Table 3. Birth weight percentiles among stillbirths and live births using different equations to estimate individualized expected weight.
| Birth Weight Norms and Percentiles | SB | LB | Crude OR for SB (95% CI)a | Adjusted OR for SB (95% CI)b |
| Weighted sample size, number c | 528 | 1,382 | ||
| Individualized norms, percent d (original 19 variables) | ||||
| <5th percentile | 33 | 9 | 6.47 (4.91–8.53) | 6.01 (4.41–8.20) |
| 5th–<10th | 8 | 6 | 2.05 (1.35–3.12) | 1.84 (1.13–2.98) |
| 10th–90th | 42 | 72 | Reference | Reference |
| >90th–95th | 4 | 5 | 1.39 (0.85–2.26) | 1.48 (0.87–2.52) |
| >95th | 13 | 7 | 3.04 (2.14–4.30) | 2.57 (1.73–3.81) |
| Individualized norms, percent e (11 variables) | ||||
| <5th percentile | 33 | 10 | 5.82 (4.38–7.71) | 5.33 (3.92–7.26) |
| 5th–<10th | 8 | 6 | 2.33 (1.55–3.50) | 2.07 (1.29–3.32) |
| 10th–90th | 43 | 73 | Reference | Reference |
| >90th–95th | 3 | 4 | 1.33 (0.78–2.25) | 1.27 (0.69–2.31) |
| >95th | 12 | 7 | 2.67 (1.88–3.80) | 2.21 (1.49–3.28) |
| Individualized norms, percent f (6 variables) | ||||
| <5th percentile | 30 | 7 | 6.80 (5.08–9.12) | 6.24 (4.49–8.67) |
| 5th–<10th | 8 | 5 | 2.33 (1.48–3.65) | 2.06 (1.26–3.35) |
| 10th–90th | 43 | 72 | Reference | Reference |
| >90th–95th | 4 | 6 | 1.17 (0.73–1.89) | 1.45 (0.83–2.54) |
| >95th | 15 | 9 | 2.58 (1.87–3.56) | 2.21 (1.54–3.17) |
| Individualized norms, percent g (5 variables) | ||||
| <5th percentile | 30 | 7 | 6.64 (4.96–8.88) | 6.25 (4.51–8.67) |
| 5th–<10th | 7 | 6 | 2.04 (1.30–3.20) | 1.74 (1.06–2.84) |
| 10th–90th | 44 | 71 | Reference | Reference |
| >90th–95th | 4 | 7 | 0.95 (0.58–1.57) | 1.04 (0.58–1.87) |
| >95th | 14 | 9 | 2.66 (1.93–3.68) | 2.39 (1.67–3.44) |
Birth weight for GA at death (stillbirths) or delivery (live births) by the SCRN algorithm [17]. Percentages may add to slightly more or less than 100% because of rounding.
Unadjusted OR for stillbirth for infants with birth weight in the percentile group shown compared to infants in the reference group from a logistic regression model that included effects for percentile group only.
Adjusted OR for stillbirth for infants with birth weight in the percentile group shown compared to infants in the reference group from a logistic regression model that in addition to the percentile group indicators included study site number; paternal age (<20, 20–34, 35–39, ≥40 y); the following maternal variables (categorized as shown in Table 1 or as noted): maternal age (<20, 20–34, 35–39, ≥40 y), race/ethnicity, education, marital status, insurance/method of payment, family income, smoking during the 3 mo prior to pregnancy, alcohol use during the 3 mo prior to pregnancy, drug use, BMI, blood type, Rh factor, pregestational hypertension, pregestational diabetes, seizure disorder, and pregnancy history; and infant sex.
Analysis weights that accounted for the basic study design plus other aspects of the sampling were used.
Unweighted sample sizes were 527 stillbirths and 1,821 live births. Unweighted (weighted) sample sizes included in computation of adjusted ORs were 452 (451) stillbirths and 1,665 (1,261) live births.
Individualized norm percentiles were derived using the fetal weight equation from Bukowski et al. [15].
All 19 variables were used in the fetal equation here. (Fetal heart rate was included in the original Bukowski equation but was not collected by the SCRN study. GA in days minus 280 d drops out of the equation when predicting birth weight at 280 d.)
Individualized norm percentiles were derived using the subset of 11 variables largely non-missing in the SCRN cohort in the fetal weight equation to predict term birth weight: maternal weight, height, race/ethnicity, education, marital status, number of prior term pregnancies, number of prior abortions, altitude of residence, use of ovulation induction to become pregnant, cigarettes smoked per day during the first trimester, and male fetus.
Individualized norm percentiles were derived using the subset of six variables suggested by Gardosi et al. [20] in the fetal weight equation to predict term birth weight (excluding GA, which drops out): maternal weight, height, race/ethnicity, number of prior term pregnancies, cigarettes smoked per day during the first trimester, and male fetus.
Individualized norm percentiles were derived using a subset of five variables (the six variables above minus number of cigarettes smoked) in the fetal weight equation to predict term birth weight: maternal weight, height, race/ethnicity, number of prior term pregnancies, and male fetus.
LB, live birth; SB, stillbirth.