Optimization of dabi teff-field pea based energy and protein dense novel complementary food with improved sensory acceptability using D-optimal mixture design

Abstract

Protein-energy malnutrition is unacceptably high among children in developing countries due to inadequate required nutrients and poor quality of complementary foods characterized by low protein and energy density and often monotonous. Thus, this research was aimed at examining the potential of including dabi teff, the underutilized/forgotten crop into pre-processed local food crops viz., germinated maize, roasted barley, roasted field pea, dehulled oats and linseed to develop energy and protein-dense optimized novel complementary food with improved sensory acceptability. Nutrisurvey software was employed to define ranges and they were constrained at 20–35% dabi teff, 0–30% field pea and 5–20% maize, while the rest were set constant at 25% barley, 15% oats and 5% linseed. Eleven experimental runs were generated from the six mixture components using D-optimal mixture design, Stat-Ease Design Expert ® software version 11. A 5-point Hedonic scale was used to evaluate the sensory attributes. ‘Scheffe’ regression was used to fit and test model adequacy and numerical multi-response optimization was performed to identify optimal points using the Design expert. Field pea and linseed contained significantly higher (P < 0.05) protein at 20.95% and 20.57%. The newly formulated products contained significantly higher protein (1.4–1.6 times) and protein density (1.31–1.56 times) as compared to the control and fulfilled the recommended standard. The optimal was identified at 34.66% dabi teff, 25% barley, 15% oats, 15.34% field pea, 5% linseed and 5% maize flour ratios with response values at overall optimization to be 5.57% moisture, 15.74% protein, 5.09% fat, 2.26% ash, 2.88% fiber, 73.05% carbohydrate, 380.43 kcal/100 g energy and 4.12 sensory acceptability score and it contained an energy density of 1.27 kcal/g and protein density of 4.14 g/100kacl. These findings showed that optimized dabi teff-field pea based novel complementary food can be used as a sustainable food-based strategy to combat protein-energy malnutrition among children in developing countries.

Highlights

• •Food-based approach is a better strategy where it uses a combination of production and consumption of nutrients rich foods among children, pregnant and lactating women.
• •FAO/WHO guideline describes mixtures of cereals, legumes; pulses/oilseed can constitute an appropriate source of nutrients and energy, essential fatty acids with many health benefits.
• •Optimized dabi teff-field pea based novel complementary food can be used as a sustainable food-based strategy to combat protein-energy malnutrition among children
• •Dabi teff, field pea, linseed crops which were super source of protein, fat and ash (minerals) combined with barley, maize and oats were found to exhibit superior quality over the traditional complementary foods and were shown to develop protein and energy dense optimized novel complementary foods to be used by children in poor setting.

1. Introduction

Malnutrition, in all its forms, remained a major public health problem globally and the world is off-truck or at slow momentum to achieve the Zero-Hunger target by 2030 [1]. The problem is further worsened by the COVID-19 pandemic where the number of undernourished people continued to rise since then. In most African countries, the vicious cycle of malnutrition is endemic due to poverty, food insecurity and infectious diseases. Protein-energy malnutrition, ‘the silent emergency of the world’ is an immediate cause of child mortality, morbidity and anthropometric deficits; stunting, wasting and underweight. It had hunted and continued to hunt mankind since the dawn of history and is by far the most lethal form of malnutrition [2,3].

Indeed, under-nutrition affects several millions of people worldwide where mainly children under-five years are among the most visible victims to nutritional deficiencies and most susceptible to protein-energy malnutrition-related growth impairments due to their high demand for protein and energy during this period [4]. In 2020, 22% of under five children were stunted, 6.7% were wasted and 5.7% were overweight worldwide [1,5]. Overall under-nutrition contributes to about 45% of child deaths in under-five years resulting in 3.1 million deaths every year [6]. In Africa, 61.4 million of under-five children were stunted, 12.1 million were wasted and 10.6 million were overweight [5]. According to the Ethiopian Public Health Institute and the International Classification of Functioning, Disability and Health (ICF) survey [7], 37% of children under-five years were stunted, 21% were underweight, 7% were wasted, and 2.9% were overweight.

The importance of proper nutrition in the first 1000 days of life (conception up to 24 months) and beyond is well-recognized in children. Growth faltering occurred during this critical period would end up with irreversible stunting, impaired cognitive development, diminished learning and earning capacity, greater susceptibility to infectious disease and increased morbidity and mortality as shown by many longitudinal studies [8,9]. Catch-up growth can be attempted but it is mostly at the later stage of growth faltering which is commonly observed during recovery periods after illness, after addressing causes of growth failure and post-natally after severe intrauterine growth restriction [10] and ending impractical and even poses a later-life risk of chronic disease in most cases [11].

The significantly higher prevalence of malnutrition in under-five children occurs during or after the introduction of complementary foods because of the poor quality and inadequate required nutrients for optimal growth and development [12]. Traditional complementary and weaning food in developing countries including Ethiopia are characterized by low nutritive value (often monotonous), low protein and energy density, high bulk density, high viscosity, and high anti-nutrients and there is a suboptimal complementary feeding practice [13,14].

Commercial fortified food products such as powdered milk and peanut butter and supplementation such as ready-to-use therapeutic foods (Plumpy’Nut, F-75, F-100) were proposed as effective strategies of combating child malnutrition where these strategies posed limitations in developing countries. In Ethiopia, efforts were being made in implementing a multi-sectorial plan of high impact nutrition-specific and nutrition-sensitive interventions such as the [15] and the [16] to end child under-nutrition including stunting by 2030 [17]. However, the country is still experiencing one of the worst situations where protein-energy malnutrition is unacceptably high among under-five children.

The food-based approach has currently received much attention as a better strategy where it uses a combination of production and consumption of nutrients rich foods among children, pregnant and lactating women [18,19]. It is considered as the best approach since it is cost-effective, sustainable and can be adapted to different cultural, dietary traditions and locally feasible strategies. This approach focuses on the combination of diverse staple/indigenous foods that can meet nutrient requirements and is based on the fact that people eat foods, not nutrients or supplements [20]. Diverse foods also give infants and young children a chance to appreciate various food flavors and textures, all of which are vital for later developing healthy eating habits [21]. The Food and Agriculture Organization and World Health Organization (FAO/WHO) [22] codex alimentarius on complementary feeding guidelines describes that mixtures of cereals, legumes, pulses/oilseed can constitute an appropriate source of nutrients and energy, essential fatty acids and limiting amino acids with many functional and health benefits as well as improved organoleptic characteristics. Therefore, there is a need to assure both nutrient densities and sensory acceptability when developing new complementary products.

Local foods can be modified through the use of various processing techniques like germination, soaking, roasting, fermentation, dehulling and extrusion cooking which can result in the reduction of anti-nutrients, improves energy and nutrients densities of complementary foods, increase bioavailability of key micronutrients as well as enhance sensory and palatability of foods [2].

Across the world, many plant-foods are underutilized/forgotten while they can play a substantial role in ensuring food and nutrition security, having medicinal value and sources of income generation [23]. The sub-Saharan African region is endowed with a rich diversity of such plant-foods, where little or no attention was given in terms of research and development and policy framework that could promote their extensive agricultural production, marketing, industrial utilization as well as home consumption.

In Ethiopia, the food culture of teff is both a historical, integral part of the country's antiquity and mostly traded domestically where about more than 70% of the Ethiopian population use teff as a traditional staple meal [24]. The gluten free property of teff grains as reported by Bultosa [25] qualifies the crop as a medicinal food to be used as a good alternative diet for celiac disease patients. Dabi teff (Eragrostis teff), a farmer variety teff, is one of the underutilized/marginalized or forgotten food crops grown in Ethiopia (Oromia region) that could be utilized to combat protein-energy and micronutrient malnutrition and can be used for the preparation of important traditional healthy foods such as injera/budena (pancake-like local bread), kitta/maxino/cacabsa (unleavened bread), atmit or muk/mooqa (gruel), fetfet/fafato, cumboo and cafaqoo (the favorite cultural dishes) and bread for all age groups including children in Ethiopia [26,27].

Dabi teff is the ‘afaan oromoo language ' name for an early maturing variety of dark red teff. The early maturing property makes the crop to be harvested twice within one rainy season (at early rain fall called “daabi gannoo” and late rain fall called “daabi birraa”) (personal experience). An early maturing and low rainfall requirement makes dabi teff a unique variety teff amenable to climate-smart agriculture. Farmers in Wollega and Illuababor, western Ethiopia, cultivate dabi teff either for its grain seeds or its straw. The contribution of dabi teff grains to household food and nutrition security is considerable when viewed through a food security lens while the straw is composed of fine stems used for plastering (finishing) mud hut walls during house construction and rated higher among others for animal fodder, especially oxen feeding (personal experience). There are many social beliefs regarding the nutritional and health benefits of dabi teff among rural elderly people in particular and the consumers in general where they prize it as a medicinal food (Personal communication).

The key underlying cause of child malnutrition among households with low socio-economic status is a lack of purchasing power of quality foods for their children and they consume foods of low nutritive value. Price-wise, the red teff varieties grown in Ethiopia including dabi teff are the cheapest. For example [28], reported that the market price of mixed teff and red treff were 24% and 55% less than the white teff. This statement specifies that the price of dabi teff is quite affordable for all socio-economic classes. In addition to lower prices, red teff varieties have been implicated in the low incidence of anemia in Ethiopia which is presumed to be due to the grain's high iron content [24].

Some previous studies have been conducted on the nutrient contents of different teff varieties grown in Ethiopia with comparable results with the present finding [25,29,30] and the formulation of complementary food by inclusion of the different teff varieties has also been conducted [31,32] with lower results with the current finding. These scholars obtained the teff varieties from Ethiopian agricultural research centers some of which are released and others on the accession stage. Dabi teff is however traditionally known and acceptably consumed among the producers and the samples were directly collected from the popular market. It is therefore, we suggest the crop needs to be integrated into the Ethiopian agricultural research program for its intensive research, extensive production and industrial utilization.

Despite the social beliefs then, to date, the potential of including dabi teff into complementary flour formulations in combating protein-energy malnutrition in infants and younger children is not yet examined. Thus, this research was aimed at examining the potential of including dabi teff, the underutilized/forgotten food crop into pre-processed local food crops viz., germinated maize, roasted barley, roasted field pea, dehulled oats and minimally cooked linseed to develop energy and protein dense optimized novel complementary food with improved sensory acceptability and finally to suggest the potential of dabi teff-field pea based novel complementary food to be used as a sustainable food-based strategy to combat protein-energy malnutrition among children in less developed and resource-limited countries including Ethiopia.

2. Materials and methods

2.1. Food crops sample collection

The food crops viz., dabi teff (Eragrostis teff (Zucc.) farmer variety), maize (Zea mays L.), barley (Hordeum vulgare), white field pea (Pisum sativum), oats (Avena. sativa) and linseed (Linum usitatissimum) were purchased from an open market of Nedjo town, Oromia, Ethiopia which is located at 575 km away to the west of Addis Ababa where Nedjo district is a potential dabi teff growers. The availability of these crops in the area was verified by an inspection survey of Nedjo market as well as by asking mothers about the traditional usage of the crops as complementary foods and their affordability. Literature-based nutrient content of each crop was taken into account during the selection process which was later verified through compositional analysis (Table 1) and the crops were wisely selected to add one or more essential nutrients to their mixture for complementation. Apparently, uninfected samples were purchased from the center and corners of the market on two different market days to assure representativeness. Each piece of the six crops purchased from the different corners was packed separately in polyethylene bags and transported to Ethiopian Agricultural Research Institute (Food Science and Nutrition Laboratory) and mixed using a laboratory mixer (Hot Sale DZ-10 L laboratory mixer, China) to get a homogeneous, 3 kg of composite crop samples each. All laboratory analyses of the macro-compositions and sensory acceptability were conducted in the stated laboratory which was certified by the International Organization for Standardization (ISO), ISO-17025:2017 by the International Laboratory Accreditation Cooperation (ILAC).

Table 1.

Macro-composition and energy values of the processed individual flours.

Individual flours							Gross Energy (kcal/100 g)	PER (g/100 kcal)	PEL
	Macro-composition (%)
	Moisture	Protein	Crude Fat	Total Ash	Crude Fiber	Carbohydrate
Dabi teff	9.03 ± 0.10a	10.74 ± 0.04a	3.94 ± 0.58a	3.96 ± 0.03a	3.76 ± 0.16a	68.57 ± 0.51a	352.70a	3.045	12.180
Roasted Barley	4.32 ± 0.33b	11.62 ± 0.03b	2.58 ± 0.43a	1.73 ± 0.09b	2.12 ± 0.07b	77.63 ± 0.06b	380.22b	3.056	12.224
Dehulled Oats	7.16 ± 0.06c	12.45 ± 0.06bc	5.69 ± 0.03a	1.92 ± 0.01bc	2.78 ± 0.07c	70.00 ± 0.09a	381.01bc	3.268	13.071
Roasted Field pea	5.80 ± 0.01bc	20.95 ± 0.02d	2.74 ± 0.01a	2.60 ± 0.03d	7.46 ± 0.07d	60.45 ± 0.08c	350.26a	5.981	23.925
Cooked Linseed	6.32 ± 0.25ce	20.57 ± 0.30de	36.08 ± 1.45b	3.45 ± 0.01e	5.34 ± 0.11e	28.24 ± 1.61d	519.96d	3.956	15.824
Germinated Maize	7.68 ± 1.13ac	8.58 ± 0.13f	4.61 ± 0.95a	1.42 ± 0.01bf	2.47 ± 0.03bcf	75.24 ± 0.03be	376.77be	2.277	9.109

Key: Values are means ± standard deviation of the triplicate determinations. Values in the same column followed by different superscripts are significantly different at P < 0.05. PER-protein energy ratio, PEL-protein energy level.

2.2. Samples processing

It is not only the proper selection of nutritionally promising crops, but it is also the processing techniques applied that determine the quality of the final product. Accordingly, the collected crop samples underwent various controlled processing techniques. In brief, dabi teff was manually cleaned by winnowing to remove chaffs, straw, dust and other extraneous and washed with tap water and sundried while the other cereals (maize, barley and oats) and the legumes (field pea and linseed) were sorted out from sands, sticks, stones and defective seeds, later washed and sundried for two days and were ready for further individual processing (Fig. 1). Due to the small size of dabi teff seeds, it was made into whole-seed milled flour which could be the reason for the higher nutrient contents of the crop. Barley and oats samples were soaked in clean tap water for 2 h and the water used for soaking was drained off and the crops were immediately decorticated (while the seeds were still wet) using a wooden decorticator and the hulls were removed by winnowing.

Fig. 1.

Process flowchart diagram of the crop samples and Formulations.

Germination of maize seeds was adopted from the method previously described by Rasane et al. [33] with little modification. Briefly, the maize grains were soaked in water (1:3 w/v) for 3 h to achieve hydration, drained, spread on a clean jute sack placed on a wooden platform and covered with another jute sack for germination at room temperature (25 °C ± 2 °C) for 72 h s and water was sprayed every 12hr to stay humid (60% relative humidity). This 72 h s was used because Rasane et al. [33] reported that the maximum amylase activity was obtained at 72 h s of germination time. After 72 h s, the germinated seeds were rinsed, drained for 10 min, transferred to aluminum trays and dried in an air oven at 40 °C for 5hr to terminate the germination process. The dried germinated sample was further roasted at 120 ± 5 °C for 10 min and allowed to cool.

The crop samples viz., barley, oats, field pea and germinated maize were minimally roasted in an oven at 120 °C for 20 min until light brown colored and then cooled to room temperature (25 ± 2 °C) as described by Rasane et al. [33]. The roasting process was carefully controlled to prevent over-roasting and the formation of undesirable Maillard reactions that may lead to protein quality damage. The linseed sample was minimally cooked for 5 min at 90 °C as described by Ref. [34] with little amount of water to condition the seeds to release oil from the oil cells and later sundried without draining the water used for cooking.

2.3. Flour preparation and handling

All the six processed samples were milled into flours using a standard miller (Cyclotec 1093 sample mill, Foss Analytical, Sweden) to obtain smooth and consistent particle sizes and sieved through a 0.5 mm mesh sieve size. The flours were then packed in air-tight high-density polyethylene bags (AACC, 2000), separately coded and stored safely at room temperature till formulation.

2.4. Experimental design

2.4.1. Flour formulation

About 100 g of flour from each of the six components was taken separately before generating the formulations to determine the macro-compositions of each sample (Table 1).

In the current study, Stat-Ease Design Expert® software version 11 (D-optimal, Minneapolis, USA, 2018) was used to generate the formulation design matrix (Table 2). Defining the range of each component in the formulation was based on three considerations including, first targeting to attain the FAO/WHO [22] codex alimentarius guideline secondly targeting the Ethiopian complementary feeding guideline [35] and thirdly based on the macro-nutrient contents of the individual components (Table 1).

Table 2.

D-optimal mixture design matrix, formula code, mixture ratio, control and individual flours and constraints with their limits.

Std. order	ID	Run order	Mixture ratio (%)						Limits of mixture components
Standard	ID	Formulated code	X1	X2	X3	X4	X5	X6	Total	Constraints	Lower	Upper
2	0	F1	27.5	25.0	15.0	15.0	5.0	12.5	all	X1	20	35
7	10	F2	26.123	25.0	15.0	23.877	5.0	5.0	100	X2	25	25
10	4	F3	30.457	25.0	15.0	19.543	5.0	5.0		X3	15	15
11	5	F4	20.000	25.0	15.0	29.988	5.0	5.012		X4	0	30
1	2	F5	28.793	25.0	15.0	6.207	5.0	20.0		X5	5	5
9	3	F6	35.000	25.0	15.0	11.884	5.0	8.116		X6	5	20
8	1	F7	21.705	25.0	15.0	13.295	5.0	20.0
4	7	F8	34.385	25.0	15.0	6.476	5.0	14.139
5	8	F9	20.000	25.0	15.0	19.54	5.0	15.46
3	6	F10	22.735	25.0	15.0	22.515	5.0	9.75
6	9	F11	29.604	25.0	15.0	10.076	5.0	15.32
		Control	0	80	0	0	0	20
		Dabi teff	100	0	0	0	0	0
		Barley	0	100	0	0	0	0
		Oats	0	0	100	0	0	0
		Field pea	0	0	0	100	0	0
		Linseed	0	0	0	0	100	0
		Germinated maize	0	0	0	0	0	100
		Optimized flour	34.66	25	15	15.34	5	5

Key: Std-Standard, ID-design identity, x₁-Dabi Teff, x₂-Barley, x₃-Oats, x₄-Field Pea, x₅-Lin Seed, x₆-Germinated Maize, F1–F11-Formula Code for each Formulation.

The macronutrient contents analysis results of individual components (Table 1) were recorded (customized) into Nutriurvey analysis (Version 2007) (Table 3). By estimating the amount of a meal to be consumed by 1–3 years old children to be 75 g (solid portion) per meal and adjusting for the required number of meals per day, several trials (iterations) were made to define ranges of major components (dabi teff, filed pea and germinated maize) by entering range related amounts into the software combined with the constant components and examine the output where finally the generated output showed the percentage fulfillment by the meal of the mix as compared to the recommended daily energy and recommended nutrients intake (RNI) by Pan American Health Organization and World Health Organization (PAHO/WHO) [36] for 1–3 years children to be 84–152%, 90–188%, 107–195%, 41–61%, 612–1045%, 133–225% and 229–400% for energy, protein, carbohydrate, fat, iron, zinc and for magnesium, respectively (Table 3) which corresponded to 20–35% of dabi teff, 0–30% of field pea, 5–20% of germinated maize, 25% of barley, 15% of oats and 5% of linseed meal mixture, respectively and these were used as constraints for generating the formulation matrix (Table 2).

Table 3.

Nutrisurvey analysis of the food records (customized) to define the range of each component in the formulations.

		Calculated energy and nutrient values of the formulated meal by Nutrisurvey
Components	Proportion (%) (lower-upper)	Amount eaten range/day (g)	Energy range (kcal)	CHO range (g)	Protein range (g)	Fat range (g)	Iron range (mg)	Calcium range(mg)	Zinc range (mg)	Magnesium Range (mg)
Dabi teff	20–35	45–79	158.7–278.7	30.9–54.2
Roasted Barley	25	56	212.9	43.5
Dehulled Oats	15	34	129.5	23.8
Roasted Field pea	0–30	0–68	0–238.3	0–41.1
Cooked Linseed	5	12	62.4	3.4
Germinated Maize	5–20	11–45	41.4–169.5	8.3–33.9
Analyzed value (range) for the formulated meal using our components			604.9–1091.5a	109.9–199.9 (73–74%)	19–39.8 (13–15%)a	10–14.8 (15-12%)	49–83.6a	87.8–176.1	4–6.7a	183–320.3
Recommended value for 1–3 years old child as generated by the ‘Nutrisurvey’ software			718.3	102.5 (>55%)	21.2 (12%)	24.4 (<30%)	8	600	3	80
Percentage fulfillment from the formulated meal (range) (lower-upper) %			84-152a	107-195a	90-188a	41–61	612-1045a	15–29	133-225a	229–400%

^aShowing the blended complementary flours would be promising to attain the FAO/WHO recommendations.

D-optimal mixture design was run to generate the formulation matrix using the range defined and later used to optimize the formulation. The six components were constrained to generate a total of 16 experimental runs with 6 central points and 5 replicates to provide 11 experimental runs with an estimate of pure error and 5 lack of fit points. The amount in a gram of individual component generated in each formulation (experimental run) was carefully weighed on a digital balance gravimetrically and put together after which it was thoroughly mixed using an electrical blender for 3 min at 200 rpm to homogenize and finally packed and sealed in high-density polyethylene bags and stored in a refrigerator at 4 °C till analysis.

2.4.2. Porridge preparation

A thick and consistent porridge was prepared by mothers acquainted with good cooking skills from all the formulations (Table 2) and the control flour following the method outlined by Onabanjo et al. [37] with slight modifications. Briefly, 300 g of the composite flour was mixed with 500 mL of clean tap water in a saucepan to make a slurry and put aside. 800 mL of the water was boiled in a stainless steel pan and once the water reached boiling point, the previously prepared slurry was added to the boiled water and allowed to cook for 10–15 min on an electric stove with continuous stirring using a wooden stick (special stirring wood, named ‘ boojjitoo in afaan oromoo language’) to avoid coagulation until the desired consistency was reached. The porridge was allowed to cool to a mild temperature and kept in thermo-flasks to maintain the serving temperature around 40 °C. The usual (commonly consumed cereal-based complementary food in the sample collection area) was used as a control constructed from 80% barley flour and 20% un-geminated maize flour (personal communication with some caretakers/mothers). All the prepared porridges were subjected to sensory evaluation as fresh as possible by mother panelists.

2.4.3. Macro-compositions analysis of individual components and the formulations

In this study, Associations of Official Analytical Chemists (AOAC) [38] modified methods were used to determine the macro-compositions of the individual components and formulated flours. The moisture content was determined by air convection oven drying method (Model No. DHG-9123 A, Sweden) using the method described by 925.10 [38], for 1 h at 130 ± 3 °C. Crude protein content was determined by Kjeldahl (Kjeltec 8400, Auto Sample Systems, Foss Analytical, Sweden) using a nitrogen conversion factor of 6.25 following the official method 954.10, [38]. Soxhlet method (Soxtec 8000, Tecator Line, Foss Analytical, Sweden) was used to determine crude fat content by N-Hexane extract according to the method number 2003.06, [38]. Ash content was determined using the combustion method by box type muffle furnace (Model No. SX2-4-10 GJ, China) at 550 °C for 4 h following method 923.03, [38]. Crude fiber content was determined by Fibertec 8000, Foss analytical, Sweden, following official method 978.10, [38]. Utilizable carbohydrate was estimated by difference: 100 - (% Moisture +% Crude protein +% Crude fat +% Crude fiber + % Ash) as indicted by Ayele et al. [39]. Gross energy (kcal/100 g) of individuals and formulated flours was computed based on FAO/WHO [40] recently amended in 2021 codex guideline by multiplying the values obtained for energy-yielding nutrients (crude protein, crude fat, and utilizable carbohydrate) with Atwater conversion factors where E (kcal/100 g) = [ (% crude fat x 9) + (% crude proteins x 4) + (% utilizable carbohydrates × 4)].

2.4.4. Determination of energy and protein density of the optimized formula

The energy density of complementary foods, expressed as kcal/g, can be calculated by dividing the total energy from complementary foods by the amount of complementary foods served to children, taking into account the theoretical gastric capacity of infants and young children to be 30 g/kg body weight per meal and the required number of meals to be served to the children per day for the respective age group. But the theoretical gastric capacity may not usually be functional due to a lack of responsive feeding, inappropriate consistency of foods, or due to underlying health conditions where the children take less [41].

The amount of food needed per day by children in their respective age groups to satisfy their energy and nutrients demand is a function of the amount of energy and nutrient densities that can be provided by the complementary foods. Abeshu et al. [42] estimated that for complementary food having an energy density range of 0.6–1.0 kcal/g, the amount of food served to provide the daily energy requirement is 263.3 g for 6-8-month-old, 317.09 g for 9–11month old, and 334.92 g for 12–23month old children for the study done in Wolayita Zone of Ethiopia and the children could consume an average amount of 212.82 g of the complementary food per day due to their low functional gastric capacity. Another report by Wasswa et al. [43] stated that the amount of porridge eaten by children in rural settings of Uganda was estimated at 200 g per day for aged 6–12 months old children. Based on these two reports, we have estimated the amounts of our optimized novel complementary foods to be served to an age group of 6–8months, 9–11 and 12–23 months old children to meet their daily requirements to be 225 g, 300 g and 450 g, respectively (i.e 75 g, 100 g and 150 g per meal) for the respective age group and 3 meals/day. Protein-density or protein energy ratio was calculated by dividing crude protein content of the optimal flour by its corresponding optimal energy value expressed as g/100 kcal.

2.4.5. Sensory evaluation of the porridges

The sensory quality of the formulated flour porridges was examined for its color, aroma, taste, mouth-feel and overall sensory acceptability by the mother panelist. Each porridge sample was evaluated by 48 untrained healthy panelists comprising mothers/caretakers having babies between 6 and 24 months. The selection of the mothers was purposive and assisted by the community health extension workers where mothers with at least primary education and who are familiar with quality characteristics of porridge for children were targeted. The evaluation was conducted at around 9.00 a.m. (morning) to ensure that they were not hungry and that it was not immediately after a meal. The porridges were served on clean, odor-free identical plates coded with 3-digit random numbers in randomly arranged order at the serving temperature. The panelists were requested to evaluate each sample after they were briefed about scoring of a sensory attribute using a 5-point hedonic scale representing; 5-like very much, 4-like moderately, 3-neither like nor a dislike, 2-dislike moderately and 1-dislike very much. The panelists were provided with mild-heated water before and after each tasting of the porridge to rinse (expectorate) their mouths to prevent bias of evaluation and avoid the carry-over effect and waited for 1 min before testing the next sample. The traditional way to report on the hedonic scale is to sum the values given to each selected food attribute and create a score for each respondent and used to represent a particular trait. The evaluation process was attentively instructed and controlled by the researcher.

2.5. Statistical analysis and model evaluation

All the laboratory analysis results of the 11 formulations (Table 2) were subjected to ‘Scheffe’ polynomial multiple regression analysis using the Stat-Ease Design Expert® software version 11 (D-optimal mixtures design). Mixture components were considered as model terms and mixture regression was designated as model fitting. Linear, quadratic, cubic and special cubic models and interactive effects of the independent variables were fitted for evaluation of macro-compositions and overall acceptability (Table 4). Analysis of variance (ANOVA) of the design expert was performed to develop models and determine the goodness of fit (significance) of the models developed. Linear and polynomial regression models were judged (verified) to be adequate and significant using F-statistic at a probability (P) of 0.05, 0.01, and 0.001 and the coefficient of determinations R². R²% is the percentage variation of the response variable that can be explained by its linear relationship with the independent variable and measures the degree of fitness of a regression model. The closer the R² value is to unity (1) and the closer R²_adj to R²_pred (the difference less than 0.2) and the adequate precision value greater than 4 shows the better the model fits the actual data ensuring satisfactory fitted models adequate to specify the relationship between dependent and the independent variables. The R²% value should be at least 80% to have a good fit of a regression model [44]. Normality and constant variance assumptions of the error terms were checked to determine whether a model meets the assumptions of the analysis. On the other hand, datasets of macro-compositions of the individual flours, the formulated flours and sensory evaluation of the prepared porridges were statistically analyzed using One-way ANOVA of statistical packages for social services (SPSS) (IBM version 24, Chicago, USA) to declare statistically significant mean differences between the formulated flours as compared to the control and the Cerifam® faffa flour (the popular commercial complementary/weaning flour in Ethiopia). All the data collected were in triplicate except for the sensory evaluation. Levene's test was used to check the equal variance assumptions (P > 0.05, should be non-significant). Tukey honestly significant difference (HSD) post hoc test was used for pairwise multiple comparisons (mean difference separation test) and the significant differences were declared at P < 0.05.

Table 4.

Models fitted for macro-nutrients, sensory evaluation and statistical outputs showing significance and model adequacy.

Model's P-value (model terms)	Macro-Compositions							SE
	MC	Protein	Fat	Ash	Fiber	CHO	Energy	OA
Linear (x1,x4,x6)	0.0049 #**	0.0005 #***	0.334	0.0211#*	0.3442	0.1790	0.3108#^	0.2196
Quadratic (x12, x42, x62)	0.3313	0.4721	0.0030#**	0.7957	0.9237	0.0545	0.8144	0.0977
Special Cubic	0.5132	0.5436	0.9767	0.8995	0.2875#^	0.4498	0.3075	0.9245
Cubic	0.3819	0.4444	0.8637	0.9694	0.6022	0.0483#*	0.7794	0.0452
Sp. Quartic vs Quadratic	0.1780	0.4675	0.5789	0.8310	0.3177	0.3527	0.5419	0.0311#*
Interaction terms
x1x4	–	–	0.0102	–	0.5564	0.0560	–	0.3471
x1x6	–	–	0.0064	–	0.2962	0.1521	–	0.0319
x4x6	–	–	0.0009	–	0.5077	0.0406	–	0.5901
x1x4x6	–	–	–	–	0.2875	0.9808	–	0.0308
F-statistics (ANOVA)
Model F-value	11.15	22.23	16.61	6.49	0.64	257.55	1.36	31.58
Lack-of-Fit	PE0	PE0	PE0	PE0	PE0	PE0	PE0	PE0
R2	0.75	0.8487	0.9432	0.6188	0.4899	0.9996	0.2534	0.9921
R2%	75	84.87	94.32	61.88	48.99	99.96	25.34	99.21
ADP	8.53	12.961	12.72	6.979	2.856	55.029	2.76	19.95
Model	0.0049*	0.0005*	0.0039*	0.0211*	0.7025^	0.0483*	0.3108^	0.0311*

Key: Sp.-special, vs-versus, PE0-pure error zero, R²%-coefficient of determination, ADP-adequate precision, MC-moisture content, CHO-carbohydrate, SE-sensory evaluation, OA-overall acceptability, #-suggested, #*, #** and #***-model is suggested and significant at P < 0.05, at P < 0.01 and P < 0.001, respectively, #^ -suggested and not significant, * and ^ model is significant and not significant respectively, x₁-dabi teff, x₄-field pea, x₆-germinated maize.

2.6. Variables optimization and validation

After regression analysis of the experimental data, numerical and graphical optimization techniques were employed to identify the optimum formula (sweet spot) using deign expert (D-optimal mixture design). Simultaneous numerical values of independent variables and multi-responses optimization were performed by setting the desired goals for each independent and dependent (response) variable. In brief, the major components used; dabi teff flour, field pea flour and germinated maize flour were set in range and the responses were set to maximize energy, protein, ash, fat, and carbohydrate; whereas to minimize moisture and crude fiber (Table 5). In addition to goal setting, the relative importance of each dependent variable in the overall optimization solution was set to be ‘++++’ for protein and carbohydrate while it was set as ‘+++++’ for overall acceptability while the remaining were kept at default which is ‘+++‘. These settings are important because the overall desirability function value of an optimization process is majorly ‘Goal or criteria setting dependent’.

Table 5.

Goals set, the relative importance of each variable and the optimal values at optimal conditions identified.

Name	Goal	Lower Limit	Upper Limit	Importance	Optimum value
Independent variables
A:Dabi teff Flour	In range	20	35	3	34.66
D:Field pea Flour	In range	0	30	3	15.34
F:Maize Flour	In range	5	20	3	5
Barley flour	Constant	25	25	3	25
Oats flour	Constant	15	15	3	15
Linseed flour	Constant	5	5	3	5
Dependent variables
Moisture (%)	minimize	4.41	5.74	3	5.57
Crude Protein (%)	maximize	14.58	17.21	4	15.74
Crude Fat (%)	maximize	4.22	5.59	3	5.09
Ash (%)	maximize	2.01	2.6	3	2.26
Crude Fiber (%)	minimize	2.68	3.96	3	2.88
Carbohydrate (%)	maximize	66.91	70.76	4	73.05
Energy (kcal/100 g)	maximize	378.82	386.9	3	380.43
Overall Acceptability	maximize	3.4	4.4	5	4.12

The graphical optimization was carried out by superimposition of contour plots for all the responses with respect to component ratios. For confirming the validity (accuracy) of the models, a laboratory experiment was conducted on the component ratios identified at the optimal formula and the obtained laboratory results were then compared with optimal values at the optimal condition and with values of the control flour and finally reported as the formula with an optimal nutritional and sensory profile of all the possible combinations or solutions.

3. Results

3.1. Macro-composition of the individual processed flours

The mean and standard deviation of the macro-composition of the individual components that underwent the various processing steps were presented in Table 1. A significant difference (P < 0.05) was observed in mean values of the macro-compositions among the individual components and ranged between 4.32 and 9.03% for moisture, 8.58–20.95% protein, 2.58–36.08% fat, 1.42–3.96% ash, 2.12–7.46% fiber, 28.24–77.63% utilizable carbohydrate and 350.26–519.96 kcal/100 g for gross energy, respectively. The fat content was similar (P > 0.05) in all the components except in linseed flour where it was significantly higher (P < 0.05) than the others (Table 1). This table indicted that field pea and linseed had significantly higher protein content at 20.95% and 20.57%, respectively with their highest protein density at 5.98 and 3.96 g/100 kcal, respectively followed by oats flour at 12.5% protein content while the protein content of dabi teff was determined at 10.74%.

In addition to the significantly higher protein content, linseed flour was determined to contain significantly higher fat content at 36.08% as expected from an oil seed. The fat content of all the other components was significantly lower-far below 6%. The ash content of dabi teff flour was significantly higher (P < 0.05) at 3.96% among the others followed by linseed flour at 3.45% while germinated maize showed the lowest ash at 1.42%. The fiber content of linseed was significantly higher at 5.34% preceded by field pea with the highest fiber content at 7.46%. Dabi teff flour had shown fiber content at 3.76% while the remaining components contained fiber blow 3%. The carbohydrate content of roasted barley was significantly higher at 77.63% followed by germinated maize flour at 75.24% with the lowest value for linseed at 28.24% and the carbohydrate content of dabi teff was determined at 68.57% which was comparably high as that of oat flour at 70%.

A relatively lower energy value was computed for dabi teff flour at 352.7 kcal/100 g as compared to the other components which could be attributed to its higher moisture content (9.03%). The lowest energy value was computed for field pea flour at 350.26 kcal/100 g while roasted barley, oats and germinated maize flours had shown comparably higher energy values at 380 kcal/100 g whereas the highest energy value at 519.96 kcal/100 g was determined for linseed flour (Table 1).

3.2. Macronutrient content of the formulated complementary flour

3.2.1. Model fitting and testing model adequacy

The fitted models were found to be adequate and significant for most response variables based on the F-statistic (the ANOVA regressions outputs); the P-value, the coefficient of determinations R² and the agreement between R²_adj and R²_pred. Table 4 showed that the linear models were adequately fitted for moisture, protein and ash contents and significant at P-values of <0.01, <0.001 and < 0.05. This shows that x₁, x₄ and x₆ (dabi teff, field pea and germinated maize) were the significant model terms for these compositions. The quadratic model was adequately fitted for fat and significant at a P-value of <0.05. This shows that the linear model terms (x₁, x₄, and x₆) and the interactive model terms (x₁x_4, x₁x₆ and x₄x₆) were significant model terms for fat content.

On the other hand, the cubic model was significantly fitted for carbohydrates and significant at P-values of <0.05 where the linear models terms (x₁, x₄, x₆) and the interactive model terms x₁x₆, x₄x₆ (x_1- x₆), and x₄x₆(x_4-x₆) were the significant model terms for carbohydrate at p-value less than 0.05. Special cubic and linear models were suggested by the software for fiber and energy contents, respectively, but the models were non-significant (P >0.05).

3.2.2. Macro-composition of the formulated complementary flours

Mean moisture, protein, fat, ash, fiber, carbohydrate and energy contents of the new formulations were determined to be 4.89, 15.49, 4.91, 2.23, 3.48, 68.99% and 382.14 kcal/100 g, respectively (Table 6). When compared to the [22] guideline for the nutrient content of formulated complementary foods, the mean value of the new formulations fulfilled the critical limits for all the macro-composition except for fat where the guideline recommends moisture, protein, fat, ash, fiber, carbohydrate and energy contents to be ≤ 5, ≥15, 10–25, ≤3, ≤5, 64 ± 4% and 400–425 kcal/100 g, respectively.

Table 6.

Macro-composition and energy values of the formulated fours with their corresponding protein-energy density (PER) and FAO/WHO (1991) recommendation.

Formulations	Macro-composition (%)						Gross Energy (Kcal/100 g)	PER (g/100 kcal)	PEL (%)
	Moisture	Protein	Crude Fat	Total Ash	Crude Fiber	Carbohydrate
F1	4.58 ± 0.01a	15.08 ± 0.25a	4.22 ± 0.04a	2.31 ± 0.08a	3.68 ± 0.06a	70.13 ± 0.26a	378.82a	3.981	15.293
F2	4.61 ± 0.25a	16.51 ± 0.16b	4.75 ± 0.01a	2.45 ± 0.01a	3.56 ± 0.03a	68.12 ± 0.13a	381.27a	4.330	17.321
F3	5.74 ± 0.10a	15.93 ± 0.05ab	4.78 ± 0.06a	2.25 ± 0.06a	3.22 ± 0.01b	68.08 ± 0.04a	379.06a	4.203	16.810
F4	4.68 ± 0.54a	17.21 ± 0.17bc	5.14 ± 0.40a	2.60 ± 0.01ab	3.46 ± 0.04a	66.91 ± 0.29b	382.74a	4.496	17.986
F5	5.05 ± 0.16a	14.63 ± 0.02a	5.59 ± 0.57a	2.14 ± 0.01ac	3.82 ± 0.06a	68.77 ± 0.33a	383.91b	3.811	15.243
F6	5.52 ± 0.42a	15.27 ± 0.32ab	4.92 ± 0.69a	2.20 ± 0.11a	2.68 ± 0.01c	69.41 ± 048a	383c	3.987	15.948
F7	4.50 ± 0.01ab	15.17 ± 0.04ab	5.32 ± 0.06a	2.05 ± 0.02ad	3.74 ± 0.03a	69.22 ± 0.08a	385.44d	3.936	15.743
F8	5.45 ± 0.03a	14.65 ± 0.35a	5.30 ± 0.11a	2.09 ± 0.18ae	3.96 ± 0.06ag	68.55 ± 0.14a	380.5a	3.850	15.401
F9	4.41 ± 0.03ab	14.80 ± 0.09a	4.62 ± 0.33a	2.35 ± 0.02a	3.66 ± 0.06a	70.16 ± 0.19a	381.42a	3.880	15.521
F10	4.43 ± 0.04ab	16.53 ± 0.25bd	4.34 ± 0.12a	2.11 ± 0.03af	3.78 ± 0.06a	68.81 ± 0.08a	380.42a	4.345	17.381
F11	4.88 ± 0.04a	14.58 ± 0.04a	5.06 ± 0.19a	2.01 ± 0.19ag	2.71 ± 0.32cd	70.76 ± 0.03a	386.9e	3.768	15.074
Mean	4.8955	15.4873	4.913	2.233	3.479	68.993	382.135	4.053	16.156
Control	6.15 ± 0.82c	10.77 ± 0.52e	3.78 ± 0.79a	2.24 ± 0.01a	2.65 ± 0.03ce	74.41 ± 2.15c	374.74f	2.874	11.496
Cerifam®Faffa	6.5c	14a	4a	2.5a	3b	76c	396g	3.54	14.14
Optimized	5.57	15.74	5.09	2.26	2.88	73.05	380.43	4.14	16.55
FAO/WHO	≤5	>15	10–25	≤3	≤5	64 ± 4	400–425	<5.5	6–15

Key: Values are means ± standard deviation of the triplicate determinations. Values in the same column followed by different superscripts are significantly different at P < 0.05. The corresponding component ratio for each formula code (F1– F11) is well described in Table 2 earlier. PER-protein energy ratio; PEL-protein energy level.

The moisture content of the formulations ranged from 4.41 to 5.74% and the mean moisture was significantly lower (P < 0.05) than that of the control and ‘Cerifam® faffa’ flour at 6.15 and 6.5%, respectively. The protein content of the new formulations ranged from 14.58 to 17.21% where it had shown significantly higher (P < 0.05) protein content (1.4–1.6 fold) and protein density (1.31–1.56 fold) than the control flour and slightly higher than the Cerifam® faffa flour. In this study, fat contents of the formulations ranged from 4.22 to 5.59% and there was no significant difference (P > 0.05) between the mean fat content of the formulations, the control at 3.78% and the Cerifam® faffa flour at 4%.

The ash content of the formulated flours ranged from 2.01 to 2.60% where some of the formulated flours had significantly higher (P < 0.05) ash content as compared to the control. The fiber content of the new formulations ranged from 2.68 to 3.96% which was slightly higher than that of the control and Cerifam® faffa flour at 2.65% and 3%, respectively. The carbohydrate content of the new formulations ranged from 68.08 to 70.76% which it was slightly lower than that of the control and Cerifam® faffa flour at 74.41% and 76%, respectively and this could be due to the barley and maize which made the control and the wheat and soybean in the Cerifam®. The energy content of the formulations ranged from 378.82 to 386.9 kcal/100g and it was significantly higher (P < 0.05) than the control flour at 374.74 kcal/100 g.

3.3. Sensory evaluation results

The mean values of the sensory evaluation score were significant difference (p < 0.05) among the formulated porridges (as affected by components ratio variations) and ranged between 3.4 and 4.8 for color, 2.8–4.8 for aroma, 3.0–4.4 for taste, 3.2–4.5 for mouth feel and 3.4–4.4 for overall acceptability, respectively with mean overall sensory acceptability score of 3.99 among all the porridges (Table 7). The control porridge had 3.5, 4.2, 3.8, 3.9 and 3.4 for color, aroma, taste, mouth feel and overall acceptability, respectively. With these determined sensory mean scores, it can be judged that the porridges prepared from the formulated complementary foods were liked much and acceptable as high as 96% by mothers/caregivers who participated as sensory panelists.

Table 7.

The mean score of the sensory evaluation of porridges prepared from the formulated complementary flours.

Porridge from Formulations	Mixture components ratio (%)						Sensory Attributes
	DTF	BF	OF	FPF	LSF	GMF	Color	Aroma	Taste	MF	OA
FP1	27.50	25	15	15.00	5	12.50	3.5 ± 0.97a	3.8 ± 0.92a	3.4 ± 0.52a	3.8 ± 1.32a	3.9 ± 0.74a
FP2	26.123	25	15	23.877	5	5.00	3.7 ± 1.06a	3.7 ± 0.82a	3.4 ± 1.17a	4.2 ± 0.79a	4.2 ± 1.05a
FP3	30.457	25	15	19.543	5	5.00	3.9 ± 0.57a	3.3 ± 0.82a	3.2 ± 0.92a	3.6 ± 0.84a	4.2 ± 1.08a
FP4	20.00	25	15	29.988	5	5.012	3.4 ± 0.82a	3.6 ± 0.82a	3.4 ± 0.70a	3.3 ± 1.20a	4.0 ± 0.94a
FP5	28.793	25	15	6.207	5	20.00	4.0 ± 1.2a	4.8 ± 1.02b	4.1 ± 1.10a	4.1 ± 0.99a	3.4 ± 0.82a
FP6	35.00	25	15	11.884	5	8.116	4.8 ± 0.74b	4.0 ± 0.99a	4.2 ± 0.92a	4.0 ± 0.92a	4.4 ± 0.92b
FP7	21.705	25	15	13.295	5	20.00	4.0 ± 1.35a	4.2 ± 0.92a	3.8 ± 1.14a	4.2 ± 0.57a	3.9 ± 0.99a
FP8	34.385	25	15	6.476	5	14.139	4.7 ± 0.99c	4.4 ± 0.52a	4.1 ± 0.95a	4.2 ± 1.14a	3.8 ± 0.52a
FP9	20.00	25	15	19.540	5	15.460	4.1 ± 0.57a	3.8 ± 1.03a	3.9 ± 0.88a	4.5 ± 0.95a	4.3 ± 1.16a
FP10	22.735	25	15	22.515	5	9.750	4.0 ± 0.82a	2.8 ± 0.88a	3.0 ± 0.94b	3.2 ± 1.32b	3.8 ± 1.05a
FP11	29.604	25	15	10.076	5	15.320	4.3 ± 0.62a	4.4 ± 0.70a	4.4 ± 0.97a	4.2 ± 12a	4.0 ± 1.03a
						Average	4.04	3.89	3.72	3.94	3.99
Control porridge	0	80	0	0	0	20	3.5 ± 1.25a	4.2 ± 0.63a	3.8 ± 0.63a	3.9 ± 0.88a	3.4 ± 0.63a

Key: FP1-FP11- porridges of the coded formulations, DTF- dabi teff flour, BF-barley flour, FPD-field pea flour, LSF-linseed flour, GMF-germinated maize flour, MF-mouth feel, OA-overall acceptability.

Values are means ± Standard deviation of the sensory scores. Values in the same column followed by different superscripts are significantly different at P < 0.05.

3.4. Variables optimization and validation

An optimal complementary food that can meet energy and protein density requirements can be obtained from nutritious cereal-legume blends to develop a super-quality complex child food. In the present study, from the multi-response numerical optimization, the optimal formula (sweet spot) at overall optimization was identified to be at 34.66% dabi teff flour, 25% barley, 15% oats, 15.34% field pea, 5% linseed and 5% germinated maize flours with the predicted optimal value for the macronutrients to be 5.57% for moisture, 15.74% for protein, 5.09% fat, 2.26% ash, 2.88% fiber, 73.05% carbohydrate, 380.43 kcal/100 g for energy and 4.12 for overall sensory acceptability score, respectively (Table 5) where the 3-D surface graph showing the maximum selected desirability function at overall optimal condition was found to be 0.651 (Fig. 2). Results obtained at optimum conditions showed significantly higher protein (1.46 fold), and fat (1.35 fold) as compared to the control (Table 8).

Fig. 2.

3-D surface graph showing the desirability function of each response variable and the combined desirability function.

Table 8.

Experimental validation of optimal condition and comparison with the control flour.

Response variables	Optimized value	Validation value	Contro value
Moisture	5.57	5.34	6.15
Protien	15.74	15.86	10.77
Fat	5.09	5.12	3.78
Ash	2.26	2.21	2.24
Fiber	2.88	2.85	2.65
Carbohydrate	73.05	68.82	74.41
Energy	380.43	384.80	374.74
Overall acceptability	4.12	4.34	3.45

The following figure shows the material balance design of the defined ranges of the components, their nutrient contents and the optimal condition identified (Fig. 3).

Fig. 3.

Material balance design showing individual flour ranges, their nutrient contents and the optimal condition.

Key: dt = dabi teff, bf = barley, of = oats, fp = field pea, lf = linseed, mf = maize flours.

To validate the optimal condition, the ratio of each component at the optimized formula was blended and laboratory analysis of nutrient compositions of the blend was conducted where the validation analysis result was 5.34%, 15.86%, 5.12% 2.21%, 2.85%, 68.82%, 384.80 kcal/100 g for moisture, protein, fat, ash, fiber, energy, respectively. This showed there was a good agreement (accuracy of the model solution) between the optimal values at optimal conditions and the validation results (Table 8).

The optimization process was verified using the graphical optimization technique (Fig. 4) where the overlay counter plot was generated for the most important variables to identify the optimum regions. Any design points that fall within the yellow region (with the very conservative silver-gray region) in the overlay plot represent an optimal combination of dabi teff flour, field pea and germinated maize flours in combination with the other constant components that can bear optimal macronutrient contents and overall sensory acceptability within 95% confidence interval (CI).

Fig. 4.

Graphical overlay contour plot for the overall graphical optimization.

When compared to the [22] guideline for nutrient content requirements of formulated complementary foods, the optimized complementary flour adequately met the recommended critical limits for all the macro-compositions except for fat content at ≤5, ≥15, 10–25, ≤3, ≤5, 64 ± 4% and 400–425 kcal/100 g for moisture, protein, fat, ash, fiber, carbohydrate and energy contents, respectively.

The ‘novelty’ of the optimized complementary flour developed would be accredited to the incorporation of dabi teff flour containing high iron at 86.5 mg/100 g (determined by the authors) with good bioavailability (Ph:Fe molar ratio far below the critical limit) that greatly evidenced the social beliefs of this typical farmer variety teff. In addition, linseed flour which is a leading source of α-Linolenic acid, omeg-3 polyunsaturated fatty acid would make the product super.

3.5. Energy and macro-nutrient densities of the optimized complementary flour

The calculated energy and nutrient densities of the optimized flour was presented in Table 9. As described earlier, we have estimated the amount of complementary food to be served to children per day at 6–8 months, 9–11 and 12–23months old to be 225 g, 300 g and 450 g, respectively. After adjusting for these daily amounts, the energy density of our optimized complementary flour for all the age groups was computed to be 1.27 kca/g. The protein and carbohydrate densities were determined at 4.14 g/100 kcal and 19.20 g/100 kcal, respectively.

Table 9.

Energy and nutrient densities of complementary flours and Codex standards.

Parameters	Optimal flour	Control flour	PAHO/WHO	Cerifam®
Energy density (kcal/g)	1.27	1.25	≥0.8	1.32
Protein density (g/100 kcal)	4.14	2.87	>5.5	3.54
Fat density (g/100 kcal)	1.34	1.01	2.5–4.5	1.01
Carbohydrate density (g/100 kcal)	19.20	19.86	–	19.19

According to PAHO/WHO standard, the energy density of processed cereal-based complementary foods for older infants and young children should be ≥ 0.8 kcal/g [36] and from Table 9 it can be shown that the newly optimized novel complementary flour can considerably meet the PAHO/WHO standards for energy, protein and carbohydrate densities of complementary foods required for infants and young children.

Codex Alimentarius committee recommends the protein density of cereal-based formulated complementary foods for infants and young children should not exceed 5.5 g/100 kcal [22] and the protein density of our optimized complementary flour was 4.14 g/100 kcal which was near to the upper threshold (5.5 g/100 kcal) and it was 1.44 times higher than the protein density of the control (traditional) and 1.2 times higher than the Cerifam® faffa flours.

4. Discussion

Macro-composition are among the key nutritional quality indicators of food components and are usually the basis for establishing the nutritional profile and overall sensory acceptability of the final product [45]. In this study, the macro-compositions of each component were examined to justify the bases for the inclusion of the crop as a potential candidate in the formulations and it was aimed to add one or more essential nutrients to the mixture. The compositions of each component varied significantly and this could be attributed to group differences (some are cereals and others are pulses/legumes) and due to the individual samples pre-processing techniques applied.

In agreement with the present findings, Munshi et al. [46] reported the protein content of linseed at 18.2%. Our result indicated that field pea and linseed were super sources of protein which justified their inclusion in the formulation to develop protein-dense complementary food. This higher protein content of field pea and linseed was attributed to the legume and pulse crops naturally containing high amounts of protein than cereals. Further, the incorporation of field peas in the lysine-limited cereal is an excellent complement which further justifies the use of field peas as a feasible protein supplement in the development of cereal-based complementary foods. The mean protein content at 10.74% of dabi teff flour was in agreement with Bultosa [25] who reported a range of 8.7–11.1% with a mean of 10.4% for 13 teff varieties grown in Ethiopia and also with Daba [29] who reported protein content at 10.7 for DZ-01-2423 teff variety. In the contrary, the observed protein content of dabi teff was higher than the report by Agza et al. [47] which was 8.43% for ‘kora-DZ-Cr-438’ teff variety.

The present observed findings of fat and energy content of linseed were slightly lower than that reported by Munshi et al. [46] at 42% and 534 kcal/100 g, respectively. Linseed is a popularly cultivated oil crop by the community in the sample collection area and its oil is among the leading source of α-Linolenic acid, omega-3 polyunsaturated fatty acid from plant foods [34,48] which is crucially important for brain neural and cognitive development in utero and early childhood [49]. Such super quality and accessibility of linseed signifies its selection as a potential crop in the current formulations. The fat content of dabi teff flour was 3.94% which is relatively higher than the fat contents of teff grains reported by Daba [29] at 3–3.2%.

The higher ash content is an indicator of the higher total mineral content of the crop where the ash content of dabi teff in the present finding was high which was comparably agreed with the report by Woldemariam et al. [50] that red teff had an ash content of 3.2% whereas the report by Bultosa [25] showed lower mean ash content at 2.45% of 13 teff varieties. The observed fiber in all the individual components reasonably agreed with the codex standard where the value for dabi teff was consistent with a previous study by Bultosa [25] that showed the mean fiber content of 13 teff varieties at 3.3% and Woldemariam et al. [50] also reported a similar result at 3.51% of fiber content in red teff.

The higher carbohydrate content was observed in the cereal components where cereals are naturally known to contain higher starch than legumes and oil seeds. The lower carbohydrate value observed for linseed was in agreement with the report by Munish et al. [46] at 29% which could probably be compromised by its higher fat and protein content.

The energy density of a food is critical for energy intake by infants and young children and this can be maintained by selecting energy-rich food components to increase the energy density of formulated products. The comparably higher observed energy values of the individual components justified the selection of the food crops in our formulations to develop energy-dense complementary food. The highest amount of energy content computed for linseed may perhaps be due to its higher fat content than the other components because fat provides double as much energy (9 kcal/g) as compared to protein and carbohydrate.

Even though, it may not make sense to compare the nutrient contents of cereals and legumes because of their different natural physiognomies, in the present study, we have compared them simply because they were our components and to justify that cereals and legumes need to be mixed to nutritionally complement each other to develop wholesome and nutritionally complete end product. Aligned with this, the nutrient content of dabi teff such as protein, fat, carbohydrate, and energy, though lower than some of our current components, results were incomparably higher (in some cases similar) than other cereals such as sorghum [24].

In model fitting, almost all the models generated were declared to be adequate to describe the effects of independent variables (all the possible model terms) on the response variables. Meaning changes in the response variables could adequately be described (adequate predictive power) by the model terms in the developed mathematical model as a function of component ratio variations.

A linear model was adequately fitted and significant to describe changes (adequate predictive power) in moisture, protein and ash contents as a function of blending ratio variations whereas the quadratic model was adequately fitted and significant for fat content and the cubic model adequately fitted for carbohydrate content. Meaning changes in fat can adequately be described by the linear and quadratic model terms as a function of the blending ratio variations and the cubic medal terms for carbohydrates. Though a linear model was suggested by the software for energy contents, it was non-significant, and has no predictive power in describing changes in energy content as a function of components ratio variations in the blends. This finding was similar to the report by Tadesse et al. [51] that the combination variation of bulla, pumpkin and germinated amaranth had no predictive effect in describing the energy contents of complementary mixes.

The response of a mixture system does not depend on the total amount of the mixture, but it depends on the proportions of the individual components constituting the mixture. This study has shown that incorporating dabi teff which is a nutritious forgotten/underutilized food crop into other pre-processed local food crops (filed pea, germinated maize, barley, oats and linseed) to formulate and develop energy and protein-dense novel complementary food could be a potentially promising alternative than traditional (monotonous) child food. The lower moisture content observed in the newly formulated complementary food could be attributed to the processing techniques applied. A lower moisture content of less than 5% is recommended for formulated complementary flour [22] which is an indicator of the flour has longer storage stability that is required particularly among rural mothers in developing countries because there is a usual practice of preparing a relatively large amount of complementary flour and use it for a longer time to avoid frequent preparation.

The current observed protein content of our newly formulated complementary food was higher than the report by Desalegn et al. [31] for the formulation of complementary food from Chickpea (kabulli variety), red teff (DZ-0199 variety) and quality protein maize (BHQPY-545 variety and Ipomea batata, Tulla variety) that protein ranged from 10.42 to 12.70%. The higher protein content of the new formulations could be attributed to the higher protein content of field pea and linseed flours in the formulations as well as due to the germinated maize which could be attributed to the mobilization of storage nitrogen and the synthesis of enzymatic proteins by the sprouting maize seeds during germination. Mariam [52] reported that complementary food products from cereal-legume combinations of two or more components have better protein (overall nutritive value) than products from single plant food.

Regarding fat content of the currently new formulations, comparatively higher fat content was reported by Walle and Moges [32] from the formulation of complementary food from maize, wheat, barley, chick pea, soya bean and lentil ranging from 5.8 to 8.22%. The lower fat content may call for finding other sources and there is a common practice among Ethiopian mothers/caretakers that they add a spoonful of spiced refined cow butter during child food preparation that may match the fat content. In consistent to the present finding, Birhanu et al. [31] reported a similar result for the formulation of complementary food from Chickpea (kabulli variety), red teff (DZ-0199 variety) and quality protein maize (BHQPY-545 variety and Ipomea batata, Tulla variety) that ash ranged from 1.61 to 3.07%.

In agreement with the present findings, Mezgabo et al. [30] reported that the fiber content ranged between 2.33 and 3.6% for complementary porridge formulation from red teff, malted soybean flour and papaya fruit powder. The fiber content was in the recommended range and although consumption of dietary fiber-rich products provides various health benefits, the importance of food fiber is not over-emphasized in child nutrition and is required to be very low or nil in complementary food [53].

The contribution of carbohydrates as an energy source in complementary foods is considerably high and it must be digestible enough for infants and young children to be able to obtain the amount of energy required. The current finding fairly agrees with the report by Mezgabo et al. [30] that carbohydrate content ranged from 55.43 to 69.68% for complementary porridge formulation from red teff, malted soybean flour and papaya fruit powder. The present observed value of energy content was amenable with the report by Mezgabo et al. [30] that the energy content of complementary porridge formulation from red teff, malted soybean flour and papaya fruit powder ranged from 376.30 to 385.56 kcal/100 g. On the other hand, there was a slightly lower energy content of the current formulations than the FAO/WHO [22] recommendation (400 kca/100 g) and the Cerifam® faffa flour (396 kcal/100 g) which may be attributed to the lower fat content in the new formulations.

Sensory attributes are of paramount concern especially when new food product development is targeting infants and young children. In the first place, there is a problem of ‘picky eating and food jags’ in this age group and if the food is not pleasant or good taste, they may refuse to consume it and this is why mothers most often tend to add more sugar to make child porridge sweeter. In this study, color, aroma and overall sensory acceptability attributes of the formulated porridges was more preferred than that of the control porridge by mothers panelists and there would be a high likelihood that the new products will also be liked much by their infants and young children. The observed optimal overall sensory acceptability score at 4.12 out of the five hedonic scales of the formulated porridge was judged to be high and much liked. This enhanced sensor acceptability could be attributed to the multiple grains and the processing technique applied which imparted good flavor, aroma and mouth feel to the newly formulated porridges.

Dewey and Brown [41] reported that cereals contain a limited amount of essential amino acid (lysine) while pulses and peas are lacking in cysteine, methionine and tryptophan therefore giving the bases for complementing each other by blending and optimizing to correct for the essential amino acid deficits to be able to develop fairly complete protein (containing all essential amino acids) product. Equally, child nutrition emphasizes the importance of quality protein, omega-3 polyunsaturated fatty acid and adequate iron during complementary feeding [49] where our new optimized product would certainly contain a good amount of these nutrients as a result of the presence of field pea, linseed and dabi teff flours in the formulations.

Regarding the overall desirability function of the present study, it was in agreement with Talabi et al. ([54] who has reported a similar desirability function of 0.65 during optimization of complementary meal from a mix of yellow maize, sorghum, millet, soybean, groundnut, crayfish and fish, additionally, Keyata et al. [55] has reported a similar desirability function of 0.625 for the overall optimization of complementary food from sorghum, soybean, karkade and premix.

Feeding energy and nutrient-dense foods provides children with the required energy/nutrients for proper growth and development as well as their daily demands would be fulfilled by fewer meal frequencies. The energy and nutrient intake/kg body weight by children would be maintained by increasing the energy and nutrient density of their diet and not by an increase in the intake amount or volume. In the current study, it was unable to compare the energy density of our optimized complementary food with some past reports because of the wrong energy density calculation which said, “Simply dividing the energy contents of the complementary foods by 100”. For example, Tenagashaw et al. [56] reported energy density of complementary food formulated from blends of teff, soya bean and orange flashed sweet potato ranged from 3.70 to 3.76 kcal/g and

Ayele et al. [39] reported the energy density of complementary foods from kocho, pumpkin fruit, red kidney bean and maize at 3.3 kcal/g following the same calculation which was over-exaggerated. Even higher energy-dense foods have energy density of 1.07–1.46 kcal/g where in this case the approximate quantity of complementary food that would meet the energy needs of the children would be 137–187 g/day at 6–8 months, 206–281 g/d at 9–11 months, and 378–515 g/d at 12–23 months [36,57]. Conversely, the current energy density of our optimized novel complementary food was higher than the energy density reported by Abebe et al. [58] where the traditional corn-based porridge consumed as complementary foods in rural villages of Sidama zone of Southern Ethiopia contains the energy density of 0.53 kcal/g, while kocho-based contained 0.49 kcal/g, both of which contained very low energy levels.

In agreement to the present finding, Tenagashaw et al. [56]) reported that the protein density of complementary foods from blends of teff, soya bean and orange flashed sweet potato was in the range of 3.50–4.79 g/100 kcal. It also agreed with the report by Ayele et al. [39] where the protein density of an optimized formula from kocho, pumpkin fruit, red kidney bean and maize was 3.0 g/100 kcal. Alternatively, the protein density of homemade complementary foods in Wolayita zone of Ethiopia contained lower protein density ranging from 2.13 to 2.48 g/100 kcal as reported by Abeshu et al. [57]. This showed that the newly optimized novel complementary food is superior to the traditional ones, met the recommended standard and can provide the daily protein and energy demands of 6–23 months old children if served the amount estimated for the respective age group and fed three times per day.

The current study would have practical application in protein and energy-dense baby food industrial production, application in dabi teff and field pea enriched homemade complementary food preparation targeting infants and young children, school going children and whole family set up aiming to meet optimal levels of protein and energy dense food products.

5. Limitations and future scope of the study

Of paramount, budget constraints and the current security challenge in Ethiopia were the biggest limitations of the study. The major findings of this study need to be substantiated with further studies including microbial safety and shelf life study, the effect of combined and varying processing conditions in keeping the quality against aflatoxin and rancidity, fatty acid profiling (the mono, poly and omega3), amino acid profiling, the effect of roasting on trypsin inhibitor and protein digestibility , conducting randomized controlled feeding trail to determine the efficacy of the newly developed dabi teff-field based optimized novel complementary food in improving nutritional outcomes (anthropometric indices) among infants and young children.

6. Conclusion

This study demonstrated that optimized energy and protein-dense novel complementary food with improved sensory acceptability can be developed by the inclusion of dabi teff, the underutilized/forgotten food crop into locally available pre-processed food crops viz., germinated maize, roasted barley, roasted field pea, dehulled oats and linseed flours that can fulfill the FAO/WHO standard and exhibit superior quality over the traditional complementary food and the Cerifam® faffa flour. Each of the components used contributed an important nutrient of interest to develop a nutritionally complemented optimized product. The results indicated that the optimized condition was identified at 34.66% dabi teff, 25% barley, 15% oats, 15.34% field pea, 5% linseed and 5% germinated maize flour ratios which were found to contain 15.74 g/100 g protein, 5.09 g/100 g fat, 2.26 g/100 g ash, 2.88 g/100 g fiber, 73.05 g/100 g carbohydrate, 380.43 kcal/100 g energy, 4.14 g/100kacl protein density and 1.27 kcal/g energy density. The lower fat content can be matched by looking for additional fat-energy-rich legumes. These findings suggested that the developed dabi teff-field based optimized novel complementary food can be used as a sustainable food-based strategy to combat protein-energy malnutrition among children in less developed countries like Ethiopia.

Funding statement

This work was supprted by Jimma University, Institute of Health Science (JUIHS) [IHRPGS/428/19].

Additional information

No additional information is available for this paper.

Ethical approval and consent

Ethical clearance for the sensory analyses was obtained from the Institutional Review Board (IRB) of Jimma University, Institute of Health Science, Ethiopia (Ref No. IHRPGS/395/2019). Informed verbal consent was obtained from each of the sensory panelists.

Author contribution statement

Diriba Chewaka Tura: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper. Tefera Belachew, Dessalegn Tamiru, Kalkidan Hassen Abate: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper.

Data availability statement

Data included in article/supp. Material/referenced in article.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.