Comment on Chang et al.: The safety and efficacy of minimally invasive discectomy: a meta-analysis of prospective randomised controlled trials

To the Editor, We read with great interest the article by Chang et al. [1] entitled “The safety and efficacy of minimally invasive discectomy: a meta-analysis of prospective randomised controlled trials” published in International Orthopaedics. In this meta-analysis, Chang et al. compared the safety and efficacy of minimally invasive discectomy with standard discectomy. Liu et al. [2] provided some questions regarding this article. However, after thoroughly reading the article and analysing the included studies, we found other problems that should be of great concern. First, Chang et al. stated that only randomised controlled trials comparingminimally invasive discectomywith standard discectomy for lumbar disc herniation were included. However, this meta-analysis incorrectly included a nonrandomised controlled trial. Michael et al. [3] reported their study was a prospective controlled trial rather than a randomised controlled trial. Second, repetitive data existed in this meta-analysis. Arts et al. [4–6] performed a randomised controlled trial and reported their study three times. The differences among the three articles were the periods of follow-up and variables. However, Chang et al. included them as three randomised controlled trials and repeatedly pooled them in Figure 4 and Figure 6. Obviously, this was not reasonable. Third, the number of patients was not in accordance with the original study in one study. Michael et al. [3] reported that 20 patients were actually included in their study, ten patients in the minimally invasive lumbar microdiscectomy group and ten patients in the traditional open discectomy group. However, Chang et al. inputted the number of patients as 30 respectively in the two groups. Fourth, methodological quality of all 16 included studies was assessed as either low risk or unclear risk, which was incredible. In fact, the assessment of methodological quality was inappropriate in this meta-analysis. For example, Huang et al. [7] did not report the allocation concealment and blinding of participants and personnel in their study. However, Chang et al. incorrectly assessed the selection bias and performance bias as low risk. In conclusion, the results of the meta-analysis by Chang et al. should be interpreted with caution. They should carefully and scientifically perform the meta-analysis and solve all these questions in their article.