Dietary Patterns, Nutrient Intakes and Metabolic Conditions Among Agro-Pastoralists in Monduli District, Tanzania

Background

Obesity and metabolic conditions are associated with many non-communicable diseases (NCDs) such as cardiovascular diseases, type 2 diabetes and cancer.¹ These conditions are associated with poor quality of life, disability and mortality. They also increase the cost of the health care system significantly. Nowadays, most of the low and middle income countries (LMICs) are affected by rapidly increasing prevalence of these conditions.² These including the rising prevalence of high blood pressure and obesity which are the main risk factors for cardiovascular diseases such as heart diseases and stroke worldwide.³ Globally, it was estimated that a quarter of the world’s adults population had high blood pressure in 2002, and the prevalence will increase to 29% by 2025.⁴ Tanzania is one of several African countries facing increasing rates of NCDs. The prevalence of obesity and other metabolic syndromes in the country has dramatically increased in the past decade.⁵ More studies are therefore needed to inform public health actions to prevent these conditions and eventually to reduce the NCDs.

Poor dietary habits play an important role in the development of metabolic conditions such as obesity and high blood pressure. Studies that examined this relationship are very scarce among indigenous population of Africa, especially those living in poor resource settings such as in pastoral and agro-pastoral communities. Monduli District is one of five districts of Arusha region, Northern Eastern Tanzania. It is occupied mostly by the “Maasai” communities whose main livelihood is pastoralism. They are deep attached to their livestock and undertake minimal crop production mainly on maize which is low in micronutrients. In general, this could lead to cases of under-nutrition and the diets consumed being relatively low in diversity and thus lack essential micronutrients.⁶ The food intake is mainly from grains and maize as staple foods. Meat products are rarely consumed, where animals are slaughtered for specific occasions or social obligations. Most of the milk, which is one of the most common animal products available among the community, is transported and sold to other neighboring areas. On the other hand, previous cross-sectional studies showed that 27% of adults had high blood pressure and 20.2% of them were obese in this district.^7,8 We therefore sought to identify the dietary patterns which may be potential dietary factor to prevent the rise of metabolic conditions. Dietary pattern reflects a combination of foods and may be easier to interpret and translate by the general public, potentially improving the development of and adherence to food-based dietary guidelines and recommendations.⁹ This analysis has emerged as an alternative approach to examining the relationship between diet and the risk of metabolic syndromes.¹⁰

Very few studies regarding dietary patterns have been conducted in Tanzania.^11–16 For example, Katalambula et al¹¹ examined the dietary patterns that are associated with high blood pressure among adults in Arusha city in Tanzania. In addition, Keding et al¹² described and characterized the dietary patterns of rural women in three districts from northeastern and central Tanzania. Moreover, other researchers have explored the association of dietary patterns with various clinical and health outcomes in different part of the country.^13,14,17 None of these studies, to our knowledge, assessed the relationship between dietary patterns and metabolic health of agro-pastoralists in Tanzania. Examination of the types of dietary patterns is very important to provide specific dietary recommendations to inform tailored interventions. Therefore, we evaluated the types of dietary patterns existed in agro-pastoral areas in the Monduli district. We also determine the associations between dietary patterns and some selected metabolic conditions which are high blood pressure, general obesity and abdominal obesity.

Materials and Methods

Study Population

This study involves adults 18 years of age or older residing in Monduli district in northern part of Tanzania. The participants were permanent residents in previously sampled villages in the district. Five (5) villages were selected randomly in the district as explained previously.⁷ Briefly, participants were recruited in their traditional homesteads known as “bomas”. We obtained the list of bomas in each village and selected them randomly. Usually, one boma may contain up to 10 households. A total of 15 bomas were selected from previously sampled villages. In each “boma” all adults were invited to participate until the required number is reached. With the help of research assistants, all participants were interviewed based on the interviewer-administered questionnaires and completed mandatory measurements.

Dietary Assessment

Food consumption data were collected by using a quantitative food frequency questionnaire (FFQ). The FFQ contained 27 food items featuring typical food items consumed in the community. We earlier examined the validity of this food frequency questionnaire (FFQ) and found that it can be useful in diet related disease studies.¹⁸ All participants were required to indicate how many times they consumed each food item in the past month. The frequency of consumption was measured by selecting one of the following nine options: (1) never; (2) 1–3 per month (3) once per week; (4) 2–4 times per week; (5) 5–6 times per week; (6) once per day; (7) 2–3 times per day (8) 4–5 times per day (9) 6+ times per day.¹⁹

Dietary Pattern and Nutrients

In this study, the number of foods used to derive the dietary patterns was 25 by excluding some foods with consumption of less than 10% of participants due to lack of their variation. The quantity of food intake in g/day was calculated for each participant and was used to determine the dietary pattern. Dietary patterns were derived from factor analysis using principal component analysis (PCA). The number of patterns were selected based on the Scree plot and Eigenvalues of more than 1 using orthogonal transformed with varimax options.²⁰ The factor scores for each pattern was calculated and assigned to each participant. Main dietary patterns were labeled based on the highly loaded food items for each factor.²⁰ Food items with factor loadings less than 0.3 were disregarded for defining the dietary patterns.¹³ The factor scores of each participants were then categorized into quartiles for analysis. For each participant, the factor scores indicated the extent to which his/her diet conformed to the respective dietary patterns.¹¹ Furthermore, nutrient values for each food were taken from the Tanzania Food Composition Table (TFCT).²¹ The TFCT provides the amounts of each nutrient per 100 grams for individual foods, beverages and for mixed dishes.

Anthropometric Measurements

Weight and height were taken using a digital weighing machine, and a non-stretchable measuring tape, respectively. Body weight was measured to the nearest 0.1 kg precision. Height was measured using a measuring tape fixed on a height board to the nearest 0.1 cm precision. Participants were asked to remove shoes and heavy clothes before weight and height were measured. All anthropometric measurements were repeated at least two times for accuracy. Body mass index (BMI) was calculated as weight (kilograms) divided by squared height (meters) and then categorized as < 18.5 kg/m² as underweight, 18.5–25 kg/m² as normal, 25–30 kg/m² as overweight, and above 30 kg/m² as obesity. Waist circumference was measured at the thinnest point of the abdomen at the end of a normal expiration. Hip circumference was measured at the maximum circumference over the buttock horizontally. Both waist and hip circumferences were measured using a measuring tape with 1 mm accuracy. Waist-hip ratio (WHR) was calculated by dividing the waist by the hip circumference. Abdominal obesity was defined based on the WHO cut-offs, whereby measurement of WHR ≥0.90 in males and WHR ≥0.85 in females indicates higher risk for cardiovascular diseases.²² Abdominal obesity by WC was determined by using a cut-off point of ≥ 102 cm for males and ≥ 88 cm for females.²²

Measurement of Blood Pressure

Blood pressure was measured at least three times by trained health professionals. The participants rested for few minutes before their blood pressure was measured. All three measurements were performed in a sitting position for at least five minutes apart. The average of last two readings was used for analysis. Blood pressure was measured using a validated digital automatic blood pressure machine. Both systolic blood pressure (SBP) and diastolic blood pressure (DBP) were measured. High blood pressure was defined as SBP ≥ 140 mmHg and/or DBP ≥ 90 mmHg.²³

Statistical Analysis

Characteristics of all participants including demographic and anthropometric measurements were described by using means and standard deviations. Normality of data was checked using Kolmogorov–Smirnov test. The associations of dietary patterns quartile score with high blood pressure and obesity measures were evaluated by means of adjusted odds ratio (AOR) from multivariable logistic regression models, with quartiles of dietary patterns’ scores as independent variables. The Statistical Package for the Social Sciences (version 23) was used for all analysis, and a p-value of <0.05 was considered significant.

Results

Dietary Patterns of the Study Population

A total of 283 participants were recruited in the study, of whom 181 (63.9%) were females and 102 (36.1%) were males. Their mean BMI was 32.5 (31.7). About 21% had high blood pressure, 24% general obesity, 35% abdominal obesity by WHR and 31% abdominal obesity by WC (Table 1). Three major components or dietary patterns were extracted from principal component analysis (PCA). The data met all statistical assumptions for carrying factor analysis: Bartlett’s test of Sphericity confirmed that our data showed a patterned relationship among variables (p<0.001); Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy (MSA) was 0.689 which is greater than the recommended minimum value of 0.5.

Table 1 Demographic, Blood Pressure and Anthropometric Measures of the Study Participants (N=283)

Three dietary patterns were retained as suggested by the scree plot and Eigenvalues. The rotated components accounted for total variance of 74.6% of the food intake. The first dietary pattern is characterized by traditional foods consumed frequently in the community including porridge, maize stiff porridge, beans, milk, yoghurt and tea, we called this as “Maize, beans and dairy” explained 35.8% of the total variance. Pattern two was highly loaded by animal sourced foods such as beef and goat, we called this as “Meat based” accounted for 25.3% of the variance. The third dietary pattern accounted for 13.5% of variance characterized by high consumption of fruits and vegetables, and therefore we called this as “Fruits and Vegetables” (Table 2).

Table 2 Factor Loading Scores of the Types of Dietary Patterns Derived in Study Population (N=283)

Nutrient Intakes

The association of each dietary pattern with energy-adjusted nutrient intakes are illustrated in Table 3. The maize, beans and dairy dietary pattern was associated with higher intake of Iron, vitamin C, vitamin B6 and vitamin B12. The meat based dietary pattern was associated with higher intake of calcium and protein (p<0.001). Significantly higher intakes of protein, fiber and Vitamin B2 were found in the fruits and vegetables dietary pattern. Furthermore, the mean percentage of energy (% energy) from carbohydrates was higher in maize, beans and dairy dietary pattern. Percentage of energy from fat and protein were higher in fruits and vegetables dietary patterns and meat patterns compared to maize, beans and dairy dietary patterns.

Table 3 Nutrients Intake According to types of Dietary Patterns Among the Study Participants (N=283)

Association of Dietary Patterns and Metabolic Conditions

Table 4 reports the adjusted odds ratio (AOR) from multivariable logistic regression analysis after controlling for age, gender and energy intake of the participants. The analysis shows that participants in the third quartile (Q3) of maize pattern had higher odds of having abdominal obesity measured by WC (AOR=2.81; 95% CI: 1.09–7.26) compared to first quartile (Q1). Moving from first quartile (Q1) to fourth quartile (Q4) of the maize pattern score decreased the prevalence of high blood pressure (P-trend=0.019) and general obesity (P-trend=0.002), however, the odds of individual quartile were not significant. Similarly, participants in the third quartile (Q3) of meat pattern have increased odds of having abdominal obesity by WHR (AOR=5.03; 95% CI: 2.31–10.9). The analysis also found that participants in the third quartile (Q3) of fruits and vegetables dietary pattern have lower odds of general obesity (AOR = 0.25, 95% CI: 0.07, 0.9) and abdominal obesity by WC (AOR = 0.25, 95% CI: 0.09, 0.71). The odds of high blood pressure were not significant for all types of dietary pattern as shown.

Table 4 Association of Dietary Patterns Quartile Scores with Metabolic Conditions in Agro-Pastoral Communities (N=283)

Discussion

This study sought to identify dietary patterns among adults in agro-pastoral communities and examine whether these patterns were associated with some selected metabolic outcomes of high blood pressure and obesity. We derived three types of dietary patterns named as (1) maize, beans and dairy; (2) meat based; and (3) fruits and vegetables. Our findings suggest that the observed dietary patterns contributed to the development general and abdominal obesity in this community. We found that adherence to maize and meat based dietary patterns was associated with increased risk of obesity. In contrast, fruits and vegetables dietary pattern was associated with reduced risk of obesity. This study did not find the relationship between dietary patterns and the risk of high blood pressure. These findings were evident even after the control of effect of age, gender and total energy intake.

In comparison, the dietary patterns derived in this community showed some similarities with previous studies conducted among adults in rural and urban areas in Tanzania. The maize dietary pattern derived here is similar to the “traditional-inland” pattern identified by Keding et al,¹² and the “complex carbohydrate” pattern identified by Katalambula et al.¹¹ Although the maize pattern derived in this study is characterized by higher loadings of milk and dairy products, which is different with those reported in other studies in Tanzania. This could be due to ethnic differences as “Maasai” populations preferred milk and dairy products in their diet.²⁴ However, the reliance of milk in the traditional diet of pastoralists have been reduced over the years, and nowadays they are found to adhere to maize staples.¹⁵ The second dietary pattern in this study contained meat products which accounted for 25.8% of the total variance of foods. Meat consumption is one of the key characteristic of the previously called “Western” dietary pattern as reported in other studies.²⁵ Similarly, meat and meat products dietary pattern was previously reported in rural households in Kilosa district.¹³ Furthermore, we identified another dietary pattern dominated by fruits and vegetables food items. This was similar to “healthy” pattern identified by Katalambula et al,¹¹ and “Mchicha diet” pattern reported by Jordan et al.¹⁴

In general, this study suggests that the derived dietary patterns might have significant contribution to the development of metabolic conditions in agro-pastoral communities. We found a significant positive association between maize dietary pattern with obesity. The maize dietary pattern in this study is rich in beans, dairy products and many of carbohydrates rich foods such as maize meal and porridge. Similarly, Yu et al²⁶ and Zhang et al²⁷ have both reported that dietary patterns dominated by carbohydrates rich foods were associated with higher odds of obesity among adults and adolescents in China. In Tanzania, a study found positive association dietary pattern characterized by bread, black tea and cakes on BMI among women aged 16 to 45 years.¹² Therefore, a positive association between this dietary pattern and the risk of obesity was expected due to high kilocalories and nutrient content from these foods. Nutrients analysis shows that participants adhere to this pattern have higher percentage of energy from carbohydrates than other patterns. Foods that have high amount of sugars in this pattern contributed to higher energy intake such as black tea.

This study indicates that there may be negative health consequences in adherence to meat dietary pattern on obesity. This is consistent with a body of literature conducted in different countries.^25,28 The increased risk of obesity with high meat consumption seen in this study might be related to high fat and added salt or simply that regular meat consumption is an indication of wealth in this community.⁶ Moreover, vegetables dietary pattern appeared to be a protective factor for obesity. Similarly, a study done by Haidari et al in West Indian reported similar findings.²⁹ Previous studies found that fruits and vegetables dietary pattern are inversely related to body weight, waist-to-hip ratio and other metabolic conditions.³⁰ A protective effect from fruits and vegetables is probably due to higher intake of dietary fiber which can reduce glycemia from lower absorption and digestion of carbohydrates. However, in this study, vegetables dietary pattern is limited by higher intake of fats which may have counteracted the benefits that this pattern may had on obesity and blood pressure. Higher intake of fats from fruits and vegetables dietary pattern can be partly explained by cooking methods. Most of the participants prepared their vegetables by mixing with cooking oils. Therefore, vegetable intake may be acting as a marker for fats intake in this community.

The association between dietary patterns and high blood pressure was not significant in this study. Although the prevalence of high blood pressure decreased significantly across the maize pattern scores. Perhaps higher loadings of milk and dairy products in this pattern contributed to lowering the blood pressure levels. Previous body of literature showed that milk and dairy products is very important in the prevention of cardiovascular disease and lowering of blood pressure.^31–33 Diverse studies also showed suggestive evidence that dietary pattern dominated by milk and dairy products may reduce the risk of high blood pressure.^34,35 In addition, meat pattern was not found to contribute to risk of high blood pressure. This finding was not surprising as it was previously found that red meat consumption was not associated with hypertension among adults in Ngorongoro district in Tanzania.³⁶ Therefore, the association between dietary patterns and the risk of high blood pressure remains uncertain in this study population. Further longitudinal studies would be needed to understand the relationship between dietary factors and high blood pressure.

The strength of this study including the use of a validated food frequency questionnaire (FFQ) to determine the dietary patterns. The dietary patterns derived in this study were data-driven, which derived more reproducible and interpretable results than other types of dietary patterns. There are some important limitations in this study. First, this study was cross-sectional in nature, thus the causal relationship between dietary patterns and metabolic conditions could not be established. Second, the FFQ can give the reliable information on habitual consumption of foods, but the recall bias could not be avoided. Although we made efforts to control for confounders in this study, the effect of residual confounding could not be eliminated. Example, there may have been a healthy-adherer effect in people who adhered to a healthier diet. Such confounding factor was not measured in this study.

Conclusion

This study found three major types of dietary patterns among adults in agro-pastoral communities named as maize, beans and dairy; meat based and; fruits and vegetables. The traditional maize dietary pattern dominated maize, beans and dairy products was associated with increased risk of obesity. Higher adherence to meat based dietary pattern may increase the rise in prevalence of obesity. Only fruits and vegetables dietary pattern was found to protect against excess weight gain. This study did not show the association between dietary patterns and the risk of high blood pressure. However, the analysis suggests that there may be a negative association between maize, beans and dairy dietary pattern with high blood pressure, which needs to be confirmed in further studies. The findings may be used to guide the preventive nutrition interventions to curb the rise in obesity in Tanzania.