Study	Reason for exclusion
Du 2003	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were alternately allocated to the intervention and the control group by their visiting sequence.
Fei 2004	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The allocation procedure was described as that the patients decided which treatment would be accepted by themselves.
Feng 1998	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were alternately allocated to the intervention and the control group by their visiting sequence.
Ge 2004	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The trialist decided which remedy would be allocated to the patient in a non‐random way.
Hu 2005	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. Chinese medicine was used by in‐patients and western medicine was used by out‐patients.
Huang 2002	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The trialist decided which remedy would be allocated to the patient in a non‐random way.
Huang 2004	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were alternately allocated to the intervention and the control group by their visiting sequence.
Huang 2005	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were allocated to the intervention and control group by their registration sequence, depending on the even or odd number.
Li 2003a	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. Which remedy would be taken was decided by the patients.
Liang 2000	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were alternately allocated to the intervention and the control group by their visiting sequence.
Liao 1999	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The trialist decided which remedy would be allocated to the patient in a non‐random way.
Liao 2000	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The trialist decided which remedy would be allocated to the patient in a non‐random way.
Lv 1997	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were alternately allocated to the intervention and the control group by their visiting sequence.
Nie 1997	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were alternately allocated to the intervention and the control group by their visiting sequence.
Shi 2002	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were allocated to the intervention and control group by their registration number at libitum.
Sun 1999	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were alternately allocated to the intervention and the control group by their visiting sequence.
Tang 2003	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were alternately allocated to the intervention and the control group by their visiting sequence.
Wang 1999a	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The trialist decided which remedy would be allocated to the patient in a non‐random way.
Wang 2001a	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were allocated to the intervention group and the control group by the method of self‐drawing lots.
Wang 2004a	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were alternately allocated to the intervention and control group by their registration sequence.
Xiao 2004	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. Which remedy would be taken was decided by the patients.
Xing 2003	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were alternately allocated to the intervention and the control group by their visiting sequence.
Xu 2003a	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were alternately allocated to the intervention and the control group by their visiting sequence.
Yang 2004a	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were allocated to the intervention and the control group by the date of visit.
Yao 1997	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were alternately allocated to the intervention and control group.
Yuan 2002	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The trialist decided which remedy would be allocated to the patient in a non‐random way.
Zeng 1998	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were allocated to the intervention and the control group by the method of drawing lots.
Zhang 2000	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were alternately allocated to the intervention and the control group by their visiting sequence.
Zhang 2001	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were alternately allocated to the intervention and the control group by their visiting sequence.
Zhang 2002	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were alternately allocated to the intervention and the control group by their visiting sequence.
Zhang 2003b	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The trialist decided which remedy would be allocated to the patient in a non‐random way.
Zhang 2004a	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. Which remedy would be taken was decided by the patients.
Zhang 2005	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were allocated to the intervention and control group by their registration sequence depending on the even or odd number.
Zhao 2001	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were alternately allocated to the intervention and the control group by their visiting sequence.
Zhao 2003	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were alternately allocated to the intervention and the control group by the date on which they visited the doctor.
Zhao 2004a	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were alternately allocated to the intervention and the control group by the date on which they visited the doctor.
Zhao 2004b	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were alternately allocated to the intervention and the control group by the date on which they visited the doctor.
Zhao 2005	Authors described their study as a randomised controlled trial. We interviewed the original author(s) by telephone to identify which method of randomisation was used. According to authors' description, we decided that this trial was of a non‐randomised nature. The patients were alternately allocated to the intervention and the control group by the date on which they visited the doctor.