Chinese medicine and acupuncture are widely used in the treatment of myasthenia gravis (MG), but there is a lack of high-quality evidence to verify their effectiveness and safety. One of the reasons is the lack of recognized and unified outcomes in the randomized controlled trial (RCT) design, which leads to the inability to combine data in secondary studies such as meta-analysis. The construction of the core outcome set (COS) of traditional Chinese medicine (TCM) for MG is conducive to the design of clinical research and the accumulation of high-quality evidence.

To describe the current situation of outcome selection in RCTs of MG treated with TCM, and analyze the existing problems, then put forward suggestions, promoting the construction of a COS of TCM.

Electronic databases of CNKI, Wanfang Data, VIP, SinoMed, Medline, PubMed, Cochrane Library, and Web of Science were systematically searched for RCTs of MG treated with TCM from inception to April 1, 2022. Two researchers independently screened the literature, extracted data, classified the outcomes and counted their frequency of use, and divided the outcomes into different outcome domains according to the functional attributes. Then, we analyzed the current status and existing problems in the selection of outcomes, and put forward corresponding suggestions.

A total of 186 RCTs were included, involving 71 outcomes, with a frequency of use of 724 times in total. The roughly classified seven categories of all outcomes with frequency of use are as follows: MG severity scale (133 times), MG quality of life scale (20 times), TCM syndrome score (70 times), safety outcomes (66 times), blood biochemical outcomes (224 times), muscle strength changes (three times), and others (208 times). According to the functional attributes, the outcomes can be classified into seven outcome domains. A single study selected at least one and at most 12 outcomes. The most frequently used outcomes include effective rate, MG absolute and relative score, and TCM syndrome score. Among the included RCTs, only two reported the primary and secondary outcomes explicitly. The number of follow-up time point ranged from two to seven, and most RCTs chose three evaluation time points. Most RCTs did not report the use of blinding in the process of outcome evaluation.

At present, there are many problems in the selection of outcomes in RCTs of MG treated with TCM, including unclear distinction between primary and secondary outcomes, lack of internationally recognized outcomes, insufficient attention given to safety, large inter-study differences in the number of outcomes and evaluation time points, evaluation bias, and failure to include health economic outcomes, and non-standardized TCM curative effect evaluation. According to the existing problems, we proposed the following specific recommendations: promoting the establishment of a COS for TCM clinical research, selecting internationally recognized outcomes (including safety outcomes) with clear defining of the primary and secondary ones, selecting the appropriate treatment cycle and evaluation time point, completing clinical research registration, using blinding method, and standardizing the report of study.