Relationships between prenatal smoking cessation, gestational weight gain and maternal lifestyle characteristics

Introduction 

Maternal smoking has been well documented as a risk factor for low gestational weight gain, low birth-weight, fetal growth restriction and preterm birth,1, 2, 3 and may explain why the spontaneous quit rate during pregnancy exceeds that in the general population of women.4

In non-pregnant women, smoking cessation is associated with an average weight increase of 5.0kg after 1 year of quitting.5 Despite the established effect of smoking on weight gain, relatively little attention has been devoted to the relationship in the context of pregnancy. Most studies regarding maternal smoking habits have focused on the comparisons between smokers and non-smokers and only few have considered smoking cessation in their analyses.6, 7, 8, 9 These studies found that women who quit smoking after conception gained from 1.2 to 3.5kg more during gestation than continuing smokers.6, 7, 8, 9 Therefore, maternal weight concerns might be an obstacle of maintaining smoking abstinence over the postpartum period after spontaneous quitting during pregnancy.10, 11

Women naturally gain weight during pregnancy, typically 12–14kg, and most of them gradually lose this weight postpartum.12 However, for some women pregnancy-related weight gain may trigger obesity development. Thus, it is important to advise pregnant women to gain weight within recommended ranges, which secures a healthy fetal development without causing postpartum weight retention and consequently maternal obesity.13, 14

In 1990, the American Institute of Medicine (IOM) published guidelines for weight gain during pregnancy based on pre-pregnancy body mass index (BMI).15 After the guidelines were published, systematic reviews including studies from different populations have supported the soundness of the IOM recommended ranges of weight gain for decreasing the occurrence of negative pregnancy outcomes.12, 16 However, at least 30% of Western women gain weight above the recommendations.16

It is known that gestational weight gain is influenced by numerous lifestyle changes during pregnancy, such as physical activity level, food intake, alcohol consumption and smoking habits.17, 18 Although previous studies have indicated that smoking cessation during pregnancy is a potential predictor of high gestational weight gain,6, 7, 8, 9 changes in other lifestyle habits during pregnancy were not taken into consideration in the analyses. Women who quit smoking during pregnancy may also adopt other healthy lifestyle measures because of concerns about fetal well-being.4

This study, therefore, aims to describe maternal socio-demographic and lifestyle characteristics associated with smoking habits during pregnancy and to evaluate the independent effect of quitting smoking in early pregnancy on gestational weight gain, regardless of other maternal lifestyle habits. It is hypothesised that women who quit smoking are (1) prone to adhere to a generally healthy lifestyle, regarding eating regularity, physical activity and alcohol consumption and (2) more likely to gain weight above the IOM recommendations.

Methods 

The Stockholm Pregnancy and Weight Development Study (SPWDS) is a 1 year follow-up study conducted to identify risk factors for postpartum weight retention. Women who gave birth in 1984/1985 in 14 selected maternity units in Stockholm, Sweden were invited to participate in the study by the units’ staff after birth. Women who agreed to participate attended two weigh-in sessions at 6 months and 1 year postpartum. On these occasions, they filled out questionnaires about their lifestyle before, during, and after pregnancy. Information about gestational weight gain and gestational age was collected from obstetrics records. Detailed information about the year-long study is described elsewhere.19, 20 The data used in this study were restricted to information collected at the 6-month follow-up visit. Over the years, numerous post hoc analyses have been conducted, extracting the salient information of this database.13, 19, 20, 21, 22

Initially, 2342 women agreed to participate in the study. Forty-seven women were excluded due to complicated pregnancies and lack of pre-pregnancy weight reports, leaving a sample of 2295 women. After 348 losses to follow-up (15% of 2295), 1947 women completed the initial 6-month follow-up. For the purpose of this study, 159 women were excluded owing to missing information on gestational weight gain, infant birth-weight and smoking habits. Due to known ethnic differences in the relation between weight gain and obesity, non-Nordic women (n=35) were excluded, leaving a final sample of 1753 women.

Results 

Mean maternal age at birth was 29.6 (SD 4.7), ranging from 18 to 45 years. In the study sample, 22.2% of the women identified themselves as current smokers and 11.6% reported smoking cessation during pregnancy. As shown in Table 1, quitters were younger than non-smokers. Compared to non-smokers and quitters, smokers were more likely to be single, have infants with the lowest birth-weight and gestational age, have a previous child and were less likely to give birth through normal vaginal method. There were no significant differences in pre-pregnancy BMI and occupation across groups.

Table 1. Maternal socio-demographic and biological characteristics according to prenatal smoking status.

		Non-smokers	Quitters	Smokers	p-Value
		n=1159	n=390	n=204
	Socio-demographic characteristics
	Age (years)a	29.9 (4.5)	28.6 (4.9)	29.0 (4.8)	<0.001
	Civil statusb				<0.001
	Single/divorced	5.2%	8.4%	14.8%
	Occupation during pregnancyb				0.381
	Unemployed/student	10.3%	13.2%	9.7%
	Biological and reproductive characteristics
	Parityb				<0.001
	No previous child	51.6%	65.7%	48%
	Birthing methods				0.05
	Normal	89.7%	86.3%	85.6%
	Excess weight gainb				<0.05
	Yes	27.4%	37.2%	27.8%
	Total weight gain (kg)a	14.1 (4.0)	15.3 (4.4)	13.8 (4.3)	<0.001
	Pre-pregnancy BMI (kg/m2)a	21.6 (2.7)	21.2 (2.4)	21.5 (2.8)	0.181
	Gestational age (weeks)a	39.6 (1.7)	39.4 (1.7)	39.3 (1.8)	<0.01
	Infant birth-weight (kg)a	3.5 (0.5)	3.4 (0.5)	3.3 (0.5)	<0.001

a ANOVA. b Chi-square test.

Quitters gained significantly more weight during pregnancy than both non-smokers (1.2kg; p<0.001) and smokers (1.5kg; p<0.001), but the weight gain difference between non-smokers and smokers (0.3kg; p=0.7) was not statistically significant. Although quitters had the highest gestational weight gain, they gave birth to infants with almost the same mean birth-weight as non-smokers. In total 37.2% of quitters versus 27.4% of non-smokers gained weight above IOM recommendations (p<0.01). There was no significant difference in the percentage of women who gained excess weight between non-smokers (27.4%) and smokers (27.8%).

Table 2 compares maternal lifestyle characteristics according to prenatal smoking status. Non-smokers were more likely to be physically active before and during pregnancy and have regular eating habits (meals frequency). Smokers had the highest frequency of low breakfast quality score.

Table 2. Maternal lifestyle characteristics according to prenatal smoking status.

		Non-smokers	Quitters	Smokers	p-Value
		n=1159	n=390	n=204
	Physical activity
	Level before pregnancy				<0.001
	Inactive	14.5%	19.8%	25.5%
	Medium	49.8%	50.8%	53.5%
	High	35.7%	29.4%	21.0%
	Level during pregnancy				<0.01
	Inactive	27.5%	36.6%	36.6%
	Medium	59.9%	52.1%	54.8%
	High	12.6%	11.3%	8.6%
	Activity practice before and during pregnancy				<0.001
	Inactive before and during pregnancy	12.5%	17.5%	21.9%
	Active before and inactive during pregnancy	15.4%	18.7%	14.9%
	Inactive before and active during pregnancy	2.1%	2.8%	3.8%
	Active before and during pregnancy	70.0%	61.0%	59.4%
	Food habits (frequency/quality)
	Food craving during pregnancy				0.08
	Yes	50.6%	58.68%	53.6%
	Having breakfast every-day during pregnancy				<0.01
	Yes	93.6%	90.1%	88.0%
	Having lunch every-day during pregnancy				<0.001
	Yes	72.8%	69.0%	60.0%
	Consumption of snacks/sweets between meals during pregnancy				0.426
	≥3 times/day	37.1%	41.7%	36.7%
	Frequency of cooked meals during pregnancy				0.364
	≥2 times/day	59.7%	59.3%	55.6%
	Breakfast quality during pregnancy				<0.001
	Low	22.9%	24.7%	40.5%
	Medium	48.6%	42.9%	39.6%
	High	28.5%	32.4%	19.8%
	Breakfast quality before and during pregnancy				<0.001
	Low quality before and during pregnancy	20.7%	22.0%	37.0%
	Medium/high quality before and low during pregnancy	1.2%	2.2%	2.1%
	Low quality before and medium/high during pregnancy	1.8%	3.5%	2.7%
	Medium/high quality before and during pregnancy	76.3%	72.3%	58.1%
	Alcohol consumption during pregnancy				<0.01
	No	26.1%	19.1%	20.9%
	Quit	58.5%	64.8%	68.2%
	Yes	15.4%	16.1%	10.9%
	Dietary advice during pregnancy
	Dietary advice				0.132
	Yes	62.2%	54.6%	60.2%
	Lack of information about diet during pregnancy				0.527
	Yes	30.0%	32.5%	28.0%

Chi-square test performed.

The percentage of women who received dietary advice or reported lack of information about diet during pregnancy did not differ across smoking groups. Among those who received dietary advice during prenatal care, there were no differences between non-smokers, quitters or smokers whether advice was given by midwives, dietitians or via brochures. The majority (68%) received advice from midwives (results not shown). Similarly, no differences were seen across smoking groups regarding food craving and any particular food group (results not shown).

Lifestyle changes during pregnancy were more frequent (p<0.05) among quitters who reported the highest frequency (3.5%) of improvement in their breakfast quality from low quality before pregnancy to medium or high quality during pregnancy, compared to non-smokers (1.8%) and smokers (2.7%). However, the group of quitters also had the highest proportion of women (18.7%) who were previously active, but stopped exercising during pregnancy. Conversely, alcohol consumption throughout pregnancy was more frequent among quitters than was in others.

Interestingly, alcohol abstinence was more frequent among smokers (68.2%) than non-smokers (58.5%). However, smoking abstinence was similarly frequent among smokers and quitters (68.2% versus 64.8% p=0.2). Among women who stopped drinking only 10% reduced smoking intensity from >10 cigarettes/day before pregnancy to ≤10 cigarettes/day during pregnancy (results not shown).

Table 3 displays the results of the logistic regression models. As expected, in the full model, women who quit smoking during pregnancy had greater risk (OR=2.0; 95%CI 1.4–3.0) of gaining weight above IOM recommendations compared to women who never smoked, after adjustment for infant birth-weight, gestational age, maternal parity, pre-pregnancy BMI, alcohol consumption, physical activity and breakfast frequency. Unexpectedly, smokers had 30% greater risk of having excess weight gain than non-smokers, though this difference was not statistically significant (OR=1.3; 95%CI 0.9–1.8).

Table 3. Association between excess gestational weight gain and prenatal smoking status.

	Smoking status	Crude			Adjusteda
		OR	95%CI	p-Value	OR	95%CI	p-Value
	Quit smoking	1.6	1.1/2.1	<0.01	2.0	1.4/3.0	<0.0001
	Smoking	1.0	0.8/1.3	0.922	1.3	0.9/1.8	0.09

Reference category: never smoking.

a Adjusted for birth-weight, gestational age, parity, pre-pregnancy BMI, alcohol consumption, physical activity before and during pregnancy (active/inactive) and breakfast frequency during pregnancy.

Discussion 

In this study, prenatal smoking cessation was an independent predictor of gaining weight above IOM guidelines. This finding is consistent with other studies,6, 8 but not all.23 In the latter study,23 smoking cessation doubled the risk of gaining excess weight, but the association was no longer significant after adjustment for dietary factors. In addition, it is possible that failure to adjust for alcohol consumption, food habits and physical activity in the former studies6, 8 may have either under- or overestimated the true association between smoking cessation and weight gain.

The multivariate model showed that smokers had a slightly higher chance of gaining excess weight than non-smokers, but the estimate was not statistically significant and this finding was consistent with previous reports.1, 6 There is evidence that any reduction in smoking habits has the potential to result in additional weight gain.6, 24 Thereby, this finding might reflect the decrease in numbers of cigarettes smoked during pregnancy, especially among heavy smokers, and inclusion of women who either tried to quit smoking in early gestation, but relapsed or quit smoking at advanced stage of pregnancy.25 Stratifying the analysis according to the time of quitting (early versus late pregnancy) was not possible, due to the few number of women included in these groups. Furthermore, it was not possible to assess whether the relationship between smoking and weight gain was dose-dependent. The dose–response analysis had limited statistical power to detect difference between groups because few women reduced smoking intensity from>10 cigarettes/day before pregnancy to ≤10 cigarettes/day during pregnancy.

Maternal characteristics differed significantly according to prenatal smoking status. As expected, smokers were more likely to report low quality of breakfast and irregular eating habits, while non-smokers were more likely to have a healthier lifestyle. Quitters also adopted healthier eating habits (e.g. improvement in their breakfast quality). Surprisingly, alcohol abstinence was more frequent among smokers than non-smokers. There are two possible interpretations of this finding. The first is that smokers underreported their consumption of alcohol and the second is that they were concerned about fetal health. Smoking is a well-known risk factor for negative pregnancy outcomes and smoking mothers possibly tried to avoid another risk factor by suspending alcohol consumption during pregnancy.

About one third of women reported that they would like to have received more information about diet during pregnancy. Self-perceived lack of dietary advice was significantly associated with excess weight gain. However, the association was abolished after adjustment for pre-pregnancy BMI. It is unclear whether the advice given was ineffective, or whether women did not accurately recall the advice they had actually received. Dietary advice from a midwife or from elsewhere was not statistically significant in explaining excess weight gain during pregnancy. Regrettably, a detailed description of the dietary advice and further information on the women's views on dietary advice was not available. Reverse causality is also a potential concern because women were asked about dietary advice after birth. Women who gained excess weight or were overweight prior to gestation might be inclined to report lack of dietary information during prenatal care.

This study included a large number of women and hence required a simplified dietary and physical activity questionnaire. It remains uncertain if the significant association between smoking cessation and maternal weight gain after controlling for lifestyle factors was due to limitations of the questionnaire in capturing maternal food habits. Although an accurate measure of energy intake was not used in this study, breakfast frequency and quality were found to be inversely associated with obesity and chronic disease in a recent review.26 This might indicate that mealtime regularity and quality can be used as surrogate measures for maternal dietary habits.

Recall bias is also a concern as women might not have reported accurately or remembered their dietary habits or degree of physical activity before and during pregnancy. Therefore, the method used might have given an inaccurate picture of maternal habits. Nevertheless, lifestyles matters are better recalled in connection with pregnancy than during other periods in life.27 It is possible that some women have under-reported their dietary intake and over-reported their physical activity level. However, the tendency for women to give an embellished picture of their habits during pregnancy is as likely to occur when using short-method as with more detailed questionnaires.

Gestational weight gain was calculated using maternal weight measured at the end of gestation at the maternity unit. This procedure ensured that women did not experience additional weight gain that was not accounted for. Although one important limitation was the use of self-reported pre-pregnancy weights, which is prone to bias, our previous validation study showed that the information was of good quality.19

Smoking habits were also self-reported. Reliance on self-reported smoking status may result in exposure misclassification because of increased concern of the negative effects of smoking during pregnancy.28 In this study, smoking prevalence was lower (22.2%) than the national statistics (31%) for 1983.29 About 34% of previous smokers (n=594) quit smoking upon pregnancy. This estimate is in accordance with previous literature reporting that 20–40% of smokers quit after conception and that the majority do so in early pregnancy.4 If women reported smoking cessation while continuing to smoke, it would underestimate our results.

Additionally, smoking rate was higher among non-completers than completers and it may indicate the presence of selection bias or a ‘healthy participant effect’. However, the extensive analyses of losses to follow-up showed that, for most variables under study, the remaining women were reasonably representative of the initial cohort.19, 21

The prevalence of obesity in this study (2.0%) was lower than the corresponding national prevalence (5.6%) at that time.30 Since obesity prior to pregnancy is a strong predictor of excess gestational weight gain,31 it seems that the higher risk group was underrepresented in this study. This may have led to an underestimation of our findings.

The prevalence of obesity in Sweden is lower than in many other Western countries. Meanwhile, the increasing trend of obesity is alarming. In 2004/2005 about 8% of women of childbearing age were obese.32 Similarly, smoking prevalence in Sweden is lower compared to some Western countries. In 1993, a comprehensive Tobacco Act was implemented and the adverse effects of cigarette smoking during pregnancy have been well publicised since. Consequently, the prevalence of daily cigarette smoking among pregnant women has been declining steadily. In 2003, 10% of pregnant women in Sweden were daily smokers in early pregnancy.29 This decline was due to decrease in smoking initiation and increase in smoking cessation in pregnancy. Furthermore, the weight gain data were recorded before the release of IOM guidelines. There were no formal recommendations for gestational weight gain in Sweden in 1984/1985. In light of these particularities, it is unknown to what extent these results could be generalized to other populations.

Notwithstanding our data were collected in 1984/1985, the topic of gestational weight gain associated with smoking cessation during pregnancy is still timely due to the scarcity of studies addressing this issue. The findings preseted in this study could be used as a groundwork for future studies where detailed information on important counfounder, such as, dietary habits and physical activity level, might be available. Additionally, there is no reason to suspect that biological mechanisms behind post-cessation weight gain have changed from the 1980s to 2000s. Since our data are from the 1980s, when smoking cessation during pregnancy and maternal obesity were less prevalent, it is speculated that the risk of gaining excess weight during pregnancy might be of greater magnitude among contemporary Swedish women.

Despite the potential limitations, this is one of a few studies with information on prenatal smoking cessation together with other relevant maternal lifestyle factors both before and during pregnancy in a large sample.

Conclusion 

This study indicated that smoking cessation in early pregnancy doubled the likelihood of gaining weight above IOM recommendations even after controlling for other lifestyle factors. Excess gestational weight gain may ultimately increase the risk of postpartum weight retention and long-term obesity.13, 33 Nonetheless, the literature shows that smoking cessation might reduce the risks of premature birth, intrauterine growth restriction and maternal smoking-related cancers and cardiovascular diseases.9 Therefore, an increased weight gain among quitters does not override the potential benefits of smoking cessation during pregnancy.

Our findings highlight the need for supportive measures to control weight gain among women who quit smoking during pregnancy. Epidemiological studies suggest that pregnant women are more likely to gain weight within the recommended guidelines if their physicians or midwives instruct them to do so.34, 35 If health professionals identify women who quit smoking in early pregnancy and give individual advice regarding healthy weight gain during pregnancy the number of quitters and long-term abstainers may increase with a reduction in the number of women resuming smoking following birth.36 Interventions designed to prevent excess weight gain associated with smoking cessation should encourage women to remain physically active and adopt healthy eating habits during pregnancy.