Plant Foods for Human Nutrition 57: 365–376, 2002.

© 2002 Kluwer Academic Publishers. Printed in the Netherlands.

A preliminary study on the use of tempe-based

formula as a weaning diet in Nigeria

OLUWATOOYIN F. OSUNDAHUNSI1 and OGUGUA C. AWORH2

1 Author for correspondence: Department of Crop Production, Food Science and Technology,
The Federal University of Technology, PMB 704, Akure, Ondo State, Nigeria; 2 Department
of Food Technology, University of Ibadan, Nigeria

Abstract. Tempe, an Indonesian mold fermented food, was prepared from cowpeas and
soybeans using the traditional oriental process with modifications where appropriate. Four
complementary foods were developed from whole maize meal or dehydrated fermented maize
(ogi) flour fortified with either cowpea tempe or soybean tempe. Wholesomeness and poten-
tials of the vanilla-flavored foods as weaning diets were determined. The nutrient content of
all the developed products were within the range prescribed by the FAO/WHO pattern for
processed weaning foods. The products had loose bulk densities between 0.40–0.55 g/ml and
packed bulk densities between 0.68–0.75 g/ml; reconstitution indexes were between 92.30–
104.00 g/ml; viscosities were 34.2–65.0 CP at 70 ◦ C and 45.0–76.9 CP at 30 ◦ C. Total
plate counts ranged from 3.2 × 103 – 4.3 × 104 cfu/g; coliforms staphylococcus, mold, and
yeast were absent in some of the products. The results on sensory attributes showed that
tempe-based weaning diets were comparable with Mameal, a commercial product. Maize and
maize-fortified diets were slightly higher in packed bulk density. Fortification reduced the vis-
cosity of the diets and the values reduced with increasing temperature. Tempe-based weaning
foods reconstituted easily in hot water, while cooking destroyed most of the microorganisms
present. Maize-based tempe fortified foods were relatively inexpensive and have potential as
weaning foods.

Key words: Tempe, Weaning diets

Introduction

Malnutrition still persists as a principal health problem among children below

the age of five in Nigeria [1]. Most traditional weaning foods in developing
countries are made from cereals, starchy fruits, roots and tubers. Such foods
are often of poor protein quality and have high paste properties [2]. Grain
legumes are promising materials to increase the supply of vegetables protein
that the world will need in the near future [3]. Legume – fortified weaning
foods are of good nutritive value and have proved to prevent protein-energy
malnutrition [4, 5]. Fermentation of either legumes or cereals or both before
use in weaning food, increases digestibility, and reduces antinutritional and
366

flatulence factors [6, 7]. Tempe is an Indonesian fermented food tradition-

ally made from soybeans using mold of the genus Rhizopus. Preparation of
tempe and activity of the mold during storage have been investigated [8–
10]. The nutritional value of tempe has been studied by Murata et al. [11,
12], Omotola [13], Wang et al. [14], Winarno & Reddy [15] and Zamora
& Veum [16]. The nutritive value of maize-soybean tempe as weaning food
was evaluated by Mugula [17]. In another work, sorption characteristics of
tempe-based weaning foods were determined to ascertain storage stability
[18]. However, in establishing the acceptability of tempe – fortified maize-
based weaning foods, Egounlety & Aworh [19] reported that although the for-
tified foods were acceptable, most mothers preferred a commercial weaning
food ‘Cerelac’ because of its vanillia flavor and sweet taste.
This study was undertaken to develop low cost weaning formulations with
tempe as the protein source. An attempt was made to evaluate the acceptabil-
ity based on some quality parameters.

Materials and methods

Production of tempe and weaning diets

White maize (Zea mays) was obtained from a local market in Ibadan. Tempe
was produced using basically the traditional Indonesian technology with some
modification. Dehulled, cooked soybeans and cowpeas were inoculated with
powdered Rhizopus oligosporus culture, provided by National Nutrition Re-
search and Development Centre, Bogor, Indonesia, at the level of 2 g dry
inoculum powder per kg beans. The inoculated beans were packed in per-
forated heat – sealed polyethylene bags and incubated at room temperature
(27–30 ◦ C) for 48 hours. After fermentation, the fresh tempe was cut into
small pieces, blanched for 15 min in boiling water and dried in a force draught
oven (Gallenkamp, Model OV – 160) at 60 ◦ C. The dried product was milled
into flour in a hammer mill (Christy and Morris) (100µ mesh size). ‘Ogi’, a
traditional Nigerian fermented weaning food, was prepared from maize using
the improved method of Akingbala et al. [20]. The fresh ‘ogi’, was dried
at 60 ◦ C and milled. Shelled melon seeds (Citrullus vulgairs Schrad) were
milled into flour in a disc mill.
Four weaning foods, containing 50–80% maize flour or ‘ogi’ flour, were
prepared from blends of maize flour, ‘ogi’ flour and tempe flour with the
addition of melon seed flour as an additional source of fat in the case of
products made with cowpea tempe flour as shown in Table 1.
From a preliminary sensory evaluation, vanilla flavor was found to be the
most preferred of all the flavors tested. Vanilla flavor (powdered form) ob-
 367
Table 1. Formulation of maize-based weaning foods fortified with soybean and cowpea
tempe

Ingredients Maize-soy ‘Ogi’-soy Maize-cowpea ‘Ogi-cowpea

tempe tempe tempe tempe

Maize flour (%) 80 – 50 –

Soybean tempe flour (%) 20 20 – –
‘Ogi’ flour (%) – 80 – 50
Cowpea tempe flour (%) – – 40 40
Melon seed flour (%) – – 10 10

tained from Roche (Nigeria) Limited, Lagos was incorporated into the foods
at the level of 0.002% (w/w). Sucrose was added at 6% of the total weight.
Mameal, a soy-based weaning product of Glaxo Nigeria Plc, was used as the
reference diet.

Proximate analysis. Moisture, crude protein (N × 6.25), fat and ash were
determined by standard procedures [21]. Total carbohydrates were calculated
by difference. Energy was calculated using Atwater factors.

Bulk densities. The procedure of Akpapunam & Markakis [22] was used
in determining bulk density. The flour of known weight was transferred into
a 250 ml graduated glass cylinder and the volume was determined. The bulk
density of packed flour was calculated after tapping the cylinder until the flour
stopped settling, after about 2 min.

Reconstitution index. The reconstitution index was determined by mixing

flour with boiling water for 90 sec. The mixture was then poured into a 250
ml graduated glass cylinder and the amount of clear supernatant read after 10
min.

Viscosity. The viscosities of the weaning foods were evaluated in a Rheo-

meter (Model 286 Baroid_Hauston, Texas, USA). Slurries of 8% and 10%
(w/v) (dry solid basis) were prepared. The well – mixed porridge was poured
into the rheometer cup at 70 ◦ C and 30 ◦ C and the viscosity readings were
ascertained.

Microbiological analysis. Tempe-based weaning diets were examined mi-

crobiologically after serial dilution for total aerobic bacteria using plate count
agar (Oxoid), coliforms, presumptive pathogenic staphylococci, molds and
368

yeasts using violet red bile agar (Oxoid), mannitol salt agar (Oxoid) and
ozytetracyline glucose yeast extract agar (Oxoid), respectively.

Sensory evaluation. The paired preference test is designed to determine

which, if any, of two products is liked more. Two samples were presen-
ted simultaneously to assessors who were asked to indicate which, if any,
of the samples is preferred. Assessors were allowed to make a ‘no prefer-
ence’ response if both were disliked. The tempe – based weaning foods were
evaluated for aroma, color, consistency, taste and overall acceptability by a
fifty-member nursing mother panel. A hedonic rating a scale of 1 to 9 where
1 = dislike extremely; 5 = neither like nor dislike; 9 = like extremely [23] were
used. For the multiple comparison difference tests, porridge R(Mameal) was
used as a reference against which the assessors were asked to compare the
aroma, color, consistency, taste and overall acceptability of the tempe-based
weaning foods.

Cost of products. The cost of production of different tempe-based weaning

diets was computed, using the price of each ingredient in the local market.
Cost of the tempe based product was compared with the cost of commercial
weaning foods available in the market.

Statistical analysis. The results of the paired preference test were expressed
as the number of judges that preferred one sample to the other and then
calculated from the appropriate table of Roessller et al. [24] The results of
the hedonic rating test and multiple comparison for each diet for each sensory
characteristic were averaged to give mean scores. Differences between means
obtained from the ANOVA were ascertained using the Duncan’s multiple
range tests. The computer program used in this analysis was the Statistical
Analysis System [25]. Significance was accepted at p ≤ 0.05.

Results and discussion

The proximate composition results of maize, ‘Ogi’, tempe – based weaning

diets and Mameal are presented in Table 2. The tempe-based weaning diets
had an average protein content of 19.05% which was much higher than the
9.45% and 8.75% for maize flour and ‘ogi’, respectively. Protein content
ranged from 18.60%–19.70% while fat ranged from 8.20% (maize-cowpea
tempe) to 9.43% (maize-soy tempe). The gross energy value of 433 kcal/
100 g was better than that of maize (413 kcal/ 100 g) and ‘ogi’ (414 kcal/
100 g). The protein, fat and energy values of tempe-based weaning foods
were above the 16.7%, 6.0% and 375 kcal/ 100 g, respectively, which are the
Table 2. Chemical composition of ogi, maize, and tempe-based weaning diets versus Mameal (Dry weight basis except moisture)

Maize Ogi Maize-soy Maize-cowpea Ogi-soy Ogi-cowpea Mameal FAO/WHO

tempe tempe tempe tempe Patterns

Moisture (%) 10.13 ± 0.3 7.82 ± 0.1 8.17 ± 0.17 8.70 ± 0.1 6.70 ± 0.2 7.61 ± 0.4 5.0 <10
Protein (%) 9.49 ± 0.1 8.75 ± 0.1 19.70 ± 0.4 19.24 ± 0.3 18.60 ± 1.1 18.62 ± 0.2 18.0 >16.70
Fat (%) 3.98 ± 0.1 3.95 ± 0.1 9.43 ± 1.2 8.20 ± 0.2 8.27 ± 1.4 8.83 ± 0.2 8.0 >6.00
Ash (%) 1.58 ± 0.0 1.06 ± 0.1 1.90 ± 0.1 1.97 ± 0.1 1.72 ± 0.2 1.83 ± 0.1 3.0 –
Carbohydrate (%) 84.82 ± 0.3 86.24 68.97 70.59 71.41 70.72 66.0 –
Gross energy 413 414 426 434 435 437 408 >375
(kcal/100 g)

Mean ± standard deviation for at least 3 determinations.

369
370
Table 3. Physical properties of tempe-based weaning foodsa

Mixture Bulk density (g/ml) Reconstitution Viscosity (centipoise)

Loose Packed index (ml) 70 ◦ C 30 ◦ C

Maize flour 0.54 ± 0.02 0.89 ± 0.01 99.10 ± 0.22 50.3 ± 0.5 58.5 ± 0.4
Maize-soy tempe 0.51 ± 0.02 0.73 ± 0.01 92.30 ± 0.95 38.4 ± 0.5 45.0 ± 0.2
Maize-cowpea tempe 0.55 ± 0.01 0.75 ± 0.02 97.00 ± 0.81 37.3 ± 0.9 46.7 ± 0.6
Ogi 0.40 ± 0.01 0.69 ± 0.02 104.00 ± 0.22 65.0 ± 0.5 76.9 ± 0.7
Ogi-soy tempe 0.42 ± 0.02 0.68 ± 0.01 100.45 ± 1.20 37.0 ± 0.4 48.0 ± 0.6
Ogi-cowpea tempe 0.48 ± 0.01 0.73 ± 0.02 96.78 ± 1.01 34.2 ± 0.6 48.0 ± 0.6
FAO/WHO Patternb 0.75

a – Mean value of six replicates. b – Mitzner et al. [26].

minimum recommendations of FAO/WHO pattern for weaning foods [26].

Similar results of higher protein content than maize or ogi have been reported
[19]. The results were also comparable to the protein (20.7%), fat (8.3%), and
energy (397.1 kcal/ 100 g) reported by Mugula [17].
Losses and packed bulk densities, reconstitution index and viscosity of
tempe-based weaning foods are shown in Table 3. In general, all products
reconstituted easily in hot water. The fortified foods were comparable with
maize and ‘ogi’ in loose and packed bulk densities, although maize and maize
fortified foods were slightly higher in packed bulk density. The relatively
high bulk density of the foods indicate that packaging would be economical.
However, if the bulk densities are higher than for maize, it would suggest
unsuitability for infant foods. At 70 ◦ C, the viscosity of the products ranged
from 34.2 centipoise for ogi-cowpea tempe to 65.0 centipoise for ogi; the
range was between 45.0 centipoise (maize-soy tempe) and 76.9 centipoise for
ogi at 30 ◦ C. A similar low viscosity was reported in ogi supplemented with
30% soy tempe flour [27]. These were very low when compared with values
(200–800 CP) reported for some weaning foods based on malted grains [2].
Fortification reduced the viscosity of the diets and the values also de-
creased with increasing temperature due to a cooling effect at lower temperat-
ures. The apparent low viscosity is nutritionally beneficial in infant formulas.
Reduction in energy and nutrient density has been implicated in high pre-
valence of infant malnutrition in developing countries. Customary weaning
foods have a high paste viscosity and require dilution with water prior to
feeding [2].
Coliforms, staphylococcus, molds, yeasts and total plate counts are presen-
ted in Table 4. Coliforms and molds were absent in maize, maize-soy tempe,
ogi and ogi-soy tempe. Presumptive staphylococcus varied from 3.0 × 101
Table 4. Coliforms, staphylococcus, molds, yeasts, and total plate count of tempe-based weaning diets (cfu/g)

Counts Maize Maize-cowpea Maize-soy Ogi Ogi-cowpea Ogi-soy Cooked

tempe tempe tempe tempe porridges

Coliforms Absent <3.0 × 102 Absent Absent <3.0 × 102 Absent Nil
Staphylococcus Absent 3.4 × 102 3.0 × 101 3.0 × 102 2.0 × 103 <3.0 × 102 Nil
Molds Absent <3.0 × 101 Absent Absent <3.0 × 102 Absent Nil
Yeast Absent <3.0 × 101 Absent <3.0 × 102 <3.0 × 102 <3.0 × 102 Nil
Total plate count 3.2 × 103 4.3 × 104 4.2 × 103 3.3 × 104 3.5 × 104 3.0 × 103 <3.0 × 101

371
372
Table 5. Results of paired preference test of vanilla-flavored tempe-fortified maize-based
weaning porridges

Vanilla-flavored foods Preferring each Preferring neither Positive Significance

sample (n) sample (n) response (%)

Ogi-soy tempe 5 0 50 NS
Ogi-cowpea tempe 5 0 50
Ogi-soy tempe 6 0 60 NS
Maize-soy tempe 4 0 40
Ogi-soy tempe 6 0 60 NS
Maize-cowpea tempe 4 0 40
Ogi-cowpea tempe 6 0 60 NS
Maize-soy tempe 4 0 40
Ogi-cowpea tempe 6 0 60 NS
Maize-cowpea tempe 4 0 40
Maize-soy tempe 5 0 50 NS
Maize-cowpea tempe 5 0 50
Ten panelists.
NS at 0.01 level of probability.

to 2.0 × 103 cfu/g in maize fortified with soy-tempe and ogi fortified with
cowpea tempe, respectively. Yeasts were absent in maize and maize forti-
fied with soy-tempe but were present in other foods at a relatively low level
(<3.0 × 102 cfu/g). Total plate count ranged between 3.2 × 103 –4.3 × 104 cfu/g
and 3.0 × 104 –3.5 × 104 cfu/g, respectively, in maize and ‘ogi" fortified
with bean-tempe. Maize-based weaning foods fortified with cowpea-tempe
showed slightly higher counts to maize or ‘ogi’. This may have been due to
incorporation of melon seed flour. However, melon seeds were milled in a
disc mill in an urban market with poor environmental sanitation.
However, all the tempe – based weaning diets required cooking before
feeding to children. Most of the microorganisms were killed upon cooking
(Table 4). The thermal destruction of Staphylococcus aureus cells is con-
siderably easier to accomplish than the destruction of enterotoxin [28]. The
presence of pathogenic microorganisms in weaning foods is not desirable
[29], because they usually cause diarrhea and vomiting in children potentially
leading to dehydration.
Preliminary results on sensory evaluation revealed that vanilla flavor was
preferred out of all the flavors in the tempe-based weaning diets. Table 5 data
show the result of the paired preference test of tempe-based weaning foods.
For each pair of the tempe-based diets, there was no significant (p >0.01) dif-
ference in all cases. The results on sensory attributes showed that tempe-based
 373
Table 6. Sensory attributes of vanilla-flavored tempe-fortified maize-based weaning
porridges

Product Color Aroma Taste Consistency Overall

acceptability

Ogi-cowpea tempe 8.2a 8.2a 8.0a 7.8a 8.2a

Ogi-soy tempe 8.0a 8.1a 8.1a 7.9a 8.2a
Maize-cowpea tempe 7.9a 7.8a 7.7a 7.4a 7.7ab
Maize-soy tempe 7.8a 8.0a 7.9a 7.5a 7.8ab
Mameal 8.0a 7.9a 7.9a 7.7a 7.9ab
Mean values of fifty (50) scores.
Mean for each attribute followed by same letter are not significantly different.
1 = Dislike Extremely 9 = Like Extremely.

Table 7. Multiple comparison of organoleptic properties for tempe-fortified

maize-based weaning porridges

Product Color Aroma Taste Consistency Overall

acceptability

Ogi-cowpea tempe 5.0 7.3 5.0 6.8a 7.1

Ogi-soy tempe 5.0 7.1 5.1 7.0a 7.0
Maize-cowpea tempe 4.9 6.9 5.3 4.4b 6.7
Maize-soy tempe 4.8 6.9 5.3 4.4b 6.9
Mameal 5.0 5.5 5.0 5.0a 5.0
Mean for each attribute followed by the same letter are not significantly different at
5% level.
1 = Dislike extremely 9 = Like Extremely.

weaning diets were comparable with Mameal (Table 6). The multiple compar-
ison difference test (Table 7) showed that there were no significant (p >0.05)
difference in quality attributes assessed, except consistency. The food made
from maize received lower scores relative to ogi-fortified foods and Mameal
(control). This could have been due to the fact that most mothers are used to
the consistency of ogi as the customary weaning food. However, the lower
scores did not reduce preference in terms of the overall acceptability.
The comparative cost of producing tempe-based foods and some com-
mercial weaning foods purchased from the market are given in Table 8. It is
evident that the commercial formulas are up to six times or more costly than
the tempe-based foods. Only the cost of packaging material was not included
in the formulated foods; cans were used to package the commercial weaning
foods which undoubtedly were included in the price. A low-income (N1,000
374
Table 8. Cost of tempe-based diets and
commercial formulas

Name of Product Cost/400 g

N:K

Maize – cowpea tempe 21: 00

Maize – soy tempe 15: 00
Ogi – cowpea tempe 21: 00
Ogi – soy tempe 15: 00
Mameal 120: 00
Cerelac 150: 00
Babeena 130: 00
Nutriend 120: 00

N = Naira, K = Kobo (100k = N).

per month) family cannot afford to buy the commercial weaning foods. Thus,
a simple method to ensure safety, acceptability and low-cost of production is
greatly needed. Because the tempe-based foods are adapted from culturally
acceptable foods, they have the potential of fulfilling such a need.

Conclusion

In conclusion, this preliminary study shows that there is a potential for using
tempe-based formulas in weaning diets.

References

1. United Nations Children Emergency Fund (1996) Child Nutrition. New York, USA:
Published for UNICEF by Oxford Universities Press, pp 82.
2. Desikachar HSR (1979) Development of weaning foods with high caloric density and
low hot paste viscosity using traditional technologies. Food Nutr Bull 2: 21–33.
3. National Academy of Sciences (1979) Tropical Legumes: Resource for the Future. USA
331.
4. Ojofeitimi EO, Afolabi OA, Fapoluno OO, Grisson FE, Oke OL (1984) The use of
black – eyed cowpea-maize gruel mixture, ‘Ewa-Ogi’ in the treatment and prevention
of infantile protein malnutrition. Nutr Reports Int 30: 841–852.
5. Abiodun PO (1991) Use of soyabean for the dietary prevention and management of
malnutrition in Nigeria. Acta Pgediatric Scand Suppl 374: 175–185.
6. Liener IE, Kakade ML (1980) Protease inhibitor. In Liener IE (ed), Toxic Constituents
of Plants Food Stuffs, 2nd edn, New York: Academic Press.
 375

7. Zamora AF, Fields ML (1979) Nutritive quality of fermented cowpea (Vigna sinensis)
and chickpea (Cicer arientum). J Food Sci 44: 234–236.
8. Hermana O, Roedjito SW, Karyadi D (1972a) Preparation of tempeh mold inoculum and
observation on its activity during storage. Paper presented at the Fourth International
Fermentation Symposium, Kyoto, Japan 19–25 March 1972, pp 9.
9. Hesseltine CW (1967) Fermented products – miso, sufu and tempeh. In proceeding
of International conference on soyabean protein foods. Washington DC: Peoria U.S.
Department of Agriculture, pp 170–180.
10. Ko SD, Hesseltine CW (1979) Tempe and related foods In Economic Microbiology,
Microbial Biomas vol 4 London: Academic Press, pp 115–140.
11. Murata K, Ikehata H, Myamoto T (1967) Studies on the nutritional value of tempeh I. J
food Sci 32: 580–586.
12. Murata K, Ikehata H, Edani Y, Koyanagi K (1971) Studies on the nutritional value of
tempeh II Rat feeding test with tempeh, unfermented soybeans and tempe supplemented
with amino acids. Agric Biochem 35(2): 233–241.
13. Omotola D (1984) Cowpea Tempe and its Nutritional Value. Paper presented at the
International course in Food Science and Nutrition, 23rd May, 1984. Wageningen, The
Netherlands.
14. Wang HL, Ruttle DI, Hesseltine CW (1968) Protein quality of wheat and soybeans after
Rhizopus oligosporus fermentation. J Nutr 96: 109–114.
15. Winarno FG, Reddy NR (1986) Tempe. In Reddy NR, Pierson HD, Salunkhe DK (eds),
Legume-based Fermented Foods. Boca Raton: CRC Press Inc, pp 85–117.
16. Zamora RG, Veum JL (1979) The nutritive value of dehulled soybeans fermented with
Aspergillus oryzae or Rhizopus oligosporus as evaluated by rats. J Nutr 109: 1333–1339.
17. Mugula JK (1992) Evaluation of the nutritive value of maize – Soybean tempe as a
potential weaning food in Tanzania. Int J Food Sci Nutr 43: 113–119.
18. Osundahunsi OF, Aworh OC (2000) Moisture equilibria of tempe – fortified maize-based
complementary foods. Appl Trop Agric 5(1): 41–47.
19. Egounlety M, Aworh OC (1991) Production and physico-chemical properties of tempe
– fortified maize-based weaning foods. In Sefa-Dedeh S (ed), Energy Foods from Grain
Legumes. Proc seminar on development of high protein. Accra, pp 141–153.
20. Akingbala JO, Onochie EU, Adeyemi IA, Oguntimein GB (1987) Steeping of whole and
dry millsed maize kernels in Ogi preparation. J Food Proc Pres 11: 1–11.
21. AOAC (1984) Official Methods of Analysis. 14th ed, Washington DC: Association of
Official Agricultural Chemists.
22. Akpapunam MA, Markakis P (1981) Physicohemical and nutritional aspects of cowpea
flour. J Food Sci 46: 972–973.
23. Larmond E (1977) Laboratory methods for sensory evaluation of foods. Ottawa, Canada:
Department of Agriculture Publication. no 1673, 74 pp.
24. Roessler EB, Pangborn RM, Sidel JL, Stone H (1978) Expanded statistical tables for
estimating significance in paired preferenced, paired – difference, duo – trio and triangle
tests. J Food Sci 43: 940–943 & 947.
25. SAS (1985) Statistical analysis system user’s guide: Statistics. Cary NC: SAS Institute.
26. Mitzner K, Scrimshaw N. Morgan R (1984) Improving the nutritional status of children
during the weaning period. HOVIPREP (Home – and – village – prepared weaning
foods). Cambridge Massachusetts: Publication, Massachusetts Institute of Technology.
27. Egounlety M, Syarief R (1992) Study on the supplementation of ogi with tempe. Nig
Food J 10: 92–102.
376

28. Nwosu VC (1986) The incidence of coagulase – positive staphylococcus in market meat
in Awka, Anambra State Nigeria. Nig Food J 4: 106–111.
29. PAG (1972) Protein advisory group guideline for the production and use of dried protein
foods. New York: United Nations PAG Guideline No 11.