Variations in Nutrients Composition of Most Commonly Consumed Cassava ( Manihot esculenta ) Mixed Dishes in South-Eastern Nigeria

Variations in nutrient composition of most commonly consumed cassava (Manihot esculenta) mixed dishes in South-eastern Nigeria were determined. Four communities were randomly selected from each of the five states in the South-east. Focus group discussions (FGD) were conducted in each of the communities to determine commonly consumed foods and variations in recipes. 24-Hour dietary recall was conducted using 50 randomly selected households in those communities. Recipes collected during the FGDwere standardized, prepared, and chemically analysed using standardmethods. Data were analysed using descriptive statistics. Cassava-based dishes were the most commonly consumed in South-eastern Nigeria. Ninety-two percent of the study population ate cassava either in form of “fufu” (fermented cassava meal)/garri (fermented and roasted cassava meal) with soup or as “abacha” (tapioca salad). Commonly consumed soups were melon (Citrullus vulgaris) seeds, “ora” (Pterocarpus soyauxii), and vegetable soups. Seven melon seed, six “ora,” and four vegetable soups and five “abacha” variations were identified. Except for vegetable soup, coefficient of variation for moisture was <10%, while large variations (19–71%) were observed for energy and nutrients.These variations in cassava-based dishes need to be reflected in the country-specific food composition database to enable nutrient intake assessment or provision of dietary guidance using such food composition database as a reference material to be more effective.


Introduction
The need for good quality food composition data on commonly consumed dishes cannot be overemphasized.They are needed for a variety of purposes.For the nutritional epidemiologist, food composition data are used to estimate nutrient intakes of individuals or population groups in order to establish diet-disease relationships [1].They are used for establishment of dietary goals [2], while the dietitian employs them for dietary counselling [1].Inaccurate data will lead to overestimation or underestimation of dietary intakes.
Foods can be consumed as single foods or as mixed dishes/multi-ingredient foods and as such any comprehensive food composition database or table should contain the nutrient composition of foods as consumed by the population it serves.The nutrient composition of foods varies as a result of several factors.For single foods, nutrient variability has been attributed to inherent (age, maturity, genus, species, variety, and cultivar), environmental (soil, climate, pesticides, etc.), and processing (cooking methods, preservation methods, etc.) factors.However, for mixed dishes or multiingredient foods, the causes are much more complex [3].
Studies have shown that most commonly consumed traditional mixed dishes in Nigeria differ in several ways depending on the socioeconomic status of the family food provider, the prevailing food taboos, food prices, cultural and religious practices, level of nutritional knowledge, and food availability and seasonality [4,5].Variations in Nigerian local dishes are also due to the different ingredients (with different composition) used in preparing them, for example, type and quantity of animal protein, type of cooking oil, type of thickening agent, use of vegetables, and use of potash and seasoning/flavouring agents [6].These mixed dishes vary from simple (few ingredients) to very complex (many ingredients) ones.Holden [7] noted that a particular product can have 1-30 different ingredients.According to Vasilopoulou et al. [8], composite dishes show the greatest variations and, therefore, represent the least reliable data in food composition database.
Variation can be defined as deviation from a central position or value.Measures of nutrient variability are useful not only to database compilers, food analysts, and food manufacturers but also to dieticians, nutritionists, and researchers who use food composition database to plan and evaluate the diets of patients and clients with modified diets [9].In this situation, dietitians seek for mean values of nutrients in foods.The presence of a measure of nutrient variability such as standard deviation (SD) or coefficient of variation (CV) can be very useful to nutrition professionals to assess the reliability of mean values and provide practical dietary advice and guidance [3].Measurement of nutrient variability is of particular importance if it occurs in foods that are relied upon as major sources of nutrients for a specific population.Therefore, determining the nutrient variability in commonly consumed Nigerian dishes is a right step towards the development of a comprehensive, country-specific nutrient database that can be used to obtain good estimate of the dietary intake of individuals and/or population groups.

Area of the Study.
The study was carried out in the Southeast geopolitical zone of Nigeria.The zone comprises Anambra, Enugu, Abia, Imo, and Ebonyi states.It lies between the North-central and South-South zones of Nigeria.The Southeast zone is the home of the Ibos with a total land area of about 158,000 square miles (about 41,000 sq km).The Ibo population is estimated to be about 18 million [10].
Apart from trading, which is an old occupation among the Ibos, the traditional Ibo economy depends largely on root-crop farming where yam, cassava, and taro are the chief root crops.Hence, a typical Ibo meal includes a starch and a soup/stew prepared with vegetables to which pieces of fish, beef, goat, or other animal food sources are added.
Presently, an increasing number of Ibos engage in wage labour.Growing cities, expanding roads, construction of new industries, and oil exploration are creating many job opportunities.

Study Design.
A cross-sectional survey was adopted for food/recipe documentation, while quasi-experimental design was adopted for recipe harmonization/standardization and food analysis as well as dietary evaluation.

Study Population.
The population for this study consisted of all the communities in South-east geopolitical zone of Nigeria.

Study Sample.
Stratified sampling technique that involved two stages of purposive sampling and one stage of randomization was used.South-east zone was grouped into three clusters differing considerably in terms of food production and consumption pattern.The three clusters formed three strata based on similarities as follows: (1) Anambra/Enugu, (2) Abia/Imo, and (3) Ebonyi.In the first stage, four Local Government Areas (LGAs) were purposefully selected from each of the first two strata (Anambra/Enugu and Abia/Imo) and two were selected from the last stratum (Ebonyi).In the second stage, two communities were also purposively selected from each of the chosen LGAs (one urban and rural community).This implies that the documentation exercise was carried out in ten LGAs and in twenty communities in the study area.In the third stage, women were randomly selected from their clusters in each of the community's major market for the 24-hour dietary recall.One in every three women in each cluster was selected until the required number was obtained.

24-Hour Dietary Recall.
Food intake pattern at household level was determined with the use of 24-hour dietary recall.The 24-hour dietary recall was conducted on 50 randomly selected women representing 50 households in the community making a total of 1000 women/households.The information so gathered was used to produce food consumption statistics (list of foods/dish and frequency of consumption by families).From the food consumption statistics, the most frequently consumed foods/dishes were identified.The data presented in this study are based on the most commonly consumed food/dishes in each of the selected communities.

Focus Group Discussion
. The focus group discussion was used for documenting the recipes of the commonly consumed dishes and also to validate the information gathered from the consumption survey.The focus group discussion was conducted in all the selected communities.It comprised small groups of about 6-8 adult women selected with the help of community based facilitators.Each session lasted for 45-60 minutes.The commonly consumed dishes, as obtained from the food consumption statistics and a form containing all the food groups, served as the focus group discussion guide.Recipes of the commonly consumed dishes, their variations, and traditional methods of preparations were the major points supplied by the focus group.

Recipe Standardization.
The recipes collected were standardized using a modified National Food Service Management Institution (NFSMI) [11] method.According to NFSMI, a recipe standardization process can be summarized in three phases: recipe verification, product evaluation, and quantity adjustment.In this study, the two major stages of recipe verification phase (recipe review and recipe preparation) were particularly of interest.Recipe harmonization stage, though not part of recipe verification phase, was included in the study due to its relevance.
(a) Recipe Review.This was done by working on only one recipe at a time to find out if the recipe contained the following information: (a) recipe title, (b) recipe category, (c) ingredients, (d) weight/volume for each ingredient, (e) preparation instructions (directions), (f) cooking temperature and time, if appropriate, (g) serving size, (h) recipe yield, and (i) equipment and utensils to be used.
(b) Recipe Harmonization.Collected recipes of a particular cassava dish were critically examined.Recipes that were significantly different from others in terms of ingredients, quantity, quality, type, and number served as variations.Weights of each ingredient in similar recipes (recipe of a particular dish containing similar ingredients in terms of number, quantity, quality, type, and amount) were collated and harmonized.This involved determining the mean values of each ingredient in each set of similar recipes: mean = summation of ingredient weight (g) number of times it (ingredients) occured .
(c) Recipe Preparation.The standardized recipes of the most commonly consumed dishes were prepared using the traditional method obtained from FGD.

Method of Preparation.
Fry onion in heated palm oil for a minute and add the seasonings (for variations 1 and 3) or prepare an emulsion with palm oil, potash, and seasons (for variation 2).Add the soaked, washed, and drained cassava chips and mix properly.Serve hot garnished with prawn (for variation 1) or cold garnished with fermented oil bean seed (Pentaclethra macrophylla) and garden egg leaves or ukazi leaves (Gnetum africanum) (for variation 3).For variations 4 and 5, add soaked, washed, and drained cassava chips, onion, palm oil, and seasonings to the cowpea (boiled till tender), boil for four minutes, stir properly, and serve hot garnished with prawn.

Method of Preparation.
Boil beef till tender seasoning with salt.For variations 2, 4, and 6, melt the ground "ogbono" seed (Irvingia gabonensis) in the heated palm oil, add water (for variation 1) or meat and stock (for variations 4 and 6), and bring to boil.For variations 1, 3, and 5, mix "ukpo" (Mucuna urens) with palm oil, add to boiling soup water (for variation 1) or boiled meat and stock (for variations 3 and 5), and bring to boil again.Add stock fish head (all the variations) and stock fish (for variation 7) and boil for three minutes.Add melon seeds (gradually to avoid lumping) and bitter leaves (all the variations), "okpei" (Prosopis africana) (variations 1 and 2), onion (variations 3-7), smoked fish (variations 1 and 2), pumpkin leaves, crayfish, and seasonings (all the variations).
Stir until the mixture becomes homogenous and simmer for five minutes.Serve hot with cassava "fufu" or garri.
Method of Preparation.Boil, peel, and pound cocoyam into paste.Add salt to the beef and boil till it is done.Boil washed stock fish head for three minutes, add "ukpo" (Mucuna urens) mixed with palm oil (for variation 1) or add the stock fish head and the mixed "ukpo" (Mucuna urens) to meat and stock (for variations 3 and 5), and boil for five minutes.Add cocoyam paste in balls (for variations 2, 4, and 6) to meat and stock (for variations 4 and 6) or boiling soup water (for variation 2) and boil for five minutes.Add dry fish and boil for another five minutes (for variations 5 and 6).Add smoked fish (for variation 2), "ukazi" leaves (Gnetum africanum) (for variation 5), "uziza" leaves (Piper guineense), and "ogiri" (Ricinus communis) (for variation 6) and palm oil, crayfish "ora" leaves (Pterocarpus soyauxii), and seasonings (for all variations).Stir until the mixture becomes homogenous and simmer for three minutes.Serve hot with cassava "fufu" or garri.

Method of Preparation.
Season raw beef (for variation 3) and ice fish (for variation 4) and boil the beef till it is done.Tie the pumpkin leaves firmly with cord, boil in a potash and water solution for eight minutes, and drain and cut into tiny pieces (for variations 1 and 2).Bring the soup water to boil, add stock fish bone (for variation 1), stock fish head (for variations 2 and 4), boiled meat and stock (for variation 3) or boiled ice fish and stock (for variation 4), and boil for five minutes.Add the precooked pumpkin leaves (for variations 1-2) or sliced raw pumpkin leaves and okra fruits (for variation 3).Add smoked fish and dry fish (for variation 2), garden egg leaves, African spinach, and melon seeds (for variation 4), crayfish (for variations 3 and 4), onion (for variations 1, 2, and 4), palm oil, "okpei" (Prosopis africana), and seasonings (for all variations).Simmer for five minutes.Stir to mix well and serve hot with cassava "fufu" or garri.The prepared dishes were cooled, properly packaged, and taken to the laboratory for chemical analysis.Preparations were done in the Department of Home Science, Nutrition and Dietetics, University of Nigeria, Nsukka.

Sample Preparation.
The prepared dishes were homogenised with the use of electric blender.Five grams of each sample was used to determine the actual moisture at 100 ∘ C, while the rest were dried at 60 ∘ C in an air oven drier, milled, packaged in an air-tight container, and taken stored in the freezer until being ready for chemical analysis.

Chemical Analysis.
Protein, moisture (actual and residual), fat, ash, soluble and insoluble dietary fibre, minerals (calcium, magnesium, iron, zinc, sodium, potassium, copper, and phosphorus), and vitamins (A, B 1 , B 2 , and B 3 ) were determined according to the method of AOAC [12].Moisture was determined using the air oven method.Crude protein and fat were determined by Kjeldahl procedure and Soxhlet solvent extraction method, respectively.Dietary fibre (soluble and insoluble) was determined by enzyme gravimetric method of Prosky et al. [13].Ash was determined by incineration of samples in a muffle furnace at 550 ∘ C for six hours.Available carbohydrate was calculated by difference       100 − (moisture + protein + fat + ash + dietary fibre).Mineral elements were determined using the Atomic Absorption Spectrophotometer (Perkin-Elmer model 3110, USA).Phosphorus was determined using the vanadomolybdate method.Vitamins were determined using high-performance liquid chromatography (HPLC).All samples were analysed in triplicate.The proximate, mineral, and vitamin values in Tables 5-7 are, therefore, means of three analyses.

Determination of the Nutrient Content of the Recipes as
Consumed.Water conversion factor was calculated using the following formula: Values obtained through chemical analysis were multiplied with the WCF to obtain nutrient content of the recipes as consumed.
2.9.Statistical Analysis.The data collected were analysed using frequencies, percentages, and means.
Table 7 shows the energy and nutrient content of six variations of "ora" (Pterocarpus soyauxii) soup.CV% was least for moisture (4.5%) and thiamin (4%) and highest for phosphorus and potassium (45%) followed by protein and calcium (38%) and magnesium (35%).CV% for other nutrients ranged from 25 to 29%.The protein and all the mineral and vitamin contents of variation 1 of "ora" soup cooked without beef were lower than those of variation 3 cooked with beef.Variations 2, 4, and 6 were all thickened with cocoyam paste; however, the protein contents varied.The protein contents of variations 4 (4.2%) and 6 (4.3%) cooked with beef were more than that of variation 1 (1.8) cooked without meat.A similar pattern of variation was observed for these three variations in energy and all the other nutrients.Variation 2 also had more moisture (89%) compared to variations 4 (82%) and 6 (80%).Of all the "ora" soups thickened with "ukpo" (Mucuna urens) (variations 1, 3, and 5), variation 5, with additional protein source and less moisture content, had a higher concentration of all nutrients.Soups thickened with "ukpo" (Mucuna urens) had better mineral and vitamin contents compared to those thickened with cocoyam paste.The variations were wider among macroelements.The variation in protein content appeared to be related to the protein sources rather than the thickeners in this case.
The nutrient composition of the four varieties of vegetable soups differed widely (see Table 8).CV% ranged from 17% for moisture to 57% for carbohydrate.Again, vegetable soups with lower moisture (58-60%) content (variations 1 and 2) had higher nutrient content than variations 3 and 4 with higher moisture (>80%) content.

Discussion
The high frequency of cassava-based dishes in this study is not surprising.A similar report of the frequency of consumption of cassava among children (92%) and women (95%) in Akwa Ibom State in the South-South zone of Nigeria was given by De Moura et al. [14].According to FAO and IFAD [15], cassava food products are the most important staples for rural and urban households in Southern Nigeria.This is because cassava as a food crop fits well into the farming system of the small-holder farmers in Southern Nigeria and is available all the year round, thus providing household food security.Compared to grains, cassava is more tolerant to low soil fertility and more resistant to drought, pest, and diseases [16].It is also worthy of note that rice-based dishes were as popular as yam-based dishes.Although yam is one of the most prized and cherished root crops in South-eastern Nigeria, it appears  (1) vegetable soup with pumpkin leaves and stock fish bone; (2) vegetable soup with pumpkin leaves, stock fish head, dry fish, and meat/ice fish/smoked fish; (3) vegetable soup with pumpkin leaves and okra with crayfish and meat/fish; (4) vegetable soup with spinach and garden egg leaves with melon seed, crayfish, and ice fish.CV%: coefficient of variation.that its consumption is dwindling.The popularity of rice according to Deslisle [17] has been attributed to urbanization and ease of preparation.Bitter cassava, which is predominantly the variety cultivated and consumed in South-eastern Nigeria, is rarely eaten alone.When used as snack food (e.g., soaked garri or tapioca salad, "abacha"), it is eaten with coconut, fish, milk, groundnuts, or pork meat.Tapioca ("abacha") salad is sometimes eaten as a snack or a full meal.As observed from this study, the most popular cassava-based dishes are those eaten with a variety of soups (e.g., garri or "fufu").Although cassava by itself is known to have a low protein/energy ratio, its contribution to nutrient intake cannot be viewed in isolation.Soups and other ingredients are the major nutrient contributors to cassava-based dishes; thus, their contribution will depend on the type of soup and other food materials accompanying the cassava products.
Nigerian soups vary in complexity.They are made with multiple ingredients of varying proportion and types as well as varied methods of preparation.The observed variations in the commonly consumed soups in this study confirm the observation made by Ene-Obong et al. [6], where sources of variation were observed to be in the use, type, and quantity of animal protein and use of vegetables and thickeners.Others included seasonings and ingredient number as well as quantity.
The observed high variability in the nutrient composition of these soups suggests the need to include these different variations in a country-specific food composition table in order to avoid overestimation or underestimation of energy and nutrient intakes.According to Williamson [18], variation in nutrient composition limits the usefulness of FCDBs for both scientific and regulatory purposes.Thus, documenting and including the same dishes cooked with varying proportions and types of ingredients will help to reduce the errors in the estimation of nutrient intakes of population groups.
Based on the data presented, it was observed that moisture content, protein source, and quantity/type of ingredients and thickeners used appeared to cause the greatest variations in the energy and nutrient contents of the dishes.Generally, soups with less moisture had higher energy and nutrient contents compared to those with higher moisture content.This result is expected, since the nutrient composition of a food is inversely related to its water content.This shows that the common practice in poor households of adding water to soups and stews "so that it will be enough" according to a focus group discussant dilutes the nutrient concentration of the soup or stew and so should be discouraged.Cassava is a very poor source of most micronutrients of public health importance [19].Overdilution of the accompanying soup could, therefore, have serious implication on the nutrient intake of household member in South-east Nigeria, particularly the vulnerable groups.Sanghvi and Murray [20] observed that infants and young children are frequently fed with these meals.According to Okeke et al. [21], such foods are not suitable for feeding young children with small stomach capacity.Another implication of this finding is that there will always be need to correct for the moisture content of these soups even when the ingredients used in preparation are similar in order to avoid overestimation of intakes.
The use of animal protein source in Nigerian soup preparation deserves special mention as a result of varying practices.Some prepare soups with little or no animal protein, while others prepare soups with an assortment of animal proteins.Apart from some religious groups like the Seventh Day Adventist, most households in South-eastern Nigeria use ground crayfish in cooking.This is a major animal protein source in the South-eastern cuisine.The ways in which the proteins are served also differ, since members of the household do not always have equal portions to consume during meals.This makes it extremely difficult to estimate what a standard portion is for any age group.Therefore, only the liquid part of the soups was analysed, while the big lumps of meat/fish were excluded.This was done to take care of individuals who may consume a cassava-based meal without any visible serving of meat/fish.The implication of this (analysing only the liquid part of the soups) is that the nutrient content of whatever quantity of meat/fish consumed by an individual will have to be calculated separately and added to whatever the soups contribute.Despite the fact that meat/fish was not included in the analyses, it was observed, however, that the number of animal protein sources present in a recipe still determined the nutrient composition of the prepared dish.Soups cooked with smoked fish had less protein than other variations prepared with beef, dry fish, and/or stock fish.The meat/stock fish/nitrogenous extractives as well as small pieces of fishes that mixed up with the soup at the preparation stage could be responsible for the higher protein values in soups prepared with more animal protein sources.Soups prepared with meat, dry fish, and stock fish also had higher mineral values, particularly calcium, iron, potassium, magnesium, and phosphorus, since they are excellent sources of these nutrients.The implication of this finding is that women and children in some traditional societies who are denied meat and fish because of food taboos [22] are not completely left out so long as the meat or fish was among the ingredient used in preparing the meal even though the quantity may still not be adequate.
The similarity in the protein and fat content of melon ("egusi") soup thickened with "ukpo" (Mucuna urens) and the one cooked with melon seed alone could be attributed to the fact that both "ukpo" (Mucuna urens) and melon are from the same food group (oil seeds and nuts) and the fact that both variations of melon soup were cooked with substantial amounts of animal proteins.However, the mineral profile of melon soup thickened with "ukpo" (Mucuna urens) was higher."Ogbono" (Irvingia gabonensis) also shares the same food group with "ukpo" (Mucuna urens) and melon seed.However, the protein content of the soups thickened with "ogbono" (Irvingia gabonensis) was lower.None of them contained meat.
There was not much difference in the energy content between soup thickened with "ukpo" (Mucuna urens) and those thickened with "ogbono" (Irvingia gabonensis).Soups thickened with "ukpo" (Mucuna urens) in most cases contained less carbohydrate than those thickened with cocoyam paste.Cocoyam as a tuber contains more carbohydrate than seeds and nuts.Furthermore, large quantity of cocoyam was used per pot of soup."Ukpo" (Mucuna urens), because of its high dietary fibre content [23], has a greater water absorbing capacity and is more viscous than cocoyam paste.It is, therefore, added in smaller quantities.The cocoyam thickened soups were, therefore, more energy-dense than the variations thickened with "ukpo" (Mucuna urens).Type of soup thickener should, therefore, be an important consideration when planning low-/high-energy diet.
A slight change in the ingredient composition of a recipe automatically reflects in its nutrient composition as seen in nutrient values of different recipe variations.The quantity of ingredient used reflected in the nutrient content of different variation of a particular recipe.This explained why the nutrient content of "abacha" prepared with more cowpea had more protein than the variation with less or without cowpea."Abacha" prepared with more cowpea contributes 27.5 and 30% of the protein required daily by adult males and females, respectively, while another variation of the same dish ("abacha" and vegetable) contributes only 4.5 and 4.9% of the protein required by the same population groups.This is understandable, since legumes (cowpea) are better protein sources than fresh vegetables and tubers.Increasing the number of ingredients in a recipe should, therefore, be encouraged over the use of basic ones.This will also encourage dietary diversity in cassava-consuming communities, since it has been shown that increased reliance on cassava leads to less dietary diversity, an association that identifies cassava consumers as a vulnerable population that may require interventions to improve nutrition [16].

Conclusion
Cassava dishes in form of "abacha" (tapioca salad) and "fufu"/"garri" with soups were the most commonly consumed in South-eastern Nigeria and their recipes showed great variability.Factors causing the variability included the moisture content of the dish and the type and number of ingredients and the thickening agents used.These variations in commonly consumed dishes cannot be neglected.They should be reflected in the country-specific food composition database so that nutrient intake assessment or provision of dietary guidance using such food composition database as a reference material could be more effective.

Table 1 Ingredient
Quantity (g) in different variations

Table 8 :
Energy and nutrients composition of variations of vegetable soup.