Table 4

Subgroup analyses by study design and type of control

Subgroups

SBP

DBP

Mean difference in mmHg (95% CI)

P

I²

Mean difference in mmHg (95% CI)

P

I²

Study design

Crossover [10,12,13,18,20,21,23]

-4.91 (-7.94, -1.88)

0.002

80%

-3.24 (-5.36, -1.12)

0.003

67%

Parallel [11,15-17,21-23]

-2.01 (-3.42, -0.60)

0.005

9%

-1.20 (-2.29, -0.11)

0.03

12%

Parallel^a [11,15,17,21-23]

-0.43 (-2.80, 1.94)

0.72

0%

-0.25 (-1.83, 1.33)

0.75

0%

Type of control

Flavanol-free: white chocolate, milk, pills [11,15,17,20,21]

-4.21 (-6.65, -1.77)

0.0007

82%

-2.81 (-4.52, -1.10)

0.001

70%

Low-flavanol: drink, bar, tablet [10,13,16,18,22-24]

0.74 (-2.92, 4.40)

0.69

53%

-0.38 (-2.00, 1.24)

0.65

0%

SBP, systolic blood pressure; DBP, diastolic blood pressure; mmHg, millimetre mercury; CI, confidence interval; P, probability value (P < 0.05 statistically significant); I², heterogeneity test.

^aSensitivity analysis excluding trial [16], as this trial used a notably lower dose and longer duration than the other trials included in the meta-analysis.

Ried et al. BMC Medicine 2010 8:39   doi:10.1186/1741-7015-8-39