Author's reply: “Serum bile acids in cystic fibrosis patients—Glycodeoxycholic acid as a potential marker of liver disease”

We have read with interest the comments [ ] on our recent article, which showed that glycodeoxycholic acid discriminates between cystic fibrosis (CF) patients with non-cirrhotic liver involvement and no detectable disease (AUC: 0.924, 95%CI 0.822–1.000, p < 0.001) [ ]. Given this result, we do not share the concern that sample size is an issue. Please note that to include 25 patients with liver cirrhosis, as defined in our study, we needed to screen almost 1000 patients in total (nearly half of the national population). Our prospective work on the bile acid profile is also not easily comparable with the retrospective clinical analysis by Colombo et al. (in which one of the authors was involved), where the bile acids were not measured [ ]. Interestingly, that research revealed the non-significantly higher crude cumulative incidence of portal hypertension in patients treated in centers prescribing ursodeoxycholic acid (UDCA) as compared to those followed by centers not applying UDCA ((10.1% (95% CI: 7.9–12.3) vs. 7.7% (95% CI: 4.6–10.7)). Second, key clinical characteristics proposed in the correspondence are already reported in the article, including exocrine pancreatic status and individual CFTR mutations (all the mutations in all patients). Besides, we did assess whether bile acid concentrations differ between CF patients with cirrhosis, other liver involvement and no liver disease – in groups that, by definition, could not be equal and should not be matched. Third, it was suggested that we did not present data on bile acids other than glycodeoxycholic acid. Please note that seven other bile acids were also measured, and their levels were summarized in table 2 in the original article. Fourth, the research on ivacaftor [ ] cited in the letter does not directly apply to our study because there were no patients with CFTR mutations S1251N or G551D. In addition, our work was conducted at a time when combined therapy was not yet available. Fifth, the current study only included patients in stable condition who did not require antimicrobial therapy in the previous six weeks (specified in the methods section). The article referenced in the correspondence does not refer to bile acids measured in the serum, but in the broncho-alveolar lavage (BAL), and the bile acids were likely to “reach the lower airways through duodenogastroesophageal reflux” [ ], which is not immediately related to the current study. Documented “temporal links between bile acids and inflammatory markers” reflected a different aspect of the disease. We hope that the above discussion will help to lift the presented concerns.