Prenatal Tobacco Smoke Exposure and Risk for Brain Abnormalities on MRI at Term in Infants Born Very Preterm

Abstract

Background:

Prenatal tobacco smoke exposure (TSE) and preterm birth are associated with abnormal brain and neurodevelopmental outcomes in infants. Studies that can disentangle indirect mediating effects from direct effects of prenatal TSE on sensitive early brain MRI biomarkers in very preterm infants are needed.

Objectives:

We sought to determine whether prenatal TSE in preterm infants posed any direct effects on MRI-determined global brain abnormality score and secondary measures of brain abnormalities after removing any indirect mediating effects of preterm birth on neurostructural outcomes.

Study Design:

We examined brain MRI findings collected at 39–44 weeks postmenstrual age from a prospective cohort of 395 infants born very preterm (≤32 weeks of gestational age). The primary outcome was global brain abnormality score and secondary outcomes were: global efficiency of structural connectome, diffuse white matter abnormality volume, total brain tissue volume, total gray and white matter volume, and cerebellar volume. Maternal reports of smoking during pregnancy were obtained. We performed multivariable linear regression analyses to examine the association between prenatal TSE and our MRI outcomes, controlling for prospectively collected confounders. We performed a mediation analysis to estimate the direct effects of prenatal TSE on brain abnormalities and any indirect effects through preterm birth.

Results:

In total, 12.6% of infants had prenatal TSE. Compared to nonexposed infants, infants with prenatal TSE had a higher median(range) global brain abnormality score of 7(0, 41) vs. 5(0, 34) (p=<.001); findings remained significant (p<.001) after controlling for antenatal confounders. Global efficiency (p<.001), diffuse white matter volume (p=.037), and total brain tissue volume (p=.047), were significantly different between TSE groups in multivariable analyses. On mediation analysis, preterm birth mediated between 0 and 29% of the indirect effect of prenatal TSE on several measures of brain abnormality outcomes. Thus, prenatal TSE had a direct adverse effect between 71% and 100% on brain injury or abnormal development.

Conclusions:

We identified multiple adverse effects of prenatal TSE on sensitive and objective measures of neonatal brain injury/abnormal development; most appear to be a direct effect of prenatal TSE on fetal brain development. Results underscore the significant adverse neurostructural effects of prenatal TSE to tobacco smoke pollutants.

Condensation:

Prenatal TSE affects sensitive and objective measures of neonatal brain injury/abnormal development; the majority appear to be a direct effect of TSE on preterm birth.

Introduction:

Although rates of tobacco use in the United States (U.S.) have recently declined,¹ rates of smoking during pregnancy are as high as 12.8% in the Mid-Western U.S. compared to national rates of 6.0%.² Maternal tobacco use during pregnancy is a leading, preventable risk factor for preterm births³ as up to 16.6% of women who smoke during their third trimester have preterm births⁴ compared to overall preterm birth rates of 10.1% in the U.S.⁵ When mothers actively smoke during pregnancy, their fetuses are exposed to nicotine and other tobacco smoke pollutants via the placenta.⁶ Infants who experienced prenatal tobacco smoke exposure (TSE) in-utero, are at increased risk for a myriad of adverse health outcomes including preterm birth and neurodevelopmental conditions.^{7, 8}

Since both prenatal TSE and preterm birth are associated with adverse neurodevelopmental outcomes,^9–14 it can be difficult to disentangle the effects of each of these prenatal insults on neurodevelopmental outcomes. Further complicating the examination of these effects is the fact that infants with prenatal TSE are commonly at risk to have been exposed to opioids and other substances^{15, 16} which are also associated with poor neurodevelopmental outcomes.^17–19

Magnetic resonance imaging (MRI) studies conducted at term-equivalent age in infants born very preterm (≤32 weeks of gestational age [GA]) are an important tool that can objectively examine the results of prenatal insults on neurodevelopment.²⁰ Abnormal macrostructural and morphometric MRI findings are posited to represent biomarkers that indicate brain injury and maturational delays that are associated with subsequent neurodevelopmental outcomes.^20–22

The global brain abnormality score is a composite measure of brain injury and abnormal maturation on term-equivalent age MRI that is predictive of long-term neurodevelopment up to 7 years of age.^{22, 23} In addition, reduced structural network connectivity/efficiency and diffuse white matter abnormality (DWMA), which is a hyperintense white matter MRI signal on T2-weighted MRI at term-equivalent age, are additional sensitive measure of brain injury/development that are associated with poor neurodevelopmental outcomes.^24–28 However, it is unknown if these MRI findings are present in very preterm infants with prenatal TSE. Thus, the primary objective was to determine whether preterm infants with prenatal TSE would have higher MRI-determined global brain abnormality scores, compared to infants with no prenatal TSE. We expected to observe a significant direct effect of prenatal TSE on brain injury and/or abnormal maturation even after accounting for the indirect mediating effects of preterm birth.

Materials and Methods:

Subjects

We conducted a prospective multisite regional cohort study in very preterm infants born at ≤32 weeks GA without congenital or chromosomal abnormalities that affect the central nervous system. Infants were recruited from September 2016 and November 2019 from five level III/IV neonatal intensive care units from the Greater Cincinnati region; details are elsewhere.²⁹ The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines for observational studies were followed.³⁰ Institutional review board approval from all participating hospitals was obtained and parents provided written informed consent.

Brain MRI

All infants underwent a brain MRI between 39- and 45-weeks postmenstrual age on a 3T Philips Ingenia scanner (Philips Healthcare, Best, Netherlands) while using a 32-channel head coil, as previously described.²⁹ Axial T2-weighted sequence parameters were as follows: echo time (TE) 166 milliseconds (ms), repetition time (TR) 18567 ms, flip angle (FA) 90°, and voxel dimensions 1.0×1.0×1.0 mm; MPRAGE T1-weighted images (three-dimensional fast field echo): TR/TE/inversion time = 8.5/3.4/1610 ms, FA 13°, in-plane resolution = 1.0×1.0×1.0 mm; diffusion MRI: TE 88 ms, TR 6972 ms, FA 90°, voxel dimensions 2.0×2.0×2.0 mm; 34 directions of diffusion gradients were applied with a b-value of 800 s/mm², and 4 b0 images were acquired. Brain MRIs were used to assess the global brain abnormality scores, the primary outcome, and the following secondary outcomes: global efficiency of whole brain structural connectivity, DWMA, and regional brain volumes (i.e., total brain tissue volume, total gray and white matter volume, cerebellar volume).

Global Brain Abnormality Score

A pediatric neuroradiologist (BMKF), who was blinded to clinical history, performed qualitative and quantitative assessments of the MRI scans to delineate brain injury or abnormal brain maturation. As previously described,^{29, 31} we used the standardized Kidokoro et al.,²² preterm brain MRI scoring system that evaluates four of the key brain regions –white matter, cortical gray matter, deep gray matter, and cerebellum – to derive composite scores for each region and a global brain abnormality score (sum of all four regions).

Global Efficiency

Diffusion MRI scans were processed as previously described in Kline et al.³² Briefly, we used the developing Human Connectome Project (dHCP) diffusion preprocessing pipeline.³³ We used Diffusion Toolkit to derive fractional anisotropy (FA) maps and performed whole-brain, deterministic fiber tracking. We aligned the 122-region of interest John’s Hopkins University (JHU)³⁴ neonatal brain template to create parcellated brain maps in diffusion space in a multi-step process. Brain structural connectome construction was performed in MRTRIX3 (http://www.mrtrix.org,version 0.3.0), as described previously.³² We generated symmetrical, undirected connectivity matrices, with edge weights corresponding to the mean FA for all streamlines connecting each pair of regions. We used Brain Connectivity Toolbox (brain-connectivity-toolbox.net) to calculate global efficiency, defined as average inverse shortest path length for that network.³⁵ Higher global efficiency represents fewer overall steps between nodes in a network, and higher FA values equate to higher global efficiency.

Brain Volume Measurements

As previously described,^{32, 36} we used the dHCP structural MRI pipeline to automatically segment the structural T2-weighted images and compute total brain tissue volume, total gray matter volume, total white matter volume, and cerebellar volume. We divided brain volumes (originally in mm³) by 1000 to transform them to cm³(mL) to facilitate interpretability of the reported coefficients.

Normalized Diffuse White Matter Abnormality Volume

We objectively quantified volume of DWMA using our published algorithm.³⁷ Briefly, our software labels white matter voxels with signal intensity >1.8 standard deviation(SD) above the mean intensity for all gray and white matter voxels. Volume of DWMA was calculated as the product of voxel volume and total number of voxels in the detected DWMA regions. Normalized DWMA volume was calculated by dividing DWMA volume by total white matter volume to correct for the effect of varying head sizes.

Maternal and Infant Clinical Assessments

Information about maternal prenatal TSE, marijuana, and opioid use pertaining to this pregnancy was assessed via questionnaire at the MRI visit. Mothers were asked: 1) if they were current or former smokers; current smokers reported number of packs smoked/day. Infants of current smokers were categorized as having prenatal TSE; 2) if they used street drugs including marijuana; 3) if they used narcotics or narcotic-strength pain medications and if they were obtained with or without a prescription.

Maternal characteristics, pregnancy/delivery data, and infant data were collected prospectively by research staff prior to NICU discharge; see Parikh et al.²⁹

Statistical analysis:

We compared baseline maternal and infant characteristics between preterm infants with and without prenatal TSE using Wilcoxon rank sum test for continuous variables and Fisher’s Exact test for categorical variables. We conducted multivariable linear regression analyses while controlling for the effects of five maternal confounders: opioid use, marijuana use, absent/ incomplete course of antenatal steroids, magnesium therapy, and hypertensive disorders of pregnancy(HDP). Postmenstrual age at MRI scan was included as a confounder in all analyses. We included intracranial volume as a confounder for all regional volume outcome analyses. We tested the social risk score,³⁸ a composite measure of socioeconomic status, in our primary mediation model; it did not exert a significant effect on global brain abnormality and was not included as a confounder.

Since GA was significantly different between infants with and without prenatal TSE and preterm birth temporally occurs after TSE, GA is likely an indirect mediator between smoking and brain abnormalities. Thus, we performed a causal inference mediation analysis³⁹ for outcomes that were significant in multivariable analyses. This allowed us to separate out the direct effect of prenatal TSE on the brain abnormalities from the indirect effect attributed to preterm birth. This involved two multivariable linear regression models; (1)predict mediator from TSE adjusting for known confounders, (2)predict outcome from TSE adjusting for the mediator and known confounders. We investigated interactions between prenatal TSE and GA on brain MRI outcomes. Since no interaction was detected, we examined the effect of prenatal TSE on MRI outcomes without considering the interaction term in the outcome model. The indirect (i.e., mediated) effect is the product of the TSE coefficient in the mediator model times the GA coefficient in the outcome regression model. The natural direct effect is the coefficient of TSE from the outcome model that includes the mediator. The total effect is the sum of the direct and indirect effects. The percentage of mediated effect was estimated dividing the indirect effect by the total effect and multiplying by 100; confidence limits were calculated using a normal approximation for the Wald test. Statistical significance was considered p<0.05; SAS (SAS Institute Inc., Cary, NC) version 9.4 was used.

Results:

The mean(SD) GA and birthweight of our cohort (n=395) was 29.3(2.5) weeks and 1294.3 (448.9) grams, respectively. Of these 395 infants, 50(12.7%) had prenatal TSE and of the 50 mothers who smoked, 40 smoked <½ pack/day. Thirty-five(9.2%) mothers used opioids of which 27 were prescribed opioids. Thirteen mothers used opioids and smoked cigarettes during pregnancy; 22(5.6%) used marijuana, of which 10 concurrently smoked cigarettes. We identified significant group differences in these two key maternal characteristics (Table 1).

Table 1.

Maternal and infant baseline characteristics of very preterm infants with and without prenatal TSE during pregnancy

Baseline variables n (%)	No prenatal TSE (N=345)	Prenatal TSE (N=50)	P
Maternal age, years, mean (SD)	29.0 (5.5)	30.5 (4.4)	0.039
Maternal antenatal opioid use	18 (5.2%)	14 (28.0%)	<0.001
Maternal antenatal marijuana use	12 (3.5%)	10 (20.0%)	<0.001
Antenatal steroids (incomplete course)	86 (24.9%)	17 (34.0%)	0.172
Antenatal magnesium	291 (84.3%)	41 (82.0%)	0.672
Histologic chorioamnionitis	96 (27.8%)	17 (34.0%)	0.367
Hypertensive disorder of pregnancy	150 (43.5%)	17 (36.0%)	0.318
Gestational age, weeks, mean (SD)	29.4 (2.5)	28.5 (2.4)	0.028
Birth weight Z-score, mean (SD)	0.090 (0.982)	−0.044 (0.828)	0.302
Small for gestational age	24 (7.0%)	3 (6.0%)	1.000
Apgar score <5 at 5 min‡	42 (12.2%)	7 (14.6 %)	0.647
Caffeine therapy	240 (69.8%)	41 (82.0%)	0.074
Postnatal sepsis	41 (11.9%)	3 (6.0%)	0.334
Bronchopulmonary dysplasia (any severity)	146 (42.3)%	21 (42.0%)	1.000
Postnatal corticosteroids for BPD	35 (10.2%)	6 (12.0%)	0.628
Surgery for NEC or SIP	20 (5.8%)	2 (4.0%)	1.000
Surgery requiring general anesthesia	38 (11.0%)	8 (16.0%)	0.343
Retinopathy of prematurity (any grade)	113 32.8%)	21 (42.0%)	0.204
PMA at MRI, weeks, mean (SD)	42.7 (1.3)	42.5 (1.3)	0.327

^‡Data unavailable for 4 infants

Abbreviations: PMA – postmenstrual age; NEC – necrotizing enterocolitis; SIP – spontaneous intestinal perforation

There were no statistically significant group differences in variables known to be associated with brain structural/neurodevelopmental outcomes: histologic chorioamnionitis, magnesium therapy, absent/incomplete antenatal steroid course, or HDP.^{17, 40–42} Because the latter two variables exhibited clinically significant group differences, they were considered confounders and adjusted for in all multivariable analyses as expected from prior studies.^{7, 8, 43} GA was significantly different between groups and thus we examined its effect as a mediator between prenatal TSE and brain abnormalities (Table 1).

We were unable to examine the potential confounding effects of alcohol or other street drugs because only 2/29 mothers who reported drinking alcohol also smoked during pregnancy. Of the 35 mothers who reported using street drugs, only 13 used anything other than marijuana.

Our primary outcome was significantly different between groups in univariate analysis, with infants with TSE having a median(range) global brain abnormality score of 7(0, 41) and non-exposed infants having a score of 5(0, 34) (p<.001). In multivariable analyses, global brain abnormality score remained significant (p<.001) after controlling for our five antenatal confounders (Table 2). Regional brain abnormality subscores were also different between groups in secondary analyses except for cerebellar scores.

Table 2.

Relationship between smoking during pregnancy and brain abnormalities at term-equivalent age in infants born very preterm in unadjusted analyses and analyses adjusted for confounders

	Unadjusted		Adjusted*
MRI outcome measure	Coefficient (95% CI)	P value	Coefficient (95% CI)	P value
Global brain abnormality score	2.94 (1.41, 4.48)	<.001	2.83 (1.18, 4.49)	<.001
Cortical Gray Matter score	0.51 (0.14, 0.88)	.007	0.58 (0.18, 0.98)	.005
Deep Gray Matter score	0.49 (0.12, 0.87)	.010	0.49 (0.09, 0.89)	.017
Total White Matter score	1.65 (0.81, 2.5)	<.001	1.42 (0.51, 2.32)	.002
Cerebellar score	0.29 (−0.22, 0.79)	.264	0.35 (−0.20, 0.89)	.210
Global efficiency	−0.008 (−0.011, −0.004)	<.001	−0.006 (−0.009, −0.003)	<.001
Diffuse white matter abnormality volume	0.019 (0.004, 0.035)	.015	0.016 (0.001, 0.032)	.037
Total brain tissue volume (mL)	−19.75 (−35.49, −4.00)	.014	−16.51 (−32.78, −0.25)	.047
Total gray matter volume (mL)**	−115.92 (−205.67, −26.16)	.012	−24.76 (−72.80, 23.28)	.311
Total white matter volume (mL)**	−60.23 (−140.54, 20.09)	.141	24.91 (−24.68, 74.50)	.324
Total cerebellar volume (mL)**	−18.22 (−33.99, −2.45)	.024	−0.98 (−11.38, 9.43)	.854

^*Adjusted for opioid use during pregnancy, marijuana during pregnancy, absent or incomplete course of antenatal steroids, hypertensive disorders of pregnancy, and postmenstrual age at MRI scan

^**Additionally adjusted for intracranial volume

In secondary analyses, we found that global efficiency (p<.001), DWMA (p=.015), total brain tissue volume (p=0.014), total gray matter volume (p=0.012), and cerebellar volume (p=.024), were each significantly different between TSE groups. In multivariable linear regression analyses, global efficiency (p<.001), DWMA (p=.037), and total brain volume (p=.047) remained significant after controlling for opioid use, marijuana use, HDP, incomplete course of antenatal steroids, and PMA(Table 2).

Birth GA was significantly lower in infants with TSE (p=.028). Thus, we conducted mediation analyses to determine if preterm birth/GA exhibited indirect adverse effects on brain MRI outcomes. For our primary outcome, global brain abnormality score, GA mediated 18.6% of the effect of prenatal TSE on global brain abnormality score, though there was only a trend towards significance (p=.080) (Table 3). The remaining 81.4% of the effect can be attributed as a direct harmful effect of smoking on brain abnormalities (p=.003). Thus, most of the adverse effects of prenatal TSE is a direct effect on the preterm brain and independent of the effect of preterm birth. For global efficiency, smoking exerted both a significant direct (81.1%) and indirect effect (18.9%) through preterm birth on structural connectivity global efficiency.

Table 3.

Results of mediation analysis to examine the indirect effect of smoking during pregnancy on preterm birth and its direct effect on MRI brain abnormalities at term-equivalent age in infants born very preterm.

MRI outcome measure	Total Effect Coefficient (95%CI) P value	Direct Effect Coefficient (95% CI) P value	Indirect Effect Coefficient (95% CI) P value	% Mediated
Global brain abnormality score	2.83 (1.20, 4.47) P<.001	2.31 (0.77, 3.84) P=.003	0.53 (−0.06, 1.12) P=.080	18.6%
Global efficiency	−0.0059 (−0.0089, −0.0028) P<.001	−0.0048 (−0.0077, −0.0018) P=.002	−0.0011 (−0.0022, −0.0001) P=.038	18.9%
Total brain volume (mL)	−16.51 (−32.57, −0.46) P=.044	−12.65 (−27.88, 2.57) P=.103	−3.86 (−9.21, 1.49) P=.157	23.4%
Diffuse white matter abnormality	0.016 (0.001, 0.032) P=.035	0.017 (0.001, 0.032) P=.033	−0.0002 (−0.002, 0.001) P=.822	−1.1%

All analyses were adjusted for the following confounders: opioid use during pregnancy, marijuana during pregnancy, absent or incomplete course of antenatal steroids, hypertensive disorders of pregnancy, and postmenstrual age at MRI scan.

Although the effect of TSE on preterm birth is known,^{3, 4} there was no indirect effect of GA on DWMA volume. The effect of smoking on increased DWMA volume appeared to be an exclusively direct effect(Table 3). Last, we did not observe a significant indirect effect of preterm birth on total brain tissue volume and the direct effect also was nonsignificant.

Discussion:

Principal Findings:

This is the first study that has objectively examined global brain injury and brain maturation on brain MRIs of very preterm infants who were exposed to tobacco smoke in-utero compared to unexposed infants. Moreover, we used multiple sensitive and validated measure of abnormalities on term-equivalent age MRI, including Kidokoro’s global brain abnormality score,²² global efficiency of the whole brain structural connectome,³² and objectively defined DWMA volume.²⁸ Our results indicate that in this cohort of 395 very preterm infants, prenatal TSE was independently associated with increased risk of brain injury or brain maturational abnormalities by term-equivalent age. On secondary analysis, after controlling for variables commonly associated with abnormal brain MRI and/or neurodevelopmental outcomes,^{17, 40–42} we found that prenatal TSE was associated with global efficiency, DWMA volume, and total brain tissue volume. These findings are significant and concerning as global brain injury/maturational abnormalities as assessed on term-equivalent age MRI is a biomarker of poor neurodevelopmental outcomes, including cognitive, language, and motor impairments up to age 7.^{23, 27} Similarly, decreases in global efficiency may be a marker of early diagnosis of cerebral palsy at 3–4 months corrected age and internalizing symptoms at 7 years of age in very preterm infants.^{23, 32} Further, DWMA was found to be an independent biomarker of long-term motor development at age 3 in a subsample of this cohort.²⁴ The objective radiologic biomarkers examined provide further evidence of the serious neurobehavioral consequences potentially associated with prenatal TSE.

In order to account for the potential indirect effect of TSE on preterm birth-associated brain abnormalities through the degree of preterm birth, we conducted a mediation analysis. This analysis revealed a persistent, significant, and sizable direct effect of prenatal TSE on brain injury of 81.4%, global brain efficiency of 81.1%, and DWMA of 100%. These findings indicate that most of the adverse effects of prenatal TSE is a direct effect on the preterm brain at term-equivalent age and independent of the effect of TSE on preterm birth or associated low birth weight.

Results in the Context of What is Known:

This study’s results expand upon the limited extant literature that has examined the MRI-confirmed effects on prenatal TSE in preterm and term infants. In a study of preterm infants <32 weeks gestation or birth weight <1500 grams, infants with prenatal TSE had significantly smaller frontal lobe and cerebellar volumes on MRIs at term equivalent age compared to unexposed infants.¹³ However, this study considered GA as a confounder and inappropriately adjusted for it along with other downstream neonatal diseases, thus likely overcorrecting for the effect of TSE on neonatal brain volumes. In another study that obtained MRIs on 6–8-year-olds, children with prenatal TSE had smaller total brain volumes and cortical gray matter volumes compared to children with no prenatal TSE.⁴⁴ Finally, in a study of 10–14-year-olds, prenatal TSE was associated with reductions in cortical gray matter, total parenchymal volumes, and head circumference.⁴⁵ Our results on brain volume findings are in agreement with prior work as we also found that total brain tissue volume, total brain gray matter volume, and cerebellar volume were each significantly different between TSE groups; however, total white matter volume was not different between the TSE groups. Moreover, even after controlling for several confounders, global brain abnormality, global efficiency, DWMA volume, and total brain tissue volume remained significant.⁴⁶ This is noteworthy because these sensitive surrogate measures are not confounded by socioeconomic and other environmental exposures that plague studies examining the association between prenatal TSE and neurodevelopmental outcomes.^{8, 47} Furthermore, our mediation analyses confirmed a robust direct adverse effect of TSE on several key brain structural outcomes that are independent of the effects of TSE on preterm birth. These findings confirm that prenatal TSE is not only associated with brain abnormalities through induction of preterm birth and/or lower birth weight, but also that prenatal TSE appears to exhibit even stronger direct effects on the developing preterm brain. Considering that all of our evaluated brain biomarkers are associated with significant long-term adverse neurodevelopmental and behavioral outcomes, our findings should renew public health⁴⁸ and medical intervention efforts^{49, 50} to mitigate the harmful effects of TSE.

Clinical Implications:

Similar to prior work,^{51, 52} we observed that birth GA was significantly lower in infants with prenatal TSE. The mechanisms underlying development of direct brain injury/aberrant maturation may be due to the effects of nicotine on exposed infants. Nicotine crosses the placenta at 7 weeks gestation and it is present in the amniotic fluid and umbilical cords of neonates.^{53, 54} Nicotine can activate and then desensitize nicotinic acetylcholine receptors.^{55, 56} Animal studies indicate that acetylcholine plays an important role in brain development as it is involved in the proliferation, maturation, and differentiation of several types of brain cells.⁵⁷ Activation of nicotinic acetylcholine receptors interferes with the development of neurotransmitter systems which include dopamine, norepinephrine, and serotonin, which may result in changes in neurobehavioral development.^55–62 Given research indicating that prenatal TSE is associated with adverse neurodevelopmental effects including epilepsy, behavioral conditions, and delays in cognition, motor, and language skills,^{7, 8} our findings add credence that nicotine is a neuroteratogen that is unequivocally harmful to unborn children.⁴⁶

Strengths and Limitations:

This study has numerous strengths including the examination of objective and sensitive MRI biomarkers in a large cohort of infants born very preterm who were recruited from level III and IV NICUs in the southwestern Ohio region. The primary outcome of MRI-confirmed brain injury and abnormal maturation was obtained by a neuroradiologist blinded to the infants’ clinical history and in a manner that was highly reliable.⁶³ Further, the mediation analysis revealed strong direct effects of prenatal TSE on brain abnormalities independent of the infants’ degree of preterm birth. Limitations include maternal-reported smoking and other substance use. However, self-report most likely was underreported; thus, the actual effects of prenatal TSE may have been even stronger if validation of TSE was confirmed with cord blood analyses. Further, mothers were not asked whether they used electronic cigarettes (e-cigarettes), thus, the potential effects of e-cigarette exposure on infants in this cohort are unknown. Additionally, our results infants may not generalize to other populations and may have underestimated the effects of prenatal TSE on abnormal brain MRI findings.

Conclusions

In conclusion, in a regional cohort of infants born very preterm, we identified multiple adverse effects of prenatal TSE on sensitive and objective measures of neonatal brain injury/abnormal development. Additionally, we performed mediation analyses to disentangle the indirect effects of prenatal TSE on preterm birth /growth restriction from any direct effects on brain injury/abnormalities and demonstrated that a majority of the adverse effects of TSE on brain development are directly mediated. Our unique findings advance our understanding of the significant adverse neurostructural effects of prenatal TSE and underscores the importance of administering maternal interventions to reduce this exposure.

AJOG MFM at a Glance.

1. Why was this study conducted? This study was conducted to explore whether prenatal tobacco smoke exposure (TSE) in very preterm (VPT) infants directly affects brain abnormalities on term MRI.

2. What are the key findings? We examined the MRIs of a prospective cohort of 395 VPT infants and found that compared to nonexposed infants, infants with prenatal TSE had higher global brain abnormalities scores, more diffuse white matter abnormality, and smaller brain tissue volumes. Prenatal TSE had a direct adverse effect between 71% and 100% on brain injury or abnormal development.

3. What does this study add to what is already known? These findings confirm that prenatal TSE affects sensitive and objective measures of neonatal brain injury/abnormal development; the majority appear to be a direct rather than an indirect effect of TSE on preterm birth.

Glossary:

TSE

Tobacco smoke exposure

DWMA

Diffuse white matter abnormality

GA

Gestational age

HDP

Hypertensive disorders of pregnancy