Table 1

Subgroup meta-analysis of studies included in the quantitative analyses

Analysis	Number of RCTs	Meta-analysis				Heterogeneity	Trim and fill effect size (95% CI) (adjusted studies)
Analysis	Number of RCTs	SMD	95% CI		P value	I²	Trim and fill effect size (95% CI) (adjusted studies)
Main analysis	41	−0.946	−1.18	−0.71	<0.001	82.490	Unchanged
Depression classification
MDD*	20	−0.998	−1.39	−0.61	<0.001	84.746	−1.167 (−1.59 to −0.74)(2)
Depressive symptoms	21	−0.915	−1.21	−0.62	<0.001	78.974	Unchanged
Risk of bias
Low risk of bias	12	−0.666	−0.99	−0.34	<0.001	80.218	−0.367 (−0.72 to −0.01) (4, right of mean)
Some concerns	7	−0.829	−1.49	−0.16	0.015	86.926	−1.151 (−1.86 to −0.45) (2)
High risk of bias	22	−1.199	−1.59	−0.81	<0.001	81.484	Unchanged
Type of control condition
Usual care	22	−0.949	−1.25	−0.65	<0.001	82.951	Unchanged
Wait list	11	−1.238	−1.69	−0.79	<0.001	73.749	−1.008 (−1.43 to −0.59) (3, right of mean)
Health education	3	−1.123	−2.56	0.31	0.125	93.955	Unchanged
Intensity of exercise†
Light	2	−1.041	−2.53	0.45	0.170	88.867	N/A
Moderate	26	−1.132	−1.45	−0.81	<0.001	84.972	−1.403 (−1.80 to −1.00) (4)
Vigorous	10	−0.924	−1.47	−0.38	0.001	85.472	Unchanged
Exercise type†
Aerobic	30	−1.156	−1.46	−0.85	<0.001	81.392	−1.264 (−1.59 to −0.94) (2)
Resistance	7	−1.042	−1.87	−0.22	0.013	87.944	Unchanged
Mixed‡	10	−0.455	−0.80	−0.11	0.009	80.237	Unchanged
Group exercise†
Yes	27	−0.848	−1.10	−0.59	<0.001	80.503	Unchanged
No	13	−0.802	−1.24	−0.36	<0.001	83.312	−0.881 (−1.32 to −0.44) (1)
Supervision†
Yes§	40	−1.026	−1.28	−0.77	<0.001	83.926	−1.115 (−1.38 to −0.85) (2)
No	6	−0.451	−0.91	0.00	0.052	63.553	−0.122 (−0.63 to 0.39) (3, right of mean)
Type of supervision†
Exercise professional	18	−0.984	−1.34	−0.63	<0.001	78.879	Unchanged
Other professional/student	14	−1.278	−1.81	−0.74	<0.001	85.356	Unchanged
Sample size intervention arm†
n≥25	12	−0.532	−0.81	−0.26	<0.001	81.445	Unchanged
n<25	35	−1.166	−1.49	−0.84	<0.001	81.239	−1.370 (−1.72 to −1.02) (4)
n≥10	34	−0.868	−1.11	−0.63	<0.001	84.568	−0.913 (−1.16 to −0.67) (1)
n<10	13	−1.281	−1.87	−0.70	<0.001	75.817	−1.568 (−2.22 to −0.91) (2)

See online supplemental tables 3–5 for detailed categorisation for each study.
*Including three studies with MDD and dysthymia.67 81 82
†Data for a second intervention group within the same study were included for these analyses.
‡Mixed exercise=aerobic and resistance exercise.
§Two studies were partly supervised, see online supplemental table 5).
MDD, major depressive disorder; RCT, randomised clinical trials; SMD, standard mean difference.