Table 3.

Pooled results for the association of DAPK1 promoter hypermethylation with HSIL risk.

Comparisons	Studies (N)	Sample size (HSIL/controls)	Heterogeneity		Pmeta-regression	Modela	Effect size
			I2(%)	PQ-test			OR (95% CI)	P
Total	18	733/561	75	<0.001	–	R	7.62 (3.51–16.57)	<0.001
Ethnicity					0.610
Asian	6	304/211	72	0.003		R	7.76 (2.76–21.86)	<0.001
Caucasian	9	342/186	71	0.001		R	4.95 (1.50–16.31)	0.008
Other ethnicities	3	87/164	49	0.140		F	21.73 (7.41–63.70)	<0.001
Source of controls					0.487
Healthy	11	409/275	78	<0.001		R	10.53 (3.40–32.59)	<0.001
Non-healthyb	7	324/286	71	0.004		R	5.10 (1.67–15.54)	0.004
Study quality					0.004
High (>11)	11	420/248	5	0.761		F	8.09 (4.71–13.88)	<0.001
Low (≤ 11)	7	313/313	11	0.312		F	2.10 (1.35–3.27)	<0.001
Primer set					0.90
1	10	454/402	81	<0.001		R	8.14 (2.68–24.69)	<0.001
2-7	7	257/135	67	0.006		R	6.99 (1.87–26.08)	0.004

N, number; F, fixed-effects model; R, Random-effects model; HSIL, high-grade squamous intra-epithelial lesion.

^aIf significant heterogeneity was found (I² ≥ 50% or P_Q−test ≤ 0.1), a random-effects model with the inverse variance method was applied.

^bNon-healthy controls included autologous controls, controls with benign gynecological diseases, and mixed controls.

Bold values indicate significant results with P < 0.05 in meta-regression.