Comment
Dynamics of liver stiffness in chronic hepatitis B patients with concurrent metabolic-associated fatty liver disease
February 19, 2021
To the Editor:
We read with interest the article by Mak et al.1 and commend their work exploring the diverse relationship between hepatic steatosis and chronic hepatitis B (CHB). Through prospective follow-up of 330 patients with normal alanine aminotransferase (ALT) and low viraemia, Mak et al. showed persistent severe steatosis (controlled attenuation parameter ≥280 dB/m) was associated with progression of fibrosis category after three years. As modern nucleos(t)ide analogues are able to achieve effective long term viral/biochemical suppression and attenuate fibrosis development,2 the focus of CHB management has shifted to address non-viral risk factors such as concurrent steatosis—with approximately 30% prevalence amongst patients with CHB3 —and metabolic-associated fatty liver disease (MAFLD).4 Although Mak et al. provided useful insight into the effects of severe steatosis on virologically quiescent CHB, the broader relationship between these two conditions remains elusive. We aimed to provide additional insight by modelling serial liver stiffness measurements (LSM) over time and comparing their trajectories amongst patients with or without concurrent MAFLD.
We retrospectively identified all non-cirrhotic CHB patients at a tertiary Australian centre. Patients were followed up from first review between 01/01/2010–31/12/2016 until 31/08/2020 or loss-to-follow-up. Steatosis was diagnosed radiologically (diffusely increased echogenicity on ultrasound) and MAFLD was diagnosed using new criteria. 4 Transient elastography was performed by a certified operator in accordance with best clinical practice for quality and probe selection. 5
The median of ≥10 successful measurements was recorded. Multivariable general linear mixed-effects regression (random intercept at the patient level) was used to model LSM, BMI, ALT and viral load using optimal order polynomial time covariates (by Akaike’s information criterion) and time interaction terms. Each model was controlled for baseline age, sex, LSM, ALT, BMI, viral load, antiviral status, prior antiviral exposure and metabolic risk factors. Analysis was performed in Stata/IC 16.1 (StataCorp LP, USA, 2020).
Of 660 patients (median follow-up 6.0 years; IQR 4.1–8.3), 172 (26%) had concurrent MAFLD. Data comprised 1,997 LSM, 10,647 ALT and 8,223 DNA measurements. MAFLD patients were of similar age (median 45.5 vs 43.0 years, P=0.09), but were more commonly male (65% vs 44%, P<0.001). There was no significant difference (MAFLD vs non-MAFLD) in the proportion who were HBeAg positive (19% vs 25%, P=0.21), on antiviral therapy at baseline (7% vs 12%, P=0.18) or commenced on therapy during follow-up (25% vs 29%, P=0.49). Most (61%) were never on antiviral therapy. Interestingly, we found that although average LSM was 15% higher initially (95% CI 7–24%, P<0.001) amongst patients with concurrent MAFLD, this improved over time such that there was no significant difference after 2–3 years (Figure 1). Although LSM is known to be elevated in obesity, the trajectory of LSM in MAFLD patients did not appear to correlate with BMI, but instead mirrored the trajectory of HBV DNA and ALT, reflective of the underlying virological and biochemical response. On average, DNA was 35% lower in MAFLD patients (95% CI 14–51%, P=0.003), and this difference remained over time. ALT decreased in all patients over time but was 13% higher in MAFLD patients on entry (95% CI 4–23%, P=0.003), and remained higher throughout follow-up. Of note, a similar degree of improvement in LSM was not observed in patients without MAFLD, despite no difference in the rate of change in ALT or viral load between MAFLD and non-MAFLD patients.
Perhaps the most striking finding of these data is the lack of any apparent worsening of LSM on average over time—in fact, LSM improved in MAFLD patients, with similar findings when analysing only untreated patients (data not shown). However, our cohort was substantially different to the cohort by Mak et al. as we included all CHB patients, the majority of whom had elevated HBV DNA and ALT throughout follow-up. This perhaps suggests viraemia and/or raised ALT has a more pronounced influence on LSM, and that the effect of steatosis is unmasked only after viral and biochemical suppression. Another explanation could be the influence of improved lifestyle changes not captured in our data, however BMI did not substantially change over time. Although we were limited by the lack of CAP measurements to quantify steatosis degree, ultrasound is less sensitive for steatosis so the patients in our cohort were likely to have moderate-to-severe steatosis. Thus, our results are markedly different from Mak et al. who showed the rate of fibrosis progression appeared highest in patients with persistent severe steatosis (41%) and new onset severe steatosis (35%). However, data for milder degrees of steatosis were not provided. It remains speculative whether the negative effects of steatosis on liver fibrosis outweigh the apparent protective effects, since concurrent steatosis attenuates viraemia,6,7
accelerates HBsAg seroclearance,1and possibly even improves the rate of response to antiviral therapy.8 Additional prospective studies are required to further clarify the unique relationship between steatosis and CHB.