Discussion

This study, involving data from past five demographic and health surveys over 13 years, found that the prevalence of undernutrition is declining, but the prevalence of being overweight/obese is increasing among the ever-married adolescent girls in Bangladesh. Although the proportion of being underweight is decreasing, almost one-fourth of the ever-married adolescent girls had a BMI less than 18.5 kg/m² in 2017.

Our results indicate that both under- and over-nutrition are extant among the ever-married adolescent girls in Bangladesh. We observed that the prevalence of underweight among adolescent girls is decreasing, although the burden is still substantial. On the other hand, the prevalence of overweight/obesity is increasing significantly in the country. This finding is consistent with the studies conducted in neighbouring countries such as India, Pakistan, Nepal, and Myanmar (IIPS, Reference Iips2017; Macro, 2006). Studies showed that co-existence of both underweight and overweight/obesity – commonly referred to double burden of malnutrition (DBM) – is evident in low- and middle-income countries (LMICs), and Bangladesh is not an exception because of the nutritional transitions as well as changes in the food habits over the last couple of years (Das et al., Reference Das, Fahim, Islam, Biswas, Mahfuz and Ahmed2019; Kapoor and Anand, Reference Kapoor and Anand2002; Ramirez-Zea et al., Reference Ramirez-Zea, Kroker-Lobos, Close-Fernandez and Kanter2014; Sekiyama et al., Reference Sekiyama, Jiang, Gunawan, Dewanti, Honda, Shimizu-Furusawa, Abdoellah and Watanabe2015). It is evident that LMICs such as Vietnam, Indonesia, and Ghana are also experiencing nutritional transitions and suffering from the DBM (Doku and Neupane, Reference Doku and Neupane2015; Khan and Khoi, Reference Khan and Khoi2008; Winkvist et al., Reference Winkvist, Nurdiati, Stenlund and Hakimi2000). In India, the rate of overweight and obesity has increased from 2.4% to 4.2% within a time period of 10 years (2006–2016) among the adolescent girls (15–19 years), whereas the rate of undernutrition has declined from 46.8% to 41.9% in the same population during the same duration (IIPS, Reference Iips2017). The rate of overnutrition has been increased from 6.7% to 24.7%, and the prevalence of underweight decreased from 20.3% to 18.6% among the adolescent girls in Pakistan during 2013–2018 (Macro, 2006). The prevalence of overweight/obesity was 4.5% among the adolescent girls of Myanmar in 2016 (ICF, 2017). In Nepal, the proportion of being overweight/obesity among the same age population has increased from 2.1 to 3.3 during 2006–2016, and the prevalence of underweight in the same group has reduced from 30.3% to 25.8% from 2011 to 2018 (Macro, 2006). Perhaps, rapid urbanisation, economic development, consumption of less plant-based diets, more refined and energy-dense foods, and decreased physical activity are the factors responsible for the increasing trend of overnutrition and reduction of undernutrition among the adolescent girls in the South Asian countries including Bangladesh (Das et al., Reference Das, Fahim, Islam, Biswas, Mahfuz and Ahmed2019; El Kishawi et al., Reference El Kishawi, Soo, Abed and Muda2016).

In this study, we observed that education, wealth index, and parity were the factors found to be associated with being underweight. Additionally, adolescent girls from younger age group were more likely to be underweight. Previous study conducted in Bangladesh reported that poor dietary intake and discriminatory food allocation for females at household levels were responsible for higher prevalence of undernutrition among women (Kamal and Islam, Reference Kamal and Islam2010). It can be assumed that discriminatory behaviour towards the adolescent girls as well as less support from their family members to avail health care services are attributable to greater prevalence of underweight among ever-married adolescent girls in Bangladesh (Ahmed et al., Reference Ahmed, Adams, Chowdhury and Bhuiya2000; Kamal and Islam, Reference Kamal and Islam2010; Nguyen et al., Reference Nguyen, Sanghvi, Kim, Tran, Afsana, Mahmud, Aktar and Menon2017). Moreover, adolescent girls with lower academic qualifications and poorer economic conditions are more likely to have inadequate knowledge about healthy dietary practise and proper nutrition (Ahmed et al., Reference Ahmed, Adams, Chowdhury and Bhuiya2000; Kamal and Islam, Reference Kamal and Islam2010; Nguyen et al., Reference Nguyen, Sanghvi, Kim, Tran, Afsana, Mahmud, Aktar and Menon2017). This can be an explanation for higher nutritional ailment among the adolescents with less education and from poorest wealth index (Tareke and Abate, Reference Tareke and Abate2020). Studies conducted in Bangladesh and India identified that women of younger age, living in rural and urban slum areas, having a lower level of education and lower wealth index were more likely to suffer from different types of nutritional deficiencies (Goli et al., Reference Goli, Rammohan and Singh2015; Islam et al., Reference Islam, Islam, Bharati, Aik and Hossain2016; Subramanian and Smith, Reference Subramanian and Smith2006). Our study results are also analogous to those findings. Previous studies among young Bangladeshi women indicate that age at first marriage, age at first birth, and total number of children were important factors associated with nutritional status (Islam et al., Reference Islam, Islam, Bharati, Aik and Hossain2016; Khan and Kraemer, Reference Khan and Kraemer2009). Studies also indicate that economic status and education level are correlated with early childbirth and more number of childbirth (Islam et al., Reference Islam, Islam, Bharati, Aik and Hossain2016). Consistent with those reports, our study results also demonstrated that adolescent girls with low education, poorer economic condition, and multi-parity are more likely to be underweight.

The proportion of adolescent girls with a BMI of more than 25 kg/m² was higher among those who were from richest wealth quintiles, dwellers of urban areas, received higher education, used contraceptives, read newspapers, and watched television regularly. We also observed that age of the respondents, place of residence, wealth index, partner’s occupation, contraceptive use, age of first child birth, and number of ever born child were the factors demonstrating significant relationship with overweight/obesity among the adolescent girls included in this analysis. A study conducted among Bangladeshi ever-married women showed that women from an older age group were more likely to be overweight (Biswas et al., Reference Biswas, Uddin, Al Mamun, Pervin and Garnett2017). We found that girls who were living in the urban areas and hailing from richer and richest wealth quintiles were more likely to be overweight/obese. Evidence shows that adolescent girls from the comparatively better economic conditions were more likely to be overweight and obese than the women who are from poorer economic status (Ly et al., Reference Ly, Ton, Ngo, Vo and Fitzpatrick2013; Ramesh and Jareena, Reference Ramesh and Jareena2009; Subramanian et al., Reference Subramanian, Kawachi and Smith2007, Reference Subramanian, Perkins and Khan2009). It is evident that better economic status is associated with an increased intake of energy and fat, and consumption of animal and processed foods, all of which were found to be associated with overweight and obesity (Du et al., Reference Du, Mroz, Zhai and Popkin2004). Studies confirmed that the prevalence of overweight/obesity among women is higher in the developing countries due to rapid urbanisation (Mendez et al., Reference Mendez, Monteiro and Popkin2005). Another study conducted in Bangladesh showed that the prevalence of overweight/obesity is higher among the urban residents (Tanwi et al., Reference Tanwi, Chakrabarty and Hasanuzzaman2019). Report shows that residents from the urban areas have access to advanced technology, which help them to perform daily activities without losing substantial amount of energy (Monteiro et al., Reference Monteiro, Moura, Conde and Popkin2004). A study documents that shift in dietary habits towards western foods containing high protein and energy may also contribute to the increase in BMI (Pingali, Reference Pingali2007). Urban people are more likely to consume readily available processed foods and avail better transportation system resulting in less physical activities, which ultimately contribute to increase in the prevalence of overweight obesity among them (Monteiro et al., Reference Monteiro, Moura, Conde and Popkin2004). Adolescent girls from comparatively richest economic conditions were more likely to be overweight/obese in our study. Previous studies conducted in Bangladesh also found the similar result (Tanwi et al., Reference Tanwi, Chakrabarty and Hasanuzzaman2019). Perhaps, adolescent girls from richer families have access to western diet with high ingredients of energy and proteins and that is responsible for an increasing trend of being overweight/obese among the ever-married adolescent girls from better economic conditions (Tareke and Abate, Reference Tareke and Abate2020). Our results highlight that adolescent girls married to workers and small business owners were less likely to be overnourished. This finding can also be linked with wealth index as workers or small entrepreneurs are expected to be from lower economic conditions. The findings on relationship between parity and overweight/obesity are also supported by previous reports (Heliövaara and Aromaa, Reference Heliövaara and Aromaa1981; Huayanay-Espinoza et al., Reference Huayanay-Espinoza, Quispe, Poterico, Carrillo-Larco, Bazo-Alvarez and Miranda2017). It is evident that weight gain during pregnancy enhances the risk of postpartum weight retention (Butte et al., Reference Butte, Ellis, Wong, Hopkinson and Smith2003; Linne et al., Reference Linne, Dye, Barkeling and Rössner2003). Perhaps, gestational weight gain and certain puerperal conditions could play a potential role in the pathogenesis of parity-related overnutrition. The finding on negative relationship between older age (16–19 years) at first childbirth and overweight/obesity is analogous to previous reports, although the underlying mechanism is not yet clear (We et al., Reference We, Han, Kwon and Kil2016). However, studies revealed that younger age at first childbirth is associated with multi-parity and restriction of physical activity (Gunderson et al., Reference Gunderson, Murtaugh, Lewis, Quesenberry, West and Sidney2004; Lao et al., Reference Lao, Thomas, Jiang, Zhang, Yin, Schooling, Heys, Leung, Adab and Cheng2006). Apart from behavioural and genetic factors, these could be the probable causes for developing overweight/obesity among the girls with early childbearing age. Our result also indicates that adolescent girls who used contraceptives had lower risk of being overweight/obese. Although it has been assumed that contraceptive use may have role in a high BMI, a study reported to have no association between use of contraceptives and BMI in women (Schraudenbach and McFall, Reference Schraudenbach and Mcfall2009). However, the negative relationship between contraceptive and overweight/obesity observed in this analysis can be a spurious finding and needs to be elucidated further. We have documented that adolescent girls who could avail entertainment facilities including newspapers and television had higher prevalence of being overweight/obese. This finding is consistent with previous studies done in South Asian countries including Nepal and India (Das Gupta et al., Reference Das Gupta, Haider, Hashan, Hasan, Sutradhar, Sajal, Joshi, Haider and Sarker2020; Das Gupta et al., Reference Das Gupta, Haider, Hashan, Hasan, Sutradhar, Sajal, Joshi, Haider and Sarker2019).

This is the first study to describe the trends and risk factors of overweight/obesity and undernutrition among the ever-married adolescent girls in Bangladesh using national data. The use of data from well-designed nationally representative surveys is the major strength of our analysis. Moreover, we can assure that the respondents included in this analysis are nationally representative because of the sampling strategy used in the BDHS. Therefore, we believe that the findings from this study would help the policymakers to yield proper initiative and take appropriate measures to fight against both under- and over-nutrition among the most vulnerable age group of the population. However, there are several limitations. We could only include the ever-married adolescent girls aged between 15 and 19 years in this analysis. The unmarried adolescent girls and those with an age between 10 and 14 years were not included owing to unavailability of data. Since the surveys were cross-sectional, the relationship between the exposures and outcome variables does not infer causality. Further studies are required to elucidate the causality of the risk factors identified in this study. Additionally, the BDHS data do not include information on nutrient intake as well as dietary diversity and physical activities of the adolescent girls, although these variables are considered to be important determinants of nutritional insufficiencies and malnutrition. Moreover, BMI – the indicator that we have used to determine nutritional status of the girls – cannot differentiate between fat and fat-free mass. However, BMI is a widely recognised indicator to assess nutritional status in adolescents and adults (Fahim et al., Reference Fahim, Das, Gazi, Alam, Mahfuz and Ahmed2020).