Globally, in 2015, it was estimated that 9% of the world's adult population were underweight and 30–40% were overweight or obese (1, 2). Women had a slightly higher prevalence of overweight and obesity (OWOB) (1). Although there has been a marginal decline in the prevalence of underweight (1), the rise in the proportion of people being overweight or obese is currently being described as a global pandemic. There has been a 50–80% increase in OWOB in the past 30 years (2). Despite some variability, the increase in the prevalence of OWOB is seen across all countries, ages, socio-economic strata, and in both sexes.

Both underweight and overweight affect the functioning as well as structure of body organs, increasing the risk of mortality (3, 4). A nutritional status outside of the normal range is one of the leading risk factors for premature death and loss of disability adjusted life years (DALY) (5). Overweight and obesity are associated with several other adverse health outcomes, including diabetes mellitus type 2, cardiovascular disease, certain types of cancer, musculoskeletal and mental disorders as well as pregnancy complications (3, 6, 7). Amongst others, undernutrition is associated with an increased risk of infection (8) as well as pre-term birth/ low birth weight (7, 9).

Malnutrition, in all its different forms, affects all countries of the world (10). Many countries face a double burden of malnutrition (DBM), where both undernutrition and OWOB exist in the same population, household, or even individual (11). While in high-income countries, OWOB is usually more prevalent among populations with a low socio-economic status (SES), the opposite has been observed for low-income countries (2, 12, 13). The results from a recent analysis including 126 low- and middle-income countries (LMIC) showed that the increase in the global DBM was driven by the countries with a low gross domestic product, as these tended to have a greater increase in OWOB and a slow decrease in the prevalence of undernutrition. This increase in the DBM was particularly observed in Asia (11).

Studies undertaken in Nepal have shown that although the prevalence of undernutrition in the population has generally decreased over the last decades, it has remained high among women of reproductive age. During this period the prevalence of OWOB has increased substantially (14, 15). Three national surveys (Demographic Health Survey (DHS) Nepal 2016, Micronutrient Survey 2016 and the STEPS survey 2019), estimated that overall 14.5–17% of the Nepalese adult women were underweight, while 22–25% were OWOB (16–18). Among men, 17% were underweight and 17–23.4% were OWOB (16, 17). The increase in OWOB is an important driver for the increase in the DBM in Nepal (11). In the Nepal DHS 2016, there was no difference in the prevalence of OWOB between ecological zones or between rural vs. urban populations (19). However, Nepal is a country with a large diversity in both populations and ecology, and thus nation-wide estimates or even broad comparison groups such as ecological zones will dilute important local differences. It is important to remember that the 2020 Lancet series on the DBM stresses the importance of describing the full distribution of anthropometric status of populations instead of assessing undernutrition and OWOB in their own silos (11, 20–22). Currently, there are a few local studies on the prevalence of OWOB in Nepal (23–25); however, these do not report on the burden of undernutrition in the same population.

We used high quality data on the anthropometric status from young adults whose child was enrolled in a community-based trial. Participants lived in the peri-urban municipality of Bhaktapur, Nepal. The study period was from 2015–2017. Our aim was to describe the whole distribution of BMI and the prevalence of both underweight, overweight, and obesity in this poor population, including associated factors.

In total, weight and height were measured for 600 mothers and 445 fathers. In general, the fathers were slightly older (mean 30.2 vs. 27.3 years). The majority of the were unemployed (58%), while only 2.7% of the fathers were without work. Of the 322 fathers who reported drinking alcohol, 70 (22%) reported that this involved drinking daily, and 139 (43%) reported drinking once a month or less. Most of the participants were from the Newar caste (70%). Details of the participants' characteristics are shown in Table 1.

The mean BMI of the mothers and fathers were similar, 23.7 kg/m² for both (SD 3.6 and 3.7, respectively). The distributions with the commonly used WHO cut-offs are shown in Figure 1. The proportion categorized as underweight, overweight, and obese were similar between the groups of mothers and fathers with around 5% being underweight, 30% being overweight and 5% being obese (Figure 1). In the additional categories with trigger points suggested for Asian populations, around 20% were at increased risk (BMI 23–24.9), and around 10% were at high risk (BMI 27.5–29.9). A table including this information can be found in the (Supplementary Table 1).

Results from the linear regression models can be found in Table 2. For the mothers, age was positively associated with BMI. Those mothers who were self-employed had a significantly higher BMI and those in the group of daily wage earners had a lower BMI compared to those without formal work. Women with previous pregnancies had a higher BMI compared to those being pregnant the first time. Those belonging to the Chhetri and the Tamang/Lama caste had significantly lower mean BMI compared to those belonging to the Newar caste. However, in the multivariable analyses, only age, and occupational group were retained as statistically significant. We could not find any association between maternal BMI and level of education, smoking status, family type, or number of family members. For the fathers included in this analysis, age, education level, and number of family members were positively associated with BMI. The BMI was higher in all occupational categories compared to daily wage earners, although it was not statistically significant for carpet workers. Fathers had a higher BMI if they lived in a joint family compared to a nuclear family. They had a lower BMI if they were in the Tamang/ Lama caste compared to the Newar caste. In the multivariate analysis, the results were similar to those for the mothers, where age and occupational group remained significantly associated with BMI. All other characteristics were not significantly associated with paternal BMI. Associations with BMI that were categorized as undernutrition or OWOB compared to a normal BMI showed a similar pattern (see Supplementary Tables 2, 3). These results should be interpreted cautiously due to the small sample size in some categories.

Anthropometric data were only available for 445 of 600 fathers. Those with missing BMI were slightly younger than those with information available (28.7 and 30.3 years; p = 0.05). They also had a lower education level with 46% being illiterate or with finished primary school compared to 31.7% in those with BMI available (p = 0.001).

In this study, we showed that in a group of young adults living in the municipality of Bhaktapur in Nepal, 5% were underweight (BMI <18.5) and 35% were overweight or obese (BMI ≥25). The anthropometric status was similar for men and women. There was a difference in mean BMI between different occupational groups and age was positively associated with BMI for both men and women.

Nepal is categorized as a low-income country according to the World Bank. However, in the last few decades, health indicators in Nepal have improved tremendously. For example, life expectancy at birth has increased from 54 years in 1990 to 70 years in 2017, and infant mortality rate (death per 1,000 live births) has decreased from 97.2 in 1990 to 26.7 in 2017 (30). At the same time, Nepal has still one of the highest prevalences of childhood stunting in the world. Nearly 40% of children under the age of five were reported with stunting in the Demographic Health Survey (DHS) 2016 (16). In addition to the very slow decline in stunting rates, there are reports of an increase in overweight and obesity (OWOB) (14, 31) and non-communicable diseases (NCDs) in Nepal (23, 24, 32). Comparing the prevalence of OWOB between the DHS published in 1995 and 2016, the prevalence was 4.2 times higher in 2016 with a higher increase noted for rural areas compared to urban (32). However, there remains a paucity of comprehensive and high-quality data on the double burden of malnutrition including both undernutrition and overweight in LMICs such as Nepal (33). Such information is critical for informing evidence-based health policies at the local and national levels.

In 2004, a WHO expert panel suggested additional trigger points to categorize BMI for Asian population based on higher observed risk for cardiovascular disease and diabetes mellitus type 2 compared to Caucasian populations at the same BMI (29). Possible reasons for an increased risk in Asian populations are a comparably higher percent of total body fat and a decreased fat-free mass at the same level of BMI (34). Asians have also been shown to be more prone to abdominal obesity and fat deposition in various organs such as the liver, which is strongly associated with insulin resistance (35, 36). At the same time, Zheng et al. (37) find that among 1 million Asian adults, those with a BMI between 22.6 and 27.5 had the lowest risk of mortality. In Indian and Bangladeshi cohorts, the risk of death was elevated for BMI <20 but not elevated at higher BMI levels. Despite these conflicting results, many Asian countries, such as India, have lowered their cut-off for defining obesity (38). In our study, 57% of the women and 54% for men would be defined as OWOB according to the cut-off of BMI ≥23 that is currently also used in the neighboring country India. This would not only mean an increase in the prevalence of OWOB, but also that more people at risk of obesity-related diseases. These would not be identified for treatment if only the WHO definition was used.

Our data is from adults measured between 2015 and 2017 who lived in the peri-urban municipality of Bhaktapur, which lies in the Kathmandu valley. Compared to other reports from Kathmandu and province 3 (including Kathmandu) (24, 39), the prevalence of both under- and overweight seems somewhat lower in our sample. This could be due to differences in mean age. We included relatively young adults with a mean age of around 30 years (compared to 40 years in the other reports). BMI usually increases with age, something we also showed in our analyses. The eligibility criteria of the original study might have also contributed to our findings, as only children with a height-for-age Z-score of <-1 were enrolled into the original study. However, there were few, only 41, children who did not comply with this inclusion criterion. Another consideration is that although women that are underweight have a higher risk of having offspring that is underweight, stunted, or micronutrient deficient, a study by Popkin et al. (11) showed that the global double burden of malnutrition within a household is primarily driven by the combination of overweight women with stunted children. Their result was also found for Nepal. Thus, we cannot infer a clear effect direction due to the exclusion of these 41 children and their parents.

In addition to age, occupational group was identified as being significantly associated with BMI in all our multivariate regression models. Occupation is closely related to other indicators of socio-economic status such as wealth, education, and caste and might not only indicate actual resources and thus access to health care and food, but also the level of physical activity/sedentary lifestyle. Daily wage earners, who on average had a lower BMI in our sample, usually have fewer resources and a more labor intense work compared to other occupational groups in Nepal. Education level was positively associated with BMI in the bivariable analyses, which agrees with other studies from Nepal (31, 39). However, we could only show an association between level of education and BMI in the fathers and not in the mothers. In a systematic review, Wei et al. (15) found that among women in Nepal, OWOB increased at all educational levels and that there was no clear pattern in underweight. Our findings could also indicate that in Nepal, men are often the main source of household income and therefore paternal education and occupation might contribute to a larger extent to the individual BMIs in the household compared to women's education level.

In our study sample, child undernutrition is still highly prevalent with 17% underweight and 32% stunted. Others report that 6.6% of 13–17 year old school children were overweight or obese in a national school survey (40) and 25.9% of 6–13 year old students in a private school in Lalitpur (41). Neither of these studies reported indicators of underweight. In the Lancet series on the double burden of malnutrition (DBM), Popkin et al. (11) describe a substantial increase in the prevalence of obese mothers with children that are stunted in Nepal. This emphasizes not only that malnutrition should be reported over the whole spectrum including both extremes of nutritional status, but also that intergenerational processes may play an important role (42), and thus, it underlines that we have an the opportunity and a need to explore critical windows for potential interventions so as to address several forms of malnutrition simultaneously.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Nepal Health Research Council and Regional Committee for Medical and Health Research Ethics Norway. The patients/participants provided their written informed consent to participate in this study.

CS, RC, MU, and TS: conception/design of the work. RC, MU, and TS: data acquisition. CS: data analysis, data interpretation, and drafted the manuscript. RC, MU, MH, MS, TS, and SR: substantially revised the manuscript. All authors: approved the submitted version and agreed to be personally accountable for any part of the work.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.