The widespread increase of the prevalence of overweight and obesity has long been described as a global pandemic( Reference Popkin, Adair and Ng 1 ). The total number of adults with overweight (BMI ≥ 25·0 kg/m2) and obesity (BMI ≥ 30·0 kg/m2) was estimated to be 2·01 billion and 677 million, respectively, in 2016( 2 ). Overweight and obesity are well known risk factors for type 2 diabetes, hypertension, CVD and cancer( Reference Haslam and James 3 ), accounting for an estimated 7·1 % of all deaths and 3·6 % of years lived with disability among adults in 2015( Reference Afshin, Forouzanfar and Reitsma 4 ).
The WHO Stepwise Approach to Chronic Disease Risk Factor Surveillance (STEPS) provides a standardized protocol for the surveillance of non-communicable diseases and their risk factors, including obesity, to guide monitoring efforts on a regular basis, mostly in low- and middle-income countries( 5 ). It is based on a sequential process, starting with the collection of key information on risk factors using a questionnaire, and moving on to simple physical measurements and collection of samples for biochemical analysis( 5 ). In Mozambique, the first WHO STEPS survey conducted in 2005 showed a low prevalence of obesity and overweight, although higher in urban areas and among women, mostly in those with higher socio-economic status( Reference Gomes, Damasceno and Azevedo 6 ). According to these results, Mozambique was in stage 1 of the recently proposed conceptual model of the stages of the obesity transition( Reference Jaacks, Vandevijvere and Pan 7 ). However, an increase in the prevalence of obesity among adults, as well as a narrowing of the gap between genders and of the socio-economic differences among women may be expected, given the quick demographic, nutritional and epidemiological transitions occurring in this context. This highlights the importance of monitoring the prevalence of overweight and obesity, to guide local public health policies.
Therefore, the present study aimed to quantify the prevalence of overweight and obesity in the Mozambican adult population using data from the most recent STEPS survey conducted in Mozambique in 2014/2015, as well as to compare the estimates with those obtained in the 2005 STEPS evaluation.
Methods
The present study is based on a cross-sectional analysis of a representative sample of the Mozambican population aged 15−64 years, conducted between December 2014 and February 2015.
Selection of participants
The study design and implementation have been described in detail elsewhere( Reference Jessen, Damasceno and Silva-Matos 8 ). Participants were selected through a complex sampling design, to obtain a representative sample at the national and provincial levels, and according to the place of residence (urban v. rural); the homeless and people living in collective residential institutions (e.g. hotels, hospitals, military facilities) were not eligible. Sampling procedures started with the selection of 120 primary sampling units (i.e. geographical units including 400 to 600 households), with a probability proportional to the number of households, stratified according to province, urban/rural areas and socio-economic strata. One enumeration area (i.e. geographical unit including eighty to 150 households) was then randomly selected within each primary sampling unit, and twenty-four households were randomly and systematically selected, from updated lists of households, in each enumeration area. A maximum of two participants were selected from each selected household, one aged 15–44 years and one aged 45–64 years, whenever available; whenever there was more than one household member in each of these age groups, only one per group was randomly selected using a Kish selection grid. A total of 3277 individuals were invited and 3119 (95·2 %) agreed to participate in the survey. Sampling weights were computed considering the number of participants evaluated in each stratum in relation to the number expected per stratum, according to the population projection for the same period. A total of 2756 participants were eligible for the present study, after excluding those below 18 years of age (n 248) and pregnant women (n 115).
A STEPS survey was previously conducted in Mozambique in 2005, as described in detail elsewhere( Reference Gomes, Damasceno and Azevedo 6 ). It was based on essentially the same methods used in 2014/2015 and the data set was available for comparison with the most recent data.
Evaluation of the participants
Participants were evaluated using standardized methods, following the WHO STEPS( 5 ). The Portuguese version of the WHO STEPS instrument for non-communicable disease risk factors (Core and Expanded version 3.0) was used by trained interviewers for data collection.
Anthropometric measurements were obtained with participants wearing light clothing and no footwear. Body weight was measured to the nearest 0·1 kg using a digital scale and height to the nearest 0·1 cm in the standing position using a portable stadiometer. BMI was calculated as weight (in kilograms) divided by the square of height (in metres) and stratified into the categories: <18·5 kg/m2, 18·5–24·9 kg/m2, 25·0–29·9 kg/m2 and ≥30·0 kg/m2, corresponding to underweight, normal, overweight and obesity, respectively, as defined by the WHO( 9 ). Waist circumference was measured to the nearest 0·1 cm, using a constant tension tape, directly over the skin or over light clothing, at the level of the midpoint between the inferior margin of the last rib and the iliac crest in the mid-axillary line. Participants were classified as having abdominal obesity if their waist circumference was greater than 88 and 102 cm in women and men, respectively, as defined by the WHO( 9 ).
The classification of the place of residence as urban or rural and the definition of the highest level of education attained were conducted in accordance with the 2007 census( 10 ). Monthly income per capita in meticais, per household member aged ≥18 years, was converted into US dollars, taking into account the exchange rate in the midpoint of the time period of the participants’ evaluation.
Statistical analysis
Means with 95 % CI were computed for BMI and waist circumference, and prevalence estimates as percentages with 95 % CI for different categories of BMI and waist circumference. Age-, education- and income-adjusted regression coefficients were computed using linear regression to estimate the strength of the association between sociodemographic characteristics and waist circumference, according to place of residence. Logistic regression was used to assess the association of the same characteristics with overweight and obesity. In addition, models including the same independent variables were fitted to quantify the association of place of residence and gender with overweight and obesity and abdominal obesity.
The 2005 results were compared with those observed in the current STEPS survey, after direct standardization using the 2005 population as reference. Since in 2005 only adults aged 25–64 years were evaluated, all comparisons with the 2014/2015 data referred to the same age strata. Hypothesis tests for no change between the two periods were conducted by dividing the difference in proportions or means estimated in the two surveys by its se, and calculating a two-sided P value using the Student’s t distribution( Reference Sunghee, William and Hoang 11 ). A P value less than 0·05 was considered to be statistically significant.
Analyses were conducted using the statistical software package Stata version 11.2, considering the sampling weights, so that the computed estimates reflect the means and prevalence in Mozambique, and adjusting for stratification by province and clustering at the primary sampling unit level.
Results
A total of 2595 participants were considered for the present analysis, after excluding those from whom data on weight, height or waist circumference were unavailable (n 161). The latter were more often dwellers in urban areas than those included in the analysis (57·1 v. 47·0 %, P = 0·013), but there were no statistically significant differences regarding gender (proportion of women: 55·8 v. 59·9 %, P = 0·317), age (mean age: 34·6 v. 36·5 years, P = 0·077) and education (proportion with ≥1 year of education: 73·8 v. 73·7 %, P = 0·986).
In 2014/2015, just over a third of the population lived in urban areas and approximately half were aged 18−34 years. Approximately one in three women and one in six men had no formal education, and one-third reported a monthly income of less than $US 16 per adult household member (Table 1).
Sociodemographic characteristics of the participants: representative sample of the Mozambique population aged 18–64 years (n 2595), 2014/2015
* The sum of the number of participants in each category is lower than 1554 for women and 1041 for men due to missing data.
BMI in 2014/2015
The mean BMI (kg/m2) was 23·5 (95 % CI 23·1, 23·9) among women and 22·1 (95 % CI 21·7, 22·5) in men. The prevalence of overweight was 17·5 (95 % CI 14·9, 20·1) % and 12·1 (95 % CI 9·5, 14·7) % among women and men, respectively, and higher in urban than rural areas in both genders (Table 2). Concerning obesity, the prevalence was 10·4 (95 % CI 8·4, 12·5) % among women and 4·2 (95 % CI 2·6, 5·8) % in men, also higher in urban settings and among women. Underweight affected approximately one in ten participants, in both genders and settings (Table 2).
Mean BMI and waist circumference, and prevalence of BMI categories and abdominal obesity, according to gender and place of residence, among the Mozambique population aged 18–64 years, 2014/2015*
* Results are presented as mean or as percentage, with 95 % CI, as specified.
† Defined as waist circumference greater than 88 and 102 cm in women and men, respectively.
In rural areas, in both women and men, there was a tendency for overweight and obesity prevalence to increase with age. Among those aged 18–24 years, the prevalence of overweight and obesity was highest in urban women (22·0 %) and lowest in rural men (5·8 %; Table 3). The prevalence of overweight and obesity was highest among women (urban, 44·4 %; rural, 21·4 %) and urban men (32·8 %) with 1–5 years of education, and in the more educated rural men (12·4 %). In men, in both settings, the prevalence of overweight and obesity increased with income, while no such tendency was observed among women (Table 3).
Prevalence of overweight and obesity, and mean waist circumference, among women and men from urban and rural areas, according to age, education and income, in the Mozambique population aged 18–64 years, 2014/2015
Ref., reference category.
* Derived from models including age, education and income.
† Difference between the mean value of each category and the mean value of the reference category.
The association between place of residence and overweight and obesity was attenuated after adjustments for age, education and income. The association between gender and overweight and obesity was stronger after adjustment for sociodemographic characteristics (Table 4).
OR and 95 % CI for the association between place of residence and overweight and obesity, and abdominal obesity, according to place of residence and gender, in the Mozambique population aged 18–64 years, 2014/2015
Ref., reference category.
* Defined as waist circumference greater than 88 and 102 cm in women and men, respectively.
† Derived from models including age, education and income; P interaction = 0·980.
‡ Derived from models including age, education and income; P interaction = 0·681.
Waist circumference in 2014/2015
The mean waist circumference (cm) was 78·1 (95 % CI 77·4, 78·9) among women and 76·6 (95 % CI 75·8, 77·4) in men, and higher in urban areas. The prevalence of abdominal obesity was 16·5 (95 % CI 13·8, 19·3) % among women and 1·8 (95 % CI 0·9, 2·7) % in men, and was also higher in urban settings (Table 2).
There was an increase of waist circumference with age, except among the oldest women in urban areas. The differences between settings were highest at age 45–54 years in both genders; in relation to rural dwellers, the mean waist circumference of urban women and men was 9·9 and 7·4 cm larger, respectively (Table 3).
Waist circumference was highest in urban and rural women, and in urban men with 1–5 years of education. In men, it increased with income (Table 3).
The prevalence of abdominal obesity was higher in urban than rural areas. The relationship between gender and abdominal obesity was stronger after adjustments for age, education and income, in a similar extent in both settings (Table 4).
Comparison between the results obtained in the 2005 and 2014/2015 surveys
Between 2005 and 2015, there were increases in mean BMI and mean waist circumference. The prevalence of overweight and obesity increased from 18·3 to 30·5 % among women and from 11·7 to 18·2 % in men. For obesity, there was an increase from 7·0 to 13·0 % among women and from 2·3 to 5·0 % in men. The prevalence of abdominal obesity in 2014/2015 was more than twofold higher among women (Table 5).
Overweight and abdominal obesity prevalence in Mozambique in 2005 and 2014/2015, among the population aged 25–64 years*
* Results are presented as mean or as percentage, with 95 % CI, as specified.
† Defined as waist circumference greater than 88 and 102 cm in women and men, respectively.
Discussion
In Mozambique, the prevalence of obesity and overweight increased significantly in just a decade. In 2014/2015, almost one in five men and one in three women were affected. Among the latter, there was a twofold higher prevalence of abdominal obesity than 10 years before, being observed in one out of five women in the most recent years. The prevalence of overweight/obesity was more than twice as high in urban than rural areas.
Previous studies analysing overweight and obesity trends worldwide showed a slowing down in the rate of increase of the prevalence of overweight and obesity in high-income countries and a more rapid increase in the proportion of those living with these conditions in low- and middle-income countries( Reference Ng, Fleming and Robinson 12 , Reference Kelly, Yang and Chen 13 ). In sub-Saharan Africa, nationally representative data of overweight or obesity trends are scarce and focus mainly on women of childbearing age(Reference Amugsi, Dimbuene and Mberu14–16). Nevertheless, findings from these studies showed that, in the past two decades, the prevalence of overweight and obesity more than doubled in most sub-Saharan African countries with data available( Reference Amugsi, Dimbuene and Mberu 14 , Reference Abubakari, Lauder and Agyemang 15 ), being higher among women and urban residents( Reference Abubakari, Lauder and Agyemang 15 , 16 ), which is in accordance with our results.
The trend towards the urbanization of the Mozambican population is likely to have contributed to the increase in the prevalence of overweight and obesity observed in the present study. The estimated urban Mozambican population was about 30 % in 2005 and 34 % in 2015( 17 ), being projected to increase to 37 % by 2030( 18 ). Furthermore, a transition from a predominantly rural environment to a more urban ‘westernized’ society may have occurred within regions formally classified either as urban or as rural. Urbanization is associated with lifestyle changes, such as an increased consumption of energy-dense ultra-processed foods and sugar-sweetened beverages, along with a decrease in physical activity levels( Reference Popkin, Adair and Ng 1 ). In 2005, most adult Mozambicans were classified as having a high physical activity level( Reference Padrao, Damasceno and Silva-Matos 19 ), according to the WHO criteria( 20 ); however, the prevalence of low level of physical activity was higher in urban areas and increased with education( Reference Padrao, Damasceno and Silva-Matos 19 ). A shift towards more sedentary behaviours may have occurred over the past years, as a result of increasing urbanization and economic development, at least in those previously living in rural settings. Among Mozambicans, the use of motor vehicles and computer and television ownership increased between 2008/2009 and 2014/2015, from 5·6 to 12·2 %, from 1·5 to 5·6 % and from 12·4 to 24·2 %, respectively( 21 ). Also, a decline in physical activity was previously reported for children and adolescents( Reference dos Santos, Maia and Gomes 22 ). In addition to the decrease in energy expenditure, the acquisition of new dietary habits, as demonstrated by the steep increase in the proportion of meals consumed outside the home( 23 ) and the greater availability of fast-food restaurants and street food including fried and sweet meals( 24 ), may have contributed to the pattern found. Furthermore, the consumption of vegetables and fruit, known to contribute to the prevention of obesity and other non-communicable diseases, tends to decrease with urbanization, as reflected in the lower consumption of those foods in urban compared with rural areas in Mozambique( Reference Padrao, Laszczynska and Silva-Matos 25 ).
Young adults are in a period of transition from adolescence to adulthood, and therefore increasing decision-making autonomy, which makes them a vulnerable group for the acquisition of unhealthy lifestyles that may lead to overweight and obesity( Reference Nelson, Story and Larson 26 ). Despite the 2005 STEPS survey lacking data for the 18–24 years age group, precluding comparisons among younger individuals, in 2014/2015 the prevalence in this age group was close to that observed in 2005 among participants aged 25–34 years( Reference Gomes, Damasceno and Azevedo 6 ). The already high prevalence of overweight and obesity among some subgroups of young adults shows that further increases in the burden of overweight and obesity may be expected in the coming years.
Mean waist circumference increased between 2005 and 2014/2015. When compared with other sub-Saharan African countries, the mean values observed in Mozambique are lower than those reported for Benin, Kenya and Swaziland, both in women and men( 27 ). Although waist circumference is recognized as an important indicator complementing BMI to identify individuals at increased risk of obesity-related morbidity( 9 ), the most common cut-offs used throughout the world to define abdominal obesity were based on a population of predominantly European origin( Reference Han, van Leer and Seidell 28 ). In fact, it was previously shown that black women from South Africa have less abdominal adipose tissue at the same waist circumference than European women( Reference Rush, Goedecke and Jennings 29 ). Therefore, in the absence of ethnic-specific cut-off points to define abdominal obesity in this setting, caution is needed when discussing prevalence data of abdominal obesity. Still, the prevalence, as defined using the cut-offs adopted by the WHO( 9 ), more than doubled among women in just a decade and so efforts need to be made to understand its possible impact on health outcomes, particularly on cardiometabolic diseases.
Similar to the findings from the first WHO STEPS survey conducted in Mozambique( Reference Gomes, Damasceno and Azevedo 6 ), our results showed a higher prevalence of overweight and obesity in urban settings, both among women and men. A higher prevalence of overweight and obesity in urban areas was also reported in other sub-Saharan African countries such as Malawi( Reference Price, Crampin and Amberbir 30 ) or Uganda( Reference Kirunda, Fadnes and Wamani 31 ), while equal or higher prevalence has been found in rural areas of European countries( Reference Peytremann-Bridevaux, Faeh and Santos-Eggimann 32 ) and the USA( Reference Befort, Nazir and Perri 33 ). At the same time, the gap between the prevalence of underweight and overweight/obesity has increased sharply as expected in countries under epidemiological and nutrition transitions.
Although in the 2005 survey the prevalence of overweight and obesity tended to increase with education, both in rural and urban areas( Reference Gomes, Damasceno and Azevedo 6 ), in 2014/2015 the association between obesity and education was weaker. This may reflect a more advanced stage of the transition than that found in the earlier survey, as in the more advanced stages of the obesity transition there is an inverse relationship between socio-economic status and obesity prevalence( Reference Jaacks, Vandevijvere and Pan 7 ).
The present study was based on a large representative sample of the adult population, and it allowed a direct comparison with a previous survey conducted in Mozambique, using essentially the same standardized procedures and following the WHO STEPS methodology. However, some limitations need to be discussed. Participants for whom data were unavailable on anthropometric measures were excluded from the current analysis. Nevertheless, given the low proportion of participants with missing data and the fact that they were not significantly different from those included in the analysis regarding gender, age or education, the impact of their exclusion on the validity of our results is expected to be low. Also, despite the large overall sample size, the precision of the estimates obtained for some subgroups considered in the stratified analyses was lower than desirable, especially among rural men. The lack of data in the 2005 and 2014/2015 STEPS surveys for younger children and adolescents precludes the study of the prevalence of overweight and obesity among these age groups, which would provide a better understanding of the burden expected for the next years in this population.
Conclusion
To conclude, in Mozambique there was a steep increase in the prevalence of overweight and obesity among adults in the past 10 years. The prevalence of obesity and overweight is higher in urban areas and in women, already affecting one in every four urban women aged 18–24 years. These results suggest that culturally adapted interventions are needed to tackle the increasing morbidity and mortality burden that may be expected in the next years in Mozambique, especially among women and urban dwellers with higher socio-economic status.
Acknowledgements
Financial support: This study was supported by the Mozambican Ministry of Health and by the WHO. The study design and data collection were performed according to WHO guidelines. In addition, this study was funded by FEDER through the Operational Programme Competitiveness and Internationalization and national funding from the Foundation for Science and Technology – FCT (Portuguese Ministry of Science, Technology and Higher Education) under the ‘Unidade de Investigação em Epidemiologia – Instituto de Saúde Pública da Universidade do Porto (EPIUnit)’ (POCI-01-0145-FEDER-006862; ref. UID/DTP/04750/2019). The funders had no role in the design, analysis or writing of this article. Conflict of interest: None. Authorship: A.D., C.S.-M. and N.L. conceived and designed the study. F.F. and N.L. analysed data. All authors were responsible for data interpretation. F.F. wrote the first version of the manuscript. A.D., N.J., A.P., C.S.-M., P.P. and N.L. critically revised the manuscript for relevant intellectual content. All authors approved the final version for submission. Ethics of human subject participation: This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by the National Bioethics Committee for Health. Written informed consent was obtained from all subjects.
