Contents - Previous - Next

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

Clinical nutrition

Symptoms of primary adult-type lactose maldigestion in Tswana children

François P. R. de Villiers


We investigated the milk-drinking habits in a population with a high frequency of lactose maldigestion to establish the clinical (rather than biochemical) importance of this condition. The occurrence of symptoms after milk intake was established. The research took place at a rural hospital and school in Bophuthatswana. The relevance to lifestyle of milk intolerance was established using symptoms reported in a quesfionnaire by 486 subjects, and symptom scores were calculated after standard milk meals for 92 subjects. A further 32 hospitalized children, age three to seven years, had symptom scores calculated after they underwent milk tolerance tests. Although only 38% of teenagers had no symptoms after a 500-ml milk meal, almost 80% had previously ascribed no symptoms to milk drinking. A higher proportion (61 %) of younger children had no symptoms after a standard milk meal. The symptoms produced by lactose maldigestion were not very troublesome in this population. Most lactose maldigesters said they would continue drinking milk, and only 20% of those who suffered symptoms stated they would not Subjects with a symptom score of 3 or more (out of 12) and those who reported bloating tended to drink less milk than others. The total symptom score (>3) or the severity of bloating (2 or 3 out of a possible 3) influenced subjects to assert that they would reduce future milk consumption. Symptoms were less likely to occur in younger children, 61% having no symptoms after a large milk load, compared with 38% of teenagers.


It is possible to have lactase non-persistence without suffering milk intolerance; this is one of the reasons that diverse opinions exist about whether milk should be excluded from such people's diets [1-4]. An important aspect of this study was the investigation of the occurrence of symptoms in a population with a high frequency of primary adult-type lactose maldigestion to establish how important this condition is clinically (rather than biochemically). The prevalence of lactose maldigestion in the rural Tswana is 90% to 95% in adults and children from the age of 10 years [5]. In younger children (age 3-8 yr), the prevalence is 87.5%. In addition, most Tswanas lose lactase activity before age 3 years, although in some this occurs between age 3 and 7 years [5]. Yet many adult Tswanas reported having developed symptoms of lactose maldigestion only in late adolescence and early adulthood [5]. Thus, the occurrence of symptoms after milk intake had to be established.

Since the biochemical abnormality (enzyme nonpersistence) might have been present before symptoms manifested themselves, adverse effects were also determined to find out whether younger subjects, although lactose maldigesters, might nevertheless tolerate milk without difficulty.


The study was carried out in Saulspoort, a rural town about 280 km west-north-west of Johannesburg. The protocol was approved by the committee for research on human subjects of the University of the Witwatersrand.


Secondary-school children

The study population consisted of all Tswana children in grades 5 and 6 at the Moruleng Middle School in Saulspoort. The procedure involved was explained to all 486 pupils, informed consent was obtained, and a questionnaire was administered under the supervision of a teacher and the researcher. Subjects who did not have at least one parent who was a Tswana were excluded. The final sample consisted of 469 children.

Two groups were established based on age: those 10 to 12 years old and those 13 to 16 years old. Forty-eight pupils in each group were randomly selected to take pan in the lactose maldigestion-prevalence study. Four declined to take pan. Six subjects were tested at each session. After an overnight fast, a milk tolerance test was performed on each subject.

Young hospitalized children

Most Tswana children in the area stay at home until they reach primary school age. Accordingly, children admitted to the children's wards of George Stegmann Hospital who were between three and eight years old were enrolled. Patients with the following problems were excluded: acute febrile illness, malnutrition (kwashiorkor, marasmic kwashiorkor, marasmus), acute gastroenteritis, and current severe illness. Also excluded were children due for discharge before the end of the study period and those whose parents refused to give informed consent.

The subjects generally were the least ill patients in the ward and did not have an obvious cause of secondary lactose maldigestion. Each week between four and six children were selected to take part in the study. Their sex, age, and date of birth were recorded. On Monday morning a standard lactose tolerance test was performed on each child, and on Tuesday morning a milk tolerance test was performed. A non-lactose-containing diet was given over this period.

Milk and lactose tolerance tests

For the milk tolerance test (MTT) [6], blood specimens were obtained for glucose estimations before, and 15 and 30 minutes after, administration of a test meal to a fasting subject. The meal consisted of 20 ml of milk per kilogram body weight to a maximum of 500 ml. A blood glucose rise of more than 0.5 mmol/L (9 mg/dl) over fasting value indicates a lactose digester.

This test is similar to the standard lactose tolerance test (LTT) in which the test meal is 2 g lactose per kilogram body weight to a maximum of 50 g lactose, administered as a 10% aqueous solution. The cut-off point is 1.1 mmol/L (20 mg/dl).

Symptom scores

The symptom score was devised as more accurate and objective than symptom reports for measuring and comparing symptoms developing after lactose intake. The day after the LTT, subjects were questioned about their symptoms during the 24 hours after the test. Bloating, cramps, diarrhoea, and flatulence were assessed on a scale of zero to 3 (0= absence of the symptom, to 3 = severe, interfering completely with the subject's usual activities).

The scores for each symptom were added to give a total symptom score (range 0-12). This symptom score follows that of Stephenson and Latham [7]; however, their classification (mild, moderate, severe) depended only on the subjects' interpretations. To maximize objectively, the level of interference with daily activities by symptoms was defined. This means that the classification is less dependent on individual interpretation.

Symptoms due to lactose maldigestion

The questionnaire administered to the schoolchildren contained questions about milk preference and use. The responses were correlated with the symptom scores obtained after carbohydrate tolerance tests.


Symptoms experienced by adolescents after drinking milk

Table 1 shows the symptoms experienced by pupils drinking fresh milk as reported in the questionnaire; the 92 who took the MTT were compared with the whole group of 469. Most (79.3% and 78.4%, respectively) reported no symptoms at all. Bloating and diarrhoea were the most commonly reported.

The symptom score was calculated for subjects who underwent the MTT (in three children the score was not obtained). No subject attained a score higher than 5 out of a possible 12. Table 2 presents the score reported by each subject, classified by maximum glucose rise values into three groups: lactose maldigesters (rise <=0.3 mmol/L), probable lactose maldigesters (rise 0.4-0.5 mmol/L), and lactose digesters (rise >0.5 mmol/L).

The mean symptom score attained by lactose digesters (0.833) was lower than that of lactose maldigesters (1.276), but this difference was not statistically significant (p > .1). Fifty percent and 36.8%, respectively, had no symptoms at all (symptom score of zero).

The symptoms reported in the questionnaire were compared with those experienced after the MTT (i.e., symptom score; table 3). The symptom scores attained by the subjects (columns) are compared with the symptoms described by the same subjects (rows). There was some correspondence between the two, but also many cases where they differed.

Many more subjects reported symptoms after the MTT (61.8%) than during day-to-day living (20.7%). Of the 19 who reported symptoms in the questionnaire, the symptoms were elicited during the MTT in only 5 (26.3%) of them. The others experienced different symptoms from those originally reported, or none at all.

TABLE 1. Symptoms experienced after drinking milk: Comparison of total group with those who had the MTT



Total group

No. of


% No. of


None 73 79.3 368 78.4
Cramps 1 1.1 8 1.7
Abdominal pain 2 2.2 11 2.3
Diarrhoea 5 5.4 21 4.5
Flatulence 1 1. 1 6 1.3
Bloating 8 8.7 36 7.7
Nausea or vomiting 0 0 4 0.8
Two symptoms 2 2.2 14 3.0
Three or more symptoms 0 0 1 0.2
Total 92 100 469 100

TABLE 2. Symptom score after the ingestion of 500 ml milk for the MTT

Symptom score Group A Group B Group C Total
0 28 3 3 34
1 23 2 2 27
2 8 2 0 10
3 11 0 1 12
4 5 0 0 5
5 1 0 0 1
Total 76 7 6 89
Mean score 1.276 0.857 0.833 1.213
Standard deviation 1.343 0.900 1.169  

Group A, maximum glucose rise <=0.3 mmol/L; group B. maximum glucose rise 0.4-0.5 mmol/L; group C, maximum glucose rise <=0.6 mmol/L.

TABLE 3. Symptoms reported in the questionnaire compared with symptom scores after the MTT

Symptoms reported Symptom score
0 1 2 3 4 5
None 24 24 8 8 5 0
Cramps 0 0 0 1 0 0
Abdominal pain 1 0 0 1 0 0
Diarrhoea 1 2 0 1 0 1
Flatulence 1 0 0 0 0 0
Bloating 6 1 0 1 0 0
Two symptoms 0 0 2 0 0 0

Symptoms and future milk consumption In adolescents

Subjects were asked whether, after their experience with the MTT, they would still drink milk in the future. Most (86.8%) reported that they would continue to drink milk, 11% would never drink milk again, and 2.2% said they might drink milk, but in small amounts only.

Table 4 shows the effect of the symptom score of the 76 lactose maldigesters on their decision whether or not to drink milk in the future. Most (86%) would continue drinking milk. Only 20.8% of maldigesters who suffered symptoms during the MTT would not drink milk in the future. Those with a symptom score of zero or 1, and 7 of the 8 subjects with a score of 2, were prepared to continue drinking milk; but of the 17 subjects with a symptom score of 3 or more, 53% would not drink milk again.

The maximum glucose rise value during the MTT was compared with the subjects' decision on whether to drink milk in future, after experiencing the effects of the test (table 5). No correlation was seen. Similarly, the effect of the maximum glucose rise value during the MIT on the subjects' preference for sour or fresh milk (table 6) was not statistically significant.

Symptoms experienced by young hospitalized children after tolerance tests

One subject did not take part in the complete study, so his symptom scores are not available. The symptom score had a denominator of 6 if diarrhoea and abdominal cramps only were recorded, or 12 if flatulence and bloating were also recorded. Symptom scores obtained after the MTT and LTT are shown in table 7. After the MTT, 61.3% had no symptoms, and 41.9% had none after the LTT.

We could not correlate symptom score with future milkdrinking plans, since very few of the children could envisage having this choice.

Fewer younger children than teenagers had symptoms after the MTT: 36.8% and 61.3%, respectively.

TABLE 4. Effect of symptom score on the decision by lactose maldigesters to drink milk in future

Symptom score


Percentage who will drink milk
No Maybe Yes
0 0 0 28 100
1 0 0 23 100
2 1 0 7 87.5
3 5 0 6 54.5
4 3 1 1 20
5 1 0 0 0
Total 10 1 65 85.5

No = will not drink milk in future; maybe = may possibly drink milk, or will drink very small amounts; yes = will drink milk in future.

TABLE 5. Effect of maximum glucose rise during the MTT on the decision to drink milk in future

Maximum glucose rise



No Maybe Yes
0 7 0 35 42
0.1-0.3 3 1 29 33
0.4-0.5 0 0 7 7
20.6 0 0 6 6
Total 10 1 77 88

X2 = 4.649; df = 6; p = .5895 (NS).

TABLE 6. Effect of maximum glucose rise during the MTT on subjects' milk preference

Maximum glucose rise


Milk preference

Sour Fresh Both Neither
0 5 8 28 3 44
0.1-0.3 2 3 27 3 35
0.4-0.5 1 1 5 0 7
20.6 0 2 4 0 6
Total 8 14 64 6 92

X2 = 5.465; df = 9; p = .792 (NS).

TABLE 7. Comparison of the maximum glucose rise and symptom score found during the MTT with that of the LTT



Maximum rise Symptom score Maximum rise Symptom score
0.1 0/6 0 0/6
0 1/6 0 1/6
0.2 0/6 0.6 0/6
0.4 0/6 0.7 0/6
0.3 0/6 0.4 0/6
0.1 0/6 0.4 0/6
0.3 0/12 0.4 1/12
0 0/12 0.2 0/12
0.4 0/12 0 2/12 (0/6)
0.1 1/12 (1/6) 0.4 3/12 (1/6)
0.2 2/12 (2/6) 0.4 0/6
0.1 0/12 0.3 2/12 (1/6)
1.6 1/12 (1/6) 1.9 2/12 (2/6)
0.6 0/12 0.5 0/12
0.4 0/12 0.9 0/12
0.5 2/12 (1/6) 0 4/12 (2/6)
0 0/12 0 1/12 (0/6)
0.2 0/12 0 2/12 (2/6)
0 0/12 0 2/12 (2/6)
0.2 1/12 (0/6) 0 2/12 (0/6)
0.3 0/12 0.6 0/12
0.3 0/6 0.1 0/6
0.6 0/12 0.5 1/12 (0/6)
0.4 1/12 (1/6) 0.3 2/12 (216)
0.2 2/12 (1/6) 0.9 2/12 (1/6)
1.9 0/12 3.6 1/12 (0/6)
0.2 3/12 (1/6) 0.2 2/12 (2/6)
0.8 4/12 (2/6) 0.5 4/12 (2/6)
0.1 5/12 (2/6) 0.1 6/12 (2/6)
0.3 0/12 0 5/12 (2/6)
0.1 4/12 (1/6) 0.3 0/12
0.3 _ 0.7 _

Correlation coefficient = 0.858988; df = 30; t value = 9.1893; p < .001.


Symptoms experienced by adolescents after drinking milk

Of the 469 pupils tested, 78.4% reported that drinking fresh milk caused no symptoms. Thus the high prevalence of lactose maldigestion in Tswana adolescents would not have been expected on the basis of the symptoms reported in the questionnaire.

Despite random selection, bias may still have occurred. The symptoms reported in the questionnaire by those who were tested for lactose maldigestion were accordingly compared with those of the total school population. A remarkable consistency was seen in the symptoms between the two groups (X2 = 0.3574, not significant), showing that the MTT group was a true subset of the population. Thus it would seem that most pupils do not ascribe any symptoms to the consumption of milk, even though most are lactose malabsorbers.

After a large (500-ml) milk load for the MTT, 50% of the lactose digesters and 36.8% of the maldigesters had no symptoms; 38.2% of the whole group had no symptoms. It is clear that symptom recall over 24 hours elicits a higher frequency of reported symptoms than was yielded in answer to the question of whether milk causes symptoms. The population under study did not seem to have a general awareness of symptoms caused by lactose maldigestion, and those who did, did not expect them to manifest until adulthood. This may account for the low percentage (20.7%) ascribing symptoms to milk drinking, even though 93.5% were lactose maldigesters and 61.8% actually reported symptoms after a provocation test.

It must be noted that even if one event follows another, the second event is not necessarily caused by the first; for example, it is unlikely that the three lactose digesters who suffered symptoms did so because of lactose remaining in the gut. In some lactose-intolerant children with recurrent abdominal pain, the pain is not caused by lactose maldigestion [8]. In addition, eliminating lactose does not affect the overall frequency of improvement in recurrent abdominal pain [9]. Ascribing symptoms after lactose ingestion as solely due to lactose is thus an example of the logical fallacy of post hoc ergo propter hoc [10]. If one accepts that 50% of lactose digesters have symptoms similar to those caused by lactose intolerance but not attributable to the condition, it is possible that 50% of lactose maldigesters have similar symptoms that cannot be attributed to lactose maldigestion. Thus, the symptoms of only 13.2% of lactose maldigesters would be due to lactose maldigestion. However, the sample of lactose digesters (6) was too small for this calculation to be reliable. What is certain is that some of the symptoms reported by both digesters and maldigesters after the milkloading test were not due to lactose maldigestion.

Symptoms experienced in day-to-day living compared with the symptom scores show many differences. Nine (47.4%) of the 19 subjects who reported symptoms in the questionnaire experienced none after the MTT. Sixty-nine subjects reported no symptoms in the questionnaire; 24 (34.8%) actually experienced no symptoms after the test. Six (75%) of the eight subjects who said they experienced bloating after drinking milk reported no symptoms after drinking 500 ml milk.

The MTT involves the maximum intake of milk that any of these children could be expected to drink all at once, which is much more than their recorded habitual total daily intake (101 ml) [11]. The way it is administered, furthermore, ensures that the worst possible symptoms will be elicited. It is thus not surprising that in many subjects the symptom score was higher than would be expected from the questionnaire. It is less clear why there was such a poor correspondence between the types of symptoms reported in the two different settings, unless the explanation offered above is in fact true: that is, that the symptoms are often not caused by lactose maldigestion at all.

In general, those diagnosed as lactose maldigesters had higher symptom scores than lactose-tolerant subjects. The mean symptom scores were 1.276 and 0.833, respectively (not significant). The degree of lactose digestion, therefore, has some bearing on the severity of symptoms experienced, but the symptom score did not accurately reflect a subject's lactosedigestion status.

Several workers have tried to predict lactosedigestion status from symptoms [7, 12-14]. Our results showed that neither a history of symptoms experienced in day-today living nor a symptom score after a milk load will accurately predict biochemical lactose-digestion status. The importance of symptoms is in how they determine the subject's attitude toward drinking milk.

Symptoms and future milk consumption among adolescents

The subjects' symptom scores were compared with their decisions whether or not to drink milk in future. Only 11% of the whole group would never drink milk again. Of the 76 lactose absorbers, those with minimal or no symptoms after the MTT were prepared to continue drinking milk. This decision was not influenced by the composition of the score; several mild symptoms were as important in influencing the decision as one severe symptom.

The type of symptom, however, did make a difference. The 3 subjects who scored 2 for bloating would not drink milk again, and only 4 of the 10 who scored 2 for diarrhoea were prepared to continue their normal milkdrinking habits. The 3 subjects who scored 2 or 3 for cramps were all prepared to drink some milk, whereas only one of the 2 who had moderate flatulence would drink milk. Therefore, bloating seems to be the most troublesome and diarrhoea the second most troublesome symptom, and cramps, surprisingly, seem not to have influenced the subjects' future milk-drinking habits.

The maximum glucose rise value during the MTT did not affect the decision whether or not to drink milk in future. This is not surprising, since patients perceive symptoms, but not adjustments in homeostatic or absorptive mechanisms, unless they cause symptoms. Lactose digestion status per se did not, therefore, influence the decision whether or not to drink milk, but symptoms caused by it did. Similarly, and for the same reasons, the height of the plasma glucose rise during the MTT did not affect pupils' preference for sour or fresh milk, both, or neither.

Ascribing symptoms to the consumption of milk

Most people, even those who were lactose maldigesters in a community with a high prevalence, did not ascribe any symptoms to the consumption of milk; nor was the type of symptom consistent when reported. Possible reasons include the absence of symptoms despite enzyme nonpersistence, little or no milk intake, lack of insight to relate milk intake causally with symptoms, and masking of symptoms when milk was taken only with meals. Possibly all these explanations contributed to the low prevalence of symptoms ascribed to the consumption of milk.

Symptoms due to lactose maldigestion in younger children

The symptom score could not always be calculated accurately. Although the meaning of diarrhoea is obvious both to the children and to the nursing staff, and the presence of abdominal pain is fairly easy to establish, some of the younger children could not be certain whether they had more flatulence than usual or whether bloating occurred. It is also possible that such small children could not discern the difference between bloating and abdominal pain or cramps.

Thus, in some cases only two symptoms were recorded. However, the exact types of symptoms or the exact severity may not be completely certain, but the presence or absence of the symptoms of lactose intolerance was, in our opinion, accurately noted.

After the MTT, 61.3% of subjects had no symptoms, compared with 41.9% after the LTT. In the group aged 10 to 16 years, 38.2% had no symptoms after the MTT. One of nine subjects (11.1%) had no symptoms after the LTT, but this sample was too small to draw accurate conclusions.

Symptoms, therefore, were more likely to occur in older than younger children, even though the frequency of lactose maldigestion may not differ greatly between the groups. This agrees with our observation that most of the Tswana people who are milk intolerant become aware of this in adolescence or early adulthood. The same phenomenon was also noted in Native Americans [15].


In the questionnaire, 78% of the pupils claimed that drinking fresh milk caused no symptoms. After a test meal, 38% said the same. However, half of the lactose digesters also claimed to have symptoms attributable to the test meal of milk.

Those who suffered bloating asserted that they would not drink milk again, whereas those who suffered cramps would continue to drink milk. Both the total symptom score (>3 out of a possible 12) and the severity of the bloating affected the subjects' decisions to reduce their future milk consumption.

Symptoms were less likely to occur in younger than in older children; 61% had no symptoms after a large milk load, compared with 38% of teenagers. Finally, symptoms ascribed to milk drinking are often reported for reasons other than lactose maldigestion.


I thank Mr. J. Mahuma, principal of the Moruleng Middle School in Saulspoort, for permission to carry out this project in his school; his staff members for their help; and the nurses in the paediatric wards of George Stegmann Hospital for their assistance in the study of the hospitalized children.


1. Commentary. Lactose milk intolerance and feeding problems (including statement of the Food and Nutrilion Board, National Research Council). J Am Diet Assoc 1972;61:241-2.

2. American Academy of Pediatrics, Committee on Nutrition. Should milk drinking of children be discouraged? Pediatrics 1974;53:576-82.

3. Kretchmer N. Lactose and lactase. Sci Am 1972;227(4): 70-8.

4. Simons FJ, Johnson JD, Kretchmer N. Perspective on milk-drinking and malabsorption of lactose. Pediatrics 1977;59:98-109.

5. De Villiers FPR. The prevalence of lactose maldigestion in the rural Tswana. S Aft J Food Sci Nutr 1990; 2:81-4.

6. De Villiers FPR. A standardized milk tolerance test. J Clin Gastroenterol 1987;9:320-3.

7. Stephenson LS, Latham MC. Lactose intolerance and milk consumption: the relation of tolerance to symptoms. Am J Clin Nutr 1974;27:296-303.

8. Barr RG, Levine MD, Watkins JB. Recurrent abdominal pain of childhood due to lactose intolerance. N Engl J Med 1979;300:1449-52.

9. Lebenthal E, Rossi TM, Nord KS, Branski D. Recurrent abdominal pain and lactose absorption in childhood. Pediatrics 1981;67:828-32.

10. Copi IM. Introduction to logic. 5th ed. New York: Macmillan, 1978.

11. De Villiers FPR. Milk and lactose intolerance in the rural Tswana. Doctoral thesis, University of the Witwatersrand, Johannesburg, 1990.

12. Leichter J. Comparison of whole milk and skim milk with aqueous lactose solution in lactose tolerance testing. Am J Clin Nutr 1973;26:393-6.

13. Newcomer AD, McGill DB, Thomas PJ, Hofmann AF. Tolerance to lactose among lactase deficient American Indians. Gastroenterology 1978;74:44-6.

14. Roggero P. Offredi ML, Mosea F. Perazzini M, Mangiaterra V, Ghislanzoni P. Marenghi L, Careddu P. Lactose absorption and malabsorption in healthy Italian children: Do the quantity of malabsorbed sugar and the small bowel transit time play roles in symptom production? J Pediatr Gastroenterol Nutr 1985; 4:82-6.

15. Newcomer AD, Thomas PJ, McGill DB, Hofmann AF. Lactase deficiency: a common genetic trait of the American Indian. Gastroenterology 1977;72:234-7.

Contents - Previous - Next