Letter: Association of serum growth differentiation factor 15 and liver-related outcomes in patients with MASLD—Authors' reply

We thank Wang et al.¹ for their interest and comments on our research.²

As Wang et al. noted, among the biopsy-MASLD cohort, the MASLD patients with a GDF15 >1.75 ng/mL (n = 153) were older (median [interquartile range] 67 [60–74] years versus 58 [47–67] years, p < 0.0001 by the Mann–Whitney U-test) and had a greater MASH frequency (Non-MASH/MASH 48/105 vs. 179/186, p < 0.0001 by the chi-squared test) than those without (n = 365). Age and the presence of MASH are well-known risk factors for hepatocellular carcinoma in patients with steatotic liver disease.³ The serum GDF15 level gradually increases with increasing age⁴ and is significantly greater in MASH patients.² However, both GDF15 concentrations and liver fibrosis status were independent risk factors for liver cancer occurrence according to multivariate analysis after considering age or the presence of MASH (Table 1). Furthermore, even when baseline age and MASH status were matched by propensity score in the biopsy-MASLD cohort, patients with GDF15 <1.75 ng/mL (n = 143) had a significantly greater incidence of liver cancer than did those without (n = 143) (8.1% vs. 0% at 5 years, p < 0.0001 by log-rank test). High-GDF15 patients have a significantly greater risk for liver cancer occurrence independent of age and the presence of MASH.

Second, among the biopsy-MASLD cohort, the MASLD patients with a GDF15 >1.75 ng/mL included more patients with metformin use (taking/not taking 44/109 vs. 36/329, p < 0.0001 by the chi-squared test). It has been reported that metformin use elevates serum GDF15 by approximately 1.4-fold.⁵ To exclude the effect of metformin use, we examined only patients without metformin use among the total cohort (a combination of the biopsy-MASLD cohort and the validation cohort). The incidence of liver cancer was significantly greater in patients with GDF15 >1.75 ng/mL (n = 160) than in those without (n = 454; 11.9% vs. 0% at 5 years, p < 0.0001 according to the log-rank test). In addition, the GDF15 concentration, as well as the Fib-4 index, was an independent risk factor for liver cancer occurrence according to multivariate analysis, even after metformin use was taken into account (Table 1). Although metformin use is associated with increased GDF15 levels, high-GDF15 patients are at high risk for liver cancer independent of metformin use.

Third, Wang et al. suggested the need to follow the changes in cardiometabolic risk factors and alcohol intake over time. This point is not limited to the present study but applies to any study analysing the occurrence risk of liver disease-related events in MASLD. Since the influence of these factors increases with the length of the follow-up period, future analyses may need to account for these aspects.

The GDF15 value itself is influenced by various factors, including the degree of liver fibrosis, age, presence of MASH and metformin use. Importantly, however, it can be used for risk stratification independently of these factors as a marker for predicting liver-related incidence, including liver cancer occurrence. However, prospective studies are needed to validate the predictive performance of GDF15 concentrations. To this end, we initiated the enrolment of a prospective cohort.

Shusuke Kumazaki: Writing – original draft; writing – review and editing; conceptualization. Hayato Hikita: Writing – original draft; writing – review and editing; conceptualization. Yuki Tahata: Writing – review and editing. Tetsuo Takehara: Writing – review and editing; conceptualization.

This study was partly supported by a Grant-in-Aid for Research from the Japan Agency for Medical Research and Development (JP23fk0210121 to T. Takehara) and a Grant-in-Aid for Scientific Research (JP23H02894 to H.H.) from the Japan Society for the Promotion of Science, Japan.

This article is linked to Kumazaki et al. papers. To view these articles, visit https://doi.org/10.1111/apt.18063 and https://doi.org/10.1111/apt.18150