patients and community case-matched and AIH controls were invited to complete a questionnaire (which can be accessed following links from liver) containing the previously validated FIS. 11 The FIS, which consists of 40 questions addressing the impact of fatigue on aspects of daily life (maximum score 160), contains 3 intermixed domains addressing physical, cognitive (maximum score 40 each), and psycho-social (maximum score 80) elements of fatigue. PBC subjects also completed a visual analogue scale (VAS) for pruritus assessment. Given the established association between fatigue and depression, all subjects were also assessed for the likelihood of depressive illness by use of the Rand MOS Depression Screener ( org/health/surveys/screener.html). This 8-item screener does not place undue emphasis on factors that could be influenced by the presence of physical illness. It has a proven history of use in chronic disease and community-based populations. 16 Scores (between 0 and 1) indicate the probability of the presence of depressive disorder. This study was approved by all relevant local ethical committees. Subjects All diagnosed PBC and AIH control patients in the study areas were identified as previously described. 13,17 All enrolled PBC patients had definite (all 3 of antimitochondrial antibody at a titer of 1:40 by immuno-fluorescence, cholestatic liver function tests, and compatible liver histology) or probable (2 out of the 3 above criteria) disease as previously defined. 17 All AIH controls had definite or probable disease as defined by the International Autoimmune Hepatitis Group. 18 Standard clinical data relating to therapy history, liver synthetic function, immunologic status and histologic stage were collected for all patients. Serum biochemical data were available for all subjects from within the year immediately before symptom assessment. Histologic data, where available, were historical as no patients underwent liver biopsy for the purposes of this study. Age- and sex-matched community controls were selected from the primary care practice age and sex registers of appropriate general practices situated within the NE1 NE25 postcode districts. In each case, the 2 normal subjects of the same sex with birth dates closest to that of the index PBC/AIH patient were approached by their primary care physician and invited to participate in the study. Only subjects identified as being willing to participate were subsequently sent symptom assessment tools. In cases in which individuals were unwilling to participate, the normal control with the next closest birth date was identified and approached. This process was continued until 2 controls were identified for each PBC/ AIH patient (2 sex- and age-matched controls were approached for each PBC and AIH patient to allow for the anticipated lower final response rate from population controls). Clinical details of specific conditions previously associated with fatigue (diabetes and thyroid disease) were supplied by participants, as were details of current medical therapies. Statistical Analysis Fatigue scores for PBC and community and AIH control subjects were compared using non-parametric testing. The numbers of subjects reporting high and low fatigue scores in the PBC and control populations were compared using the chi-squared test. The relationship between FIS scores and demographic and clinical variables (including depression scores) was analyzed using univariate and multivariate analyses. FIS scores were logarithmically transformed before inclusion in the regression equation. Results Fatigue Impact in PBC and Community and Chronic Liver Disease Controls Symptom assessment tools were sent to all 194 PBC patients identified in the study area on the census date of December 31, Responses were received from 157 of 194 (81%). Of these assessment tools, 136 of 157 (87%) were satisfactorily completed. The PBC subjects who returned completed symptom assessment tools, and who were therefore included in the study, were found to be representative, as a group, of the study population as a whole. There were no statistically significant differences between the whole PBC patient cohort and the group returning completed symptom assessment tools with regard to patient age (median 67.5 years [range, years] vs years [range, years]), proportion of female patients (91% vs. 91%), case status (66% definite disease and 34% probable vs.

3 May 2002 FATIGUE IN PBC % and 36%, respectively), and histologic stage (14% histologically proven advanced disease vs. 11%). Both the geographically based source population and the group returning symptom assessment tools were predominantly made up of patients with non-advanced disease. The median Childs-Pugh score in the non-transplanted patients at the time of the study was 5 (range, 5 10). At the study point, 109 of 125 (87%) of nontransplanted patients had normal serum bilirubin and albumin levels. Only 10% of patients had elevated serum bilirubin. The vast majority of the non-transplanted patients (94%) presented either after discovery of abnormal liver biochemistry or the presence of anti-mitochondrial antibody on screening, or investigation for the symptoms fatigue, pruritus, or joint pains. Only 7 patients (6%) presented with the clinical features of advanced disease (hepatocellular dysfunction or the features of portal hypertension). Each PBC patient enrolled into the analysis phase of the study was paired with the first community control matched to them at the outset of the study who responded with a completed assessment tool. Matching was effective with 124 of 136 (91%) of subjects being female in both the PBC patient and the control groups. Mean age at the time of the study was 65.5 for the PBC patient group and 66.0 for the controls. The PBC population contained 11 patients who had undergone liver transplantation a median of 3.5 years (range, years) before the census date. The median FIS score was significantly higher in the PBC patient group than in the matched community control group both for the whole PBC population (40 [range 0 138] vs. 28 [0 156], P 0.005) and the non-transplanted patient group (n 125) (40 [0 138] vs. 29 [0 156], P 0.005). Score distribution was skewed towards the physical domain in both PBC patients and community controls. Scores in each domain were significantly higher in the PBC patients than in controls (Table 1). The overall distribution of fatigue scores in the 2 groups is shown in Figure 1. It is notable that the PBC population included a sizeable group who reported little or no impact of fatigue on their daily lives (Figure 1). This group was, however, significantly smaller in the PBC patient than in the control cohort group; 61 (45%) PBC patients had scores below 35 (the mean value for the control population) compared with 83 (61%) controls (P 0.01). Conversely, a higher proportion of the PBC patient group had high scores (total score mean 1 SD for the control population [67]) (42 [31%] vs. 19 [14%], P 0.001). Table 1. Comparison of Total and Individual Domain Fatigue Impact Scores Between PBC Patients and Matched Community AIH Controls Cognitive (max 40) Median fatigue scores (range) Physical (max 40) Psycho-social (max 80) Total FIS (max 160) PBC (n 136) 9 (0 39) 15 (0 38) 19 (0 68) 40 (0 138) Normal control (n 136) 5 (0 39) 11 (0 38) 11 (0 79) 28 (0 156) P P P P AIH control (n 38) 4.5 (0 30) 8.5 (0 34) 10 (0 68) 20.5 (0 129) P ns P 0.05 P 0.05 P 0.05 NOTE. All statistical comparisons are between the observed value for the relevant control group and its equivalent in the PBC group. Completed symptom assessment tools were received from 38 of 58 (66%) AIH patients identified in the geographical study area. The mean age of the AIH patients returning completed symptom assessment tools was 57.5 years, with 33 of 38 (87%) of returnees being female. Mean revised International Autoimmune Hepatitis group score at diagnosis for the AIH group was Mean doses of prednisolone and azathioprine at the study point were and mg/d, respectively. Median FIS was significantly lower in the AIH patient group than in the PBC patients (AIH 20.5 [range, 0 129], P 0.05 vs. PBC). Median FIS was the same in the AIH patients as in the whole community control group (P 0.9) and community controls specifically matched to the AIH patients for age and sex (median FIS 24.5 [range, 0 145], P 0.9). Distribution of FIS scores was similar in the whole community and AIH control groups (Figure 1). No correlation was seen between FIS and patient age, prednisolone dose, or azathioprine dose in the AIH patients. Disease Phenotype and Fatigue Severity in PBC No difference was seen between the fatigue scores in the PBC patients with Childs-Pugh scores of 5 and those with Childs-Pugh scores of 5 (median 40 [0 138] vs. 37 [0 136], P 0.9). None of the patients participating in the study had a clinical history of encephalopathy. No relationship with FIS was found for any parameter of disease activity (including serum albumin, serum bilirubin, prothrombin time, liver histologic stage, Mayo risk score, and pruritus severity) using univariate analysis (Table 2). Although biochemical data were current for all patients, histologic data referred to the most recent biopsy performed for clinical manage-

4 1238 GOLDBLATT ET AL. GASTROENTEROLOGY Vol. 122, No. 5 ment purposes. This limits the accuracy of the histologic fatigue association data. No relationship was found between total fatigue scores and age, sex, and presence of thyroid disease in the PBC patient group (Table 2). In contrast with previous studies, only a weak positive correlation (r , P 0.001) was seen between fatigue (FIS) and depression (Rand Depression Screener [RDS]) scores (Figure 2A and Table 2). A similar weak correlation (r , P 0.001) was also seen in the community control population (Figure 2B). Rate of reported anti-depressant use was also similar between patients and community controls (20 of 136 vs. 17 of 136). To examine whether the association of fatigue and depression in the PBC patients might be a result of confounding factors, multivariate analysis was performed, including the clinical parameters outlined in Table 2. Depression scores remained significantly associated with fatigue scores (P 0.002). None of the other clinical parameters was significantly associated with the fatigue scores. The use of ursodeoxycholic acid (UDCA) was reported by 40 (29%) PBC patients. There were no significant differences in fatigue scores between those taking UDCA and those who were not (P 0.11). The median FIS in the 11 PBC patients in the study population who had undergone liver transplantation was 38 (range, 0 103). No difference was seen between this subgroup of patients and either the un-transplanted PBC group as a whole or, more specifically, the subgroup of un-transplanted PBC patients with histologically advanced disease (Figure 3). Figure 1. Distribution of total FISs in (A) a geographically based PBC population, (B) age- and sex-matched geographic normal, and (C) AIH control subjects. Discussion In this study, we have shown that the degree of fatigue experienced by a representative cohort of PBC patients is significantly greater than that experienced by age- and sex-matched community controls, and by a second control group of patients with AIH. Although the return of completed symptom assessment tools by the study PBC population was incomplete, the group returning assessment tools was found to be representative of the whole geographically defined study population with regard to both demographic and disease severity parameters. In this study we also showed, for the first time, the extent to which fatigue is a variable symptom in PBC, with a significant minority of patients apparently untroubled by energy-related problems. We believe that Table 2. Univariate Analysis of Transformed PBC Patient Fatigue Scores (Ln) Against Relevant Clinical Parameters Variable r 2 Significance Age Sex Depression score Itch VAS TSH Albumin Bilirubin Alkaline phosphatase Prothrombin time AMA titer Histologic stage

5 May 2002 FATIGUE IN PBC 1239 Figure 2. Correlation between FISs and Rand Depression Scores in (A) PBC patients and (B) geographic controls Figure 3. FISs in the PBC patients who had undergone liver transplantation (Trans) in comparison to the non-transplanted PBC patients (Non-Trans) and the subgroup of non-transplanted PBC patients with histologically proven advanced disease (Scheuer stage III-IV). Minimum, 25% percentile, median, 75% percentile, and maximum scores are indicated for each group. the heterogeneity in symptom severity is likely to result from the inclusion of prevalent patients with the disease within a given geographic area, rather than a serial group attending a specialist clinic. Recruitment of clinic-based series may result in overestimation of clinical problems such as fatigue, as patients with difficult to control problems may be attending the clinic more frequently, and thus be more likely to be enrolled in studies, than asymptomatic individuals. A comparison between the findings of the current study and one previously published by our group makes this point emphatically. In the earlier study, where patients were recruited from a PBC patient support group and a tertiary referral clinic, the median FIS was In the current geographically based study, the median FIS was 40. This observation emphasizes the importance of appropriate and meaningful population selection in studies of PBC symptomatology. We believe that, as a result of its design, the current study gives a truer reflection of the impact of fatigue in the whole population with PBC. The observation that a significant proportion of PBC patients is not troubled by fatigue raises important questions regarding the cause of this symptom. The advent of tools such as the FIS, and data regarding the normal distribution of fatigue severity in the whole PBC population, will greatly assist in future studies of fatigue cause in PBC by allowing identification of high and low fatigue patient subgroups for direct comparison. This approach has already proved fruitful in clinical studies. 19,20 The findings of the current study do, however, already provide some insight into cause. Depression, present to a similar extent in the PBC and control populations, cannot fully explain fatigue in the PBC, although it may contribute to the fatigue experienced by some people. Although the results of the current study concur with the findings of other groups, 12,21 in showing a statistically significant correlation between depression and fatigue scores in PBC, the strength of the association is markedly weaker in the current study than has previously been suggested to be the case. The use of depression scoring systems weighted towards physical manifestations of depression (and which were not designed for use in patients with chronic physical illness) may have, in previous studies, led to overestimation of the contribution of depression to fatigue in PBC. Liver disease severity seems not to be directly linked to fatigue severity. Although the limited number of patients seen in current practice suffering from end-stage disease may conceivably suffer from fatigue resulting from hepatic encephalopathy (there were, however, no patients with encephalopathy in the current study cohort), liver disease severity, as measured by conventional

6 1240 GOLDBLATT ET AL. GASTROENTEROLOGY Vol. 122, No. 5 markers, seems to be entirely unrelated to FIS in the broader PBC population. This echoes the findings of earlier studies performed on different samples of PBC patients. 3,12,21 Likewise, fatigue does not appear to be related to other obvious features such as patient age and/or the presence of relevant intercurrent disease, such as ineffectively controlled thyroid disease. The observation that fatigue scores in patients with AIH are significantly lower than those in PBC patients, and similar to those in matched community controls, is of significance in 2 respects. First, it would suggest that the elevated fatigue scores reported by PBC patients do not simply result from increased symptom reporting by patients diagnosed with chronic disease. Second, it suggests that fatigue, although associated with several forms of chronic liver disease (including hepatitis C infection, 22 non-alcoholic steato-hepatitis, 23 and alcoholic liver disease [CP Day, Newcastle; unpublished data]) is not a universal feature of chronic liver disease. A major issue for PBC patients, in our practice at least, is the question of therapy for fatigue. Fatigue impact scoring systems will be of value as outcome measures for trials of therapy in PBC aimed at the symptom of fatigue specifically and disease modulation in general. The FIS itself has already been used as an outcome measure in one therapeutic trial in PBC. 24 Those therapies aimed at modification of the disease course that have entered widespread clinical use in PBC (including UDCA and corticosteroids) have not been fully studied with regard to their actions on fatigue severity. 3 Our data suggest that UDCA use is not associated with fatigue of reduced severity. Prospective studies specifically designed to address this question, using tools such as the FIS, need to be performed. An increasing awareness among patient groups of the availability and potential benefits of liver transplantation has led, in the experience of the authors, to pressure to consider transplantation for purely symptomatic reasons. The limited published data available suggest a significant improvement in fatigue (as assessed using a 5-point scale) at 1 year post liver transplantation for PBC. 25 This finding has contributed to a patient desire for transplantation for fatigue. Although clearly not designed to address the question of response of fatigue to transplantation, our study, in which we found no difference in fatigue severity between transplanted and non-transplanted patients, raises concerns regarding the effectiveness of transplantation as a treatment modality for fatigue in PBC. It may be significant that the transplanted patients in our cohort were studied a median of 3.5 years after transplantation as opposed to 1 year as reported by Gross et al. 25 One interpretation of the apparent differences between the 2 studies is that any benefits of transplantation with regard to fatigue are not sustained. Long-term prospective studies using more sensitive Fatigue impact scoring systems of the type used here are required to address this important question. In conclusion, we have shown that fatigue is a true and specific feature of PBC, and one that can adversely affect quality of life. It does not, however, affect all patients, nor does it appear to simply result from the presence of advanced disease. An appreciation of the true spectrum of fatigue impact in PBC, and the advent of reproducible measures such as the FIS, will contribute significantly to future studies of the cause and treatment of fatigue in PBC. References 1. Owens DK. In the eye of the beholder: assessment of healthrelated quality of life. Hepatology 1998;27: Neuberger JM. Primary biliary cirrhosis. Lancet 1997;350: Prince MI, Jones DEJ. Primary biliary cirrhosis: new perspectives in diagnosis and treatment. Postgrad Med J 2000;76: Fahey S. The experience of women with primary biliary cirrhosis: a literature review. J Adv Nurs 1999;30: Witt-Sullivan H, Heathcote J, Cauch K, Blendis L, Ghent C, Katz A, Milner R, Pappas SC, Rankin J, Wanless IR. The demography of primary biliary cirrhosis in Ontario, Canada. Hepatology 1990;12: Metcalf JV, Bhopal RS, Gray J, James OFW. Incidence and prevalence of primary biliary cirrhosis in the city of Newcastle-upon- Tyne, England. Int J Epidemiol 1997;26: Kingham JG, Parker DR. The association between primary biliary cirrhosis and coeliac disease: a study of relative prevalences. Gut 1998;42: Christensen E, Crowe J, Doniach D, Popper H, Ranek L, Rodes J, Tygstrup N, Williams R. Clinical pattern and course of disease in primary biliary cirrhosis based on an analysis of 236 patients. Gastroenterology 1980;78: James OFW, Macklon AF, Watson AJ. Primary biliary cirrhosis: a revised clinical spectrum. Lancet 1981;i: Fisk J, Ritvo P, Ross L, Haase D, Marrie T, Sclech W. Measuring the functional impact of fatigue: initial validation of the Fatigue Impact Scale. Clin Infect Dis 1994;18:S Prince MI, James OFW, Holland NP, Jones DEJ. Validation of a fatigue impact score in primary biliary cirrhosis: towards a standard for clinical and trial use. J Hepatol 2000;32: Huet PM, Deslauriers J, Tran A, Faucher C, Charbonneau J. Impact of fatigue on the quality of life in patients with primary biliary cirrhosis. Am J Gastroenterol 2000;95: Baragiotta M, Craig W, James OFW, Mitchison H, Burke D, Bateson M, Trewby P, Bassendine MF. Is the revised international autoimmune hepatitis group (IAHG) scoring system useful (abstr)? J Hepatol 2001;34: James OFW, Bhopal R, Howel D, Gray J, Burt AD, Metcalf JV. Primary biliary cirrhosis once rare, now common in the UK? Hepatology 1999;30: Jones DE, Watt FE, Grove J, Newton JL, Dakly AK, Gregory WL, Day CP, James OF, Bassendine MF. Tumour necrosis factor-alpha promoter polymorphisms in primary biliary cirrhosis. J Hepatol 1999;30: Burnham MA, Wells KB, Leake B, Landsverk J. Development of a

7 May 2002 FATIGUE IN PBC 1241 brief screening instrument for detecting depressive disorders. Med Care 1988;26: Metcalf JV, Howel D, James OFW, Bhopal RS. Primary biliary cirrhosis: epidemiology helping the clinician. BMJ 1996;312: International Autoimmune Hepatitis Group. Review of criteria for the diagnosis of autoimmune hepatitis. J Hepatol 1999;31: Goldblatt J, James OFW, Jones DEJ. Grip strength and subjective fatigue in patients with primary biliary cirrhosis. JAMA 2001;285: Johns E, Goldblatt J, Jones DEJ, Newton JL. Fatigue in primary biliary cirrhosis is associated with early autonomic failure (abstr). Hepatology 2001;34:371A. 21. Cauch-Dudek K, Abbey S, Stewart DE, Heathcote EJ. Fatigue in primary biliary cirrhosis. Gut 1998;43: Foster GR, Goldin RD, Thomas HC. Chronic hepatitis C infection causes a significant reduction in quality of life in the absence of cirrhosis. Hepatology 1998;27: Talwalker J, Keach J, Angulo P, Lindor KD. Health-related quality of life assessment in nonalcoholic steatohepatitis (abstr). Hepatology 2001;2001:251A. 24. Watson JP, Jones DEJ, James OFW, Cann PA, Bramble MG. Case report: oral anti-oxidant therapy for the treatment of primary biliary cirrhosis: a pilot study. J Gastroenterol Hepatol 1999;14: Gross CR, Malinchoc M, Kim WR, Evans RW, Wiesner RH, Petz JL, Crippin JS, Klintmalm GB, Levy MF, Ricci P, Therneau TM, Dickson ER. Quality of life before and after liver transplantation for cholestatic liver disease. Hepatology 1999;29: Received June 27, Accepted January 10, Address requests for reprints to: David E. J. Jones, M.D., Ph.D., Centre for Liver Research, 4th Floor William Leach Building, The Medical School, Framlington Place, Newcastle-upon-Tyne, NE2 4 HH, United Kingdom. D.E.J.Jones@ncl.ac.uk; fax: