Meta-analysis of duloxetine vs. pregabalin and gabapentin in the treatment of diabetic peripheral neuropathic pain

Table 2 Random-effects pooled results: duloxetine vs. placebo.

Outcome	Treatment effect (θ)	95% CI for θ	U-test (p-value)	τ ²	NNT/NNH (95% CI)
*Efficacy*
Reduction in 24-hour pain intensity	-1.128	(-1.364; -0.891)	<0.001	0	-
Response	0.856	(0.628; 1.085)	<0.001	0	5(3;7)
PGI	-0.756	(-1.004; -0.508)	<0.001	0	-
*Tolerability*
Premature study discontinuation due to:
- Lack of efficacy	-0.962	(-1.800; -0.124)	(0.024)	0
- Adverse events	1.077	(0.663; 1.490)	<0.001	0	11 (7; 23)
- Other	-0.278	(-0.636; 0.079)	(0.127)	0
Diarrhoea	0.233	(-0.436; 0.903)	(0.307)	0.307
Dizziness	0.817	(0.398; 1.235)	<0.001	0
Headache	0.468	(0.090; 0.845)	(0.015)	0
Nausea	1.306	(0.942; 1.669)	0.039	0.039
Somnolence	1.472	(1.044; 1.900)	<0.001	0

Note:
θ is absolute difference for 24-hour pain intensity.
θ is log-odds ratio for Response, PGI and all tolerability analyses.
τ ² is between-study heterogeneity.
NNT/NNH: Number needed to treat, number needed to harm. NNTs were calculated for response rate and NNHs were calculated for discontinuation due to adverse events.

ISSN: 1471-2377