Table 4.
Meta-analysis of the association between various birth size measurements and percent breast density, stratified by potential sources of between-study heterogeneity
| Perinatal factor | Number of studiesa | Average relative changeb (95 % CI) | z-score p value | I 2 statistic (%) | References |
|---|---|---|---|---|---|
| Birth size measurements: | |||||
| Birthweight | |||||
| Overallc | 9 | 1.59 (1.58–1.59) | <0.001 | 100.0 | [21, 22, 25, 26, 33, 35, 49, 50]; present study |
| Using Andersen et al. [35] OR2d | 9 | 1.59 (1.58–1.59) | <0.001 | 100.0 | |
| Excluding present study | 8 | 1.63 (1.62–1.63) | <0.001 | 100.0 | |
| Menopausal status | |||||
| Pre-menopausal women | 3 | 1.03 (1.02–1.04) | <0.001 | 38.1 | [21, 22]; present study |
| Post-menopausal women | 3 | 1.72 (1.71–1.72) | <0.001 | 99.7 | [21, 22, 26] |
| Source of birthweight data | |||||
| Self-/parent report | 4 | 0.99 (0.99–1.00) | 0.117 | 91.0 | [21, 22, 35, 50] |
| Hospital records | 5 | 1.67 (1.66–1.67) | <0.001 | 10.0 | [25, 26, 33, 49] present study |
| Breast density assessment method | |||||
| Radiographer-assessed | 4 | 1.05 (1.00–1.10) | 0.066 | 49.7 | [25, 33, 35, 49] |
| Computer-assisted | 5 | 1.59 (1.58–1.59) | <0.001 | 100.0 | [21, 22, 26, 50] present study |
| Restricted to hospital records | |||||
| Radiographer-assessed | 3 | 1.07 (1.02–1.13) | 0.013 | 33.3 | [25, 33, 49] |
| Computer-assisted | 2 | 1.67 (1.66–1.67) | <0.001 | 100.0 | [26]; present study |
| Quality scoree | |||||
| Highest tertile (≥50) | 3 | 1.67 (1.66–1.67) | <0.001 | 100.0 | [26, 35]; present study |
| Middle tertile (40–50) | 3 | 0.99 (0.99–1.00) | 0.142 | 94.7 | [22, 48, 50] |
| Lowest tertile (<40) | 3 | 1.04 (0.99–1.09) | 0.160 | 26.3 | [21, 25, 35] |
| Birth length | |||||
| Overall | 3 | 1.02 (1.00–1.04) | 0.051 | 89.2 | [25, 26]; present study |
| Head circumference | |||||
| Overall | 2 | 1.11 (1.11–1.11) | <0.001 | 100.0 | [26]; present study |
| Maternal age | |||||
| Overall | 5 | 1.01 (1.01–1.02) | <0.001 | 31.7 | [20, 22, 25, 50]; present study |
| Excluding present study | 4 | 1.01 (1.01–1.02) | <0.001 | 21.9 | |
| Menopausal status | |||||
| Pre-menopausal women | 3 | 1.00 (0.99–1.02) | 0.539 | 0.0 | [20, 22]; present study |
| Post-menopausal women | 3 | 1.01 (1.01–1.02) | <0.001 | 0.0 | [20, 22, 50] |
| Quality score (maximum 59)d | |||||
| Highest (≥40) | 3 | 1.01 (1.01–1.02) | <0.001 | 17.4 | [22, 50]; present study |
| Lowest (<40) | 2 | 1.00 (0.97–1.04) | 0.898 | 67.2 | [20, 25] |
aLope et al. [20] was not included in the birthweight meta-analysis, owing to concerns about the validity of a summary trend measure across the limited number of categories (three groups)
bDue to the high between-study heterogeneity in most strata, these average estimates should be interpreted simply as indicators of the direction of the trend in breast density with increasing birth size
cMeta-analysis uses OR1 from Andersen et al. [35] as reported in Table S3, which is adjusted for age at screening and birth cohort: OR 0.98; 95 % CI 0.90–1.07
dMeta-analysis uses OR2 from Andersen et al. [35] as reported in Table S3, which is adjusted for age at screening, birth cohort and BMI at age 13 years: OR 1.11; 95 % CI 1.02–1.22
eRange 0–59; see Methods section of main text and Additional file 1: Methods 3 for description of how study quality scores were developed