Nutritional status of Tajik children and women: Transition towards a double burden of malnutrition.

The double burden of malnutrition, an emerging concern in developing countries, can exist at various levels: individual, household, and population. Here, we explore the nutritional status of Tajik women (15-49 years) and children (5-59 months) focusing on overweight/obesity along with undernutrition (underweight, stunting, and micronutrient deficiencies). For this, nutritional markers (haemoglobin (Hb), transferrin receptor (TfR), serum ferritin (Sf), retinol binding protein (RBP), vitamin D, serum folate, and urinary iodine), height, and weight were assessed from 2,145 women and 2,149 children. Dietary intake, weaning, and breastfeeding habits were recorded using a 24-hr recall and a questionnaire. Overweight (24.5%) and obesity (13.0%) are increasing among Tajik women compared with previous national surveys (2003 and 2009). Prevalence of iron deficiency and anaemia was 38.0% and 25.8%, respectively; 64.5% of women were iodine deficient, 46.5% vitamin A deficient, and 20.5% had insufficient folate levels. Women in rural areas had significantly lower iron status and body mass index and higher iodine intake compared with urban areas; 20.9% of children were stunted, 2.8% wasted, 6.2% underweight, 52.4% iron deficient, and 25.8% anaemic; all more prominent in rural areas. Dietary diversity was higher among urban women. Intraindividual or household double burden was not seen. In summary, double burden of malnutrition constituted an increase in overweight among women, especially in urban areas, and persisting levels of undernutrition (stunting, iron, and vitamin A deficiency), predominately in rural areas. A holistic, innovative approach is needed to improve infant and young children feeding and advise mothers to maintain an adequate diet.


| INTRODUCTION
Undernutrition derives from insufficient energy and nutrients intake to meet an individual's needs and is often associated with underweight, stunted, wasted, or suffering from micronutrient deficiencies (Maleta, 2006;World Health Organization [WHO], 2018b). Undernutrition is a widespread health problem, particularly affecting the low-and middle-income countries. Adequate maternal nutrition during pregnancy and lactation is important for the promotion of optimal health during early development of the infant. Stunting, as a result of chronic malnutrition, can only be fully reversed if addressed efficiently within the first 2 years of life (Black et al., 2008). Iron deficiency (ID) and the consequential ID anaemia (IDA) in early life can lead to poor cognitive and motor development (Sachdev, Gera, & Nestel, 2005). Iodine deficiency leads to insufficient production of thyroid hormone, which may result in foetal brain damage and impairs cognition and growth in early life (Zimmermann & Boelaert, 2015), mental function, reduced work productivity, and toxic nodular goitre in adults (Taylor et al., 2018;Zimmermann & Boelaert, 2015). Folic acid plays a major role in foetal neural tube formation (van Gool, Hirche, Lax, & De Schaepdrijver, 2018). Emerging evidence on nutrition during pregnancy and early development suggests undernutrition in utero and in early life may predispose an individual to be overweight and susceptible to the development of noncommunicable diseases (diabetes and heart disease) later in adult life (WHO, 2018a).
Paradoxically, many parts of the world are currently exposed to an overweight and obesity epidemic, with over 39% of adults being overweight worldwide, caused by an energy imbalance between consumption and expenditure (WHO, 2018c). As a consequence, noncommunicable diseases like cardiovascular diseases, diabetes, musculoskeletal disorder, and some cancers are increasing (Collins et al., 2018;Igel, Saunders, & Fins, 2018;Steele et al., 2017). This fast increase in overweight and obesity along with a slow decrease or stagnancy in undernutrition (wasting, stunting, and micronutrient deficiencies) experienced in many low-and middle-income countries (Freire, Silva-Jaramillo, Ramirez-Luzuriaga, Belmont, & Waters, 2014;Haddad, Cameron, & Barnett, 2015;WHO, 2018a) leads to the socalled double burden of malnutrition (Kosaka & Umezaki, 2017). It can exist at various levels: At individual level, for example, as overweight along with nutritional deficiencies; at household level, with underweight or stunted children and overweight adults; or at population level, with the coexistence of undernutrition and overweight.
Previous Tajik National Nutritional Surveys (TNNS) have been con-  Tajikistan et al., 2013;2018). In the 2016 TNNS, a large panel of nutritional indicators was assessed together with anthropometrics, offering the opportunity for a solid analysis of malnutrition in women and children. Here, we explore the nutritional status of children under 5 years and women of childbearing age and patterns of the double burden of malnutrition that Tajikistan is facing.

| Study population and design
This survey was based on a national cross-sectional cluster sample design using population estimates from the 2010 census extrapolated to 2016. Sampling was conducted in a two-stage approach to obtain representative data for Tajikistan (stratified by urban and rural areas) and among the four administrative regions (oblasts) of the country  Tajikistan and UNICEF, 2010), using the formula: c = (((t 2 ·p (1 − p))/m 2 )·d)/nh + 10%, where c = required number of cluster, t = confidence level at 95% (standard value of 1.96), p = estimated 50% prevalence of micronutrient deficiencies, m = margin of error at 6.5% , d = design effect of 1.75, and nh = number of household by cluster. Therefore, the necessary sample size was estimated to be 2,160 children and 2,160 women including anticipated 10% dropouts or sample loss (Tajikistan, 2010).
A total of 36 clusters per oblast were selected with the probability of selection within strata being proportional to size and clusters

Key messages
• This national survey reports the nutritional status of Tajik women of childbearing age and children under 5 years.
• There is a high prevalence of micronutrient deficiencies (iron, iodine, and vitamin A) among the study population.
• Over the last decade and along persistent undernutrition in children, a marked increase in overweight and obesity in women is observed leading to a double burden of malnutrition in the Tajik population.
• The double burden of malnutrition is mainly represented by increasing overweight and obesity in women and ongoing undernutrition in children.
• The rural population is more affected by undernutrition and the urban population by the increase of overweight.
• A multidisciplinary collaborative approach is needed to tackle the public health problem of malnutrition in the Tajik maternal and child population.
consisting of urban and rural domains proportional to their distribution in the respective oblast. In every selected cluster (village and community), 12 children between 6 and 59 months and 12 nonpregnant women of childbearing age (15-49 years) were randomly selected. This selection relied on two mechanisms, depending if the cluster was involved in the World Bank's "Poverty Diagnostics of Water Supply, Sanitation and Hygiene Condition in Tajikistan" survey (World Bank, 2017). If this was the case, the same households were visited, and if needed, a random walk was done to achieve the envisaged sample size. If the respective cluster was not part of the WB survey, a local census list was used to randomly select 20 households (UNICEF, 2016).
The survey instruments, based on the questionnaires used in

| Data collection
In November 2016, 12 teams (each including a clinician, an interviewer for the face-to-face interviews, a laboratory technician in charge of the biological samples, and a driver) collected data, blood, and urine samples. Data were collected using electronic tablets and the Open Data Kit software (https://opendatakit.org/).
The face-to-face interviews covered the topics on household characteristics, food security, child health, infants feeding, and weaning practices, as well as dietary habits/intakes of the women using a 24-hr recall questionnaire. Further, household salt was tested for its iodine content using the MBI rapid test kit (UNICEF supply cat nr: S0008193).
The weight (electronic scale mother/child, 150 kg × 100 g; UNICEF supply cat nr: S0141021) and height/length (portable measuring board; UNICEF supply cat nr: S0114530) of women and children were recorded. Capillary blood (approximately 400 μl) was collected using serum separating gel tubes (Microvette®, Sarstedt) and stored in a cooling box until centrifugation at the nearest district or oblast hospital the same day. The serum was transferred to light-protected tubes (Microtainer™, Sarstedt) and frozen at −20°C until analysis. Hb was measured on spot using a HemoCue device (HemoCue AB, Ängelholm, Sweden). The clinicians provided clinical feedback regarding the anaemia status (Hb value), and anaemic women and children were advised to attend nearby health facilities.
For the urinary iodine concentration (UIC) analysis, containers were distributed to the women or the guardians for urine collection. Urine sampling from children (6-59 months) was done, whenever possible, from the traditional cradle (Gavora, where the urine is collected in a pot under the cradle). The urine samples were transferred to a labelled 20-ml falcon tubes and also frozen at −20°C until analysis.

| Biochemical analyses
The laboratory analyses were carried out by the Research Laboratory UIC in spot urine was measured using the Pino modification of the Sandell-Kolthoff reaction with spectrophotometric detection (Caldwell, Makhmudov, Jones, & Hollowell, 2005;Pino, Fang, & Braverman, 1996). All analyses were done using nanopure grade water, and all laboratory glassware and plasticware were acid washed before use. Internal controls were used for the UIC analysis.

| Statistical analysis
Z-scores were calculated with the World Health Organization (WHO) reference for children aged 0-5 years, using the WHO Anthro software v3.2.2. Stunting was defined as a low length/height-for-age z-score (<−2 for moderate or <−3 for severe), wasting as a low weight-for-height z-score (<−2 for moderate or <−3 for severe), and underweight as low weight-for-age z-score (<−2 for moderate or <−3 for severe). The body mass index (BMI) was calculated using weight and height: BMI = kg/m 2 . Anthropometrical data of previous large surveys (TNNS 2003 and2009;MICS 2005;DHS 2012DHS , 2017 were used to show time trends.
Abbreviations: CRP, C-reactive protein; Hb, haemoglobin; ID, iron deficiency; IDA, iron deficiency anaemia;SF, serum ferritin; TfR, transferrin receptor; UIC, urinary iodine concentration. associations and differences of malnutrition indicators, with a special focus on overweight and undernutrition (underweight, stunting, wasting, and micronutrient deficiencies). Differences in prevalence were assessed using Pearson's chi-square tests and differences of absolute variables using t tests. To display nationally representative data, the demographic weight of the different oblast in their rural and urban areas was taken into account, and national data weighted accordingly. To assess double burden at household level, regressions were conducted looking into correlations of BMI of women and anthropometrical or nutritional markers of children of the same household (generalised least squares random effect model).

| Dietary intake and nutritional habits
The diet of Tajik women of childbearing age was rich in cereals (wheat, bread, rice, pasta, and biscuits), potatoes, and other roots or tubers, as well as fats and oil, vegetables, and sweets, which were consumed by over 80% of the women in the previous 24 hr. Further, over 60% reported to have eaten dairy products and meat, but only 3.9% con-

| Nutritional status of women 15-49 years
Overall, 8.0% of the women were underweight (BMI < 18.5 kg/m 2 ), 24.5% overweight (BMI ≥ 25 to >30 kg/m 2 ), and 13.0% obese (BMI ≥ The national prevalence of ID and anaemia in women was 38.0% and 25.8%, respectively, and varied between the oblasts ( fruit juices (p = .012), and less plain water (p = .003). The reported consumption of eggs was significantly higher in overweight and obese women (p = .004), whereas fat and oil consumption was lower (p < .000) when compared with normal weight women.   (Figure 4), whereas the prevalence of underweight shows a less clear trend ( Figure 5). Interestingly, GBAO and Dushanbe experience an increase in underweight and stunting in recent years.
The national prevalence of ID (52.4%) and anaemia (25.8%) in children varied between the regions and 16.9% of children have IDA (

| Intraindividual double burden
Women's BMI was negatively correlated with ID (r = −.067, p = .003) and IDA (r = −.064, p = .005), but positively with inflammation (r = .199, p < .000). No correlation between the women's BMI and vitamin A, folic acid, or iodine deficiency was found. Underweight women more frequently reported consuming iron supplements compared with overweight women (20.1% vs 13.1%, p = .009). For the children, stunting, wasting, and underweight were not correlated with any nutritional deficiency measured. Furthermore, rising prevalence of overweight and obesity (to 38% in 2016) in Tajik women of reproductive age also have to be recognized as an important public health problem.

| Household double burden
In 2016, out of three Tajik women, one was overweight or obese and one was iron deficient. According to our data, this was most probably not the same woman as no intraindividual double burden of malnutrition was identified. A study in Vietnamese women showed that micronutrient deficiencies were observed among all weights (Laillou et al., 2014), whereas in this current study, ID and anaemia were slightly more prominent in underweight women. A study in the Greater Tunis area in Tunisia reported intra-household double burden of malnutrition, most prominently anaemic children and overweight mothers (Sassi et al., 2018). In the Tajik National Nutrition Survey 2016, there was a strong positive correlation of women's BMI and children's iron status in the same household, and therefore, no evidence of a double burden at household level on this or other assessed indicators.
In Tajik women of reproductive age, BMI was positively correlated with inflammation. This could be due to the so-called chronic inflammation of obesity, deriving from stimulation of inflammatory mediators through adipose tissue and resulting pro-inflammatory state and oxidative stress, and finally an increase in CRP (Ellulu, Patimah, Khaza'ai, Rahmat, & Abed, 2017). Inflammation has been shown to decrease ferroportin, through increased production of hepcidin, and therefore reduce iron absorption, as well as the release of storage iron (Nemeth et al., 2004). This can lead to ID and finally to anaemia of inflammation (Cook, 2005). However, this effect could not (yet) be shown in this population, as the obese women indeed had a tendency for increased CRP levels but rather have a better iron status compared with the women with lower BMI. Furthermore, the better iron status in overweight and obese women could not be attributed to more frequent consumption of iron-rich foods, such as meat, beans, or green leafy vegetables, but may be the result of an overall more stable food availability throughout the year. Of note, the rate of elevated CRP and therefore inflammation was extremely high in the study population. This was partly described in the field reports by many study participants presenting with obvious infections. On the other hand, capillary blood samples are slightly more prone for haemolysis (Heenan, Lunt, Chan, & Frampton, 2017); although this was not noted during handling of the blood specimen, we would like to mention this limitation and its potential impact on increased CRP values.
The national MDD score was adequate with 80.7% of women consuming at least five out of 10 food groups over 24 hr prior to the interview. This was comparable with another study conducted in spring 2016 in Tajikistan Tajikistan et al., 2018). Wealth brings lower food insecurity and the possibility to access qualitatively better and more diverse nutrition (fresh nutrient-rich food [e.g., meat and green leafy vegetables], fortified food, and supplements). Further, working outside in the sun induces cutaneous production of vitamin D; this could be a reason for higher vitamin D values in rural areas (Jablonski & Chaplin, 2018;Webb & Holick, 1988). Then, the better iodine status in rural settings may be due to the more frequent preparation of home-made meals using iodised salt. A study in China reported people in urban areas preferred noniodised salt in recent years (Zou et al., 2014). This could not be underlined by the current study; In Tajikistan Breastfeeding has been reported to reduce the risk of a childmother pair being undernourished and obese (Oddo et al., 2012).
The average time of breastfeeding was slightly longer in rural Tajikistan (11 months) but did not differ statistically from the urban population (10 months).
The high prevalence of micronutrient deficiencies (iron, iodine, and Overall, the double burden of malnutrition not only adds a layer of complexity to tackle in a population but also gives the opportunity for actions to simultaneously address obesity and undernutrition. For example, promoting breastfeeding and reducing the infant's intake of calories without micronutrients as with breastmilk substitutes and processed complementary foods, can counteract obesity in women and children, as well as chidren's undernutrition and stunting at the same time (Dietz, 2017). Further, emphasis should be given to multidisciplinary collaborative programmes across different sectors, namely, agriculture, education, trade and economy, and health (including WASH). Differences in the rural and urban population seem to be contributing to the national double burden of malnutrition. This calls for an adapted focus on the improvement of infant and child nutrition especially in rural areas but not excluding the urban population. Furthermore, there is a need for a country-wide information and education campaign to promote a healthy lifestyle, in order to tackle the most important nutritional issues for children and women in Tajikistan.

ACKNOWLEDGMENTS
We would like to thank UNICEF for funding the study and UNICEF and MOHSP for allowing this further analysis of the TNNS results, the Research Laboratory of Preventive Medicine of the Republic of Tajikistan in Dushanbe for their efforts during the lab analysis, everyone involved in the data collection, and most of all study participants.
The information of this document expresses authors' personal views and opinions and does not necessarily represent UNICEF's position.

CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.

CONTRIBUTIONS
TBJ, LZ, MB, SK, SR, and KW designed research (project conception, development of overall research plan, and study oversight); LZ and SK conducted research (hands-on conduct of the experiments and data collection); TBJ and KW analysed data or performed statistical analysis; TBJ and KW wrote paper (only authors who made a major contribution); all others have reviewed the final content.