Table 2.

Summary of methodology used in the meta-analyses

	Cochrane	Winblad	Doody	PenTAG	IQWiGa
Selection of studies
Included studies	MRZ-9605	MRZ-9605	MRZ-9605	MRZ-9605	MRZ-9605
	MEM-MD-01	MEM-MD-01	MEM-MD-01	MEM-MD-01	MEM-MD-01
	MEM-MD-02	MEM-MD-02	MEM-MD-02	MEM-MD-02	MEM-MD-02
	Lu-99679	Lu-99679	Lu-99679	MEM-MD-12b	Lu-99679
	MEM-MD-10	MEM-MD-10	MEM-MD-10		MEM-MD-10
	MEM-MD-12	MEM-MD-12	MEM-MD-12		MEM-MD-12
					Lu-10116
Use of posthoc analysis of subgroups	No	Yes	No	No	Yes
Selection of patients
Includes patients across the licensed population for memantine	Yes	Yes	Yes	No (exclusion of MMSE 14–19)	Yes
Inclusion of off-label patients	Yes (MMSE >19)	No	Yes (MMSE >19)	No	No
Analysis strategy
Monotherapy and combination studies	Grouped	Grouped	Separated and grouped	Separated	Separated and grouped
Moderately severe to severe AD and mild to moderate AD	Separated	Grouped	Separated and grouped	Separated—only moderately severe to severe	Grouped
Analysis method
Management scores (MD or SMD)	MD	SMD	SMD	MD	MD and SMD if necessary
Management of missing data (OC or LOCF)	LOCF	OC (plus LOCF supportive analysis)	LOCF (plus OC supportive analysis)	ITT, as reported in individual studies	ITT, as reported in individual studies

AD Alzheimer’s disease, ITT intent to treat, LOCF last observation carried forward, MD mean difference, OC observed cases, SMD standardized mean differenc

^aExcluded severe patients (identified as those having a baseline MMSE below 10) from study MEM-MD-02

^bIncluded in original PenTAG analysis but excluded here to ensure consistency between monotherapy and combination analyses