TABLE 3.

Summary risk estimates of the association between the total duration of breastfeeding and ovarian cancer risk: longest compared with shortest durations¹

	Studies	Summary RR (95% CI)	Q statistic	I2	Ph2	Ph3
	n			%
Overall	26	0.65 (0.55, 0.78)	70.26	64.4	<0.001	—
Subgroup analyses
Study design						0.511
Cohort studies	3	0.80 (0.66, 0.98)	1.69	0	0.429
Case-control studies	23	0.63 (0.52, 0.78)	67.25	67.3	<0.001
Exposure assessment						0.790
Trained interviewer	14	0.61 (0.49, 0.76)	49.54	73.8	<0.001
Self-administered questionnaire	9	0.75 (0.63, 0.88)	9.33	14.3	0.315
Type of control subjects						0.185
Population based	14	0.57 (0.45, 0.71)	27.98	53.5	0.009
Hospital based	8	0.81 (0.53, 1.21)	31.11	77.5	<0.001
Study population						0.365
Asians	3	0.66 (0.43, 1.00)	1.36	0	0.505
Americans	11	0.55 (0.43, 0.71)	25.93	61.4	0.004
Europeans	9	0.81 (0.59, 1.10)	27.46	70.9	0.001
Cancer grading						0.291
Invasive	4	0.55 (0.36, 0.84)	7.95	62.3	0.047
Borderline	5	0.41 (0.28, 0.60)	1.67	0	0.797
Cancer histotype						0.258
Serous	6	0.75 (0.59, 0.96)	1.78	0	0.879
Mucinous	4	0.61 (0.19, 1.94)	12.04	75.1	0.007
Endometrioid	3	0.59 (0.35, 0.98)	2.64	24.4	0.267
Clear cell	1	0.24 (0.06, 0.97)	NA	NA	NA
Adjustment for confounders
Parity						0.318
Yes	21	0.68 (0.59, 0.82)	50.60	60.5	<0.001
No	5	0.53 (0.30, 0.94)	17.53	77.2	0.002
BMI						0.406
Yes	5	0.81 (0.67, 0.98)	3.44	0	0.486
No	21	0.63 (0.50, 0.78)	64.46	69.0	<0.001
OC use						0.428
Yes	16	0.62 (0.50, 0.77)	48.25	68.9	<0.001
No	10	0.73 (0.53, 1.00)	22.00	59.1	0.009
Smoking						0.521
Yes	6	0.58 (0.40, 0.85)	11.15	55.1	0.049
No	20	0.67 (0.55, 0.83)	59.07	67.8	<0.001

¹NA, not available; OC, oral contraceptive.

²P value for heterogeneity within each subgroup.

³P value for heterogeneity between subgroups with meta-regression analysis.