Table 2.

Overview of the assessment of the ROBINS-I risk-of-bias tool and the certainty of evidence of the included studies

Year	Author	Bias due to confounding	Bias in classification of interventions	Bias in participant selection	Bias due to missing data	Bias in measurement of the outcome	Bias in selection of reported result	Overall risk of bias	Certainty of the evidence (GRADE)
2024	Cao et al.	Low	Low	Moderate	Low	Low	Low	Moderate	High certainty
2013	Cummings et al.	Low	Low	Low	Moderatea	Moderate	Low	Moderate	Low certainty
2023	He et al.	Low	Low	Low	Low	Moderate	Low	Moderate	Low certainty
2020	Kanellopoulos	Low	Low	Low	Low	Moderate	Low	Moderate	Low certainty
2024	Kanellopoulos	Low	Low	Moderate	Low	Moderate	Low	Moderate	Low certainty
2024	Kanellopoulos et al.	Low	Low	Seriousb	Low	Moderate	Low	Serious	Very low certaintyc
2025	Khoramnia et al.	Low	Low	Low	Low	Low	Low	Low	High certainty
2025	Luo et al.	Low	Low	Seriousb	Low	Moderate	Low	Serious	Very low certaintyc
2025	Luo et al.	Low	Low	Seriousb	Low	Moderate	Low	Serious	Very low certaintyc
2012	Schumacher et al.	Low	Low	Moderate	Low	Moderate	Low	Moderate	Low certainty
2025	Sitto et al.	Low	Low	Low	Low	Moderate	Low	Moderate	Low certainty
2025	Tan et al.	Low	Low	Low	Low	Moderate	Low	Moderate	Low certainty
2025	Thananjeyan et al.	Low	Low	Low	Low	Moderate	Low	Moderate	Low certainty
2024	Yuan et al.	Low	Low	Low	Low	Low	Low	Low	High certainty

^aFor the 6-month follow-up

^bExclusion of patients with adverse events

^cDowngraded owing to serious overall risk of bias

NA, not applicable; NR, not reported