Myostatin serves as a negative feedback molecule during the processes of muscular growth and regeneration in order to limit the final muscle mass. In this model, myostatin is switched on only during muscle growth or regeneration. In the Pompe disease patients included in this study, suggesting a low level of muscle growth or regeneration. After ERT completion, myostatin levels in the Pompe disease patients increased significantly. This elevation in myostatin may have occurred as a signals of muscle regeneration. ERT stops the process of muscular destruction in order that new muscle fibers can be generated. Since myostatin negatively regulates human myoblast proliferation, antagonism of myostatin may enhance the therapeutic effect of ERT in those patients who respond poorly. Follistatin was originally found in the ovarian follicular fluid, and broader expression was later shown in the reproductive, endocrine, digestive, and neurological systems of humans. Follistatin binds to members of the transforming growth factor-b superfamily, and exerts an inhibitory effect on these growth factors. The follistatin reference value for a term neonate is 0.4360.02 ng/ml, which is correlated with fat mass and gestational age. In our study, baseline serum levels of follistatin were higher than normal for both the Pompe disease group and the control group, and increased at follow-up. A possible explanation could be high inter-individual variation resulting from a small sample size. In addition, because follistatin augments adipogenesis, higher adipose tissue growth might correlate with higher follistatin serum levels. Follistatin is a cytokine that does not originate in muscle; we therefore regarded it as a control marker. This study has some limitations. Due to the rarity of the disease, the sample was small and heterogeneous and duration of follow-up was short. However, we have still obtained positive results from a small sample size. By recruiting more patients for a larger sample size, it should be possible to obtain more detailed information in the future through subgroup analyses, e.g. by looking at IOPD versus LOPD or good responders versus poor responders. In addition, due to ethical considerations, the present study does not include a Pompe disease cohort that did not receive ERT. Therefore, it is not known if the effects are due to the treatment or related to the disease’s natural course. Furthermore, assessments of muscle 
mass and strength may be needed to delineate the correlation between muscle regeneration and serum marker levels. A future longitudinal study is also needed to analyze dynamic changes in these markers over the disease course. In conclusion, serum IGF-1 levels were significantly lower in Pompe disease patients relative to normal control subjects, and serum myostatin also showed a trend for lower levels.