The effects of maternal smoking on fetal cranial development. Findings from routine midtrimester sonographic anomaly screening

Abstract

The aim of this study was to assess the effect of continued smoking before and during pregnancy on mid-trimester fetal head development. A total of 250 pregnant women enrolled in the study. All participants were confirmed to be smokers or non-smokers by verifying breath carbon monoxide readings. Biparietal diameter (BPD), head circumference (HC), lateral ventricle (LV), and cisterna magna (CM) were evaluated by ultrasound between 20–22 weeks of pregnancy. Gender and gestational age-adjusted BPD z- scores were not statistically different between smokers and non-smokers (−0.75 ± 1.6 vs −0.51 ± 1, p = .3). HC measurements and z- scores were significantly lower in the smoking group than in the non-smoking groups (183.38 ± 14.56 vs. 189.28 ± 12.53, p = .003, 0.18 ± 1.39 multiple of median (MoM) vs. 0.56 ± 0.92, respectively, p = .023). At linear regression analysis, maternal smoking was the only independent factor associated with fetal HC z score (p = .041). In conclusion, continued smoking during pregnancy reduces fetal HC and has no effect on BPD, LV, or CM measurements at mid-gestation.

IMPACT STATEMENT

• What is already known on this subject? Smoking during pregnancy is one of the most common environmental factors affecting fetal and neonatal growth and well-being. Despite the well-known effects of smoking on somatic growth, current studies have shown that it selectively affects some parts of the fetal brain, even in appropriately growing fetuses.

• What do the results of this study add? Continued smoking during pregnancy reduces fetal HC and has no effect on BPD, LV or CM measurements at mid-gestation. Since smoking is well known for its early and late childhood behavioral and neurological consequences, smaller mid-trimester fetal HC measurements should bring maternal smoking to mind as one of the potentially reversible causes.

• What are the implications of these findings for clinical practice and/or further research? The harmful effects of smoking start before the third trimester and antenatal counseling should be started early in the gestation. Every effort should be made to quit smoking before or early in pregnancy.

Keywords:

• Breath carbon monoxide
• fetal cranial measurements
• fetal development
• pregnancy
• smoking
• ultrasound

Introduction

Smoking during pregnancy is one of the most common environmental factors affecting fetal and neonatal growth and well-being. Smoking is thought to cause fetal growth restriction through its toxic ingredients. Cotinine is a long-acting metabolite of tobacco smoke and easily passes through the placenta and concentrates in the developing fetus (Luck et al. 1985, Jauniaux et al. 1999). This ingredient of tobacco has a myriad of untoward effects on cellular growth, uterine perfusion, the epigenetic process and ultimately, fetal growth. On the other hand, carbon monoxide is an avid substrate for hemoglobin, creates a carboxyhemoglobin (COHb) complex, and leads to tissue hypoxia (Longo and Hill 1977, Law et al. 2003, Godding et al. 2004). Despite the well-known effects of smoking on somatic growth, current studies have shown that it selectively affects some parts of the fetal brain, even in appropriately growing fetuses. Animal and human studies have demonstrated that smoking reduced overall fetal brain volume and this effect was prominent in the frontal lobe (Ekblad et al. 2010, Abraham et al. 2017). Other researchers have shown that the amygdala and pallidum regions bear the brunt of abnormal growth and the development process (Liu et al. 2011, Haghighi et al. 2013).

Fetal cranial measurements reflect the global growth of the fetal brain. Therefore, the measurements of fetal biparietal diameter (BPD) and head circumference (HC) are good indicators of fetal brain development (Durmus et al. 2011, Abraham et al. 2017). Similarly, the widths of the lateral ventricle (LV) and cisterna magna (CM) are products of the relative growth of the cortical regions, cerebrospinal fluid circulation and ependymal cell function, and thus, are markers for the anatomical and functional abnormalities in these regions (D'Addario and Kurjak 1985, Lavezzi et al. 2010, Kutuk et al. 2013).

Second-trimester fetal anomaly screening ultrasound (US) is a globally accepted method for detecting fetal anomalies and is implemented as a routine part of antenatal follow-up in all developed countries, and even in many developing countries. Fetal skull, CM, and LV measurements are all essential parts of this screening (Salomon et al. 2011). In this study, we retrospectively compared BPD, HC, LV and CM measurements taken during the second trimester US between smoking and non-smoking women in a highly selective group of patients. We intend to observe the effects of smoking on the development of this structure and the potential implication of smoking on the interpretation of routine second trimester US screening.

Materials and methods

This study was conducted in a university hospital serving as a tertiary referral center. The study group consisted of 54 cases who were smokers before and during pregnancy. The control (non-smoking) group (n = 196) consisted of people who had never been smokers. Smoking was determined based on self-reports from the pre-pregnancy period to the day of ultrasonography. An individual was considered a non-smoker if they declared that they had never smoked, including the current pregnancy, and this statement was confirmed by a biochemical marker (BCO readings ≤ 3 ppm). A smoker was defined as someone who smokes ≥1 cigarette before and during pregnancy and whose BCO test reads ≥ 3 ppm (at least one year of a regular smoking history before pregnancy). Social smokers, former smokers, and those cases with indefinite smoking statements were not included in the study. All the participants were pregnant women who had attended our antenatal clinic for pregnancy follow-ups and were accepted to enroll in a prospective study assessing fetal hepatic circulation with Doppler sonography (ClinicalTrials.gov Identifier: NCT04721782). The second trimester detailed ultrasound data of the main project were used for this study.

Patients with chronic hypertension, pregestational diabetes, or fetal structural anomalies were not included in the study. After the routine detailed ultrasonography between 19 and 24 weeks, all pregnant women who were willing to enroll in the study underwent breath carbon monoxide (BCO) readings to determine their smoking status. BCO levels were performed using the inexpensive, handheld TABATABA CO-Tester V2 (FIM Medical, Lyon, France). To perform the breath test based on the manufacturer’s recommendation, the women were asked to exhale completely, inhale fully and breath-hold for 15 s. At the end of the breath hold, the women were asked to exhale slowly (over 15 s) and fully into the disposable mouthpiece adapter of the CO monitor. Non-smokers with BCO readings > 3 ppm and smokers with BCO readings < 3 ppm were not included in the final analysis, as previously recommended (Javors et al. 2005, Reynolds et al. 2018). Necessary safety protocols were put in place. The peak BCO reading (the monitor was watched until the CO reading declined) was taken as the subject’s BCO level. The device measured BCO levels in parts per million (ppm). Breath- holding allows the CO in the blood to form an equilibrium with the CO in the alveolar air. This technique is responsible for the high level of correlation between breath CO levels and COHb concentration. Approval was obtained from the local ethics committee for the study (BEAH 2020 06/109). All participants gave informed consent during enrollment into the study.

Gestational age at the time of the US examination was established according to the first day of the last menstrual period and was confirmed by the first trimester US. In the case of discordance between menstrual history and the US, the first trimester of US-based gestational age was accepted. BPD, HC, atria of the posterior LV, and CM were measured and compared between the groups. Gestational age and sex-adjusted z-scores for BPD and HC measurements were created based on the previously established international reference chart (Papageorghiou et al. 2018). All sonographic measurements were performed by a highly experienced specialist (Kutuk MS). BPD represents the widest diameter of the fetal head in the transverse plane, measured outer to outer of the fetal skull, perpendicular to the midline. The HC was measured at the level of the BPD and represented the outer perimeter of the skull. The atrial width of the LV is the widest diameter of the atrium of the posterior ventricle. The width of CM is the measurement of the distance between the posterior edge of the cerebellar vermis and the interior surface of the occipital bone in the transverse plane.

The baseline characteristics of the groups are presented as mean ± standard deviation (SD), median (Q1–Q3), and percentages. The distribution of the data of the groups was assessed with Shapiro–Wilk tests. Normally and asymmetrically distributed variables were compared with the independent samples t-test and Mann–Whitney U test, respectively. BPD and HC measurements were transformed to the gestational age and gender-based z-scores and compared with independent samples- t test between groups. Categorical variables were compared with the Pearson Chi-square and Fisher exact tests. The variable with p value < .25 in the univariate analysis was included in the multivariate analysis. Linear regression analysis was performed to assess the independent variables. Enter method was used as a variable selection method. Results were evaluated at a p < .05 significance level within a 95% confidence interval. Data analysis was performed using the IBM SPSS program, version 26.0 (IBM Corp., Armonk, NY, USA).

Results

A total of 253 pregnant women enrolled in the study. Two women from the non-smoker group and one from the smoker group were not included in the final analysis since their BCO tests did not meet the inclusion criteria. As a result, overall, 250 pregnant women (smoker, n = 54, and non-smoker, n = 196) continued in the study. The baseline demographic and clinical characteristics of the groups are presented and compared in Table 1. The median number of cigarettes women smoked per day was 7/day (Q1- Q3: 4-10/day), and the mean duration was 11.08 ± 5.72 years. The two groups were similar for maternal age, and fetal gender, gravida and, parity. Body Mass Index was higher in the non-smoking group (26.86 ± 4.44 vs 25.33 ± 3.83, p = .013). CO breath test measurement was higher in the smoking group than the non-smoking group, as expected (7.18 ± 3.8 vs. 1.23 ± 0.6, p = <.001). While the rates of preterm birth and gestational hypertension were higher in smoking women, the rates of preeclampsia, intrahepatic cholestasis of pregnancy, and growth restriction were similar between groups (Table 1).

Table 1. Comparison of baseline demographic and clinical characteristics between smoking and non-smoking women.

Variable	Smokers (n: 54)	Non-smokers (n:196)	p
Age (years)	29.05 ± 4.66	29.66 ± 4.73	.4
Female fetus (%)	27 (%50)	97 (%49.5)	.535
BMI kg/m^2	26.86 ± 4.44	25.33 ± 3.83	.013
Gravida med (Q1–Q3)	2 (1–7)	1 (1–2)	.051
Parity med (Q1–Q3)	0 (1–3)	0 (0–1)	.082
CO (ppm)	7.18 ± 3.8	1.23 ± 0.6	<.001
Hematocrit (%)	35.88 ± 1.68	36.3 ± 1.61	.103
Education			.285
Uneducated or preliminary	11 (19.6)	33 (17.1)
High school	27(48.2)	89(46.1)
College	18 (32.2)	71(36.8)
Paternal smoking (Q1–Q3) (number/day)	14 (10–20)	0 (0–2)	<.001
Paternal smoking status	46 (82.1)	49(25.3)	<.001
Single motherhood	2 (3.6)	1 (0.5)	.127
ICP	0 (0)	1(0.5)	.776
Gestational diabetes	4 (7.1)	7 (3.6)	.272
Gestational hypertension	5(8.9)	2 (1)	.007
Preeclampsia	1 (1.8)	1(0.5)	.399
Preterm birth	8 (14.3)	6 (3.1)	.004
Growth restriction	3(5.4)	5(2.6)	.383

Notes: Values are expressed as n (%), mean ± standard deviation or median (min–max). p < .05, statistically significant difference. BMI: body mass index; CO: exhaled carbon monoxide concentration; ICP: intrahepatic cholestasis of pregnancy.

The gestational ages at the US examination were similar between the groups. While mean BPD measurement was lower in the smoking group (49.2 ± 3.96 vs. 50.59 ± 3.84, p = .024), gestational age-adjusted z- scores were not statistically different (−0.75 ± 1.6 vs −0.51 ± 1, p = .3). HC measurements and z- scores were significantly lower in the smoking group than in non-smoking groups (183.38 ± 14.56 vs. 189.28 ± 12.53, p = .003, 0.18 ± 1.39Mom vs. 0.56 ± 0.92, respectively, p = .023). Comparison of LV (6.18 ± 0.73 vs. 6.18 ± 0.78, p = .95), and CM (5.59 ± 1.02 vs. 5.66 ± 1.13, p = .651) revealed no significant differences between the smokers and non-smokers. The comparisons between the groups are presented in Table 2.

Table 2. Comparison of gestational age at ultrasound examination and sonographic measurements between smoking and non-smoking women.

Variable	Smokers (n: 54)	Non-smokers (n: 196)	p
Gestational age at US (days)	145.64 ± 7.72	147.34 ± 6.55	.10
BPD (mm)	49.2 ± 3.96	50.59 ± 3.84	.024
BPD (MoM)	−0.75 ± 1.6	−0.51 ± 1	.3
HC (mm)	183.38 ± 14.56	189.28 ± 12.53	.003
HC (Z-score)	0.18 ± 1.39	0.56 ± 0.92	.023
Lateral ventricle (mm)	6.18 ± 0.73	6.18 ± 0.78	.95
Cisterna magna (mm)	5.59 ± 1.02	5.66 ± 1.13	.651

Note: MoM: multiple of median. Values are expressed as mean ± standard deviation. p < .05, statistically significant difference. BPD: biparietal diameter; HC: head circumference.

Linear regression analysis showed that maternal smoking was found to be independent factor affecting fetal HC Z-score (Table 3). The relationship between maternal smoking and other cranial measurement was not statistically significant.

Table 3. Multiple linear regression analysis of independent effect of smoking on fetal cranial measurement.

	B	Std. error	p	95.0% Confidence interval for B
				Lower bound	Upper bound
HC Z-score	−0.405	0.197	.041	−0.794	−0.016
HC score	−5.153	2.406	.033	−9.892	−0.414
BPD Z-score	−0.310	0.220	.161	−0.743	0.124
BPD score	−0.309	0.221	.162	−0.744	0.126
LV	0.049	0.146	.738	−0.238	0.336
CM	0.061	0.210	.772	−0.352	0.474

Notes: Body mass index, gravida, paternal smoking, gestational hypertension, preterm birth, hematocrit level, educational status and single motherhood were taken as independent variables. HC: head circumference; BPD: biparietal diameter; LV: lateral ventricle; CM: cisterna magna.

Discussion

Our results demonstrate that continued smoking during pregnancy affects fetal skull growth with a significant reduction in fetal HC, and has no effect on BPD, LV, and CM measurements at routine second trimester US examination at mid-gestation. Previous studies have shown conflicting results on the effect of smoking on different fetal brain regions. Roza et al. assessed the development of BPD, HC, LV, and transcerebellar diameter longitudinally as part of a Generation R study and showed that smoking during pregnancy reduced fetal head growth (Roza et al. 2007). In contrast, Jaddoe et al. found no association between maternal smoking and fetal head growth in the second trimester of pregnancy in a population-based prospective cohort study (Jaddoe et al. 2007). In that study, fetal head growth significantly slowed after 25 gestational weeks. However, abdominal circumference, and femur length growth also showed decreased growth and reduced fetal head growth rate was regarded as a part of the general growth restriction process. Lampl et al. assessed the effect of smoking on the growth of different fetal parts and concluded that fetal circulation selectively spared the proximal body and relatively tended to neglect the distal extremity with resultant larger fetal head and growth-restricted leg development (Lampl et al. 2003). Hanke et al. used cotinine concentration as a proxy to tobacco exposure at 20–24 weeks of pregnancy, and found significant shortening of BPD, HC, and femur length (FL) when the measurements were adjusted for gender and gestational age. The authors did not, however, mention the ascertainment of gestational age, which limited the interpretation of their results (Hanke et al. 2004). Abraham et al. (2017) studied the existing literature regarding the effect of smoking on fetal biometric measurements throughout pregnancy, and their results showed that smoking during pregnancy reduced fetal measurements after the first trimester; specifically, reduced head size and femur length. Some authors have suggested that smoking-induced body and head growth restrictions are more prominent in male fetuses (Spinillo et al. 1995, Zarén et al. 2000). Our results supported the later studies by demonstrating that fetal HC, not BPD, reduced at midgestation in smoking mothers (Hanke et al. 2004, Abraham et al. 2017). As it is clear from the aforementioned discussion, greater body of work demonstrate detrimental effect of smoking during pregnancy on fetal head growth. However, timing and mechanism of this effect has not been proven. According to some, increased fetal hemoglobin concentration in fetus of smoking mothers increase oxygen carrying capacity and resultant normal or overgrowth of fetuses during the first half of pregnancy (Semenza 1985). Another hypothesis suggested that increased usage of intrafetal shunts such as ductus venosus and foramen ovale spare upper extremity and head growth (Lampl et al. 2003). In contrast, our results showed that reduction in the fetal head growth started before the mid gestation and affects the fetal biometric measurements. We also demonstrated that gestational age adjusted BPD was not different in fetuses from the smoking mother. This finding may indicate that smoking affect the formation of fetal head shape by modulating the closure of cranial sutures. Kozieł et al. demonstrated that smoking leads to dolicocephalic head shape postnatally and suggested that premature closure of the suturas may be responsible for this effect (Alderman et al. 1994, Kozieł et al. 2018). Though we did not measure and calculate the cephalic index, it is evident that the combination of small fetal HC with a normal BPD strongly suggests a normal or brachycephalic head shape. Therefore, our results do not confirm their conclusions. However, racial differences, methodology, gender differences, and the growth pattern after the mid-gestation may affect the interpretation of our results.

The effect of smoking on the ventricular system has not yet been largely studied. Lavezzi et al. studied the effects of smoking on the structure of the ependymal cells lining the ventricular system in the pathological specimen of victims of sudden infant death syndrome. They suggested that nicotine and its metabolites cause ependymal damage and potentially cause changes in the cerebrospinal fluid constitution (Lavezzi et al. 2010). It was also demonstrated that maternal smoking increased intracranial hemorrhage in term and preterm infants and acknowledged this effect by smoking-induced hypoxemia, intracranial hypertension and altered function of the blood-brain barrier (Spinillo et al. 1995, Matturri et al. 2011). Roza et al. were the first to study the effect of smoking on the width of the LV atrium as part of a population-based, longitudinal study. They observed that the atrial width of the LV did not change significantly throughout pregnancy between smoking and non-smoking women (Roza et al. 2007). Similarly, our results demonstrated for the first time in the literature that continued smoking during pregnancy does not change LV and CM width in pregnant women, whose smoking conditions were objectively confirmed.

The major limitation of this study was its cross-sectional design and inherent inability to demonstrate temporal change. However, our study population consisted of women who smoked until the second trimester US. We observed the effect of smoking periconceptionally and during the first and mid-trimester. Another limitation was that transcerebellar diameter (TCD) was not included in the study. Though the cerebellum was routinely evaluated in our center, TCD is not routinely recorded in electronic form if its diameter and anatomy are normal. Previous studies evaluating the effect of fetal growth restriction and smoking on the cerebellum showed that it was not affected by environmental insults. Because of its resistance to environmental factors and unhindered growth patterns, TCD is widely used for estimating gestational age during the third trimester (Campbell et al. 1994).

This study also had some strong points. Firstly, it did not include cases with other medical risk factors, such as pregestational diabetes, poor obstetrical history, multiple gestations, hypertension, drug use, and maternal epilepsy. Another advantage is that, contrary to previous reports indicating a high personal and social deprivation index in the smoking population in western societies, smoking pregnant women in this cohort were taken good family care, covered by a social security system. They did not use alcohol or any recreational drugs. Also, due to them being part of a longitudinal study, the entire participants were taken good antenatal care, nutrition counseling, supplementation and psychological support. In addition, all the cases were screened with CO breath tests, and those cases with inconsistent results were not included in the study. An experienced fetal specialist performed all the measurements in this study with the same US device and, as a result, interobserver discordance resulting from operator and US device differences was excluded. It should be stressed here that the pre-pregnancy smoking condition of our cohort could not be assessed objectively and based on the participants’ statements. This fact may have affected our results. In addition, excluding of patients with smoking-related chronic diseases (hypertension, heart disease etc.) may have caused the selection of healthier smoking women and diluted the harmful effects of smoking on fetal development.

The potential implications of our results are many. First, small fetal HC relative to BPD without other etiological factors should raise suspicion of smoking. Since it was repeatedly demonstrated that quitting smoking even at midgestation positively affects fetal growth and development (Brand et al. 2019), detection of relatively small HC may provide an opportunity for clinicians to search for smoking and give appropriate counseling at second trimester anomaly screening. Cochrane review revealed that non-pharmacological interventions to quit smoking in late gestation are effective and decrease the rate of growth-restricted neonates (Chamberlain et al. 2017). Secondly, our results have shown that fetal head growth starts to lag before the third trimester, and compensatory mechanisms believed to protect the fetal head are not effective during the first two trimester (Lampl et al. 2003). This finding may change the discourse of maternal counseling about smoking and its fetal harm. It is obvious that detection of smoking-related feal harm at midgestation may prompt or motivate mothers to quit smoking with appropriate counseling, support and follow-up. The third point that our outcomes implied is that, effect of smoking on fetal growth should always be based on an objective assessment of maternal smoking. Shisler et al. (2017) elegantly demonstrated that maternal smoking based on the self- report was not related to fetal growth. However, the more intensive calendar-based self-report measure or biological assays were significant predictors of fetal growth. In most population-based studies, a smoker is defined as a woman who smoked at least one cigarette during pregnancy, either daily (at least one cigarette every day) or occasionally (at least one cigarette per occasion, which indicates less than daily) (Lange et al. 2018). On the other hand, nearly 20% of smoking women hide their smoker status due to fear of being stigmatized (Grant et al. 2020). Consequently, it is possible that some of the patients in these cohorts did not represent the group they were supposed to.

In conclusion, continued smoking during pregnancy reduces fetal HC, and has no effect on BPD, LV or CM measurements at mid-gestation. Since smoking is well known for its early and late childhood behavioral and neurological consequences, smaller mid-trimester fetal HC measurements should bring maternal smoking to mind as one of the potentially reversible causes. With appropriate counseling and support, it is possible for a mother to quit smoking and consequently to stop or reverse the harmful effects of tobacco.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.