BREATH HYDROGEN IN HYPOSUCRASIA

119 rate among infants less than 2500 g decreased slightly between 1956-59 and 1971-74 but the incidence among necropsies doubled owing to reduction in the size of the necropsy population (table v). The death-rate from un-

complicated

H.M.D. or I.V.H.

fell and for H.M.D./I.V.H.

rose.

BREATH HYDROGEN IN HYPOSUCRASIA GEOFFREY METZ ALVIN NEWMAN*

DAVID J. A. JENKINS LAURENCE M. BLENDIS

Department of Gastroenterology and Medical Research Council Gastroenterology Unit, Central Middlesex Hospital, London NW10

Discussion

two-year periods with a four-year gap between, Simmons et al. found a lower incidence of intracranial haemorrhage in the second when sodium bicarbonate therapy was restricted. In contrast we did not find a lower incidence of I.V.H. among preterm infants in the pre-sodium bicarbonate era, nor did cases developing i.v.H. appear to receive more sodium bicarbonate prior to the onset of the I. v .H. The total quantity of sodium bicarbonate given to infants who developed I.V.H. was greater than that given to those with H.M.D. alone, but it is possible that much of this was given after intracranial haemorrhages had occurred and because of the clinical state these haemorrhages engendered. This and other factors’ 8 may have accounted for the difference between Simmons’ two groups of cases. Our H.M.D./I.v.H. infants were more ill than those with H.M.D. alone. Respiratory therapy was invariably necessary, hypothermia was commoner, and pulmonary lymphatic dilatation (a characteristic of H.M.D.)9 greater. More bicarbonate may have been given in con-

During

two

A simple and reliable test for the diagnosis of hyposucrasia is required, since this may be an unsuspected cause of long-standing gastrointestinal disorder. Furthermore little has been done to define the epidemiology of this condition, possibly because of the limitations of multiple blood-sampling. Breath hydrogen (H2) production after lactose ingestion is a reliable test for hypolactasia, and has now been measured after sucrose ingestion in eleven patients with various gastrointestinal symptoms. Six who had normal sucrase activity on jejunal biopsy produced no H2 after taking 50 g of sucrose. No H2 was produced in three patients with borderline hyposucrasia, either after 50 g sucrose or when retested using 100 g sucrose (two patients). However, the two patients with low jejunal sucrase activity showed rises of breath H2 after only 25 g sucrose. Breath H2 measurement is a simple, accurate,

Summary

and non-invasive test for diagnosing symptoms due to hyposucrasia. Introduction

sequence.

Simmons et al. found a correlation between sodiumbicarbonate dosage and hypernatraemia, and, on the basis of the evidence that serum hyperosmolarity may induce cerebral haemorrhage,lO proposed that sodium-bicarbonate-induced hypernatraemia might cause cerebral haemorrhage. Although we found a correlation between maximum serum-sodium concentration and the lifetime dose of sodium bicarbonate, hypernatraemia was uncommon in our series and mean serum-sodium higher in H.M.D. than H.M.D./I.V.H. infants. An alternative factor promoting bleeding might be hypervolaemia. Cole et al."I have recently argued in favour of venous hypertension secondary to cardiac failure as the most important factor in the pathogenesis of i.v.H. While alkali treatment may increase pulmonary perfusion’2 and lessen cardiac embarrassment, the injected fluid could significantly swell the venous compartment." The association of breech delivery with I.V.H. 14 might also be explained by venous hypertension resulting from compression of the thorax at birth. No specific relationship was found between coagulation disorders and I.V.H. 14 11 but all infants with I.V.H. had some abnormality of hmmostasis. H.M.D. infants who develop i.v.H. form a severely ill group with little chance of survival and might be expected to receive larger doses of alkali buffer thera-

gastrointestinal

deficiency, in the form of hypolactaof gastrointestinal symptoms.’ In contrast, hyposucrasia is rare in most parts of the world although it is present in more than 10% of Greenland Eskimos.2 Although the diagnosis is considered in inDISACCHARIDASE

sia, is

a common cause

fants with diarrhoea and failure to thrive,3 it is seldom suspected in patients with intractable gastrointestinal symptoms who may suffer from flatulence, diarrhoea, and weight loss for several years before being diagnosed.4S A safe, simple, yet accurate screening test would therefore be of value in patients with gastrointestinal symptoms of uncertain aetiology. Hydrogen (H2) is evolved when carbohydrate is fermented by bacteria.66 Thus, when malabsorbed disaccharide reaches the carcum it is fermented with production of H2.A proportion of the gas diffuses into the circulation to be carried to the lungs and exhaled.8 Breath H2, collected by end-expiratory sampling,9 has been found to be a simple and accurate method for diagnosing hypolactasia.1O We have measured end-expiratory breath H2 after an oral sucrose load in patients known to have or suspected of having hyposucrasia. "Present address: New Mount Sinai Hospital, Toronto, Ontario, Canada.

peutically. Requests for reprints

should be addressed

to

F.

C., Simpson

Memorial Maternity Pavilion, Royal Infirmary, Edinburgh EH3 9EF. REFERENCES

1. Machin, G. A. J. clin. Path. 1975, 28, 428. 2. Gruenwald, P. Am. J. Obstet. Gynec. 1951, 61, 1285. 3 Larroche, J-C. Biol. Neonat. 1964, 7, 26. 4. Towbin, A. Am. J. Path. 1968, 52, 121. 5 Leech, R. W., Kohnen, P. Am. J. Path. 1974, 77, 465.

6. Simmons, M. A., Adcock, E. W., Bard, H., Battaglia, F. C. New Engl. J. Med. 1974, 291, 6. 7. Volpe, J. ibid. p. 43. 8. Gilles, F. H., Leviton, A., Dooling, E. C. ibid. p. 1088. 9. Lauweryns, J. M. Pediatrics, Springfield, 1969, 44, 126. 10. Luttrell, C. N., Finberg, L., Drawdy, L. P. Archs Neurol., Chicago, 1959,

1, 153. 11.

12. 13. 14. 15.

Cole, V. A., Durbin, G. M., Olafson, A., Reynolds, E. O. R., Rivers, R. P. A., Smith, F. F. Archs Dis. Childh. 1974, 49, 722. Gupta, J. M., Dahlenburg, G. W., Davis, J. A. ibid. 1967, 42, 416. Baum, J. D., Roberton, N. R. C. J. Pediat. 1975, 87, 255. Chessels, J. M., Wigglesworth, J. S. Archs Dis. Childh. 1972, 47, 564. Hambleton, G., Appleyard, W. J. ibid. 1973, 48, 31.

120 Discussion

Patients and Methods Two patients previousl) diagnosed as having hyposucrasia the basis of very low jejunal disaccharidase activity and three patients with diarrhoea thought possibly to be due to hyposucrasia and with sucrase activity at or just below the lower limit of normal were studied. Six patients with gastrointestinal symptoms requiring jejunal biopsy, but who had a normal sucrase activity on assay were used as controls. After a 9 h overnight fast all patients had an oral sucrose-tolerance test (S.T.T.). This consisted of ingestion of 25 to 100 g sucrose loads dissolved in 200 ml of water. The known hyposucrasia patients were given only 25 g sucrose loads to see if breath hydrogen could be detected from the lower dose of sugar without causing the diarrhoea usually associated with SO g loads in sucrose-intolerant patients. The three suspected hyposucrasic and six control patients had SO g sucrose loads; two of the three possible hyposucrasic patients later had 100 g S.T.T.S.

on

,

Any gastrointestinal symptoms such as borborygmi, excessive flatus, pain, or diarrhoea were noted. Samples of end-expiratory alveolar air from a single breath using a modified Haldane-Priestley tube were collected every 30 min, and the Hz concentration was measured by gas chromatography.9 The results were expressed as rise in breath H2 concentration above the fasting level and 20 parts/106 (p.p.m.) was taken as abnormal. Disaccharidase activities were assayed independently by a modification of the techniques of Dahlqvist.11 12 Patients failing to produce H2 during the sucrose test were given a 2S g lactulose tolerance test to ensure that their colonic flora was capable of producing H2 from malabsorbed disaccharide. 13

Results None of the six controls developed symptoms or produced H2 for the 3 h after oral sucrose (table). Two of the three patients with borderline hyposucrasia (G and J) produced no H2 after infesting 50 g of sucrose, but both wondered if there may have been an increase in borborygmi and flatus during the test (table). Therefore they were both subsequently given 100 g of sucrose which produced neither symptoms nor H2 in the breath in either patient. Both patients with proven hyposucrasia had symptoms after the test (table). In patient K, H2 first appeared in the breath at 60 min, in a concentration of 4 p.p.m., rising to 62 p.p.m. at 120 min, whilst in patient L, H2 first appeared in the breath at 90 min at a concentration of 9 p.p.m. rising to 72 p.p.m. at 180 min.

Hyposucrasia, although far less common than hypolactasia in the U.K., should be considered in the differential diagnosis of patients with gastrointestinal symptoms of unknown origin. On the other hand, hyposucrasia is much more common in Greenland Eskimos,’ and this may be true of other ethnic groups. In epidemiological studies a simple non-invasive screening test would be useful.

Although patients

Hl

hyposucrasia

may

give

a

his-

meal, gives a very good diagnostic yield in hyposucrasia. 15 However, this requires X-ray facilities and involves irradiation of the abdomen, and in the presence of hyposucrasia the barium is diluted by the osmotic effect of the sucrose which may prevent an adequate demonstration of the anatomy of the small intestine." The most conclusive diagnosis is likely to be obtained by measuring sucrase activity in the jejunum, an invasive technique requiring intubation, jejunal biopsy, and highly specialised biochemical analyses. Breath H2 is a simple and accurate method for the diagnosis of hypolactasia. It has not been previously tested in the diagnosis of hyposucrasia, and in this study it has diagnosed both cases of known hyposucrasia, clarified the status of three patients with borderline enzyme estimations, and given no false-positive results in six patients known to have normal enzyme levels. It would therefore seem to be a simple, accurate, and noninvasive test in the diagnosis of hyposucrasia. We thank Dr N.

RESULTS OF BREATH

with

tory of intolerance to sweets, they frequently do not. Furthermore, sugar often aggravates diarrhoea so that a sucrose-free diet may be tried without objective testing with non-specific benefit and thus delaying the correct diagnosis. In this study symptoms during an S.T.!. clearly did not correlate with H2 production, a finding previously noted in hypolactasia." The monitoring of the blood-glucose rise after a 50 g S.T.T. is commonly used for the diagnosis of hyposucrasia. In a previous study only half the patients with blood-glucose rises of less than 25 mg/dl developed diarrhoea.2 This suggests that several may have had falsepositive low blood-glucose rises, as often happens when a low blood-glucose rise is used as the criterion of abnormality in hypolactasia. Another non-invasive technique, the sucrose/barium

Coghill,

West Middlesex

Hospital, and

Dr G.

Neale, Royal Postgraduate Medical School for allowing us to study their patients, and Dr T. J. Peters, Royal Postgraduate Medical School, whose laboratory did the enzyme analyses. G. M. is in receipt

TESTS

of the T. K. Stubbins Fellowship of the Royal College of Physicians. Requests for reprints should be addressed to L. M. B., Gasuw terology Unit, Central Middlesex Hospital, London NW10 7NS. REFERENCES

M., Rothfield, B., Massa, C., Wise, L., Paige, D., Bedine, M. S Med. 1975, 292, 1156. 2. McNair, A., Gudman-Hoyer, E., Jarnum, S., Orrild L. Br. med. J. 1972,ii 3. Anderson, C. M., Messer, M., Townley, R. R. W. Pediatrics, 1963, 31, 1003 4. Neale, G., Clark, M., Levin, B. Br. med. J. 1965, ii, 1223. 5. Pink, I. J. Gut, 1967, 8, 373. 6. Calloway, D. H., Colasito, D. J., Mathews, R. D. Nature, 1966, 212, 1238 7. Levitt, M. D. New Engl. J. Med. 1969, 281, 122. 8. Levitt, M. D., Ingelfinger, F. J. Ann. N. Y. Acad. Sci. 1968, 150, 75. 9. Metz, G. L., Gassull, M. A., Leeds, A. R., Blendis, L. M., Jenkins, D. J. A. Clin. Sci. molec. Med. (in the press). 10. Metz, G. L., Jenkins, D. J. A., Peters, T. J., Newman, A., Blendis, L. M Lancet, 1975, i, 1155. 11. Guilbault, G. G. Analyt. Chem. 1966, 38, 527. 12. Dahlqvist, A. Enzym. biol. clin. 1970, 2, 52. 13. Bond, J. H. Jr., Levitt, M. D. J. Lab. clin. Med. 1975, 85, 546. 14. Isokoski, M., Jussila, J., Sarna, S. Gastroenterology, 1972, 62, 28. 15. Laws, J. W., Neale, G. Lancet, 1966, ii, 139. 16. Laws, J. W., Spencer, J., Neale, G. Br. J. Radiol. 1967, 40, 594. 1.

Bayliss, New

-

*

---L-

I

I

I

I

mild abdominal symptoms However, neither had any symptoms with the 100 g load.

These

two

patients suspected

I

during

the 50 g

test.

T.

Engl. J.