Table 2.

Differences in the main findings and study characteristics among previous systematic reviews and meta-analyses and the current systematic review and meta-analysis of clinical trials on calcium supplementation and the risk of cardiovascular disease.

		2010, Bolland et al. [4]	2011, Bolland et al. [5]	2013, Mao et al. [28]	2015, Lewis et al. [6]	2016, Chung et al. [10]	Current Meta-Analysis
Conclusion on Calcium Supplementation and Risk of CVD		Increase	Increase	Might Increase	Not Increase	Not Associated, Small Risk and Not Clinically Important, if Any	Increase
Main Findings: RR (95% CI), Number of Included Trials (Reference No.) *, Interpretation in Each Article
Myocardial Infarction (MI)		-1.27 (1.01–1.59) -7 (16, 19, 21, 22, 25, 26) -Increased risk	-1.24 (1.07–1.45) -8 (16, 19, 21, 22, 24, 25, 26) -Increased risk	-1.28 (0.97–1.68) -8 (16, 19, 21, 22, 24, 25, 26) -Non-significantly increased risk	-1.08 (0.93–1.25) -8 (19, 21, 24, 25, 26, 2004 Larsen, 2012 Sambrook) -No increased risk	n.a.	-1.25 (1.07–1.45) -9 (16, 18, 19, 21, 22, 24, 25, 26) -Significantly increased risk
Stroke		-1.12 (0.92–1.36) -8 (1993 Reid, 16, 19, 20, 21, 25, 26) -No increased risk	-1.15 (1.00–1.32) -9 (1993 Reid, 16, 19, 20, 21, 24, 25, 26) -Increased risk	-1.14 (0.90–1.46) -Not specified -Non-significantly increased risk	n.a.	n.a.	-1.13 (0.97–1.31) -12 (14–18,20–22,24–26) -Non-significantly increased risk
Cardiovascular disease (CVD): coronary heart disease (CHD) plus stroke		-1.12 (0.97–1.30) -8 (1993 Reid, 16, 19, 21, 22, 25, 26) -No increased risk (composite end point of MI, stroke, and sudden death)	-1.15 (1.03–1.27) -10 (1993 Reid, 16, 19, 20, 21, 22, 24, 25, 26) -Increased risk (MI or stroke)	-1.16 (0.97–1.40) -Not specified -Non-significantly increased risk (major CV events)	-1.02 (0.96–1.09) -6 (19, 24, 25, 2004 Larsen, 2012 Sambrook) -No increased risk (CHD)	-No meta-analysis performed -4 (2011 Lewis, 24, 25, 26) -No statistically significant difference	-1.15 (1.06–1.25) -14 (14–26) -Significantly increased risk (CHD plus stroke)
Selection Criteria	Type of Trials	Randomized, double-blind, placebo-controlled trials with >1 year of trial duration	Randomized, double-blind, placebo-controlled trials	Randomized, placebo-controlled trials with at least one year of follow-up	Randomized placebo-controlled trials and open-label trials	Randomized controlled trials	Randomized, double-blind, placebo-controlled trials
	Inclusion of Trials with No Use of Placebos	No	No	No	Yes	No	No
	Study Participants	Participants aged >40 years	Participants aged >40 years	Not described	A mean cohort age >50 years	Generally healthy adults	Adults
	Inclusion of Unpublished Data	Yes	Yes	Yes	Yes	No	Yes
Subgroup meta-analysis		Type of endpoints (MI, stroke, and death)	Type of endpoints (MI, stroke, and MI + stroke)	Type of endpoints (major cardiovascular events, MI, stroke) and sex	Type of endpoints (MI, angina pectoris and acute coronary syndrome, chronic coronary heart disease, and all-cause mortality)	n.a.	Type of endpoints (MI, angina pectoris, coronary revascularization, stroke, coronary heart disease, CVD), low risk of bias, population, age, gender, region, dosage, duration of supplementation, and data source (published or unpublished)
Funding Source		The Health Research Council of New Zealand and the University of Auckland School of Medicine Foundation	The Health Research Council of New Zealand and the University of Auckland School of Medicine Foundation	National “Eleven Five” “Significant new drugs creation” special science and technology major, a major national science and technology projects, etc.	Not described	National Osteoporosis Foundation through Pfizer Consumer Healthcare in U.S.	None

* Ref. [15]—1995 Reid et al. [16]—1999 Baron et al. [18]—2005 Brazier et al. [19]—2006 Prince et al. [20]—2007 Bonnick et al. [21]—2007 Lappe et al. [22]—2008 Reid et al. [24]—2011 Bolland et al. (WHI data) (=2006 Jackson et al.), [25]—2012 Avenell et al. [26]—2013 Bolland et al. (=2006 Reid et al.); Bolland et al.’s meta-analysis included Grant et al.’s trial, which is the first report of the RECORD trial. Ref. [25]—2012 Avenell et al. is the long-term follow-up report for the same trial. In Bolland et al.’s meta-analyses in 2010 and 2011, Grant et al.’s trial (=Ref. [25] 2012 Avenell et al.) was counted as two trials because it reported two findings from the RECORD trial calcium vs. placebo arms and calcium plus vitamin D vs. placebo plus vitamin D arms. 2004 Larsen et al. (open-label trial: a non-placebo control group used), 2012 Sambrook et al. (open-label trial a non-placebo control group used); n.a., not available.