Prevalence and trends of coexisting forms of malnutrition and its associated factors among children aged 6–59 months in South and Southeast Asia, 1996–2022: a cross-sectional time series study

Abstract

Background

South and Southeast Asia face severe malnutrition in children under five, with coexisting forms of malnutrition (CFM) exacerbating mortality risks and posing greater challenges than isolated forms of malnutrition. We aimed to assess the prevalence, trends, and factors of CFM among children aged 6–59 months in South and Southeast Asian countries and the entire region.

Methods

We used anthropometric and hemoglobin data from 515,170 children aged 6–59 months in seven low- and middle-income South and Southeast Asian countries, based on Demographic and Health Surveys conducted between 1996 and 2022. Weighted multivariable logistic regression was performed to identify the sociodemographic factors associated with seven forms of CFM: coexistence of underweight with wasting, coexistence of underweight with stunting, coexistence of underweight with both wasting and stunting, coexistence of stunting with overweight/obesity (CSO), coexistence of anemia with overweight/obesity, coexistence of anemia with underweight, and coexistence of anemia with stunting (CAS).

Results

The overall pooled prevalence of child CFM ranged from 0.8% for CSO to 23.4% for CAS. Most countries showed a declining trend in CFM, except for Timor-Leste, India, and the Maldives. Higher maternal education and being male were associated with lower odds of CFM. Compared to children aged 6–23 months, those aged 24–59 months had a higher risk of CUS but a lower risk of CSO and CAO. Children in India had higher odds of experiencing CAU and CAS compared to those in the Maldives, Myanmar, Nepal, and Timor-Leste.

Conclusion

The coexistence of undernutrition and/or anemia in South and Southeast Asia remains a public health problem. Although the prevalence of child CFM has decreased in most countries, it remains higher in Timor-Leste and India. It is necessary to consider multi-faceted nutritional interventions for children with CFM in this region, taking into account the impact of children’s gender, age, and maternal education, to further reduce child CFM.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-025-23482-w.

Introduction

Childhood malnutrition is a leading contributor to the global burden of disease, disability, and mortality among children; however, only about a quarter of countries are on track to meet Sustainable Development Goal 2, which aims to end all forms of malnutrition by 2030 [1, 2]. Malnutrition refers to deficiencies, excesses, or imbalances in an individual’s energy and/or nutrient intake [1]. The World Health Organization (WHO) classifies malnutrition into three primary groups: undernutrition, including wasting (low weight-for-height), stunting (low height-for-age), and underweight (low weight-for-age); micronutrient-related malnutrition, involving deficiencies or excesses of vital vitamins and minerals, with iron deficiency being the most prevalent globally and anemia serving as a common indicator; and overnutrition, including overweight/obesity (high weight-for-height), and diet-related noncommunicable diseases [1]. Children are the most severely affected by malnutrition, with an estimated 200 million children under the age of five impacted globally [3]. The link between child nutrition and health risks is evident. Currently, almost 8 million children are at risk of dying from severe malnutrition without immediate treatment, and these mostly occur in low- and middle-income countries [4, 5].

In Asia, particularly in low- and middle-income countries of South and Southeast Asia, malnutrition among children under five years old is notably severe [6]. In South Asia, the prevalence of stunting, wasting, and underweight among children under five years old in 2022 was 30.5%, 14.3%, and 26%, respectively, all exceeding global averages [6, 7]. Concurrently, the prevalence of stunting, wasting, and overweight in Southeast Asia surpasses global averages, estimated at 26.4%, 7.8%, and 7.4%, respectively [6]. Furthermore, the overall prevalence of anemia among children aged 6–59 months in these regions was alarmingly high at 57.3% in 2019, significantly exceeding the WHO threshold of 40% for a severe public health concern [8, 9]. Thus, South Asia and Southeast Asia experience common challenges in child malnutrition.

Research has long treated wasting, stunting, and other forms of malnutrition separately, typically measured at one point in time. While this helps assess immediate needs, it limits the understanding of how these conditions are connected and hinders effective responses [10]. It is increasingly recognized that different forms of malnutrition share common causes, including poor dietary intake, infections, and socio-economic factors [10, 11], meaning they can coexist in individuals, a phenomenon called coexisting forms of malnutrition (CFM). First introduced in the 2014 Global Nutrition Report [12], CFM is associated with higher risks for mortality and morbidity. Children with CFM face a 2- to 8-fold higher risk of death than those with only one form of malnutrition [13, 14]. Recognizing CFM helps improve understanding and guide more effective interventions. Gausman et al. [5] analyzed three types of CFM among children under 5 years old using Demographic and Health Surveys (DHS) data from 2005 to 2018. this study showed that in eight South and Southeast Asian countries (Bangladesh, Cambodia, India, Maldives, Myanmar, Nepal, Pakistan, Timor-Leste), the prevalence of coexisting underweight with stunting (CUS), underweight with wasting (CUW), and underweight with both wasting and stunting (CUWS) ranged from 7.5 to 28.6%, 3–8.2%, and 1.6–7.1%, respectively. The prevalence of CFM exceeds global averages (CUS: 13.9%, CUW: 5%, CUWS: 4.1%), suggesting CFM remains a public health concern in the region. However, the study did not examine all forms of CFM, such as CFM related to anemia, nor did it provide an overall pooled prevalence for the region.

Notably, the trend of CFM in some countries in South and Southeast Asia is declining. A study was conducted in Pakistan, and results indicate a decline in pooled CFM, including CUS, CUW, CUWS, and the coexistence of stunting with overweight/obesity (CSO) from 30.6% in 2012 to 21.5% in 2017 among children under 5 years old [15]. Conversely, in Indonesia, the prevalence of CSO among children aged 2–5 years increased from 5.2% in 2000 to 7.2% in 2007 [16]. Evidence on trends of CFM primarily focuses on individual countries and often lacks the examination of CFM related to anemia among children under 5 years old [5, 15, 16]. Findings from some studies in some South and Southeast Asian countries, such as Pakistan and Indonesia, showed household wealth status, maternal age, education, and children’s age as associated factors [15–17], but there remains a significant gap in regional analyses of factors affecting child CFM.

In Asia, the burden of CFM remains a public health concern in some countries of South and Southeast Asia, with certain types of CFM still on the rise. However, the prevalence and contributing factors of various forms of CFM at both regional and national levels remain unclear, hindering efforts to address CFM among children effectively. Analysing data from these two subregions, which bear the highest burden of child malnutrition in Asia, is essential for identifying the prevalence and shared influencing factors of child CFM. These findings can provide a stronger evidence base for policymakers at both national and subregional levels to develop targeted strategies aimed at reducing CFM among children in South and Southeast Asia. To address this research gap, we conducted this study to (1) assess the prevalence and trends of CFM among children aged 6–59 months in low- and middle-income South and Southeast Asian countries, as well as across the overall region, from 1996 to 2022, and (2) identify the factors associated with child CFM within this region.

Methods

Data source and study design

We used data from standard DHS surveys, which consist of nationally representative cross-sectional surveys conducted in low- and middle-income countries by ICF International. This data is publicly available on the DHS Program website (https://dhsprogram.com/). The DHS used a two-stage stratified random sampling design for each survey to select enumeration areas and households randomly [18]. In the DHS Program, primary sampling units (PSUs) were selected with a probability proportional to the population size. In the second stage of sampling, on average 25 households within the PSU were randomly selected for an interview by equal probability systematic sampling. Within each sampled household, members were listed, and females eligible for a more detailed interview were identified. Typically, this was a woman between the ages of 15 and 49. However, in some surveys of the DHS Program, the interview was limited to ever-married women; in others, it was women aged 10–49 or 12–49. At the time of the interview, the height and weight were recorded for children born in the last 3 to 5 years.

To keep the scientific presentation of the study, we used the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline [19]. At the beginning, we examined the children’s recode (KR) datasets from a total of 15 South and Southeast Asian countries included in the DHS in April 2025. These KR datasets contain health and nutrition information on living children born to interviewed women within the five years preceding the survey, typically covering children aged 0–59 months. The criteria for included countries were having at least one survey with simultaneous weight, height, and hemoglobin measurements in children aged 6–59 months. Since standard DHS surveys are highly standardized across countries and survey years, the definitions and measurement methods for anthropometric indicators and anemia remain consistent. All datasets follow uniform protocols for sampling, data collection, anthropometric measurements (weight and height), and biomarker assessments (hemoglobin) [20]. This ensures that children’s weight, height, and hemoglobin values are comparable across different countries and survey years. In the standard DHS surveys, weight was measured using a digital scale, length was collected by two trained individuals using a portable measuring board, and hemoglobin concentrations were measured using capillary blood in the HemoCue^® machine [20]. Because in these DHS surveys, anthropometric data such as height and weight data were obtained from children aged 0–59 months, while hemoglobin values related to anemia were only obtained from children aged 6–59 months, we excluded children aged 0–5 months to make meaningful comparisons between different forms of child CFM. Children with missing anthropometric and biochemical variables (age, sex, height/length measurement, weight measurement, and hemoglobin measurement) were excluded. Children with anthropometric outliers were also excluded. Biologically implausible values are defined by the WHO for height-for-age z score (HAZ) as < − 6 or > 6; and weight-for-age z score (WAZ) as < − 6 or > 5; weight-for-height z score (WHZ) as < − 5 or > 5. These anthropometric outliers occur due to measurement or recording errors [21]. Thus, we excluded 8 countries where no surveys included anthropometric and hemoglobin data (Supplementary Table 1). The final sample for analysis consisted of 28 surveys conducted across seven countries from 1996 to 2022: Cambodia, Myanmar, Timor-Leste, Bangladesh, India, the Maldives, and Nepal. The prevalence assessment included data from a total of 515,170 children aged 6–59 months from seven countries. The analysis comparing the prevalence of child CFM was based on the latest available data from these seven countries.

When assessing trends of child CFM over time, the study sample was restricted to countries with at least two DHS surveys, including anthropometric or hemoglobin data. Myanmar is excluded because there is only one DHS survey. Ultimately, from 1996 to 2022, we included six countries to evaluate trends, with four countries having two or more surveys that included data on both anthropometric and hemoglobin.

The associated factors were analyzed based on the most recent surveys from these seven countries, which captured data on anthropometric and hemoglobin variables. Data with missing associated factor variables were excluded, which enables the integration of data from diverse countries, aiding in a comprehensive assessment across South and Southeast Asia. Finally, the latest surveys of five countries were selected for the analysis of associated factors from 2015 to 2022, with a sample size of 184,507 (Myanmar: 2015-16, Timor-Leste: 2016, India: 2019-21, Maldives: 2016-17, Nepal: 2022). Figure 1 shows a flow chart inclusion/exclusion of DHS surveys and participants.

Fig. 1.

Flow chart of inclusion/exclusion of DHS surveys and participants

Outcome variables

Currently, the standard DHS surveys do not provide a specific definition and data for CFM. However, the standard DHS uses the same definitions and calculation methods for individual forms of malnutrition (such as stunting, wasting, overweight, and anemia) as the WHO standards, providing data on weight, height, z-scores for height-for-age, weight-for-height, and weight-for-age, altitude-adjusted hemoglobin, and children’s age and gender [20]. Based on WHO definitions, these indicators help identify stunting, wasting, overweight, and anemia, and allow for classifying CFM in children. Child stunting was defined as HAZ <–2 standard deviation (SD); child underweight was defined as WAZ <–2 SD; child wasting was defined as WHZ <-2 SD; child overweight was defined as WHZ > 2 SD; child obesity was defined as WHZ > 3 SD; and child anemia was defined as altitude-adjusted hemoglobin < 11 g/dL as per WHO standards [21, 22].

The outcome variable, CFM was defined as a child being simultaneously affected by two or more different types of malnutrition at the time of data collection. Guided by the WHO classification of malnutrition, which includes undernutrition, micronutrient-related malnutrition, and overnutrition, we systematically reviewed existing literature and combined it with DHS data to identify seven specific types of CFM. These included: (1) CUW: coexistence of underweight and wasting; (2) CUS: coexistence of underweight and stunting; (3) CUWS: coexistence of underweight, wasting, and stunting; (4) CSO: coexistence of stunting and overweight/obesity; (5) CAO: coexistence of anemia and overweight/obesity; (6) CAU: coexistence of anemia and underweight; and (7) CAS: coexistence of anemia and stunting. Since stunting and wasting (without underweight) is theoretically implausible [23], only CUWS was included to represent this combination, and the coexistence of stunting and wasting categories was excluded in this study. These CFMs were categorized into the following forms: undernutrition combinations (CUW, CUS, CUWS); undernutrition with overnutrition (CSO); overnutrition with micronutrient deficiencies (CAO); and micronutrient deficiency with undernutrition (CAU, CAS). In summary, to CFM, undernutrition includes wasting, stunting, and underweight; micronutrient-related malnutrition is represented by anemia; and overnutrition refers exclusively to overweight/obesity.

Covariates

Factors potentially associated with CFM related to household, parental, and child characteristics were identified from previous literature. These household-level factors included residence (urban or rural), the number of living children, and the wealth index: an indicator of economic status classified into five categories: poorest, poorer, middle, richer, and richest. The Wealth index has been used as a proxy for economic status in household surveys and is computed based on household assets and possessions using a statistical technique called Principal Component Analysis in DHS surveys [24]. These parental factors included the age of mothers, maternal work status (employed or unemployed), maternal education (no education, primary, secondary, higher), and paternal education (no education, primary, secondary, higher). Some children’s factors (children’s sex, age of children, birth order number, breastfed or not) were also included as covariates.

Statistical analysis

Before data analysis, pre-processing procedures were conducted to address missing values in covariates across surveys. The imputation strategy aimed to maximize statistical power and minimize potential bias from the exclusion of incomplete socio-demographic data, particularly information related to children’s households or parents. Different imputation methods were applied based on the extent of missingness. For categorical variables with less than 10% missing data, mode imputation was used to maintain internal consistency [25]. For variables with more than 10% missing data, multiple imputation using predictive mean matching (PMM) was performed in R (version 4.3.1) with the mice package [26]. Twenty imputed datasets were generated to ensure the stability of the results and reduce the potential bias associated with single imputation. Imputation quality was assessed through visual inspection, including frequency distribution plots for mode-imputed variables and density plots for PMM-imputed variables, which showed similar distributions before and after imputation, indicating consistency and model robustness. Variables with more than 40% missing data were excluded from the analysis to reduce potential bias [27]. Imputation was limited to socio-demographic characteristics and did not involve the primary study outcomes, such as the prevalence and associated factors of CFM. Sampling weights were applied after imputation to account for the complex survey design and ensure nationally representative estimates, allowing the imputation process to address missingness independently of survey weighting.

Given that all continuous variables were non-normally distributed, the Kolmogorov–Smirnov test was used to assess their distribution. This test is robust in large samples and flexible in assessing fit against any theoretical distribution, making it more appropriate for this study than other non-parametric tests such as the Shapiro–Wilk test, which is more suitable for smaller samples (n < 50) [28, 29]. Accordingly, weighted medians and interquartile ranges (IQRs) were reported. The weighted prevalence and corresponding 95% confidence intervals (CIs) of child CFM in each country were calculated using the original sampling weights, strata, and primary sampling units provided in the datasets [30]. To examine potential factors associated with child CFM, weighted multivariable logistic regression models with country fixed effects were applied. Fixed effects were used because the number of countries was relatively small (only five countries), which helps reduce errors and improve the accuracy and robustness of the estimates [31]. These weighted multivariable logistic regression models allowed for the simultaneous evaluation of multiple predictors while adjusting for potential confounders and accounting for the complex survey design of the DHS. This approach ensured nationally representative and credible effect estimates [29, 32]. A key assumption underlying the analysis was that sociodemographic characteristics (such as child’s age, sex, parental education, household wealth, and number of children) and feeding practices (e.g., breastfeeding status) are important factors of child CFM. Based on prior evidence, these variables were included as covariates, and their associations with CFM were examined. To control for potential confounding, adjustments were made for socioeconomic status (household wealth index) and regional differences (urban/rural). In addition, parental education and maternal work status were incorporated as indirect indicators of healthcare access, reflecting variations in health literacy and utilization of health services [15, 33, 34]. Child nutrition was captured by breastfeeding status, as information on breastfeeding duration was excluded due to a high proportion of missing data (over 70% in most countries). Similarly, vaccination history was initially considered but excluded due to excessive missingness (over 90% in countries such as India and Cambodia), to maintain data quality and analytical validity. We performed a weighted logistic analysis that accounted for the stratified, two-stage cluster sampling design of DHS. Specifically, we used the design variables—sampling weights, strata, and clusters—following DHS guidelines [30]. To ensure equal contribution from each survey, regardless of sample size, we rescaled the sampling weights within each country, which was necessary to prevent larger countries with disproportionately large samples, like India or Bangladesh, from unduly influencing the results. All analyses were carried out using the survey package in R, which allowed us to incorporate these design elements and produce design-consistent estimates. We examined and reported multicollinearity among the predictor variables using variation inflation factors (VIF) to prevent statistical bias in the multivariable logistic regression model. In this study, we used “5” as a cut-off value for the maximum acceptable level of VIF [35]. Sensitivity analyses were conducted by removing country fixed effects to test the stability of the findings. P < 0.05 was initially considered statistically significant, and to control for Type I error due to multiple comparisons, Bonferroni correction was applied by dividing 0.05 by the number of comparisons (7 models), resulting in a corrected threshold of P < 0.007. Results were presented as odds ratios (ORs) with 95% CI. Data analyses were conducted on R (version 4.3.1), and figures were produced using the R statistical program’s package ggplot2 (R Foundation for Statistical Computing) or Microsoft Excel.

Ethical approval

The DHS received government permission, used informed consent, and assured respondents of confidentiality. We did not require ethics approval for this study since it involved secondary data analysis.

Results

Sociodemographic characteristics of the study sample

We included data from a total of 515,170 children across seven countries, with mothers aged 10 to 49 years (see Supplementary Table 2 for a detailed list of the survey years and eligible females by country). Among the households, 23.9% of children were in the poorest quintile and 15.9% in the richest quintile. Myanmar had the highest proportion of children in the poorest (31.1%), while Timor-Leste had the highest in the richest (19.2%). Urban households accounted for 22.8%, ranging from 18.7% in Cambodia to 28.6% in the Maldives. The median number of living children was 2 (IQR: 2–3), and the median age of mothers was 27 years (IQR: 23–32). Mothers’ employment ranged from 25.6% in Bangladesh to 67.4% in Cambodia. About 31.4% of mothers had no formal education, with Nepal having the highest proportion (53.6%), while 6.5% had higher education, the highest in the Maldives (11.3%). These children consisted of 51.1% boys, with a median age of 31 months (IQR: 18–45), and a median birth order of 2 (IQR: 1–3) (Table 1).

Table 1.

Socio-demographic characteristics of the study population¹

Country	Year	Sample size(unweighted)	Household characteristics(weighted)				Parental characteristics(weighted)						Children characteristics(weighted)
			Wealth index(%)		Urban(%)	Number of living children	Age of mother (year)	Employed mothers(%)	Maternal education(%)		Paternal education(%)		Boys (%)	Age of children (mo.)	Birth order number
			Poorest	Richest					No education	Higher	No education	Higher
Cambodia	2021-22	3421	22.9	18.6	38.5	2 (1,3)	30 (26,35)	65.0	10.3	6.6	8.6	9.6	50.4	32 (18,46)	2 (1,3)
	2014	3825	25.6	18.2	13.7	2 (1,3)	28 (24,33)	65.7	13.5	2.4	10.7	5.3	50.9	31 (18,45)	2 (1,3)
	2010	3309	26.3	16.0	15.1	2 (1,3)	28 (25,32)	69.2	19.2	1.6	11.1	4.0	52.2	32 (18,46)	2 (1,3)
	2005	3046	25.9	16.2	13.0	3 (2,4)	30 (25,35)	64.6	24.4	0.3	14.1	2.1	49.5	32 (19,45)	3 (1,4)
	2000	1466	24.4	11.1	13.2	3 (2,5)	31 (27,36)	72.4	32.7	0.1	17.5	0.6	51.0	35 (19,46)	3 (2,5)
	Pooled	15,067	25.0	16.0	18.7	2 (2,4)	29 (25,34)	67.4	20.0	2.2	12.4	4.3	50.8	32 (18,46)	2 (1,4)
Myanmar	2015-16	3460	31.1	12.1	20.9	2 (1,3)	31 (26,36)	57.3	15.4	7.2	15.1	5.1	51.3	34 (19,46)	2 (1,3)
Timor-Leste	2016	1612	19.7	18.9	27.1	3 (2,5)	30 (26,35)	32.0	24.9	9.0	21.9	13.7	52.1	33 (19,46)	3 (2,4)
	2009-10	2226	21.2	19.5	22.0	4 (2,6)	31 (26,37)	38.8	34.3	1.6	26.5	4.7	50.1	33 (20,46)	4 (2,6)
	Pooled	3838	20.4	19.2	24.6	3 (2,5)	30 (26,36)	35.4	29.6	5.3	24.2	9.2	51.1	33 (20,46)	3 (2,5)
Bangladesh	2022	3653	21.3	18.7	25.9	2(1,3)	26(22,31)	26.7	6.3	17.0	15.0	17.7	51.0	32(18,46)	2(1,3)
	2017-18	6889	21.8	18.6	26.3	2 (1,3)	25 (22,30)	42.2	7.3	15.1	14.8	16.7	52.3	32 (18,46)	2 (1,3)
	2014	6407	22.8	18.7	25.2	2 (1,3)	25 (21,29)	27.0	16.7	9.0	26.4	13.4	51.4	31 (18,46)	2 (1,3)
	2011	2232	24.6	17.1	21.5	2 (1,3)	25 (21,29)	10.1	19.4	6.6	29.7	12.4	50.9	34 (18,47)	2 (1,3)
	2007	4819	22.5	17.9	21.2	2 (1,3)	25 (21,29)	27.7	27.2	6.3	35.4	11.0	49.5	32 (19,45)	2 (1,3)
	2004	5327	25.6	16.5	19.6	2 (2,4)	25 (21,30)	18.2	38.3	4.8	40.5	9.1	50.9	32 (18,46)	2 (1,4)
	1999-00	4717	25.1	16.2	16.6	2 (2,4)	25 (21,30)	18.4	47.4	3.9	44.3	9.7	50.6	32 (18,46)	2 (1,4)
	1996-97	4237	21.6	16.0	9.1	2 (2,4)	25 (21,30)	34.8	56.1	2.3	47.6	7.1	49.7	33 (18,46)	3 (1,4)
	Pooled	38,381	23.2	17.5	20.7	2 (1,3)	25 (21,30)	25.6	27.3	8.1	31.7	12.1	50.8	32 (18,46)	2 (1,3)
India	2019-21	175,252	24.5	15.3	26.2	2 (1,3)	26 (24,30)	N/A*	21.3	15.1	N/A*	N/A*	51.7	33 (20,46)	2 (1,3)
	2015-16	200,602	25.3	14.5	27.8	2 (1,3)	26 (24,30)	N/A*	30.2	10.1	N/A*	N/A*	52.0	33 (19,46)	2 (1,3)
	2005-06	34,285	25.2	14.0	24.3	2 (2,4)	26 (23,30)	30.9	49.4	4.9	28.1	9.7	52.9	33 (19,46)	2 (1,4)
	1998-99	20,571	20.8	16.8	24.2	2 (1,3)	24 (21,28)	30.7	50.1	8.7	26.8	18.6	51.7	19 (13,28)	2 (1,4)
	Pooled	430,710	23.9	15.1	25.6	2 (2,3)	26 (23,29)	NA*	37.8	9.7	NA*	NA*	52.0	28 (17,42)	2 (1,3)
Maldives	2016-17	1879	23.4	10.0	26.5	2 (1,3)	30 (27,34)	34.9	1.3	18.3	3.3	14.2	53.2	34 (20,46)	2 (1,3)
	2009	2126	19.5	20.2	30.7	2 (1,3)	28 (25,33)	33.2	13.0	4.3	20.2	5.6	50.3	31 (17,45)	2 (1,3)
	Pooled	4005	21.4	15.1	28.6	2 (1,3)	29 (26,34)	34.1	7.1	11.3	11.3	10.2	51.7	33 (19,46)	2 (1,3)
Nepal	2022	2304	24.3	14.8	64.7	2 (1,3)	26 (23,30)	55.0	22.3	4.4	10.5	6.6	51.2	33 (20,46)	2 (1,3)
	2016	2095	20.2	12.9	52.4	2 (1,3)	26 (22,30)	52.6	36.1	12.9	15.8	17.0	52.1	32 (19,46)	2 (1,3)
	2011	2066	26.7	13.4	8.7	2 (1,3)	26 (23,30)	59.6	48.3	5.0	22.1	9.7	50.5	33 (19,45)	2 (1,3)
	2006	4645	25.8	14.4	11.3	2 (2,4)	26 (23,30)	71.5	61.2	2.1	24.6	6.7	50.9	33 (20,46)	2 (1,4)
	2001	5574	25.6	14.4	6.9	3 (2,4)	27 (23,32)	83.5	73.9	0.9	37.8	5.7	49.3	32 (19,46)	3 (2,4)
	1996	3125	25.6	13.9	6.6	3 (2,4)	26 (22,31)	77.4	79.7	0.9	38.0	9.3	50.9	19 (12,27)	3 (2,5)
	Pooled	19,809	24.7	14.0	25.1	2 (2,3)	26 (23,30)	66.6	53.6	4.4	24.8	9.2	50.8	29 (17,43)	2 (1,4)
Overall pooled		515,170	23.9	15.9	22.8	2 (2,3)	27 (23,32)	N/A*	31.4	6.5	N/A*	N/A*	51.1	31 (18,45)	2 (1,3)

Abbreviations: NA: not available. ¹ Values are presented as the weighted median (interquartile range) for continuous variables or weighted percentages (%) for categorical variables. *Date of missing exceeds 80%

Prevalence of CFM among children aged 6–59 months

Among the 515,170 children included in the pooled analysis, the prevalence of CUS was highest at 20.7% (95% CI: 20.4–21.1), ranging from 6.4% in the Maldives (2016–17) to 33.5% in Bangladesh (1996–97), as detailed in Supplementary Table 3. This was followed by CUWS at 6.1% (95% CI: 5.9–6.3), CUW at 4.9% (95% CI: 4.7–5.0), and CSO, which had the lowest prevalence at 0.8% (95% CI: 0.7–0.9).

For anemia-related CFM, 444,304 children were analyzed due to missing hemoglobin data in some surveys. CAS had the highest prevalence at 23.4% (95% CI: 22.9–23.9), ranging from 7.9% in the Maldives (2016-17) to 38.4% in India (2005-06), followed by CAU at 18.9% (95% CI: 18.4–19.4) and CAO at 0.9% (95% CI: 0.8-1.0). Figure 2 visualizes the differences, showing Timor-Leste and India with a higher prevalence of CFM than other countries.

Fig. 2.

Comparison of the prevalence (%) of coexisting forms of malnutrition among children aged 6–59 months in South and Southeast Asian countries, based on the latest available data. Circles represent prevalence estimates, and horizontal bars indicate 95% confidence intervals (CI). CUW: Prevalence of coexistence of underweight and wasting (CUW) in Timor-Leste (2016), India (2019-21), Cambodia (2021-22), Maldives (2016-17), Bangladesh (2022), Myanmar (2015-16), and Nepal (2022) with 95% CI. CUS: Prevalence of coexistence of underweight and stunting (CUS) in Timor-Leste (2016), India (2019-21), Cambodia (2021-22), Maldives (2016-17), Bangladesh (2022), Myanmar (2015-16), and Nepal (2022) with 95% CI. CUWS: Prevalence of coexistence of underweight with both wasting and stunting (CUWS) in Timor-Leste (2016), India (2019-21), Cambodia (2021-22), Maldives (2016-17), Bangladesh (2022), Myanmar (2015-16), and Nepal (2022) with 95% CI. CSO: Prevalence of coexistence of stunting with overweight/obesity (CSO) in Timor-Leste (2016), India (2019-21), Maldives (2016-17), Bangladesh (2022), Myanmar (2015-16), and Nepal (2022) with 95% CI. CAO: Prevalence of coexistence of anemia with overweight/obesity (CAO) in Timor-Leste (2016), India (2019-21), Maldives (2016-17), Myanmar (2015-16), and Nepal (2022) with 95% CI. CAU: Prevalence of coexistence of anemia with underweight (CAU) in Timor-Leste (2016), India (2019-21), Maldives (2016-17), Myanmar (2015-16), and Nepal (2022) with 95% CI. CAS: Prevalence of coexistence of anemia with stunting (CAS) in Timor-Leste (2016), India (2019-21), Maldives (2016-17), Myanmar (2015-16), and Nepal (2022) with 95% CI

Trends of CFM among children aged 6–59 months from 1996 to 2022

Generally, most of these countries experienced a decline in the prevalence of CFM among children, except in Timor-Leste, India, and the Maldives, which had either an increase or stagnation in certain types of CFM, particularly in Timor-Leste, which showed up to five types of CFM (CUW, CUWS, CSO, CAU, CAO) exhibiting this trend (Fig. 3). Timor-Leste had an increasing trend in CUW increase of 2.8%. Regarding CUS, the prevalence of CUS decreased by 8.9% (Timor-Leste), 10.6% (India), 16.7% (Cambodia), 1.2% (Maldives), and 22.1% (Bangladesh). For CUWS, Timor-Leste had an increase of 0.8%. India and the Maldives showed a decreasing trend of 5% and 0.2%, respectively. For CSO, Timor-Leste increased by 0.2%, while the Maldives decreased by 0.5%. Regarding CAO, Timor-Leste and India had an upward trend, with increases of 0.3% and 1%, respectively. For CAU, Timor-Leste remained unchanged, but India and Cambodia experienced a decline in the CAU prevalence of 10.6% and 11.9%, respectively. A downward trend was further observed in CAS prevalence in Timor-Leste (3.4%) and Cambodia (13.7%).

Fig. 3.

Trends of coexisting forms of malnutrition among children aged 6–59 months in South and Southeast Asian countries from 1996 to 2022. A: Trend in prevalence (%) of CUW; B: Trend in prevalence (%) of CUS; C: Trend in prevalence (%) of CUWS; D: Trend in prevalence (%) of CSO; E: Trend in prevalence (%) of CAO; F: Trend in prevalence (%) of CAU; G: Trend in prevalence (%) of CAS. CAO: coexistence of anemia with overweight/obesity; CAU: coexistence of anemia with underweight; CAS: coexistence of anemia with stunting; CSO: coexistence of stunting with overweight/obesity; CUS: coexistence of underweight with stunting; CUW: coexistence of underweight with wasting; CUWS: coexistence of underweight with both wasting and stunting; CSO: coexistence of stunting with overweight/obesity

Factors associated with CFM among children aged 6–59 months

Table 2 presents the factors associated with CFM among children aged 6–59 months, based on the most recent adjusted dataset used in the analysis. Maternal education was negatively associated with CAS. Boys had lower odds of CUWS (OR = 0.67, 95% CI = 0.54–0.84) than girls. Children aged 24–59 months had higher odds of CUS (OR = 1.84, 95% CI = 1.56–2.18), and lower odds of CSO (OR = 0.42, 95% CI = 0.30–0.59), and CAO (OR = 0.35, 95% CI = 0.25–0.49) compared to children aged 6–23 months. Relative to India, children in Maldives, Myanmar, Nepal, and Timor-Leste had significantly lower odds of experiencing CAU and CAS. The unadjusted results for each outcome are presented in Supplementary Table 4. After countries were excluded in the sensitivity analysis (Supplementary Table 5), the originally identified predictors remained statistically significant and showed consistent directions of association with child CFM, indicating the robustness of the main findings. In addition, some variables that were not significant in the adjusted analysis with country fixed effects, such as residential type, maternal employment status, and whether breastfeeding or not, became significantly associated with the odds of CFM. These findings suggest that country was an important variable controlling for between-country differences, but it may have also reduced the apparent effect of individual and household characteristics.

Table 2.

Adjusted associations with coexisting forms of malnutrition among children aged 6–59 months in South and Southeast Asian countries, based on the most recent data

Variable	Category	CUW	CUS	CUWS	CSO	CAO	CAU	CAS
		OR (95%CI)	OR (95%CI)	OR (95%CI)	OR (95%CI)	OR (95%CI)	OR (95%CI)	OR (95%CI)
Wealth Index	Poorest	Ref	Ref	Ref	Ref	Ref	Ref	Ref
	Poorer	1.09 (0.84–1.41)	0.88 (0.74–1.06)	0.88 (0.66–1.18)	1.02 (0.59–1.78)	1.28 (0.76–2.17)	0.98 (0.83–1.16)	0.88 (0.75–1.03)
	Middle	0.96 (0.73–1.26)	0.85 (0.70–1.02)	0.67 (0.48–0.92)	0.89 (0.47–1.66)	0.98 (0.54–1.79)	0.84 (0.69–1.02)	0.86 (0.72–1.02)
	Richer	0.83 (0.58–1.19)	0.77 (0.61–0.97)	0.95 (0.66–1.37)	0.79 (0.34–1.80)	1.01 (0.47–2.18)	0.86 (0.68–1.09)	0.82 (0.65–1.02)
	Richest	0.60 (0.39–0.92)	0.70 (0.50–0.98)	0.92 (0.55–1.56)	0.91 (0.42–1.96)	1.31 (0.57-3.00)	0.68 (0.49–0.94)	0.70 (0.52–0.93)
Residence	Rural	Ref	Ref	Ref	Ref	Ref	Ref	Ref
	Urban	0.80 (0.61–1.06)	0.96 (0.77–1.20)	0.99 (0.73–1.34)	1.38 (0.79–2.44)	1.06 (0.57–1.99)	0.77 (0.62–0.95)	0.92 (0.76–1.13)
Number of living children	1	Ref	Ref	Ref	Ref	Ref	Ref	Ref
	2	1.30 (0.83–2.05)	0.94 (0.71–1.26)	1.68 (0.91–3.10)	0.57 (0.14–2.27)	0.46 (0.14–1.52)	1.13 (0.82–1.55)	0.85 (0.64–1.14)
	≥ 3	1.25 (0.68–2.31)	1.18 (0.73–1.90)	1.72 (0.79–3.74)	0.52 (0.10–2.73)	0.26 (0.08–0.85)	1.20 (0.73–1.97)	0.90 (0.56–1.43)
Age of mother(years)	<20	Ref	Ref	Ref	Ref	Ref	Ref	Ref
	20–34	1.43 (0.67–3.06)	0.81 (0.53–1.25)	1.04 (0.47–2.29)	1.36 (0.34–5.39)	0.75 (0.25–2.25)	1.01 (0.64–1.61)	0.83 (0.58–1.19)
	≥ 35	1.48 (0.66–3.35)	0.62 (0.39–0.99)	1.03 (0.43–2.47)	0.65 (0.14–3.02)	0.60 (0.17–2.09)	0.84 (0.51–1.39)	0.65 (0.43–0.98)
Maternal work status	Unemployed	Ref	Ref	Ref	Ref	Ref	Ref	Ref
	Employed	1.01 (0.81–1.27)	0.95 (0.81–1.11)	0.78 (0.59–1.02)	0.88 (0.58–1.33)	0.89 (0.55–1.46)	0.89 (0.76–1.04)	1.05 (0.91–1.20)
Maternal education	No education	Ref	Ref	Ref	Ref	Ref	Ref	Ref
	Primary	1.25 (0.89–1.74)	0.88 (0.71–1.08)	0.90 (0.63–1.30)	0.38 (0.19–0.74) *	0.50 (0.28–0.88)	0.87 (0.71–1.07)	0.74 (0.62–0.89) *
	Secondary	1.21 (0.89–1.65)	0.72 (0.59–0.89) *	0.69 (0.47–1.01)	0.64 (0.37–1.11)	0.76 (0.44–1.29)	0.71 (0.58–0.87) *	0.60 (0.50–0.73) *
	Higher	1.10 (0.71–1.71)	0.56 (0.39–0.80) *	0.38 (0.19–0.77) *	0.77 (0.39–1.54)	0.65 (0.28–1.55)	0.58 (0.41–0.82) *	0.44 (0.32–0.60) *
Paternal education	No education	Ref	Ref	Ref	Ref	Ref	Ref	Ref
	Primary	0.74 (0.53–1.03)	0.78 (0.61–0.99)	0.73 (0.51–1.04)	1.10 (0.57–2.15)	0.75 (0.41–1.38)	0.70 (0.57–0.87) *	0.82 (0.68–0.99)
	Secondary	0.89 (0.65–1.23)	0.78 (0.62–0.98)	0.90 (0.64–1.25)	1.19 (0.64–2.22)	1.00 (0.58–1.73)	0.66 (0.54–0.81) *	0.73 (0.60–0.88) *
	Higher	0.66 (0.41–1.08)	0.81 (0.56–1.18)	0.50 (0.28–0.91)	1.73 (0.86–3.47)	2.02 (0.96–4.23)	0.70 (0.50-1.00)	0.91 (0.66–1.25)
Sex of children	Female	Ref	Ref	Ref	Ref	Ref	Ref	Ref
	Male	0.95 (0.79–1.14)	1.12 (0.98–1.28)	0.67 (0.54–0.84) *	1.02 (0.71–1.47)	0.72 (0.50–1.02)	0.95 (0.84–1.08)	0.88 (0.78–0.99)
Age of children (months)	6–23 mo.	Ref	Ref	Ref	Ref	Ref	Ref	Ref
	24–59 mo.	0.89 (0.73–1.08)	1.84 (1.56–2.18) *	0.84 (0.66–1.06)	0.42 (0.30–0.59) *	0.35 (0.25–0.49) *	0.91 (0.78–1.07)	0.86 (0.75–0.99)
Birth order number	First	Ref	Ref	Ref	Ref	Ref	Ref	Ref
	Second	0.90 (0.59–1.37)	1.24 (0.95–1.62)	0.64 (0.37–1.09)	1.65 (0.43–6.38)	2.5 (0.76–8.24)	1.04 (0.77–1.41)	1.31 (0.99–1.73)
	Third & more	1.04 (0.62–1.76)	1.16 (0.77–1.74)	0.78 (0.40–1.54)	1.19 (0.24–5.86)	2.63 (0.78–8.86)	1.18 (0.78–1.78)	1.38 (0.91–2.09)
Ever breastfed	No	Ref	Ref	Ref	Ref	Ref	Ref	Ref
	Yes	0.75 (0.47–1.19)	1.08 (0.79–1.47)	1.19 (0.73–1.95)	1.00 (0.44–2.29)	0.68 (0.32–1.44)	1.02 (0.78–1.33)	0.96 (0.77–1.19)
Country	India	Ref	Ref	Ref	Ref	Ref	Ref	Ref
	Maldives	0.55 (0.37–0.82) *	0.42 (0.30–0.58) *	0.40 (0.22–0.70) *	0.36 (0.18–0.72) *	0.91 (0.52–1.59)	0.40 (0.30–0.53) *	0.28 (0.21–0.37) *
	Myanmar	0.39 (0.29–0.53) *	0.83 (0.69–0.99)	0.27 (0.19–0.38) *	0.45 (0.18–1.12)	0.85 (0.46–1.56)	0.50 (0.43–0.60) *	0.59 (0.51–0.69) *
	Nepal	0.40 (0.27–0.59) *	0.66 (0.51–0.85) *	0.68 (0.46–0.98)	0.00 (0.00–0.00) *	0.23 (0.08–0.66) *	0.45 (0.35–0.58) *	0.45 (0.37–0.56) *
	Timor-Leste	1.19 (0.96–1.49)	1.50 (1.19–1.89) *	1.49 (1.14–1.94) *	2.48 (1.58–3.90) *	1.36 (0.78–2.39)	0.71 (0.56–0.91) *	0.65 (0.51–0.83) *

Abbreviations: CUW: coexistence of underweight with wasting; CUS: coexistence of underweight with stunting; CUWS: coexistence of underweight with both wasting and stunting; CSO: coexistence of stunting with overweight/obesity; CAO: coexistence of anemia with overweight/obesity; CAU: coexistence of anemia with underweight; CAS: coexistence of anemia with stunting. Weighted multivariable logistic regression models with country fixed effects were used to assess the association of potential factors with child CFM. P<0.05. *Bonferroni-corrected P < 0.007

Discussion

In this cross-sectional time series study, we found the overall pooled prevalence of various forms of CFM among children aged 6–59 months in South and Southeast Asia to be: CAS (23.4%), CUS (20.7%), CAU (18.9%), CUWS (6.1%), CUW (4.9%), CAO (0.9%), and CSO (0.8%). The coexistence of undernutrition and/or anemia, particularly CAS, CUS, and CAU, remain health problems in the region. While most countries have experienced a decline in CFM, Timor-Leste, the Maldives, and India have shown that the prevalence of certain forms of CFM remains stagnant or increasing. Based on the more robust adjusted results from the weighted multivariable logistic regression analysis, higher maternal education and being male were associated with lower odds of CFM. Compared to children aged 6–23 months, those aged 24–59 months had a higher risk of CUS but a lower risk of overnutrition related to CFM (CSO and CAO). Children in India had higher odds of experiencing anemia-related CFM (CAU, CAS) compared to those in the Maldives, Myanmar, Nepal, and Timor-Leste.

The coexistence of anemia and undernutrition (CAS, CAU) remains a health problem in South and Southeast Asia. Previous research shows that anemia is a key factor in stunting and being underweight in children [36, 37], making them more vulnerable to these conditions. Additionally, anemia and undernutrition share common risk factors [38], which together may contribute to the higher prevalence of undernutrition with anemia (CAS, CAU). Meanwhile, nearly one in four children in South and Southeast Asia suffered from CUS, a higher prevalence than other forms of coexisting undernutrition (CUW: 4.9%, CUWS: 6.1%) and above the global CUS prevalence of 13.9% [5]. These findings underscore the importance of targeted prevention and intervention for CFM related to undernutrition and/or micronutrient deficiencies, particularly CAS, CUS, and CAU. Despite rising overweight prevalence in Bangladesh, Nepal, Vietnam, and the Philippines, with some countries exceeding the global 13.9% prevalence [39], CAO and CSO remain uncommon and may not yet be major public health concerns. However, given the common drivers behind undernutrition, micronutrient-related malnutrition, and overweight, multi-faceted nutritional interventions during childhood are crucial, such as micronutrient supplementation, use of complementary and therapeutic foods, and regulation of the marketing of unhealthy foods.

We found that the trend of CFM among children aged 6–59 months was declining in most South and Southeast Asian countries. However, in Timor-Leste, the Maldives, and India, the prevalence of certain forms of CFM remained higher than in other countries and showed a stagnant or increasing trend. In addition to the effects of economic development in recent years, the decline in the prevalence of child CFM in many South and Southeast Asian countries may also be closely linked to the positive impact of national and community-based nutrition programs [40]. Countries like Bangladesh and India have implemented community-based programs that successfully reduced undernutrition, while the Homestead Food Production program, introduced in Bangladesh, Nepal, and Cambodia between 2003 and 2007, improved access to micronutrient-rich foods, increased household income, and reduced childhood anemia [41, 42]. In Bangladesh, the National Food Policy (2006) and participation in the Scaling Up Nutrition movement since 2013 have supported efforts to enhance food access and nutrition [40, 43]. Additionally, programs promoting exclusive breastfeeding, complementary feeding, and vitamin A supplementation have contributed to better child nutrition [43]. These initiatives, backed by strong political commitment and collaboration with non-governmental organizations and international organizations, have significantly improved child nutrition, contributing to a reduction in child CFM.

Trends of stagnation (CAU) or even increase (CUW, CUWS, CSO, CAO) were observed in Timor-Leste and the Maldives. These findings are consistent with previous studies, which reported a high and rising prevalence of CUW and CUWS in Timor-Leste between 2009 and 2016 [5], as well as a CUW prevalence of 4.5% among children under five in the Maldives during 2016–2017, mirroring the result observed in our study [5, 44–46]. These country-specific trends may be closely linked to the more severe and prolonged reductions in rainfall and extreme drought experienced during the 2015–2016 El Niño event, particularly in Timor-Leste and the Maldives [47]. As a result, these climatic disruptions led to widespread crop failures, significant declines in agricultural productivity, and acute water shortages [48]. These environmental factors further strained the already vulnerable nutrition and public health systems, ultimately contributing to a sustained deterioration in child nutritional status [49]. Taken together, it is urgent to enhance disaster preparedness and build climate-resilient systems to safeguard child health and nutrition in the face of escalating climate threats. Compared to other countries, India showed an upward trend only in CAO. Some studies have confirmed that the prevalence of anemia and overweight/obesity among Indian children has continued to rise since 2015, thereby contributing to the growing burden of CAO [12, 50]. This trend may be partly attributed to the concentration of economic development in urban areas, minimal economic advancements for workers in the agricultural sector, persistently high rates of female illiteracy, and the shifting diets toward nutrient-poor, high-calorie processed foods [51]. Addressing underlying factors such as improving women’s education, strengthening rural economies, and enhancing child nutrition programs is essential to tackle anemia and obesity in Indian children, alongside ongoing surveillance to guide interventions. Although national-level factors like health policy changes, economic development, and environmental shocks may influence these trends, they are not captured in the DHS datasets. Future research could explore these influences using cross-country panel data or other external sources.

This study found that higher maternal education protects against CAS. Previous research also showed that maternal education is negatively correlated with child CFM types like CUW and CUWS [15, 52, 53]. In many households, mothers are primary caregivers, so their education could directly affect child nutrition. Higher maternal education is linked to better knowledge of child nutrition, healthier feeding practices, improved living environments, and greater use of healthcare services [15, 33]. These advantages enable mothers to adopt appropriate feeding strategies, recognize early signs of malnutrition, and maintain better hygiene, thereby helping to prevent multiple forms of malnutrition simultaneously. Consequently, higher maternal education may play a critical role in reducing the risk of CAS among children. It is essential to further improve women’s education levels in the region, particularly among families in impoverished areas, and to integrate nutrition-related knowledge into the school curriculum. This approach could not only help prevent malnutrition more effectively but also break the intergenerational cycle of malnutrition, ultimately improving the nutritional status of the next generation.

It was also found that children’s gender and age significantly influenced different forms of CFM. Boys showed a lower prevalence of CUWS compared to girls. A similar finding was reported in a nationally representative survey conducted in Bangladesh, where the prevalence of undernutrition among preschool-aged girls was higher than that among boys [54]. This disparity may be attributed to discriminatory practices by parents. Consistent results were also observed in Kenya, where girls under five years old, due to persistently lower food intake, were more likely than boys to experience stunting, underweight, and wasting [55]. This may partly explain why boys are less likely than girls to experience CUWS, which is primarily associated with undernutrition. In the future, efforts to reduce CFM should place greater emphasis on addressing gender-based disparities by promoting equitable caregiving practices, improving maternal education, and raising community awareness to ensure that girls receive equal access to nutrition and healthcare. Additionally, older children were found to have a higher prevalence of CUS and a lower prevalence of overweight-related CFM (CSO and CAO). A previous cross-sectional study on child CFM in Pakistan showed that children over the age of one are at a higher risk for rapid increases in various forms of undernutrition, including CFM [15]. This may be due to a shift in parental resources towards younger siblings, increasing the risk of undernutrition in older children [56]. Therefore, future interventions should focus on improving nutrition, addressing food insecurity, healthcare access, and resource distribution for older children, particularly in vulnerable communities, to reduce the risk of CFM and its impact on health.

Significant cross-country differences in the odds of CFM were observed. Compared to India, children in the Maldives, Myanmar, Nepal, and Timor-Leste had lower odds of experiencing CAU and CAS, indicating a higher risk of anemia-related CFM (CAU, CAS) among Indian children. The higher odds in India may be due to factors such as poverty, food insecurity, poor sanitation, and regional disparities in healthcare access. Despite large-scale nutrition programs like Integrated Child Development Services, gaps in implementation and coverage may limit their effectiveness [57]. Additionally, India’s large population and socioeconomic inequalities contribute to the ongoing burden of undernutrition and micronutrient deficiencies, such as stunting, underweight, and anemia, especially in vulnerable communities [58]. Some studies have also confirmed that the prevalence of anemia among Indian children has continued to rise since 2015 [12, 50]. These overlapping issues may exacerbate the burden of burden of anemia-related CFM (CAU, CAS), making the situation in India more severe than in other Asian countries. Strengthening the implementation of community programs—particularly by ensuring regular iron supplementation, promoting the use of fortified foods, and enhancing food security—is essential to reducing the prevalence of anemia and improving children’s overall nutritional status. Additionally, the inclusion or exclusion of country fixed effects altered the significance and magnitude of several predictors, suggesting that country may act as a confounder in the associations between individual- or household-level factors and child CFM. In this context, countries are likely to capture a range of unmeasured or complex contextual factors, such as differences in national health policies, socioeconomic conditions, and cultural norms, which can be interpreted as contextual or structural factors. Including country fixed effects improves the model’s ability to control for between-country variability, resulting in more conservative and potentially more reliable estimates of individual-level predictors. Conversely, excluding a country’s fixed effects, it uncovers associations that are otherwise masked by broader structural influences. These findings underscore the importance of considering individual-, household-, and country-level factors when examining determinants of child CFM across countries.

One of the strengths of this study is the simultaneous inclusion of seven types of child CFM related to anthropometric and hemoglobin data. Additionally, we utilized multi-year data to analyze relevant prevalence, trend, and associated factors based on nationally representative DHS data. This approach provides a more comprehensive insight into child CFM both regionally and in terms of content across seven low- and middle-income countries in South and Southeast Asia. However, our study also has limitations. First, due to the cross-sectional study design, we cannot prove any causal relationship between the identified factors and child CFM. Future research using longitudinal designs could help address this limitation and better clarify causal links. Secondly, this study excluded eight countries due to missing anthropometric and hemoglobin data, which may have introduced selection bias and limited the generalizability of the findings. Future studies should use mixed data sources and include more countries to improve the representativeness of the results. Third, an important limitation is the lack of data on household food security and child dietary diversity score, and some key variables, such as the child’s vaccination history and breastfeeding duration, had high proportions of missing data, limiting their inclusion in the analysis. Despite adjusting for a range of sociodemographic factors, limited control of confounders restricts our understanding of the complex factors of child CFM. Although the standard DHS surveys followed consistent protocols for measuring anthropometry and anemia, variations in data availability and potential measurement errors, such as differences in equipment, training, or timing, may have affected the results. However, measurement validity and reliability have improved over time due to better standardization, enhanced training, and stronger quality control in DHS. These improvements are likely to reduce bias in recent surveys, resulting in more accurate estimates of child nutrition. Any remaining errors are likely random and unlikely to have significantly impacted the overall conclusions. Additionally, differences in sample characteristics, such as the age distribution of child mothers and regional representation, may affect the comparability of results. However, only 11 children were born to mothers aged 10–14 years in this study, and this had a negligible impact on the prevalence, trends, and factors of child CFM. We recommend incorporating the assessment of CFM into the DHS program and other national nutrition programs to better understand the status of children experiencing various types of CFM across different regions worldwide.

Conclusion

In the South and Southeast Asia region, the prevalence of different types of child CFM differed greatly, and the coexistence of undernutrition and/or anemia, namely CAS, CUS, and CAU, remains a public health problem. While most countries have shown a downward trend in child CFM, it remains higher in Timor-Leste and India. Maternal education levels and children’s gender, and age were closely related to child CFM. To effectively address child CFM, it is essential to develop multifaceted intervention strategies based on children’s age and gender characteristics, while fully considering their developmental needs and differences in nutritional intake. These strategies may include addressing food insecurity, micronutrient supplementation, using complementary and therapeutic foods, and regulating the marketing of unhealthy foods. Additionally, efforts should focus on improving women’s education levels, particularly among families in impoverished areas, and integrating nutrition-related knowledge into school curricula to help break the intergenerational cycle of malnutrition.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

CAO

Coexistence of anemia with overweight/obesity

CAU

Coexistence of anemia with underweight

CAS

Coexistence of anemia with stunting

CFM

Coexisting forms of malnutrition

CSO

Coexistence of stunting with overweight/obesity

CUS

Coexistence of underweight with stunting

CUW

Coexistence of underweight with wasting

CUWS

Coexistence of underweight with both wasting and stunting

CI

Confidence interval

DHS

Demographic and Health Surveys

HAZ

Height-for-age z score

OR

Odds ratio

PMM

Predictive mean matching

PSU

Primary sampling unit

IQR

Interquartile range

WAZ

Weight-for-age z score

WHZ

Weight-for-height z score

Author contributions

The authors’ responsibilities were as follows—WG, SC, YS: designed the study; WG, SC: analyzed the data; WG: principal investigator and wrote the first draft of the manuscript; WG, SC, MZ, YS: participated in the interpretation of the data, reviewed the manuscript for intellectual content, and read and approved the final version of the manuscript.

Funding

This work was supported by JST SPRING, Grant Number JPMJSP2132.

Data availability

The data used in this article were available on http://dhsprogram.com.

Declarations

Ethical approval and consent to participate

The DHS received government permission, used informed consent, and assured respondents of confidentiality. Since this study involved secondary data analysis, we did not require ethics approval.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.