Non-invasive assessment of liver function: trying not to miss an opportunity
Article first published online: 18 MAR 2004
Alimentary Pharmacology & Therapeutics
Volume 19, Issue 7, pages 827–828, April 2004
How to Cite
Giannini, E. G. and Testa, R. (2004), Non-invasive assessment of liver function: trying not to miss an opportunity. Alimentary Pharmacology & Therapeutics, 19: 827–828. doi: 10.1111/j.1365-2036.2004.01915.x
- Issue published online: 18 MAR 2004
- Article first published online: 18 MAR 2004
Sirs, We read with interest the paper by Saadeh et al. regarding the usefulness of the 13C-galactose breath test (13C-GBT) for the measurement of liver function.1 In particular, we appreciated the thorough methodological part of their study, although we did not agree with the majority of their clinical conclusions.
First of all, it is not true, as the authors state in their introduction, that: ‘To date, no quantitative liver function test has been demonstrated to be significantly altered in the early stages of disease’.1 In fact, to our knowledge, it has already been proven that both the monoethylglycinexylidide test and 13C-aminopyrine breath test (13C-ABT), just to name two, are altered in the early phases of chronic liver disease.2, 3 In particular, in chronic hepatitis C, 13C-ABT identified patients with low fibrosis scores, and was able to distinguish between normal subjects and chronic hepatitis patients, as well as between patients with chronic hepatitis and those with initial, compensated cirrhosis.3 Moreover, other studies have already shown that 13C-GBT is altered in patients with chronic hepatitis,4 and the results obtained by Saadeh et al. cannot be considered to have been obtained in patients with ‘early’ stages of liver disease, but rather in the early stages of liver cirrhosis.
Secondly, in their study, the authors identified (by means of receiver operating characteristic curve analysis) a given 13C-GBT %dose/h result cut-off at 120 min to distinguish between healthy subjects and Child–Pugh class A patients. However, they did not provide any accuracy data regarding this cut-off. Furthermore, the cut-off that distinguishes between normal subjects and patients with liver cirrhosis as a whole (6.3%) is not very different from the cut-off that has been identified to differentiate between Child–Pugh class A patients and normal subjects (6.4%), while the cut-off used to distinguish between compensated and decompensated cirrhotic patients is very low indeed (1–1.5%).
Thirdly, the authors showed that, in healthy subjects, the 13C-GBT %dose/h results at 120 min were significantly different from those of cirrhotic patients. Although this result is interesting, it is not new, as the same result was obtained by Shreeve et al. more than 20 years ago.5 More importantly, it is not the role expected for such a test. Indeed, the role of non-invasive, quantitative liver function tests in the work-up of patients with chronic liver disease is to diagnose two different conditions that would not otherwise be easily differentiated on clinical grounds alone. As far as this topic is concerned, we do not need any further studies to show that 13C-GBT can distinguish between a healthy subject and a patient with (decompensated) liver cirrhosis, as this has already been established.5 Moreover, we do not need to use an important and sophisticated tool, such as 13C-GBT, to differentiate between compensated and decompensated cirrhotics, as clinical judgement is more than sufficient in this setting. Rather, we need a test that can help the clinician to make difficult decisions, that can eliminate the need for invasive procedures, such as liver biopsy, and that can aid in the assessment of the prognosis in particular situations. We need tests that can differentiate between well-compensated, initial liver cirrhosis and chronic hepatitis, two entities that are hard to distinguish on clinical grounds alone. In this setting, in a population of 61 patients with chronic liver disease, we found that 13C-GBT showed a discrete accuracy in differentiating between patients with chronic hepatitis (n = 40) and those with Child–Pugh class A liver cirrhosis (n = 21). We found that a 13C-GBT %dose/h of 1.2 at 30-min had 76.2% sensitivity (95% confidence interval, 52.8–91.7), 80.0% specificity (64.3–90.9) and 0.804 accuracy (s.e., 0.055) in differentiating between chronic hepatitis and well-compensated cirrhosis. Indeed, these are the first available data to report the usefulness of 13C-GBT as a diagnostic aid in these two subgrops of patients.
To conclude, we believe that non-invasive assessment of liver function by means of 13C breath tests is an important aid for the clinician, and a combination of multiple tests may even provide a deeper insight into various aspects of liver disease.6 However, in order to avoid missing an opportunity, we believe that their widespread use should not be hampered by their inappropriate use.7