Table 2.

Subgroup analyses of the association between vitamin D and OSA in this meta-analysis

Subgroup	No. of studies	Sample		Mean difference (95% CI)	P for Heterogeneity	I-squared Value	P Value between groups
		Experiment	Control
Ethnic
Caucasian	16	3547	1707	− 0.59[− 0.9, − 0.27]	0.0003	93%	P < 0.00001
Asian	10	662	382	− 1.28[− 1.92, − 0.64]	P < 0.0001	94%	P < 0.00001
BMI
Normal range	3	176	129	− 1.66[− 2.73, − 0.59]	0.002	93%	P < 0.00001
Overweight	4	1983	1186	− 1.33[− 2.55, − 0.1]	0.03	98%	P < 0.00001
Obesity	17	1981	634	− 0.57[− 0.93, − 0.21]	0.002	92%	P < 0.00001
PSG degree
I	19	3750	1833	− 0.55[− 0.85, − 0.26]	0.0002	93%	P < 0.00001
II	4	277	157	− 1.4[− 2.21, − 0.58]	0.0008	91%	P < 0.00001
III	3	182	99	− 1.93[− 2.28, − 1.58]	P < 0.00001	21%	0.28
Study quality
High/A	18	3640	1769	− 0.61[− 0.92, − 0.31]	P < 0.0001	93%	P < 0.00001
Medium/B	8	569	320	− 1.37[− 2.07, − 0.66]	0.0001	95%	P < 0.00001
Study latitude
Low latitude	4	336	117	− 1.24[− 2.25, − 0.23]	0.02	94%	P < 0.00001
Middle latitude	22	3873	1972	− 0.8[− 1.11, − 0.49]	P < 0.00001	95%	P < 0.00001

OSA Obstructive sleep apnea, BMI body mass index, PSG polysomnogram