Article Text

Download PDFPDF

Fatigue in primary biliary cirrhosis
  1. D M Forton1,
  2. N Patel2,
  3. A Oatridge2,
  4. G Hamilton2,
  5. J V Hajnal2,
  6. H C Thomas3,
  7. S D Taylor-Robinson3,
  8. M Prince4,
  9. J Goldblatt4,
  10. M Bassendine4,
  11. D E J Jones4
  1. 1Liver Unit, Faculty of Medicine, St Mary’s Hospital Campus, Imperial College London, London, UK
  2. 2Robert Steiner MR Unit, MRC Clinical Sciences Centre, Hammersmith Hospital Campus, Imperial College London, London, UK
  3. 3Liver Unit, Imperial College London
  4. 4Centre for Liver Research, University of Newcastle, Newcastle, UK
  1. Correspondence to:
    D M Forton
    Hepatology Section, Faculty of Medicine, Imperial College London at St Mary’s Hospital, 10th floor QEQM Building, South Wharf Rd, London W2 1NY, UK;

Statistics from

We would like to take the opportunity to clarify some of the points in response to the recent leading article (Gut 2004;53:475–7) which accompanied our report1 of reduced globus pallidus (GP) magnetisation ratios (MTRs) in patients with fatigue and primary biliary cirrhosis (PBC).

As we stated in the paper, fatigue in PBC is a subjective multidimensional symptom with many potential determinants, including sleep disturbance, depression, and personality, in addition to a potential central neurological cause.1 We therefore wholeheartedly concur with Drs Milkiewicz and Heathcote when they state that brain manganese (Mn) deposition is certainly not the cause of fatigue in all patients with PBC. We certainly do not believe that we drew this conclusion. However, we do believe that our findings of reduced GP MTRs in patients with stage I–II disease, which were associated with hypermanganesaemia and measured fatigue, do open up a novel avenue of research into a poorly understood symptom in patients with PBC.

In order to control for inter-examination system variability, it is necessary to normalise the raw MTRs against an internal region of interest (ROI). Although it might initially appear easier to analyse the raw MTR data, normalisation to an internal standard allows external sources of variation, unrelated to the patient, to be removed. We followed previously published protocols to calculate GP indices, normalised to the putamen and to the frontal white matter,2,3 and these were used to test associations with fatigue and blood Mn levels. The raw MTR data were used for the primary comparison between PBC patients and healthy volunteers. We chose two rather than one internal control ROI because, contrary to the assertion in the editorial, there is evidence for Mn accumulation in brain structures, other than the GP, in patients with cirrhosis.4,5 Rose et al reported significantly elevated Mn concentrations in the frontal and occipital cortex, pallidum, putamen, and caudate4 while Maeda et al showed elevated Mn concentrations in the GP, putamen, and frontal white matter.5 In both series, the highest Mn concentration was in the GP. Our choice of two standard ROIs was made to maximise the interpretation of the raw data although we accept that the a priori assumption that pathology is absent from these regions in this and all relevant magnetic resonance studies to date, which have used internal controls, may be false. This may explain the unexpected trend towards a positive association between blood Mn and the putaminal index normalised to white matter.

Drs Milkiewicz and Heathcote have expressed concern about an apparent autocorrelation in our data that did not equal 1. Table 2 in our paper1 shows the correlation coefficients between individual MTR indices and blood Mn level. We did not compare the normalised putamen index against the normalised putamen index.

We are grateful to the two commentators for extending our interpretations and naturally agree that bile duct loss, rather than liver fibrosis, governs the severity of cholestasis and that there may be dissociation between these features in PBC. For the purposes of this study, we chose to examine patients with stage I–II disease to remove the possibility of hepatic encephalopathy or cirrhosis as a cause for the MTR findings. We believe that both this patient selection and the demonstration of normal cerebral magnetic resonance spectroscopy (MRS) in these patients, compared with healthy volunteers, does indeed achieve this. We found reduced GP MTRs in patients with stage I–II disease, which were associated with hypermanganesaemia and measured fatigue, but we also studied four patients with stage III–IV disease and, as a group, there were no significant differences in GP MTR indices compared with stage I–II patients. Although this may be due to the small number of individuals studied, the lack of clear distinction between stage I–II and stage III–IV disease may also reflect a process that adversely affects the brain long before the development of cirrhosis, owing to early bile duct loss.

The commentators point out that the value of liver biopsy staging of PBC is limited owing to sampling error and that there may not have been a true distinction between the stage I–II and III–IV groups. We accept the possibility of sampling error but, in our view, liver biopsy still remains the gold standard for diagnosing cirrhosis. We disagree with the suggestion that cerebral MRS would have been useful in supporting the histological diagnoses as cerebral MRS abnormalities are only seen in a minority of patients with Child-Pugh A cirrhosis.6,7 We did not assume that MRS would be abnormal in stage III–IV patients; in fact, there were no significant differences between these patients and stage I–II patients.

Fatigue in PBC merits further research. We hope that we will be able to take further “steps in the right direction”.



  • Conflict of interest: None declared.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.