Table 2

Study quality—risk-of-bias assessment

Study	Sequence generation	Allocation concealment	Blinding	Adequate handling of incomplete outcome data	Total drop out from drug assignment	No selective reporting	Groups comparable at baseline	Adequate power	Funder
Dapagliflozin
Bailey et al8	Yes	Yes	Yes (double blind)	Yes—last observation carried forward	12%	Yes	Yes	Yes—0.5% HbA1c difference detectable	Astra-Zeneca and Bristol-Myers-Squibb
Bolinder et al9/Ljunggren et al10	Yes	Yes	Yes (double blind)	Yes—last observation carried forward	7.1%	Yes	Yes	Unclear for primary endpoint, 2% BMD difference detectable	Astra-Zeneca and Bristol-Myers-Squibb
Nauck et al11	Yes	Yes	Yes (double blind and double dummy)	Yes—last observation carried forward	22.1%	Yes	Yes	Yes—0.35% HbA1c difference detectable	Astra-Zeneca and Bristol-Myers-Squibb
Rosenstock et al12	Not reported	Not reported	Yes (double blind)	Not reported	8% at 24 weeks, 19% at 48 weeks	Yes	Unclear	Not reported	Astra-Zeneca and Bristol-Myers-Squibb
Strojek et al13	Yes	Yes	Yes (double blind and double dummy)	Yes—last observation carried forward	8.5%	Yes	Yes	Yes—0.5% HbA1c difference detectable	Astra-Zeneca and Bristol-Myers-Squibb
Wilding et al14	Not reported	Not reported	Yes (single blind during lead in, double blind during study)	Yes—last observation carried forward	7%	Yes	Partially; matched for patient demographics, not for prior medications	Yes—0.5% HbA1c difference detectable	Astra-Zeneca and Bristol-Myers-Squibb
Wilding et al15	Yes	Yes	Yes (double blind and double dummy)	Yes—last observation carried forward	11% at 24 weeks, 15.5% at 48 weeks	Yes	Yes	Yes—0.5% HbA1c difference detectable	Astra-Zeneca and Bristol-Myers-Squibb
Canagliflozin
Rosenstock et al16	Not reported	Not reported	Yes (double blind)	Yes—last observation carried forward	10.9%	Yes	Yes	Yes—0.55% HbA1c difference detectable	Janssen Global Services

BMD, bone mineral density.