In many low- and middle-income countries, the persistence of undernutrition in children is accompanied by increasing overweight and obesity in older children, adolescents and adults. This phenomenon, known as “double-burden of malnutrition” (DBM), is emerging within countries, communities and even households [1–5]. National surveys from six low- to middle-income countries showed that 3.7–15.5 % of households were with underweight and overweight members [1], while national surveys from 42 developing countries in Africa, Asia and Latin America reported that there were 0.9–16.0 % of households with a stunted child and an overweight or obese mother [2]. This DBM phenomenon is an emergent public health problem that could lead to impaired child growth and cognitive development which in turn will dramatically increase the risk of obesity and non-communicable chronic diseases as well as impaired social and economic productivity in adulthood [6–8]. In addition, DBM profoundly complicates nutritional interventions because under- and over-nutrition have so far been treated as distinct nutritional problems when in fact they may share similar determinants [3, 9].

The DBM phenomenon is due to interaction of changes related to demographic and socio-economic status, traditional food patterns and lifestyle [10]. Studies have shown that DBM is more prevalent in urban than rural areas of developing countries, particularly among low income households, but gradually increasing in the rural areas [1, 3, 9, 11]. Factors such as household size, number of children, maternal (age, marital status, education and employment status) and child factors (age and sex) were also associated with risk of DBM households [3, 9, 11–13].

Indigenous peoples make up approximately 4.5 % (~300 million) of the global population and they are still among the poorest and most marginalized groups with poor health and social outcomes [14]. Indigenous peoples are not spared from DBM. For example, over half of the indigenous adults in Australia were overweight and obese despite the persistence of undernutrition in children (10–14 %) [15]. Likewise, the First National Survey of Indigenous Peoples in Brazil reported that about 6 % and 26 % of children were underweight and stunted, while 46.1 % of adults were overweight and obese [16, 17]. The increasing prevalence of DBM phenomenon is parallel to the rising prevalence of diabetes mellitus, hypertension and cardiovascular diseases among indigenous peoples in Brazil [17], Australia [18, 19] and Canada [20].

The indigenous peoples of Peninsular Malaysia (Orang Asli) are also experiencing a high rate of obesity, hypertension and diabetes even though they contribute to 0.7 % of the total population in Peninsular Malaysia [21]. Several studies showed that approximately 10–35 % of Orang Asli adults in Peninsular Malaysia were overweight and obese, yet more than half of the children were underweight and stunted [22–25]. In an earlier study, the prevalence of diabetes mellitus (1.3 %) and impaired glucose tolerance (10.7 %) was significantly higher among Orang Asli living in rural area than those living in the remote forest (0.0 % and 3.3 % respectively) [26]. Wan Nazaimoon & Suraiami [27] reported that 22.7 % of female Orang Asli adults in Peninsular Malaysia had metabolic syndrome, with the prevalence of diabetes mellitus and impaired fasting glucose was 8.4 % and 16.8 %, respectively. A recent study among the Orang Asli of Peninsular Malaysia showed that the prevalence of obesity, abdominal obesity, hypertension and diabetes mellitus was 16.8 %, 38.4 %, 29.6 % and 4.6 %, respectively [28].

While undernutrition still prevails in Orang Asli children [22, 25], obesity and metabolic risks are increasingly prevalent among the Orang Asli adults. This study aimed to assess the presence of double-burden of malnutrition at the community and household levels by assessing the nutritional status of Orang Asli adults and children. In addition, we further investigated the demographic and socio-economic characteristics that differentiate normal and double-burden of malnutrition households in this community.

Similar to other indigenous communities worldwide [15, 17, 25], DBM occurs in the Orang Asli community of the current study, with a high prevalence of adult overweight/obesity coexisting with a very high prevalence of child undernutrition. This DBM phenomenon could be partially due to high proportions of stunted children and short stature of adults (i.e. women) that were observed in this Orang Asli community. In Guatemala, more indigenous women were reported to have short stature compared to non-indigenous women [36] and maternal short stature was significantly associated with overweight mother-stunted child pairs [9, 11]. The intergenerational transmission of stunting from mother to child, in which poor intrauterine growth brought on by a malnourished/stunted mother is associated with low birth weight and stunting in early childhood [6–8]. Another possible explanation for DBM is that indigenous populations are undergoing nutrition transition [37], whereby they experience a shift from highly mobile lifestyle (subsistence farming and hunting-fishing-gathering) and traditional diets high in plant-source foods to westernized diets high in fat, sugars, salt and animal-source foods, and sedentary lifestyle [15, 25, 28, 38, 39]. These changes could contribute to child undernutrition and adult overweight/obesity due to impaired control of dietary energy, lower energy expenditure and a lower rate of fat oxidation [40].

At household level, we reported a higher prevalence of households with OWOBM/UWSTC (20 %) and OWOBM/STC (19 %) than OWOBM/UWC (13 %). Similarly, Lee et al. [11] reported that 21.5 % of indigenous households in Guatemala comprised of OWOBM/STC. Several other studies have reported lower prevalence of double-burden households. In poor urban area of Benin (West Africa), the prevalence of households with maternal overweight/obesity and child stunting/wasting was 16.2 % [41] while households with OWOBM/STC in rural Indonesia and Bangladesh were 11 % and 4 %, respectively [9]. Jehn & Brewis [3] showed that there were less than 10 % of households with OWOBM/STC in the middle and lower income countries, except in Egypt (12.5 %), Ghana (12.5 %), Nicaragua (12.5 %), Bolivia (15 %), Peru (16 %) and Guatemala (23 %). For households with OWOBM/UWC, the reported prevalence was lower than the prevalence reported in previous local studies [12, 13, 42]. Saibul et al. [12] reported that 25.8 % of Orang Asli households in semi-urban areas had OWOBM/UWC while 26–29.6 % of Malay households in rural and semi-urban areas were with OWOBM/UWC [13, 42]. While a study among urban poor in Haiti (14.3 %) [43] reported a similar prevalence, the Demographic and Health Surveys (DHS) database from 18 middle and lower income countries in South Asia, Africa and Latin America (0.3–5.3 %) [3] showed much lower prevalence of households with OWOBM/UWC as compared to the present study. The different prevalence of DBM across studies could be due to the different proportions of maternal overweight and child underweight/stunting across regions, countries or populations.

While several studies showed that DBM was associated with higher income and wealthier households [1, 9], others have reported that DBM was prevalent among the lower and middle income households [11, 42]. Similarly, we observed that double-burden households with OWOBM/STC and OWOBM/UWSTC were more likely to occur in lower income households (households income per capita < USD 29.01). Inadequate resources to purchase quality foods may contribute to the purchase of cheaper energy-dense and nutrient-poor foods by lower income households to maintain families’ daily dietary intakes at less cost [44]. The consumption of these foods may predispose children to poor growth and development due to inadequate important nutrients and energy, particularly if small amounts are consumed, but may provide excess calories for adults [37].

We found that the OWOBM/UWC households were more prevalent in the non-Jah Hut (Temuan and Che Wong). For Jah Hut, while there is increasing dependence on rubber/oil palm cultivation, subsistence farming and hunting-fishing-gathering are still important food activities [25]. The involvement of Jah Hut women in these activities may sustain the dependence on traditional food sources and physical activity level. Living near to the town area might explain the easier access to processed foods by the Che Wong and Temuan, although the Che Wong (mostly men) still maintains hunting-fishing-gathering activities [24, 38]. Studies have shown that Che Wong and Temuan tend to depend on relatively inexpensive processed foods (e.g. sweetened condensed milk, sugar and cooking oil) that can be purchased at the nearest sundry shops/town when income is inadequate and forest resources for food and economy dwindle due to the monsoon season [12, 24, 39].

In the present study, boys were more likely to be in DBM households with OWOBM/STC and OWOBM/UWSTC. However, several studies demonstrated that female children were more likely to be in households with OWOBM/STC [4, 9]. A plausible explanation for the inconsistent findings is sex differences in child nutritional status across countries. For example, boys were more likely to be stunted than girls in Sub-Saharan Africa [45] whereas girls tend be more stunted than boys in Bangladesh [46]. A higher stunting prevalence among boys than girls in the present sample might explain the association between DBM and male child gender. In addition, cultural influence on child feeding and caring due to power relations and social norms that perpetuate discriminatory attitudes and practices may also contribute to the sex differences in DBM [47].

Several maternal characteristics were associated with DBM, including higher maternal education and older mother. Numerous studies have reported that higher maternal education was protective against DBM [3, 9, 11, 42, 48]. However, the present study and a study in rural Bangladesh [9] showed that double-burden of OWOBM/STC was more likely to occur in households where mothers had higher education level. The conflicting findings underscore the complexities of the association between maternal education and DBM. Maternal education may not necessarily be sufficient to initiate behavioral change related to health and nutrition [49]. Lacking of health and nutrition knowledge could contribute to women being less sensitive to child and own health or less responsive to health and nutrition issues due to barriers such as food cost, accessibility and availability, lack cooking skill and equipment [44, 49]. On the other hand, our findings, together with previous studies [3, 9] showed that older mothers were associated with double-burden of OWOBM/STC. Reduced physical activity due to physical exertion decreases with age may contribute to overweight and obesity among older mothers in DBM households [10].

We showed that OWOBM/UWC was less likely to occur in households with higher number of children. In contrast, Jehn & Brewis [3] reported that having more children in the households was associated with increased risk of OWOBM/UWC. However, it should be noted that we considered children < 18 years while Jehn & Brewis [3] investigated young siblings < 5 years. Different age ranges in defining “number of children in the same household” in both studies may explain these different findings. We also found that in this sample, number of children in the household increased with household size, number of schooling children and household income (data not shown). As many of schooling Orang Asli children are staying at school hostel during school terms, the number of children living in the household is lesser which could mean more resources for family members (i.e., food, time and money) [48]. It could also mean that having many children, particularly older ones, provides extra hands to care for younger siblings or contribute to household income.

There are several limitations to the present study. The cross-sectional study design to investigate the association of demographic and socio-economic factors with DBM could not determine the cause and effect relationship. In addition the small sample size, inclusion of only three sub-tribes and specific study location could limit the generalization of study findings to indigenous peoples of Peninsular Malaysia. The small sample size could also limit additional multivariate analyses to compare DBM households with other household phenotypes and limit the number of variables that could be included in the multivariate models. The present study did not examine other important determinants of DBM such as dietary intake, physical activity, caregiving practices, cultural influences, pregnancy and birth information. Finally, as the operational definition of “double-burden of malnutrition” varied across studies, comparing the prevalence and determinants of DBM with other countries and within different contexts is inevitably difficult. Despite these limitations, the present study has provided important information on DBM in indigenous communities and its socio-demographic characteristics which could be the basis for strategies to improve health and nutrition of the Orang Asli.

DBM is a public health concern in lower income communities undergoing nutrition and socio-economic transition, and the Orang Asli population in Peninsular Malaysia is not spared. Undernutrition in early childhood needs to be addressed as underweight or/and stunted children may be predisposed to be overweight and obese in adulthood [50, 51]. The insights on demographic and socio-economic factors associated with DBM underscore the importance of having effective and culturally sensitive strategies to simultaneously address undernutrition and over-nutrition in children and adults, respectively. Poverty reduction, access to quality diet and improved health and nutrition literacy would be important and useful strategies to address the coexistence of DBM in this population.