Table 2. Univariable and Multivariable Analyses of Association of Individual- and Cohort-Level Factors With IQ z Scores Among Adults With Very Preterm Birth or Very Low Birth Weight (N = 1068).
| Factor | 1-Stage univariable analysis | 1-Stage multivariable analysis | ||
|---|---|---|---|---|
| IQ z score estimate (95% CI) | P value | IQ z score estimate (95% CI) | P value | |
| Individual-level | ||||
| Gestational age, wk | 0.04 (0.02 to 0.06) | <.001 | 0.11 (0.07 to 0.14) | <.001 |
| Male sex | 0.07 (−0.05 to 0.20) | .26 | 0.09 (−0.03 to 0.20) | .16 |
| Birth weight z score per 1 SD | 0.05 (−0.01 to 0.11) | .08 | 0.21 (0.14 to 0.28) | <.001 |
| Maternal educational levela | 0.25 (0.17 to 0.34) | <.001 | 0.26 (0.17 to 0.35) | <.001 |
| Presence of neonatal BPDb | −0.37 (−0.51 to −0.23) | <.001 | −0.16 (−0.30 to −0.02) | .02 |
| BPD defined as oxygen after 28 d postbirthc | −0.34 (−0.56 to −0.12) | .003 | NA | NA |
| BPD defined as oxygen after 36 wk postmenstrual agec | −0.40 (−0.56 to −0.23) | <.001 | NA | NA |
| Presence of neonatal IVH | ||||
| Any graded | −0.27 (−0.40 to −0.13) | <.001 | −0.19 (−0.33 to −0.05) | .007 |
| Grade 3 or 4e | −0.66 (−0.92 to −0.41) | <.001 | NA | NA |
| Multiple birthf | 0.01 (−0.13 to 0.15) | .86 | 0.00 (−0.13 to 0.14) | .95 |
| Cohort-level | ||||
| Attrition, % | ||||
| VPT/VLBW participants | 0 (−0.01 to 0) | .08 | −0.02 (−0.06 to 0.02) | .32 |
| VPT/VLBW participants with childhood NSI | 0 (−0.01 to 0) | .25 | 0.02 (−0.03 to 0.08) | .39 |
| Mean cohort age at IQ assessment, y | 0 (−0.01 to 0.02) | .60 | −0.10 (−0.26 to 0.06) | .28 |
| Year of birth | −0.02 (−0.03 to 0) | .02 | −0.03 (−0.11 to 0.05) | .48 |
Abbreviations: BPD, bronchopulmonary dysplasia; IVH, intraventricular hemorrhage; NA, not applicable; NSI, neurosensory impairment; VPT/VLBW, very preterm or very low birth weight.
Level 1 indicates low, 2 indicates medium, and 3 indicates high.
Compared with participants with no BPD. Participants from the Kroll et al33 cohort were not included in the univariable estimate, but their BPD values were imputed for the multivariable estimate.
The Heinonen et al,29 Madzwamuse et al,8 Pyhala et al,30 and Lærum et al31 studies used the criterion of 28 days’ after birth. The Linsell et al,7 Darlow et al,32 and Doyle et al34 studies used the criterion of 36 weeks’ postmenstrual age.
Compared with participants with no neonatal IVH.
Compared with all other grades. Participants in the Darlow et al study32 could not have IVH harmonized into grade 3 or 4 and were therefore not included in the subanalysis.
Compared with singleton birth.