Table of Contents

HK J Paediatr (New Series)
Vol 2. No. 1, 1997

HK J Paediatr (New Series) 1997;2:29-34

Original Article

Breath Hydrogen Response to Lactose in Chinese Neonates

FHW Wong, CY Yeung, AYC Tam, HC Kwan


Breath hydrogen response to oral lactose was studied in 89 neonates of Chinese origin, 54 term and 35 preterm. End-expiratory air samples were collected with an automatic electronic device newly developed by the authors. Breath H2 was analyzed with Shimadzu gas chromatograph. Post-lactose rise in breath H2 of 20 ppm appears to distinguish two populations of infants, namely the lactose-absorbers with lower rises and the mal-absorbers with higher rises. Preterm infants showed much higher frequency of lactose mal-absorption as indicated by excessively high breath H2 response, a feature compatible with immature lactase development. Although given similar amount of lactose for the test, infants taking formula showed significantly lower breath H2 response than taking freshly prepared lactose solution. Slower gastric emptying with formula may be a reason for the difference. 97% of healthy term infants and 92% of healthy preterm neonates showed no evidence of lactose-malabsorption when fed the usual 7%-lactose-containing formula by the breath H2 response test.

Keyword : Breath hydrogen test; Chinese infants; Milk versus pure lactose; Neonatal lactose malabsorption

Abstract in Chinese

Nature provides lactose as the main form of carbohydrate in milk which has been acknowledged to be the most appropriate source of nutrition for the infant.1-10 Yet several studies have indicated that malabsorption of lactose is a common phenomenon in the human neonates. The clinical significance of these observations deserves further examination if 'Nature really knows best'.10 After all, infants have been feeding and thriving on lactose-containing food for thousands of years without much related health problems.

Lactose malabsorption in the milk feeding infants may be inferred from the finding of reducing substances in the faeces11-13 or lactose14 and reducing substances15 in the urine. This pattern is consistent with incomplete digestion of lactose. As the unabsorbed sugar in split by the intestinal bacterial flora into its component monosaccharides,16 hydrogen is generated which is absorbed and excreted by the lungs. This forms the basis of the breath H2 test for lactose mal-absorption.

Various methods have been developed to sample the expired air for breath H2 analysis as a non-invasive means to detect lactose-malabsorption.2,17-21 Most workers have obtained the end-expiratory samples manually.2,17-21 Such attempts in the rapidly breathing small infants are not only clumpsy but also difficult. We utilized an electronic automatic device recently developed by us22 for interval sampling of the end-expiratory air to examine the response of breath hydrogen production to oral lactose in the term and preterm neonates. This paper reports the results of the study. We want to examine the frequency and extent of excessive H2 production which has been utilized to indicate lactose mal-absorption in infants.

Materials and Methods

(A) Subjects studied

45 healthy Chinese term infants and 39 healthy preterm Chinese neonates of the maternity ward and the Special Care Baby Unit of the Queen Mary Hospital were studied. The demographic data of the infants were shown in Table I. Informed consent was obtained from all the parents before each study. All infants were on their routine formula feeding (S26, Wyeth) at 3-hourly intervals. None of them had been receiving antibiotics.

(B) Breath hydrogen test (BHT)

In order to obtain maximal cooperation from the mothers and to produce minimal interruption to the nursing procedures, none of the infants was subjected to prolonged fasting. The preprandial sample of the BHT was taken just before meal time i.e. 3-4 hours from the last meal. 45 term and 27 preterm infants were given a test meal of 7% freshly prepared aqueous lactose solution (lactose 1 g/kg). 37 of the term infants and 12 other preterm infants were also tested with their usual amount of formula milk (20± ml/K/feed). The actual amount consumed was measured to the nearest milli-litre (ml) and the lactose load calculated. Breath samples were taken at half-hourly intervals for 3 hours using an automatic end-tidal sampler which has been validated in a recent study.22 25 term infants had a 4-hour test.

(C) Analysis of sample

Samples were collected in water-sealed 2 ml-glass syringes and analyzed immediately in the laboratory close to the ward. Gas chromatography equipped with a thermal conductivity detector (Shimadzu GC8-APT) was used for determining the gaseous contents of the samples. It allowed a fast, interference-free and sensitive detection for H2 and CO2 concentration in duplicate 025-ml breath samples. The method was highly reproducible with coefficient of variation (CV) of < 2% for the gases studied.23

(D) Data analysis

The automatic end-tidal air sampling device offers a precise sampling method with reproducibility of more than 90%,22 as measured by intra-individual coefficient of variation of the concentrations of respiratory gases. When compared with the expected alveolar concentration of carbon dioxide of about 5%, the values of breath H2 obtained correspond to 72-100% of the end tidal air samples. The normalised concentration24,25 of breath H2 could thus be calculated by the following formula

the expected concentration
of end-tidal CO2
the observed concentration
of CO2 in samples
X observed concentrations
of H2 sample

In our laboratory, the CV for hydrogen analysis between 1 and 100 ppm is less than 2%. The arithmetic mean, standard deviation, coefficient of variation of grouped data, student's t-test and regression analysis were all computed as descriptive statistics. Where paired samples were compared, the Student's t-test for paired data was adopted.


A total of 921 end-expiratory air samples were collected from 45 term infants (45 on aqueous lactose and 37 also on milk) and 39 (27 on aqueous lactose and 12 on milk) preterm infants (Tables I & II). Following lactose ingestion (including 12 preterm neonates who were given formula), the maximum rises of normalised breath H2 in the 84 Chinese neonates are shown in Figure 1. Other results are shown as follows:

Table I Demographic Data of 45 Term and 39 Preterm Chinese Infants
Item Term Preterm
Number 45 39
Sex 18 : 27 22 : 17
Gestation (wk) 38 - 43 25 - 37
Body weight (kg) 2.5 - 3.95 1.17 - 3.55
Post-natal age (days) 4.89 ± 5.7 11.97 ± 13.45*
Developmental age 279.5 ± 9.08 257 ± 14.2+
* Range = 1.5 - 58; +Range = 202 - 297.5


Table II Lactose Malabsorption in Chinese Neonates
Infant Lactose Malabsorption in
Milk Aqueous Total
Preterm - Number 1 in 12 17 in 27 18 in 39
Frequency % 8.3 63 46.2
Term - Number 1 in 37 8 in 45 9 in 82
Frequency % 2.7 17.8 11.0
Lactose malabsorption = Breath H2 response > 20 ppm above basal
Note: much higher frequencies of malabsorption rate when tested with pure lactose solution compared with milk, although the actual amount of lactose contained in the milk consumed was higher.


Fig. 1 Breath H2 response to lactose in 84 neonates (NRBH in ppm) 45 term and 39 preterm neonates (all term and 17 preterm infants were given milk).

1. Breath-H2 response in the healthy term infants

82 test-results were obtained from 45 term infants (38-42 weeks). Nine (11%) showed an elevation of normalised maximal rise in breath H2 greater than 20 ppm above the pre-prandial base line (NBBH), suggestive of lactose malabsorption.16-21 When the type of lactose containing test-meals were analyzed separately, the frequency of suggestive lactose malabsorption was found to be 2.7% (1 in 37) for milk and 17.8% (8 in 45) for aqueous lactose solution (Table II).

The mean rise in normalised breath H2 (NRBH) above the basal in response to aqueous lactose administration was significantly higher than that of formula milk as shown in Table III. In addition, the time for the peak rise in breath H2 was also significantly affected by the type of test-meal. For formula-fed infants it was 136.5 ± 39.3 minutes, with a range of 165 minutes while aqueous lactose solution produced a smaller range of 150 minutes, with a mean peak time of 119.7 ± 34.7 minutes. The difference between these two groups was statistically significant (t=2.03, P<0.05).

Table III Breath H2 Response to Lactose in Infants Milk vs Aqueous Lactose
  Infant Mean Rise in NRBH2 (ppm) Time for Peak Rise (min)
Feed   Term Preterm Term Preterm
Milk 5.86 ± 8.15 8.1 ± 8.8 136.5 ± 39.3 131.3 ± 23.2
AQ Lactose 16.56 ± 17.09 25.8 ± 14.3 119.7 ± 34.7 128.3 ± 36.6
Number (M : L) ( 37 : 45 ) ( 12 : 27 ) ( 37 : 45 ) ( 12 : 27 )
t-value 3.72 4.72 2.03 0.30
p-value < 0.001 < 0.001 < 0.05 < 0.77
Lactose in Milk intake (M) = 1.44 ± 0.48 gm/kg/feed
Lactose in Aqueous Lactose (L) = 1.08 ± 0.29 gm/kg/feed
NRBH2 = Normalised rises in breath-H2
Note that milk (with same amount of lactose) produces much lower peak H2 and takes longer to reach the peak, when compared with aqueous lactose solution.

2. Breath H2 response in the healthy preterm infants

Table II shows that 46.2 % of the BHT results in the preterm neonates were > 20 ppm above the basal concentration indicative of lactose malabsorption. Taking the type of test-meals into consideration, the percentages of results inferring lactose malabsorption became 8.3% (1 in 12) for milk and 63.0% (17 in 27) for aqueous lactose respectively. The maximal rise in breath H2 was also significantly higher (Table III) with aqueous lactose. However, no significant difference was found on the time for the peak rise in the breath H2.

3. Effect of gestational age, developmental age and body weight

The postprandial maximal breath H2 responses of the 84 infants bear a negative linear relationship with gestational age, developmental age and body weight, with regression coefficients of -0.35796 (P=0.0253), -0.37747 (P=0.0178) and -0.48463 (P=0.001 8) respectively. Figures 2 & 3 are representative illustration with such correlations in the preterm neonates.

Fig. 2 Effect of developmental age on breath H2 response to lactose in 39 preterm infants
Developmental Age = gestation + post-natal age


Fig. 3 Effect of body weight on breath H2 response to lactose in 39 preterm infants

4. The preprandial breath H2

The mean preprandial breath H2 concentrations of the term and the preterm infants were 9.80 ± 4.8 ppm (range 1.41 to 26.32) and 16.20 ± 12.82 ppm, (ranging from 2.87 to 67.7) respectively. A linear correlation was also observed between the maximal postprandial rise in breath H2 (NBBH) and the preprandial breath H2 (NBBH) with a regression coefficient of 0.60480 (P<0.0001). Figure 4 shows the significant correlation between the normalised preprandial (basal) and the rise in breath H2 of all the infants tested: 121 tests in 45 term infants and 39 in preterm neonates.

45 term and 27 preterm infants were tested on aqueous lactose; 37 of the term infants and an additional 12 preterm infants were tested with formula-milk also.
Fig. 4 Basal vs rises in breath H2 after lactose in 121 infant-test


This study shows that many neonates, particularly the preterm infants may have raised breath H2 response to oral lactose suggestive of some degree of lactose malabsorption as reported suggested by earlier workers.11-14 Our findings however do not show as high a prevalence rate of excessive breath hydrogen production by term infants fed lactose-containing milk as shown by these earlier works.1-9,21 It appears that there might be a difference between the Oriental infants in our series and those reported in the Caucasians, and our findings agree with the results of a small study on Japanese infants.26

We were happy to find that over 97% of healthy term infants studied did not show malabsorption of the usual amount of lactose contained in the milk feed (20 ± ml/K/feed). Even in the premature neonates, only 8.3% showed excessive H2 production in the breath suggestive of malabsorption, the majority were apparently able to assimilate the usual lactose load in the milk feeds. Although our findings are different from those reported from Caucasian infants1-9,25,27 our results are compatible with nutritional sense; as all neonates normally thrive on milk which contains only lactose as the sole source of carbohydrate. It is difficult to comprehend if Nature produces lactose in milk only to be partially absorbed and to be wasted in the excreta.

As shown in our study (Table II), lactose is well 'absorbed' when given in the usual amount of 7% as a constituent component of milk (Wyeth formula). However, when it is given on its own as a 7% aqueous solution, 17.8% of term infants and 63% of preterm neonates showed high H2 breath response suggestive of its malabsorption. Apparently, other ingredients in milk may contribute to lactose digestion and assimilation. Lactose on its own however, is not well absorbed and appears not to be a suitable supplementary source of food for infant feeding.

One of the mechanisms for enhanced lactose absorption with milk could be related to a prolonged contact of lactase of the intestine with its substrate. This possibility is partially supported by our observation of a significantly longer transit time to reach the peak rise of breath H2 in the formula-fed infants (Table III). The fat and protein contents of milk could have slowed down the gastric emptying.28 Slow streams of small amount of lactose are presented to the intestine allowing for more contact time with the lactase for digestion; absorption of the sugars is thus more complete. Although the lactose load in the formula feeds (1.44 ± 0.48 gm/K/feed) was significantly larger than the aqueous solution (1.08 ± 0.29/K/feed), the maximal rise in breath H2 in both the term and preterm neonates (Table III) was significantly less in the formula taking infants. It would be interesting to study human milk to see if it might even be more superior to humanised formulae in promoting better digestion and absorption of lactose.

The much higher frequency of apparent lactose malabsorption in the preterm infants is related to the immature development of intestinal disaccharidases.3-9 In human, the disaccharidases are detected early in gestation. They increase rapidly after 20 weeks; and except for lactase, their activities reach adult level by 27-28 gestational weeks. Lactase however, lags behind sucrase and does not reach "normal" or term levels until 36 weeks of gestation. Our findings are in agreement with these observations; a significant negative correlation between the maximal NRBH and gestational age of the infants was demonstrated (r=-0.24083; P=0.0273). The decrease in breath H2 response with gestational age was apparently due to increasing lactase activities in the small intestines of the infants, as they become progressively more mature. 21 of the 25 infants with maximum NRBI-I greater than 20 ppm were below 39 weeks gestational age. We have also found that the developmental age (postnatal age + gestational age) of the preterm neonates bears a significantly stronger negative correlation with the NRBH. (Fig. 2). This finding suggests that the development of lactase in the prematurely born infants may be a function of their post-conceptual maturity, and not a post-natal adaptation.

The maximum NRBH was also negatively correlated to the body weight (r = -0.32357, P = 0.0027, Fig. 3). For infants with body weight less than 2.5 kg, 60% (15 out of 25) responded to oral lactose with H2 greater than 20 ppm. For those between 2.5 and 3.5 kg, the frequency decreased to 22.9% (11 out of 48), and above 3.5 kg body weight, none of the 11 infants demonstrated excessively high H2 excretion. Apparently, lactose malabsorption as inferred by high breath H2 response to lactose is a developmental phenomenon, prevalent mainly in the prematurely born and the low birth weight infants. There was also significantly more girls with high breath H2 responses to lactose. 22 of the 44 female infants demonstrated maximal breath H2 rises of>20 ppm (27 term and 17 preterm). Only 5 of the 40 male infants (18 term and 22 preterm) showed high responses (Table IV).

Table IV Prevalence of Neonates with NRBH > 20 PPM After Oral Lactose (1 gm/kg)
Neonates Total H2 > 20 ppm Prevalence
< 2.5 kg 25 15 60%
< 3.5 kg 48 11 22.9%
> 3.5 kg 11 0 0
Girls 44 22 50%
Boys 40 5 12.5%
Total 84 27 32.1%
Note the much increased frequency of lactose malabsorption in girls and the progressively decreasing rate with increasing birth weights.
NRBH = normalised rises in breath H2

Although we would not suggest using a single 3-hour post-prandial (or preprandial) breath H2 analysis to diagnose lactose malabsorption, our study has shown that this test is of some predictive value for the Breath H2 profile results in both the term and preterm neonates (Fig. 4). Figure 1 shows the histogram distribution of the values of the maximal NRBH in the infants studied. A visibly nice bell-shaped curve occupies the left third of the graph; and this intersects with a more widely spread and irregular shaped curve on the right. The cutoff point of the 2 curves around 20 ppm H2 which is the conventional appears 2 diagnostic level for lactose malabsorption. The population with over 20 ppm breath H2 responses corresponds to that of lactose malabsorbers and covers 30.9% of the infant-tests studied.

Our study has also suggested that malabsorption of pure lactose and increased breath H2 response is a normal event in many healthy infants, particularly in the premature (Fig 5). An elevated breath H2 concentration in an infant may be "physiologic".5-8 It cannot be taken as prima facie evidence of lactose intolerance, but is more likely indicative of a desirable adaptation of the gastro-intestinal tract to the establishment of the faecal flora in the newborn infant. Indeed as Maclean has commented earlier5 that the decision whether lactose is being tolerated by the infant is a clinical one, based mainly on symptoms of abdominal distention, stool output, and rate of weight gain. An elevated breath H2 concentration should not be the cause for a change in the dietary management so long as the infant is doing well. Conversely, reduced production of H2 in the lactose-malabsorbing infant may be a reflection of the failure of colonic microflora to adapt to the unabsorbed sugar.4 It may be necessary to examine the stools for sugar content and microbial counts while performing sugar mal-absorption studies in order to understand the underlying problem more fully.

Note a degree of fluctuations of H2 concentration between 1 1/2 - 2 1/2 hours. The dip at 2 hours was probably the result of passage of gas before breath sampling.
Fig. 5 A typical breath H2 profile of a preterm infant

Different frequencies of lactose malabsorption have been reported with different lactose-containing foods.9 Many reports have included the use of hypertonic lactose solution5,17,18,20 which might have caused osmotic effects, with intestinal hurry and increased amount of unabsorbed sugars reaching the colon. This could have resulted in the excessive breath H2 production besides inducing symptoms of colic or even diarrhoea in some infants. Even using 7% aqueous lactose, we have obtained totally different prevalence rate of malabsorption from formula milk. We suggest that breast milk or humanised formula, which is the usual lactose containing food, may be the more suitable dietary challenge to test for lactose malabsorption in infants in order to obtain clinically meaningful results. When aqueous lactose is used, an isotonic or hypotonic solution should be freshly made up for the test.

We cannot explain entirely the discrepancies in the low prevalence rate of abnormal breath H2 test results in our Chinese neonates compared with other reports on Caucasian infants. The difference could be ethnic or genetic;29 or it could be due to a difference in methodology used in the studies or a difference in intestinal microbial colonization.30 Indeed, the occurrence of genuine lactose malabsorption in the Chinese newborn infants is uncommon as shown in our studies, even though it is highly prevalent in older Chinese children23,31-33 and adults.34


1. Maclean WC, Fink BB, Schoeller DA, Wong W, Klein PD. Lactose assimilation by full-term infants relation of [13C] and H2 breath tests with fecal [13C] excretion. Pediatr Res 1983;17:629-33.

2. Perman JA, Barr RG, Watkins JB. Sucrose malabsorption in children a noninvasive diagnosis by interval breath hydrogen determination. J Pediatr 1978;93: 17-22.

3. Pennan JA, Waters LA, Heldt GP, Rosental E. Carbohydrate Absorption in Premature Infants. Gastroenterology 1979;76: 1216. (Abstr)

4. Grand RJ, Watkins JB, Torti FM. Development of the Human Gastrointestinal Tract A Review. Gastroenterology 1976;70:790-810.

5. MacLean WC, Fink BB, Beverly BF. Lactose malabsorption by premature infants magnitude and clinical significance. J Pediatr 1980;97:383-8.

6. Chiles C, Watkins JB, Barr R, Tsai PY, Goldmann DA. Lactose utilization in the newborn : role of colonic data. Pediatr Res 1979;13:365. (Abstr)

7. Lifschitz CH , O'Brien-Smith E, Garza C. Delayed complete functional lactase sufficiency in breast-fed infants. J Pediatr Gastroenterol Nutr 1983;2:478-82.

8. Barr RG, Hanley J, Kingsworth Patterson D, Woolridge J. Breath Hydrogen excretion in normal newborn infants in response to usual feeding pattern: evidence of "functional lactase insufficiency" beyond the first month of life. J Pediatr 1984;104:527-33.

9. Douwes AC, Oosterkamp RF, Fernandes J, Los T, Jongbloed AA. Sugar malabsorption in healthy neonates estimated by breath hydrogen. Arch Dis Child 1980;55:512-5.

10. Committee on Nutrition, American Academy of Pediatrics. Commentary on breast-feeding and infant formulas. Pediatrics 1976;57:278-85.

11. Davidson AGF, Mullinger M. Reducing substances in neonatal stools detected by clinitest. Pediatrics 1970;46:632-5.

12. Counahan R, Walker-Smith J. Stool and urinary sugars in normal neonates. Arch Dis Child 1976;51:519-20.

13. Whyte PK, Homer R, Pennock CA. Faecal excretion of oligosaccharides and other carbohydrates in normal neonates. Arch Dis Child 1978;53:914-5.

14. Bickel H. Mellituria: a paper chromatographic study. J Pediatr 1961 ;59:641-56.

15. Haworth JC, McCredie D. Chromatographic separation of reducing sugars in the urines of newborn babies. Arch Dis Child 1956;31:189-90.

16. Lindquist BC, Wrann L. Problems in analysis of faecal sugar. Arch Dis Child 1976;51:319-21.

17. Levitt MD. Production and excretion of hydrogen gas in man. N Engl J Med 1969;281:122-7.

18. Calloway DH, Murphy EL, Bauer D. Determination of lactose intolerance by breath analysis. Am J Dig Dis 1969;14:811-5.

19. Bond JH, Levitt MD. Use of pulmonary hydrogen measurements to quantitate carbohydrate absorption : Study of partially gastrectomizeal patients. J Clin Invest 1972;51:1219-25.

20. Grand Ostrander CR, Cohen RS, Hopper AO, et al. Breath hydrogen analysis : a review of they methodologies and clinical applications. J Pediatr Gastroenterol Nutr 1983;2:525-33.

21. Moore DJ, Fink Robb TA, Davidson GE Breath hydrogen response to milk containing lactose in colicky and non-colicky infants. J Pediatr 1988;113:979-84.

22. Yeung CY, Ma YP, Wong FHW, et al. An automated end-expiratory tidal sampling device for breath tests in small infants. Lancet 1991;337:90-3.

23. Wong FHW, Yeung CY. Breath hydrogen response to lactose in Chinese children. (In preparation)

24. Niu HC, Schoeller DA, Klein PD. Improved gas chromatographic quantitation of breath H2 by normalisation to respiratory CO2. J Lab Clin Med 1979;94:755-63.

25. Kien CL, Liechty EA, Myerbery DZ, Mullett M D. Effects in premature infants of normalising breath H2 concentration with CO2: increased H2 concentration and reduced interaliquot variation. J Pediatr Gastroenterol Nutr 1987;6:286-9.

26. Nose O, Lida Y, Kai H, Harada T, Ogawa M, Yabunchi H. Breath hydrogen test for detecting lactose malabsorption in infants and children. Arch Dis Child 1979;54:436-40.

27. Cheu HW, Brown DR. Breath hydrogen excretion in the premature neonate. Am J Dis Child 1990;144:197-202.

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

29. Yeung CY. Chinese neonates are different. Chap 89 In: Yu V, Tsang R, Feng ZK, Yeung CY, editors. Textbook of Neonatal Medicine: a Chinese perspective. Hong Kong University Press, 1996;877-84.

30. Stevenson DK, Shahin SM, Ostrander CR, et al. Breath hydrogen in preterm infants : correlation with changes in bacterial colonization of the gastro intestinal tract. J Pediatr 1982;101:607-10.

31. Chung MH, Hsu HY, Chen CJ, Lee CH, Hsu JY. Lactose malabsorption and small intestinal lactase in normal Chinese children. J Ped Gastroent Nutr 1987;6:369-72.

32. Huang SS, Bayliss TM. Milk and Lactose intolerance in healthy Orientals. Science 1968;160:83-4.

33. Chung MH, McGill DB. Lactase deficiency in Orientals. Gastroenterology 1968;54:225-6.

34. Sung JL, Shih PL. The jejunal dissacharidases activities and lactose intolerance of Chinese adults. Asian J Med 1972;8:149-51.


This web site is sponsored by Johnson & Johnson (HK) Ltd.
©2021 Hong Kong Journal of Paediatrics. All rights reserved. Developed and maintained by Medcom Ltd.