Observational studies have linked elevated homocysteine to vascular conditions. Folate intake has been associated with lower homocysteine concentration, although randomised controlled trials of folic acid supplementation to decrease the incidence of vascular conditions have been inconclusive. We investigated determinants of maternal homocysteine during pregnancy, particularly in a folic acid-fortified population.

Data were from the Ottawa and Kingston Birth Cohort of 8085 participants. We used multivariable regression analyses to identify factors associated with maternal homocysteine, adjusted for gestational age at bloodwork. Continuous factors were modelled using restricted cubic splines. A subgroup analysis examined the modifying effect of MTHFR 677C>T genotype on folate, in determining homocysteine concentration.

Participants were recruited in Ottawa and Kingston, Canada, from 2002 to 2009.

Women were recruited when presenting for prenatal care in the early second trimester.

In 7587 participants, factors significantly associated with higher homocysteine concentration were nulliparous, smoking and chronic hypertension, while factors significantly associated with lower homocysteine concentration were non-Caucasian race, history of a placenta-mediated complication and folic acid supplementation. Maternal age and BMI demonstrated U-shaped associations. Folic acid supplementation of >1 mg/d during pregnancy did not substantially increase folate concentration. In the subgroup analysis, MTHFR 677C>T modified the effect of folate status on homocysteine concentration.

We identified determinants of maternal homocysteine relevant to the lowering of homocysteine in the post-folic acid fortification era, characterised by folate-replete populations. A focus on periconceptional folic acid supplementation and improving health status may form an effective approach to lower homocysteine.

To investigate key risk factors associated with undernutrition in the first few years of life.

A cross-sectional household survey was conducted in January 2018 collecting anthropometric data and other information on household, caregiver and child characteristics. Crude and adjusted odds ratios were calculated to assess the association of these characteristics with stunting and underweight outcomes.

Kitui and Machakos counties in south-east Kenya.

Caregivers and their children aged 0–23 months in 967 beneficiary households of the Government of Kenya’s cash for orphans and vulnerable children (CT-OVC) social protection scheme.

Twenty-three per cent of the 1004 children with anthropometric data were stunted, 10 % were underweight and 6 % experienced wasting. The strongest predictors of stunting and underweight were being in the second year of life and being born with a low birth weight. Residing in a poor household and having more than one child under 2 years of age in the household were also significant risk factors for being underweight. Although 43 % of children did not receive the minimal acceptable diet, this was not a significant factor associated with undernutrition. When age was removed as a covariate in children aged 12–23 months, being male resulted in a significantly higher risk of being stunted.

While only 9 % of children were born with a low birth weight, these were four to five times more likely to be stunted and underweight, suggesting that preventive measures during pregnancy could have significant nutrition and health benefits for young children in this study area.

The objective of this study was to assess the prospective association between diet quality, as well as a 6-year change in diet quality, and risk of incident CVD and diabetes in a community-based population.

We used Cox regression models to estimate the prospective association between diet quality, assessed using the Healthy Eating Index (HEI)-2015 and the Alternative HEI (AHEI)-2010 scores, as well as change in diet quality, and incident CVD and diabetes.

The ARIC Study recruited 15 792 black and white men and women (45–64 years) from four US communities.

We included 10 808 study participants who reported usual dietary intake via FFQ at visit 1 (1987–1989) and who had not developed CVD, diabetes, or cancer at baseline.

Overall, 3070 participants developed CVD (median follow-up of 26 years) and 3452 developed diabetes (median follow-up of 22 years) after visit 1. Higher diet score at the initial visit was associated with a significantly lower risk of CVD (HR per 10 % higher HEI-2015 diet quality score: 0·90 (95 % CI: 0·86, 0·95) and HR per 10 % higher AHEI-2010 diet quality score: 0·96 (95 % CI: 0·93, 0·99)). We did not observe a significant association between initial diet score and incident diabetes. There were no significant associations between change in diet score and CVD or diabetes risk in the overall study population.

Higher diet quality assessed using HEI-2015 and AHEI-2010 was strongly associated with lower CVD risk but not diabetes risk within a middle-aged, community-based US population.

To cope with the pressure of modern life, consumer demand for convenience foods has increased in the last decades. The current study set out to compare the costs of buying industrially processed dishes and of preparing them at home.

Direct purchase costs of industrially processed dishes frequently consumed in France (n 19) and of the ingredients needed for their home-prepared counterparts (n 86) were collected from four major food retailers’ websites in Montpellier, France. Mean prices and energy density were calculated for four portions. Costs related to energy used by cooking appliances and time spent preparing dishes were further estimated.

Montpellier, France.

Not applicable.

Based on the costs of ingredients and energy used for cooking, dishes prepared at home cost less (–0·60 € per four portions, P < 0.001) than industrially processed dishes, but when the cost of time was taken into account, the industrially processed dishes were much cheaper (–5·34 € per four portions, P < 0.001) than their home-prepared counterparts. There was no difference in energy density between industrially processed and home-prepared dishes.

Our findings suggest that industrially processed dishes are more profitable to consumers when the cost of time for preparing dishes at home is valued. Given the ever greater demands of everyday life, more account should be taken of the additional cost to consumers of the time they spend preparing meals at home.

To identify factors influencing dietary behaviours in urban food environments in Africa and identify areas for future research.

We systematically reviewed published/grey literature (protocol CRD4201706893). Findings were compiled into a map using a socio-ecological model on four environmental levels: individual, social, physical and macro.

Urban food environments in Africa.

Studies involving adolescents and adults (11–70 years, male/female).

Thirty-nine studies were included (six adolescent, fifteen adolescent/adult combined and eighteen adult). Quantitative methods were most common (twenty-eight quantitative, nine qualitative and two mixed methods). Studies were from fifteen African countries. Seventy-seven factors influencing dietary behaviours were identified, with two-thirds at the individual level (45/77). Factors in the social (11/77), physical (12/77) and macro (9/77) environments were investigated less. Individual-level factors that specifically emerged for adolescents included self-esteem, body satisfaction, dieting, spoken language, school attendance, gender, body composition, pubertal development, BMI and fat mass. Studies involving adolescents investigated social environment-level factors more, for example, sharing food with friends. The physical food environment was more commonly explored in adults, for example, convenience/availability of food. Macro-level factors associated with dietary behaviours were food/drink advertising, religion and food prices. Factors associated with dietary behaviour were broadly similar for men and women.

The dominance of studies exploring individual-level factors suggests a need for research to explore how social, physical and macro-level environments drive dietary behaviours of adolescents and adults in urban Africa. More studies are needed for adolescents and men, and studies widening the geographical scope to encompass all African countries.

To examine differences in eating and physical activity behaviours among ethnic groups in Queensland, Australia, and differences in those behaviours due to the duration of residency in Australia.

Cross-sectional study using baseline data collected for the Living Well Multicultural–Lifestyle Modification Program between October 2014 and June 2017.

Culturally and linguistically diverse communities (CALD), including Afghani, Somali, Burmese, Pacific and South Sea Islander, Sri Lankan, Sudanese and Vietnamese, living in Queensland, Australia.

People were recruited if they were ≥18 years old and living in the targeted CALD communities.

Burmese/Vietnamese, on average, had better eating scores in line with Australian dietary guidelines, compared with Afghani/Arabic-speaking (difference = 2·05 points, 95 % CI 1·39, 2·72), Somali/Sudanese (difference = 1·53 points, 95 % CI 0·79, 2·28) and Pacific Islander (difference = 1·46 points, 95 % CI 0·79, 2·13). Association between ethnicity and meeting the physical activity guideline was not significant. Those who stayed in Australia longer than a year were less likely to meet the physical activity guideline than those staying <1 year (OR = 0·51, 95 % CI 0·31, 0·84). There was no significant association between duration of residency in Australia and eating scores.

Eating behaviours were significantly different among the ethnic groups in Queensland with Burmese/Vietnamese and Sri Lankan/Bhutanese having the healthiest diets. All ethnic groups were less likely to meet the physical activity guideline compared with the general Australian population. People with duration of residency of at least 1 year in Australia were less likely to meet the physical activity guideline compared with those who had shorter stays.

To investigate whether vitamin D status predicts weight gain or increase in waist circumference during the 11-year follow-up in general adult population.

A population-based longitudinal study.

The study was conducted using data from the nationally representative Health 2000/2011 Survey. The analyses were based on regression models adjusted for sociodemographic and lifestyle factors.

Weight, waist circumference and vitamin D status (serum 25-hydroxyvitamin D concentration analysed with radioimmunoassay) were measured from 2924 participants aged 30–64 years at baseline.

In men, low vitamin D status at baseline predicted ≥10 % increase in waist circumference during the follow-up when adjusted for age only (OR for sufficient v. deficient S-25(OH)D 0·41; 95 % CI 0·25, 0·67; P for trend <0·01), but the association with weight gain was only borderline significant. After adjustment for potential confounders, low vitamin D status remained a significant predictor of increase in waist circumference, but the association with weight gain was further attenuated. In women, vitamin D status at baseline did not predict weight gain or increase in waist circumference.

Our results suggest that vitamin D insufficiency may be a risk factor of abdominal obesity among men but not among women. In men, it may also increase the risk of weight gain. Further studies are required to confirm these findings and examine potential mechanisms behind them. There is also a possibility that vitamin D is a biomarker of healthy lifestyle rather than an independent risk factor for obesity.

To describe the Na concentration of pre-packaged foods available in Hong Kong.

The Na concentrations (mg/100 g or mg/100 ml or per serving) of all pre-packaged foods available for sale in major supermarket chains in Hong Kong were obtained from the 2017 Hong Kong FoodSwitch database. Median and interquartile range (IQR) of Na concentration for different food groups and the proportion of foods and beverages considered low and high Na (<120 mg/100 g or mg/100 ml and >600 mg/100 g or mg/100 ml, respectively) were determined.

Hong Kong.

Not applicable.

We analysed 11 518 pre-packaged products. ‘Fruit and vegetables (including table salt)’ had the highest variability in Na concentration ranging from 0 to 39 000 mg/100 g, followed by ‘sauces, dressings, spreads and dips’ ranging from 0 to 34 130. The latter also had the highest median Na concentration (mg/100 g or mg/100 ml) at 1180 (IQR 446–3520), followed by meat and meat products (median 800, IQR 632–1068) and snack foods (median 650, IQR 453–926). Fish and fish products (median 531, 364–791) and meat and meat products (median 444, IQR 351–593) had the highest Na concentration per serving. Overall, 46·7 and 26·7 % of products were low and high in Na, respectively.

Our results can serve as a baseline for food supply interventions in Hong Kong. We have identified several food groups as priority areas for reformulation, demonstrating the potential of such initiatives to improve the healthiness of the food supply in Hong Kong.

To examine changes in sales of highly processed foods, including infant formulas, in countries joining free trade agreements (FTAs) with the US.

Annual country-level data for food and beverage sales come from Euromonitor International. Analyses are conducted in a comparative interrupted time-series (CITS) framework using multivariate random-effects linear models, adjusted for key confounders: gross domestic product (GDP) per capita, percent of the population living in urban areas and female labor force participation rate. Memberships in other FTAs and investment treaties are also explored as possible confounders.

Changes are assessed between 2002 and 2016.

Ten countries joining US FTAs are compared with eleven countries without US FTAs in force; countries are matched on national income level, world region and World Trade Organization membership.

After countries join a US FTA, sales are estimated to increase by: 0·89 (95 % CI 0·16, 1·6; P = 0·016) kg per capita per annum for ultra-processed products, 0·81 (95 % CI 0·47, 1·1; P < 0·001) kg per capita per annum for processed culinary ingredients and 0·17 (95 % CI 0·052, 0·29; P = 0·005) kg per capita under age 5 per annum for baby food. No significant change is estimated for minimally processed foods. In statistical models, large unexplained variations in country-specific trends suggest additional unmeasured country-level factors also impact sales trends following entry into US FTAs.

These findings strongly support the conclusion that joining US FTAs can contribute to detrimental changes in national dietary consumption that increase population risk of non-communicable diseases.

The present study aimed to identify whether discretionary food consumption declined in an intervention focused primarily on promoting fruit and vegetable consumption. We also aimed to identify potential mediators explaining intervention effects on discretionary food consumption.

Secondary analysis of data from the ShopSmart study, a randomised controlled trial involving a 6-month intervention promoting fruit and vegetable consumption. Linear regression models examined intervention effects on discretionary food consumption at intervention completion (T2). A half-longitudinal mediator analyses was performed to examine the potential mediating effect of personal and environmental factors on the association between the intervention effects and discretionary food consumption. Indirect (mediated) effects were tested by the product of coefficients method with bootstrapped se using Andrew Hayes’ PROCESS macro for SPSS.

Women were recruited via the Coles FlyBuys loyalty card database in socio-economically disadvantaged suburbs of Melbourne, Australia.

Analyses included 225 women (116 intervention and 109 control).

Compared with controls, intervention participants consumed fewer discretionary foods at T2, after adjusting for key confounders (B = −0·194, 95 % CI −0·378, −0·010 servings/d; P = 0·039). While some mediators were associated with the outcome (taste, outcome expectancies, self-efficacy, time constraints), there was no evidence that they mediated intervention effects.

The study demonstrated that a behavioural intervention promoting fruit and vegetable consumption among socio-economically disadvantaged participants was effective in reducing discretionary food intake. Although specific mediators were not identified, researchers should continue searching for mechanisms by which interventions have an effect to guide future programme design.

To gain more comprehensive understanding of metabolic syndrome (Mets) among in general Chinese population.

Cross-sectional study. Mets was defined by three widely accepted definitions including modified Adults Treatment Panel (ATP) III criteria, International Diabetes Federation (IDF) criteria and harmonized definition. Risk factors were evaluated by using multivariate logistic regression.

Nineteen rural villages in northeast China.

The survey was conducted in September 2017 and May 2018 on 10 926 individuals.

According to modified ATP III criteria, IDF criteria and harmonised definition, the overall prevalence of Mets was 41·3 % (95 % CI 40·3, 42·2), 34·2 % (95 % CI 33·2, 35·1) and 44·1 % (95 % CI 43·1, 45·1), respectively. Females had a higher prevalence, and elevated blood pressure was the most frequent. Age, female sex, non-peasant worker, higher BMI and lower-annual income were independent risk factors of Mets in all three definitions (all ps < 0·05). Based on modified ATP III criteria and harmonised definition, heavy drinking was positively correlated with Mets. In contrast, former drinking was inversely associated with Mets.

Mets is highly prevalent in rural areas of northeast China. Its independent risk factors include higher age, female sex, non-peasantry worker, higher BMI and lower-annual income. Modified ATP III criteria and harmonised definition may be superior definitions of Mets.

The present study analysed and compared the nutritional qualities of private brand (PB) and national brand (NB) ready meal (RM) products in South Korea.

Data on RM products were collected via a food information database (EatSight) between February and May of 2017. A total of ninety-one PB products from a leading discount store chain in South Korea and 415 NB products were selected and divided into four main categories of ‘refrigerated soups/stews’, ‘refrigerated noodles’, ‘frozen rice/noodles’ and ‘frozen dumplings’. Nutritional information was extracted for each product and analyses were performed to compare nutritional differences.

South Korea.

A total of 506 RM products.

PB products explained the highest percentage of refrigerated RM products while NB products occupied the highest percentage of frozen ones. Notably, main PB and NB products showed evident nutritional problems. Among PB products, refrigerated soups/stews were significantly higher in energy, fat, Na, saturated fat and trans-fat. Likewise, in the case of NB products, frozen rice/noodles were significantly higher in Na and saturated fat. Particularly, overall Na contents were high. When the Korean guidance on traffic-light labelling was applied, a mere 0·2 and 5·9 % of total RM products showed all four and three green lights, respectively. Meanwhile, those with no green lights accounted for 42·3 %.

Although it is difficult to draw a definitive conclusion on the comparative nutritional qualities due to the lack of consistency in the results, the study suggests that there is room to improve the nutritional qualities of overall RM products.

Decades of marginalization have led Palestinian refugees living in Lebanon to experience multigenerational poverty and food insecurity. The Healthy Kitchens, Healthy Children programme implemented and examined the impact of a two-pronged intervention that employed women through community kitchens to deliver a subsidized healthy daily school snack to elementary-school children in Palestinian refugee camps in Lebanon. We describe the rationale, study design, theorized impact pathways, and discuss lessons learned.

The programme was quasi-experimental. We conducted formative and process evaluation of both components of the intervention to elucidate the pathways to programme impact.

Palestinian refugee camps in Lebanon.

Thirty-three women participated in the kitchens and provided subsidized snacks to 714 children.

Snacks were healthy, traditional Palestinian recipes designed by women and a nutritionist. Participation fluctuated but eventually increased after modifying the meals to ensure acceptability by children. The main challenges to sustainability related to the need for subsidization of the meals and the lack of school policies around the regulation of sales of school food, which together led to fluctuations in programme participation.

The study provides lessons learned on the potential of this model to improve the human capital of two generations of protracted refugees. The availability of schools as a constant market for these social enterprises offers an opportunity for sustainable livelihood generation and food security gains. Challenges to sustainability remain and could be addressed through social (subsidies to support the programme) and structural (policies to restrict unhealthy food sales) measures.

To determine the prevalence and predictors of hypocalcaemia in under-five children (1–59 months) hospitalised with severe acute malnutrition (SAM).

A cross-sectional study was designed to determine the prevalence of hypocalcaemia among children hospitalised with SAM. Serum Ca and 25-hydroxycholecalciferol (25-(OH)D) were estimated. Hypocalcaemia was defined as serum Ca (albumin-adjusted) <2·12 mmol/l. To identify the clinical predictors of hypocalcaemia, a logistic regression model was constructed taking hypocalcaemia as a dependent variable, and sociodemographic and clinical variables as independent variables.

A tertiary care hospital in Delhi, between November 2017 and April 2019.

One-hundred and fifty children (1–59 months) hospitalised with SAM were enrolled.

Hypocalcaemia was documented in thirty-nine (26 %) children hospitalised with SAM, the prevalence being comparable between children aged <6 months (11/41, 26·8 %) and those between 6 and 59 months (28/109, 25·7 %) (P = 0·887). Vitamin D deficiency (serum 25-(OH)D <30 nmol/l) and clinical rickets were observed in ninety-eight (65·3 %) and sixty-three (42 %) children, respectively. Hypocalcaemia occurred more frequently in severely malnourished children with clinical rickets (OR 6·6, 95 % CI 2·54, 17·15, P < 0·001), abdominal distension (OR 4·5, 95 % CI 1·39, 14·54, P = 0·012) and sepsis (OR 2·6, 95 % CI 1·00, 6·57, P = 0·050).

Rickets and hypocalcaemia are common in children with SAM. Routine supplementation of vitamin D should be considered for severely malnourished children. Ca may be empirically prescribed to severely malnourished children with clinical rickets, abdominal distension and/or sepsis.

Drinking water instead of beverages with added sugar can help prevent obesity and cavities and promote overall health. Children spend much of their day in school, where they have variable access to drinking water. In 2010, federal and state law required California public schools to provide free potable water to students in areas where meals are served and/or eaten. The current study aims to identify factors associated with an excellent drinking water culture in schools.

A qualitative assessment of barriers and facilitators to providing excellent water quality and access in a purposive sample of California schools. In-depth interviews with key informants were conducted using a snowball sampling approach, after which data were analysed using both inductive and deductive methods.

California public elementary, middle/junior and high schools.

Knowledgeable individuals involved in initiatives related to school drinking water accessibility, quality or education at each selected school.

Thirty-four interviewees participated across fifteen schools. Six themes emerged as prominent facilitators to a school’s success in providing excellent water access to students: active and engaged champions, school culture and policy, coordination between groups, community influences, available resources and environmentalism.

While policy is an important step for achieving minimum standards, resources and interest in promoting excellence in drinking water access and quality can vary among schools. Ensuring that schools have dedicated staff committed to advancing student health and promoting the benefits of water programs that are more salient to schools could help reduce disparities in drinking water excellence across schools.

To examine associations between economic residential segregation and prevalence of healthy and unhealthy eating markers.

Cross-sectional. A stratified sample was selected in a three-stage process. Prevalence of eating markers and their 95 % CI were estimated according to economic residential segregation: high (most segregated); medium (integrated) and low (less segregated or integrated). Segregation was measured at the census tract and assessed using the Getis–Ord local $G_i^{\rm{\ast}}$ statistic based on the proportion of heads of household in a neighbourhood earning a monthly income of 0–3 minimum wages. Binary logistic regression using generalized estimating equations were used to model the associations.

Belo Horizonte, Brazil.

Adults (n 1301) residing in the geographical environment (178 census tracts) of ten units of the Brazilian primary-care service known as the Health Academy Program.

Of the 1301 participants, 27·7 % lived in highly segregated neighbourhoods, where prevalence of regular consumption of fruit was lower compared with more affluent areas (34·6 v. 53·2 %, respectively). Likewise, regular consumption of vegetables (70·1 v. 87·6 %), fish (23·6 v. 42·3 %) and replacement of lunch or dinner with snacks (0·8 v. 4·7 %) were lower in comparison to more affluent areas. In contrast, regular consumption of beans was higher (91·0 v. 79·5 %). The associations of high-segregated neighbourhood with consumption of vegetables (OR = 0·62; 95 % CI 0·39, 0·98) and beans (OR = 1·85; 95 % CI 1·07, 3·19) remained significant after adjustments.

Economic residential segregation was associated with healthy eating markers even after adjustments for individual-level factors and perceived food environment.

To (i) determine the proportion of deaths from CVD that could be avoided in both rural and metropolitan Australia if public health recommendations were met; (ii) assess the impact on the rural CVD mortality; and (iii) determine if policy priorities should be different by rurality for CVD prevention.

A macro-simulation modelling study of population data. Population, risk factor and CVD death data stratified by rurality were analysed using the Preventable Risk Integrated Model. The baseline scenario was the current risk factor levels (including physical activity, smoking, diet and alcohol). The counterfactual scenario was the population levels of these risk factors expected if public health recommendations were met.

Metropolitan and rural Australia.

Rural- and metropolitan-dwelling adults in Australia.

Both populations would experience similar relative declines in the proportion of deaths from CVD. A total of 14 892 deaths from CVD would be avoided annually; with similar declines in the proportions of deaths by rurality. Critically, the order of policy priorities for public health recommendation attainment would differ by rurality CVD prevention, with addressing fat intakes being a higher priority in rural areas.

Achieving public health recommendations in Australia would result in large declines in CVD mortality. Despite declines in overall CVD mortality under this scenario, an inequality in CVD burden would persist for rural populations. The order of risk factor priorities would differ by rurality.

To (i) identify and synthesise findings from interventions to improve the dietary intake, physical activity and weight status of children aged 0–6 years attending family day care services; and (ii) assess the impact of interventions on family day care environments, intervention cost and adverse outcomes.

Medline in Process, PsycINFO, ERIC, Embase, CINAHL, CENTRAL and Scopus databases were searched in March 2019. Studies were included if they (i) evaluated an intervention to improve the diet, physical activity and/or weight of children aged 0–6 years; (ii) were delivered in family day care services; (iii) targeted child diet, physical activity and/or weight; and (iv) used a parallel control group design. Screening was undertaken by two reviewers with disagreements resolved by a third reviewer.

Family day care services, also known as family childcare homes.

Children aged 0–6 years attending family day care services.

In total, 8977 titles were retrieved, and 199 full-texts reviewed. No studies met the inclusion criteria for the primary outcome; however, two studies reporting on the secondary outcome of family day care environments were included. The 4-year community-wide obesity prevention programme and the 12-month train-the-trainer programme both reported statistically significant improvements in the healthy eating and physical activity environments of family day care, compared to cross-sectional state-average control groups.

Findings highlight few existing interventions in family day care services and a need for high-quality controlled trials to identify effective interventions to improve children’s diet, activity and weight in this setting.

Evidence on the relationship between maternal Hb concentration and spontaneous abortion (SA) risk is limited and conflicting. The purpose of the study was to evaluate whether maternal preconception anaemia or high Hb concentration is associated with risk of SA.

A population-based cohort study established between 2013 and 2017.

Local maternal and child care service centres in each county.

In total, 3 971 428 women aged 20–49 years, who participated in National Free Pre-Pregnancy Checkups Project from 2013 to 2016 and successfully got pregnant before 2017 in rural China.

A total of 101 700 (2·56 %) women were recorded having SA, with highest SA rate in women with severe anaemia (4·58 %). Compared with women with Hb of 110–149 g/l, the multivariable-adjusted OR for SA was 1·52 (95 % CI: 1·25, 1·86) for women with Hb < 70 g/l, 0·92 (0·84, 1·01) for 70–99 g/l, 0·80 (0·77, 0·83) for 100–109 g/l, 1·11 (1·08, 1·15) for 150–159 g/l, 1·12 (1·04, 1·20) for 160–169 g/l and 1·02 (0·93, 1·12) for ≥ 170 g/l, respectively. An approximate U-shaped curve for the risk of SA with Hb concentrations was observed when Hb concentrations less than 145 g/l, above which the association plateaued (Pnon-linear < 0·001).

Severe anaemia and high Hb concentration before pregnancy were associated with an increased risk of SA. Women with mild anaemia prior to pregnancy had lower risk of SA. Underlying mechanisms need to be further studied.