SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. References
  9. Appendix

Objective To test the stress hypothesis that women who give birth to small for gestational age infants lack important psychosocial coping resources, such as a sufficient social network, social support and control in daily life.

Design A prospective cohort study of nulliparous pregnant women.

Setting Antenatal care units in the city of Malmö, Sweden.

Population All women (n= 994) during a one year period (1991–1992) were invited, and 872 (87.7%) participated. This study was restricted to pregnancies resulting in singleton live birth (n= 826); 6.7% of infants were classified as small for their gestational age.

Methods Self-administered questionnaires were given to all women at the time of their first antenatal visit.

Main outcome measures The classification of small for gestational age was based on a gender-specific intrauterine growth reference curve. Newborn babies were classified as being small for gestational age if their birthweight was > 2 SD below the mean weight for gestational age.

Results Lack of psychosocial resources, such as social stability, social participation, emotional and instrumental support, all increased the likelihood of delivering an infant that was small for gestational age. The odds ratios when controlled for demographic background factors, lifestyle factors and anthropomorphic measures were: OR 1.7 (95% CI0.9–3.3) for women with poor social stability; OR 2.2 (95% CI 1.1–4.4) for women with poor social participation; OR 2.6 (95% CI 1.2–5.7) for women with poor instrumental support; and OR 1.5 (95% CI 0.8–2.8) for women with poor emotional support. Simultaneous exposure to a poor total network index, as well as a poor total support index showed a significantly increased odds ratio for having a small for gestational age baby: OR 3.3 (95% CI 1.6–6.7) and OR 2.7 (95% CI 1.3–5.6), respectively. A synergy index of 9.0 and 6.8 supported the assumption of an interaction between immigrant status and poor total network or poor total support, in a synergistic direction.

Conclusions The findings support the stress-hypothesis that a lack of psychosocial resources might increase the risk of giving birth to a baby that is small for gestational age.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. References
  9. Appendix

Intrauterine growth retardation, resulting in low birth-weight for a given gestational age, is a main concern in obstetrics. Small for gestational age birthweight has frequently been used to define intrauterine growth retardation, but we still lack sufficient insight into the causes of why babies are born small for their gestational age. Different maternal factors are found to be related to intrauterine growth retardation, such as parity1, age1, race1, prepregnancy weight1–3, height1,4, educational level5,6, income3,4, as well as alcohol intake1,6, and smoking1–3,7–9.

Psychosocial resources, such as social network and social support, are important for maintaining good health10. A major area of research on social networks, social support and health is stress and stressful life events11–13. Selye11 described the General Adaptation Syndrome in 1946 which states that an individual's relation to his/her environment may be viewed as a dynamic process, since the environment changes all the time, requiring continuous adaptation by the individual. Stress was seen by Selye as a primary and nonspecific tation. In 1976 Cassel14 presented his theory of general susceptibility, which emphasised that negative psychosocial influences in the environment may decrease an individual's resistance to pathogenic agents. Cassel's theory provided an important theoretical link between stress research and social epidemiology. The availability of psychosocial resources, such as social network and social support, has been shown to be important for individuals, enabling them to cope with different, potentially stressful, situations in daily life15–18.

Pregnancy, especially the first one, is a new and potentially stressful situation with demands, constraints and opportunities which require adaptation19–20. Although pregnancy itself could be considered to constitute a stressful experience, it is not uniformly stressful among women20.

Many investigators have studied psychosocial factors related to birthweight, but research findings in this area have not been consistent and comparisons across studies are difficult to make because of discrepancies both in design and methodology3–9,21.

The objective of this prospective study was therefore, to test the hypothesis that women who give birth to small for gestational age infants during their first pregnancy lacked important psychosocial coping resources, such as a sufficient social network, social support and control in daily life, and by an analysis of interaction, as proposed by Rothman22, identify susceptibility among subgroups of pregnant women. The model chosen to represent the hypothesis is demonstrated in Fig. 1.

image

Figure 1. A model illustrating the relation between psychosocial resources, lifestyle factors and the risk for small for gestational age (SGA) children.

Download figure to PowerPoint

METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. References
  9. Appendix

During the period September 1991-September 1992, all nulliparous women at four antenatal care clinics in the city of Malmö (240,000 inhabitants), were invited to participate in the study. Malmö was chosen because 85%–90% of all pregnant women in the city could be covered by including one private and three public antenatal care clinics. A total of 872 of the 994 invited women (87.7%) agreed to participate. One hundred and and of these, 94 (47.7%) were European and 103 (52.3%) were non-European. At the time of recruitment 115 of the immigrants (58.3%) had lived in Sweden for 12 months or less. Interpreters were used for about half (45.2%) of the immigrant women. This study included all pregnancies resulting in a singleton live birth (n= 826). Women with multiple pregnancies n= 12 (1.4%), those with subsequent miscarriages n= 17 (1.9%), and finally those lost to follow up at delivery n= 17 (1.9%), were excluded.

The baseline investigations of the nulliparous women took place at their first antenatal visit (on average, in their 12th postmenstrual week). Each woman was asked, confidentially and of her own free will, to complete a questionnaire before seeing the midwife. The questionnaire contained demographic background factors (e.g. age, educational level, country of birth, marital and cohabiting status), and lifestyle factors and psychosocial factors, such as social network, social support and control in daily life. Data on maternal biomedical risk factors for having a baby that was small for gestational age were collected from the perinatal database; data on maternal prepregnancy weight and height was taken from the medical records at the Department of Obstetrics and Gynaecology, Malmö University Hospital.

All pregnant women at the Department of Obstetrics and Gynaecology, Malmö University Hospital, are routinely examined by ultrasound twice during the pregnancy at 16 to 18 and at 32 weeks of gestation. In this study group 806 women (97.6%) were examined twice. At the first ultrasound examination, the pregnancy was dated and fetal anatomy examined. At the second examination, fetal weight was estimated and the deviation from the expected weight was calculated. Data on the outcome of pregnancy, such as birthweight and gestational age, were collected from the perinatal database of the Department of Obstetrics and Gynaecology at Malmö University Hospital, and from medical records at other Swedish hospitals where 83 (10.0%) women gave birth.

Gestational age was established on the basis of ultrasound examination performed before 20 weeks of gestation. The reliability of ultrasound fetometry for estimating gestational age was found to be high with a standard deviation (SD) from true gestational age of only 2.4 days:

  • image

where GA = gestational age, BPD = biparietal diameter and FL = femur length.

In order to examine the small for gestational age variable, a deviation from the expected birthweight was assessed according to an intrauterine growth reference curve24. This reference curve is based on longitudinal ultrasound estimations of fetal weights in uncomplicated pregnancies and data are collected from four perinatal centres in Sweden and Denmark. The SD value was determined at 12% for both boys and girls uniformly for all gestational weeks. If the newborns had a birthweight of > 2 SD below the mean on the gender specific reference curve, they were classified as being small for gestational age24. The small for gestational age variable was analysed as a dichotomous variable.

Maternal age was classified into three groups: 15–19, 20–29 and 30–44 years of age. Data on prepregnancy maternal weight and height were obtained from the women at their first antenatal visit. Maternal weight was divided into low (< 50 kg), normal (50–75 kg) and high prepregnancy weight (> 75 kg), and maternal height into normal (> 157 cm) and short height (≤ 157 cm)25. Country of origin was based on whether the women were born in Sweden or not. Educational level was based on the women's years of education and divided into two groups: 12 years or fewer, and more than 12 years. Cohabitation status depended on whether the woman was living with the child's father or not.

Maternal smoking was divided into four groups based on the numbers of cigarettes smoked per day: 1. non-smokers (women who at their first antenatal visit reported that they did not smoke); 2. women who reported smoking 1–9; 3. 10–19; and 4. ≥ 20 cigarettes per day.

Alcohol consumption was defined by those women who at the first antenatal visit reported that they did not drink alcohol at all (nonconsumers) and those who were drinking (consumers).

Physical exercise was assessed by one question. High physical exercise was defined as running, swimming, gymnastics and playing tennis or badminton for at least two hours per week. Moderate physical exercise was defined as engaging in activities such walking or biking or light gardening for at least four hours a week and low physical exercise was defined as mostly sedentary activities, such as reading or watching TV26. The low and moderate categories were combined in the analyses.

Based on a scoring system for selection of pregnancies that were at risk of delivering a small for gestational age infant, women with a history of hypertension, urinary tract infection, bleeding and/or preterm contractions during the current pregnancy were identified from the perinatal database of the Department of Obstetrics and Gynaecology, Malmö University Hospital27.

Psychosocial variables

A model based on a perspective of social resources which has been validated and described elsewhere was used15,17,28,29. The model consists of two main concepts, social network and social support.

A person's social network is considered as a structural concept and was here defined using two sub-concepts:

  • 1
    Social stability: this describes the degree to which a person belongs to, and has stability within, formal and informal groups such as family, club memberships, neighbourhood, friends and the feeling of membership in these-groups [four items].
  • 2
    Social participation: this describes how actively the person takes part in social activities of formal and informal groups in society [13 items].

Social support is regarded as a function of the person's interactions with her social network and was divided into five sub-concepts:

  • 1
    Emotional support: this reflects the person's experience of receiving care, encouragement of personal value and feelings of confidence and trust from relatives, friends, neighbours and colleagues [three items].
  • 2
    Instrumental support: this measures a person's access to advice, information, and practical service [one item].
  • 3
    Support from the child's father refers to the degree of perceived support from the child's father in a general sense and whether he was expected to accompany his partner to childbirth classes [two items].
  • 4
    Maternal support refers to the degree of support a woman receives from her own mother [one item].
  • 1
    Job support: this describes the work place atmosphere and conditions including the understanding and support received from supervisors and workmates, was defined by Johnson30 and assessed by an instrument modified by Theorell et al.31 [six items].

Based on the items of these concepts, seven social network and social support indices were created. The scores from each index were divided into high/low, as close to the lowest tertile as possible. The lowest third of the distribution was defined as low. However, since the distributions of some of the indices were skewed and based on a categorical scale the prevalence of the low categories varied somewhat between the indices: social stability (25.1%); social participation (39.6%); emotional support (30.0%); instrumental support (13.8%); support from the child's father (14.2%); maternal support (28.5%); and job support (44.8%). Control of daily life is a measurement of chronic stress and refers to the woman's experiences of being able to influence important aspects of her daily life, including how she deals with personal problems, according to a method described and validated by Cohen and Williamson32 [four items]. All items in the questionnaire are presented in Appendix 1.

Statistical methods

The t test for equality of means was used to analyse differences in maternal weight and height between women who gave birth to babies that were small for gestational age (SGA) and those who did not. Maternal age, height, weight, country of birth, educational level, physical exercise, alcohol consumption and smoking were regarded as potential confounders because they are associated with both the dependent variable (e.g. SGA) and the independent psychosocial variables (e.g. social anchorage, social participation, emotional support and instrumental support). A series of multiple logistic regression analyses were performed to assess the contribution of the psychosocial variables to the SGA risk, after controlling for potential confounding factors. Age, height and weight were included in the model (Model I) as a first step in the analysis. In the second step, the background variables country of origin and educational level were added (Model II). Finally, the lifestyle factors physical exercise, alcohol consumption and maternal smoking were added since we assume that these factors should be considered, at least partially, as intermediate factors between psychosocial resources and SGA according to the model in Fig. 1 (Model III). The effect of maternal biomedical risk factors was assessed in a fourth step. In order to estimate the effect of simultaneous exposure to low social anchorage and low social participation (total network index), as well as low instrumental and emotional support (total support index) on SGA, logistic regression analyses were performed with the additional psychosocial variables. To identify susceptibility among subgroups of pregnant women, such as those who were not born in Sweden or who had a low level of education, the interaction effect was analysed as proposed by Rothman22. The variables assessing country of birth and psychosocial resources (total social network index and total social support index, respectively) were combined into dummy variables, so that women born in Sweden who were not exposed to low psychosocial resources constituted the reference category. The other exposure categories consisted of women of foreign origin who were not exposed to low psychosocial resources, women of Swedish origin who were exposed to low psychosocial resources, and women who were both of foreign origin and were exposed to low psychosocial resources. The same procedure was made for educational level and psychosocial resources. The following algorithm was used to determined interaction, resulting in a synergy index (SI), where departure from one signified an interaction effect22:

  • image

where RR = risk ratio; Ab = exposed to one of the two factors; aB = exposed to the other factor; AB = exposed to both factors.

Maternal age was introduced in the analyses as a continuous variable since there was a linear association between age and SGA risk. Maternal smoking was assessed at four levels in order to estimate the effect of different amounts of smoking. All other potential confounding factors were dichotomised. Differences at the 0.05 level were considered statistically significant. The statistical analyses were performed using the SPSS programme33.

The study was approved by the Ethics Committee of Lund University and by the Swedish Data Inspection Board.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. References
  9. Appendix

Background characteristics of the study population are presented in Table 1. The mean birthweight (826 infants) was 3394 ± 573 g and the mean gestational age at birth was 39.4 ± 2.0 weeks. Of these infants, 55 (6.7%) were classified as small for gestational age, 31 of whom were female and 24 male; 44 (80%) of the infants who were small for gestational age were born at term (≥ 37 gestational weeks) and 26 (47.3%) had a birth-weight below 2500 g. The absolute small for gestational age birthweights and gestational age birthweights varied between 785 to 2960 g and 30 to 42 weeks with a median of 2420 g and 39 weeks, respectively.

Table 1.  Background and life-style characteristics of the 826 women in their first pregnancy.
Background and life-style characteristicsn(%)
Maternal age (years) 
 15–1919 (2.3)
 20–29581 (70.3)
 30–44226 (27.4)
Maternal weight (kg) 
 < 5077 (9.6)
 50–75674 (83.8)
> 7553 (6.6)
Missing22
Maternal height (cm) 
 > 157735 (91.1)
 ≤ 15772 (8.9)
 Missing19
Country of origin 
 Sweden637 (77.5)
 Other185 (22.5)
 Missing4
Educational level (years) 
 > 12315 (38.8)
 ≤ 12496 (61.2)
 Missing15
Cohabiting status 
 Cohabiting716 (88.0)
Not cohabiting98 (12.0)
Missing12
Maternal smoking 
 Nonsmokers574 (69.5)
 1–9 cigarettes103 (12.5)
 10–19 cigarettes109 (13.2)
 ≥20 cigarettes40 (4.8)
TOTAL826 (100.0)

The unadjusted mean birthweight of infants to nonsmoking mothers (n= 574; 69.5%) was 3429 g (557 g). For mothers smoking 1–9 cigarettes per day (n= 103; 12.5%), it was 3353 g (627 g). For mothers smoking 10–19 cigarettes per day (n= 109; 13.2%) unadjusted mean birthweight was 3216 g (603 g). For women smoking 20 cigarettes per day and more (n= 40; 4.8%) it was 3471 g (469 g).

Univariate analysis

The associations between the various background variables and SGA are presented in Table 2. The crude odds ratios for SGA deliveries were increased and reached borderline statistical significance for women with short height (≤ 157 cm) [OR 2.1 (95% CI 1.0–4.6)], and for women of foreign origin [OR 1.7 (95% CI 1.0–3.1)]. Mean prepregnancy weight for women who were delivered of SGA infants was 58.4 kg (9.1 kg), compared with 61.0 kg (9.8 kg) for women who gave birth to infants with normal birthweight (P= 0.056). Mean height for women with SGA babies and women with non-SGA babies was 163.7 cm (7.1 cm) and 166.6 cm (6.6 cm), respectively (P= 0.002).

Table 2.  Associations between background and life-style characteristics and the odds for giving birth to small for gestational age (SGA) infants, presented as crude odds ratios (OR) with 95% confidence intervals (CI).
Background and life-style characteristicsSGA/AllCrude OR (95% CI)
Maternal age (years)  
 15–191/190.8 (0.1–6.3)
 20–2937/5811.0
 30–4417/2261.2 (0.7–2.2)
Maternal weight (kg)  
 < 508/771.6 (0.7–3.6)
 50–7545/6741.0
 > 752/530.5 (0.1–2.3)
 Missing0/22 
Maternal height (cm)  
 > 15746/7351.0
 ≤ 1579/722.1 (1.0–4.6)
 Missing0/19 
Country of origin  
 Sweden37/ 6371.0
 Other18/1851.7 (1.0–3.1)
 Missing0/4 
Educational level (years)  
 > 1215/3151.0
 ≤ 1238/4961.7 (0.9–3.1)
 Missing2/13 
Cohabiting status  
 Cohabiting46/7161.0
 Not cohabiting7/981.1 (0.5–2.6)
 Missing2/10 
Maternal smoking  
 Nonsmokers31/5741.0
 1–9 cigarettes9/1031.7 (0.8–3.6)
 10–19 cigarettes13/1092.4 (1.2–4.7)
 > 20 cigarettes2/400.9 (0.2–4.0)
TOTAL55/826 

Women who smoked 1–9 cigarettes per day had an increased odds ratio for SGA [OR 1.7 (95% CI 0.8–3.6)], and women smoking 10–19 cigarettes per day and women smoking ≥ 20 cigarettes per day [OR 2.4 (95% CI 1.2–4.7)] and [OR 0.9 (95% CI 0.2–4.0)], respectively, compared with nonsmokers. Persistent alcohol consumption and a high physical exercise rating were not significantly associated with a SGA risk [OR 1.4 (95% CI 0.8–2.6)] and [OR 1.2 (95% CI 0.6–2.2)], respectively.

Table 3 shows the association between the psychosocial variables and the risk of SGA. All odds ratios for giving birth to a small for gestational age baby are in the expected direction except for the variable control in daily life (i.e. women with low psychosocial resources have a higher odds of giving birth to a small for gestational age baby). Having a poor social network (i.e. poor social stability and poor social participation resulted in statistically significant increased odds ratios in the univariate analysis [OR 2.0 (95% CI 1.1–3.5)] and [OR 2.2 (95% CI 1.2–3.8)], respectively. Having poor emotional support reached borderline significance [OR 1.7 (95% CI 1.0–3.0)].

Table 3.  Associations between psychosocial characteristics and the odds for giving birth to small for gestational age (SGA) infants, presented as crude odds ratios (OR) with 95% confidence intervals (CI).
Psychosocial characteristicsSGA/A11Crude OR (95% CI)
Social stability  
 High31/5671.0
 Low21/2072.0 (1.1–3.5)
 Missing3/49 
Social participation  
 High22/4811.0
 Low31/3272.2 (1.2–3.8)
 Missing2/16 
Instrumental support  
 High42/7041.0
 Low12/1141.9 (0.9–3.6)
 Missing1/7 
Emotional support  
 High32/5681.0
 Low23/2481.7 (1.0–3.0)
 Missing0/10 
Support from the child's father  
 High33/5931.0
 Low9/1171.4 (0.7–3.0)
 Missing13/103 
Maternal support  
 High36/5711.0
 Low16/2351.1 (0.6–2.0)
 Missing3/17 
Job support  
 High18/3281.0
 Low25/3701.2 (0.7–2.3)
 Missing12/116 
Control of daily life  
 High24/5021.0
 Low19/3040.9 (0.5–1.6)
 Missing2/18 
TOTAL55/826 

Multivariate analysis

The results of the multivariate analyses are presented in Tables 4–7. Since we assume that the lifestyle factors (e.g. physical exercise, alcohol consumption and smoking) should be considered as intermediate factors between the psychosocial resources and SGA (Fig. 1), the multivariate models were run with these lifestyle factors absent (Model I and Model II) as well as present (Model III).

Table 4.  Logistic regression analysis of the association between social stability and the odds for giving birth to small for gestational age (SGA) infants, presented as crude odds ratio (OR) or *adjusted OR with 95% confidence intervals (CI).
 Crude ORModel I*Model II*Model III*
Social stability (low vs high)2.0 (1.1–3.5)2.0 (1.1–3.6)1.7 (0.8–3.3)1.7 (0.9–3.3)
Maternal age (per year) 1.0 (1.0–1.1)1.1 (1.0–1.1)1.1 (1.0–1.2)
Maternal height (≤ 157 cm vs others) 1.7 (0.7–4.1)1.7 (0.7–4.3)1.8 (0.7–4.7)
Maternal weight (≤ 50 kg vs others) 1.6 (0.7–3.4)1.2 (0.5–2.8)1.3 (0.6–3.2)
Country of origin (others vs Sweden)  1.5 (0.8–2.8)1.8 (0.8–4.0)
Educational level (≤ 12 years vs others)  1.7 (0.8–3.3)1.7 (0.8–3.7)
Physical exercise (high vs low/moderate)   2.0 (0.9–4.2)
Alcohol consumption (consumers vs nonconsumers)   1.6 (0.8–3.1)
Smoking (smokers vs nonsmokers (included as dummy variables)    
  1–9 cigarettes   2.1 (0.9–5.1)
  10–19 cigarettes   2.7 (1.2–6.1)
≥ 20 cigarettes   0.6 (0.1–4.3)
Table 5.  Logistic regression analysis of the association between social participation and the odds for giving birth to small for gestational age (SGA) infants, presented as crude odds ratio (OR) or *adjusted OR with 95% confidence intervals (CI).
 Crude ORModel I*Model II*Model III*
Social participation (low vs high)2.2 (1.2–3.8)2.2 (1.3–4.0)1.9 (1.02–3.5)2.2 (1.1–4.4)
Maternal age (per year) 1.0 (1.0–1.1)1.1 (1.0–1.1)1.1 (1.0–1.1)
Maternal height (≤ 157 cm vs others) 1.5 (0.6–3.6)1.6 (0.7–4.0)1.8 (0.7–4.6)
Maternal weight (≤ 50 kg vs others) 1.5 (0.7–3.1)1.3 (0.6–2.8)1.3 (0.6–3.1)
Country of origin (others vs Sweden)  1.1 (0.6–2.2)1.6 (0.7–3.5)
Educational level (≤ 12 years vs others)  1.5 (0.7–3.0)1.6 (0.7–3.4)
Physical exercise (high vs low/moderate)   2.5 (1.2–5.2)
Alcohol consumption   1.9 (1.0–3.8)
  (consumers vs nonconsumers)    
Smoking (smokers vs nonsmokers    
  (included as dummy variables))    
  1–9 cigarettes   1.7 (0.7–4.2)
  10–19 cigarettes   2.8 (1.3–6.1)
  ≥ 20 cigarettes   0.4 (0.1–3.3)
Table 6.  Logistic regression analysis of the association between instrumental support and the odds for giving birth to small for gestational age (SGA) infants, presented as crude odds ratio (OR) or *adjusted OR with 95% confidence intervals (CI)
 Crude ORModel I*Model II*ModeIII*
Instrumental support (low vs high)1.9 (0.9–3.6)1.9 (0.9–3.8)1.9 (0.9–4.0)2.6 (1.2–5.7)
Maternal age (per year) 1.0 (1.0–1.1)1.1 (1.0–1.1)1.1 (1.0–1.1)
Maternal height (< 157 cm vs others) 1.7 (0.7–3.9)1.8 (0.8–4.2)2.1 (0.8–5.2)
Maternal weight (< 50 kg vs others) 1.6 (0.8–3.4)1.3 (0.6–2.8)1.4 (0.6–3.2)
Country of origin (others vs Sweden)  1.2 (0.6–2.3)1.5 (0.7–3.3)
Educational level (< 12 years vs others)  1.8 (0.9–3.6)1.9 (0.9–4.0)
Physical exercise (high vs low/moderate)   2.4 (1.2–4.9)
Alcohol consumption (consumers vs nonconsumers)   1.7 (0.9–3.3)
Smoking (smokers vs nonsmokers (included as dummy variables)    
1–9 cigarettes   1.9 (0.8–4.6)
10–19 cigarettes   3.2 (1.5–6.8)
≥ 20 cigarettes   0.5 (0.1–3.9)
Table 7.  Logistic regression analysis of the association between emotional support and the odds for giving birth to small for gestational age (SGA) infants, presented as crude odds ratio (OR) or *adjusted OR with 95% confidence intervals (CI).
 Crude ORModel I*Model 11*Model III*
Emotional support (low vs high)1.7 (1.0–3.0)1.7 (1.0–3.0)1.4 (0.8–2.6)1.5 (0.8–2.8)
Maternal age (per year) 1.0 (1.0–1.1)1.1 (1.0–1.1)1.1 (1.0–1.1)
Maternal height(≤ 157 cm vs others) 1.7 (0.7–3.9)1.8 (0.8–4.2)2.1 (0.8–5.1)
Maternal weight(≤ 50 kg vs others) 1.6 (0.7–3.3)1.3 (0.6–2.8)1.3 (0.6–3.1)
Country of origin(others vs Sweden)  1.3 (0.7–2.5)1.7 (0.8–3.7)
Educational level(≤ 12 years vs others)  1.7 (0.9–3.4)1.8 (0.8–3.7)
Physical exercise (high vs low/moderate)   2.2 (1.1–4.5)
Alcohol consumption (consumers vs nonconsumers)   1.6 (0.8–3.1)
Smoking (smokers vs nonsmokers (included as dummy variables))    
 1–9 cigarettes    2.0 (0.8–4.7)
 10–19 cigarettes   3.2 (1.5–6.9)
 ≥ 20 cigarettes   0.5 (0.140)

The SGA odds ratio for women with low social anchorage did not change in Model I, but decreased somewhat and became statistically nonsignificant when adjustment was made for the background variables, country of origin and education in Model II and the lifestyle factors in Model III [OR 1.7 (95% CI 0.9–3.3)] (Table 4). Poor social participation remained as a statistically significant risk factor associated with SGA risk in all models (Model I, II and III) and the odds ratio varied around 2.0 (Table 5). Inadequate instrumental support showed an increased but nonsignificant association with SGA (OR 1.9; 95% CI 0.9–4.0) in Model I and II. When lifestyle factors were added in Model III the odds ratio increased and became statistically significant [OR 2.6 (95% CI 1.2–5.7)] (Table 6). Finally, the odds ratio for exposure to poor emotional support did not change in Model I [OR 1.7 (95% CI 1.0–3.0)] and still showed a borderline significance. After adjustment for background and lifestyle factors the odds ratio decreased and became nonsignificant (Table 7).

Biomedical risk factors for SGA, such as hypertension, urinary tract infection, bleeding and/or preterm contractions during current pregnancy, were present in 118 women (14.3%). The univariate SGA odds ratio for having at least one of these risk factors was estimated to [OR 1.6 (95% CI 0.8–3.1)]. In order to evaluate the effect of pregnancy complications on the SGA births in the multivariate analyses, these factors were analysed with the other confounding variables in a fourth step. The odds estimates for the psychosocial variables did not change when applying biomedical risk factors in Model III.

Antenatal clinic attendance rates (mean/days) varied very little between women with different psychosocial characteristics. Women with low social participation had an increased odds ratio for an elective delivery [OR 1.6 (95% CI 1.0–2.5)] (Table 8).

Table 8.  Associations between psychosocial characteristics and antenatal clinic attendance rate and the odds of an elective delivery, presented as crude odds ratios (OR) with 95% confidence intervals (CI).
Psychosocial characteristicsSGNAlI (n)Antenatal clinic attendance rate (days)(mean)Elective delivery (Crude OR (95% CT))
Social stability   
  High31/56711.31.0
  Low21/20711.40.9 (0.5–1.5)
  Missing3/49  
Social participation   
  High22/48111.41.0
  Low31/32711.21.6 (1.0–2.5)
  Missing2/16  
Instrumental support   
  High42/70411.41.0
  Low12/11410.70.7 (0.3–1.4)
  Missing1/7  
Emotional support   
  High32/56811.41.0
  Low23/24811.20.9 (0.5–1.5)
  Missing0/10  
TOTAL55/826  

The additive effect of being exposed to both poor social stability and poor social participation (a low total network index), in contrast to being exposed to none of them, showed a significantly-increased odds ratio for SGA births in Model I (adjustment for age, height and smoking) of OR 3–3 (95% CI 1.6–6.7). Additionally, low total support index, defined as low access to both instrumental and emotional support showed an adjusted SGA risk of OR 2.7 (95% CI 1.3–5.6).

Figure 2 shows the interaction between country of origin or educational level and psychosocial resources (total network index and total support index, respectively). Being born in another country than Sweden and having a low total network increased the SGA odds significantly [OR 5.7 (95% CI 2.5–13.1)], compared with women of Swedish origin with a high network index. Similarly, foreign born women with low total support also were at higher odds for bearing an SGA child [OR 4.5 (95% CI 1.9–10.7)], these estimates were adjusted for age, height and maternal smoking. The magnitude of the synergy index (SI) indicated increased susceptibility for giving birth to SGA infants for foreign born women with a poor social network and poor social support: SI = 9.0 and SI = 6.8, respectively. The interaction was absent for low educational level and low total network, (SI = 1.9) and low educational level and low total support (SI = 0.5).

image

Figure 2. Odds ratios of giving birth to small for gestational age (SGA) infants from interactions between country of birth or educational level and pyschosocial resources (total network index and total support index, respectively). Adjusted for age, height and smoking (95% CI). SI = synergy index.

Download figure to PowerPoint

Only 6% (n= 3) of the women with foreign origin with a low total network score smoked ≥ 10 cigarettes per day, and the same figures applied to women with foreign background with a low total support score. Among women with a low educational level and a low total network score the smoking prevalence was 25.5% (n= 24) and among women with low education an a low total support score it was 16.7% (n= 11).

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. References
  9. Appendix

The main findings of this study support the stress hypothesis since psychosocial factors seem to influence intrauterine growth and increase the risk of giving birth to SGA infants independent of background, lifestyle and biological risk factors. There were also indications of an interaction between foreign origin and low access to psychosocial resources, so that these factors effected the risk for SGA in a synergistic manner.

The present results could be biased by selection, misclassification, or confounding. This study was designed as a prospective study, and 85% to 90% of all mothers having their first babies in Malmö were invited to participate. The 10% to 15% of women not approached were those attending two private clinics. There was good reason to believe that almost all foreign-born women were approached, since women who see private obstetricians are usually of Swedish origin. The participation rate in the study was high (87–7%). An analysis of the nonresponders revealed some minor differences regarding age and country of origin but not regarding pregnancy outcomes, such as birthweight and gestational age34. Therefore, there is no reason to believe in a selection bias to any important degree. Anyhow, the high response rate should minimise this problem.

The SGA variable was based on an estimation of the birthweight in relation to the gestational age, which is a valid measurement of pregnancy outcome because it is not likely to be influenced by medical decision making21. Most of the women (97.6%) in this study were dated according to ultrasound examination twice during pregnancy. This method is regarded as the most valid one in measuring gestational age23. A newborn baby was classified as small for gestational age if its birthweight was > 2 SD below the mean weight for gestational age. The SGA classification was based on gender-specific growth curves constructed for use in perinatological practice and research24. Older standard intrauterine growth curves did not reflect fetal growth continuing in the uterus and thus under-estimated fetal growth in the preterm period24. In this study population of nulliparous women, 6.7% of the infants were classified as small for gestational age, compared with 6.3% in a Danish study6 and 10.5% in a Scandinavian multi-centre study using the 10th centile as cutoff2. Per definition, one can expect 2.5% of the infants in a normal population to be classified as small for gestational age. Our somewhat higher prevalence could be explained by the fact that SGA deliveries are somewhat more common in first pregnancies1 and that our growth curve classified more babies as small for gestational age in the preterm period.

The instrument to measure social network and social support has in various studies been shown to reveal association to cardiovascular risk factors, lifestyle factors and to predict mortality15,17,28,34–37. An analysis of the construct validity showed that the different indices measured different aspects of the psychosocial environment28. The reliability (test-retest stability) of the instrument has also been found to be sufficient15,17,28 likewise the internal consistency15,17,28. However, some modifications of the instrument have been made since specific support from the child's father and from the pregnant woman's own mother have been found to be of importance during pregnancy38.

Another potential validity problem is confounding. In this study, adjustments were made for maternal age, height, weight, country of origin, educational level, physical exercise, alcohol consumption, smoking and pregnancy complications in the logistic regression analyses. In most previous studies adjustments for maternal characteristics and smoking have been carried out3–9. We found small changes in our SGA estimates when maternal age, height and weight were taken into account (Model I). Moreover, controlling for country of origin and educational level resulted in an only modest decrease in the odds of having a baby that is small for gestational age. Maternal smoking of 10–19 cigarettes per day was statistically significantly associated with SGA deliveries; however, inclusion of a graded smoking variable in the multivariate analysis did not change the increased SGA risk associated to low psychosocial resources.

Women with a low degree of participation in society and those lacking instrumental support shared the greatest risk of giving birth to small for gestational age babies in our study. We also noted that women with a low social participation showed an increased odds ratio for elective delivery. We considered the risk of differential misclassification of SGA status (i.e. the pregnancies of women in underprivileged groups were examined in a different manner from other women, and thus the chance of detecting SGA status could be different). However, since there was no difference in the attendance rate of the antenatal clinic visits, we believe that this supports our hypothesis rather than being an artefact due to dependent misclassification. Surprisingly, poor support from the woman's own partner or mother did not show up as a significant risk factor for having a small for gestational age infant. Simultaneous exposure to poor social stability and poor social participation, as well as poor emotional support and poor instrumental support showed the highest odds ratios (3.3 and 2.7, respectively). Recent research concerning effects of psychosocial resources on fetal growth has been divergent3–9. Three studies which reported positive associations found that poor psychosocial factors and maternal stress affected growth, independent of maternal smoking4,8,9, while others found no effect from psychosocial factors when smoking was controlled for3,5,7.

In this study nearly a quarter (23%) of the women were born in countries other than Sweden. The immigrant women were a very heterogeneous group and originated from 52 different countries. Genetic difference in birthweight is one of several factors that have to be observed39. However, 50% of the Asian women (from Asia and the Pacific Islands) were classified as having a low height and a low prepregnancy weight, but none of these women gave birth to small for gestational age babies. Our analyses show an interactional effect of country of origin and poor social network or poor social support for the risk of giving birth to small for gestational age babies, since the estimated synergy index were increased. However, there was weak evidence for a synergistic relation in this respect between educational level and psychosocial resources.

The methods used to assess psychosocial influences on birthweight differ considerably in this study compared with many others3–9. While we measured ‘stress’ as lack of psychosocial resources using a questionnaire with 34 different questions, many studies assessed stressful life events only4,5,7,8. Our study consisted of a representative random population of women in their first pregnancy. Some other studies might have had difficulties in finding an association between social support and risk of SGA since they used selected populations in which the variation was too small in the independent variable3. Another reason for the inconsistencies might be that most studies in this area include both nulliparous and multiparous women. It is during the first pregnancy that the demands for adaptation are greatest and perhaps the variation in the level of stress is biggest.

Possible pathways for the influence of psychosocial stressors on fetal growth can be direct or indirect. An indirect effect on fetal growth may be mediated through adverse health habits, such as smoking and alcohol consumption. Stressors may also act more directly through the release of stress hormones, such as catecholamines, cortisol, neuropeptides or by altering the immune system21. The stress hormones can decrease the blood flow to the uterus, giving rise to hypotension and fetal bradycardia. However, another plausible mechanism for the relation between psychosocial factors and low birth-weight has been propounded by Picone et al.40 They stated that an increased secretion of catecholamines, corticoids, growth hormone and prolactin may effect maternal weight gain by reducing the utilisation of nutrients and thereby affecting the birthweight40.

CONCLUSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. References
  9. Appendix

The findings supported the stress model that the lack of psychosocial resources influence the risk of delivering SGA infants. There is also reason to believe that the effect of an inadequate social network and weak social support on intrauterine growth may be greater among women who are already subjected to some sort of social deprivation (i.e. immigrant women and women with little education).

The identification of SGA is important to antenatal health care, not only because it is of importance for children's health but it may also influence adult health41. Pregnant women with insufficient psychosocial resources are a high risk group that should be identified early in pregnancy and offered specific intervention programmes with individualised support from mid-wives. In the Nordic countries, compliance with the antenatal routine programme is extremely high. In Sweden women have a very high mean number of antenatal visits, even among very healthy women. By redistributing antenatal health care resources, it ought to be possible to develop a more effective antenatal care in order to decrease the number of SGA births.

Acknowledgements

The authors would like to thank all the pregnant women who took part in the study, and the midwives and the medical staff for their help in recruitment. This study was financed by a grant from the Swedish Council for Social Research, the Swedish Council for Planning and Coordination of Research, the National Institute of Public Health, the Vådal Foundation, the Medical Faculty at Lund University and the Malmö University Hospital.

References

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. References
  9. Appendix
  • 1
    Wen WS, Goldenberg RL, Cutter GR, Hoffman HJ, Cliver SP. Intrauterine growth retardation and preterm delivery: prenatal risk factors in an indigent population. Am J Obstet Gynecol 1990; 162: 213218.
  • 2
    Bakketeig LS, Jacobsen G, Hoffman HJ et al. Pre-pregnancy risk factors of small-for-gestational-age births among parous women in Scandinavia. Acta Obstet Gynecol Scand 1993; 72: 273279.
  • 3
    Aarts MCG, Vingerhoets AJJM. Psychosocial factors and intrauterine fetal growth: a prospective study. J Psychosom Obstet Gynaecol 1993; 14: 249258.
  • 4
    Pagel MD, Smilkstein G, Regen H, Montano D. Psychosocial influences on new born outcomes: a controlled prospective study. Soc Sci Med 1990; 30: 597604.
  • 5
    Jacobsen G, Schei B, Hoffman J. Psychosocial factors and small-for-gestational-age infants among parous Scandinavian women. Psychosocial factors and small-for-gestational-age infants among parous Scandinavian women 1997; 76 (Suppl 165): 1418.
  • 6
    Nordentoft M, Lou HC, Hansen D et al. Intrauterine growth retardation and premature delivery: The influence of maternal smoking and psychosocial factors. Am J Public Health 1996; 86: 347354.
  • 7
    Brooke O, Anderson R, Bland M, Peacock J, Stewart M. Effects on birth weight of smoking, alcohol, caffeine, socioeconomic factors and psychosocial stress. BMJ 1989; 298: 795801.
  • 8
    Mutale T, Creed F, Maresh M, Hunt L. Life events and low birth weight analysis by infants preterm and small for gestational age. Br J Obstet Gynaecol 1991; 98: 166172.
  • 9
    Cliver SP, Goldenberg RL, Cutter GR et al. The relationships among psychosocial profile, maternal size and smoking in predicting fetal growth retardation. Obstet Gynecol 1992; 80: 262267.
  • 10
    House JS, Landis KR, Umberson D. Social relationship and health. Science 1988; 241: 540544.
  • 11
    Selye H. The general adaptation syndrome and the disease of adaptation. J Clin Endocrinol 1946; 6: 112230.
  • 12
    Holmes TH, Rahe RH. The social readjustment rating scale. J Psychosom Res 1967; 11: 213218.
  • 13
    Nuckolls KB, Cassel J, Kaplan BH. Psychosocial assets, life crisis and the prognosis of pregnancy. Am J Epidemiol 1972; 95: 431441.
  • 14
    Cassel J. The contribution of the social environment to host resistance. Am J Epidemiol 1976; 104: 107123.
  • 15
    Hanson BS. Social Network, Social Support and Health in Elderly Men. A Population Study. [MD Thesis] Malmö: University of Lund, 1988.
  • 16
    Syme L. Control and health: a personal perspective. In: SteptoeA, AppelsA, editors. Stress, Personal Control and Health. Chichester : John Wiley & Sons, 1989: 318.
  • 17
    Ouml;stergren P-O. Psychosocial Resources and Health. With Special Reference to Social Network, Social Support and Cardiovascular Disease. [MD Thesis] Malmö: University of Lund, 1991.
  • 18
    Hanson BS, Ostergren P-O. Social networks and health: theories and empirical evidence from Malmö, Sweden. In: Lykke MortensenE, EgsgaardJ, editors. Social Networks and Health. Copenhagen : Hans Jörgensens Bogtrykkeri/Offset, 1995: 5282.
  • 19
    Mercer RT. A theoretical framework for studying factors that impact on the maternal role. Nurs Res 1981; 30: 7377.
  • 20
    Norbeck JS, Tilden VP. Life stress, social support, and emotional disequilibrium in complications of pregnancy: a prospective, multivariate study. J Health Soc Behav 1983; 24: 3046.
  • 21
    Paarlberg KM, Vingerhoets AJJM, Passchier J, Dekker GA, van Geijn HP. Psychosocial factors and pregnancy outcome: a review with emphasis on methodological issues. J Psychosom Res 1995; 39: 563595.
  • 22
    Rothman KJ. Modern Epidemiology. Boston : Little, Brown & Co, 1986.
  • 23
    Persson P-H, Weldner B-M. Reliability of ultrasound fetometry in estimating gestational age in the second trimester. Acta Obstet Gynecol Scand 1986; 65: 481483.
  • 24
    Marsal K, Persson P-H, Larsen T, Lilja H, Selbing A, Sultan B. Intrauterine growth curves based on ultrasonically estimated foetal weights. Acta Paediatr 1996; 85: 843848.
  • 25
    Bergsjö P, Hoffman HJ, Davies RO et al. Preliminary results from the collaborative Alabama and Scandinavian study of successive small for gestational age births. Acta Obstet Gynecol Scand 1989; 68: 1925.
  • 26
    Grimby G, Wilhelmsen L, Björnstorp P, Saltin B, Tibblin G. Habitual physical activity. aerobic power and blood lipids. In: PernowB, SaltinB, editors. Muscle Metabolism During Exercise. New York : Plenum Press, 1971: 469481.
  • 27
    Wennergren M, Karlsson K, Olsson T. A scoring system for antenatal identification of fetal growth retardation. Br J Obstet Gynaecol 1982; 89: 520524.
  • 28
    Hanson BS, Östergren P-O. Different social network and social support characteristics, nervous problems and insomnia: theoretical and methodological aspects on some results from the population study ‘Men born in 1914’, Malmö, Sweden. Soc Sci Med 1987; 25: 849859.
  • 29
    Hanson BS, Östergren P-O, Elmstahl S, Isacsson S-O, Ranstam J. Reliability and validity assessments of measures of social networks, social support and control—results from the Malmö Shoulder and Neck study. Scand J Soc Med 1997; 25: 249257.
  • 30
    Johnson JV. The Impact of Workplace Social Support, Job Demands and Work Control upon Cardiovascular Disease in Sweden. [PhD Thesis] (Environmental and Organizational Psychology Monograph Series 1986; 1: 1–204) Stockholm : University of Stockholm, 1986.
  • 31
    Theorell T, Harms-Ringdahl K, Ahlberg-Hultén G, Westin B. Psychosocial job factors and symptoms from the locomotor system—a multicausal analysis. Scand J Rehabil Med 1991; 23: 165173.
  • 32
    Cohen S, Williamsson GM. Perceived stress in a probability sample of the United States. In: SpacapanS, OskampS, editors. The Social Psychology of Health. The Claremont symposium on applied social psychology. Newbury Park : Sage Publications, 1988.
  • 33
    Norusis MJ. The SPSS Guide to Data Analysis for Release 4. Chicago , Ill : SPSS Inc. 1990.
  • 34
    Dejin-Karlsson E, Hanson BS, Ostergren P-O, Ranstam J, Isacsson S-O, Sjöberg N-O. Psychosocial resources and persistent smoking in early pregnancy—a population study of women in their first pregnancy in Sweden. J Epidemiol Community Health 1996; 50: 3339.
  • 35
    Dejin-Karlsson E, Hanson BS, Östergren P-O. Psychosocial resources and persistent alcohol consumption in early pregnancy—a population study of women in their first pregnancy in Sweden. Scand J Soc Med 1997; 25: 280288.
  • 36
    Hanson BS. Social network, social support and heavy drinking in elderly men—a population study of men born in 1914, Malmö, Sweden. Addiction 1994; 89: 725732.
  • 37
    Hanson BS, Isacsson SO, Janzon L. Social support and quitting smoking for good. Is there an association? Results from the population study, ‘Men born in 1914’, Malmö, Sweden. Addict Behav 1990; 15: 221233.
  • 38
    Norbeck JS, Anderson JN. Life stress, social support and anxiety in mid- and late pregnancy among low income women. Res Nurs Health 1989; 12: 281287.
  • 39
    Gardosi J. Ethnic differences in fetal growth. Ultrasound Obstet. Gynecol 1995; 6: 7374.
  • 40
    Picone TA, Allen LH, Schramm MM, Olsen PN. Pregnancy outcome in North American women. 1. Effects of diet, cigarette smoking, and psychological stress on maternal weight gain. Am J Clin Nutr 1982; 36: 12051213.
  • 41
    Barker DJP. Mothers, babies and health in later life. Glasgow : Bell and Bain Ltd, 1998.

Appendix

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. References
  9. Appendix

Appendix 1.

  • Social stability (4 items)

  • 1
    Do you feel a strong affinity to your relatives (apart from spouse, cohabitee)?
  • 2
    If you belong to an association, are you able to say that you feel a strong affinity to this association and its members?
  • 3
    Are you rooted in and feel a strong affinity to your residential area?
  • 4
    Do you belong to a group of friends who have something in common or do something together such as play cards, listen to music, make excursions?
  • Social participation (13 items)

  • Which of the following activities have you taken part in during the last 12 months?

  • 5
    Participated in a study circle/course at work?
  • 6
    Participated in a study circle/course in your spare time?
  • 7
    Participated in a union meeting?
  • 8
    Participated in a meeting of an association?
  • 9
    Visited the theatre/cinema?
  • 10
    Visited an art exhibition?
  • 11
    Been to church?
  • 12
    Been to a public sports event?
  • 13
    Written a letter to a newspaper/magazine?
  • 14
    Taken part in a demonstration of some sort?
  • 15
    Visited a place of public entertainment (e.g. night club, dance, etc.)?
  • 16
    Taken part in a large family gathering?
  • 17
    Been to a party at someone's home?
  • Emotional support (3 items)

  • 18
    Do you feel you have someone (or several people) who can give you real personal support in order to cope with stresses and strains of life?
  • 19
    Do you have anyone whom you feel you can be totally yourself in front of, someone who accepts you with all your virtues and shortcomings?
  • 20
    How many people do you think you know well, and with whom you can talk about almost anything (relatives, friends, neighbours, workmates)?
  • Instrumental support (1 item)

  • 21
    If you were ill or needed assistance with some practical problem (e.g. needed to borrow an appliance, needed help with repairs or to write a letter, needed advice or information) is there anyone (privately) who could help you?
  • Support from the child's father (2 items)

  • 22
    Do you feel you have support from your baby's father?
  • 23
    Do you plan to attend childbirth classes during pregnancy? If yes, do you expect accompany from your baby's father?
  • Maternal support (1 item)

  • 1
    Do you feel you have support from your mother?
  • Job support (6 items)

  • Agree or disagree in following statements.

  • 25
    There is a peaceful and pleasant atmosphere at my job.
  • 26
    Everyone is helpful and cooperative at work.
  • 27
    The people I work with support me.
  • 28
    My colleagues understand if I am having a bad day.
  • 29
    I get on well with my supervisors.
  • 30
    I get on well with people I work with.
  • Control of daily life (4 items)

  • 31
    How often during the past month have you felt unable to exercise control over important matters in your life?
  • 32
    How often during the past month have you felt uncertain of your ability to handle your personal problems?
  • 33
    How often during the past month have you felt that things have not turned out as you had wished?
  • 34
    How often during the past month have you felt that your problems have been more than you have been able to cope with?