Background
In Ethiopia, recent analyses revealed that most of the millennium development goals were achieved [
1], but not the goal for reducing maternal mortality. Although maternal mortality significantly declined, it is still high, especially in rural areas of Ethiopia [
1,
2], with maternal nutritional deficiencies being one of the underlying causes [
3,
4]. It was estimated that for every maternal death, 1000 women suffer from stunting and/or anemia [
5]. These nutritional disorders act synergistically with multiple other factors leading to nutrition-related adverse pregnancy outcomes [
5,
6]. Low hemoglobin concentrations, for instance, are associated with postpartum hemorrhage, the leading cause of maternal death [
7], due to its attenuating effect on uterine contraction [
8,
9]. Apart from its impact on maternal morbidity and mortality, taking into account that the mother and her fetus or newborn are an inseparable dyad, nutritional deficiencies of mothers can cause undernutrition to fetuses and children [
5,
10‐
12]. This in turn can affect their survival and well-being even later during adulthood [
13,
14]. Thus, ensuring maternal nutritional well-being plays a critical role for preventing morbidity and mortality of mothers and their children [
6].
Maternal undernutrition, however, has remained one of the most serious public health problems to Ethiopian mothers [
2,
15‐
17]. Even though the level of maternal short stature (height less than 145 cm) reported in 2000 was reduced by 41.7% in 2016 in Tigray regional state, there was almost no reduction in the burden of maternal undernutrition [
2,
15]. Added to this existing challenge is the paradoxically rising phenomenon of maternal overweight and obesity both nationally and in the regional state, making the efforts of addressing maternal nutritional problems even more complex [
2,
18]. Hence, unless maternal undernutrition is alleviated, the health problems resulting from both under- and over-nutrition, which were associated with an increasing burden of non-communicable diseases and intergenerational transmission of poverty [
14,
19‐
21], will be an increasing challenge to public health in the future.
Family planning, antenatal care visit, in-facility delivery, skilled birth attendance and preventing nutritional deficiencies are key inputs to enhancing the survival and well-being of women of reproductive age and their children [
22‐
24]. In contrast, morbidity and mortality of adults from chronic diseases could impose high and regressive costs that could significantly cause household poverty [
25‐
29], thereby influencing maternal nutritional status.
This study was carried out to examine factors associated with undernutrition among lactating mothers using epi-demographic, socio-economic and agricultural datasets from Kilte Awlaelo-Health and Demographic Surveillance Site (KA-HDSS), eastern Tigray, Ethiopia.
Discussion
This study presents evidence on the level and associated factors of maternal undernutrition in eastern part of Tigray, Ethiopia. The mean MUAC (± SD) of the participants was 23.0 ± 1.6 with an overall prevalence of maternal undernutrition (MUAC< 23 cm) of 38% and severe undernutrition (MUAC< 21 cm) with 6.6%. This study used high quality socio-economic, agro-ecologic and epidemiologic data from community-based surveys implemented in a health and demographic surveillance system platform. Our analysis reveals associations of household histories of adult morbidity and mortality, diverse food crop production, maternal health-seeking practice, housing and environmental factors index with maternal undernutrition.
A study from South Eastern Tigray reported a mean maternal MUAC value of 23.2 cm, which is similar to our finding (MUAC 23.0 cm) [
17]. However, the level of severe maternal undernutrition in the cited study area in 2011 was about twice as high as in our study, with a magnitude of 13% vs. 6.6% respectively. Betemariam et al. found even a 24% prevalence of severe maternal undernutrition in Bale Zone of Ethiopia in 2013 [
51]. This is nearly four-fold higher than the burden reported in our study, but the study population in that study was from a community that had a marginal agricultural production with cyclic food insecurity, and this may partially explain the observed difference. Similarly, a study based on a large representative survey sample from two Gojjam zones of the Amhara region found 52.9% maternal undernutrition (MUAC< 23 cm), a level much higher than the current finding [
52]. The difference in the implementation period of the cited research works, agro-ecologic and cultural variations of the study areas might have contributed to the observed differences. The relatively lower burden of maternal undernutrition in the current study could partly be explained by the fact that the study area is known for being a hot-spot for various health and developmental interventions [
53‐
55]. Nonetheless, the prevalence of maternal undernutrition is still high in our study community.
Morbidity and mortality of household members, particularly from chronic diseases and in developing country settings, could predispose households into a poverty trap [
25,
29,
47]. Our data potentially hinted that mothers living in households that experienced adult mortality attributed to chronic diseases had increased risk of undernutrition, compared to those living in households without history of adult mortality. As shown in Fig.
2, the association of households’ experience of chronic diseases attributed adult mortality with severe maternal undernutrition, however, is strong. In our previous work, however, we identified that nutritional insecurity of children aged 6 to 23 months did not vary by households’ history of adult mortality from chronic diseases [
56]. This might be explained by the possibility of a buffering effect made by the mothers on the caloric intake of children, which in turn may have resulted in their increased maternal wasting [
57]. Therefore, the current finding supports our proposed research hypothesis that the extent of maternal undernutrition, among mothers who lived in households that experienced adult mortality from chronic diseases, may be higher than among those mothers who did not live in such households. In general, chronic diseases are characterized by long duration of illness, with high out-of-pocket medical expenditure. This would more negatively impact nutrition security of poor and uninsured patients, and affected households could fall into a poverty trap [
27,
58‐
60]. Loss of income, due to lower employment rate of the affected family members or chronic illness and mortality of the affected adults, is also another unavoidable negative consequence imperiling the economic welfare of households [
58,
61‐
63]. Our data suggest that the synergistic negative economic impact of such household-level shocks might ultimately result in undernourishment of mothers, who were living in the households that experienced adult mortality attributed to chronic diseases. Most notably, a study from Ethiopia found a decline in dietary diversity and increase in mean dependency ratio following prime age adult mortality in poor households, regardless of the sex and position of the deceased adult [
64]. In a rural setting of South Africa, household food security was affected by adult mortality, particularly by the death of a male wage-earner [
65]. A survey conducted 3 years after identification of households affected by adult illness or mortality from HIV/AIDS, in comparison to the non-affected households, were found to have less production of food crops [
63]. Furthermore, longitudinal studies linked the effect of adult mortality on the well-being of older household members; accordingly, a sharp drop in body mass index in the short term (possibly related grief and depression) and an increased probability of acute illnesses (explained by increased working hours in the field) in the long-term, were reported [
62,
66]. We assume that this effect could plausibly be more intense for the lactating mothers, considering the cultural context in the study community that the mothers are the pillars and main caregivers to family members in addition to bearing the different laborious out-door responsibilities, especially if the husband is deceased. While this assumption needs to be tested in independent studies, we may conclude that lactating mothers living in households that experienced adult death due to chronic diseases are vulnerable to undernutrition. Thus, we strongly recommend that those lactating mothers, and their children, need targeted nutritional screening with a subsequent intervention for those who are already undernourished or found to be at a high risk of undernutrition.
Mortality of adult household member, especially if the deceased one is of prime-age, is associated with adverse income and assets shocks. Illness or mortality of male household head has been shown to lead to a lower crop production, severe impacts on farm production and livestock assets [
63,
67]. A study conducted in a district proximal to our study community reported that lower farmland size and not cultivating maize were associated with severe undernutrition (MUAC< 21 cm) of lactating mothers [
17]. Our observation, of a 28% lower risk of being undernourished for lactating mothers living in households which produced diversified food crops accords with these research findings. Another study reported a 27% reduction in total cultivated farmland following death of male head or spouse [
67]. In our study, the mean household farmland size was significantly lower among the undernourished mothers than those who were not (mean difference was 0.28 ha; T
calc = 4.6189,
p < 0.001). Additionally, HIV/AIDS-related adult mortality is associated with poor agricultural and resource management, such as watershed and soil conservation, diminished care to household family members, mostly felt by women, change in crop mix and lower capacity to ensure food security [
67,
68]. This may impede the capacity to cultivate nutritious and diverse food crops and result in undernourishment of the lactating mothers.
We could show that there was a 14% lower risk of maternal undernutrition given the condition that the lactating mother had achieved a good maternal health seeking practice. The nutritional benefit of utilizing maternal health service has also been demonstrated by other studies [
17,
51,
69]. Access to such infrastructural services, especially in rural areas, are clear determinants of maternal and child health [
70,
71]. Our study revealed that better housing and environmental factor index scores are negatively associated with maternal undernutrition. This composite index was computed based on a number of relevant public health measures such as quality of housing materials, media access and cleanliness of cooking fuel, availability of environmental health services like access to improved water and latrine services. Media exposure increases maternal awareness, and so could positively influence health service utilization, such as antenatal care service [
72], which in turn may increase the mothers’ nutritional knowledge. Poor access to water and sanitation services were shown to exert the burden of water collection for the mother and reduces time for mother-child interaction, endangering nutritional well-being of both the mothers and the children [
71,
73]. Given that the lactation period is already characterized by higher energy and nutrient requirements of the mother [
74‐
76], the workload to fetch water may further increase energy requirements and enhance the risk of undernutrition. Our data indicate that the levels of moderate and severe maternal undernutrition were higher for mothers who did not have access to improved water relative to those who had access to improved water (X
2calc = 20.8,
p < 0.001), which is in line with earlier research showing the protective effect of improved water and handwashing on maternal undernutrition [
77,
78].
Our study poses some limitations. Firstly, the surveillance site, source of the study data, was established not only with the aim of generating scientific health and demographic information, but also to serve as a platform for the implementation of various researches. Due to its proximity to the regional city and thus to Mekelle University, the study area has also been one of the sites for the community-based training program (CBTP) of the College of Health Sciences of the University. In this practical attachment program teams of health sciences and medical students are routinely involved in assessing the general health status of the community, designing, and implementing interventions for sets of identified and prioritized public health problems. This has been the tradition for many years to equip graduating students with various medical and public health skills. Because of these exposures, we may assume that the frequently exposed part of the study site’s population could have a better health literacy level and healthy behavior compared to other communities. If this assumption holds true, the strength of the association of adult mortality from chronic CoD with maternal undernutrition might have been attenuated. However, we do not have the data to evaluate if this bias is induced to the estimates. Secondly, the association of maternal undernutrition with adult mortality attributed to chronic CoD is only at aggregate level and does not give detailed information of each of the chronic causes of adult death. Thirdly, the degree of maternal undernutrition may vary by the length of duration since the occurrence of adult death and the current analyses did not take into account this probable source of variation. Notwithstanding these limitations, our study is based on extensive and high-quality data, and its findings could be relevant to maternal nutrition and to the public health actors in general.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.