Fibromax-based nonalcoholic fatty liver disease in chronic obstructive pulmonary disease patients with obstructive sleep apnea: Methodological considerations

The relationship between nonalcoholic fatty liver disease (NAFLD) and obstructive sleep apnea (OSA) has been well demonstrated, but remains to be evidenced in chronic obstructive pulmonary disease (COPD). Recently, Viglino et al. (Eur Respir J, 2017) attempted to determine the prevalence of liver fibrosis, steatosis and nonalcoholic steatohepatitis (NASH) in COPD patients, some of whom had OSA, basing the NAFLD diagnostic on three circulating biomarker-based liver scores: the FibroTest, SteatoTest and NashTest, from the Fibromax® panel. Among the main findings, the absence of OSA treatment emerged as independently associated with liver fibrosis and steatosis, when compared to effective treatment. However, besides the low number of treated patients, no polysomnographic respiratory data was provided, making it difficult to differentiate the impact of OSA from that of COPD in NAFLD prevalence. Furthermore, NAFLD diagnosis relied exclusively on circulating biomarker-based liver scores, without histological, imagery or other liver exploratory methods. Therefore, in this article, some methodological points are reminded and discussed, including the choice of OSA measurements, and the significance of ActiTest and AshTest scores from Fibromax® in this pathophysiological context.

The relationship between nonalcoholic fatty liver disease (NAFLD) and obstructive sleep apnea (OSA) has been well demonstrated [1][2][3] , but remains to be evidenced in chronic obstructive pulmonary disease (COPD). To this end, using biomarkerbased Fibromax® scores, Viglino et al. recently attempted to determine the prevalence of liver fibrosis, steatosis and nonalcoholic steatohepatitis (NASH) in COPD patients, which they found at nearly 61%, 41%, and 37%, respectively 4 . Interestingly, the absence of OSA treatment emerged as independently associated with liver fibrosis and steatosis, when compared to effective treatment. However, the number of treated patients was low (10 versus 38 untreated), and no polysomnographic respiratory data was provided, making it difficult to differentiate the impact of OSA from that of COPD in NAFLD prevalence. Furthermore, NAFLD diagnosis relied exclusively on circulating biomarkerbased liver scores, without histological, imagery or other liver exploratory methods. It is, therefore, the opportunity to remind and discuss some methodological points, especially concerning the choice of OSA measurements, and the significance of ActiTest and AshTest scores from Fibromax® in this pathophysiological context.

1) Impact of OSA on NAFLD in COPD patients
Recently, Pépin's team showed a prevalence of liver steatosis of about 40-45% in moderate-to-severe OSA patients, with nearly 40-60% of patients displaying borderline NASH 5 . They also showed in obese OSA patients that fibrosis and NAFLD-related lesions, like hepatocyte ballooning and lobular inflammation, were more severe in those with the highest nocturnal oxygen desaturation 6 . Accordingly, nocturnal time spent at <90% oxygen saturation was independently associated with liver fibrosis in patients with suspicion of OSA 7 . In the Viglino et al. study, the absence of OSA treatment emerged as an independent factor of liver fibrosis and steatosis, as compared to treatment. Including the absence versus effective OSA treatment in the multivariate model is not the most appropriate criterion, since the authors recently showed that 6-12 weeks of CPAP treatment did not reduce steatosis, NASH or liver fibrosis 5 . This is in addition to the very low number of treated patients (10 vs 38 untreated). Instead, and according to the increasingly obvious hypothesis of chronic intermittent hypoxia (CIH) on NAFLD, the authors could have chosen the oxygen desaturation index as a criterion (e.g. with a cut-off < or ≥15 events/h), which they elsewhere claimed to be "a good marker of CIH" 6 . Studies displaying detailed results for both OSA and liver scores are few to date, and do not allow an in-depth analysis of their relationships. Therefore, in such NAFLD/COPD/OSA-related studies, polysomnographic respiratory profiles should be provided, including oxygen desaturation measurements, along with liver scores and detailed biology, for the overall COPD group and for OSA patients (all, treated and untreated), in order to strengthen conclusions and enable comparison with further studies.

2) Fibromax® scores
Viglino et al. focused on the three most appropriate NAFLD scores (i.e. FibroTest, SteatoTest, and NashTest). However, they did not provide ActiTest, another Fibromax® score proposed for the estimation of liver necroinflammatory activity in chronic hepatitis C and B 8 , which is based on the measurement of the five Fibrotest® parameters plus alanine aminotransferase. Interestingly, ActiTest has been shown as highly accurate for the diagnosis of NASH and steatosis in patients with severe obesity, with notably an excellent negative predictive value for NASH at 96% using a cut-off at 0.29 9 . In another study on patients with suspected NAFLD, ActiTest showed a significant diagnostic value for NASH, which was not shown for FibroTest 10 . The informative value of ActiTest on the inflammatory component of NAFLD in COPD patients with or without OSA remains questionable; it could therefore be provided and discussed with regards to other inflammatory biomarkers (tumour necrosis factor-α and leptin in the Viglino et al. study).
Furthermore, alcohol consumption at ≥20g/day (women) and ≥30g/day (men) was chosen by the authors as exclusion criteria to discard potential alcoholic steatohepatitis (ASH). However, alcohol consumption may vary over time, even in abstainers or occasional drinkers, and may thus introduce a misclassification bias 11 . Self-reported alcohol consumption remains subjective and should ideally be evaluated using a reliable and objective measure. In this way, the AshTest score -the fifth of Fibromax® -is proposed for the detection of alcoholic steatohepatitis 12 , and thus could be provided as a control for non-excessive alcohol consumption in such pathophysiological contexts.

3) Statistical consideration
Depending on the score, age, sex and/or weight and height are included in the Fibromax® calculation formulas. Therefore, the multivariate analyses from Viglino et al, which included age, sex and/or BMI as independent variables in addition to the Fibromax® score as a dependent variable may induce multicollinearity, and thus cause imprecise estimates of coefficient values or introduce large prediction errors in the case of extrapolation.
Consequently, multicollinearity must be tested in such models, and controlled as much as possible.

4) Biomarker standardization
Viglino et al. did not provide any information about the methods used for Fibromax® parameters. Analytically, standardization allows the reduction of inter-laboratory variability. It is of particular importance for gamma-glutamyl transferase, known for its high inter-method variability, as well as for transaminases, which are measurable with or without pyridoxal 5-posphate as a coenzyme activator 13 . Fibromax® proteins also need standardization given their weight in score calculation, especially α2-macroglobulin 14 . Comparison with peer and method groupsthrough programs of quality control -allows inter-laboratory variation assessment; it is an analytical requirement of the ISO15189 standard for accreditation of medical laboratories, which is a strong guarantee of result reliability 15,16 . If ISO15189 certified methods are used for Fibromax® assays, it must be mentioned, along with methods and analyzers, in order to strengthen the biomarker component, to make it sufficiently informative to be compared with further studies.
To conclude, in such studies evaluating NAFLD, based exclusively on combined-biomarker scores without clinical, histological, imagery or other liver exploratory methods, information on assay methods, analyzers, and guarantees of analytical performance are required, which requires a strong collaboration between clinicians and lab practitioners.

Competing interests
No competing interests were disclosed.

Grant information
The author(s) declared that no grants were involved in supporting this work. In section 1, I think the article may benefit from a very brief mention of the biological differences between intermittent and sustained hypoxia -experimentally these exposures yield very different metabolic outcomes, and the mechanisms which underlie these effects (sympathetic activation, oxidative stress, tissue oxygen profiles, etc.) may be quite different in sustained hypoxia as in COPD, versus chronic intermittent hypoxia, as in OSA.

Is the conclusion balanced and justified on the basis of the presented arguments? Yes
No competing interests were disclosed. This letter is interesting, written in good and fluent English, clearly addressing the limits of a paper attempting to find a link between obstructive sleep apnea (OSA) and Non alcoholic Fatty Liver, in patients with chronic obstructive pulmonory disease (COPD). I agree with Monneret that one of the main limit of this paper, among others, is that it is quite difficult to differentiate the impact of OSA from that of COPD in fatty liver prevalence, because of the common respiratory nature of OSA and COPD and the common negative effect in oxygen saturation. Therefore, although the paper has the merit of pointing out that non alcoholic fatty liver disease can be present -and maybe more often present-in patients with COPD, more polysomnographic respiratory profiles should be provided, including oxygen desaturation measurements.

Is the conclusion balanced and justified on the basis of the presented arguments? Yes
No competing interests were disclosed.

Competing Interests:
I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.

The benefits of publishing with F1000Research:
Your article is published within days, with no editorial bias You can publish traditional articles, null/negative results, case reports, data notes and more The peer review process is transparent and collaborative Your article is indexed in PubMed after passing peer review Dedicated customer support at every stage For pre-submission enquiries, contact research@f1000.com