Area, time, and data source

This study was conducted in Sambas District, which has total area about 6.395,70 km², comprising 4.36% of the total area of West Kalimantan Province. Sambas District is classified as a rural and border area of Indonesia and Malaysia. Study participants resided in 17 of the 19 subdistricts within 30 villages throughout the district.

The data for this study were derived from two cross-sectional studies in 2016 and 2017. The data management for 2016 was conducted from April–August 2016 over several successive phases: (1) preparation from April 11–24; (2) data collection from April 25–May 11; (3) data entry, data cleaning, and analysis in June 2016; and (4) the report writing phase through August 2016.

This follow-up study in 2017 was conducted from April–June 2017 over several phases: (1) preparation from April 3–14; (2) data collection May 17–April 2; (3) data entry and data cleaning and analysis in May 2017; and (4) the report writing phase in late May–June 8, 2017.

Study design

This study was designed as a repeated cross-sectional study of child birth outcomes and nutritional status related to the conditions during pregnancy determined from a previous study in Sambas District. The report focused on the prenatal, perinatal, and postnatal factors and their association with stunting in infants age 0–11 months. This study is part of a larger study, “Association Between Maternal Nutritional Status During Pregnancy and Birth Outcome and Nutritional Status of Their Infant Age 0–11 Months in Sambas District, West Kalimantan Province” conducted by South East Asian Ministers of Education Organization (SEAMEO) Regional Center for Food and Nutrition (RECFON) and Department of Nutrition, Faculty of Medicine, Universitas Indonesia.

Ethics statement

This study was conducted after receiving ethical clearance from the Ethical Committee of Faculty of Medicine of the University of Indonesia (Number 329/UN2.F1/ETIK/2016, April 25, 2016). Permission from the local government (provincial and district) as well as the local health authority were also obtained prior to conducting this study. To follow the basic principles of bioethical research, the survey team explained the details of the study, emphasized the voluntary nature of the participation, collected data safely, and ensured confidentiality of the data. Written consent was obtained from study participants.

Study population and subjects

The target population for the study was mothers who joined the previous study in April 2016 during their pregnancy. The study participants were both the mothers of caregivers of the infants and the infants whose mothers participated in the previous study. Inclusion criteria were Sambas District residence; being either the mother or the caregiver of the infant or being the infant whose mother was interviewed in the previous study; and willingness to participate in the study. Exclusion criteria were current residence of the potential participant in an area outside Sambas District and inability to be contacted.

Sample size calculation and sampling procedure

Data collection procedure

Variables and operational definition

In this study, the prenatal and postnatal variables were considered explanatory variables, for which the corresponding coding definitions are shown in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Variables and operational definition.

https://doi.org/10.1371/journal.pone.0254662.t001

Anthropometric measurements for children and mothers

To assess the nutritional status of the child, anthropometric data were collected by measuring the child’s weight and length. Child length was measured using a length board, and weight was measured by weight scale (SECA). Date of birth was input by asking the mother and checking in the Health of Mother and Child Book. Both length and weight were assessed twice, and then the average was noted. The procedure for performing anthropometric measurements was used as specified by WHO. The measurements were conducted from May 17–April 2, 2017.

To assess the nutritional status of the mother, anthropometric data were collected by measuring weight and height. Height was measured using length board, and mother’s weight were measure by weight scale (SECA), and date of birth were input by asking the mother and cross-referencing with the family certificate. Both the mother’s height and weight were assessed twice, and then the average was recorded. The measurements were conducted in the previous study from April 25–May 11, 2016.

Statistical analysis

Descriptive analysis was performed on participant’s characteristics and variables included in this study. Categorical data were applied to display the frequency and percentage. Numeric data were also reported as mean scores and standard deviations to determine the collected data were normally distributed. If the collected data were not normally distributed, they were presented as median, minimum, maximum value and interquartile range. To analyze normal distribution of data, the Kolmogorov–Smirnov Z test was used.

With regard to the association between several variables and stunting, bivariate analysis was performed using the chi-squared or Fisher’s exact test. The pre- and postnatal determinants of stunting were determined to be contributing factors after assessment in multivariable analysis using multiple logistic regressions. Potential confounders were first selected on the basis of their known association with child stunting in developing countries. For all tests, p < 0.05 was considered statistically significant. The odds ratios with 95% CIs were calculated in order to assess the adjusted risk of independent variables. All analyses were performed using SPSS (version 26.0, IBM, Armonk, NY, USA). For missing data, complete case analysis (listwise deletion) was used. With this approach, a case is dropped from an analysis because it has a missing value in at least one of the specified variables. The analysis is only run-on cases which have a complete set of data.

Bivariate and multivariable analysis for association of pre- and postnatal determinants of stunting

As shown in Tables 3–5, bivariate analysis results showed that short maternal stature, preterm birth, low birth weight (LBW), children age ≥6 months, diarrhea, complete immunization coverage, income per capita below the median, and poor knowledge of infant and young child feeding were associated with stunting in infants.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Child characteristics and stunting.

https://doi.org/10.1371/journal.pone.0254662.t003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Mother/caregiver characteristics and stunting.

https://doi.org/10.1371/journal.pone.0254662.t004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Household characteristics and stunting.

https://doi.org/10.1371/journal.pone.0254662.t005

Multivariable logistics regression was adjusted for sociodemographic variables including sex and age of the children. The results of multiple logistic regression models for prenatal and postnatal factors revealed that prenatal factors which was maternal stature and post-natal factors including preterm birth and complete immunization coverage were significantly related to stunting.

In the model without LBW included, children with preterm birth had higher odds of being stunted (aOR = 5.10; 95% CI: 1.04–24.8) compared to children with term birth (born alive after 37 weeks of pregnancy are completed). The odds of being stunted were higher (aOR = 3.28; 95% CI: 1.61–6.65) among children with incomplete immunization coverage compared to children who had complete immunization coverage (aOR = 2.65; 95% CI: 1.14–6.18).

Children with LBW had higher odds of being stunted (adjusted odds ratio [aOR], 4.12; 95% confidence interval [CI], 1.71–9.91) compared with children with normal birth weight. The odds of being stunted were higher (aOR, 3.28; 95% CI, 1.61–6.65) among children with diarrhea in the past 2 weeks compared with children without diarrhea. Children who had incomplete immunization coverage were likely to be stunted compared with children who had complete immunization coverage (aOR, 2.43; 95% CI, 1.03–5.76).

Findings

This study revealed that prenatal and postnatal factors were associated with infant stunting in Sambas District, Indonesia. In model where LBW wasn’t included in the multivariable analysis, preterm birth, short maternal stature, and complete immunization coverage were the key factors for the 20.8% prevalence of infant stunting in this rural area, whereas prenatal factors, including maternal CED risk during pregnancy, maternal anaemia during pregnancy, maternal dietary diversity, knowledge during pregnancy and knowledge on IYCF were not significantly associated with infant stunting in this study.

In model without LBW, it was found that preterm birth was associated with stunting. This study in line with a few studies that examined the association between preterm birth and stunting, wasting and underweight in developing countries. A study conducted in Brazil reported that at the age of 1 year, the odds ratios for stunting among the late preterm infants was 2.35 [10]. These estimates are lower than our estimates (Table 3). A cohort study in rural Malawi among 840 infants found that preterm infants were at a significantly higher risk of being wasted and underweight compared with the term infants [11].

Preterm birth and intrauterine growth restrictions (IUGR) are the two underlying biological factors leading to LBW. A baby who suffered from intra-uterine growth restriction (IUGR) as a fetus was effectively born malnourished. Growth deficits since birth seemed to significantly increase the risk of stunting up until 2 years of life and contribute to a short stature as well as increasing the risk of developing chronic diseases later in life. Link between low birth weight (LBW) and child malnutrition could possibly be described by the increased vulnerability of children with LBW to infections such as diarrhea and lower respiratory infections and the increased risk of complications including sleep apnea, jaundice, anemia, chronic lung disorders, fatigue and loss of appetite compared to children with normal birth weights [12].

In multivariable analysis (model without LBW), we found that short maternal stature had association with stunting. This study result is in line with prior longitudinal study by Mendes MSF, Villamor E, and Melendez GV using the 2006 Brazilian Demographic Health Survey data that maternal height was positively associated to children’s HAZ values. Children having mothers <145 cm tall have lower HAZ than those having ≥145 cm height mother. In addition, the results of those analysis also indicated the risk of child stunting by maternal height categories (<145, 145–149, 150–154, 155–159 and ≥160 cm) were 2.95 (1.51; 5.77), 2.29 (1.33; 3.93), 1.09 (0.63; 1.87), and 0.89 (0.45; 1.77), respectively. This suggests that mothers with the height of <145 cm were at greater risk of having stunted children (2.95 times) than other maternal height categories [13].

Based on a literature review and meta-analyzes conducted in 2012, maternal stature is associated with LBW owing to both genetic and environmental factors. Women with short stature may have passed relevant genes to their fetus. Short stature is also associated with smaller uterus and pelvic size, thus contributing to a smaller fetus size. An example of environmental factor that is associated with short stature is low socioeconomic status, which keeps women in a cycle of malnourishment [14].

In model where LBW included, LBW and diarrhea were associated with stunting in this study. LBW occurred due to intrauterine growth restriction during pregnancy, in which the fetus was effectively born malnourished. Growth deficits since birth seemed to significantly increase the risk of stunting up until 2 years of life and contributed to a short stature as well as increasing the risk of developing chronic diseases later in life. The link between LBW and childhood malnutrition may be described by the increased vulnerability of children with LBW to infections such as diarrhea and lower respiratory infections and the increased risk of complications, including sleep apnea, jaundice, anemia, chronic lung disorders, fatigue, and loss of appetite, compared with children with normal birth weights [12].

LBW is closely associated with fetal and neonatal morbidity and mortality [15]. Compared with the national data collected in 2013, the proportion of LBW in the present study is much lower. In 2013, the Indonesian Ministry of Health stated that the prevalence of LBW was 10.2% [4]. Although the percentage of both LBW and diarrhea in Sambas District was lower than that on the national level, it was still a significant concern because LBW has negative impacts on infants. At population level, the proportion of infants with a LBW is an indicator of a multifaceted public health problem that includes long-term maternal malnutrition, ill health, hard work, and poor health care in pregnancy [9].

The present study showed that diarrhea was associated with stunting. The incidence of diarrhea has been shown to lead to greater risk of child stunting. Based on the findings of a study published in 2014 by the Australian Government, infection can contribute to undernutrition, which results in children being more vulnerable to infectious diseases. This finding is consistent with the framework mentioned in the literature review, which explains that when a child has an infectious disease, the malabsorption of nutrients and metabolism disturbances are also observed [16]. Another study found that frequent diarrhea in children can lead to malnutrition [17].

This study found that that >10% of infants age 0–11 months in Sambas District had acute diarrhea. This prevalence of diarrhea is higher than the rate 3.5% reported for the national prevalence in 2013 among children age <5 years [4]; however, it is important to note that the data from the Indonesian government represent an older age group. According to other studies, the prevalence of diarrhea was the highest among children age 6–11 months, remained at a high level among children age 1 year, and then decreased as the child aged [18]. A decreased prevalence of diarrhea with increased age may be due to the fact that the immune system becomes stronger with age. The peak prevalence of diarrhea occurs at age of 1–1.5 years because child starts moving around the house and begins to eat food other than breast milk [18]. Symptoms of diarrhea in children at an older age are less severe compared with those at a younger age [19]. In terms of the cause, diarrhea is significantly associated with poor sanitation [18].

In the model of logistic regression in two models (with and without LBW), complete immunization coverage of children ≥6 months was associated with stunting. This finding similar with previous studies that showed complete coverage of child immunization explains the lower prevalence of both wasting and stunting. These studies also provided evidence that a childhood vaccination program led to improved child growth in a setting in which most children were undernourished according to WHO child growth standards [20, 21]. Moreover, a study in Papua New Guinea found that children with stunting (6 children, 13.6%) were less likely to be fully vaccinated compared with children without stunting (10 children, 16.4%) (OR, 0.805; 95% CI, 0.269–2.409) [22, 23].

The study findings indicate the need for integrated interventions to reduce the stunting rate in Indonesia. Interventions should use life-cycle approaches to address various factors from the prenatal to the postnatal determinants. A strong required effort to improve the basic needs for mothers and their children. Low level of maternal knowledge related to nutrition, lack of basic immunization coverage, and the high prevalence of diarrhea should be addressed by the local government. Health and nutrition education targeting mothers to improve healthy supplementary feeding, immunization for infants and children, and management of diarrhea among infants should be programmed and advocated with a multilevel approach involving health workers at the level of the Integrated Health Center, village midwives, and the local public health authority. Empowering women through education and economics will help improve a child’s nutritional status, which can lead to better childcare practices, and improve the household economic status, which is essential for better food intake, less exposure to infections, and better use of health care services.

Strengths and limitations

The strength of the present study is that it is a prospective study of participants enrolled in a previous study. As a result, the present study was able to capture a wide spectrum of the determinants of stunting and could assess the prenatal, perinatal, and infant periods. This study also covers 17 out of the 19 representative sub-districts in Sambas District, thus representing nearly the entire population of Sambas District. To reduce the possibility of bias and to ensure data quality, the author team trained of the researchers, conducted pretesting, used qualified enumerators, conducted daily calibration, and performed data cleaning/field checking in the field and basecamp and before data analysis. A limitation is that the statistical power of the study may have been reduced by the lost-to-follow-up rate of 10.6% (59 of 559 participants).