Department of Public HealthUniversity of Helsinki

Passive Smoking During Pregnancy and Early Childhood:

Occurrence, Determinants, Health Effects and Prevention

Niina Jaakkola

Academic Dissertation

To be presented, with the permission of the Faculty of Medicine of the University of

Helsinki, for public criticism in Auditorium XIII, University Main Building,

on October 11th, 2002, at 12 o’clock

Yliopistopaino, Helsinki 2002

ISBN 951-45-8517-8 (nid.)ISBN 952-10-0659-5 (PDF)ISSN 0355-7979

Supervisor: Professor Jouni J. K. JaakkolaEnvironmental Epidemiology UnitDepartment of Public HealthUniversity of Helsinki, andThe Nordic School of Public HealthGöteborg, SWEDEN

Reviewers: Professor Markku KoskenvuoInstitute of Clinical Medicine, Public HealthUniversity of Turku

Docent Ossi RahkonenDepartment of Social PolicyUniversity of Helsinki

Opponent: Professor Arja RimpeläTampere School of Public HealthUniversity of Tampere

To Jyrki and Simo-Pekka

Niina Jaakkola Abbreviations6

ABBREVIATIONS

BMI body mass index

CI confidence interval

ETS environmental tobacco smoke

LBW low birth weight

OR odds ratio

SD standard deviation

SGA small for gestational age

Abstract Niina Jaakkola7

ABSTRACT

Although small children rarely smoke, tobacco smoke in their environment constitutes an

extensive health hazard. The World Health Organization has estimated that, globally, almost

half of all children are exposed to environmental tobacco smoke.

The general objective of this dissertation was to address public health issues related to prenatal

and postnatal exposure to tobacco smoke products. The specific objectives were 1) to study the

occurrence, trends, and determinants of smoking during pregnancy in Finland (Study I); 2) to

study the occurrence and determinants of exposure to environmental tobacco smoke (ETS) during

childhood in Finland (Study II); 3) to study the effects of exposure to ETS during pregnancy on

fetal growth and length of pregnancy (Study III); and 4) to evaluate the effects of a population-

based smoking cessation program on smoking cessation and fetal exposure to tobacco smoke

products (Study IV).

These studies were all population-based representing geographically defined source

populations. There were three population levels, nationwide, regional, and local.

The nationwide study population consisted of 694,926 pregnant women in 1987-97 recorded in

The Finnish Medical Birth Registry (Study I). On the local level, the source population

consisted of all children in the city of Espoo born between 1.12.1984 - 31.12.1989 (Study II).

The study population included a total of 2,568 children whose parents filled in the

questionnaire (response rate 80.3%). The study focused on the 1,003 (39.1%) children whose

one or both parents or guardians were current regular smokers. On the regional level, Studies

III and IV were based on The Finnish Prenatal Environment and Health Study and included all

the 2,751 children born 1.5.1996 – 30.4.1997 in two geographically defined hospital districts in

South East Finland. Study III focused on 389 singleton newborns. In Study IV, the intervention

group consisted of all women who were smokers at the time of their first visit to maternity

health clinics and who gave birth in The Kymi Hospital District (306/1917). A reference group

was chosen using identical inclusion criteria from The Porvoo Hospital District (152/834).

Study I showed that smoking in pregnancy did not change in Finland between 1987 (15%) and

1997 (15%). Smoking was inversely related to age. It was more common in single than in

married, and in occupations with low education requirements. The prevalence of smoking was

Niina Jaakkola Abstract8

highest in northern and eastern Finland and lowest in western Finland. The determinants of

smoking remained similar, but the regional differences increased over the study period.

Study II provided evidence that low education and single parenthood increases the risk of ETS

exposure at home in children of smoking parents. The risk of exposure was found to be smaller

for atopic children, and asthmatic children were not exposed at all.

In Study III, the risk of preterm birth (<37 weeks) was increased in the presence of ETS

exposure measured as hair nicotine concentration in non-smoking mothers after delivery. There

was also an inverse linear relation between the mother’s hair nicotine concentration and the

child’s mean birthweight.

Study IV comprised a controlled evaluation of a smoking cessation intervention based on the

training of maternity care nurses and providing them with health education material. The

reference group received ordinary maternity care health education. The difference in quitting

smoking during pregnancy between the intervention group and the reference group was 4.5%

(95% CI −2.6 − 11.6%), indicating a small effect of intervention. The intervention group

indicated that they received more information on the adverse effects of smoking, studied the

material more actively, and felt that it influenced their smoking behavior more than the

reference group did. There was no fference in the experience of receiving support for attempts

to quit smoking between the two groups.

In spite of increasing knowledge of its adverse effects, smoking during pregnancy has not

declined in Finland over the past decade. The present results indicate that the resources of health

education in the prevention of passive smoking should be focused on single parents and parents

with a low educational background. Future smoking cessation programs need to offer more

support to those attempting to quit. They should also emphasize the harmful effects on

women’s own health as an additional reason for quitting. A task for the future is to find

methods that are both effective and feasible to be used as a matter of routine maternity clinics.

The effectiveness of these methods could be assessed in randomized controlled trials.

However, the effectiveness and feasibility should also be evaluated in settings in which the

influence of the evaluation itself is minimal.

Original publications Niina Jaakkola9

LIST OF ORIGINAL PUBLICATIONS

This thesis is based on the following original publications referred to in the text by the Roman

numerals I-IV.

I. Jaakkola N, Jaakkola MS, Gissler M, Jaakkola JJK. Smoking during pregnancy in

Finland: determinants and trends, 1987-1997. Am J Public Health 2001;91:284-286.

II. Jaakkola N, Ruotsalainen R, Jaakkola JJK. What are the determinants of children’s

exposure to environmental tobacco smoke at home? Scand J Soc Med 1994;21:107-

112.

III. Jaakkola JJK, Jaakkola N, Zahlsen K. Fetal growth and length of gestation in relation to

prenatal exposure to environmental tobacco smoke assessed by hair nicotine

concentration. Environ Health Perspect 2001;109:557-561.

IV. Jaakkola N, Zahlsen K, Jaakkola JJK. Effects of a population-based smoking cessation

programme on smoking in pregnancy. Eur J Public Health 2001;11:446-449.

The original publications are reproduced with the permission of the copyright holders.

Niina Jaakkola Introduction10

1. INTRODUCTION

Although small children rarely smoke, tobacco smoke in their environment constitutes a

significant health hazard. The World Health Organization has estimated that globally almost

half of all children are exposed to environmental tobacco smoke (ETS) (WHO 1999). There is

accumulating evidence of the diverse health effects of exposure to ETS in children. Exposure

during pregnancy is associated with low birth weight (LBW), small for gestational age (SGA),

perinatal death, and reduced lung function (Kramer 1987, Haddow et al. 1988). There is strong

consistent evidence that exposure to ETS in childhood increases the risk of early respiratory

infections, causes chronic respiratory symptoms, and increases exacerbation of asthma (Cook

and Strachan 1999, Nafstad et al. 2000, Jaakkola and Jaakkola 2002). There is also strong

evidence of an increased risk of asthma related to ETS exposure (U.S. EPA 1992, DiFranza et

al. 1996, Larsson et al. 2001).

This literature review first elaborates on our knowledge about prenatal exposure to tobacco

smoke products, defined as active and passive smoking in pregnant women. Second, the trends

and determinants of exposure to ETS in children are reviewed. Further, fetal effects of

exposure to ETS in pregnancy are discussed. Finally, evaluations of smoking cessation

interventions in pregnant women are reviewed.

1.1 Environmental Tobacco Smoke

ETS is a complex mixture containing over 4,000 compounds, including more than 40 known or

suspected human carcinogens as well as other toxic agents (U.S. EPA 1992). ETS or passive

smoking is divided into main stream and side stream smoke. Main stream smoke is defined as

tobacco smoke generated during puff-drawing in the burning cone of a tobacco product, which

then travels through the unburnt tobacco and is inhaled directly by the smoker (Spitzer et al.

1990). Side stream smoke is defined as a combination of smoke emitted into the air during the

burning of a tobacco product between puffs, smoke escaping into the surrounding air during

puffs, and smoke components that diffuse through cigarette paper (Spitzer et al. 1990).

Exposure from breathing airborne tobacco smoke products is what is commonly called ETS

exposure.

Introduction Niina Jaakkola11

The fetus cannot be an active smoker, but it can be considered a passive smoker as a result of a

pregnant woman’s active and/or passive smoking. There are several suggested mechanisms for

assessing the effects of tobacco smoke on fetal growth (Eskenazi et al. 1995). While most

studies have been carried out on active smokers, because the side stream smoke contains the

same compounds as main stream smoke, it is likely that the mechanisms are similar in both

active and passive smoking. Smoking in pregnancy is associated with numerous

complications for both the mother and the baby. The effects start in utero, resulting in

increased perinatal mortality and morbidity. It is a major cause of sudden infant death

syndrome (Haglund and Cnattingius 1990), and there are other well-documented health effects,

including reduced birth weight and decreased lung function (WHO 1999).

Strong public programs to protect children from exposure to ETS are essential. These programs

should aim to ensure the right of every child to grow up in an environment free of tobacco

smoke.

1.2 Smoking during pregnancy and prenatal exposure to tobacco smoke products

Smoking during pregnancy constitutes exposure to ETS for the fetus. The adverse effects on

fetal growth were first recognized by Simpson and Loma (1957), and quantified in a meta-

analysis by Kramer (1987). Prenatal exposure to tobacco smoke through maternal smoking has

also been linked to an increased risk of perinatal death (Cal EPA 1997). From the public health

point of view, smoking during pregnancy is among the most important preventable factors.

A series of studies conducted in Europe (Cnattingius et al. 1997, Eriksson et al. 1996 and 1998,

Owen et al. 1998, Chatenoud et al. 1999) and North America (Dodds 1995, Stewart et al. 1995,

Kirkland et al. 2000, Ebrahim et al. 2000), summarized in table 5.1, reported the prevalence of

smoking in pregnancy in the 1980s and 1990s. The prevalence of smoking varied from 34% in

Norway in 1987 to 12% in Italy in 1999. Some studies have also provided information on the

quantity of smoking (Eriksson et al. 1996, Eriksson et al. 1998, Ebrahim et al. 2000). The mean

number of cigarettes smoked per pregnant smoker per day decreased in Norway from 8.6 in

1987 to 7.0 in 1994 (Eriksson et al. 1996). At 18 weeks of pregnancy, the mean number was

7.3 (Eriksson et al. 1998). In the United States, the median number of cigarettes smoked per

day by pregnant smokers remained at 10 from 1987 to 1996 (Ebrahim et al. 2000).

Niina Jaakkola Introduction12

A total of six studies reported prevalences over time in a given population (Stewart et al. 1995,

Dodds 1995, Eriksson et al.1996, Cnattingius et al. 1997, Owen et al. 1998, Ebrahim et al.

2000). In 1966, 12% of women in Northern Finland continued smoking throughout their

pregnancy, and in 1985-86 the prevalence of smokers was 18% (Isohanni et al. 1995). In

1981-1982, approximately 25% of Finnish pregnant women smoked at some point in

pregnancy and 15% continued smoking throughout (Tikkanen 1984). In England, the

prevalence of smoking during pregnancy was 26% in 1992, and it did not change substantially

between 1992 and 1997 (Owen et al. 1998). These results suggest that current practices to

prevent smoking during pregnancy are not working, or that there is a lack of sufficient

investment in and prioritization of effective implementation. Similar results were seen in Nova

Scotia, where the smoking rate changed little between 1988 and 1992 (Dodds 1995). A study

based on the Swedish Birth Registry reported that the prevalence of daily smoking among

pregnant women decreased from 29% in 1983 to 22% in 1992 (Cnattingius et al. 1997). The

attributable fraction of SGA newborns due to smoking decreased from 26% to 21% during the

same time period. A similar reduction was reported in Canada from 29% in 1983 to 19% in

1992 (Stewart et al. 1995), and in the United States from 16% in 1987 to 12% in 1996

(Ebrahim et al. 2000). These differences were mainly due to a reduction in the proportion of

women who smoked before pregnancy.

Women’s smoking is an important determinant of both prenatal and postnatal exposure to

tobacco smoke products in children. In Finland, there was a steady decline in the prevalence of

smoking in men, from 37% in 1978 to 33% in 1990, but unfortunately the prevalence in

women increased from 17% to 20% during the same time period (Rahkonen et al. 1992). The

National Public Health Institute has been monitoring health behavior in Finland since 1978 by

using annual postal surveys. In 2000, men’s daily smoking prevalence was 27%, and women’s

was on the same level (20%) as in 1990 (Helakorpi et al. 2000). In 1996-97, 19% of women

were daily smokers (Helakorpi et al. 2001). Smoking in young women has an especially strong

impact on both prenatal and postnatal exposure. The Finnish Adolescent Health and Lifestyle

Survey has followed teenagers’ smoking habits since 1977 (Rimpelä et al. 1999). The

prevalence of daily smoking in 14- to 18-year-old boys was slightly lower in 1999 (26%) than

in 1981 (28%), but the prevalence in girls increased from 21% to 25%. This increase also leads

to an increase in smoking in pregnancy, unless the smoking cessation rate does not increase

correspondingly in pregnant women and women who plan to become pregnant.

Introduction Niina Jaakkola13

This review indicates that smoking in pregnancy is alarmingly common in Europe and North

America in spite of accumulating evidence of serious adverse effects on the fetus.

1.3 Occurrence and determinants of passive smoking in children

It is essential to know the occurrence and determinants of exposure to ETS during pregnancy

and early childhood to be able to develop ways to prevent it. Several studies conducted in

different parts of the world documented the occurrence of children’s exposure to ETS in the

1990s (Table 5.2), and some studies also assessed the determinants of exposure.

In most of the studies, children’s exposure to ETS was based on questionnaire information

(Eriksen and Bruusgaard 1995, Brenner and Mielck 1993, Lund et al.1998a), and in two

studies (Cook et al. 1994, Pirkle et al. 1996) both questionnaires and measurements of cotinine

were used. One of the studies used cotinine concentrations as a measure of exposure to ETS,

and changes in concentrations over time were analyzed to describe time trends (Jarvis 2000).

A cross-sectional study was conducted in Norway among families attending health care centers

for healthy-child visits (Eriksen and Bruusgaard 1995). The parents were asked to fill in a

questionnaire on smoking habits beforehand. All the information in this study was based on

parental reports, and the parents may have tended to underreport their smoking. In 48% of the

families, at least one adult smoked, and 53% of these smoking families reported that they

smoked indoors. Smoking was found to be more frequent among single parents, the less

educated and older parents.

A recent cross-sectional study assessed children’s exposure to ETS in all the Nordic countries

using anonymous questionnaires distributed to a random sample of parents with a 3-year old

child (Lund et al. 1998b). While the prevalence of household smoking was similar (25%) in all

countries, there was a great difference between them with regard to the children’s ETS

exposure. Finnish parents were more likely than all other Nordic parents to protect their

children from ETS. The situation was worst in Denmark and Iceland, where almost half of all

households, and nine out of ten households with daily smokers, exposed their children.

Approximately 80% of current and former smokers in all countries stated that they had made

efforts to change their smoking habits because of their children (Lund et al. 1998a). The data

showed that single parents and parents with lower socio-economic status exposed their children

Niina Jaakkola Introduction14

more. The study did not assess cotinine levels to validate the reported exposure to ETS, so

there may have been some under-reporting, although the anonymity probably decreased this.

The response rates were lower in Norway and Denmark than in Finland, Sweden or Iceland.

Regulations in Norway allowed only one reminder to be sent, and in Denmark they did not

permit any reminders. Two reminders were sent in all the other countries.

In West Germany, the occurrence of smoking in households was assessed in a national health

survey conducted in 1987 (Brenner and Mielck 1993). The prevalence of at least one smoking

household member was 50% among 2- to 5-year-old children, while the majority of children

aged < 2 years (59%) and 6 to 13 years (68%) were living in a household with at least one

smoker. For all age groups, the prevalence of smoking was higher among less-educated

parents.

The prevalence of exposure to ETS in children was assessed in England and Wales using a

sample of 10 primary schools in 10 towns (Cook et al. 1994). The study population consisted

of 4,043 children aged 5 to 7 years. The parents were asked about their current smoking habits,

and saliva cotinine was collected from the children. A total of 53% of the children were

exposed to cigarette smoke at home or by an outside carer. The effect of maternal smoking on

cotinine concentrations was greater than that of paternal smoking, probably because the

mothers took greater responsibility for childcare.

In England, regular national surveys of smoking are conducted among secondary-school

children aged 11 to 15 years (Jarvis et al. 2000). Since 1988, samples of saliva cotinine

concentration have been taken randomly in half of the schools surveyed. Cotinine

concentrations in all non-smoking children almost halved between 1988 and 1998. These

results show a substantial decline in children’s exposure to passive smoking in England. Such

changes could be due to several factors, the most important of which is a decrease in parental

smoking (Table 5.2). However, the continuing reductions in cotinine concentrations in children

living in non-smoking households suggest that changes in society have led to less smoky

environments. The results of this study indicate trends in passive smoking in society generally,

and are not exactly comparable to studies that focused on children’s exposure to ETS at home.

It is possible that exposure to passive smoking in homes did not change as much as in public

places where tobacco bans were imposed.

Introduction Niina Jaakkola15

Population exposure to ETS was assessed in the United States in nationally representative

cross-sectional surveys in 1988 – 1991 (Pirkle et al. 1996). These surveys consisted of a

parent-administered questionnaire for persons aged 2 months and older, and measurements of

serum cotinine from persons aged 4 to 11 years. One member of the household was asked

whether any of the household members smoked cigarettes in the home. For children from 4 to

11 years of age (n=3,011), the prevalence of reported ETS exposure at home was 43%.

Mannino and coworkers (2001) analyzed data among children aged 4 through 16 years from a

nationally representative sample of US children (n=5,653), both with and without reported

tobacco exposure in their homes, to determine what factors predict cotinine levels in children.

Most children in this sample had detectable levels of cotinine in their blood, reflecting

exposure to tobacco smoke. Among those with reported exposure, non-Mexican-American

race/ethnicity, young age, a low number of rooms in the home, low parental education, and an

increasing number of cigarettes smoked in the home were predictors of increased serum

cotinine levels. The age of the children was an important predictor in both those exposed to

smoke and those not exposed, although the effects were in different directions. In the smoke-

exposed children, the highest levels were among the youngest, while in the unexposed

children, the older ones had higher mean levels of cotinine (Mannino et al. 2001).

In Australia, trends in ETS restrictions in homes were reported between 1989 and 1997

(Borland et al. 1999). The results indicate a strong move towards smoke-free homes and

towards protecting children from tobacco smoke. Not smoking in the presence of children rose

from 14% in 1989 to 33% in 1996, and smoking outside the home from 20% in 1995 to 28% in

1997. A population-based survey in Ontario, Canada (Ashley et al. 1998) showed that the

percentage of parents who agreed that small children should not be exposed parental tobacco

smoke at home increased from 51% to 70% between 1992 and 1996. In spite of that, only 20%

of homes with children and any daily smokers were smoke-free.

Studies conducted in the 1990s indicated a large variation in the prevalence of children’s

exposure to ETS, ranging from 7% in Finland to 47% in Denmark (Lund el al. 1998a).

Although parental smoking per se is clearly the most important determinant of children’s

exposure, avoidance of smoking in the presence of children seems to be an important

behavioral factor. Only a few studies assessed the determinants of passive smoking, showing

that children’s exposure to ETS was related to lower levels of education, lower socio-economic

Niina Jaakkola Introduction16

status, younger age, and single parenthood (Jarvis et al. 1992, Cook et al. 1994, Lund et al.

1998b).

A substantial reduction in children’s exposure could thus be achieved by both reducing

parental smoking and increasing avoidance of smoking in the presence of children. Both

strategies are needed and should be encouraged for the protection of children’s health.

According to the existing literature, efforts should be focused on people who are young, have

less education, are single, and have lower socio-economic status.

1.4 Fetal effects of ETS exposure

Inhaled carbon monoxide and nicotine increase fetal carboxyhemoglobin and reduce placental

blood flow, resulting in low fetal tissue oxygenation. Structural changes in the placentas of

smoking mothers have been observed as additional evidence of a plausible biologic mechanism

(Rebagliato et al. 1995). Smoking during pregnancy increases metabolites of tobacco in the

fetus that may have toxic effects (Meber et al. 1979). Maternal smoking has also been attributed

to fetal zinc deficiency (Kuhnert et al. 1987). Zinc is considered an essential trace element for

many aspects of growth and development (Apgar 1985).

A recent meta-analysis by Windham and colleagues (1999) suggests that exposure to ETS

during pregnancy has a small effect on birth weight and risk of LBW, although individual

studies were often inconclusive. The effects on the length of gestation and risk of preterm

delivery have been studied less (Martin and Bracken 1986, Ahlborg and Bodin 1991, Mathai et

al. 1990, Fortier et al. 1994), and the evidence for any effect is weak.

As recent discussions suggest, the assessment of exposure during a relevant time period and the

control of confounding factors are the most critical issues of validity in studies on the effects of

ETS on adverse pregnancy outcomes (Jaakkola and Jaakkola 1997, Windham et al. 1999).

Table 5.3 summarizes selected studies on the relation between exposure to tobacco smoke in

pregnancy and fetal growth and prematurity. The first reports based exposure assessment only

on information on the spouse’s smoking (MacMahon et al. 1966, Comstock el al. 1967,

Underwood et al. 1967, Magnus et al. 1984, Rubin et al. 1986, MacArthur and Knox 1987,

Schwartz-Bickenbach et al. 1987, Cambell et al. 1988, Brooke et al. 1989, Chen et al. 1989,

Saito 1991, Mathai et al. 1990, Zhang and Ratcliffe 1993, Martinez et al. 1994). More recent

Introduction Niina Jaakkola17

studies have used information collected from multiple sources, such as other family members

and work exposure, or have measured quantity in terms of daily duration of exposure or

number of cigarettes smoked indoors (Martin et al. 1986, Lazzaroni et al. 1990, Ahlborg and

Bodin 1991, Ogawa et al. 1991, Fortier et al.1994, Sadler et al. 1999).

Few studies have used biomarkers of exposure, such as serum cotinine (Haddow et al. 1988,

Eskenazi et al. 1995, Peacock et al. 1998), and saliva cotinine (Rebagliato et al. 1995)

measured during pregnancy or after delivery. Both of these biomarkers share the common

feature of having a short half-life (Guerin et al. 1992), and therefore they measure exposure

only during a couple of days before sampling. Hair nicotine is a promising new biomarker; the

method is non-invasive and a sample of 0 - 2 cm from the proximal hair gives a good estimate

of exposure during the past two months (Zahlsen and Nilsen 1994, Nafstad et al. 1997). Taking

into account the time frame, a hair sample at birth would effectively describe exposure during

the last trimester, which is considered the most important period for fetal growth. In a recent

case-control study by Nafstad and colleagues (1998), the risk of SGA was related to the hair

nicotine concentration of newborns and their non-smoking mothers after delivery.

1.5 Smoking cessation programs

Two studies have shown that cessation of smoking during pregnancy can have many benefits

for both mothers and infants (Ahlsten et al. 1993, Li et al.1993). The evidence supports the

view that cognitive behavioral programs tailored to pregnancy can produce significant

improvements in the smoking cessation rates of pregnant women compared with usual care

procedures (Walsh and Redman 1993). A number of randomized controlled trials have

established the effectiveness of smoking cessation programs during pregnancy, with observed

quit rates of 14 - 27% (Windsor et al. 1985, Sexton and Hebel 1984), and associated changes in

mean birth weight from –12 g to 90 g (Lumley et al. 1998, Ershoff et al. 1989, Dolan-Mullen et

al. 1994).

There is evidence that well designed and time- and personnel-consuming smoking cessation

programs can reduce smoking in pregnancy. Dolan-Mullen and colleagues (1994) reported a

meta-analysis of 11 randomized, controlled intervention studies indicating an overall effect of

interventions during pregnancy on the proportion of quitters in smoking women. Several

individual studies following the meta-analysis have shown similar effects of interventions on

Niina Jaakkola Introduction18

smoking reduction (Valbø and Nylander 1994, Kendrick et al. 1995, Hartmann et al. 1996,

Walsh et al. 1997, Lowe et al. 1998, Gebauer et al. 1998, Wakefield and Jones 1998).

The effectiveness of smoking cessation programs integrated into routine prenatal care has been

addressed in only a few studies, summarized in Table 5.4 (Hjalmarson et al. 1991, Windsor et

al. 1993, Kendrick et al. 1995, Walsh et al. 1997, Wakefield and Jones 1998, Ershoff et al.

1999). Hjalmarson and colleagues (1991) carried out a population-based smoking cessation

program in 13 out of 14 public health maternity clinics in Gothenburg, Sweden. Over a one-

year period, all pregnant women attending a maternity clinic for their first visit were asked

about their smoking habits by a midwife, who advised smokers to stop smoking. The smoking

women were randomized into an intervention and a reference group. The intervention included

a self-help manual on stopping smoking, which was revised from the manual developed by

Windsor et al. (1985) and distributed by an obstetrician. The program in the manual included a

one-week preparatory phase with daily assignments based on the principles of behavioral

therapy, and a cessation period with new information daily for the first 5 days. This

information included coping with the most common abstinence problems, advice for special

situations, and relapse- prevention counseling. The acceptance of the program in the

intervention group was 85%. Overall, 10.4% of the intervention were not smoking up to 8

weeks after delivery compared with 5.2% of the reference group (difference 5.2%, 95%

confidence interval (CI) 1.1% - 9.3%).

Kendrick and colleagues (1995) evaluated statewide low-intensity programs in Colorado,

Maryland and Missouri in a group-randomized controlled trial. All three interventions provided

information on the effects of smoking on the fetus, the benefits of quitting, quitting techniques,

developing social support, preventing relapse, and limiting exposure to ETS, but the setting

and details of the approaches varied. The results of all three programs showed a reduction in

the reported smoking rate (overall reduction 3.5%; 95% CI 1.3% - 5.7%), but no difference in

the level of urine cotinine levels. More recent articles on the same data demonstrated a

nonlinear relation between two measures of tobacco exposure; self-reported cigarette smoking

and urine cotinine concentration, and birth weight (England et al. 2001a). Reducing cigarette

use was associated with an increase in the mean adjusted birth weight of only 32 g (p=0.33).

As third-trimester cigarette use increased, birth weight declined sharply, but leveled off at more

than eight cigarettes per day. The findings were similar when urine cotinine concentration was

used (England et al. 2001b).

Introduction Niina Jaakkola19

Walsh and colleagues (1997) evaluated the effects of a smoking cessation program with more

extensive intervention carried out at the antenatal clinic of an urban teaching hospital in

Australia. The intervention included advise from a doctor, a 14-minute video program, midwife

counseling, a self-help manual, a lottery, social support tips for an accompanying adult, and a

chart reminder in the medical record during the first visit, a reminder letter, and midwife

counseling during the second visit. In the intervention group, 13% of the women abstained

from smoking at midpoint, end of pregnancy and postpartum compared with 6% in the

reference group (difference 7.0%, p=0.035).

Windsor and colleagues (1993) conducted a randomized trial to assist smoking cessation in

women attending public maternity clinics in Alabama, 1986 - 91. The intervention consisted of

15-minute counseling by an educator, a self-help manual, clinical reinforcement, a mailed

letter, social-support methods including a letter from a supporting person (“buddy”), "a buddy"

contract, and "a buddy" tip sheet, and a quarterly newsletter. The intervention strategy was

effective; 14.3% quit smoking in the intervention group and 8.5% in the control group

(difference 5.8%, 95% CI 1.4–10.1). Ershoff and colleagues (1999) recently reported the

results of a randomized controlled trial in California in which neither a computerized telephone

cessation program nor systematic provision of motivational counseling improved cessation

rates compared with a tailored self-help booklet delivered with brief advice from prenatal

providers in a managed care organization.

Interventions integrated into routine prenatal care have had low biochemically verified quit rates.

A meta-analysis from the late 1980s of 39 smoking cessation randomized trials in general

practice settings suggests that intervention success increases with the number of intervention

modalities employed, the number of health professionals involved, and the number of follow-

up assessments conducted (Kottke et al. 1988). An integrated effort beginning during

pregnancy and continuing throughout the critical 6-month postpartum period could potentially

decrease relapse rates. Although effectiveness has been clearly documented for cognitive

behavioral programs (Walsh and Redman 1993), it is difficult to make recommendations about

what specific components should be included in an intervention because of the diversity of the

methods tested (Lumley et al. 2000). Group programs are very poorly attended and are not

recommended (Walsh and Redman 1993). Self-help materials appear to be useful, especially

Niina Jaakkola Introduction20

when used in the populations for whom they were developed (Walsh and Redman 1993, Lowe

et al. 1998).

1.6 Rationale

The first phase of the present research project was initiated as a response to findings of The

Finnish Children’s Environment and Health Project, from which it became apparent that

prenatal and postnatal exposure to ETS is a significant public health problem in Finland. The

second phase of the project was developed to provide information to promote the rational

prevention of this public health problem. The project was designed to provide information on

four types of phenomena, that were considered potentially helpful in developing preventive

measures: i) the extent of the problem, i.e. exposure distributions in the population of interest,

ii) determinants of exposure, iii) information on the health effects of exposure, and iv)

effectivity of currently used preventive measures.

Objectives Niina Jaakkola21

2. OBJECTIVES

The general objective was to study issues related to prenatal and postnatal exposure to tobacco

smoke products.

The specific objectives were:

1. To study the occurrence, trends, and determinants of smoking during pregnancy in Finland

(Study I).

2. To study the occurrence and determinants of exposure to environmental tobacco smoke

during early childhood in Finland (Study II).

3. To study the effects of exposure to environmental tobacco smoke during pregnancy on fetal

growth and length of pregnancy (Study III).

4. To evaluate the effects of a population-based smoking cessation program on smoking

cessation and fetal exposure to tobacco smoke products (Study IV).

Niina Jaakkola Methods22

3. METHODS

3.1 Study design

This dissertation is based on information collected in two consecutive research projects: The

Children’s Environment and Health Project and The Finnish Prenatal Environment and Health

Project. A general feature is that these studies are population-based representing

geographically defined source populations. There are three population levels, nationwide,

regional, and local. The nationwide study (Study I) assessed smoking habits in pregnancy in

1987 - 97, representing fetal exposure to tobacco smoke products. On the regional level, the

source population comprised all the pregnant women and their children born between May 1,

1996 to April 30, 1997 in Kymi and Uusimaa provinces. First, a cross-sectional study was

conducted in a stratified random sample of non-smoking pregnant women and their newborns

(Study III). Further, the regional-level approach comprised a controlled evaluation of the

effects of smoking cessation intervention focusing on all the pregnant women who were

smoking at the beginning of their pregnancy (Study IV). On the local level, a population-based

cross-sectional study was conducted in a random sample of day-care children living in the City

of Espoo in 1991 (Study II).

Methods Niina Jaakkola23

* 1st visit to maternity health clinic

Figure 3.1. Design of the project and a map of Finland showing the study areas.

Espoo

Helsinki

Porvoo

Kotka

Kuusankoski

Study IV

N=694,926women

1-7-year-oldchildren

All births1996 - 1997

Study I

Study II N=1,003children

Intervention group N=306

Reference group N=152

N=389women

Study III

N=2,591

Smokingwomen*N=458

Non-smoking women N=1,621

N=2,568

Non-smokingparentsN=1,565

SmokingparentsN=1,003

All births 1987- 1997

Local: Regional: National:Espoo Kymi-Porvoo Birth Registry

Niina Jaakkola Methods24

3.2 Study populations

The nationwide study (Study I) was based on information from The Finnish Medical Birth

Registry, which has been collecting information on smoking habits during pregnancy since

January 1, 1987. A standard form to be used nationwide was introduced in 1987. The primary

study population for the present analyses includes a total of 694,926 women who gave birth

between 1987 and 1997 in Finland. The Registry information was validated with information

obtained from the Finnish Prenatal Environment and Health Study, which is described in Study

III. We also retrieved information on smoking during pregnancy by the mothers concerned

from the medical records filled in by nurses during regular visits to the maternity health care

centers. During the 1996 - 97 study period there was a total of 2,751 births, and 2,591 mothers

completed the questionnaire (response rate 94%). Information from medical records and the

Birth Registry was available for all 2,591 mothers who completed the questionnaire.

In Study II, the source population, 14,749 children, included all children in the city of Espoo

born between January 1, 1984 and December 31, 1989. Espoo is an urban-suburban

municipality with a population of 170,000 in 1990 and 213,000 in 2001, located across the

western border of Helsinki. We distributed a parent-administered questionnaire in March 1991

to a random sample of children living in Espoo drawn from the roster of the Population

Register Centre of Finland. The study population included a total of 2,568 children whose

parents filled in the questionnaire (response rate 80.3%). The parents were asked about their

smoking habits and their child's exposure to ETS. Study II focused on the 1,003 (39.1%)

children whose one or both parents or guardians were current regular smokers.

Studies III and IV were based on The Finnish Prenatal Environment and Health Study. This

study focused on a source population that included all the 2,751 children born from May 1,

1996 to April 30, 1997 in two geographically defined hospital districts in South East Finland

(Kymi and Porvoo Hospital Districts). All of the mothers were asked, after delivery to fill in a

self-administered questionnaire, ‘Environment and pregnancy’ (response rate 94.2%). The

respondents, 2,591 mothers, had 2,568 singletons and 23 twin pairs. A total of 1,621 women

(62.3%) were non-smokers. They were used as a framework for selecting 472 non-smoking

women for the present study: 189 women who reported exposure to environmental tobacco

Methods Niina Jaakkola25

smoke either at home or outside the home, and 283 women who reported no exposure to ETS.

The latter group was selected randomly from the unexposed, with a 1 to 1 balance according to

the spouse's current smoking habits, so that in 141 women the spouse was a current smoker,

and in 142 women the spouse had never been a smoker, or had quit more than 12 months

previously. Study III focused on all the 389 singleton newborns of women, whose hair sample

was sufficient for the measurement of nicotine concentration. The characteristics of the study

population are presented in Table 1 of Study III.

Study IV was based on a total of 2,751 births, 1,917 in the intervention area and 834 in the

reference area. We requested the records of these women from the maternity health clinics in

the two areas. We received a total of 2,507 records (96.8%), 1,775 (97.5%) from the

intervention area and 732 (94.9%) from the reference area. The intervention group consisted of

all women who were smokers at the time of their first visit to the maternity health clinics and

who gave birth in the two birth clinics in The Kymi Hospital District from May 1, 1996 to

April 30, 1997 (306 of 1,917). Identical inclusion criteria were used in choosing the reference

group from The Porvoo Hospital District (152 of 834).

3.3 Relations of interest

Study I

In Study I, we assessed the prevalence of smoking in pregnancy over time, and compared the

prevalence by the following potential determinants: time period, geographic area, age group

and marital status of the mother, and maternal occupation.

Study II

In Study II, we assessed the role of the parents' education and socioeconomic status, single

parenthood (or guardianship), and knowledge of the child's health as determinants of the child’s

exposure to ETS. The child's reported history of allergic diseases and current asthma were

considered to indicate parental knowledge of the harmful effects of ETS and were chosen as

determinants of ETS exposure. Exposure to ETS was defined from the answers to the following

structured questions: Smoking regularly at home: 1) "Have you or anyone else smoked

regularly indoors at home during the past 12 months?" (No; Yes, number of cigarettes a day;

Niina Jaakkola Methods26

Yes, number of cigars a day; Yes, number of pipes a day). There was also another question to

verify the exposure: 2) "Has anyone ever smoked regularly indoors during the child's lifetime?"

(If Yes, number of cigarettes smoked a day per 1st to 7th year of the child's life by the mother,

father or someone else). The children were placed in an exposed or reference category

according to their current exposure to ETS.

Study III

In Study III, we assessed the relations between prenatal exposure to ETS and the risk of

different pregnancy outcomes. Exposure assessment was based on the nicotine concentration in

the mother’s hair taken after delivery. The proximal 0-2 cm segment of maternal hair was

analyzed for nicotine chromatography/mass spectrometry (Zahlsen and Nilsen 1994). The

detection limit was 0.05 µg nicotine/g hair when the sample was at least 15 mg. Nicotine

concentration in hair samples <15 mg was not reported due to uncertainty in the estimated

concentrations. We also used maternity clinic medical records, the questionnaire at the birth

clinic, and birth registry data as additional sources of information.

The primary health outcomes were fetal growth and preterm delivery based on information

retrieved from maternity health clinic records. We used three different measures of fetal

growth: birth weight in grams (g), LBW (<3,000 g), and SGA. A higher than traditional cut

point (2,500 g) for LBW, 3,000 g, was chosen to increase the number of cases from 7 to 46.

SGA was defined as birth weight in the lowest 10th percentile according to the week of

gestation, which was calculated from the source population of non-smoking women. Preterm

delivery was defined as a length of gestation of less than 37 weeks.

Study IV

In Study IV, we evaluated the effect of a health education program on quitting smoking during

pregnancy. In 1993, the Kymi Hospital District, responsible for a population of 188,000,

founded a Health Promotion Working Group with representatives from all of the primary

health care units and all four hospitals in the district. The program, initiated in September

1995, included: 1) a one-day training seminar for all the maternity clinic’s nurses; 2) health

education material for use in daily work and distributed in scheduled meetings of pregnant

women; 3) health education material for visits to the homes of families with newborns; 4)

health education material such as a set of overheads and printed material for scheduled

Methods Niina Jaakkola27

maternity training offered to both pregnant women and their husbands; and 5) posters, mobiles

and printed material on smoking cessation for the maternity clinics during the time of the

study.

The primary outcome was quitting smoking during pregnancy after the first visit to the

maternity health clinic and before the delivery. This outcome was defined according to the

participant response to the questionnaire at the birth clinic. Hair nicotine concentration was

used to identify probable smokers among the women who reported that they had quit smoking,

and to assess the role of information bias in the study.

The secondary outcome was a change in smoking rate during pregnancy. The change was

calculated by subtracting the smoking rate at the beginning of pregnancy from smoking rate

after delivery. The other outcomes included the following self-assessed measures: 1) the

amount of information received from the maternity health clinics; 2) the amount of reading or

studying the material; 3) the influence of the information/support received from the maternity

health clinics; 4) the avoidance of ETS; and 5) the number of quitting attempts during

pregnancy.

3.4 Statistical methods

Study I

We calculated the annual prevalence of smoking during pregnancy and its 95% CI based on

binomial distribution. We compared the prevalence by the following potential determinants:

time period, geographic area, age group and marital status of the mother, and maternal

occupation. We evaluated the validity of the Birth Registry Data on smoking by comparing it

with the corresponding information from the questionnaire and from the medical records in the

study population of the Prenatal Environment and Health Study in Southeast Finland. The

goodness of agreement was quantified using the Kappa coefficient (Fleiss 1981).

Study II

In the crude analysis, prevalence ratios and odds ratios (OR) of the relations of exposure to ETS

and its determinants were estimated. The relations between ETS exposure and its determinants

Niina Jaakkola Methods28

were then estimated in logistic regression models, including the determinants of interest

(education, socioeconomic status, single parenthood, atopy) and the potential confounders as

covariates (age, gender, type of building).

Study III

We estimated the prevalences (%) of the reproductive outcomes with 95% CIs based on

binomial distribution, and the mean birth weight with CIs based on normal distribution. First,

we compared the mean birth weight and the risks of LBW, SGA and preterm birth in the

different categories of exposure defined by the hair nicotine concentration and questionnaire

information. The relations between birth weight and the determinants of interest were

estimated in linear regression analysis, controlling for potential confounding factors. For the

three dichotomous outcomes, OR was the measure of effect. We used logistic regression

analysis to estimate the adjusted ORs. We adjusted for the potential determinants of the

outcomes, including sex, birth order, maternal age, body mass index (BMI) before pregnancy,

marital status, parental education, index of socioeconomic status, alcohol consumption during

pregnancy, and employment during pregnancy.

Study IV

We assessed the effects of the intervention by comparing the difference in the outcomes

between the intervention and reference groups. Chi-square statistics and unpaired t-test were

used to evaluate the statistical significance. The Cochran-Mantel-Haenszel method, based on

standardized midrank scores (Modified Ridit Scores), was applied to the ordinal scale

categorical variables (Stokes et al. 1995).

Methods Niina Jaakkola29

Table 3.1. Summary of the statistical methods used in Studies I-IV.

Study StudyPopulation

Determinants Outcomes Analytical approach:covariates

Study I Nationwide:694,926pregnanciesRegional:2,751 deliveries

Time period,geographic area,age, marital status,occupation

Prevalence ofsmoking

Prevalence ratiosKappa statistics

Study II Local:1,003 children

Education,socioeconomicstatus, singleparenthood, atopy

Exposure to ETS Logistic regression:age, gender, type ofbuilding

Study III Regional:389 newborns ofnon-smokingwomen

Prenatal exposureto ETS

Fetal growth,preterm delivery

Linear regressionLogistic regression: sex,birth order, maternal age,BMI before pregnancy,marital status, parentaleducation, alcoholconsumption andemployment duringpregnancy

Study IV Regional:458 women,smoking at 1st visitto maternity healthclinic

Intervention Quittingsmoking,change insmoking rate

Chi-square statistics,unpaired t-test, andCochran-Mantel-Haenszel method

Niina Jaakkola Results30

4. RESULTS

4.1 Trends and determinants of smoking during pregnancy in Finland

The overall prevalence of smoking in pregnancy was approximately 15% both in 1987 and in

1997 (Figure 4.1.A). There was an increase in this prevalence in the beginning of the 1990s,

and then a decrease around 1994 back to the level of 1987. The prevalence varied considerably

according to geographic area, maternal age, and marital status and the mother’s occupation

(Figure 4.1). It was the lowest in Western Finland from 1987 through 1997, ranging between

10 and 13%, and increased in Eastern Finland (Kymi) during the same period from 16% to

19%. Thus the prevalence of smoking in pregnancy in 1997 was almost as high in Eastern

Finland as in Northern Finland.

Young women smoked most often during pregnancy (Figure 4.1.C). The prevalence of

smoking in 1997 was 37% in the age groups of 19 years or younger, and 23% in the group of

20 to 24 years. The trend in smoking over time was similar in all age groups.

Marital status was a strong determinant of smoking during pregnancy (Figure 4.1.B). The

single mothers smoked three times more often than the married women. Those who were not

married but were living with the father also smoked more often than the married women. The

differences between the groups remained similar over the entire study period.

0%5%10%

15%

20%

25%

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

Year

Prevalence of smoking

Tot

alL

apla

ndV

aasa

Uus

imaa

Kym

i

0 %5 %10 %

15 %

20 %

25 %

30 %

35 %

40 %

45 %

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

Year

Prevalence of smoking

-19

20-2

425

-29

30-3

435

-39

40-4

445

-

0 %

5 %

10 %

15 %

20 %

25 %

30 %

35 %

40 %

45 %

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997 Year

Prevalence of smoking

Mar

ried

Coh

abite

dSi

ngle

0 %

5 %

10 %

15 %

20 %

25 %

30 %

35 %

1991

1992

1993

1994

1995

1996

1997 Ye

ar

Prevalence of smoking

Ent

repr

eneu

r U

pper

leve

l em

ploy

ee L

ower

leve

l em

ploy

ee W

orke

r S

tude

nts

Hou

sew

ife M

issi

ng

Figu

re 4

.1. T

rend

s and

det

erm

inan

ts o

f sm

okin

g du

ring

preg

nanc

y in

Fin

land

.

A

Are

aC

A

ge

B

Mar

ital

stat

usD

O

ccup

atio

n

Niina Jaakkola Results32

The prevalence of smoking during pregnancy was highest among workers (29% in 1991 and

25% in 1997), and among the mothers whose occupation was unknown (25% and 21%,

respectively). It was high and constant over time among the housewives, being 21% in both

1991 and 1997. The women with a higher occupational status smoked less often during

pregnancy (6% in 1991 and 4% in 1997), and the trend of the prevalence showed a clear

decrease (Figure 4.1.D).

Comparison of the information from the Finnish Medical Birth Registry on smoking during

pregnancy in Southeast Finland in 1996 - 1997 with that from the questionnaire on daily

smoking obtained in the Prenatal Environment and Health Study showed an excellent

agreement with Kappa statistics of 0.84 (95% CI 0.81 - 0.87) (Table 4.1). The agreement

between the Finnish Medical Birth Registry information on smoking during pregnancy and the

information abstracted from medical records in 1996 - 1997 was also excellent, the Kappa

statistics being 0.89 (95% CI 0.86 - 0.91) (Table 4.2).

Table 4.1. Agreement between the information from the Finnish Medical Birth Registry onsmoking during pregnancy and the questionnaire information in 1996 - 1997 in the PrenatalEnvironment and Health Study in Southeast Finland.

Birth Registry InformationQuestionnaireinformation No smoking

N (%)SmokingN (%)

TotalN (%)

No smoking 1,868 (81.2) 40 (1.7) 1,908 (82.9)Smoking daily 64 (2.8) 330 (14.3) 394 (17.1)

Total 1,932 (84.0) 370 (16.0) 2,302 (100.0)

Kappa: 0.84 (95% CI: 0.81 – 0.87)

Results Niina Jaakkola33

Table 4.2. Agreement between the Finnish Medical Birth Registry Information on smoking duringpregnancy and information from Medical Records in 1996 - 1997 in Southeast Finland.

Birth Registry InformationMedical records* No smoking

N (%)SmokingN (%)

TotalN (%)

No smoking 1,914 (80.5) 28 (1.2) 1,942 (81.7)Smoking daily 52 (2.2) 382 (16.1) 434 (18.3)

Total 1,966 (82.7) 410 (17.3) 2,376 (100.0)

Kappa: 0.89 (95% CI: 0.86 - 0.91)* Information on smoking was missing for 192 mothers

4.2 Occurrence and determinants of ETS in children

We estimated the occurrence of smokers in Espoo families with preschool children. There was at

least one smoking parent in 39% (95% CI 37.2 - 40.9) of the households. However, the majority

of parents avoided smoking in the presence of their children. According to parental reports, the

prevalence of current exposure to ETS was 9.9% (95% CI 8.8 - 11.0). The risk of exposure was

statistically significantly higher when the parents had no professional education, compared to

university or college education (adjusted OR 2.42, 95% CI 1.43 - 4.11), but socioeconomic status

was not associated with exposure (adjusted OR 1.00, 95% CI 0.63 - 1.56) when adjusting for

education. The risk was higher with a single parent or guardian than in families with both parents

(adjusted OR 2.17, 95% CI 1.36 - 3.44). The risk was statistically significantly lower for atopic

children (adjusted OR 0.61, 95% CI 0.38 - 0.98), indicating that knowledge of the child's illness

affects the smoking behavior of the parents. The estimated adjusted odds ratios and their 95%

confidence intervals are presented in Table 4.3.

Niina Jaakkola Results34

Table 4.3. Determinants of children's exposure to ETS. The estimated adjusted odds ratios forexposure and their 95% confidence intervals are from a logistic regression model, with all thevariables given in the table._____________________________________________________________________________ Determinants Crude OR Adjusted OR 95% CI_____________________________________________________________________________Education

College or university 1.00 1.00Vocational school 1.24 1.23 0.81-1.85Non-professional 2.55 2.42 1.43-4.11

Socioeconomic status*

Upper-level employees 1.00 1.00Lower-level employees 1.08 0.87 0.58-1.30Self-employed persons 1.04 0.90 0.43-1.89Manual workers 1.36 1.00 0.63-1.56Others 0.96 0.22 0.03-1.93

Single parent or guardianNo 1.00 1.00Yes 2.13 2.17 1.36-3.44

Atopy1

No 1.00 1.00Yes 0.62 0.61 0.38-0.98

Age1-2 1.00 1.003-4 1.55 1.55 1.06-2.265-7 1.55 1.56 1.08-2.26

GenderGirl 1.00 1.00Boy 1.20 1.25 0.93-1.68

Type of buildingHouse, semi-detached 1.00 1.00House, row houseBlock of flats 1.10 1.05 0.77-1.43

___________________________________________________________________________1 Doctor-diagnosed asthma, atopic eczema, allergic rhinitis or allergic conjunctivitis*The Central Statistical Office of Finland's classification of status in employment follows the UnitedNations recommendations for 1990 population censuses.

Results Niina Jaakkola35

4.3 Fetal effects of ETS exposure in pregnancy

We assessed the effects of prenatal exposure to ETS on fetal growth and length of gestation in

Study III. The primary health outcomes were fetal growth and preterm delivery. Three different

measures of fetal growth were used: birth weight in grams (g), LBW (<3,000 g), and SGA. A

higher than traditional cut point for low birth weight was chosen to increase the number of

cases from 7 to 46. The determinant of interest was prenatal exposure to ETS. The exposure

assessment was based on the nicotine concentration in the mother’s hair after delivery, which

measures exposure during the past two months (the third trimester). The exposure categories

were defined a priori as high (nicotine concentration >4.00 µg/g; N=52), medium (0.75-4.00

µg/g; N=186), and low as the reference category (<0.75 µg/g; N=151).

In logistic regression analysis, controlling for confounding the risk of preterm birth (<37

weeks) was higher in the high (adjusted OR 6.12, 95% CI 1.31 - 28.7), and medium exposure

categories (adjusted OR 1.30, 95% CI 0.30 - 5.58) compared with the reference category, and

there was 1.22 (95% CI 1.07 - 1.39) increase in adjusted OR by a unit increase in exposure.

The risks of LBW (OR 1.06 per unit increase, 95% CI 0.96 - 1.17 per µg/g nicotine) and SGA

(1.04 per unit increase, 95% CI 0.92-1.19) were only weakly related to exposure.

The greater the mother’s hair nicotine concentration was, the lower the child’s mean birth

weight was. ETS exposure at work was more strongly related to preterm birth than exposure at

home. The risk of preterm delivery was increased in the presence of both work and home

exposure (adjusted OR 8.89, 95% CI 1.05 – 75.3). The point estimate for the presence of work

exposure increased consistently, in accordance with the hypothesis, but the confidence interval

was wide and included unity (adjusted OR 2.35, 95% CI 0.50 - 11.1). The risk of preterm

delivery was not related to home

exposure only.

Results 37 Niina Jaakkola

4.4 Effects of a population-based smoking cessation program on maternal smoking

during pregnancy

The intervention and reference groups were similar with respect to age distribution, marital

status, parity, educational level and job category (Table 4.5). The average tobacco

consumption, starting age and duration of smoking (years) until delivery were also similar as

well as the smoking status of partners and friends (Table 4.6).

Table 4.5. Characteristics of the intervention group (N=306) and the reference group (N=152)__________________________________________________________________________Characteristic Intervention group Reference group

N % N %

Age (y) < 19 18 5.9 6 4.020-24 75 24.5 37 24.325-29 74 24.2 44 28.930-34 90 29.4 48 31.6> 35 49 16.0 17 11.2

Marital status Married or cohabited 263 86.2 137 90.1 Single 42 13.8 15 9.9

Parity No previous births 46 21.0 24 23.5 1-2 previous births 150 68.5 69 67.7 > 3 previous births 23 10.5 9 8.8

Educational level No schooling/incomplete 81 29.7 45 31.9 Primary studies 115 42.1 43 30.5 Secondary studies 72 26.4 49 34.8 University studies 5 1.8 4 2.8

Job category Not working 153 51.0 69 45.8 Manual job 29 9.7 13 8.6 Non-manual job 118 39.3 69 45.6__________________________________________________________________________

Niina Jaakkola Results38

Table 4.6. Smoking in the intervention group (N=306) and in the reference group (N=152)Characteristic Intervention group Reference group

N % N %

Tobacco consumptionbefore pregnancy (cigarettes per day):

1 – 9 80 26.1 32 21.110-19 148 48.4 85 55.9

> 20 78 25.5 35 23.0

Starting-age of smoking (y) < 16 223 72.9 118 77.616-18 52 17.0 25 16.5

> 19 31 10.1 9 5.9

Smoking status of partner Smoker 219 73.0 109 72.2 Non-smoker 81 27.0 42 27.8

Smoking status of friends All/Most smokers 156 56.7 79 54.5 Some smokers 94 34.2 51 35.2 Non-smokers 25 9.1 15 10.3

Hair nicotine concentration was defined for 56 (69.1%) of 81 self-reported quitters (Table 4.7),

while the hair sample was insufficient for the rest of the women. There was one quitter (1.8%)

in the reference group whose hair nicotine concentration of 23.8 µg/g clearly indicated

personal smoking.

In the intervention group, 58 (19.0%) of the women reported quitting smoking during

pregnancy, whereas in the reference group the number of quitters was 22 (14.5%) after the

exclusion of 1 probable smoker on the basis of hair nicotine analysis (Table 4.7). The

difference in the quitting rate was 4.5% (95% CI -2.6 – 11.6%).

Results 39 Niina Jaakkola

Table 4.7. Smoking cessation, hair nicotine concentration, and change in smoking rate in the

intervention (N=306) and reference groups (N=152).

Intervention Group Reference Group Effect of Intervention

N % N % ∆% (95% CI)

Quit smoking

Self-reported 58 19.0 23 15.1 3.9 (-3.3 – 11.1) Probable smoker excluded1 58 19.0 22 14.5 4.5 (-2.6 – 11.6)

Verified quitters included2 41 13.4 14 9.2 4.2 (-1.8 – 10.8)

Hair nicotine concentration (µg/g)< 2 18 43.9 10 66.62-<4 14 34.1 1 6.74-<6 3 7.3 3 20.06-<8 4 9.8 -8-<10 2 4.9 -10-<15 - -15-<20 - -20- - 1 6.7

median 2.86 0.7725th and 75th percentile 1.38, 3.56 0.35, 23.77

max 9.68 23.77missing 13 7

Reduction in smoking rate in cigarettes per day (mean, SD)

1.41 (4.60) 1.28 (4.50)

Smoking rate in cigarettes per day (mean, SD)

I trimester 9.01 (5.83) 9.61 (5.48)

II trimester 8.16 (5.40) 8.57 (5.16)

III trimester 7.93 (5.80) 8.40 (5.70)________________________________________________________________________________________________________________________________________

1 suspected smokers (hair nicotine > 10 µg/g) excluded: 0 women in the intervention group and 1 women in the

reference group2 suspected smokers and subjects with missing information on hair nicotine concentration excluded: 17 women in the

intervention group and 9 women in the reference group

The intervention group, compared with the reference group, indicated that they had received

more information on the adverse effects of active and passive smoking, had studied the

material more actively, and had felt that the material from the maternity care had influenced

their smoking behavior more, but there was no difference in the experience of receiving

support for attempts to quit smoking between the two groups (Table 4.8).

Niina Jaakkola Results40

Table 4.8. Differences in attitude, social influence, and received support between the intervention(N= 306) and reference groups (N=152) measured by the questionnaire.

Intervention Group Reference Group CMH

N % N % P-value

Did you receive information on smoking in pregnancy from maternity care?

1. None 7 2.3 6 4.0 0.074

2. A little 57 18.8 37 24.3

3. Moderately 114 37.5 54 35.5

4. Much 97 31.9 45 29.6

5. Very much 29 9.5 10 6.6

Did you receive information on ETS in pregnancy from maternity care?

1. None 46 15.2 38 25.0 0.045

2. A little 69 22.8 39 25.7

3. Moderately 118 38.9 45 29.6

4. Much 54 17.8 26 17.1

5. Very much 16 5.3 4 2.6

Did you read/study the material you received from maternity care?

1. No 24 8.0 23 15.2 0.013

2. Yes, part of it 77 25.5 43 28.5

3. Yes, at least once 173 57.3 74 49.0

4. Yes, several times 28 9.2 11 7.3

Did you avoid environmental tobacco smoke during pregnancy?

1. No 57 19.1 41 27.5 0.013

2. A little 93 31.2 47 31.6

3. Moderately 91 30.5 46 30.9

4. Much 41 13.8 9 6.0

5. Very much 16 5.4 6 4.0

Results 41 Niina Jaakkola

Table 4.8. Continues.Intervention Group Reference Group CMH

N % N % P-value

Did the material (written and oral) you received from maternity care have any effect on your smoking behavior?

1. Not at all 85 30.5 61 42.1 0.011

2. Little 113 40.5 59 40.7

3. Moderately 19 6.8 8 5.5

4. A lot 55 19.7 14 9.7

5. Very much 7 2.5 3 2.0

Quit attempts:

Before pregnancy

0 156 (55.9) 61 (44.9) 0.080

1-2 69 (24.7) 44 (32.3)

>3 54 (19.4) 31 (22.8)

During pregnancy

0 183 (62.2) 89 (61.4) 0.705

1-2 70 (23.8) 41 (28.3)

>3 41 (14.0) 15 (10.3)

Niina Jaakkola Discussion42

5. DISCUSSION

This dissertation was conducted in order to address important public health issues related to

prenatal and postnatal exposure to tobacco smoke products. The population-based approach on

the national, regional and local levels provided information, which can be transformed into

assessment of public health impact to support preventive measures. The dissertation described

the occurrence of both prenatal and postnatal exposure to tobacco smoke products, which is

necessary information in the assessment of public health impact. Two studies evaluated the

potential determinants of exposure, which is necessary information in the search for special

groups at which to target prevention. The project also provided information on the adverse

effects of exposure to ETS during pregnancy, complementing the accumulating information on

various harmful effects of tobacco smoke products. Finally, part of the project comprised a

controlled evaluation of a smoking cessation program aimed at reducing smoking in pregnancy

on the regional level.

The discussion begins by addressing the issues that crucially affect the validity of the results in

terms of three classical concerns: selection bias, information bias and confounding. Then, the

relation of the results to previous knowledge will be assessed.

5.1 Validity of the results

5.1.1 Selection bias

In general, the population-based approach, and from high to very high participation rates,

minimized the potential influence of selection bias. When assessing occurrence of phenomena

in a population, such as the prevalence of active or passive smoking in Studies I and II, any

relation between participation and the phenomena in question would bias the estimates. In the

assessment of relations, either determinants of exposure as in Study II, or the determinants of

health outcomes as in Study III, selection bias would be introduced if the participation was

related to both the determinants in question and the outcomes.

In Study I, The Finnish Medical Birth Registry provided central information. The registry

receives information on practically all births taking place in Finland, and in 1991 - 97 the

Discussion 43 Niina Jaakkola

amount of missing information on maternal smoking was 4.2% (Gissler et al. 1995). Thus, high

representativeness and completeness of data should minimize selection bias.

In Study II, the study population was a random sample of children in a designated geographical

area. The response rate to the questionnaire was reasonably high (80%). Smoking habits could

have influenced the response, but this study focused on families with at least one parent smoker.

Smoking behavior related to exposing or not exposing children to ETS was not likely to influence

participation. The study was part of the Children's Indoor Environment and Health Project, and

there was no special emphasis on the questions concerning smoking habits. This decreases the

possibility of systematic error due to awareness of the passive smoking issue in the reporting of

exposure. Thus the estimates of ETS form a reasonably valid estimate of exposure in the source

population.

In Studies III and IV, the response rates to the questionnaire at birth clinics were very high

(94%) and access to the maternity health care records (97%) was good in both the intervention

and the reference areas. Thus, selection bias is not a likely explanation for our findings.

In Study IV, the comparability of the intervention and reference groups was a critical criterion.

The selection of the reference area seemed successful, because the characteristics (Table 4.5)

and smoking behavior before the first visit to maternity care were similar in both groups. The

hospital districts chosen for the reference have a similar, nationwide health care organization as

the district that underwent intervention.

5.1.2 Information bias

In estimating the occurrence of active or passive smoking, the quality of the measurement

instruments is directly transformed to the accuracy of the estimates. In the assessment of these

relations, random measurement errors are likely to reduce the empirical relation towards the

null, whereas systematic errors may either enhance or mask the relations of interest.

In Study I, different sources of information on smoking in pregnancy provided a possibility to

evaluate the quality of the information. The agreement on smoking information between the

Niina Jaakkola Discussion44

medical records and the Birth Registry data was found to be excellent (Table 4.2). The

agreement was also excellent between the questionnaire information on daily smoking in 1996

- 97 and the Birth Registry data, and good when the Registry information was compared with

daily or occasional smoking reported in the questionnaire (Table 4.1). These findings give us

assurance that the information on smoking during pregnancy is reliable.

In Study II, children’s exposure to ETS in the home was measured by a parent-administered

questionnaire. The Study II was part of the Children's Indoor Environment and Health Project, and

there was no special emphasis on the questions concerning smoking habits. This decreases the

possibility of systematic error due to awareness of the passive smoking issue in the reporting of

exposure. The reporting of exposure to ETS rather than actual exposure could be associated with

the determinants of interest. Coultas and colleagues found that adults reported reliably (over 90%)

on smoking by household members, whereas quantitative assessments were less reliable (Coultas

et al. 1989). Cummings and coworkers validated the questionnaire reports of ETS exposure with

reports of surrogates, including parents, siblings, spouse and coworkers, showing a good

agreement between the subject and surrogate reports (Cummings et al. 1989). These studies

indicate that qualitative questionnaire information on ETS exposure (presence or absence) is

reliable, although some misclassification will occur.

Questionnaire or interview information on sources of exposure is sensitive to information bias,

in particular if carried out after delivery, and if the accuracy of the smoking information is

compromised due to inaccurate recall and variation of environmental conditions such as air

change, area and volume of space.

In Study III, the use of hair nicotine concentration provided an objective exposure assessment

with comparable information for newborns with and without normal pregnancy. This approach

minimized the problems of random and systematic error in the exposure assessment, which

could be introduced in a retrospective questionnaire administration. However, the nicotine

levels corresponded well to the reported exposure information, given that several uncontrolled

intervening factors such as air change and volume in the spaces could have influenced the

relation between source strength and exposure. The hair nicotine concentration was not

available to some of the mothers due to too small hair samples, but there is no reason to believe

that the availability was related to either exposure or the outcomes in question.

Discussion 45 Niina Jaakkola

In Study IV, hair nicotine was used in the assessment of the outcome of the information,

quitting smoking. The use of an objective biomarker of smoking in addition to questionnaire or

interview information only has been considered important for any critical assessment of

smoking cessation interventions (Jarvis 1999). Information on intervention on the group level

was quite accurate, because the factual procedures could be verified. On the individual level,

information on exposure to intervention was recorded in the questionnaires. However, the

overall evaluation of the influence of intervention was made on the group level, without taking

into account the individual variation in visits to maternity care clinics.

5.1.3 Confounding

Potential confounding was an issue in Studies II-IV, where substantial differences in the

determinants of the outcome of interests would have introduced confounding. In Studies II and

III, a large number of potential personal and environmental confounders were controlled for in

the analyses with which the determinant-outcome relations were estimated. In Study II, we

assessed the determinants of children’s exposure to ETS, including education, socio-economic

status, single parenthood and the child’s atopy. Potential confounders constituted other possible

determinants of exposure, such as age, gender and type of building. When studying the effect

of any of the studied determinants, the potential confounders included other studied

determinants and the three confounders listed above.

In Study III, the potential confounders included the suggested or shown determinants of the

outcomes. Based on the previous literature, the central determinants of the risk of low birth

weight include maternal age (Kramer 1987, Lang et al. 1996), parity (Kramer 1987, Lang et al.

1996), infant sex (Kramer 1987), pre-pregnancy weight (Kramer 1987), marital status

(Koupilova et al. 2000), maternal education (Koupilova et al. 2000), socio-economic status

(Kramer 1987), alcohol consumption (Kramer 1987), and employment (Kramer 1987). The

small-for-gestational age has similar determinants as low birth weight. The most important

determinants of preterm delivery include maternal age (Lang et al. 1996), marital status

(Peacock et al. 1995) maternal education (Peacock et al. 1995), socio-economic status

(Peacock et al. 1995), and employment (Peacock et al. 1995). We were able to adjust for the

following potential confounders: age, gender, birth order, body mass index before pregnancy,

marital status, index of socioeconomic status, alcohol consumption during pregnancy and

employment during pregnancy.

Niina Jaakkola Discussion46

In Study IV, the assessment of the effect of intervention was based on within-individual change

in smoking habits over time, which eliminated the role of a number of confounders. The

compared intervention and reference groups were found to be similar as to potential modifiers

of the intervention effects.

5.2 Synthesis with previous knowledge

5.2.1 Trends and determinants of smoking during pregnancy in Finland

Smoking in pregnancy did not change substantially in Finland between 1987 (15%) and 1997

(15%). There was an increase in prevalence during the economic recession at the beginning of

the 1990s, and a decrease around 1994 back to the level of 1987. Smoking was inversely

related to age. It was more common in single than in married women, and in occupations with

low education requirements. There were also large geographic differences, the highest

prevalence of smoking being in northern and the lowest in western Finland. These

determinants of smoking remained similar, but the regional differences increased (Study I).

Table 5.1. summarizes 11 studies on the prevalence of smoking during pregnancy. The

prevalence of any smoking and of smoking throughout pregnancy are given, if available. We

showed that the prevalence of smoking during pregnancy in Finland increased from 1987 to

1990, remained relatively stable between 1990 and 1993, and decreased between 1994 and

1997 back to the level of 1987. This trend is very similar to that reported in England in

repeated surveys between 1992 and 1997 (Owen et al. 1998). In England, the smoking rate was

highest among younger women who were unemployed or manual workers. The surveys were

carried out immediately after press advertising that was aimed at reducing smoking in

pregnancy.

A study based on the Swedish Birth Registry reported that daily smoking among pregnant

women decreased from 29% in 1983 to 22% in 1992 (Cnattingius et al. 1997). Smoking

information was gathered from self-reports of smoking habits in early pregnancy when

pregnant women registered for antenatal care. This study also showed the highest smoking

prevalence during pregnancy among young women of 15 to 24 years of age. A study in

Discussion 47 Niina Jaakkola

Norway found a similar contrast in smoking during pregnancy between single and married

women as we found in our study, the prevalences being 58% and 38%, respectively (Haug et

al. 1992).

Smoking during pregnancy is more common among the less educated and single women, and

in the lower socio-economic groups. Health care system should focus health educational effort

and support to these women who are smoking.

Niin

a Ja

akko

la

48

D

iscu

ssio

n

Tabl

e 5.

1 St

udie

s on

the

prev

alen

ce o

f sm

okin

g du

ring

preg

nanc

y.A

utho

rs,

Stud

y po

pula

tion

Stud

y pe

riod

Prev

alen

ceC

omm

ents

Yea

rC

ount

ryA

nyTh

roug

h pr

egna

ncy

Isoh

anni

et a

l.Fi

nlan

d12

,068

pre

gnan

t wom

en19

6612

Wom

en a

nd th

eir c

hild

ren

from

Nor

ther

n Fi

nlan

d.19

95 9

,362

pre

gnan

t wom

en19

85-8

618

Tikk

anen

198

4Fi

nlan

d70

1 Fi

nnis

h w

omen

1981

-82

2515

Cna

tting

ius e

t al.

Swed

enA

ll liv

e si

ngle

birt

hs19

83-9

2 29

.4 in

198

3D

aily

smok

ing

and

the

attri

buta

ble

risk

of sm

okin

g fo

r SG

A b

irths

1997

(N=1

,048

,139

)21

.8 in

199

2de

crea

sed.

Erik

sson

et a

l.N

orw

ay21

,348

wom

en19

87-9

434

in 1

987

Mea

n nu

mbe

r of c

igar

ette

s per

day

dec

reas

ed fr

om 8

.6 to

7.

1996

22 in

199

4

Erik

sson

et a

l.N

orw

ay4,

766

wom

en

1994

-95

21Lo

w c

igar

ette

con

sum

ptio

n th

e be

st p

redi

ctor

of s

mok

ing

cess

atio

n.19

98si

x N

orw

egia

n ho

spita

ls

Ow

en e

t al.

En

glan

dQ

uota

sam

plin

g of

1992

-97

26 in

199

2Pr

eval

ence

and

rate

s of s

topp

ing

did

not c

hang

e. S

mok

ing

was

1998

preg

nant

wom

en26

in 1

997

high

est a

mon

g th

e yo

ung,

the

unem

ploy

ed a

nd m

anua

l wor

kers

.

Cha

teno

ud e

t al.

Italy

1,54

2 w

omen

in se

vera

l19

90-9

512

.215

.0%

stop

ped

smok

ing,

The

pre

vale

nce

was

low

er in

the

youn

gest

1999

hosp

itals

wom

en a

ged

<25y

ears

, mor

e co

mm

on a

mon

g th

e ne

ver m

arrie

d an

dth

e le

ss e

duca

ted

wom

en.

Dod

ds 1

995

Can

ada

All

wom

en w

ho h

ad b

abie

s19

88-9

232

.4

The

smok

ing

rate

cha

nged

littl

e in

Nov

a Sc

otia

.

Stew

art e

t al.

Can

ada

3,29

6 w

omen

in 1

983

1982

-93

28.5

in 1

983

Smok

ing

afte

r the

I tri

mes

ter d

ecre

ased

.19

957,

940

wom

en in

199

218

.7 in

199

2

Kirk

land

et a

l.C

anad

a8,

528

wom

en w

hose

1997

69.9

rem

aine

d33

.1%

of w

omen

wer

e sm

okin

g be

fore

pre

gnan

cy.

2000

smok

ing

stat

us b

efor

esm

oker

spr

egna

ncy

was

kno

wn

Ebra

him

et a

l.U

nite

d St

ates

8,80

3 pr

egna

nt w

omen

1987

-96

16.3

in 1

987

Smok

ing

decr

ease

d am

ong

both

pre

gnan

t and

non

preg

nant

2000

178,

499

nonp

regn

ant

11.8

in 1

996

(26.

7% to

23.

6%) w

omen

. Cig

/day

rem

aine

d at

1.

Stud

y I

Finl

and

694,

926

wom

en19

87-9

715

.0 in

198

715

.0 in

199

7__

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

__

Discussion 49 Niina Jaakkola

5.2.2 Occurrence and determinants of ETS exposure in children

A large number of studies in the past 20 years have collected information on parental smoking

and children’s exposure to ETS, usually with the assessment of health effects as the main

objective. Prevalences of exposure are not directly comparable due to differences in exposure

assessment, but they do give a reasonable view of the levels of exposure. Table 5.2.

summarizes the results of Study II and six other studies conducted in different geographic areas

in the 1990s.

Table 5.2. Exposure to ETS in children in different countries.Authors Study population Age Study PrevalenceYear Country Period %__________________________________________________________________________

Eriksen and Norway 1,046 families 6 weeks, 2-, 1993 25Bruusgaard and 4-years1995

Lund et al. all Nordic 1,000 households/ 3 years 1995 Finland: 71998a countries country, 1,500 in 1995 Sweden: 15

Denmark 1995 Norway: 321996 Denmark: 471995 Iceland: 46

Brenner and West Germany parents living in the <2 years 1987 58.8Mielck, 1993 same household as 2-5 years 49.9

the child 6-13 years 67.5

Cook et al. England and 4,043 children 5-7 years 1990 531994 Wales

Jarvis et al. England 7,451 non-smoking 11-15 years 1988-1996 52.4 in 19882000 children 45.5 in 1996

Pirkle et al. USA 3,011 children 4-11 years 1988-1991 431996

Study II Finland 2,568 children 1-7 years 1991 9.9__________________________________________________________________________

In the 1990s, the prevalence of ETS exposure was lower in Finland than in the other Nordic

countries, Germany, England and the United States. Although the prevalence of smoking

parents was approximately 40% in Espoo, the children were reported to be exposed in only

10% of the households. The residents of Espoo are wealthier on average and better educated

than most populations in Finland, and therefore knowledge about the adverse effects of passive

Niina Jaakkola Discussion50

smoking was probably well recognized and influenced smoking behavior. It is also possible

that knowledge about the adverse effects resulted in bias towards lower prevalences.

Study II was one of the first studies to systematically assess the determinants of ETS exposure

in children of smoking parents. Low parental education and single parenthood increased the

risk of exposure, whereas socioeconomic status per se had little influence after adjustment for

parental education. Education seems combined with information on health hazards to result in

a willingness to protect children, perhaps partly through better understanding of the hazards of

passive smoking. Parental education and single parenthood were also observed to be important

determinants of children's exposure to passive smoking in Scottish children aged 7 years, but in

addition, low socio-economic status increased the risk of exposure (Jarvis et al. 1992). Both

single parenthood and low socio-economic status were identified as determinants of exposure in

the Nordic countries (Lund et al. 1998a). In agreement with our study, the Norwegian study

reported that parents were more likely to avoid smoking indoors if they had a spouse/co-habitee,

and there was also a tendency for them to be more careful and to smoke less indoors with

increasing length of education (Eriksen and Bruusgaard 1995). Study II showed that the risk of

exposure was smaller for atopic children, and asthmatic children were not exposed at all. These

results suggest that knowledge of a child’s illness affects the smoking behavior of the parents.

A more recent article on the Nordic countries reported that two in three parents who smoked

did not recognize the increased risk of inner-ear infection, and 50% were not aware of the role

of parental smoking in recruiting smokers. One in two smokers tended to agree, or agreed, that

legislation should be passed forbidding all indoor smoking in the vicinity of children. The main

findings indicated that educating parents about the established health risks to their children

may significantly reduce children’s exposure to ETS (Helgason and Lund 2001).

Knowledge about the factors that influence parents’ smoking in the presence of their children

is a key question in the planning of prevention. So far, we know about some general factors on

the population level, and these could be useful in planning focused smoking cessation and

other health education campaigns. Future studies should be directed at learning more about

individual characteristics, life situations, and details of psychological and social factors, in

order to promote better understanding of the reasons for exposing children to adverse products

of one’s own behavior. Understanding these reasons is the first stage in helping the parents and

children concerned.

Discussion 51 Niina Jaakkola

5.2.3 Fetal effects of ETS exposure in pregnancy

The results of Study III, indicating a small effect of prenatal exposure to ETS on fetal growth,

are consistent with the previous studies on the relation between exposure to tobacco smoke in

pregnancy and fetal growth and prematurity summarized in Table 5.3. Windham and

colleagues (1999) carried out a meta-analysis of their own and all 22 previous studies from

1966 - 1995, and reported a pooled effect estimate of 25 g (95% CI 16 - 41). Combining all 8

studies that examined LBW gave a pooled OR of 1.00 (95% CI 0.90 - 1.10). In the present

study, the adjusted OR for LBW was 1.55 when high exposure was contrasted to the reference

group, but the 95% CI was wide (0.55 - 4.43). The corresponding estimate from 11 studies of

SGA or term LBW was 1.19 (95% CI 1.08 - 1.32), which is similar to the results of Study III

with an OR of 1.04 (95% CI 0.92 - 1.19).

Discussion 55 Niina Jaakkola

Only six of the previous studies used an objective measure of exposure (Haddow et al. 1988,

Martinez et al. 1994, Eskenazi et al. 1995, Rebagliato et al. 1995, Peacock et al. 1998, Nafstad

et al. 1998). In a study of 1,231 women, Haddow and colleagues (1988) found an effect of 104

g (95% CI 35 - 173) in exposed non-smoking women defined by serum cotinine from 1 to 10

ng/ml compared with unexposed women (<0.5 ng/ml). Their effect estimate was not adjusted

for gestational age, and therefore it incorporated the effects through shortening on gestation, as

well as, reducing intrauterine growth rate.

Eskenazi and colleagues (1995) reported a small reduction of 45 g (95% CI –36 - 126) related

to exposure assessed by serum cotinine (2 - 10 ng/ml vs. <2 ng/ml) at around 27 weeks

gestation. There was no effect on the risk of preterm delivery, and the effect on the risk of

LBW was weak (1.35, 95% CI 0.60 - 3.30). Peacock and colleagues (1998) reported a 0.2%

(95% CI –2.4 - 2.8) reduction in birthweight in newborns of women whose serume cotinine

during pregnancy was in the upper quintile compared with newborns of women in the lowest

quintile. In the case-control study by Nafstad and colleagues (1998), the risk of SGA was

related to maternal hair nicotine concentration in both non-smoking and smoking mothers. The

adjusted OR for the medium exposure category (0.75 - 4.00 µg/g ) was 3.4 (95% CI 1.3 - 8.6)

and 2.1 (95% CI 0.4 - 10.1) for the high exposure category (>4.00 µg/g ) compared with the

reference category (<0.75 µg/g ).

The possible effect of prenatal exposure to ETS on the length of gestation and the risk of

preterm delivery has received less attention, and the results of previous studies are inconsistent.

Martin and Bracken (1986) conducted a prospective cohort study of 3,891 women in New Haven,

and found no effect of passive smoking during pregnancy on gestational age. In a cohort study of

4,687 Swedish pregnant women by Ahlborg and colleagues (1991), the adjusted risk ratio for

preterm delivery was 0.49 (95% CI 0.23 - 1.06) in women exposed in the home only, and 1.27

(95% CI 0.70 - 2.31) for those exposed in the workplace. In a study of 4,644 Canadian women by

Fortier and colleagues (1994), the adjusted OR in women exposed in the home only was 0.93

(95% CI 0.64 - 1.31), and 0.92 (95% CI 0.58 - 1.51) in those exposed in the workplace only. None

of the previous studies using biomarkers reported a relation between exposure and the risk of

preterm delivery. Study III shows, for the first time, an exposure-response relation between ETS

exposure and the risk of preterm delivery.

Niina Jaakkola Discussion56

5.2.4 Effects of a smoking cessation program on maternal smoking during pregnancy

The pregnant woman registering for antenatal care, in Finland spends approximately one hour

with the midwife. She is interviewed regarding sociodemographic factors, previous obstetric

history and lifestyle factors, all of which may have an influence on the pregnancy outcome.

She is offered information and advice verbally by the midwife, as well as through booklets, on

how she can promote her infant’s health. During the forthcoming visits (the mean number of

visits is 12), the smoking woman is generally offered more information about the health

hazards of smoking, and is encouraged to stop smoking.

The smoking cessation program reported in Study IV was based on short-term training of

public health nurses, and on providing them with health education material for different

situations of routine practice. The program did not offer structured individual-based, face-to-

face intervention, such as a risk-counseling session (Windsor et al. 1993), contact by a

volunteer smoking cessation counsellor to review the quit plan, an encouraging postcard each

week to successful quitters (Hartmann 1996), or motivational counseling session provided by

the midwife (Walsh et al. 1997, Lowe 1998). The program influenced through routine

maternity care practice all the smoking pregnant women in the target population. It was

estimated to have induced smoking cessation in 4.5% (95% CI -2.9 - 11.9%) of the women,

which was slightly smaller than the effects of the Swedish (5.2%; 95% CI 1.1 - 9.3%) and

Australian programs (9 %; 95% CI 5 - 16%). However, the Swedish (Hjalmarson et al. 1991)

and Australian (Lowe et al. 1998) programs were introduced only to some of the smoking

women, while others found the program unacceptable.

Achieving changes in smoking behavior in pregnant women by means of low-intensity

interventions was also found to be difficult in the study from the United States. Intervention was

integrated into routine prenatal care, and there was no effect in cotinine-verified quit rates

(Kendrick et al. 1995). Theoretically, pregnancy and antenatal care with repeated regular visits

over a period of months constitute a unique opportunity for successful smoking cessation

intervention. In accordance with Windsor and Orleans’ study, the results of Study IV suggest that

successful smoking intervention during pregnancy requires far more than information and

encouragement (Windsor and Orleans 1986). All midwives should offer advice and supportive

strategies to women and their partners to help to reduce the prevalence of smoking during

pregnancy.

Discussion 57 Niina Jaakkola

Table 5.4. Summary of smoking cessation programs integrated into routine prenatal care.

Reference Setting Intervention Outcomeassessment

Intervention groupCessation rate (%)

Reference groupCessation rate(%)

Measure of effect(95%CI)

Hjalmarsonet al. 1991

Public healthmaternityclinics(Sweden)

Self-help manual(Windsor)obstetrician

Smokingcessation 8weeks afterdelivery

N = 44410.4%

N = 2095.2% 5.2%

(1.1-9.3)

WindsorEt al. 1993

4 publichealthmaternityclinics(Alabama)

Experimentalgroup: healtheducation, skillscounseling session,reminders, socialsupportControl group

Pretest-posttestscreening form,questionnaireSaliva cotininetest

N = 40014.3%

N = 4148.5% 5.8%

(1.4-10.1)

Kendrick etal. 1995

Healthdptments(Colorado,Maryland,Missouri)

Low-intensityinterventionsCounselingsessions, writtenmaterials

Questionnaire,urine specimen

Self-report:N=146713.0%

VerifiedN = 8886.1%

Self-report:N = 17679.5%

Verified:N = 11775.9%

3.5%OR=1.4(1.2-1.9)

0.2%OR=1.0 (0.69-1.6)

Walsh et al.1997

Antenatalclinic,teachinghospital(Australia)

Doctor’s advice,videotape, midwifecounseling, self-help manual,lottery, socialsupport, chartreminder

Smokingcessation; self-report, cotininetest

N = 127Self-report:19.0%Verified 13.0%

N = 125Self-report: 8.0%Verified6.0%

11.0% (p=0.0113)

7.0% (p=0.035)

Wakefieldand Jones1998

Publichospitalantenatalclinic(Australia)

Demo of smokingand fetal heart rate,booklet forpregnant womenand their partners

General healthquestionnaireBiochemicalverification

110 6.4%at least24wkgestation

110 1.8%at least 24wkgestation

4.6%OR=1.5 (0.9-2.6)

Ershoffet al. 1999

Typicalclinicalpractice,prenatal care(California)

Group 1) a self-help booklet;Group 2) thebooklet plus acomputerizedtelephone cessationprogram; Group 3)the booklet plusproactivetelephonecounseling

Biochemicallyconfirmedabstinencemeasured bylevel ofcotinine inurine samplesobtained inapproximatelythe 34th weekof pregnancy

Group 1)111 16.4%

Group 2) 66 18.2%

Group 3) 60 23.3%

Non-significantdifferences betweenthe groups

Study IV Public healthmaternitycare(Finland)

Midwfivecounseling, printedmaterial

SmokingcessationBiochemicalverification(hair nicotine)

N = 30658 19.0%

N = 15222 14.5%

4.5%(-2.6 – 11.6)

Niina Jaakkola Conclusions58

6. CONCLUSIONS

In spite of increasing professional knowledge about its adverse effects, and health education

efforts, smoking during pregnancy did not decline in Finland from 1987 to 1997. The analysis

of the determinants of smoking in pregnancy revealed the most important target groups for

future action to be young, less educated, unmarried women, often living in rural areas.

Surveillance of smoking in pregnancy as part of birth registration appears to be a valid source

of information for identifying special target groups for smoking prevention programs and

evaluating the influence of preventive efforts.

Children’s exposure to ETS remains a major public health concern, which calls for strong

preventive measures. Approximately 10% of the preschool children in Espoo were exposed

during the study period. The present results indicate that the resources of health education in the

prevention of passive smoking should be focused on single parents and parents with a low

educational background.

The objective assessment of exposure to ETS during the last two trimesters of pregnancy

revealed a risk of preterm delivery. There was a trend of a risk of LBW and preterm delivery

related to maternal exposure to ETS. The risk of preterm delivery was increased in the

presence of both work and home exposure (adjusted OR 8.89, 95% CI 1.05 – 75.3). The point

estimate for the presence of work exposure increased, which was consistent with the

hypothesis, but the confidence interval was wide and included unity (adjusted OR 2.35, 95%

CI 0.50 - 11.1). The risk of preterm delivery was not related to home exposure only. Nicotine

in the hair reflects maternal ETS exposure. Hair nicotine concentration could also be used to

validate active smoking, as reported in Study IV.

In the intervention area, 58/306 women (19.0%) reported quitting smoking during pregnancy,

whereas in the reference area the numbers were 22/152 (14.5%) (difference=4.5%, 95% CI

−2.6 − 11.6%). The intervention group indicated that they had received more information on

the adverse effects of smoking, had studied the material more actively, and had felt that the

material from maternity care influenced their smoking behavior more than the reference group.

Conclusions 59 Niina Jaakkola

Future smoking cessation programs need to offer more support to quitting attempts. They

should also emphasize the harmful effects on women’s own health as an additional reason for

quitting. A task for the future is to find methods that are both effective and feasible for use in

routine maternity clinics. The effectiveness of the methods could be assessed in randomized

controlled trials. However, effectiveness and feasibility should also be evaluated in settings in

which the influence of the evaluation itself is minimal.

Young children’s greatest exposure to ETS occurs at home. Programs to raise awareness and

motivate behavioral change among pregnant women and their partners are needed to reduce the

harmful effects of prenatal and postnatal exposure to tobacco smoke. Smoking in the

workplace and public places has decreased. Smokers could be reminded that, as they do not

expose their coworkers among the general public in many places, they should extend the same

protection to their children and partners. ETS awareness and control strategies for the home, in

addition to smoking cessation strategies, should be incorporated into tobacco education

programs in schools. Otherwise, smoking young girls will be smoking mothers in the future.

Niina Jaakkola 60 Acknowledgements

7. ACKNOWLEDGEMENTS

This work was carried out at the Department of Public Health, University of Helsinki. I was

lucky to work in an inspiring and multidisciplinary environment throughout the years. I wish to

express my gratitude to the former head of the Department of Public Health, the late Professor

Olli P. Heinonen, and to Professor Mats Brommels and Professor Seppo Sarna, for the facilities

the department provided me with so that I could complete my academic work.

The research work was supported by the Ministry of Social Affairs and Health, the Doctoral

Programs in Public Health, the Ella and Georg Ehrnrooth Foundation, and the Women’s

Science Foundation. They are all gratefully acknowledged.

I wish to express my deep and sincere thanks to Professor Jouni J.K. Jaakkola for his

supervision. Without his expertise and guidance, this dissertation would not have been

completed. He introduced me to the field of scientific research, and has provided me with

inspiring guidance and friendship during the many years we have been working together. In

addition, I wish to acknowledge Professor Markku Koskenvuo and Docent Ossi Rahkonen,

who critically reviewed the manuscript and helped to improve it by giving important

suggestions. I also thank Ms. Joan Nordlund for revising the English of the thesis.

I wish to thank all my co-authors, who kindly split some of their time to read and comment on

the manuscripts. I would especially like to thank Risto Ruotsalainen, D.Sc. for his valuable

comments and suggestions. I am most grateful to all my colleagues at the Department of Public

Health and all friends who showed interest in how I was advancing with my dissertation.

I wish to express warm thanks to my Father, and to my son Simo-Pekka, for their valuable

contributions to the data gathering. The data collection would not have been possible without

their help.

Finally, my most heartfelt thanks go to my dear husband for his skillful technical assistance

with the computer technology, and for his love, understanding and support throughout these

interesting and busy years.

Helsinki, September 2002

Niina Jaakkola

Niina Jaakkola 61 References

8. REFERENCES

Ahlborg G Jr, Bodin L. Tobacco smoke exposure and pregnancy outcome among workingwomen. A prospective study at prenatal care centers in Orebro County, Sweden. Am JEpidemiol 1991;133:338-347.

Ahlsten G, Cnattingius S, Lindmark G. Cessation of smoking during pregnancy improves fetalgrowth and reduces infant morbidity in the neonatal period. A population-based prospectivestudy. Acta Paediatrica 1993;82:177-181.

Apgar J. Zinc and reproduction. Annu Rev Nutr 1985;5:43-48.

Ashley MJ, Cohen J, Ferrence R, Bull S, Poland B, Pederson I. Smoking in the home: changingattitudes and current practices. Am J Public Health 1998;88:797-800.

Borland R, Mullins R, Trotter L, White V. Trends in environmental tobacco smoke restrictions inthe homes in Victoria, Australia. Tob Control 1999; 8:266-271.

Brenner H, Mielck A. Children’s exposure to parental smoking in West Germany. Int J Epidemiol1993;22:818-823.

Brooke OG, Anderson HR, Bland JM, Peacock JL, Stewart CM. Effects on birth weight ofsmoking, alcohol, caffeine, socioeconomic factors, and psychosocial stress. BMJ 1989;298: 795-801.

California Environmental Protection Agency. Office of Environmental Health HazardAssessment. Health effects of exposure to environmental tobacco smoke. California, 1997.

Cambell MJ, Lewry J, Wailoo M. Further evidence for the effects of passive smoking onneonates. Postgrad Med J 1988;64:663-665.

Chatenoud L, Chiaffarino F, Parazzini F, Benzi G, LaVecchia C. Prevalence of smokingamong pregnant women is lower in Italy than England. BMJ 1999;318:1012.

Chen Y, Pederson LL, Lefcoe NM. Passive smoking and low birthweight. Lancet 1989;ii:54-55.

Cnattingius S, Haglund B. Decreasing smoking prevalence during pregnancy in Sweden: Theeffect on small-for-gestational-age births. Am J Public Health 1997;87(3):410-413.

Comstock GW, Lundin FE. Parental smoking and perinatal mortality. Am J Obstet Gynecol

1967;98:708-718.

Cook DG, Whincup PH, Jarvis MJ, Strachan DP, Papacosta O, Bryant A. Passive exposure totobacco smoke in children aged 5-7 years: individual, family, and community factors. BMJ1994;308:384-9.

Cook DG, Strachan DP. Health effects of passive smoking –10: summary of effects of parentalsmoking on respiratory health of children and implications for research. Thorax 1999;54:357-366.

Coultas DB, Peake GT, Samet JM. Questionnaire assessment of lifetime and recent exposure toenvironmental tobacco smoke. Am J Epidemiol 1989;130:338-347.

Niina Jaakkola References62

Cummings KM, Markello SJ, Mahoney MC, Marshall JR. Measurement of lifetime exposure topassive smoking. Am J Epidemiol 1989;130:122-132.

DiFranza JR, Lew RA. Morbidity and mortality in children associated with the use of tobaccoproducts by other people. Pediatrics 1996;97:560-8.

Dodds L. Prevalence of smoking among pregnant women in Nova Scotia from 1988 to 1992.Can Med Assoc J 1995;152:185-190.

Dolan-Mullen P, Ramirez G, Groff J. A meta-analysis of randomized trials of prenatal smokingcessation interventions. Am J Obstet Gynecol 1994;171:1328-34.

Ebrahim SH, Floyd RL, Merrit RK, Decoufle P, Holzman D. Trends in pregnancy-relatedsmoking rates in the United States, 1987-1996. JAMA 2000;283(3):361-6.

England LJ, Kendrick JS, Gargiullo PM, Zahniser SC, Hannon WH. Measures of maternaltobacco exposure and infant birth weight at term. Am J Epidemiol 2001;153:954-960. (a)

England LJ, Kendrick JS, Wilson HG, Merritt RK, Gargiullo PM, Zahniser SC. Effects ofsmoking reduction during pregnancy on the birth weight of term infants. Am J Epidemiol2001;154:964-701. (b)

Eriksen W, Bruusgaard D. Do parents take practical measures to prevent passive smoking bythe chilren? Scan J Prim Health Care 1995;13:275-80.

Eriksson KM, Salvesen KA, Haug K, Eik-Nes SH. Smoking habits among pregnant women inNorway county 1987-94. Acta Obstet Gynecol Scan 1996;75:355-9.

Eriksson KM, Haug K, Salvesen KA, Nesheim BI, Nylander G, Rasmussen S, Andersen K.Nakling JO, Eik-Nes SH. Smoking habits among pregnant women in Norway 1994-95. ActaObstet Gynecol Scan 1998;77:159-64.

Ershoff DH, Dolan-Mullen P, Quinn VP. A randomised trial of a serialized self-help smokingcessation program for pregnant women in HMO. Am J Public Health 1989;79:182-187.

Ershoff DH, Quinn VP, Boyd NR, Stern J, Gregory M, Wirtschafter D. The Kaiser Permanenteprenatal smoking-cessation trial: when more isn't better, what is enough? Am J Prev Med1999;17:161-8.

Eskenazi B, Prehn AW, Christiansson RE. Passive and active maternal smoking as measuredby serum cotinine: the effect on birthweight. Am J Public Health 1995;85:396-398.

Fleiss JL. Statistical methods for rates and proportions (2nd edition). New York: John Wiley &Sons; 1981.

Fortier I, Marcoux S, Brisson J. Passive smoking during pregnancy and the risk of delivering asmall-for-gestational-age infant. Am J Epidemiol 1994;139:294-301.

Gebauer C, Kwo CY, Haynes EF, Wewers ME.A nurse-managed smoking cessationintervention during pregnancy. JOGN Nurs 1998;21:47-53.

Gissler M, Teperi J, Hemminki E, Meriläinen J. Data quality after re-stucturing a nationwidemedical birth registry. Scand J Soc Med 1995;23:75-80

References Niina Jaakkola63

Guerin MR, Jenkins RA, Tomkins BA. Chemistry of environmental tobacco smoke:composition and measurement. Lewis Publisher, Boca Raton, 1992.

Haddow JE, Knight GJ, Palomäki GE, McCarthy JE. Second-trimester serum cotinine levels innonsmokers in relation to birth weight. Am J Obstet Gynecol 1988;159:481-484.

Haddow JE, Knight GJ, Palomäki GE, x. Cotinine-assisted intervention in pregnancy to reducesmoking and low birth-weight delivery. Br J Obstet Gynaecol 1991; 98:859-865.

Haglund B, Cnattingius S. Cigarette smoking as a risk factor for sudden infant death syndrome:a population based study. Am J Public Health 1990;80:29-32.

Hartmann KE, Thorp JM, Pahel-Short MD, Koch MA. A randomized controlled trial ofsmoking cessation intervention in pregnancy in an academic clinic. Obstet Gynecol1996;87:621-6.

Haug K, Aaro LE, Fugelli P. Smoking habits in early pregnancy and attitudes towards smokingcessation among pregnant women and their partners. Fam Pract 1992;9:494-499.

Helakorpi S, Uutela A, Prättälä R, Puska P. Health behaviour and health among Finnish adultpopulation, Spring 2000 Finnish). Publications of the National Public Health Institute.B8/2000. Helsinki 2000.

Helakorpi S, Patja K, Prättälä R, Uutela A. Health behaviour and health among Finnish adultpopulation, Spring 2001 Finnish). Publications of the National Public Health Institute.B16/2001. Helsinki 2001.

Helgason ÁR, Lund KE. Environmental tobacco smoke exposure of young children –attitudesand health-risk awareness in the Nordic countries. Nicotine & Tobacco Research 2001;3:341-345.

Hjalmarson AIM, Hahn L, Svanberg B. Stopping smoking in pregnancy: effect of self−helpmanual in controlled trial. Br J Obstet Gynaecol 1991;98:260-4.

Isohanni M, Oja H, Moilanen I, Koiranen M, Rantakallio P. Smoking or quitting duringpregnancy: associations with background and future social factors. Scand J Soc Med 1995;23:32-38.

Jaakkola MS, Jaakkola JJK. Assessment of exposure to environmental tobacco smoke. EurRespir J 1997;10:2384-97.

Jaakkola JJK, Jaakkola MS. Effects of environmental tobacco smoke on respiratory health inchildren. Scand J Work Environ Health 2002;28 suppl 2:73-86.

Jarvis MJ, Strachan DP, Feyerabend C. Determinants of passive smoking in children inEdinburgh, Scotland. Am J Public Health 1992;82:1225-9.

Jarvis MJ. Children’s exposure to passive smoking: survey methodology and monitoringtrends. Background paper. 1999. Tobacco Free initiative.(http://tobacco.who.int/toh/TF/consult.htm).

Jarvis MJ, Coddat E, Higgins V, Feyerabend C, Bryant A, Cook DG. Children’s exposure topassive smoking in England since the 1980s: cotinine evidence from population surveys. BMJ2000;321:343-5.

Niina Jaakkola References64

Kendrick JS, Zahniser C, Miller N, Salas N, Stine J, Gargiullo PM, Floyd L, Spierto FW,Sexton M, Metzger RW, Stockbauer JW, Hannon H, Dalmat ME. Integrating smokingcessation into routine public prenatal care: the smoking cessation in pregnancy project. Am JPublic Health 1995;85:217-22.

Kirkland SA, Dodds L, Bronsky G. The natural history of smoking during pregnancy in NovaScotia. Can Med Assoc J 2000;163(3):281-2.

Kottke TE, Battista RN, DeFriese GH, Krekke ML. Attributes of successful smoking cessationinterventions in medical practice. A meta-analysis of 39 controlled trials. JAMA1988;259:2882-2889.

Koupilova I, Rahu K, Rahu M, Karro H, Leon DA. Social determinants of birthweight andlength of gestation in Estonia during the transition to democracy. Int J Epidemiol 2000;29:118-124.

Kramer MS. Determinants of low birth weight: methodological assessment and meta-analysis.Bull WHO 1987;65:663-737.

Kuhnert B, Kuhnert PM, Debanne S, Williams TG. The relationship between cadmium, zinc, andbirth weight in pregnant women who smoke. Am J Obstet Gynecol 1987;157:1247-1251.

Lang JM, Lieberman E, Cohen A. A comparison of risk factors for preterm labor and termsmall-for-gestational-age birth. Epidemiology 1996;7:369-376.

Larsson ML, Frisk M, Hallström J, Kiviloog J, Lundbäck B. Environmental smoke exposureduring childhood is associated with increased prevalence of asthma in adults. Chest2001;120:711-717.

Lazzaroni F, Bonassi S, Manniello E, Morcaldi L, Repetto E, Ruocco A, Calvi A, Cotellessa G.Effect of passive smoking during pregnancy on selected perinatal parameters. Int J Epidemiol1990;19:960-966.

Li CQ, Windsor RA, Perkins L, Goldenberg RL, Lowe JB. The impact on infant birth weightand gestational age of cotinine-validated smoking reduction during pregnancy. JAMA1993;269:1519-1524

Lowe JB, Balanda KP, Clare G. Evaluation of antenatal smoking cessation programs forpregnant women. Aust N Z J Public Health 1998;22:55-9.

Lumley J, Oliver S, Waters E. Smoking cessation implemented during pregnancy (Cochranereview). The Cochrane Library, 1998, disk issue.

Lumley J, Oliver S, Waters E. Interventions for promoting smoking cessation duringpregnancy (Cochrane review). In: The Cochrane Library, 2000. Oxford: Update Software2000.

Lund KE, Skrondal A, Vertio H, Helgason ÁR. Children’s residential exposure toenvironmental tobacco smoke varies greatly between the Nordic countries? Scand J Soc Med1998;26:115-20.(a)

Lund KE, Skrondal A, Vertio H, Helgason ÁR. To what extent do parents strive to protect theirchildren from environmental tobacco smoke in the Nordic countries? A population-basedstudy. Tob Control 1998;7:56-60.(b)

References Niina Jaakkola65

MacArthur C, Knox EG. Passive smoking and birthweight. Lancet 1987;ii:37-8 .

MacMahon B, Alpert M, Salber EJ. Infant weight and parental smoking habits. Am JEpidemiol 1966;82:247-261.

Magnus P, Berg K, Bjerkedal T, Nance WE. Parental determinants of birthweight. Clin Gen1984;26:397-405.

Mannino DM, Caraballo R, Benowitz N, Repace J. Predictors of cotinine levels in US children:data from the Third National Health and Nutrition Examination Survey. Chest 2001;120:718-724.

Martin TR, Bracken MB. Association of low birth weight with passive smoke exposure inpregnancy. Am J Epidemiol 1986;124:633-642.

Martinez FD, Wright AL, Taussig LM. and the Group Health Medical Associates. The effectof paternal smoking on the birthweight of newborns whose mothers did not smoke. Am JPublic Health 1994;84:1489-1491.

Mathai M, Skinner A, Lawton K, Weindling AM. Maternal smoking, urinary cotinine levelsand birth-weight. Aust N Z J Obstet Gynaecol 1990;30:33-36.

Meber A, Sande H, Foss OP, Stenwig JT. Smoking during pregnancy - effects on the fetus and onthiocyanate levels in mother and baby. Acta Paediatr Scand 1979;68:547-552.

Nafstad P, Kongerud J, Botten G, Hagen J, Jaakkola JJK. The role of passive smoking indevelopment of bronchial obstuctions during the first 2 years of life. Epidemiology1997;8:293-297.

Nafstad P, Fugelseth D, Qvigstad E, Zahlsen K, Magnus P, Lindemann R. Nicotineconcentration in the hair of nonsmoking mothers and size of offspring. Am J Public Health1998;88:120-4.

Nafstad P, Magnus P, Jaakkola JJ. Risk of childhood asthma and allergic rhinitis in relation topregnancy complications. J Allergy Clin Immunol 2000;106:867-73.

Ockene JK. Smoking among women across the life span: prevalence, interventions, andimplications for cessation research. Ann Behav Med 1993; 15:135-148.

Ogawa H, Tominaga S, Hori K, Noguchi K, Kanou I, Matsubara M. Passive smoking bypregnant women and fetal growth. J Epidemiol Community Health 1991;45:164-8.

Olsen J. Cigarette smoking in pregnancy and fetal growth. Does the type of tobacco play arole? Int J Epidemiol 1992;21:279-284.

Owen L, McNeill A, Callum C. Trends in smoking during pregnancy in England, 1992-7:quota sampling surveys. BMJ 1998;317:728.

Panjari M, Bell R, Bishop S, Astbury J, Rice G, Doery J. A randomized controlled trial of asmoking cessation intervention during pregnancy. Aust N Z J Obstet Gynecol 1999;39:312-317.

Peacock JL, Bland JM, Anderson HR. Preterm delivery: effects of socioeconomic factors,psychosocial stress, smoking, alcohol, and caffeine. BMJ 1995;311:531-535.

Niina Jaakkola References66

Peacock JL, Cook GC, Carey IM, Jarvis MJ, Bryant AE, Anderson HR, Bland JM. Maternalcotinine level during pregnancy and birthweight for gestational age. Int J Epidemiol1998;27:647-656.

Pirkle JL, Flegal KM, Bernert JT, Brody DJ, Etzel RA, Maurer KR. Exposure to the USpopulaiton to environmental tobacco smoke. The thierd national health and nutritionexamination survey, 1988 to 1991. JAMA 1996;275:1233-40.

Rahkonen O, Berg MA, Puska P. The development of smoking in Finland from 1978 to 1990.Br J Addict 1992;87:103-110.

Rebagliato M, Florey C du V, Bolumar F. Exposure to environmental tobacco smoke innonsmoking pregnant women in relation to birth weight. Am J Epidemiol 1995;142:531-537.

Rimpelä A, Vikat A, Rimpelä M, Lintonen T, Ahlström S, Huhtala H. Nuortenterveystapatutkimus 1999. Tupakoinnin ja päihteiden käytön muutokset. (18/1999) Stakes.Helsinki, Stakesin monistamo 1999.

Roquer JM, Figueras J, Botet F, Jiménez R. Influence on fetal growth of exposure to tobaccosmoke during pregnancy. Acta Paediatr Scand 1995;84:118-121.

Rubin HD, Krasilnikoff PA, Leventhal JM, Berget A. Effect of passive smoking on birth-weight. Lancet 1986; ii:415-417.

Sadler L, Belanger K, Saftlas A, Leaderer B, Hellenbrand K, McSharry J-E, Bracken MB.Environmental tobacco smoke exposure and small-for-gestational-age birth. Am J Epidemiol1999;150:695-705.

Saito R. The smoking habits of pregnant women and their husbands, and the effect on theirinfants. Nippon Koshu Eisei Zasshi 1991;38:124-131.

Schwartz-Bickenbach D, Schulte-Hobein B, Abt S, Plum C, Nau H. Smoking and passivesmoking during pregnancy and early infancy: effects on birthweight, lactation period, andcotinine concentrations in mother’s milk and infant’s urine. Toxicol Letters 1987;35:73-81.

Sexton M, Hebel JA. Clinical trial of change in maternal smoking and its effect on birthweight.JAMA 1984;251:911-915.

Simpson WJ, Loma L. A preliminary report on cigarette smoking and the incidence ofprematurity. Am J Obstet Gynaecol 1957;73:808-815.

Spitzer WO, Lawrence V, Dales R, Hill G, Archer MC, Clark P, Abenhaim L, Hardy J, SampalisJ, Pinfold SP, Morgan PP. Links between passive smoking and disease: a best-evidence synthesis.A report of the Working Group on Passive Smoking. Clin Invest Med 1990;13:17-42.

Stewart PJ, Potter J, Dulberg C, Niday P, Nimrod C, Tawagi G. Change in smoking prevalenceamong pregnant women 1982-93. Can J Public Health 1995;86:37-41.

Stokes ME, Davis CS, Koch GG. Categorical Data Analysis Using the SAS System. Cary, NC.SAS Institute Inc., 1995.

Tikkanen J. Smoking and alcohol consumption during pregnancy in Finland (Finnish). Seriesoriginal reports 4/1984. Publications of the National Board of Health, Finland. Helsinki 1984.

References Niina Jaakkola67

Underwood PB, Kesler KF, O’Lane JM, Callagen DA. Parental smoking empirically related topregnancy outcome. Obstet Gynecol 1967;29:1-8.

U.S. Environmental Protection Agency (EPA). Respiratory health effects of passive smoking:Lung cancer and other disorders. U.S. Environmental Protection Agency, Office of Health andEnvironmental Assessment, Office of Research and Development. Washington D.C., 1992.EPA/600/6-90/006F.

Wakefield M, Jones W. Effects of a smoking cessation program for pregnant women and theirpartners attending a public hospital antenatal clinic. Aust NZ J Public Health 1998;22:313-320.

Valbø A, Nylander G. Smoking cessation in pregnancy. Intervention among heavy smokers.Acta Obstet Gynecol Scand 1994;73:215-9.

Walsh R, Redman S. Smoking cessation in pregnancy: do effective programmes exist? HealthPromot Internation 1993;8:111-127.

Walsh R, Redman S. Brinsmead MW, Byrne JM, Melmeth A. Smoking cessation program at apublic antenatal clinic. Am J Public Health 1997;87:1201-4.

Windham GC, Eaton A, Hopkins B. Evidence for an association between environmentaltobacco smoke exposure and birthweight: a meta-analysis and new data. Paediatr PerinatEpidemiol 1999;13:35-57.

Windham GC, Hopkins B, Fenster L, Swan SH. Prenatal active or passive tobacco smokeexposure and the risk of preterm delivery or low birth weight. Epidemiology 2000;11:427-433.

Windsor RA, Orleans CA. Guidelines and methodological standards for smoking cessationintervention research among pregnant women: improving the science and art. Health Educ Q1986;13:131-61.

Windsor RA, Cutter G, Morris J, Reese Y, Manzella B, Bartlett EE, Samuelson C, Spanos D.The effectiveness of smoking cessation methods for smokers in public health maternity clinics:A randomized trial. Am J Public Health 1985;75:1389-92.

Windsor RA, Lowe J, Perkins LL, Smith-Yoder, Artz L, Crawford M, Amburgy K, Boyd NR.Health education for pregnant smokers: its behavioral impact and cost benefit. Am J PublicHealth 1993;83:201-6.

World Health Organization. Division of Noncommunicable Diseases. Tobacco Free Initiative.International consultation on environmental tobacco smoke (ETS) and child health.Consultation report. Geneva: WHO, 1999. (http://tobacco.who.int/en/health/int-consult.html+)

Zahlsen K, Nilsen OG. Nicotine in hair of smokers and non-smokers: Sampling procedure andgas chromatographic/mass spectrometric analysis. Pharmacol Toxicol 1994;75:143-149.

Zhang J, Ratcliffe JM. Paternal smoking and birthweight in Shanghai. Am J Public Health1993;83:207-21

Niina Jaakkola 68 Original publications

ORIGINAL PUBLICATIONS

I Jaakkola N, Jaakkola MS, Gissler M, Jaakkola JJK. Smoking during pregnancy in Finland:determinants and trends, 1987-1997. Am J Public Health 2001;91:284-286.

II Jaakkola N, Ruotsalainen R, Jaakkola JJK. What are the determinants of children’sexposure to environmental tobacco smoke at home? Scand J Soc Med 1994;21:107-112.

III Jaakkola JJK, Jaakkola N, Zahlsen K. Fetal growth and length of gestation in relation toprenatal exposure to environmental tobacco smoke assessed by hair nicotine concentration.Environ Health Perspect 2001;109:557-561.

IV Jaakkola N, Zahlsen K, Jaakkola JJK. Effects of a population-based smoking cessationprogramme on smoking in pregnancy. Eur J Public Health 2001;11:446-449.