Table 6

Effect estimates from multiple-treatment meta-analysis (MTM) compared to direct and indirect estimates, based on node-splitting [23]

Outcome

MTM-estimate (95% Crl)

Direct estimate (95% Crl)

Indirect estimate (95% Crl)

P-value for inconsistency


Diuretics vs. beta-blockers

Diabetes- incidence

1.09

0.88

1.31

0.13

(0.80 to 1.44)

(0.59 to 1.32)

(0.88 to 2.17)


Diuretics vs. ACE-inhibitors

Diabetes- incidence

1.43

1.54

1.25

0.24

(1.12 to 1.83)

(1.14 to 2.10)

(0.89 to 1.79)


Diuretics vs. placebo

All-cause mortality

0.88

0.89

0.80

0.29

(0.80 to 0.95)

(0.82 to 0.97)

(0.68 to 0.96)


Beta-blockers vs. ARB

Stroke

1.23

1.34

1.00

0.17

(0.96 to 1.49)

(1.03 to 1.74)

(0.69 to 1.41)


Beta-blockers vs. ARB

Diabetes- incidence

1.46

1.32

1.98

0.13

(1.15 to 1.98)

(0.97 to 1.82)

(1.26 to 3.40)


ACE- inhibitors vs. CCB

All-cause mortality

1.02

1.05

0.95

0.21

(0.95 to 1.10)

(0.96 to 1.13)

(0.82 to 1.09)


ACE- inhibitors vs. CCB

Heart fail ure

0.82

0.79

0.68

0.29

(0.69 to 0.94)

(0.67 to 0.95)

(0.52 to 0.90)


ACE-inhibitors vs. placebo

All-cause mortality

0.87

1.30

0.86

0.08

(0.79 to 0.96)

(0.82 to 2.14)

(0.77 to 0.95)


Diuretics and/or beta-blockers vs. placebo

All-cause mortality

0.82

0.57

0.85

0.06

(0.73 to 0.92)

(0.34 to 0.84)

(0.75 to 0.96)


CCB vs. ARB

Stroke

0.91

0.86

1.15

0.19

(0.75 to 1.11)

(0.69 to 1.05)

(0.78 to 1.70)


CCB vs. ARB

Diabetes- incidence

1.25

1.29

0.86

0.14

(1.02 to 1.56)

(1.07 to 1.71)

(0.49 to 1.49)


Table is limited to comparisons where Bayesian "P-values" for inconsistency were < 0.3. For a complete overview of MTM-, direct, and indirect estimates see Additional file 6

Fretheim et al. BMC Medicine 2012 10:33   doi:10.1186/1741-7015-10-33

Open Data