Table 3.

Quality assessment of trials included in the meta-analysis

Trial ID	Total score	Blinding	Outcome measure: blood pressure	Loss to follow-up	Funding source	Compliance; other chocolate diary
Scores given		1: choc + control blinded	1: Primary	1: < 20%	1: Sponsor not involved in data collection, analysis	1: Compl assessed; no sign diff between groups
		0.5: control blinded	0.5: Secondary	0: ≥ 20%	0: Sponsor involved	1: Compl assessed, sign diff included in analysis
		0: No blinding				0: Compl not assessed
Taubert et al. [10]	4	0: dark/white	1	1: none	1	1
Murphy et al. [15]	3.5	1: cocoa tablets	0.5	1: 9.5% (4 of 42)	0: supported by Mars	1: tablet count, 7 day weight food records x2
Engler et al. [11]	5	1: flav bar matched	1	1: none	1	1
Fraga et al. [12]	2.5	0: dark/white	1	1: 3.6% (1 of 28)	0: 3 authors from Mars	0.5: reported what consumed, no diary
Grassi et al. [13]	3.5	0: dark/white	1	1: none	0.5: unclear	1: daily diary
Grassi et al. [13]	3.5	0: dark/white	1	1: none	0.5: unclear	1: daily diary
Taubert et al. [16]	4	0: dark/white	1	1: none	1	1: food diary
Crews et al. [17]	2.5	0.5: flav drink; > 50% assumed correct group they were in	1	0.5: 11% (11 of 101)	0.5: independent industry research grant, supplier	0: not reported
Grassi et al. [18]	3.5	0: dark/white	1	1: none	0.5: unclear	1: diary
Muniyappa et al. [20]	3	1: flav drink; blinding assessed ok	1	0: 31% (9 of 29)	1	0: not reported
Davison et al. [21]	3	1: flav drink	1	0: 21% (14 of 65)	0: Mars Financial support	1: diet + background exercise
Davison et al. [21]	3	1: flav drink	1	0: 21% (14 of 65)	0: Mars Financial support	1: diet + background exercise
Shiina et al. [22]	2	0: dark/white	0	1: none	1	0
Ried et al. [23]	4.5	0.5: control blinded	1	1: 8% (3 of 39)	1	1: diary
Monagas et al. [24]	3.5	0: cocoa powder in milk/milk	0.5	1: none	1	1: diet monitoring (3-day food questionnaires x3)

All 15 trial arms were adequately randomised.

Choc, chocolate; compl, compliance; flav, flavanol; sign diff, significant difference;