Skip to main content

Table 2 Predictive performance of a sequential algorithm to identify higher-than-normal BMI growth in offspring of mothers with obesity

From: Predicting the earliest deviation in weight gain in the course towards manifest overweight in offspring exposed to obesity in pregnancy: a longitudinal cohort study

Predictive parameter

Prediction at age 3 monthsa

Prediction at age 1 yearb

Prediction at age 2 yearsc

Higher-than-normal BMI growth in early phase (6 months–2 years)

Higher-than-normal BMI growth in late phase (3 years–5 years)

Higher-than-normal BMI growth in late phase (3 years–5 years)

Discovery cohort

Discovery cohort

Validation cohort

Discovery cohort

Validation cohort

N

711

645

670

640

666

AUROC

0.69 (0.66, 0.72)

0.73 (0.70, 0.75)

0.61

0.79 (0.76, 0.81)

0.71

Cut-off score valued

− 1.689

− 1.135

NA

− 1.133

NA

Prevalence, n (%)

140 (20.0)

194 (30.8)

221 (33.0)

192 (30.0)

223 (33.5)

Sensitivity, %

70.7 (55.5, 82.3)

73.7 (67.6, 79.0)

68.1 (62.5, 73.2)

76.0 (70.0, 81.1)

61.0 (55.3, 66.5)

Specificity, %

74.1 (61.6, 83.6)

64.7 (58.2, 70.7)

58.2 (52.5, 63.7)

78.1 (72.9, 82.5)

68.0 (63.3, 72.4)

Positive predictive value, %

40.5 (26.5, 55.6)

48.2 (41.9, 54.5)

45.3 (40.1, 50.6)

60.7 (53.5, 67.4)

49.5 (43.6, 55.3)

Negative predictive value, %

91.0 (84.7, 95.0)

84.7 (80.2, 88.3)

78.2 (73.4, 82.4)

88.0 (84.5, 90.7)

77.2 (73.3, 80.8)

Positive likelihood ratio

2.73 (1.44, 5.03)

2.09 (1.62, 2.69)

1.63 (1.32, 2.02)

3.47 (2.58, 4.64)

1.91 (1.50, 2.41)

Negative likelihood ratio

0.40 (0.21, 0.72)

0.41 (0.30, 0.56)

0.55 (0.42, 0.71)

0.31 (0.23, 0.41)

0.57 (0.46, 0.71)

  1. We used the PEACHES cohort study as the discovery cohort and the PEPO cohort study as the external validation cohort for calculation of the individual child’s risk of a “higher-than-normal BMI growth pattern” (BMI z-score >1 SD [51] at least twice). Values are predictive parameters and their 95% CI
  2. AUROC, area under the receiver operating characteristic; BMI, body mass index; CI, confidence interval; GDM, gestational diabetes; GWG, gestational weight gain; NA, not applicable; PEACHES, Programming of Enhanced Adiposity Risk in CHildhood–Early Screening; PEPO, PErinatal Prevention of Obesity; SES, socioeconomic status
  3. aPotential predictors included: maternal pre-conception BMI group, total GWG, GDM, parity, smoking during pregnancy, sex, birth weight category for gestational age and sex, SES, breastfeeding status at 1 month, breastfeeding status at 3 months, and BMI z-score >1 SD at age 3 months. External validation of models at age 3 months could not be performed due to the lack of follow-up data at age 3 months in the validation cohort PEPO
  4. bPotential predictors included: maternal pre-conception BMI group, total GWG, GDM, parity, smoking during pregnancy, sex, birth weight category for gestational age and sex, SES, breastfeeding status at 1 month, breastfeeding status at 3 months, breastfeeding status at 6 months, and BMI z-score >1 SD at age 1 year. External validation of models at age 1 year was performed in the validation cohort PEPO
  5. cPotential predictors included: maternal pre-conception BMI group, total GWG, GDM, parity, smoking during pregnancy, sex, birth weight category for gestational age and sex, SES, breastfeeding status at 1 month, breastfeeding status at 3 months, breastfeeding status at 6 months, and BMI z-score >1 SD at age 2 years. External validation of models at age 2 years was performed in the validation cohort PEPO
  6. dOffspring with a risk score above or equal to the respective cut-off score value are considered to be at risk of developing a “higher-than-normal BMI growth pattern.” The cut-off value of the score was optimized to avoid false-negative findings (sensitivity), which resulted in negative cut-off score values