Objective To evaluate the outcomes selected in the randomized controlled trials（RCTs） on traditional Chinese medicine（TCM） for acute myocardial infarction（AMI）. Methods Four Chinese databases（CNKI

VIP

Wanfang

and Sinomed） and three English databases（Cochrane Library

Embase and Medline） were searched for RCTs on TCM for treatment of AMI. The predefined inclusion and exclusion criteria were followed for literature screening and selection. Cochrane risk of bias tool was used for methodological evaluation. The outcomes in RCTs were categorized and calculated for frequency. Results Thirty-nine out of 11?277 papers were included. All the papers reported on both effectiveness and safety outcomes. The low/unclear/high risk ratio of each bias item was as follows: random sequence generation（17/18/4）

allocation concealment（3/36/0）

blinding of patients and personnel（3/36/0）

blinding of outcome assessors（4/35/0）

incomplete outcome data（39/0/0）

selective reporting（0/39/0）

and other bias（including interest conflict and sample size calculation）（0/39/0）. The outcomes in 39 RCTs were mainly categorized into 8 groups

with highest frequency of cardiac indexes（107 times）

followed by comprehensive metabolic panel indicators（60 times）

clinically important outcomes（57 times）

safety outcomes（13 times）

electrocardiogram measurements（10 times）

quality of life（5 times）

and angina indicators（4 times）

and others（16 times）. A total of 118 outcomes were reported

and the average number was six per RCT

ranging from 1 to 14. Conclusion The overall risk of bias is unclear for RCTs on TCM in the treatment of AMI. And there are problems in setting outcomes

including unclear definition of primary and secondary outcomes

overuse of surrogate outcomes

lack of endpoint outcomes

and unbalanced allocation of different outcomes

all of which need to be solved in future researches.