- Email: [email protected]
Large social disparities in spontaneous preterm birth rates in transitional Russia
Public Health (2005) 119, 77–86
Large social disparities in spontaneous preterm birth rates in transitional Russia ¨stro ¨ma A.M. Grjibovskia,b, L.O. Bygrena,c,*, A. Yngve, M. Sjo a
Unit for Preventive Nutrition, Department of Biosciences at Novum, Karolinska Institutet, 141 157 Huddinge, Sweden b Institute of Hygiene and Medical Ecology, Northern State Medical University, Arkhnagelsk, Russia c ˚ University, Umea ˚, Sweden Department of Community Medicine and Rehabilitation, Social Medicine, Umea Received 22 January 2004; received in revised form 21 May 2004; accepted 2 June 2004
KEYWORDS Preterm birth; Sociodemographic factors; Russia
Summary Objective. This study estimated the effect of maternal sociodemographic, obstetric and lifestyle factors on the risk of spontaneous preterm birth in a Russian town. Methods. All women with singleton pregnancies registered at prenatal care centres in Severodvinsk in 1999 comprised the cohort for this study (nZ1559). Analysis was based on spontaneous live singleton births at the maternity home (nZ1103). Multivariable logistic regression was applied to quantify the effect of the studied factors on the risk of preterm birth. Differences in gestation duration were studied using multiple linear regression. Results. In total, 5.6% of all spontaneous births were preterm. Increased risks of preterm delivery were found in women with lower levels of education and in students. Placental complications, stress and a history of fetal death in previous pregnancies were also associated with elevated risks for preterm delivery. Smoking, hypertension and multigravidity were associated with reduced length of pregnancy in metric form. Conclusion. In addition to medical risk factors, social factors are important determinants of preterm birth in transitional Russia. Large disparities in preterm birth rates may reflect the level of inequalities in transitional Russia. Social variations in pregnancy outcomes should be monitored. Q 2004 Published by Elsevier Ltd on behalf of The Royal Institute of Public Health.
Introduction
* Corresponding author. Address: Unit for Preventive Nutrition, Department of Biosciences at Novum, Karolinska Institutet, 141 157 Huddinge, Sweden. Fax: C46-8-6083350. E-mail addresses: [email protected] (L.O. Bygren), [email protected] (A.M. Grjibovski).
Preterm birth is a leading cause of infant mortality in industrialized societies.1 It is associated with elevated risks for perinatal mortality, serious neonatal morbidity, moderate to severe childhood disability, and high healthcare expenditures.2 These risks make preterm birth a serious public health problem.
0033-3506/$ - see front matter Q 2004 Published by Elsevier Ltd on behalf of The Royal Institute of Public Health. doi:10.1016/j.puhe.2004.06.005
78
A.M. Grjibovski et al.
A large body of evidence suggests that adverse pregnancy outcomes are more frequent among socially disadvantaged women, although the mediating factors may vary. According to Kramer, the most important causes of preterm delivery are genital tract infections, multiple birth, pregnancyinduced hypertension, low pre-pregnancy body mass index, incompetent cervix, prior preterm birth, abruptio placentae, heavy work and cigarette smoking.3 However, despite a great number of studies, many preterm births remain unexplained. Moreover, a majority of preterm births occur in low-risk groups.4 Economic reforms introduced in Russia after the break up of the Soviet Union in 1991 led to impoverishment of the majority of the population, a considerable decline in overall life expectancy, and increased social and health inequalities. In our earlier publications, we showed a clear gradient between maternal educational level and birth weight in a Russian setting,5 and we found that poor housing, self-perceived stress and smoking are important determinants of fetal growth in transitional Russia.6 However, it is now recognized that the aetiological determinants of gestational duration are different from the determinants of fetal growth.7 Countries of the former Soviet Union provide scarce information on risk factors for preterm delivery. We found only two studies on the determinants of preterm birth in these countries. A Ukrainian study concluded that the problems associated with transition did not alter this pregnancy outcome when the classical risk factors for preterm delivery were present.8
Figure 1
An Estonian study showed an independent effect of maternal education, marital status and nationality on the risk of preterm birth, although clinical data were not included in the analysis.9 This study estimated the effects of maternal sociodemographic, obstetric and lifestyle factors on preterm delivery and gestational length in a community-based cohort in a Russian town.
Methods The study was performed in Severodvinsk, a town in north-west Russia with a population of 233,800 in 1998. The study included all the pregnant women in the town who were registered in prenatal care centres (nZ1559). The cohort was followed through delivery. Pregnant women who attended antenatal clinics for abortion counselling or who were not permanent residents in Severodvinsk were excluded. The sampling material is summarized in Fig. 1 and is described elsewhere.5,10 Spontaneous and induced preterm births are likely to have different risk factors.11 As researchers understand more about the causes of the latter that mainly relate to maternal complications and/or endangered fetal wellbeing,4 all the analyses were only performed in a subgroup of spontaneous births. Delivery was classified as preterm if it occurred before the 37th completed week of gestation, measured from the last menstrual period. Data on maternal education, occupation, marital status, pre-pregnancy weight, gestational length and
Sampling procedure. Severodvinsk cohort, 1999.
Spontaneous preterm birth rates in transitional Russia pregnancy outcomes were obtained from the medical records at the prenatal care centres and at the maternity home. Smoking habits, alcohol consumption, stress and housing conditions were studied by means of a questionnaire administered during the mother’s first antenatal visit. The process of creating variables from these data is described elsewhere.5,6,10 In addition, data on maternal reproductive history and complications of the index pregnancy were abstracted from the medical files at the maternity home (Table 3). Gravidity was set as the number of pregnancies to the mother including the index pregnancy. Parity was the number of previous births. Previous induced abortions and miscarriages were also recorded. Placental complications included placental abruption, placenta previa and antepartum haemorrhage. The hypertension group consisted of women with pre-pregnancy hypertension and pregnancy-induced hypertension. Women who had haemoglobin levels lower than 110 g/l at the time of admission to the maternity home were classified as anaemic. If a woman’s records contained data on genital and/or urinary tract infections, she was classified as having a genito-urinary infection. Bivariate relationships were studied using Chisquare tests and one-way ANOVA for dichotomous and continuous data, respectively. Independent effects of the studied factors on the risks of preterm delivery were assessed by multivariable logistic regression. Multiple linear regression was used to estimate independent effects of the studied factors on gestational length. The group believed to be the most favourable was chosen as the reference group. To make our findings comparable with the previous studies from the former Soviet republics, multivariable analyses were performed in two sets. In the first set, only sociodemographic data were introduced into the multivariable models. In the second set, lifestyle data and reproductive heath data were added to the analysis. Due to a relatively small sample size, backward stepwise procedures in both logistic and linear regressions were used on this stage to obtain the best fit for the models.12 All analyses were performed using SPSS Version 10.0. Permission to perform the study was obtained from the Department of Health Care and the Mayor of Severodvinsk.
79 spontaneous (Fig. 1). Gestational length varied from 26 to 43 weeks (mean 39.34 weeks, SD 1.91). Preterm births constituted 5.6% of all spontaneous births. Bivariate relationships between maternal characteristics and preterm deliveries and average length of gestation are presented in Tables 1–3. Maternal education was the only sociodemographic factor significantly associated with gestational length and spontaneous preterm birth in the sample in bivariate analysis. Initiation of prenatal care after the first trimester and placental complications were associated with both reduced length of pregnancy and increased risks for preterm birth in crude analysis. Gravidity, history of induced abortion, smoking, hypertension and stress were associated with less extreme differences in pregnancy duration. Women with a history of previous fetal antenatal death had a higher crude risk of preterm delivery, although they had a longer average duration of pregnancy. Adjustment for other sociodemographic factors in the first set of multivariable analyses increased the risk of preterm births in mothers with a low level of education. Mothers aged 30 years or more were also at increased risk of preterm delivery. Gestational length in metric form was also reduced in all groups listed above (Table 4). Skilled bluecollar occupation was significantly associated with reduced risk of preterm delivery. After incorporating the data on maternal reproductive history, pregnancy complications and lifestyle factors in the multivariable models, placental complications, maternal education, history of antenatal death, being a student and perceived stress were independently associated with increased risk of preterm birth. Skilled blue-collar workers again had significantly lower risks for preterm delivery. They also had a longer duration of pregnancy than the reference group. Average gestational length was reduced in women with vocational education, in women who reported stress, in women with placental complications, in women with three or more previous pregnancies, in women who smoked and in women with hypertension. Results of the second set of multivariable analyses are summarized in Table 5.
Discussion Results There were 1399 live singleton deliveries in the sample. Among them, 1103 (78.8%) were
The proportion of infants born before term has increased in many industrialized countries over the past decades. For example, preterm birth rates increased from 9.4% in 1981 to 11.9% in 2001 in
80
A.M. Grjibovski et al.
Table 1
Spontaneous preterm birth rates and duration of pregnancy by maternal sociodemographic characteristics. N
Maternal age (years) 15–19 20–29 30C Education Secondary or less Vocational Unknown Universitya Occupation Unemployedb Blue-collar Skilled bluecollar Students White-collar Marital status Unmarried Married Pre-pregnancy weight Underweight Normal Overweight a b
Preterm (%)
P (c2)
Crude OR (95% CI)
Gestational length (weeks) (SD)
0.119 122 801 180
8.2 4.7 7.8
314
0.186 1.79 (0.87, 3.70) Reference 1.69 (0.90, 3.20)
39.20 (1.97) 39.41 (1.82) 39.16 (2.19)
9.2
9.12 (2.75, 30.30)
39.17 (2.01)
281 236 272
6.4 5.1 1.1
6.14 (1.79, 21.08) 4.80 (1.34, 17.23) Reference
39.19 (2.24) 39.53 (1.78) 39.54 (1.44)
373 171 120
5.9 5.8 2.5
1.09 (0.58, 2.01) 1.07 (0.49, 2.33) 0.44 (0.13, 1.51)
39.33 (1.84) 39.37 (1.76) 39.67 (1.64)
76 363
9.2 5.5
1.74 (0.71, 4.27) Reference
39.16 (1.83) 39.27 (2.12)
394 719
7.3 4.7
1.59 (0.95, 2.66) Reference
39.21 (1.90) 39.41 (1.91)
0.024
!0.001
0.389
0.292
0.078
0.093
0.569 58 1000 45
8.6 5.5 4.4
P (ANOVA)
0.488 1.62 (0.62, 4.22) Reference 0.80 (0.19, 3.39)
39.05 (2.40) 39.36 (1.89) 39.33 (1.48)
At least 3 years of university studies. Including housewives.
the USA,13 and from 6.3% in 1981 to 6.8% in 1992 in Canada.14 Although a substantial part of this increase can be attributed to the increased number of multiple pregnancies, changes in patterns of obstetric interventions and registration of early-gestation births, 14 changes in maternal sociodemographic factors have also been important determinants of the observed trends.15 Countries with lower prevalence of socio-economic disadvantage have lower preterm birth rates.7 Improvements in socio-economic characteristics were responsible for the decrease in preterm birth rates in Finland (from 9.1% in 1966 to 4.8% in 1985–1986),16 whereas in France, similar changes (from 7.9% in 1972 to 4.1% in 1988–1989) were mainly attributed to primary prevention strategies.17 In this study, the rate of spontaneous live preterm births was 5.6%, which is slightly higher than rates found in Ukraine (4.5%)8 and Estonia (5.3%).9 The Estonian study was based on a birth register and included both medically induced and spontaneous live singleton preterm births. If we use
the Estonian approach, we get a rate of 6.1%, although calculating the preterm rate as in the Ukrainian study gives a rate of 4.5%. The differences in inclusion criteria affect prematurity rates and make international comparisons difficult. Nevertheless, the rates in all these former Soviet republics are likely to be higher than in the Nordic countries, but lower than in the USA.13,18 Validity of the estimates and the relatively small sample size may be potential limitations of the study. Odland et al.,19 however, reported sufficient quality of the Russian medical documentation for epidemiological studies on reproductive health issues. The validity of the factors estimated by means of our questionnaire is discussed elsewhere.6,10 The maternity home documentation of pregnancy complications, especially genito-urinary infections, seems to greatly underestimate the rate of occurrence (9.8% in our study and 52.8% in the Ukrainian study). Previous preterm births and weight gain pattern are known to be associated with preterm delivery; however, these factors were
Spontaneous preterm birth rates in transitional Russia
81
Table 2 Spontaneous preterm birth rates and duration of pregnancy by maternal lifestyle factors and housing conditions.
Smoking No Yes No information Alcohol drinking No Occasional No information Stress No Yes No information Type of housing Own apartment Shared apartment No information Crowded housing Non-crowded Crowded Unknown
n
Preterm (%)
674 136 293
5.3 9.6 4.4
P (c2)
Crude OR (95% CI)
Gestational length (weeks) (SD)
Reference 1.87 (0.97, 3.64) 0.82 (0.43, 1.58)
39.36 (1.94) 38.93 (2.30) 39.48 (1.57)
0.089
0.019
0.622 594 195 314
6.2 4.6 5.1
431 331 341
4.2 8.5 4.7
0.347 Reference 0.73 (0.35, 1.54) 0.81 (0.44, 1.48)
39.27 (2.05) 39.45 (1.91) 39.42 (1.59)
Reference 2.12 (1.15, 3.90) 1.13 (0.57, 2.25)
39.49 (1.81) 39.02 (2.28) 39.47 (1.56)
0.086
0.001
0.351
0.217
632
6.5
Reference
39.26 (2.10)
181
4.4
0.67 (0.31, 1.45)
39.39 (1.67)
290
4.5
0.68 (0.36, 1.28)
39.49 (1.58)
0.457 405 409 289
6.7 5.4 4.5
poorly recorded in the medical files and therefore we did not include them in the analysis. We emphasize that our findings are relevant largely to preterm births occurring between 32 and 36 completed weeks of gestation. Moreover, the duration of gestation was measured from the history of the last menstrual period, and not from the ultrasound measures; therefore, the results should be interpreted with caution. The relatively small sample size might result in non-significant results for potentially important factors for preterm births. To overcome this, multiple linear regression analyses, which provide greater statistical power, were also performed. Preterm deliveries that are preceded by the rupture of membranes may have different aetiology compared with those preceded by spontaneous labour. However, many researchers suggest that these two groups may not represent aetiologically different entities;4 therefore, we did not perform separate
P (ANOVA)
0.311 Reference 0.80 (0.45, 1.42) 0.66 (0.33, 1.30)
39.27 (2.12) 39.31 (1.90) 39.49 (1.58)
analyses for these groups in order not to reduce the statistical power of the study. Maternal education was most strongly associated with the risk of spontaneous preterm delivery. Multivariable analysis with only sociodemographic factors in the model strengthened the associations. Including reproductive history, obstetric and lifestyle data in the model slightly reduced the associations, although they remained the largest associations observed in similar studies to the best of our knowledge. Large differences in the risks of preterm birth by maternal education have been found in the USA,20 the UK,21 Estonia9 and the Czech Republic.22 Considerably smaller variations were reported in France23 and Sweden.22 The mechanisms that may explain social disparities in preterm birth are still unclear. Kramer et al. proposed two sets of causal pathways. The first pathway involves unhealthy behaviours, infections,
82
A.M. Grjibovski et al.
Table 3 Spontaneous preterm birth rates and duration of pregnancy by maternal reproductive history and course of the index pregnancy. n
Preterm (%)
Reproductive history Gravidity 1 460 2–3 428 4C 215
4.6 5.1 8.8
Parity 0 1C
4.7 7.5
Gestational length (weeks) (SD)
0.88 (0.48, 1.63) Reference 1.79 (0.95, 3.83)
39.48 (1.76) 39.38 (1.80) 38.97 (2.33)
0.61 (0.37, 1.03) Reference
39.42 (1.80) 39.19 (1.80)
5.1 5.3 7.4
Miscarriages 0 1C
966 137
5.6 5.8
0.071
0.431
0.031 Reference 1.04 (0.54, 2.02) 1.49 (0.80, 2.76)
39.45 (1.81) 39.32 (1.80) 39.06 (2.24)
Reference 1.05 (0.49, 2.25)
39.37 (1.90) 39.15 (1.92)
0.906
0.214
0.001 1079 24
Index pregnancy Prenatal care initiation During first 935 trimester After first 168 trimester Hypertension No 1063 Yes 40
5.3 20.8
0.037 Reference 4.72 (1.70, 13.09)
39.36 (1.86) 38.54 (3.36)
0.006 4.8 10.1
0.029 Reference
39.40 (1.82)
2.23 (1.24, 3.99)
39.05 (2.32)
Reference 2.52 (0.95, 6.68)
39.37 (1.88) 38.50 (2.28)
0.054 5.4 12.5
0.004
!0.001 1078 25
4.9 36.0
Infection No Yes
995 108
6.0 1.9
Anaemia No Yes
701 402
5.1 6.5
Infant sex Female Male
536 567
5.8 5.3
!0.001 Reference 10.9 (4.59, 25.76)
39.40 (1.82) 36.88 (3.38)
0.29 (0.07, 1.22)
39.31 (1.96) 39.63 (1.28)
Reference 1.28 (0.76, 2.15)
39.33 (1.87) 39.35 (1.93)
Reference 0.94 (0.56, 1.57)
39.37 (1.91) 39.32 (1.90)
0.073
0.100
0.335
0.824
0.820
exposure to stress, and physiological reactions to these factors that shorten gestation. The second pathway involves a gene–environmental interaction based on mutation in the gene for
P (ANOVA)
0.004
0.062 742 361
644 244 215
Placental complications No Yes
Crude OR (95% CI)
0.069
Induced abortions 0 1 2C
Antenatal death 0 1C
P (c2)
0.674
methylenetetrahydrofolate reductase (MTHFR); if combined with low folate intake, this leads to hyperhomocysteinaemia, decidual vasculopathy and subsequent preterm delivery.1 In this study,
Spontaneous preterm birth rates in transitional Russia
83
Table 4 Adjusted odds ratios (OR) for preterm births and adjusted differences in gestational length between selected maternal sociodemographic factors and reference groups. Adjusteda OR (95% CI)
Adjusteda gestational age difference (weeks) (95% CI)
Maternal age (years) 15–19 20–29 30C
0.73 (0.29, 1.88) Reference 2.22 (1.14, 4.34)
0.01 (K0.42, 0.44) Reference K0.32 (K0.63, K0.01)
Education Secondary or less Vocational Unknown University
15.50 (4.17, 57.70) 10.53 (2.95, 37.63) 7.35 (1.91, 28.27) Reference
K0.51 (K0.91, K0.11) K0.68 (K1.04, K0.31) K0.13 (K0.51, 0.26) Reference
Occupation Unemployed Blue-collar Skilled blue-collar Students White-collar
0.68 (0.33, 0.44 (0.17, 0.21 (0.06, 1.72 (0.54, Reference
1.39) 1.12) 0.74) 5.50)
0.12 (K0.21, 0.44) 0.30 (K0.13, 0.72) 0.83 (0.39, 1.28) K0.12 (K0.64, 0.38) Reference
Marital status Unmarried Married
1.31 (0.77, 2.24) Reference
K0.16 (K0.40, 0.08) Reference
Pre-pregnancy weight Underweight Normal Overweight
1.96 (0.72, 5.34) Reference 0.75 (0.17, 3.25)
K0.37 (K0.87, 0.14) Reference 0.02 (K0.5, 0.58)
a
Adjusted for the variables in the table.
we controlled for the first mentioned set of mediators as well as for maternal reproductive history. All these factors explain only a small part of the observed differences in preterm birth rates, but it is unlikely that the second pathway alone is responsible for all social variations. Low folate intake is known to be more prevalent among the socially disadvantaged. Economic crisis in Russia led to considerable increase in inequalities. Poor compliance with the medical recommendations regarding diet and vitamin supplementation during pregnancy and lack of resources to follow such advice among low-educated women may be a contributing factor. Although there are 11–13% homozygotes for the MTHFR mutation in the USA,24 the prevalence of this mutation in northwest Russia is unknown. Among other nutritional factors, fish consumption,25 fish oil supplements, zinc, iron and calcium-rich foods26 are known to be associated with a reduced number of preterm births. In transitional Russia, seafood, fruits, meat and dairy products are relatively expensive and may not be consumed by mothers with a low level of education. High costs were identified as the main barrier to healthy eating in Ukraine, and this may be true for Russia, given that both the economic
situation and food habits are similar in these countries.27 Based on experimental data in sheep, Bloomfield et al. suggested that even a modest periconceptional reduction in food intake might lead to accelerated fetal hypothalamic–pituitary–adrenal axis maturation, increased fetal adrenocorticotrophic hormone concentrations, and subsequent preterm delivery.28 If these findings are applicable to humans, family planning issues may become more important for the prevention of preterm delivery. Unplanned pregnancies were far more common among women with a low level of education (84% in women with basic education compared with 23% in women with university education in this cohort), which might also contribute to the large social disparities in preterm birth rates. It is widely accepted that heavy work during pregnancy increases the risk of preterm delivery.3 In our analysis, a group of skilled blue-collar workers had a lower risk of preterm birth and a greater average gestational length. One of the most probable explanations of this association is that most of these women work for state-owned industrial enterprises with compulsory regular
84
A.M. Grjibovski et al.
Table 5 Results of stepwise multivariable regression analyses with preterm delivery and gestational length as dependent variables. Adjusted OR for preterm birth
Secondary or less education Vocational education Unknown education Skilled bluecollar workers Students Previous antenatal death Placental complications Stress Gravidity of 4C Smoking Hypertension
95% CI
P
Adjusted difference in gestational length (weeks)
95% CI
P
K0.44
K0.73, K0.14
0.004
0.32, 1.15
0.001
8.52
2.53, 28.73
0.001
10.21
2.82, 36.94
!0.001
6.55
1.75, 24.51
0.005
0.19
0.05, 0.75
0.017
2.53 6.15
1.00, 6.42 1.85, 20.47
0.050 0.003
11.44
4.32, 30.34
!0.001
K2.36
K3.11, K1.62
!0.001
2.15
1.20, 3.83
0.009
K0.37 K0.35
K0.61, K0.13 K0.62, K0.07
0.003 0.014
K0.37 K0.59
K0.71, K0.04 K1.18, K0.00
0.030 0.049
0.73
Reference categories are as in Tables 1–3; only the variables significantly associated with the studied outcomes are presented in the table.
medical check-ups; these might result in better general health and lower prevalence of infection. Greater levels of social security for these women compared with women employed elsewhere may also explain this finding. Students (including secondary school pupils) had a 2.5 times higher risk of spontaneous preterm delivery than the reference group. High academic demands, walking or using busy city transport to get to the educational institutions may be associated with high levels of stress, which is known to be an important determinant of preterm birth.18 Increased secretion of placental corticotropin releasing hormone and subsequent oestrogen and prostaglandin production may be the mediating factors between maternal chronic exposure to stress and preterm delivery. It has been proposed that chronic stress may be one of the most important factors that may explain large social disparities in preterm birth rates.1 In our study, we found a two-fold increase in preterm birth rates in women who reported self-perceived stress at the time of prenatal care initiation. Our estimate of stress is imprecise, and this may be why adjustment
to stress only decreased the effect of maternal education on gestational length slightly. Advanced maternal age was associated with increased risk of preterm delivery and shorter average gestation, independent of sociodemographic factors, although these differences became insignificant after adjustment for other factors. Women with three or more previous pregnancies (but not births) had significantly shorter gestation periods. These women were more likely to have had previous abortions. Multiple induced abortions were associated with reduced gestational length in bivariate analysis. Induced abortions with dilatation and curettage are common in Russia with an average rate of two abortions per birth, although a slight downward trend has been observed.29 Urinary and genital tract infections are wellknown determinants of preterm birth.1,3,7,18 Surprisingly, our study did not reveal elevated risks of preterm delivery or shortened gestational length associated with infection. Moreover, in bivariate analysis, the prevalence of diagnosed infection was considerably lower in women who delivered
Spontaneous preterm birth rates in transitional Russia preterm, which may be explained by the fact that women who deliver preterm are often admitted to a maternity home in labour and are rarely tested for infections (AN Baranov, personal communication). Placental complications, such as placenta previa, abruptio placentae and antepartum haemorrhage, are well-established risk factors for preterm delivery, although they normally result in medically induced labour. Hypertension before pregnancy or pregnancy-induced hypertension influence gestational length.18 In the Ukrainian study, women with pre-existing hypertension were 2.3 times more likely to deliver preterm.8 In our study, the small number of cases made it impossible to find significant differences in preterm birth rates in relation to hypertension, although a 2.5-fold difference was found in bivariate analysis. The difference in gestational age in metric form was 0.6 weeks or 4.1 days on average, which indicates that hypertensive disorders may shorten the duration of pregnancy in the study setting. All these complications can result in medically induced deliveries, which may explain the low prevalence of these disorders in our sample, which only included those women who had spontaneous births. Anaemia has also been reported as a factor that influences the length of gestation; however, this was not the case in our study, which did not take the decrease of haemoglobin until late pregnancy into account. This may result in lower haemoglobin values in full-term births because of longer gestation. Antenatal care can prevent pregnancy complications by identifying risk factors and detecting complications at an early stage. In Russia, antenatal care is free and covers almost all pregnancies.5 Early initiation of care was associated with a lower risk of preterm birth and a longer gestation in bivariate analysis. Given the inaccurate registration of pregnancy complications in the medical files at the maternity home, we attempted to evaluate the importance of prenatal care by excluding all pregnancy complications from the multivariable models. After such transformation, late initiation of care became a risk factor associated with increased risk of preterm delivery (ORZ2.03, 95% CI 1.09–3.80). Smoking is often cited as a risk factor of preterm delivery.18 Risks associated with smoking in relation to preterm delivery are much lower than those in relation to fetal growth restriction,1 suggesting a modest role of smoking as an explanatory variable for sociodemographic disparities in preterm birth rates. We revealed a modest reduction in duration of pregnancy among smoking
85 mothers, although this effect may be underestimated because our group of smokers only included daily smokers.6 In summary, low maternal education, being a student, perceived stress, placental complications during the index pregnancy, and a history of fetal death in previous pregnancies were associated with increased risk of spontaneous preterm delivery in an urban Russian setting. Multigravidity, smoking and hypertensive disorders reduced the average length of gestation without increasing prematurity rates. We conclude that in addition to the traditional risk factors for preterm delivery, social factors are important determinants of pregnancy outcomes in transitional Russia, and the social disparities in pregnancy outcomes may be even more profound in large cities with a higher degree of social inequality.
References 1. Kramer MS, Goulet L, Lydon J, Seguin L, McNamara H, Dassa C, et al. Socio-economic disparities in preterm birth: causal pathways and mechanisms. Paediatr Perinat Epidemiol 2001;15(Suppl 2):104–23. 2. Lumley J. Defining the problem: the epidemiology of preterm birth. Br J Obstet Gynaecol 2003;110(Suppl 20): 3–7. 3. Kramer MS. The epidemiology of adverse pregnancy outcomes: an overview. J Nutr 2003;133(Suppl 2):1592–6. 4. Moutquin JM. Socio-economic and psychosocial factors in the management and prevention of preterm labour. Br J Obstet Gynaecol 2003;110(Suppl 20):56–60. 5. Grjibovski A, Bygren LO, Svartbo B, Magnus P. Social variations in foetal growth in a Russian setting: an analysis of medical records. Ann Epidemiol 2003;13:599–605. 6. Grjibovski A, Bygren LO, Svartbo B, Magnus P. Housing conditions, perceived stress, smoking and alcohol—determinants of foetal growth in Northwest Russia. Acta Obstet Gynecol Scand 2004 in press. 7. Kramer MS, Seguin L, Lydon J, Goulet L. Socio-economic disparities in pregnancy outcome: why do the poor fare so poorly? Paediatr Perinat Epidemiol 2000;14:194–210. 8. Monaghan SC, Little RE, Hulchiy O, Strassner H, Gladen BC. Risk factors for spontaneous preterm birth in two urban areas of Ukraine. Paediatr Perinat Epidemiol 2001;15: 123–30. 9. Koupilova I, Rahu K, Rahu M, Karro H, Leon DA. Social determinants of birth weight and length of gestation in Estonia during the transition to democracy. Int J Epidemiol 2000;29:118–24. 10. Grjibovski A, Bygren LO, Svartbo B. Socio-demographic determinants of poor infant outcome in north-west Russia. Paediatr Perinat Epidemiol 2002;16:255–62. 11. Meis PJ, Michielutte R, Peters TJ, Wells HB, Sands RE, Coles EC, et al. Factors associated with preterm birth in Cardiff, Wales. II. Indicated and spontaneous preterm birth. Am J Obstet Gynecol 1995;173:597–602.
86 12. Katz MH. Performing the analysis Multivariable analysis. A practical guide for clinicians. Cambridge: Cambridge University Press; 1999 p. 84–117. 13. Martin JA, Hamilton BE, Ventura SJ, Menacker F, Park MM, Sutton PD. Births: final data for 2001. Nat Vit Stat Rep 2002; 2:1–103. 14. Joseph KS, Kramer MS, Marcoux S, Ohlsson A, Wen SW, Allen A, et al. Determinants of preterm birth rates in Canada from 1981 through 1983 and from 1992 through 1994. N Engl J Med 1998;339:1434–9. 15. Kramer MS, Platt R, Yang H, Joseph KS, Wen SW, Morin L, et al. Secular trends in preterm birth: a hospital-based cohort study. JAMA 1998;280:1849–54. 16. Olsen P, Laara E, Rantakallio P, Jarvelin MR, Sarpola A, Hartikainen AL. Epidemiology of preterm delivery in two birth cohorts with an interval of 20 years. Am J Epidemiol 1995;142:1184–93. 17. Papiernik E. Prevention of preterm labour and delivery. Baillieres Clin Obstet Gynaecol 1993;3:499–521. 18. Haram K, Mortensen JHS, Wollen AL. Preterm delivery: an overview. Acta Obstet Gynecol Scand 2003;82:678–704. 19. Odland J, Tchachtchine VP, Bykov V, Fiskebeck PE, Lund E, Thomassen Y, et al. Critical evaluation of medical, statistical, and occupational data sources in the Kola Peninsula of Russia pertinent to reproductive health studies. Int Arch Occup Environ Health 1999;72:151–60. 20. Parker JD, Schoendorf KC, Kiely JL. Associations between measures of socioeconomic status and low birth weight, small for gestational age, and premature delivery in the United States. Ann Epidemiol 1994;4:271–8.
A.M. Grjibovski et al. 21. Peacock JL, Bland JM, Anderson HR. Preterm delivery: effects of socioeconomic factors, psychological stress, smoking, alcohol, and caffeine. BMJ 1995;311:531–5. 22. Koupilova I, Vagero D, Leon DA, Pikhart H, Prikazsky V, Holcik J, et al. Social variation in size at birth and preterm delivery in the Czech Republic and Sweden, 1989–91. Paediatr Perinat Epidemiol 1998;12:7–24. 23. Foix-L’Helias L, Blondel B. Changes in risk factors of preterm delivery in France between 1981 and 1995. Paediatr Perinat Epidemiol 2000;14:314–23. 24. Ma J, Stampfer MJ, Hennekens CH, Frosst P, Selhub J, Horsford J, et al. Methylenetetrahydrofolate reductase polymorphism, plasma folate, homocysteine, and risk of myocardial infarction in US physicians. Circulation 1996;94: 2410–6. 25. Olsen SF, Secher NJ. Low consumption of seafood in early pregnancy as a risk factor for preterm delivery: prospective cohort study. BMJ 2002;324:447–50. 26. Villar J, Gulmezoglu AM, de Onis M. Nutritional and antimicrobial interventions to prevent preterm birth: an overview of randomized controlled trials. Obstet Gynecol Surv 1998;53:575–85. 27. Biloukha O, Utermohlen V. Healthy eating in Ukraine: attitudes, barriers and information sources. Public Health Nutr 2001;4:207–15. 28. Bloomfield FH, Oliver MH, Hawkins P, Campbell M, Phillips DJ, Gluckman PD, et al. A periconceptional nutritional origin for non-infectious preterm birth. Science 2003; 300:606. 29. Savelyeva GM, Gavrilova DV, Lobova TA. Family planning in Russia. Int J Gynecol Obstet 1997;58:51–7.
Large social disparities in spontaneous preterm birth rates in transitional Russia ¨stro ¨ma A.M. Grjibovskia,b, L.O. Bygrena,c,*, A. Yngve, M. Sjo a
Unit for Preventive Nutrition, Department of Biosciences at Novum, Karolinska Institutet, 141 157 Huddinge, Sweden b Institute of Hygiene and Medical Ecology, Northern State Medical University, Arkhnagelsk, Russia c ˚ University, Umea ˚, Sweden Department of Community Medicine and Rehabilitation, Social Medicine, Umea Received 22 January 2004; received in revised form 21 May 2004; accepted 2 June 2004
KEYWORDS Preterm birth; Sociodemographic factors; Russia
Summary Objective. This study estimated the effect of maternal sociodemographic, obstetric and lifestyle factors on the risk of spontaneous preterm birth in a Russian town. Methods. All women with singleton pregnancies registered at prenatal care centres in Severodvinsk in 1999 comprised the cohort for this study (nZ1559). Analysis was based on spontaneous live singleton births at the maternity home (nZ1103). Multivariable logistic regression was applied to quantify the effect of the studied factors on the risk of preterm birth. Differences in gestation duration were studied using multiple linear regression. Results. In total, 5.6% of all spontaneous births were preterm. Increased risks of preterm delivery were found in women with lower levels of education and in students. Placental complications, stress and a history of fetal death in previous pregnancies were also associated with elevated risks for preterm delivery. Smoking, hypertension and multigravidity were associated with reduced length of pregnancy in metric form. Conclusion. In addition to medical risk factors, social factors are important determinants of preterm birth in transitional Russia. Large disparities in preterm birth rates may reflect the level of inequalities in transitional Russia. Social variations in pregnancy outcomes should be monitored. Q 2004 Published by Elsevier Ltd on behalf of The Royal Institute of Public Health.
Introduction
* Corresponding author. Address: Unit for Preventive Nutrition, Department of Biosciences at Novum, Karolinska Institutet, 141 157 Huddinge, Sweden. Fax: C46-8-6083350. E-mail addresses: [email protected] (L.O. Bygren), [email protected] (A.M. Grjibovski).
Preterm birth is a leading cause of infant mortality in industrialized societies.1 It is associated with elevated risks for perinatal mortality, serious neonatal morbidity, moderate to severe childhood disability, and high healthcare expenditures.2 These risks make preterm birth a serious public health problem.
0033-3506/$ - see front matter Q 2004 Published by Elsevier Ltd on behalf of The Royal Institute of Public Health. doi:10.1016/j.puhe.2004.06.005
78
A.M. Grjibovski et al.
A large body of evidence suggests that adverse pregnancy outcomes are more frequent among socially disadvantaged women, although the mediating factors may vary. According to Kramer, the most important causes of preterm delivery are genital tract infections, multiple birth, pregnancyinduced hypertension, low pre-pregnancy body mass index, incompetent cervix, prior preterm birth, abruptio placentae, heavy work and cigarette smoking.3 However, despite a great number of studies, many preterm births remain unexplained. Moreover, a majority of preterm births occur in low-risk groups.4 Economic reforms introduced in Russia after the break up of the Soviet Union in 1991 led to impoverishment of the majority of the population, a considerable decline in overall life expectancy, and increased social and health inequalities. In our earlier publications, we showed a clear gradient between maternal educational level and birth weight in a Russian setting,5 and we found that poor housing, self-perceived stress and smoking are important determinants of fetal growth in transitional Russia.6 However, it is now recognized that the aetiological determinants of gestational duration are different from the determinants of fetal growth.7 Countries of the former Soviet Union provide scarce information on risk factors for preterm delivery. We found only two studies on the determinants of preterm birth in these countries. A Ukrainian study concluded that the problems associated with transition did not alter this pregnancy outcome when the classical risk factors for preterm delivery were present.8
Figure 1
An Estonian study showed an independent effect of maternal education, marital status and nationality on the risk of preterm birth, although clinical data were not included in the analysis.9 This study estimated the effects of maternal sociodemographic, obstetric and lifestyle factors on preterm delivery and gestational length in a community-based cohort in a Russian town.
Methods The study was performed in Severodvinsk, a town in north-west Russia with a population of 233,800 in 1998. The study included all the pregnant women in the town who were registered in prenatal care centres (nZ1559). The cohort was followed through delivery. Pregnant women who attended antenatal clinics for abortion counselling or who were not permanent residents in Severodvinsk were excluded. The sampling material is summarized in Fig. 1 and is described elsewhere.5,10 Spontaneous and induced preterm births are likely to have different risk factors.11 As researchers understand more about the causes of the latter that mainly relate to maternal complications and/or endangered fetal wellbeing,4 all the analyses were only performed in a subgroup of spontaneous births. Delivery was classified as preterm if it occurred before the 37th completed week of gestation, measured from the last menstrual period. Data on maternal education, occupation, marital status, pre-pregnancy weight, gestational length and
Sampling procedure. Severodvinsk cohort, 1999.
Spontaneous preterm birth rates in transitional Russia pregnancy outcomes were obtained from the medical records at the prenatal care centres and at the maternity home. Smoking habits, alcohol consumption, stress and housing conditions were studied by means of a questionnaire administered during the mother’s first antenatal visit. The process of creating variables from these data is described elsewhere.5,6,10 In addition, data on maternal reproductive history and complications of the index pregnancy were abstracted from the medical files at the maternity home (Table 3). Gravidity was set as the number of pregnancies to the mother including the index pregnancy. Parity was the number of previous births. Previous induced abortions and miscarriages were also recorded. Placental complications included placental abruption, placenta previa and antepartum haemorrhage. The hypertension group consisted of women with pre-pregnancy hypertension and pregnancy-induced hypertension. Women who had haemoglobin levels lower than 110 g/l at the time of admission to the maternity home were classified as anaemic. If a woman’s records contained data on genital and/or urinary tract infections, she was classified as having a genito-urinary infection. Bivariate relationships were studied using Chisquare tests and one-way ANOVA for dichotomous and continuous data, respectively. Independent effects of the studied factors on the risks of preterm delivery were assessed by multivariable logistic regression. Multiple linear regression was used to estimate independent effects of the studied factors on gestational length. The group believed to be the most favourable was chosen as the reference group. To make our findings comparable with the previous studies from the former Soviet republics, multivariable analyses were performed in two sets. In the first set, only sociodemographic data were introduced into the multivariable models. In the second set, lifestyle data and reproductive heath data were added to the analysis. Due to a relatively small sample size, backward stepwise procedures in both logistic and linear regressions were used on this stage to obtain the best fit for the models.12 All analyses were performed using SPSS Version 10.0. Permission to perform the study was obtained from the Department of Health Care and the Mayor of Severodvinsk.
79 spontaneous (Fig. 1). Gestational length varied from 26 to 43 weeks (mean 39.34 weeks, SD 1.91). Preterm births constituted 5.6% of all spontaneous births. Bivariate relationships between maternal characteristics and preterm deliveries and average length of gestation are presented in Tables 1–3. Maternal education was the only sociodemographic factor significantly associated with gestational length and spontaneous preterm birth in the sample in bivariate analysis. Initiation of prenatal care after the first trimester and placental complications were associated with both reduced length of pregnancy and increased risks for preterm birth in crude analysis. Gravidity, history of induced abortion, smoking, hypertension and stress were associated with less extreme differences in pregnancy duration. Women with a history of previous fetal antenatal death had a higher crude risk of preterm delivery, although they had a longer average duration of pregnancy. Adjustment for other sociodemographic factors in the first set of multivariable analyses increased the risk of preterm births in mothers with a low level of education. Mothers aged 30 years or more were also at increased risk of preterm delivery. Gestational length in metric form was also reduced in all groups listed above (Table 4). Skilled bluecollar occupation was significantly associated with reduced risk of preterm delivery. After incorporating the data on maternal reproductive history, pregnancy complications and lifestyle factors in the multivariable models, placental complications, maternal education, history of antenatal death, being a student and perceived stress were independently associated with increased risk of preterm birth. Skilled blue-collar workers again had significantly lower risks for preterm delivery. They also had a longer duration of pregnancy than the reference group. Average gestational length was reduced in women with vocational education, in women who reported stress, in women with placental complications, in women with three or more previous pregnancies, in women who smoked and in women with hypertension. Results of the second set of multivariable analyses are summarized in Table 5.
Discussion Results There were 1399 live singleton deliveries in the sample. Among them, 1103 (78.8%) were
The proportion of infants born before term has increased in many industrialized countries over the past decades. For example, preterm birth rates increased from 9.4% in 1981 to 11.9% in 2001 in
80
A.M. Grjibovski et al.
Table 1
Spontaneous preterm birth rates and duration of pregnancy by maternal sociodemographic characteristics. N
Maternal age (years) 15–19 20–29 30C Education Secondary or less Vocational Unknown Universitya Occupation Unemployedb Blue-collar Skilled bluecollar Students White-collar Marital status Unmarried Married Pre-pregnancy weight Underweight Normal Overweight a b
Preterm (%)
P (c2)
Crude OR (95% CI)
Gestational length (weeks) (SD)
0.119 122 801 180
8.2 4.7 7.8
314
0.186 1.79 (0.87, 3.70) Reference 1.69 (0.90, 3.20)
39.20 (1.97) 39.41 (1.82) 39.16 (2.19)
9.2
9.12 (2.75, 30.30)
39.17 (2.01)
281 236 272
6.4 5.1 1.1
6.14 (1.79, 21.08) 4.80 (1.34, 17.23) Reference
39.19 (2.24) 39.53 (1.78) 39.54 (1.44)
373 171 120
5.9 5.8 2.5
1.09 (0.58, 2.01) 1.07 (0.49, 2.33) 0.44 (0.13, 1.51)
39.33 (1.84) 39.37 (1.76) 39.67 (1.64)
76 363
9.2 5.5
1.74 (0.71, 4.27) Reference
39.16 (1.83) 39.27 (2.12)
394 719
7.3 4.7
1.59 (0.95, 2.66) Reference
39.21 (1.90) 39.41 (1.91)
0.024
!0.001
0.389
0.292
0.078
0.093
0.569 58 1000 45
8.6 5.5 4.4
P (ANOVA)
0.488 1.62 (0.62, 4.22) Reference 0.80 (0.19, 3.39)
39.05 (2.40) 39.36 (1.89) 39.33 (1.48)
At least 3 years of university studies. Including housewives.
the USA,13 and from 6.3% in 1981 to 6.8% in 1992 in Canada.14 Although a substantial part of this increase can be attributed to the increased number of multiple pregnancies, changes in patterns of obstetric interventions and registration of early-gestation births, 14 changes in maternal sociodemographic factors have also been important determinants of the observed trends.15 Countries with lower prevalence of socio-economic disadvantage have lower preterm birth rates.7 Improvements in socio-economic characteristics were responsible for the decrease in preterm birth rates in Finland (from 9.1% in 1966 to 4.8% in 1985–1986),16 whereas in France, similar changes (from 7.9% in 1972 to 4.1% in 1988–1989) were mainly attributed to primary prevention strategies.17 In this study, the rate of spontaneous live preterm births was 5.6%, which is slightly higher than rates found in Ukraine (4.5%)8 and Estonia (5.3%).9 The Estonian study was based on a birth register and included both medically induced and spontaneous live singleton preterm births. If we use
the Estonian approach, we get a rate of 6.1%, although calculating the preterm rate as in the Ukrainian study gives a rate of 4.5%. The differences in inclusion criteria affect prematurity rates and make international comparisons difficult. Nevertheless, the rates in all these former Soviet republics are likely to be higher than in the Nordic countries, but lower than in the USA.13,18 Validity of the estimates and the relatively small sample size may be potential limitations of the study. Odland et al.,19 however, reported sufficient quality of the Russian medical documentation for epidemiological studies on reproductive health issues. The validity of the factors estimated by means of our questionnaire is discussed elsewhere.6,10 The maternity home documentation of pregnancy complications, especially genito-urinary infections, seems to greatly underestimate the rate of occurrence (9.8% in our study and 52.8% in the Ukrainian study). Previous preterm births and weight gain pattern are known to be associated with preterm delivery; however, these factors were
Spontaneous preterm birth rates in transitional Russia
81
Table 2 Spontaneous preterm birth rates and duration of pregnancy by maternal lifestyle factors and housing conditions.
Smoking No Yes No information Alcohol drinking No Occasional No information Stress No Yes No information Type of housing Own apartment Shared apartment No information Crowded housing Non-crowded Crowded Unknown
n
Preterm (%)
674 136 293
5.3 9.6 4.4
P (c2)
Crude OR (95% CI)
Gestational length (weeks) (SD)
Reference 1.87 (0.97, 3.64) 0.82 (0.43, 1.58)
39.36 (1.94) 38.93 (2.30) 39.48 (1.57)
0.089
0.019
0.622 594 195 314
6.2 4.6 5.1
431 331 341
4.2 8.5 4.7
0.347 Reference 0.73 (0.35, 1.54) 0.81 (0.44, 1.48)
39.27 (2.05) 39.45 (1.91) 39.42 (1.59)
Reference 2.12 (1.15, 3.90) 1.13 (0.57, 2.25)
39.49 (1.81) 39.02 (2.28) 39.47 (1.56)
0.086
0.001
0.351
0.217
632
6.5
Reference
39.26 (2.10)
181
4.4
0.67 (0.31, 1.45)
39.39 (1.67)
290
4.5
0.68 (0.36, 1.28)
39.49 (1.58)
0.457 405 409 289
6.7 5.4 4.5
poorly recorded in the medical files and therefore we did not include them in the analysis. We emphasize that our findings are relevant largely to preterm births occurring between 32 and 36 completed weeks of gestation. Moreover, the duration of gestation was measured from the history of the last menstrual period, and not from the ultrasound measures; therefore, the results should be interpreted with caution. The relatively small sample size might result in non-significant results for potentially important factors for preterm births. To overcome this, multiple linear regression analyses, which provide greater statistical power, were also performed. Preterm deliveries that are preceded by the rupture of membranes may have different aetiology compared with those preceded by spontaneous labour. However, many researchers suggest that these two groups may not represent aetiologically different entities;4 therefore, we did not perform separate
P (ANOVA)
0.311 Reference 0.80 (0.45, 1.42) 0.66 (0.33, 1.30)
39.27 (2.12) 39.31 (1.90) 39.49 (1.58)
analyses for these groups in order not to reduce the statistical power of the study. Maternal education was most strongly associated with the risk of spontaneous preterm delivery. Multivariable analysis with only sociodemographic factors in the model strengthened the associations. Including reproductive history, obstetric and lifestyle data in the model slightly reduced the associations, although they remained the largest associations observed in similar studies to the best of our knowledge. Large differences in the risks of preterm birth by maternal education have been found in the USA,20 the UK,21 Estonia9 and the Czech Republic.22 Considerably smaller variations were reported in France23 and Sweden.22 The mechanisms that may explain social disparities in preterm birth are still unclear. Kramer et al. proposed two sets of causal pathways. The first pathway involves unhealthy behaviours, infections,
82
A.M. Grjibovski et al.
Table 3 Spontaneous preterm birth rates and duration of pregnancy by maternal reproductive history and course of the index pregnancy. n
Preterm (%)
Reproductive history Gravidity 1 460 2–3 428 4C 215
4.6 5.1 8.8
Parity 0 1C
4.7 7.5
Gestational length (weeks) (SD)
0.88 (0.48, 1.63) Reference 1.79 (0.95, 3.83)
39.48 (1.76) 39.38 (1.80) 38.97 (2.33)
0.61 (0.37, 1.03) Reference
39.42 (1.80) 39.19 (1.80)
5.1 5.3 7.4
Miscarriages 0 1C
966 137
5.6 5.8
0.071
0.431
0.031 Reference 1.04 (0.54, 2.02) 1.49 (0.80, 2.76)
39.45 (1.81) 39.32 (1.80) 39.06 (2.24)
Reference 1.05 (0.49, 2.25)
39.37 (1.90) 39.15 (1.92)
0.906
0.214
0.001 1079 24
Index pregnancy Prenatal care initiation During first 935 trimester After first 168 trimester Hypertension No 1063 Yes 40
5.3 20.8
0.037 Reference 4.72 (1.70, 13.09)
39.36 (1.86) 38.54 (3.36)
0.006 4.8 10.1
0.029 Reference
39.40 (1.82)
2.23 (1.24, 3.99)
39.05 (2.32)
Reference 2.52 (0.95, 6.68)
39.37 (1.88) 38.50 (2.28)
0.054 5.4 12.5
0.004
!0.001 1078 25
4.9 36.0
Infection No Yes
995 108
6.0 1.9
Anaemia No Yes
701 402
5.1 6.5
Infant sex Female Male
536 567
5.8 5.3
!0.001 Reference 10.9 (4.59, 25.76)
39.40 (1.82) 36.88 (3.38)
0.29 (0.07, 1.22)
39.31 (1.96) 39.63 (1.28)
Reference 1.28 (0.76, 2.15)
39.33 (1.87) 39.35 (1.93)
Reference 0.94 (0.56, 1.57)
39.37 (1.91) 39.32 (1.90)
0.073
0.100
0.335
0.824
0.820
exposure to stress, and physiological reactions to these factors that shorten gestation. The second pathway involves a gene–environmental interaction based on mutation in the gene for
P (ANOVA)
0.004
0.062 742 361
644 244 215
Placental complications No Yes
Crude OR (95% CI)
0.069
Induced abortions 0 1 2C
Antenatal death 0 1C
P (c2)
0.674
methylenetetrahydrofolate reductase (MTHFR); if combined with low folate intake, this leads to hyperhomocysteinaemia, decidual vasculopathy and subsequent preterm delivery.1 In this study,
Spontaneous preterm birth rates in transitional Russia
83
Table 4 Adjusted odds ratios (OR) for preterm births and adjusted differences in gestational length between selected maternal sociodemographic factors and reference groups. Adjusteda OR (95% CI)
Adjusteda gestational age difference (weeks) (95% CI)
Maternal age (years) 15–19 20–29 30C
0.73 (0.29, 1.88) Reference 2.22 (1.14, 4.34)
0.01 (K0.42, 0.44) Reference K0.32 (K0.63, K0.01)
Education Secondary or less Vocational Unknown University
15.50 (4.17, 57.70) 10.53 (2.95, 37.63) 7.35 (1.91, 28.27) Reference
K0.51 (K0.91, K0.11) K0.68 (K1.04, K0.31) K0.13 (K0.51, 0.26) Reference
Occupation Unemployed Blue-collar Skilled blue-collar Students White-collar
0.68 (0.33, 0.44 (0.17, 0.21 (0.06, 1.72 (0.54, Reference
1.39) 1.12) 0.74) 5.50)
0.12 (K0.21, 0.44) 0.30 (K0.13, 0.72) 0.83 (0.39, 1.28) K0.12 (K0.64, 0.38) Reference
Marital status Unmarried Married
1.31 (0.77, 2.24) Reference
K0.16 (K0.40, 0.08) Reference
Pre-pregnancy weight Underweight Normal Overweight
1.96 (0.72, 5.34) Reference 0.75 (0.17, 3.25)
K0.37 (K0.87, 0.14) Reference 0.02 (K0.5, 0.58)
a
Adjusted for the variables in the table.
we controlled for the first mentioned set of mediators as well as for maternal reproductive history. All these factors explain only a small part of the observed differences in preterm birth rates, but it is unlikely that the second pathway alone is responsible for all social variations. Low folate intake is known to be more prevalent among the socially disadvantaged. Economic crisis in Russia led to considerable increase in inequalities. Poor compliance with the medical recommendations regarding diet and vitamin supplementation during pregnancy and lack of resources to follow such advice among low-educated women may be a contributing factor. Although there are 11–13% homozygotes for the MTHFR mutation in the USA,24 the prevalence of this mutation in northwest Russia is unknown. Among other nutritional factors, fish consumption,25 fish oil supplements, zinc, iron and calcium-rich foods26 are known to be associated with a reduced number of preterm births. In transitional Russia, seafood, fruits, meat and dairy products are relatively expensive and may not be consumed by mothers with a low level of education. High costs were identified as the main barrier to healthy eating in Ukraine, and this may be true for Russia, given that both the economic
situation and food habits are similar in these countries.27 Based on experimental data in sheep, Bloomfield et al. suggested that even a modest periconceptional reduction in food intake might lead to accelerated fetal hypothalamic–pituitary–adrenal axis maturation, increased fetal adrenocorticotrophic hormone concentrations, and subsequent preterm delivery.28 If these findings are applicable to humans, family planning issues may become more important for the prevention of preterm delivery. Unplanned pregnancies were far more common among women with a low level of education (84% in women with basic education compared with 23% in women with university education in this cohort), which might also contribute to the large social disparities in preterm birth rates. It is widely accepted that heavy work during pregnancy increases the risk of preterm delivery.3 In our analysis, a group of skilled blue-collar workers had a lower risk of preterm birth and a greater average gestational length. One of the most probable explanations of this association is that most of these women work for state-owned industrial enterprises with compulsory regular
84
A.M. Grjibovski et al.
Table 5 Results of stepwise multivariable regression analyses with preterm delivery and gestational length as dependent variables. Adjusted OR for preterm birth
Secondary or less education Vocational education Unknown education Skilled bluecollar workers Students Previous antenatal death Placental complications Stress Gravidity of 4C Smoking Hypertension
95% CI
P
Adjusted difference in gestational length (weeks)
95% CI
P
K0.44
K0.73, K0.14
0.004
0.32, 1.15
0.001
8.52
2.53, 28.73
0.001
10.21
2.82, 36.94
!0.001
6.55
1.75, 24.51
0.005
0.19
0.05, 0.75
0.017
2.53 6.15
1.00, 6.42 1.85, 20.47
0.050 0.003
11.44
4.32, 30.34
!0.001
K2.36
K3.11, K1.62
!0.001
2.15
1.20, 3.83
0.009
K0.37 K0.35
K0.61, K0.13 K0.62, K0.07
0.003 0.014
K0.37 K0.59
K0.71, K0.04 K1.18, K0.00
0.030 0.049
0.73
Reference categories are as in Tables 1–3; only the variables significantly associated with the studied outcomes are presented in the table.
medical check-ups; these might result in better general health and lower prevalence of infection. Greater levels of social security for these women compared with women employed elsewhere may also explain this finding. Students (including secondary school pupils) had a 2.5 times higher risk of spontaneous preterm delivery than the reference group. High academic demands, walking or using busy city transport to get to the educational institutions may be associated with high levels of stress, which is known to be an important determinant of preterm birth.18 Increased secretion of placental corticotropin releasing hormone and subsequent oestrogen and prostaglandin production may be the mediating factors between maternal chronic exposure to stress and preterm delivery. It has been proposed that chronic stress may be one of the most important factors that may explain large social disparities in preterm birth rates.1 In our study, we found a two-fold increase in preterm birth rates in women who reported self-perceived stress at the time of prenatal care initiation. Our estimate of stress is imprecise, and this may be why adjustment
to stress only decreased the effect of maternal education on gestational length slightly. Advanced maternal age was associated with increased risk of preterm delivery and shorter average gestation, independent of sociodemographic factors, although these differences became insignificant after adjustment for other factors. Women with three or more previous pregnancies (but not births) had significantly shorter gestation periods. These women were more likely to have had previous abortions. Multiple induced abortions were associated with reduced gestational length in bivariate analysis. Induced abortions with dilatation and curettage are common in Russia with an average rate of two abortions per birth, although a slight downward trend has been observed.29 Urinary and genital tract infections are wellknown determinants of preterm birth.1,3,7,18 Surprisingly, our study did not reveal elevated risks of preterm delivery or shortened gestational length associated with infection. Moreover, in bivariate analysis, the prevalence of diagnosed infection was considerably lower in women who delivered
Spontaneous preterm birth rates in transitional Russia preterm, which may be explained by the fact that women who deliver preterm are often admitted to a maternity home in labour and are rarely tested for infections (AN Baranov, personal communication). Placental complications, such as placenta previa, abruptio placentae and antepartum haemorrhage, are well-established risk factors for preterm delivery, although they normally result in medically induced labour. Hypertension before pregnancy or pregnancy-induced hypertension influence gestational length.18 In the Ukrainian study, women with pre-existing hypertension were 2.3 times more likely to deliver preterm.8 In our study, the small number of cases made it impossible to find significant differences in preterm birth rates in relation to hypertension, although a 2.5-fold difference was found in bivariate analysis. The difference in gestational age in metric form was 0.6 weeks or 4.1 days on average, which indicates that hypertensive disorders may shorten the duration of pregnancy in the study setting. All these complications can result in medically induced deliveries, which may explain the low prevalence of these disorders in our sample, which only included those women who had spontaneous births. Anaemia has also been reported as a factor that influences the length of gestation; however, this was not the case in our study, which did not take the decrease of haemoglobin until late pregnancy into account. This may result in lower haemoglobin values in full-term births because of longer gestation. Antenatal care can prevent pregnancy complications by identifying risk factors and detecting complications at an early stage. In Russia, antenatal care is free and covers almost all pregnancies.5 Early initiation of care was associated with a lower risk of preterm birth and a longer gestation in bivariate analysis. Given the inaccurate registration of pregnancy complications in the medical files at the maternity home, we attempted to evaluate the importance of prenatal care by excluding all pregnancy complications from the multivariable models. After such transformation, late initiation of care became a risk factor associated with increased risk of preterm delivery (ORZ2.03, 95% CI 1.09–3.80). Smoking is often cited as a risk factor of preterm delivery.18 Risks associated with smoking in relation to preterm delivery are much lower than those in relation to fetal growth restriction,1 suggesting a modest role of smoking as an explanatory variable for sociodemographic disparities in preterm birth rates. We revealed a modest reduction in duration of pregnancy among smoking
85 mothers, although this effect may be underestimated because our group of smokers only included daily smokers.6 In summary, low maternal education, being a student, perceived stress, placental complications during the index pregnancy, and a history of fetal death in previous pregnancies were associated with increased risk of spontaneous preterm delivery in an urban Russian setting. Multigravidity, smoking and hypertensive disorders reduced the average length of gestation without increasing prematurity rates. We conclude that in addition to the traditional risk factors for preterm delivery, social factors are important determinants of pregnancy outcomes in transitional Russia, and the social disparities in pregnancy outcomes may be even more profound in large cities with a higher degree of social inequality.
References 1. Kramer MS, Goulet L, Lydon J, Seguin L, McNamara H, Dassa C, et al. Socio-economic disparities in preterm birth: causal pathways and mechanisms. Paediatr Perinat Epidemiol 2001;15(Suppl 2):104–23. 2. Lumley J. Defining the problem: the epidemiology of preterm birth. Br J Obstet Gynaecol 2003;110(Suppl 20): 3–7. 3. Kramer MS. The epidemiology of adverse pregnancy outcomes: an overview. J Nutr 2003;133(Suppl 2):1592–6. 4. Moutquin JM. Socio-economic and psychosocial factors in the management and prevention of preterm labour. Br J Obstet Gynaecol 2003;110(Suppl 20):56–60. 5. Grjibovski A, Bygren LO, Svartbo B, Magnus P. Social variations in foetal growth in a Russian setting: an analysis of medical records. Ann Epidemiol 2003;13:599–605. 6. Grjibovski A, Bygren LO, Svartbo B, Magnus P. Housing conditions, perceived stress, smoking and alcohol—determinants of foetal growth in Northwest Russia. Acta Obstet Gynecol Scand 2004 in press. 7. Kramer MS, Seguin L, Lydon J, Goulet L. Socio-economic disparities in pregnancy outcome: why do the poor fare so poorly? Paediatr Perinat Epidemiol 2000;14:194–210. 8. Monaghan SC, Little RE, Hulchiy O, Strassner H, Gladen BC. Risk factors for spontaneous preterm birth in two urban areas of Ukraine. Paediatr Perinat Epidemiol 2001;15: 123–30. 9. Koupilova I, Rahu K, Rahu M, Karro H, Leon DA. Social determinants of birth weight and length of gestation in Estonia during the transition to democracy. Int J Epidemiol 2000;29:118–24. 10. Grjibovski A, Bygren LO, Svartbo B. Socio-demographic determinants of poor infant outcome in north-west Russia. Paediatr Perinat Epidemiol 2002;16:255–62. 11. Meis PJ, Michielutte R, Peters TJ, Wells HB, Sands RE, Coles EC, et al. Factors associated with preterm birth in Cardiff, Wales. II. Indicated and spontaneous preterm birth. Am J Obstet Gynecol 1995;173:597–602.
86 12. Katz MH. Performing the analysis Multivariable analysis. A practical guide for clinicians. Cambridge: Cambridge University Press; 1999 p. 84–117. 13. Martin JA, Hamilton BE, Ventura SJ, Menacker F, Park MM, Sutton PD. Births: final data for 2001. Nat Vit Stat Rep 2002; 2:1–103. 14. Joseph KS, Kramer MS, Marcoux S, Ohlsson A, Wen SW, Allen A, et al. Determinants of preterm birth rates in Canada from 1981 through 1983 and from 1992 through 1994. N Engl J Med 1998;339:1434–9. 15. Kramer MS, Platt R, Yang H, Joseph KS, Wen SW, Morin L, et al. Secular trends in preterm birth: a hospital-based cohort study. JAMA 1998;280:1849–54. 16. Olsen P, Laara E, Rantakallio P, Jarvelin MR, Sarpola A, Hartikainen AL. Epidemiology of preterm delivery in two birth cohorts with an interval of 20 years. Am J Epidemiol 1995;142:1184–93. 17. Papiernik E. Prevention of preterm labour and delivery. Baillieres Clin Obstet Gynaecol 1993;3:499–521. 18. Haram K, Mortensen JHS, Wollen AL. Preterm delivery: an overview. Acta Obstet Gynecol Scand 2003;82:678–704. 19. Odland J, Tchachtchine VP, Bykov V, Fiskebeck PE, Lund E, Thomassen Y, et al. Critical evaluation of medical, statistical, and occupational data sources in the Kola Peninsula of Russia pertinent to reproductive health studies. Int Arch Occup Environ Health 1999;72:151–60. 20. Parker JD, Schoendorf KC, Kiely JL. Associations between measures of socioeconomic status and low birth weight, small for gestational age, and premature delivery in the United States. Ann Epidemiol 1994;4:271–8.
A.M. Grjibovski et al. 21. Peacock JL, Bland JM, Anderson HR. Preterm delivery: effects of socioeconomic factors, psychological stress, smoking, alcohol, and caffeine. BMJ 1995;311:531–5. 22. Koupilova I, Vagero D, Leon DA, Pikhart H, Prikazsky V, Holcik J, et al. Social variation in size at birth and preterm delivery in the Czech Republic and Sweden, 1989–91. Paediatr Perinat Epidemiol 1998;12:7–24. 23. Foix-L’Helias L, Blondel B. Changes in risk factors of preterm delivery in France between 1981 and 1995. Paediatr Perinat Epidemiol 2000;14:314–23. 24. Ma J, Stampfer MJ, Hennekens CH, Frosst P, Selhub J, Horsford J, et al. Methylenetetrahydrofolate reductase polymorphism, plasma folate, homocysteine, and risk of myocardial infarction in US physicians. Circulation 1996;94: 2410–6. 25. Olsen SF, Secher NJ. Low consumption of seafood in early pregnancy as a risk factor for preterm delivery: prospective cohort study. BMJ 2002;324:447–50. 26. Villar J, Gulmezoglu AM, de Onis M. Nutritional and antimicrobial interventions to prevent preterm birth: an overview of randomized controlled trials. Obstet Gynecol Surv 1998;53:575–85. 27. Biloukha O, Utermohlen V. Healthy eating in Ukraine: attitudes, barriers and information sources. Public Health Nutr 2001;4:207–15. 28. Bloomfield FH, Oliver MH, Hawkins P, Campbell M, Phillips DJ, Gluckman PD, et al. A periconceptional nutritional origin for non-infectious preterm birth. Science 2003; 300:606. 29. Savelyeva GM, Gavrilova DV, Lobova TA. Family planning in Russia. Int J Gynecol Obstet 1997;58:51–7.