Table 3.

Indirect comparisons calculated from the consistency equation

Ref	Plac	Study per comparison	Parameters	Consistency OR	Model 2 95% CI	Inconsistency OR	Model 3 95% CI
ASMQ	AQSP	2	$d_{27}=d_{17}-d_{12}$	0.63	0.26–1.52	0.67	0.22–1.97
DHAPa	AQSP	3	$d_{2,12}=d_{1,12}-d_{12}$	2.20	1.21–3.96	1.96	0.89–4.30
ASSP	AQSP	3	$d_{2,11}=d_{1,11}-d_{12}$	1.55	0.76–3.21	1.03	0.26–4.15
ASAQCPH	ASAQ	1	$d_{34}=d_{14}-d_{13}$	1.38	0.26–7.41	1.75	0.23–8.69
ASMQ	ASAQ	1	$d_{37}=d_{17}-d_{13}$	1.07	0.47–2.44	1.15	0.44–3.00
ASSMP	ASAQ	2	$d_{3,10}=d_{1,10}-d_{13}$	1.03	0.42–2.51	1.51	0.46–5.01
ASCD	ASAQ	2	$d_{36}=d_{16}-d_{13}$	0.52	0.23–1.15	0.69	0.25–1.94
ASSP	ASAQ	7	$d_{3,11}=d_{1,11}-d_{13}$	1.21	0.64–2.27	0.90	0.24–3.41
AQSP	ASAQ	4	$d_{32}=d_{12}-d_{13}$	0.78	0.41–1.46	0.87	0.38–2.04
DHAPa	ASAQ	6	$d_{3,12}=d_{1,12}-d_{13}$	1.70	1.10–2.64	1.72	1.01–3.07
DHAP	ASCD	1	$d_{6,12}=d_{1,12}-d_{16}$	3.25	1.46–7.25	3.17	1.21–8.23
ASSP	ASCD	1	$d_{6,11}=d_{1,11}-d_{16}$	2.31	0.95–5.61	2.49	0.78–7.86
DHAP	ASMQ	0	$d_{7,12}=d_{1,12}-d_{17}$	1.59	0.62–4.02	1.45	0.50–4.38
DHAP	ASPY	0	$d_{9,12}=d_{1,12}-d_{19}$	1.36	0.38–4.87	1.33	0.33–5.35
ASATPG	AL	0	$d_{15}=d_{35}-d_{13}$	3.40	0.42–27.31	3.10	0.36–26.62

The third column is the observed number of trials for each comparison. 0 comparison means the trial does not exist in the data set but with A one can estimate the treatment difference. $d_{kl}=d_{bl}-d_{bk}$ is the consistency equation used to derive all indirect comparisons

Ref reference group, Plac denotes the comparator. OR odds ratio is the overall effect estimated from the entire network, CI confidence interval

^aMeans the differences are significant. There was not enough evidence to support the higher efficacy of DHAP compared to ASMQ and ASPY, and there were insufficient data to compare ASATPG to AL. For loops having at least 3 treatments, the variance between treatments effects accounted for the correlation (equals to 0.5) between any pairwise contrasts. DHAP was more efficacious than AQSP and ASAQ