Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Studies on the use of soybean food in infant feeding in China and the development of formula 5410

Chi-yuan Chou
Institute of Health, Chinese Academy of Medical Sciences, Beijing, People's Republic of China


In China the soybean has been recognized as a nutritious food since ancient times. Numerous historical records have shown that not only cooked soybean curd and its products were considered delicious dishes, but also soybean milk. Soybean milk now is a popular breakfast food item sold in many small restaurants in the morning in almost every large and small city, but it had never been used to feed infants before 1927.

In the vast agricultural areas of China, animal milk is not a regular daily food item. When infants cannons obtain sufficient mother's milk, it is only natural to use cereal flours such as wheat or rice and wheat plus a little cane sugar as milk substitutes. This traditional method of feeding has been practiced over hundreds or even thousands of years.

In the light of modern knowledge of nutrition, it was realized that such cereal preparations are not adequate to meet the requirements of a growing child. TO modern medical and health scientists, it has become an urgent and interesting problem to search for an inexpensive but nutritious infant food to substitute for animal milk and the traditional cereal flour products when mother's milk is inadequate.

Dr. Ernest Tso, Professor of Pediatrics of the former Peking Union Medical College in 1928, was the first in China to use fresh soybean milk to feed an infant (1). Fresh soybean milk was supplemented with calcium lactate, common salt, and cane sugar. During the eight months of feeding with this basic diet, the infant received other foods such as egg yolk, fish liver oil, orange juice, and green vegetable puree, with soybean milk, protein furnishing about 20 per cent of the total dietary energy.

Following this successful attempt, Tso, Yee, and Chen (2), Tso and Chu (3), Tso and Chang (4), and Fan, Wu, and Chu (5) continued to work with fresh soybean milk, spray-dried soybean milk powder, or roasted soybean flour diets for a number of infants in the hospital. Guy and Yeh (6, 7), working at the Peking First Health Station, fed 15 infants fresh soybean milk and 49 infants roasted soybean flour preparations at the babies' own homes and reported that all of the infants had normal weight and height gains. Research along this line was then interrupted for 12 years.


The search for a suitable milk substitute for infant feeding, using soybean as the main source of protein, began again in 1953 with a team in the Department of Nutrition of the then National Institute of Health, now the Department of Nutrition and Food Hygiene of the Institute of Health of the Chinese Academy of Medical Sciences. On the basis of the findings of previous workers and newer knowledge of nutrition, the aim was to construct a milk substitute with easily obtainable raw materials and a simple method of preparation, nutritionally comparable to human milk or cow's milk formula, palatable and acceptable to young infants, and to be sold at a reasonably low price to meet the financial status of most families, especially those in the rural areas.

Unfavourable factors such as trypsin inhibitor, saponin, and haemagglutinin were removed by soaking the beans in water until the water content reached 35 per cent, after which the beans were steamed under ordinary pressure. The goitrogenic activity of the bean was counteracted by the addition of potassium iodide in amounts up to 0.25 mg per kg of the product, based on the results of experiments with rats (8).

Egg yolk was included in the formula to raise the methionine content. Calcium was supplied with palatable, degelatinized animal bone powder. Millet, which is rich in riboflavin, was added in the early days, but later replaced with the synthetic vitamin. Both the formula and the product are termed simply SMS-5410 throughout this paper. The control diet used in the various experiments contained cow's milk powder and is hence designated CMP. The formula SMS-5410 is shown in Table 1. For methods of preparation of the product, see Appendix I. The nutrient contents of the product are shown in Tables 2 and 3.


The infants and their foods The nutritive effect of SMS-5410 was tested on a group of 100 infants (51 boys and 49 girls), including four prematures and 10 undernourished. Another group of 43 infants (24 boys and 19 girls), serving as controls, were fed their own mother's milk and/or fresh cow's milk during the same period. Among the infants in the two groups were five sets of twins; one of each pair was fed with SMS-5410 while the other was fed with human and cow's milk formula. The great majority of the infants were cared for by their own mothers at home, while the rest (10 per cent) were kept in the nursery of the Maternal and Child Health Station.

TABLE 1. Soybean Milk Substitute: Formula 5410 (SMS-5410)

Ingredients Parts by Weight
Soybean flour 28.0
Rice flour 45.0
Cane sugar 16.5
Egg yolk powder 5.0
Soybean oil 3.0
Bone powder 1.5
Millets 0.5
Common salt 0.5
Potassium iodide** 0.25 mg/kg mixture

* Later replaced with synthetic riboflavin, 6 mg/kg
** Later supplemented

TABLE 2. Nutrient Contents of SMS-5410 per 100 g Dry weight

Protein 17.1 g
Fat 12.8 g
Carbohydrate 66.0 g
Fibre 0.8 g
Ash 3.2 g
Energy 448 keel
Calcium 692 mg
total 330 mg
phytin 97 mg
Iron 5.9 mg
Thiamine 0.11 mg
Riboflavin 0.63 mg
Niacin 1.09 mg
Carotene 40.0 mg

TABLE 3. The Amino Acid Content of SMS-5410 per 100 g

Isoleucine 726
Leucine 1,551
Lysine 889
Methionine 226
Phenylalanine 884
Threonine 758
Tryptophan 206
Valine 1,022

Most infants were under two months old at the beginning of feeding with SMS-5410; the youngest one was two weeks old. Feeding was continued until an of the infants reached their first brithday. Fish liver oil and vitamin C were supplemented to the infants in both groups from the very beginning. Vegetable soup (the water only) and fruit juice were given beginning at four months of age, followed by various other foods, including rice porridge, wheat flour noodles, bread, crackers, mashed meat and vegetables or egg yolk, to furnish not more than one-third of the total calorie intake by the age of one year, depending on the financial condition of the family.

The infants in both groups attended the well-baby clinic of the Maternal and Child Health Station at weekly intervals to get the SMS-5410, fish liver oil, and vitamin C, and to have body measurements and clinical examinations. Before infants were five months of age, be., before the addition of other foods, the average daily calorie intake was 118 calories (normal boys 123 calories, normal girls 114 calories); that of the four prematures and the 10 undernourished were 140 and 145 calories, respectively.

The appetite of those taking the SMS-5410 was good; there were no digestive disorders or abnormal stools.


Average blood haemoglobins of the babies in the experimental and control groups were 12.4 and 12.5 per cent, and red blood cell counts were 4.58 and 4.59 million per cubic millimeter, respectively. Radiographs of the wrists revealed that six infants had signs of rickets before taking SMS-5410, but all became normal by the age of one year.

The individual growth curves of the boys and girls in the experimental group, excluding the premature and undernourished infants, are shown in Figures 1 and 2. The curves all fall within the normal range (average + 2 S.D.) (9). Average growth curves indicated that the babies in the experimental group were smaller in size at the beginning, but they grew faster. Their length matched that of the controls at the age of 7 months, while their weight matched the controls at the age of one year (Figure 3).

FIG.1. Individual Growth Curves of 41 Normal Boys Fed with SMS-5410

FIG.2. Individual Growth Curves of 39 Normal Boys Fed with SMS-5410

FIG.3. The Average Growth Curves of Boys and Girls of the Experimental and Control Groups

The representative growth curves of the premature and undernourished infants showed that their size was much below the lower limit of the normal range at the beginning of the study, but all fell into the normal range at about 5 or 6 months.


Observations were made on the utilization of the nitrogen, calcium, phosphorus, and energy content of SMS-5410 in nine male infants. A cow's milk formula (CMP) was used as the control diet At the same time, a similar soybean preparation containing more soybean flour, and therefore having a higher protein content, abbreviated as SMS-HP, was also studied. The composition and nutrient contents of the control and experimental diets are shown in Tables 4 and 5, respectively. The infants were fed the experimental and control diets in alternate periods and with or without supplementation of fish liver oil (FLO).

Infants consuming the cow's milk powder diet began at an average age of two months and 12 days, weighing 5.12 kg, with an energy intake of 112 kcal/kg/day. Those fed SMS-5410 also entered at two months and 12 days, averaging 5.26 kg in weight, with an energy intake of 117 kcal/kg per day. Infants given SMS-5410 plus fish liver oil were about three months and 21 days old, weighed 6.17 kg, and had an energy intake of 108 kcal/kg per day. Those consuming cow's milk powder plus fish liver oil were also three months and 21 days old, weighed 6.21 kg, and had an average energy intake of 107 kcal/kg/day. Infants fed high-protein soybean milk and fish liver oil averaged four months and 22 days of age, weighed 6.38 kg, and consumed 108 kcal/kg/day. The results of nitrogen, calcium, and phosphorus metabolism studies are summarized in Table 6.

Nitrogen utilization. The rates of absorption (i.e., the digestibility) of nitrogen, calcium, and phosphorus in infants consuming SMS-5410 are all lower than in those on the CMP diet, as expected. However, nitrogen retention in the CMP and SMS-5410 diets were about 190 mg/kg/day, and about 160 mg/kg/day in periods when fish liver oil was added. These values are in close agreement with values observed by previous workers (10). Thus, it may be concluded that the SMS-5410 diet can provide enough nitrogen (protein) to promote nitrogen retention of infants when the nitrogen intake is 670 to 730 mg/kg/day, equivalent to about 4.2 to 4.69 protein/kg/day.

Calcium utilization. That the vitamin D in fish liver oil improves the absorption and retention of calcium in the diet is once more demonstrated in this metabolic study. Figures in Table 6 show that in the SMS-5410 period, when calcium intake was 149 mg/kg/day, calcium retention was 24 mg per kg/day and was raised to 58 mg by adding fish liver oil when calcium intake was 142 mg.

TABLE 4. The Composition of the Experimental and Control Diets Used in the Metabolic Study (Parts by Weight)

  Control Diet Experimental Diets
Ingredients CMP SMS-5410 SMS-HP
Cows milk powder 100 ----- -----
Soybean flour ----- 28.0 43.0
Rice flour ----- 45,0 30.0
Cane sugar 40 16,5 16,5
Egg yolk powder ----- 50 50
Soybean oil ----- 3,0 3.0
Bone powder ----- 1.5 1.5
Riboflavin (millet) ----- 0.5 0.5
Common salt ----- 0.5 0.5

Supplements: fish liver oil supplying vitamin A, 1500 I.U., vitamin D 500 I U. per day, ascorbic acid 25 mg per day, dissolved in distilled water. K I 0.25 mg of the mixture

TABLE 5. The Protein, Calcium, Phosphorous and other Contents (Per 100 9) of the Experimental and Control Diets Used in the Metabolic Study

  Control Diet Experimental Diets
Moisture g 2.0 3.8 3.2
Protein g 18.5 16.6 22.6
Fat g 19.6 12.8 15.5
Carbohydrate mg 55.9 62.9 53.8
Energy kcal 474 433 445
Fibre g 0 0.7 0.9
Ash g 4.0 3.2 3.9
Calcium mg 701 546 566
Phosphorus, total mg 604 491 552
Phytin mg 0 97 109
Energy distribution
Protein % 15.5 15.3 20.3
Fat % 37.1 26.6 31.3
Carbohydrate % 47.3 58.1 42.3

The literature has recorded that when infants were fed a cow's milk diet supplemented with vitamin D (340 I.U. daily), retention was 55 mg ;10, p. 83). Fan and co-workers reported that when infants were fed a soybean diet with an average daily calcium intake of 197 me, average daily retention was 53 mg (5). It seems, therefore, that a calcium retention of around 55 mg/kg/day is accept table as an adequate amount and can be provided by SMS-5410 as long as the diet is supplemented with fish liver oil or vitamin D.

A calcium retention of about 55 mg can be achieved at intakes from 100 to 150 mg/g/day when the calcium furnished by SMS-5410 is in the form of degelatinized bone powder. The average weight of infants at the age of six months is 6 to 8 kg. When calculated on this basis, the daily calcium supply would be 799 to 1200 mg/day.

Phosphorus utilization. With the increased absorption and retention of calcium, some figures indicated a corresponding increase in phosphorus, but statistical analysis showed that the difference was not significant, The ratio of calcium and phosphorus retention is 1.6:1 in both the experimental and control diet periods, and is considered normal.

The influence of a larger proportion of soybean flour in the SMS-54 10 formula on the utilization of nutrients A higher percentage of soybean flour in the higher protein product, SMS-HP, evidently lowered the absorption of both calcium and phosphorus, apparently because of greater faecal excretion. Therefore, it is not desirable to use more than 28 per cent soybean flour to raise the protein intake as some have suggested.

The stools of the infants on the various diets. One of the criteria for examining the quality of an infant food such as soybean milk is to see its effect on the stools, not by impression, but by statistical analysis of the stool record over a comparatively long period. The frequency, amount, and form of the stools of nine of the infants were closely watched during their stay in the metabolic unit for four to five months. It can be seen from Table 7 that: (ii in the SMS-5410 periods stools were fewer than in the CMP diet periods, and (ii) there were more formed and fewer liquid stools in the SMS periods than in CMP periods. These records clearly contradict the usual complaint by some mothers that the soybean infant food causes more frequent, loose stools. Of course, the weights of the stools, either wet or dry, are larger in the SMS periods (Table 7) because of the fibre content in the soybean food. How ever, the more bulky stools did not do any harm except during the higher-protein soy milk plus fish liver oil diet in which calcium and phosphorus absorption and retention were lower than those in the SMS-5410 plus fish liver oil period.

Energy utilization. The utilization of the energy in the experimental and control diets may be calculated from the energy-producing nutrients, protein, fat, and carbohydrate, as shown in Table 8a, b. Energy utilization of food means the percentage of absorption of the total energy contributed by protein, fat, and carbohydrate. It will be seen from the bottom line of Table 8a, b that, on the whole, the utilizable energy of the experimental diet (90.9 and 92.4 per cent) is only slightly lower than, or almost equal to, that of the control diet (91.7 and 93.7 per cent) in the corresponding dietary periods. When the three energy-producing nutrients are considered separately, protein is less and fat is more utilizable in the SMS-5410 than in the CMP diet, while the carbohydrate in the two diets was almost equally utilizable (digestible. In addition to the three nutrients, about one gram of fibre was ingested daily, of which about 15 per cent was digested, contributing only about 0.6 calorie per day and therefore not included in the calculation. The average daily energy intake of the individual infants in different dietary periods was between 110 to 120 calories per kilo.


Contents - Previous - Next