Ursodeoxycholic Acid Therapy in Cystic Fibrosis Liver Disease

A Retrospective Long-term Follow-up Case-Control Study

M. Kappler; C. Espach; A. Schweiger-Kabesch; T. Lang; D. Hartl; A. Hector; C. Glasmacher; M. Griese


Aliment Pharmacol Ther. 2012;36(3):266-273. 

In This Article


In a retrospective analysis of data of patients treated in our centre between 1975 and 2005, we showed that early treatment with UDCA reduces serum liver enzyme elevations significantly and persistently in CF patients and more importantly, the development of overt liver disease could be prevented to a large extent, making it a rare event nowadays. Early liver disease was detected by regular screening for liver enlargement or persistent elevation of serum liver enzymes in all CF patients.

The prevalence of CF-related liver disease was 26% in this study and is concordant with published data.[3,25,26] We saw an obvious normalisation of serum liver enzymes over the observation period under UDCA, which is in accord with data observed in smaller groups of patients over a shorter observation time.[8,16,18] In our cohort, we did not identify meconium ileus as a risk factor for the development of CF-related liver disease as reported by others,[27] but we found that liver disease started earlier in this subset of patients who therefore should be screened regularly from early childhood on.

Diagnosis of CF-related liver disease was based on (i) clinical signs and (ii) elevation of serum liver enzymes, both of which were assessed very regularly and consistently in order not to miss patients with early liver disease. Our data of annual serum liver enzyme levels were therefore almost complete with only a few missing values (i.e. 99.8% for AST, 99.1% for ALT, 98.6% for γ-GT and 78.3% for GLDH). Additional ultrasound imaging was not used to establish the diagnosis because of the limited specificity and the low sensitivity of the method.[28] Liver biopsies were not taken because fibrosis often begins focally and may be easily missed[29] and also the risks associated with a biopsy cannot be justified in this setting.[30] The non-invasive approach for diagnosing liver disease used here has previously been employed successfully in subjects with hepatitis C[31] and in CF patients.[25]

Overall, the effectiveness of UDCA therapy is supported by the very low incidence of overt liver disease in our patients. In a comparably sized group of historic CF patients, we identified within a similar period of observation 9 subjects in whom liver disease proceeded rapidly to overt liver disease ( Table 1 ). In these subjects, who were not treated with UDCA, CF-related liver disease began early (Figure 2), a fact also reported by others.[25]

Compared with the historic collective, UDCA-treated patients showed a significantly lower incidence of overt liver disease, although the control group was not matched for the presence of early liver disease, making a direct comparison difficult and consequently, conclusions must be drawn with caution. In a study by Gaskin et al.[32] overt liver disease was common and affected about 10% of the 153 CF patients included between 1984 and 1986, when UDCA was not available. The frequency of CF-related liver disease was comparable to our results (40%). This high prevalence of overt liver disease in the pre-UDCA era was also reported by Lindblad et al. for CF patients between 1976 and 1993.[33]

To investigate if long-term UDCA treatment affects lung function, UDCA-treated subjects were matched to non-UDCA-treated CF patients. Changes in lung function over time did not differ with UDCA treatment. Analysis of all values from all 98 subjects in each group (Figure 4) yielded the same results as for the paired analysis (i.e. patients with 5 or more FEV1 values available for regression analysis).

Although the cross-sectional study of Corbett et al. postulated an association between CF-related liver disease and poor FEV1,[34] the 42 patients enrolled in that study, in contrast to our patients, had impaired growth and nutrition. To the best of our knowledge, the data presented here are the first systematic investigation of long-term pulmonary outcome in UDCA-treated patients in a large cohort of CF patients.

Our study has some weak points/drawbacks that have to be discussed. Although our screening strategy and the treatment of all subjects were defined prospectively, the data collection was retrospective and no contemporary, randomised control group was established. The reason for this was that in 1989, it was assumed that UDCA was a physiological replacement of bile salts lost through fat malabsorption, and thus beneficial in CF patients with liver-related problems. Therefore, no control group receiving placebo was introduced. To partially circumvent the associated problems, we selected a case–control study design with one well-matched control group and one historic control group, using the long-term primary endpoint 'development of overt liver disease'. The advantage of such an endpoint is that it can be clearly documented clinically and is highly relevant for the patients and the caregivers. For a disease that is rare and which has a long latent period between onset of underlying disease and complications (i.e. overt liver disease among CF individuals), a case–control study is an efficient study design and probably the only feasible option.[35] In addition, this approach allowed a comparison with historic data which was justifiable as intestinal therapeutic standards have changed little in the past few years, although overall survival was worse between 1975 and 1988.[36] Because of improved survival, subjects in our study were on average older than the historic controls and therefore the probability of developing overt liver disease should have been even higher in this group.

The separate sampling of cases and controls, in this study, CF patients treated with UDCA and CF patients not treated with UDCA, usually increases the susceptibility to sampling bias in case–control studies because the disease, in our case 'liver disease', might not be systematically searched in all the study participants. Such sampling bias was excluded in our study because the complete population of CF patients was screened for liver disease on a regular basis. The number of subjects, who were not diagnosed, misdiagnosed, dead, or seen elsewhere, was thereby reduced to a negligible level.

The strength of this study lies in the fact that a large cohort of patients were handled identically and followed up closely by a rigorous treatment protocol over a very long period of time. We clearly demonstrated a low incidence of overt liver disease and a normalisation of serum liver enzymes with UDCA treatment in CF patients. By using the case–control study design, we found that CF patients treated with UDCA had stable lung function over a long observation period comparable to CF patients not treated with UDCA.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: