Table 2. Coefficients for three customization models in the NICHD Fetal Growth Studies–Singletons (N = 2,288).
| Gardosi Model a | Heteroscedastic Model | Quantile Regression Model | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Variable | Estimate | SE | P | Estimate | SE | P | Estimate 10% | Estimate 50% | Estimate 90% | SE 50% | P 50% |
| Mean Model | |||||||||||
| Intercept (Term Optimal Weight) | 3509.722 | 21.250 | < .0001 | 3510.000 | 21.264 | < .0001 | 3069.076 | 3486.616 | 4041.833 | 24.289 | < .0001 |
| Gestational age (from 280 d) | |||||||||||
| Linear term | 14.944 | 1.818 | < .0001 | 14.944 | 1.850 | < .0001 | 13.366 | 17.147 | 8.282 | 2.269 | < .0001 |
| Quadratic term | -0.282 | 0.191 | 0.1403 | -0.293 | 0.184 | 0.1103 | -0.389 | -0.403 | -0.269 | 0.206 | 0.0509 |
| Cubic term | 0.026 | 0.012 | 0.034 | 0.025 | 0.012 | 0.0379 | 0.027 | 0.008 | 0.042 | 0.016 | 0.6414 |
| Sex | |||||||||||
| Male | 67.048 | 7.911 | < .0001 | 67.048 | 8.002 | < .0001 | 55.752 | 66.499 | 71.076 | 9.517 | < .0001 |
| Female | -67.048 | 7.911 | < .0001 | -67.048 | 8.002 | < .0001 | -55.752 | -66.499 | -71.076 | 9.517 | < .0001 |
| Maternal height (from 163 cm) | |||||||||||
| Linear term | 5.864 | 1.903 | 0.0021 | 5.861 | 2.007 | 0.0035 | 9.412 | 5.306 | -4.789 | 2.129 | 0.0127 |
| Quadratic term | 0.041 | 0.117 | 0.7252 | 0.034 | 0.114 | 0.7691 | 0.173 | 0.130 | -0.081 | 0.112 | 0.2464 |
| Cubic term | -0.001 | 0.008 | 0.8934 | 0.000 | 0.009 | 0.986 | -0.018 | 0.002 | 0.020 | 0.008 | 0.7861 |
| Maternal prepregnancy weight (from 64 kg) | |||||||||||
| Linear term | 7.720 | 1.140 | < .0001 | 7.721 | 1.170 | < .0001 | 4.751 | 7.395 | 13.533 | 1.348 | < .0001 |
| Quadratic term | -0.114 | 0.063 | 0.0706 | -0.119 | 0.065 | 0.0669 | -0.086 | -0.157 | -0.203 | 0.075 | 0.0372 |
| Cubic term | 0.000 | 0.001 | 0.6579 | 0.000 | 0.001 | 0.7709 | -0.001 | 0.001 | -0.001 | 0.001 | 0.5762 |
| Race | |||||||||||
| Non-Hispanic black | -189.435 | 21.695 | < .0001 | -189.435 | 21.880 | < .0001 | -182.776 | -193.922 | -225.116 | 24.380 | < .0001 |
| Hispanic | -54.454 | 22.150 | 0.014 | -54.454 | 22.964 | 0.0177 | -77.925 | -65.586 | -74.850 | 28.232 | 0.0203 |
| Asian/Pacific Islander | -49.903 | 26.053 | 0.0556 | -49.903 | 25.876 | 0.0538 | -9.154 | -69.425 | -79.243 | 36.557 | 0.0577 |
| Parity | |||||||||||
| 1 | 92.614 | 18.070 | < .0001 | 92.614 | 18.394 | < .0001 | 72.759 | 97.666 | 59.634 | 22.623 | < .0001 |
| 2+ | 101.536 | 22.391 | < .0001 | 101.536 | 23.535 | < .0001 | 117.866 | 102.512 | 45.780 | 28.195 | 0.0003 |
| Variance Model b | |||||||||||
| Intercept | 374.598 | 14.777 | < .0001 | ||||||||
| Gestational age (from 280 d) | |||||||||||
| Linear term | -0.004 | 0.007 | 0.604 | ||||||||
| Quadratic term | 0.000 | 0.001 | 0.8844 | ||||||||
| Cubic term | 0.000 | 0.000 | 0.9483 | ||||||||
| Sex | |||||||||||
| Male | 0.025 | 0.030 | 0.4035 | ||||||||
| Female | -0.025 | 0.030 | 0.4035 | ||||||||
| Maternal height (from 163 cm) | |||||||||||
| Linear term | -0.012 | 0.007 | 0.0922 | ||||||||
| Quadratic term | 0.000 | 0.000 | 0.3882 | ||||||||
| Cubic term | 0.000 | 0.000 | 0.1498 | ||||||||
| Maternal prepregnancy weight (from 64 kg) | |||||||||||
| Linear term | 0.015 | 0.004 | < .0001 | ||||||||
| Quadratic term | 0.000 | 0.000 | 0.6577 | ||||||||
| Cubic term | 0.000 | 0.000 | 0.5303 | ||||||||
| Race | |||||||||||
| Non-Hispanic black | -0.111 | 0.082 | 0.1783 | ||||||||
| Hispanic | -0.011 | 0.082 | 0.8963 | ||||||||
| Asian/Pacific Islander | -0.041 | 0.100 | 0.6829 | ||||||||
| Parity | |||||||||||
| 1 | 0.039 | 0.070 | 0.5723 | ||||||||
| 2+ | 0.040 | 0.083 | 0.6278 | ||||||||
Note: 0.000 is used for any value <0.001.
a All three models included the same customizing variables containing cubic and quadratic terms of deviation of gestational time at delivery from the optimal 280 days mark a priori per the Gardosi model [12]. In addition to the six proposed “physiological” variables (as designated by the Gardosi method) that influence fetal growth, models also included smoking, BMI (kg/m2), gestational diabetes, gestational hypertensive disease/preeclampsia, and antepartum bleeding. Analysis was centered on 280 days’ gestation, height 163 cm, prepregnancy weight 64 kg, nulliparous, and Non-Hispanic White race/ethnicity. However, only the coefficients for the six “physiologic” variables were included in an additive model to calculate the term optimal weight percentiles.
b Variance is only for the heteroscedastic model.