Food security is defined as the physical, social and economic accessibility of sufficient, safe and nutritious foods all the times for people to meet their dietary needs and food preferences for active and healthy life(Reference Grainger1). In 2018, the FAO report showed that approximately two billion people across the world are disposed to moderate or severe food insecurity(2). In Iran, the prevalence rate of food insecurity among households, mothers and children has been reported to be 49, 61 and 67 %, respectively, with an increasing trend during 2004–2015(Reference Behzadifar, Behzadifar and Abdi3).
It is believed that food insecurity mainly affects underprivileged societies, young children and women of the reproductive age(Reference Bhattacharya, Currie and Haider4,Reference Black, Victora and Walker5) . Ample evidence suggests that food insecurity is associated with poor health outcomes in women, including higher risk of obesity(Reference Laraia, Siega-Riz and Gundersen6), higher gestational weight gain(Reference Laraia, Siega-Riz and Gundersen6,Reference Kazemi, Masoumi and Shayan7) , anaemia(Reference Anand, Hawkes and de Souza8), depression(Reference Whitaker, Phillips and Orzol9), gestational complications (diabetes, hypertension and birth defects)(Reference Anand, Hawkes and de Souza8,Reference Seligman, Bindman and Vittinghoff10,Reference Seligman, Laraia and Kushel11) and the ultimate increase in the burden of healthcare costs on the community(2). Due to increased nutritional requirements during pregnancy, not reaching the adequate nutrient supply may increase the risk of malnutrition and consequent intrauterine growth restrictions(Reference Marangoni, Cetin and Verduci12,Reference Grieger and Clifton13) which is a significant risk factor for under-five morbidities and mortalities(Reference Woldeamanuel and Tesfaye14).
Extensive research has been focussed on food insecurity, and some of the findings have confirmed the impact of maternal food insecurity on infancy and early childhood growth(Reference Mutisya, Kandala and Ngware15,Reference Kac, Schlüssel and Pérez-Escamilla16) . In this regard, a study conducted in the USA indicated that maternal food insecurity was associated with a threefold increase in the delivery of low-birth weight infants(Reference Borders, Grobman and Amsden17). Also, living in food-insecure households severely affects the health status of children in later life as they are more prone to malnutrition in growth age. Not only may food insecurity lead to increased odds of childhood obesity(Reference Papas, Trabulsi and Dahl18) and poor cognitive performance(Reference Nejati, Nemati and Rezaei Ardani19,Reference Khorramrouz, Doustmohammadian and Eslami20) , but it also is associated with psychological distress during their transition into adulthood(Reference Heflin, Kukla-Acevedo and Darolia21).
Food insecurity is assumed to affect the growth of infants by feeding practices such as breastfeeding. As the most cost-effective approach to meeting the nutritional needs of infants and reducing childhood diseases(Reference Schanler, Krebs and Mass22), exclusive breastfeeding in the first 6 months of life is adversely affected by food insecurity through changes in the initiation and duration of breastfeeding(Reference Webb-Girard, Cherobon and Mbugua23); therefore, food insecurity seems to adversely impact infant growth variably(Reference Whitaker, Phillips and Orzol9,Reference Loopstra and Tarasuk24) . Nevertheless, only limited studies have been focussed on the effect of household food insecurity (HFI) on infant growth indices, proposing conflicting results(Reference Ahmadihoseini, Omidvar and Nematy25).
This study aimed to investigate the effects of HFI during the third trimester of pregnancy on the growth indicators of infants aged less than 6 months.
Materials and Methods
Study design, sampling and inclusion criteria
In this retrospective longitudinal study, data were collected from Sina Electronic Health System (SinaEHR®) during November 2016–March 2019. SinaEHR is an integrated health information system supervised by Mashhad University of Medical Sciences, which contains the health records of more than five million people in Khorasan Razavi province, Iran.
The current research participants were selected from 15 cities (137 healthcare centres) in Khorasan Razavi province among women 18–45 years of age who referred to the healthcare centres for routine check-ups during their third trimester of pregnancy. To maximise the target population (registered in SinaEHR), we pooled data between November 2016 and March 2019, and random sampling was used for selecting study participants. We included only women who delivered singleton infants, but mothers with a history of chronic heart diseases, anaemia, type II diabetes mellitus and gestational diabetes were excluded. In addition, we excluded infants with a history of hospitalisation during the first 6 months of life.
Sociodemographic and anthropometric measurements
The sociodemographic and anthropometric data, including age, education level, smoking during pregnancy, medical history of the mother and infant, place of residence and maternal pre-pregnancy BMI were collected using a questionnaire which was developed and applied via the SinaEHR. Anthropometric variables of infants (weight, height, and head circumference at birth, second, fourth and sixth months of life) and their mothers (weight and height) were measured by trained personnel using standard protocol.
Infants’ weight was measured with minimal clothing using a Seca scale with an accuracy of 10 g (Seca 725, Hamburg, Germany). Their height (supine position) and head circumference were also measured using a standard measuring tape with 1 mm’s accuracy. Maternal weight and height were measured with light clothing and no shoes using a Seca scale and a stadiometer (Seca 755, Hamburg, Germany) with an accuracy of 1 mm for height and 100 g for weight.
The infants who weighed less than 2500 g were considered low birthweight, and those weighing ≥4000 g were classified as large for gestational age(26). The WHO Z score system was used to classify the nutritional status, and scores less than −2 of the WHO standards for height-for-age Z scores (HAZ), weight-for-height Z scores and weight-for-age Z scores (WAZ) were considered as stunted, wasted and underweight, respectively. Regarding the weight-for-height Z score standard, infants with the Z scores between −2 and 2 were considered normal, and those with the Z score of above 2 were considered overweight. In addition, infants with the HAZ between the Z-score lines −1 and 2, WAZ between the Z-score lines of −1 and 3 were classified as normal(27).
Food insecurity measurement
HFI was measured using a standard questionnaire (Household Food Insecurity Access Scale, Household Food Insecurity Access Scale), which has been developed by Food and Nutrition Technical Assistance II project in collaboration with others(Reference Coates, Webb and Houser28,Reference Coates, Swindale and Bilinsky29) . The validity of the Persian version of this questionnaire has been confirmed by Salarkia et al., who reported the acceptable internal consistency of the scale for Iranian households with the Cronbach’s alpha of 0·95(Reference Salarkia, Abdollahi and Amini30).
The Household Food Insecurity Access Scale is composed of a set of nine items specific to an experience of food insecurity occurring within the previous 4 weeks. Endorsed a standard scoring procedure was used with 1 point for occurrence and 0 for non-occurrence. The frequency scores are within the range of 0–3, with zero indicating non-occurrence, one showing rare occurrence (once/twice in the past 4 weeks), two indicating sometimes (3–10 times in the past 4 weeks) and three for often (>10 times in the past month)(31,Reference Ballard, Coates and Swindale32) . For the purpose of this article, based on the Household Food Insecurity Access Scale total score (nine items based on the frequency score), the study population was classified as food secure (scores 0–1), mildly food insecure (scores 2–7), moderately food insecure (scores 8–14) and severely food insecure (scores 15–27).
Statistical analysis
Data analysis was performed in SPSS version 22. The continuous variables were expressed as mean and sd, and the categorical variables were expressed as frequency and percentage. Independent samples t-test was used to assess the significant difference in the Z scores between the food-secure and food-insecure subjects, and one-way repeated-measures ANOVA was applied to evaluate the significant difference in the Z scores between various measurement times. The variables were categorised, and the relative risk was calculated at 95 % CI(Reference Marangoni, Cetin and Verduci12). In addition, multinomial logistic regression was applied considering food insecurity as a factor and the covariates of maternal age, pre-pregnancy BMI, education level and parity. In all the statistical analyses, the P-value of less than 0·05 was considered significant.
Results
As shown in Fig. 1, more than 175 000 pregnant women were registered in SinaEHR during the study period, and 3474 women were randomly selected for this study. Among these women, 3252 were aged 18–45 years and had live, singleton deliveries. Notably, 202 women were excluded due to medical conditions (e.g. type II diabetes mellitus, gestational diabetes mellitus, anaemia, CHD), 342 women were excluded due to data loss and 227 cases were excluded due to neonatal hospitalisation history.
The sociodemographic characteristics and anthropometric indices of the participants are presented in Table 1. The proportion of the infants by gender was reported at 50 %. In addition, 0·8 % of the women (n 20) had smoking habits within the past 12 months upon enrolment. The majority of study population were urban (95·6 %) and outskirts residents (49·7 %). In terms of education level, 37·2 % of the participants had an academic degree, 35·4 % had a high school diploma and the others had secondary education or were illiterate. Approximately 4 % of the neonates (n 100) were low birth weight, and 4·5 % (n 110) had macrosomia (weight: >4000 g). The infants’ mean birth weight was 3237 g, the mean pre-pregnancy BMI was 26·29 kg/m2 and the mean gravidity was 2·44.
As presented in Table 2, the highest prevalence of wasting was observed at birth, which gradually decreased with infants’ growth. The comparison of the mean Z scores between the food-secure and food-insecure subjects using independent t-test (Table 3) indicated that birthweight, birth height, head circumference at birth and birth’s Z scores were not significantly different between groups, although the infants in food-secure households had lower WAZ at 6 months of age (P = 0·003) and lower HAZ at the age of 2, 4 and 6 months (P ≤ 0·001).
Z scores of HAZ, WHZ and WAZ blow −2 considered stunted, wasted and underweight, respectively; WHZ of −2 to 2 considered normal and >2 classified as overweight; HAZ of −1 to 2 and WAZ of −1 to 3 considered normal. HAZ, height-for-age Z scores; WAZ, weight-for-age Z scores; WHZ, weight-for-height Z scores.
HAZ, height-for-age Z scores; WAZ, weight-for-age Z scores; WHZ, weight-for-height Z scores.
P < 0·05 considered statistically significant.
Figure 2 shows the prevalence of food insecurity in the sample population. As illustrated herewith, approximately two-thirds of the participants were food secure, while the prevalence of mild, moderate and severe food insecurity was 23 % (n 572), 7 % (n 179) and 3 % (n 70), respectively. Table 4 shows the relative risk of HFI during pregnancy for the anthropometric characteristics of the infants. Accordingly, the children born to the women in food-insecure households were, respectively, 1·96, 2·72, and 3·87 times more likely to be stunted at birth (95 % CI 1·23, 3·12), 4 months (95 % CI 1·28, 5·77) and 6 months of age (95 % CI 1·54, 9·75) compared to the infants of the food-secure women, while they were also 1·85 times more likely to be underweight at the age of 4 months (95 % CI 1·01, 3·41). After adjustment for pre-pregnancy BMI, age, education level, parity, smoking habits and place of living, the risk of stunting remained high at birth (adjusted relative risk (aRR) = 2·01; 95 % CI 1·17, 3·46), 4 months (aRR = 3·03; 95 % CI 1·21, 7·61), and 6 months of age (aRR = 3·83; 95 % CI 1·37, 10·68) in the infants born in food-insecure households compared to those in food-secure households. On the other hand, food insecurity was not significantly associated with weight-for-height Z score and WAZ during the first 6 months of life, and no correlation was observed with birth weight. These results remained persistent after adjustment for potential confounders.
Multinomial logistic regression adjusted for maternal age, pre-pregnancy BMI, education level and parity; P < 0.05 considered statistically significant. HAZ, height-for-age Z scores; LBW, low birth weight; WAZ, weight-for-age Z scores; WHZ, weight-for-height Z scores.
Discussion
Despite the fact that HFI has been associated with several adverse health outcomes among children, there is scarce literature regarding the correlation of HFI during pregnancy with the nutritional status of children. This was the first study to determine the effects of HFI during pregnancy on infants’ growth indicators in Iran. Our findings indicated that the mean WAZ and HAZ were significantly lower in infants aged 6 months born to food-insecure mothers. HFI during pregnancy may be a risk factor for stunting in infants aged less than 6 months. However, no significant association was observed between maternal food insecurity and birth weight in this study.
Our results showed that the children of food-insecure households had lower WAZ and HAZ compared to the food-secure ones, which is consistent with the studies conducted in other low- and middle-income countries, such as South Ethiopia(Reference Betebo, Ejajo and Alemseged33), Nepal(Reference Osei, Pandey and Spiro34), Brazil(Reference Kac, Schlüssel and Pérez-Escamilla16) and Kenya(Reference Mutisya, Kandala and Ngware15). Another study conducted on children of different age groups in the same region reported that HFI was significantly correlated with the mean height of the infants aged less than 6 years(Reference Shahraki, Amirkhizi and Amirkhizi35). It seems that food insecurity could affect child development in the early stages of life through at least two mechanisms. First, food-insecure mothers have limited access to nutritionally adequate and safe food, which decreases the resources for the proper growth of their infants during pregnancy and breastfeeding(Reference Boyd and Windsor36,Reference Smith37) . Another mechanism is via the maternal emotional status, as food insecurity could be the driving force behind the disruption of maternal psychological and hormonal balance, which in turn has an adverse impact on the quality and duration of breastfeeding(Reference Saha, Tofail and Frongillo38,Reference Venu, van den Heuvel and Wong39) .
The current research’s initial finding was that HFI was only associated with child stunting, and no correlations were observed with wasting and underweight. Stunting or chronic malnutrition is often an indication of long-term nutritional deprivation. However, this is an issue of greater magnitude than underweight or wasting, more accurately reflecting the nutritional deficiencies and illnesses that occur in the most critical periods of growth and development in the early stages of life(Reference Coates, Webb and Houser28). This impaired linear growth, during the first 1000 d of life particularly, tends to remain up to adulthood, which is associated with greater risk of morbidity and mortality(Reference Black, Allen and Bhutta40), poor cognition and educational performance(Reference Berkman, Lescano and Gilman41) and nutrition-related chronic diseases in adult life(Reference Dewey and Begum42).
A growing body of evidence shows that HFI is closely linked to greater risk of stunting in early age in developing countries. A study conducted in South Ethiopia reported that children aged 6–59 months living in food-insecure households were 6·7 times more at risk of stunting than their counterparts in food-secure households(Reference Betebo, Ejajo and Alemseged33). Likewise, A cross-sectional study on 2591 children aged 0–60 months confirmed the association of HFI with increased risk of stunting even after adjusting for child, mother and household confounders(Reference Sreeramareddy, Ramakrishnareddy and Subramaniam43). In a multi-country study conducted on 800 households in eight countries, Psaki et al. found that food access insecurity was associated with a significant shift in the distribution of children’s HAZ toward lower values; the risk of stunting increased by 12 % among children from food-insecure households.
Since the growth of infants aged less than 6 months largely depends on the nutritional status of mothers through breastfeeding, the adverse conditions in this period may even deteriorate with the initiation of complementary feeding(Reference Stewart, Iannotti and Dewey44). Hence, our findings highlight this point that there is a clear need to establish early and timely preventive actions to address HFI during pregnancy and the first 6 months after birth.
In contrast to our finding regarding the lack of an association between food insecurity during pregnancy and birth weight, the study by Saeed et al. on 103 pregnant women aged 19–45 years indicated that food-insecure women were at a fivefold increased risk of delivering low birth weight neonates(Reference Saeed, Naqvi and Javed45). Moreover, a prospective cohort study of 5044 pregnant women showed that severe HFI is a risk factor for the higher rate of low birth weight(Reference Bater, Lauer and Ghosh46). The discrepancy could be because the method used in the mentioned study to assess the HFI status differed from our research. In addition, the rate of food insecurity in our population was low at baseline.
There are some limitations that should be acknowledged. Firstly, there was a lack of access to reliable data on the households’ income status and the inevitable use of the habitation area as a welfare indicator. Secondly, our study was only focussed on the first 6 months of life to clarify the linkage between food insecurity during pregnancy and the growth indicators of infants when they primarily depended on the maternal nutritional status. Therefore, further longitudinal studies are recommended on various age groups in order to confirm the associations. Notably, this was the first multi-centric study regarding the effects of HFI on the growth indicators of infants aged less than 6 months in Iran.
Conclusion
The results of this longitudinal study indicated that food insecurity during pregnancy might be a risk factor for stunting in infants aged less than 6 months. Our findings could be incorporated into nutritional interventions during and before pregnancy as HFI may have detrimental effects on children’s growth and development in the early stages of life, which could become life-long.
Acknowledgements
Acknowledgements: Hereby, we extend our gratitude to the Student Research Committee of Mashhad University of Medical Sciences for the financial support of this study (grant no: 980127). Financial support: This research was supported by grants from Student Research Committee of Mashhad University of Medical Sciences (grant no: 980 127). Conflict of interest: None declared. Authorship: K.K., G.R. designed the study, F.K.H. contributed to the implementation of the research, E.M.F. analysed the data, K.K., F.K.H., N.M. and Z.S.H. contributed to the writing of the manuscript and S.R.S. contributed to the final edit of manuscript. Ethics of human subject participation: This study was conducted in accordance with the Declaration of Helsinki and all procedures involving research study participants were approved by the Ethics Committee of Mashhad University of Medical Sciences (code: IR.MUMS.MEDICAL.REC.1398·503). Written informed consent was obtained from all subjects.