Happy prenatal relationships, healthy postpartum mothers: a prospective study of relationship satisfaction, postpartum stress, and health.

OBJECTIVE: Major life transitions, such as the arrival of a new baby, can be stressful, and therefore, costly to one's health. Being in a satisfying romantic relationship has been previously linked with better health, perhaps because it provides access to a bank of psychosocial resources. The objective of the present research was two-fold. First, we examined whether prenatal relationship satisfaction benefitted mothers' postpartum health. Second, we examined whether this association was mediated by reduced postpartum stress. Design: Pregnant women (N = 431) were drawn from a larger prospective pregnancy and birth cohort study. Main Outcome Measures: Participants reported on their relationship satisfaction, perceived stress, and perceived health at two time points: 24-26 weeks of gestation and 6 months postpartum. Results: Greater prenatal romantic satisfaction predicted enhanced perceived postpartum health, and reduced perceived postpartum stress appeared to mediate this link. Conclusion: This research demonstrates that greater prenatal relationship satisfaction is associated with better postpartum health and provides evidence for one potential mechanism for this link: the reduction of perceived postpartum stress. Thus, assessments of women's prenatal relationship satisfaction may help identify those who are at higher risk of experiencing poor postpartum health. Implications for interventions are discussed.

Income inequality and social gradients in children's height: a comparison of cohort studies from five high-income countries.

BACKGROUND: Health and well-being are better, on average, in countries that are more equal, but less is known about how this benefit is distributed across society. Height is a widely used, objective indicator of child health and predictor of lifelong well-being. We compared the level and slope of social gradients in children's height in high-income countries with different levels of income inequality, in order to investigate whether children growing up in all socioeconomic circumstances are healthier in more equal countries. METHODS: We conducted a coordinated analysis of data from five cohort studies from countries selected to represent different levels of income inequality (the USA, UK, Australia, the Netherlands and Sweden). We used standardised methods to compare social gradients in children's height at age 4-6 years, by parent education status and household income. We used linear regression models and predicted height for children with the same age, sex and socioeconomic circumstances in each cohort. RESULTS: The total analytic sample was 37 063 children aged 4-6 years. Gradients by parent education and household income varied between cohorts and outcomes. After adjusting for differences in age and sex, children in more equal countries (Sweden, the Netherlands) were taller at all levels of parent education and household income than children in less equal countries (USA, UK and Australia), with the greatest between-country differences among children with less educated parents and lowest household incomes. CONCLUSIONS: The study provides preliminary evidence that children across society do better in more equal countries, with greatest benefit among children from the most disadvantaged socioeconomic groups.

How and Where Parents of Infants and Young Children Want to Receive Nutrition Information.

PURPOSE: To understand how and where parents of infants and young children (children ≤5 years old) prefer to receive nutrition information. METHODS: A 1-page survey was developed and pilot tested at 2 community agencies. The final survey was distributed at 18 community health centres (CHCs) in Calgary and surrounding rural areas. Any parent attending a well-child visit (child ≤5 years old) was able to participate. RESULTS: Five hundred and twenty-nine surveys were completed. The majority of respondents at every CHC identified online reading (79.2%) in their home (86.0%) as the preferred method and location to receive nutrition information. Almost all (99.4%) participants had internet access. Handouts (38.6%) were the second most popular way to receive nutrition information. In-person and online classes were only a preferred method by a small percentage of respondents, 10.6% and 8.1%, respectively. CONCLUSIONS: Appropriate, evidence-based nutrition websites should be promoted to parents with young children. Health professionals should be aware that parents likely access nutrition information online, and they need to provide an opportunity for parents to discuss what they found. Future research is needed to understand which websites parents access for online nutrition information and how they discern whether it is credible.

Poor, Unsafe, and Overweight: The Role of Feeling Unsafe at School in Mediating the Association Among Poverty Exposure, Youth Screen Time, Physical Activity, and Weight Status.

This study applied socioecological and cumulative risk exposure frameworks to test the hypotheses that 1) the experience of poverty is associated with feeling less safe at school, and 2) feeling less safe is associated with engaging in poorer weight-related behaviors, as well as an increased probability of being overweight or obese. Data were from the ongoing Québec Longitudinal Study of Child Development, initiated in 1998 with a population-based cohort of 2,120 Québec (Canada) infants 5 months of age and their parent or primary caregiver. Measures of youths' (age, 13 years) self-reported feelings of safety, screen time, physical activity, and objectively assessed not overweight/obese (70%), overweight (22%), and obese (8%) weight status were collected in 2011. Family poverty trajectory from birth was assessed by using latent growth modeling. As hypothesized, exposure to poverty was associated with feeling less safe at school and, in turn, with an increased probability of being overweight or obese. The association was most pronounced for youths who experienced chronic poverty. Compared with youths who experienced no poverty and felt unsafe, those who experienced chronic poverty and felt unsafe were nearly 18% more likely to be obese (9.2% vs. 11.2%). Although feeling unsafe was associated with screen time, screen time did not predict weight status.

Parenting style and obesity risk in children.

BACKGROUND: Parents play a critical role in their children's lifestyle habits. The objective was to assess the effect of parenting style on the risk of childhood obesity, and to determine whether poverty was a moderator of the association. METHODS: Participants were from the 1994-2008 cross-sectional samples of the National Longitudinal Survey of Children and Youth (NLSCY), a nationally representative survey of Canadian youth. Factor and cluster analyses identified four parenting styles consistent with Baumrind's parenting style prototypes. Multivariable logistic regression assessed the risk of obesity based on parenting style after adjusting for covariates. Analyses were stratified by age (preschool: 2-5years of age, n=19,026; school-age: 6-11years of age, n=18,551) and the moderating effect of poverty (household income

Short-term heart rate variability in a population-based sample of 10-year-old children.

Heart rate variability (HRV) is a non-invasive quantitative marker of cardiac autonomic function derived from continuous electrocardiogram (ECG) recordings. Normative HRV values and development factors have not been established in pediatric populations. The objective was to derive referent time- and frequency-domain HRV values for a population-based sample of children. Children aged 9-11 years (N = 1,036) participated in the Québec Longitudinal Study of Child Development cohort cardiovascular health screening. Registered nurses measured anthropometrics (height, weight) and children wore an ambulatory Holter monitor to continuously record an ECG signal. HRV variables included time (SDNN, pNN50, RMSSD, SDANN) and frequency (HF, LF, LF/HF ratio) domain variables. Normative HRV values, stratified by age, sex, and heart rate, are presented. Greater heart rate (ß avg  = -0.60, R avg (2)  = 0.39), pubertal maturation (ß avg = -0.11, R avg (2)  = 0.01), later ECG recording times (ß avg = -0.19, R avg (2)  = 0.07), and higher diastolic blood pressure (ß avg = -0.11, R avg (2)  = 0.01) were significantly associated with reduced HRV in 10-year-old children. The normative HRV values permit clinicians to monitor, describe, and establish pediatric nosologies in primary care and research settings, which may improve treatment of diseases associated with HRV in children. By better understanding existing values, the practical applicability of HRV among clinicians will be enhanced. Lastly, developmental (e.g., puberty) and procedural (e.g., recording time) factors were identified that will improve recording procedures and interpretation of results.

The role of the residential neighborhood in linking youths' family poverty trajectory to decreased feelings of safety at school.

Although disadvantaged youth are more likely to be victimized at school, victimization only partly explains their decreased feelings of safety at school. We applied a socioecological approach to test the hypotheses that the experience of poverty is associated with decreased feelings of safety at school, and that residential neighborhood features partly mediate the relationship between poverty and feeling less safe at school. This study draws on the Québec Longitudinal Study of Child Development (QLSCD) which began in 1998 with a representative population-based cohort of 2,120 5-month old infants (49.1% female) and their primary caregiver. The study also includes measures of ego-centred residential neighborhood exposures (based on a 500 m circular buffer zone surrounding the family's residential postal code) derived from a spatial data infrastructure. We used latent growth modeling to estimate youth's family poverty trajectory from age 5 months to 13 years, and structural equation modeling to test our hypotheses. The results suggest that youth experiencing chronic and later-childhood poverty felt less safe at school in part because they lived in neighborhoods that their parents described as being disorderly (e.g., demarked by the presence of garbage, drug use and groups of trouble-makers). These neighborhoods also tended to have less greenery (e.g., trees, parks) and more lone-parent households. Neighborhood features did not help explain the relationship between early-childhood poverty and feeling less safe at school. The findings suggest that targeting residential neighborhood features such as greenery and disorder could improve youth's felt safety at school, particularly for those experiencing chronic and later-childhood poverty.

Maternal prenatal felt security and infant health at birth interact to predict infant fussing and crying at 12 months postpartum.

Infants born with medical problems are at risk for less optimal developmental outcomes. This may be, in part, because neonatal medical problems are associated with maternal distress, which may adversely impact infants. However, the reserve capacity model suggests that an individual's bank of psychosocial resources buffers the adverse effects of later-encountered stressors. This prospective longitudinal study examined whether preexisting maternal psychosocial resources, conceptualized as felt security in close relationships, moderate the association between neonatal medical problems and infant fussing and crying 12 months postpartum. Maternal felt security was measured by assessing its indicators in 5,092 pregnant women. At birth, infants were classified as healthy or having a medical problem. At 12 months, experience sampling was used to assess daily maternal reports of fussing and crying in 135 mothers of infants who were healthy or had medical problems at birth. Confirmatory factor analyses revealed that attachment, relationship quality, self-esteem, and social support can be conceptualized as indicators of a single felt security factor. Multiple regression analyses revealed that prenatal maternal felt security interacts with infant health at birth to predict fussing and crying at 12 months. Among infants born with medical problems, higher felt security predicted decreased fussing and crying. Maternal felt security assessed before birth dampens the association between neonatal medical problems and subsequent infant behavior. This supports the hypothesis that psychosocial resources in reserve can be called upon in the face of a stressor to reduce its adverse effects on the self or others.

Impact of the 2008 economic and financial crisis on child health: a systematic review.

The aim of this study was to provide an overview of studies in which the impact of the 2008 economic crisis on child health was reported. Structured searches of PubMed, and ISI Web of Knowledge, were conducted. Quantitative and qualitative studies reporting health outcomes on children, published since 2007 and related to the 2008 economic crisis were included. Two reviewers independently assessed studies for inclusion. Data were synthesised as a narrative review. Five hundred and six titles and abstracts were reviewed, from which 22 studies were included. The risk of bias for quantitative studies was mixed while qualitative studies showed low risk of bias. An excess of 28,000-50,000 infant deaths in 2009 was estimated in sub-Saharan African countries, and increased infant mortality in Greece was reported. Increased price of foods was related to worsening nutrition habits in disadvantaged families worldwide. An increase in violence against children was reported in the U.S., and inequalities in health-related quality of life appeared in some countries. Most studies suggest that the economic crisis has harmed children's health, and disproportionately affected the most vulnerable groups. There is an urgent need for further studies to monitor the child health effects of the global recession and to inform appropriate public policy responses.

Persistent and occasional poverty and children's food consumption: evidence from a longitudinal Québec birth cohort.

BACKGROUND: Childhood poverty is associated with poorer food consumption but longitudinal data are limited. The objective was to assess if food consumption differs depending on age (6, 7, 10 and 12 years) and pattern of poverty. METHODS: Participants were from the 1998-2010 'Quebec Longitudinal Study of Child Development' birth cohort. Poverty was defined as income below the low-income thresholds established by Statistics Canada which adjusts for household size and geographic region. Multiple imputation was used for missing data, and latent class growth analysis identified poverty trajectories. Multivariable ordinal logistic regression assessed the association between poverty and greater consumption of milk, cheese, fruits, vegetables, sweets and sugar-sweetened beverages (SSB). RESULTS: Four poverty trajectories were identified: 1 reference category (stable non-poor) and 3 higher-risk categories (stable poor, increasing and decreasing risk). The probability of more frequent consumption was lower among children from stable poor households compared to children from stable non-poor households for fruit (6, 10 and 12 years), milk and vegetables (6, 7, 10 and 12 years) but was higher for SSB (10 and 12 years). Among children from increasing and decreasing poverty households compared to stable non-poor households, the probability of greater consumption of fruits and vegetables was lower and greater consumption of SSB was higher by the age of 12 years. CONCLUSIONS: While experiencing continual exposure to poverty has detrimental effects on food consumption throughout childhood, the association for milk, fruits and vegetables does not differ across age. Intermittent exposure to poverty may also have long-lasting effects.

Poverty's latent effect on adiposity during childhood: evidence from a Québec birth cohort.

BACKGROUND: Childhood poverty heightens the risk of obesity in adulthood, but the age at which this risk appears is unclear. We analysed the association between poverty trajectories with body mass index (BMI) Z-scores or the risk of being overweight or obese across four ages (6 years, 8 years, 10 years and 12 years) in childhood. METHODS: Data were from the 1998-2010 'Quebec Longitudinal Study of Child Development' cohort (n=698). Poverty was defined using Statistics Canada's thresholds, and trajectories were characterised with a Latent Class Growth Analysis. Multivariable linear and logistic regression models adjusted for sex, whether the mother was an immigrant, maternal education and birth weight. RESULTS: Four income trajectories were identified: a reference group (stable non-poor), and 3 higher exposure categories (increasing likelihood of poverty, decreasing likelihood of poverty or stable poor). Compared with children from stable non-poor households, children from stable poor households had BMI Z-scores that were 0.39 and 0.43 larger than children from stable non-poor households at age 10 years and 12 years, respectively (p<0.05). Compared with children from stable non-poor households, children from stable poor households were 2.22, 2.34, and 3.04 times more likely to be overweight or obese at age 8 years, 10 years and 12 years, respectively (p<0.05). CONCLUSION: A latency period for the detrimental effects of child poverty on the risk of overweight or obesity was detected. Whether the effects continue to widen with increasing duration of exposure to poverty as the children age should be investigated.

Comparison of three lifecourse models of poverty in predicting cardiovascular disease risk in youth.

OBJECTIVE: Childhood poverty heightens the risk of adulthood cardiovascular disease (CVD), but the underlying pathways are poorly understood. Three lifecourse models have been proposed but have never been tested among youth. We assessed the longitudinal association of childhood poverty with CVD risk factors in 10-year-old youth according to the timing, accumulation, and mobility models. METHODS: The Québec Longitudinal Study of Child Development birth cohort was established in 1998 (n = 2120). Poverty was defined as annual income below the low-income thresholds defined by Statistics Canada. Multiple imputation was used for missing data. Multivariable linear regression models adjusted for gender, pubertal stage, parental education, maternal age, whether the household was a single parent household, whether the child was overweight or obese, the child's physical activity in the past week, and family history. RESULTS: Approximately 40% experienced poverty at least once, 16% throughout childhood, and 25% intermittently. Poverty was associated with significantly elevated triglycerides and insulin according to the timing and accumulation models, although the timing model was superior for predicting insulin and the accumulation model was superior for predicting triglycerides. CONCLUSIONS: Early and prolonged exposure to poverty significantly increases CVD risk among 10-year-old youth.

Maternal stress/distress, hormonal pathways and spontaneous preterm birth.

BACKGROUND: Although second-trimester blood corticotrophin-releasing hormone (CRH) levels are robustly associated with preterm birth, the findings with respect to cortisol have been inconsistent, as have been those relating stress hormones to measured stressors and maternal distress. METHODS: We measured plasma CRH, adrenocorticotrophic hormone (ACTH), cortisol, cortisol-binding globulin, oestradiol and progesterone at 24-26 weeks in a nested case-control study of 206 women who experienced spontaneous preterm birth and 442 term controls. We also related the hormonal levels to measures of environmental stressors, perceived stress and maternal distress (also assessed at 24-26 weeks) and to placental histopathology. RESULTS: With the exception of an unexpectedly low oestradiol:progesterone ratio among cases (adjusted odds ratio = 0.5 [95% confidence interval 0.3, 0.8] for ratios above the median in controls), none of the hormonal measures was independently associated with spontaneous preterm birth; placental histopathological evidence of infection/inflammation, infarction or decidual vasculopathy; or measures of maternal stress or distress. CRH levels were positively associated with cortisol, but not with ACTH, whereas ACTH was also positively associated with cortisol. CONCLUSIONS: Our findings suggest an intact pituitary-adrenal axis and confirm the positive feedback effect of cortisol on (placental) CRH. Neither of these hormonal pathways, however, was strongly linked to maternal stress/distress or to the risk of spontaneous preterm birth.

Is the obesity epidemic worsening the cardiovascular risk factor profile of children? Evidence from two Québec samples measured 10 years apart.

BACKGROUND: The impact of the obesity epidemic on cardiovascular health in young people is of increasing concern. However, data on secular trends in CVD indicators are outdated and mixed. METHODS: This study compared lipid profiles and insulin of 9-10 year olds in 2008 (n = 605) and 1999 (n = 779). Data were drawn from two population-based samples of youth: the 1999 Québec Child and Adolescent Health and Social Survey and the 2008 Québec Longitudinal Study of Child Development. RESULTS: Mean body mass index (BMI) Z-scores were higher in 2008 than in 1999 in both boys (0.37 vs 0.12, p = 0.004) and girls (0.32 vs 0.05, p = 0.0004). After adjusting for maturity stage, height, BMI Z-score, age and household income, high-density lipoprotein cholesterol was 0.12 mmol/L (p < 0.05) and 0.10 mmol/L (p < 0.05) higher in 2008 than 1999 in boys and girls, respectively. Total cholesterol, low density lipoprotein cholesterol and insulin were not significantly different between 2008 and 1999. CONCLUSIONS: Despite higher BMI Z-scores in 2008, differences in cardiometabolic indicators between 1999 and 2008 were small and may not be clinically meaningful. Surveillance to closely monitor trends in cardiometabolic indicators in Canadian youth is needed.

Understanding the determinants of active transportation to school among children: evidence of environmental injustice from the Quebec Longitudinal Study of Child Development.

PURPOSE: To examine the combined influence of poverty and dangerousness of the neighborhood on active transportation (AT) to school among a cohort of children followed throughout the early school years. METHODS: Growth curve modeling was used to identify determinants of AT to school among 710 children participating in the Quebec Longitudinal Study of Child Development from 2003 through 2006. Parent-reported dangerousness and pedestrian-vehicle collision data were merged with travel mode and health data. RESULTS: At age 6 years, insufficient household income, having an older sibling, and living in a neighborhood that is not excellent for raising children, or characterized with high decay were predictive of greater likelihood of using AT and remained unchanged as children progressed from kindergarten through grade 2. CONCLUSION: A public health concern is children experiencing environmental injustice. Since AT is most likely to be adopted by those living in poverty and because it is also associated with unsafe environments, some children are experiencing environmental injustice in relation to AT. Interventions may be implemented to reduce environmental injustice through improvements in road safety.

Tracking exposure to child poverty during the first 10 years of life in a Quebec birth cohort.

OBJECTIVES: Early childhood poverty is associated with adult chronic diseases. The objectives of this study were to examine patterns of exposure to poverty during the first 10 years of life in the Quebec Longitudinal Study of Child Development (QLSCD) cohort according to three measures of poverty and to explore family characteristics associated with different poverty exposures. METHOD: Data from 1,334 participants from the QLSCD were collected annually at home from ages 5 months through 10 years. Household income (previous 12 months) and sources of income were recorded at each data round. Poverty status was operationalized as 1) living below the low income cut-off of Statistics Canada, 2) receiving social welfare and 3) being in the lowest quintile of socio-economic status. We plotted trends in the prevalence of child poverty over time. We used latent class growth modelling to identify subgroups with similar poverty trajectories. Duration of poverty according to each measure was computed separately for early childhood, middle childhood, and the entire 10 years of life. RESULTS: Four trajectories of poverty were identified: stable poor, decreasing likelihood, increasing likelihood, and never poor. The three measures of poverty do not cover the same population, yet the characteristics of those identified as poor are similar. Children of non-European, immigrant mothers were most likely to be poor, and there was a higher likelihood of children from single-parent families to live in chronic poverty during the first 10 years. CONCLUSION: A large proportion of children are exposed to poverty before 10 years of age. More effective public policies could reduce child poverty.

Child asthma and change in elevated depressive symptoms among mothers of children of a birth cohort from Quebec.

The authors examined the association between maternal reports of child asthma attacks since birth and occurrence of elevated maternal depressive symptoms at seventeen months postpartum in the present study. The modifying role of poverty in this association was also examined. Data from n = 1,696 mother-child dyads from the Quebec Longitudinal Study of Child Development, a birth cohort of children born in 1998, were used. Maternal depressive symptoms were measured with an abridged and validated twelve-item version of the Center for Epidemiologic Studies Depression Scale. Maternal reports of child asthma attacks since birth in relation to the occurrence of maternal depressive symptoms at 17 months postpartum and the potential modifying role of poverty were tested using multiple logistic regression models. When mothers reported child asthma attacks, those without elevated depressive symptoms at 5 months postpartum had lower odds of elevated depressive symptoms one year later (OR = 0.2, 95% CI: 0.1-0.7). Poverty was associated with increased odds of elevated maternal depressive symptoms (OR = 2.4, 95% CI: 1.5-3.9), without interacting with child asthma. Through this study, the authors suggest that in mothers without elevated symptoms at 5 months, reported child asthma attacks since birth did not contribute one year later to new occurrence of depressive symptoms.

The influence of poverty and social support on the perceived health of children born to minority migrant mothers.

OBJECTIVE: Poverty and low social support are common among minority migrant families. Little is known about their impact on the health of children of minority migrants to Canada. This study examined the associations between maternal perception of child's health and migration status, and examined the specific role of poverty and low social support in these associations. DESIGN: Data from the first two rounds of the Quebec Longitudinal Study of Child Development (QLSCD) were analysed. The sample included 1990 children at 17 months of age, classified according to their mother's migration status: children of minority migrant mothers (n=165) and Canadian-born mothers (n=1825). Maternal perception of child's health status and social support were measured at 17 months, household income was measured at 5 and 17 months. Multivariable logistic regressions were performed; interactions of migration status with poverty and social support were tested. RESULTS: Poverty and low social support were more common among minority migrant mothers than among Canadian-born mothers. Children of minority migrant mothers who were 'never poor' and reported high levels of social support were perceived in better health (OR 0.42; 95% confidence interval (CI): 0.19-0.91) than children of Canadian-born mothers (reference group). In contrast, children of minority migrant mothers who were 'always poor' and reported low social support were perceived in worse health (OR 6.32; 95% CI: 1.69-23.71) compared to the reference group. CONCLUSION: In Quebec, economic hardship and lack of social support are common realities among minority migrants with young children. Combined exposure to poverty and low social support is most detrimental to the perceived health of children of minority migrants.

Non-stress-related factors associated with maternal corticotrophin-releasing hormone (CRH) concentration.

During pregnancy, most maternal corticotrophin-releasing hormone (CRH) is secreted by the placenta, not the hypothalamus. Second trimester maternal CRH concentration is robustly associated with the subsequent risk of preterm birth, and it is often assumed that physiological and/or psychological stress stimulates placental CRH release. Evidence supporting the latter assumption is weak, however, and other factors affecting maternal CRH have received little attention from investigators. We carried out a case-control study nested within a large, multicentre prospective cohort of pregnant women to examine potential 'upstream' factors associated with maternal CRH concentration measured at 24-26 weeks of gestation. The predictors studied included maternal age, parity, birthplace (as a proxy for ethnic origin), pre-pregnancy body mass index, height, smoking, bacterial vaginosis and vaginal fetal fibronectin (FFN) concentration. Women with high (above the median) plasma CRH concentration were significantly less likely to have been born in Sub-Saharan Africa or the Caribbean, less likely to be overweight or obese, and more likely to be smokers. Associations with maternal birthplace and BMI persisted in logistic regression analyses controlling for potential confounding variables and when restricted to term controls. A strong (but imprecise and statistically non-significant) association was also observed with high vaginal FFN concentration. Further studies are indicated both in animal models and human populations to better understand the biochemical and physiological pathways to CRH secretion and their aetiological role, if any, in preterm birth.

Vaginal douching, bacterial vaginosis, and spontaneous preterm birth.

OBJECTIVE: Vaginal douching and bacterial vaginosis (BV) are independently associated with spontaneous preterm birth. Because the interrelationships among these variables remain unclear, we sought to examine the associations in a prospective study. METHODS: We conducted a nested case-control study within a prospectively recruited cohort of pregnant women. We prospectively collected demographic and health status data, data on pre-pregnancy vaginal douching, vaginal smears for bacterial vaginosis as defined by Nugent's criteria, fetal fibronectin at 26 weeks of pregnancy, and placental pathology at delivery. Spontaneous preterm births before 37 weeks' gestation were selected as cases. All spontaneous births occurring after 37 weeks were potential control subjects. To limit costs, some tests were performed only in selected control subjects. RESULTS: Preterm birth occurred in 207 of 5092 women (4.1%). In bivariate analysis, BV was not associated with preterm birth (OR 1.2; 95% CI 0.5 to 2.4). Vaginal douching was significantly associated with bacterial vaginosis (P < 0.05) and preterm birth (P < 0.05). On multivariate analysis, vaginal douching was no longer associated with preterm birth, but a significant association with early preterm birth < 34 weeks (OR, 6.9; 95% CI 1.7 to 28.2) and preterm birth due to preterm labour (OR 3.0; 95% CI 1.1 to 8.5) persisted after controlling for the presence of bacterial vaginosis and placental inflammation. CONCLUSION: Vaginal douching and bacterial vaginosis were not associated with spontaneous preterm birth overall. However, vaginal douching appears to be an independent and potentially modifiable risk factor for early preterm birth (32-34 weeks), although the mechanism remains unclear.