hypertonic. The amount of sugars in part of a 6 h urine sample has been analysed by gas liquid chromatography. To date we have studied 16 normals and 8 untreated coeliacs. The lactulose:mannitol excretion ratio in the normals ranged from 0 -008 to 0 -0 15, whereas in 7 of the 8 coeliacs it ranged from 0 - 03 to 0 - 67. However, 1 untreated coeliac had a normal excretion ratio of 0 - 009. We agree that sugar permeability ratios appear to be useful in screening for coeliac disease but it is too early to tell whether it is specific for the mucosal abnormality of coeliac disease and whether it will be valuable prospectively. We would encourage its use without cellobiose or hypertonic solutions and suggest that gas liquid chromatography is probably a more satisfactory way of measuring the quantities of sugars excreted in the urine.
University Department of Medicine, Bristol Royal Infirmary,
S. O. UKABAM B. T. COOPER
Bristol BS2 8HW
sets out the evidence that measurement of excretion ratios of inert sugars after absorption from the gut urinary is a useful method of screening adults for increased intestinal permeability. Since in adults a flat small-intestinal mucosa is almost pathognomonic of coeliac disease (CD),l it is not surprising that grossly increased intestinal permeability almost invariably correlated with CD. In children, however, severe small-intestinal mucosal damage 2 occurs not only in CD but also in other gastrointestinal conditions,2 including cow’s milk sensitive enteropathy, post-enteritis syndrome, and Giardia infections. Hence, evidence of abnormal intestinal permeability in children may provide an indication for biopsy but is not diagnostic of CD. Among the immunological procedures proposed for investigation of CD the value of screening serum for wheat protein antibodies has been underestimated, especially in children. We have developed a simple screening test for anti-gliadin antibodies which has so far given no false negative results in children with active biopsy-proven CD.3The test involves a modification of the established indirect immunofluorescence test for tissue reactive autoantibodies (e.g., antinuclear antibodies) widely used in clinical immunology laboratories,4and is based on our observation that wheat gliadin binds selectively in vitro to reticulin in cryostat sections of human or rat tissues;5 before treatment with patients’ sera the sections are incubated with a solution containing 100 pig/ml wheat gliadin.* In a survey of 192 children under investigation for gastrointestinal- disease3 we found serum anti-gliadin antibodies detectable in this way in all 32 with active CD. Of the remaining 160 with other conditions only 15% had anti-gliadin antibodies, and most of these were patients with proven transient gluten intolerance,cow’s milk sensitive enteropathy, or Crohn’s disease. In a further study we found that in children the presence of circulating gliadin antibodies of IgA class (usually accompanied by high titres of IgG antibody) correlated closely with a diagnosis of CD. A positive test for IgA antibody thus not only indicates a marked abnormality of the small-intestinal mucosa but also specifically implicates gluten. Measurement of intestinal permeability alone cannot do this. This study also showed that when gliadin antibodies are detected in gastrointestinal disease other than CD, they are invariably of IgG class only.
It is thus evident that the new immunofluorescence screening test have developed for anti-gliadin antibody is of real value in confirming a diagnosis of gluten sensitivity in children. It conforms to your criteria for the ideal screening test for enteropathy, being "safe, inexpensive, and acceptable to the patient; with no false negatives and a low false positive rate", and it requires only a single blood sample from the patient. However, neither this test nor measurement of sugar permeability eliminates the essential diagnostic role of small-intestinal biopsy in the diagnosis of childhood CD.
Bone and Joint Research Unit, London Hospital Medical College, London E1 2AD
D. J. UNSWORTH E. J. HOLBOROW G. D. JOHNSON
Queen Elizabeth’s Children’s Hospital,
*Full technical details
1. Booth CC. Definition of adult coeliac disease. In: Hekkens WThJM, Pena AS, eds. Proceedings of second international coeliac symposium. Leiden: Stenfert Kroese, 1974: 10-12. 2. Visakorpi JK. Definition of coeliac disease. In: Hekkens WThJM, Pena AS, eds. Proceedings of second international coeliac symposium. Leiden: Stenfert Kroese, 1974. 3. Unsworth DJ, Manuel PD, Walker-Smith JA, Campbell CA, Johnson GD, Holborow EJ. A new immunofluorescent blood test for gluten sensitivity. Arch Dis Childh (in
press). Johnson GD, Dorling J. Immunofluorescence and immunoperoxidase techniques. In: Thompson RA, ed. Techniques in clinical immunology. Oxford: Blackwell Scientific Publications, 1977: 85-115. 5. Unsworth DJ, Johnson GD, Haffenden G, Fry L, Holborow EJ. Binding of wheat gliadin in vitro to reticulin in normal and dermatitis herpetiformis skin. J Invest Dermatol (in press). 6. Walker-Smith JA. Transient gluten intolerance. Arch Dis Childh 1970; 45: 523-28 4.
SIR,-Your March 14 editorial comments upon the value of probe molecules in the investigation of intestinal absorption in coeliac disease. Among the probe molecules which have been used, we have found polyethyleneglycol (PEG) mixtures to have considerable advantages, especially when investigating functional disturbances which are not necessarily confined to the proximal part of the small bowel. PEG is chemically and biologically inert, is a common food additive, and can be quantified by the changes in refractive index which it produces. The degree of intestinal absorption of PEG is rapid for molecules of small molecular size, 1,2 and it is excreted unchanged in the urine. Like the St Thomas’ group3,4 we have used both a small and a large molecular probe-in our case, a PEG mixture of molecular weights 600 (PEG 600) and 4000 (PEG 4K). An important practical observation was that the commonly utilised urine collection period of 5 h proved to be inadequate for detecting changes in the handling of PEG 4K. Extending the period to 24 h resulted in an increase in our ability to detect abnormalities in the excretion of these larger molecules. In active coeliac disease we have found an abnormally low excretion of PEG 600 and an abnormally high excretion ofPEG4K, which is consistent with evidence obtained by other means. We have also found an abnormality that was confined to the excretion of the larger molecules in several other conditions, including six patients with food allergy who met the criteria discussed in your recent editorial5 and a further five patients with eczema who were not regarded as having food allergy at the time of the study. Our results will be published elsewhere.
Departments of Medicine and Chemical Pathology, Guy’s Hospital Medical School, London SE1 9RT
P. G. JACKSON R. W. R. BAKER M. H. LESSOF
TREATMENT OF MALIGNANT EFFUSIONS
SIR,-Your Jan. 24 editorial on malignant effusions discussed ways of inducing chemical pleurodesis and suggested that further randomised studies were needed to determine the most effective and best tolerated agent. We have recently reported a comparative study of intrapleural Corynebacterium parvum against mustine without complete 1. Chadwick VS, Phillips SF, Hofmann AF. Measurements of intestinal permeability using low molecular weight polyethylene glycols (PEG 400). I: Chemical analysis and biological properties of PEG 400. Gastroenterology 1977; 73: 241-47. 2. Chadwick VS, Phillips SF, Hofmann AF. Measurements of intestinal permeability using low molecular weight polyethylene glycols (PEG 400). II: Application to normal and abnormal permeability states in man and animals. Gastroenterology 1977; 73: 247-51. 3. Creamer B. Intestinal permeability as a screening test for mucosal damage. In: Hekkens WTJM, Pena AS, eds. Coeliac disease. Leiden: Stenfert Kroese, 1974. 348-49 4. Menzies IS, Laker MF, Pounder R, Bull J, Heyer S, Wheeler PG, Creamer B. Abnormal intestinal permeability to sugars in villous atrophy. Lancet 1979; ii. 1107-109. 5. Editorial. Food allergy and intolerance. Lancet 1980; ii: 1344-45.
727 in the management of recurrent malignant effusions.1 Intrapleural injection of C. parvum after pleural aspiration induced a satisfactory pleurodesis in seven of the nine patients surviving long enough (more than four weeks) for adequate assessment of response, although two patients required a second intrapleural injection. This response was superior to that seen in the controls treated with intrapleural mustine. C. parvum was also effective in three out of four patients deemed "mustine failures" because of recurrent effusions uncontrolled by intrapleural mustine. Subsequent experience with a further five patients has confirmed this high success rate. Intrapleural C. parvum was well tolerated and we believe that, although its mode of action is uncertain, it offers a promising new approach to the difficult problem of recurrent malignant effusions. The mean survival times in our small series were 6 months (C. parvum) and 3 months (mustine). Further randomised studies of the various forms of treatment now available for malignant pleural effusion should include comparison of survival times, though this may be difficult to assess if several forms of treatment have been given.
City Hospital, Edinburgh EH10 5SB
J. W. MILLAR A. M. HUNTER N. W. HORNE
SERRATIA MARCESCENS CONTAMINATED DISINFECTANTS
StR,-In their investigation of antibiotic sensitive Serratia infections complicating cardiopulmonary operations in 1978 Dr Ehrenkranz and his colleagues (Dec. 13, p. 1289) tested the
ability of S. marcescens to grow in four formulations of quaternary ammonium disinfectant, including one that we no longer market. The ’Hi-Tor’ referred to in the Lancet article carries Environmental Protection Agency registration number 303-94. In August, 1980, we introduced a revised formula (E.P.A. 303-91). Hi-Tor 303-94 contains a "quat" system ofn-alkyl(60% C14, 30% CI6’ 5% C18’ 5% CI2) dimethylbenzylammonium chloride 6.75% and n-alkyl (68% Ci2, 32% C14) dimethylethylbenzylammonium chloride 6-75%. Hi-Tor 303-91 contains didecyldimethylammonium chloride 7-5% and n-alkyl (Ct4 50%, C12 40%, C16 10%) dimethylbenzylammonium chloride 50%. The two formulations are completely different. The quat system in ’TOR’ (1 :64 dilution) (another Huntington product mentioned) is: n-alkyl (60% C14 30% C16, 5% Cl8’ 5% C12) dimethylbenzylammonium chloride 1 - 607o and n-alkyl (50% C12 30% C14 17% C16) 3% CIs) dimethylethylbenzylammonium chloride I6%, and the parts per million (p.p.m.) of quat in TOR required to kill the following organisms are: Pseudomonas aeruginosa 500, Staphylococcus aureus 400, and Salmonella choleraesuis 400. The quat system in Hi-Tor 303-94 (1 :256 dilution) is shown above and the p.p.m. required to kill are: Ps. aeruginosa 527, Staph. aureus 450, and S. choleraesuis 450. The quat systems for Hi-Tor 303-94 and TOR are much the same, the first quats being identical and the second identical in chemical nomenclature though differing in carbon chain distribution. In our experience of microbiology testing we have found that both TOR and Hi-Tor 303-94 disinfectants take approximately the same p.p.m.s consistently to kill certain test organisms, as indicated above. If I received a microbiology test report with the results as given in the Dec. 13 Lancet article I would ask for the tests to be done again. In use-dilution Hi-Tor 303-94 and TOR are too close to being identical, in regards to the quat system, to have the contrasting results described. The test procedure described by Ehrenkranz et al. consisted of taking 0 -05-0 -1 ml as an inoculum for 5-10 ml of disinfectant. We take that to mean that a certain number of organisms per ml was mixedwithup to 10 ml of disinfectant. The test procedure described not official or recognised, and if it is a modification of the 1 Millar JW, Hunter AM, Horne NW. Intrapleural immunotherapy with Corynebacterium parvum in recurrent malignant effusions. Thorax 1980; 35: 856-58
A.O.A.C. phenol coefficient test, this test is no longer recognised by the Environmental Protection Agency in that it does not simulate in-use conditions. E.P.A. requires the A.O.A.C. use-dilution confirmation test which does simulate in-use conditions. In other words, a disinfectant is designed to disinfect on hard non-porous surfaces. In the A.O.A.C. use-dilution confirmation test a hard non-porous surface is inoculated with the test organism (107- 109/ml) and then exposed to the disinfectant solution. Then we check for growth or no growth, and E.P.A. requires complete killing to pass the test. Once we have E.P.A. approval we can market that particular disinfectant with the disinfectant claims that we have established. Thus Ehrenkranz’s paper is misleading because the test procedure is not an official one while our test data are generated by official A.O.A.C. test methods required by the E.P.A. Even so, the test results are not consistent: TOR killed S. marcesens in all instances and Hi-Tor 303-94 killed S. marcescens only occasionally but killed the same organism along with three additional organisms in all instances in other hospitals. Furthermore Hi-Tor 303-94 and TOR are much the same in respect of "quat" chemical composition and p.p.m. of quat in the use-dilution. We would question not only the procedure but also the technique. Ehrenkranz et al. state that "The danger of using quaternary ammonium compounds as disinfectants rather than cleansers is reemphasised," and they refer to an article written by Center for Disease Control staff on Use of Quaternary Ammonium Compounds. The article consistently and specifically refers to benzylalkonium chlorides and concludes that this type of disinfectant should be used in non-critical areas of the hospital. We no longer use benzylalkonium chlorides in our disinfectant products; we use ammonium chlorides. Both types of chemicals are referred to as quats-therein lies the confusion. Benzylalkonium chlorides are not effective against some hospital pathogens, including Ps. aeruginosa. Ammonium chloride type quats are effective against hospital pathogens at 500 p.p.m. or less, including Ps. aeruginosa. The C.D.C. article on benzylalkonium chlorides seems out of date. The technology of quats for disinfection has improved with the advent of ammonium chlorides. -
Huntington Laboratories Inc., Huntington, Indiana 46750, U.S.A.
RICHARD D. SHEETS
TRILOSTANE INTERFERENCE WITH STEROID ASSAYS
SIR,-Like Professor Mattingly and Christine Tyler (March 7, p. 561) we have measured steroid hormone levels in serum and urine from patients being treated for Cushing’s syndrome with the new adrenal blocking agent trilostane and whilst we agree that the interference of trilostane with the fluorimetric assay for serum 11-hydroxycorticosteroids ("cortisol") may be eliminated by the inclusion of an alkali wash before the addition of fluorescence reagent we have evidence that interference with the fluorimetric assay for urinary free "cortisol" persists despite the alkali wash stage. Furthermore, patients being treated with trilostane show unexpected rises in serum levels of testosterone and androstenedione which may also result from interference of the drug with the assays. We have studied five patients (three male, two female), with Cushing’s syndrome taking trilostane in doses ranging from 240 to1 1080 mg per day. Serum "cortisol" results derived by fluorimetry, using the modification suggested by Mattingly, were lower than the results obtained by a radioimmunoassay (RIA) known not to crossreact with trilostane (’Immophase’; Corning Medical) and there was a reasonable correlation between methods: modified fluorimetry 0-71 RIA + 46(nmol/1); r = 0-75, n 28. Addition of trilostane to serum in concentrations up to 100 mol/1 did not alter the "cortisol" results obtained by either RIA or the modified fluorimetric technique. However, fluorimetric urinary free "cortisol" results were appreciably higher than the corresponding RIA results in the patients taking trilostane and there was a poor correlation between methods: =
1. Mattingly D. A simple fluorimetric method for the estimation 11-hydroxycorticoids in human plasma. J Clin Path 1962; 15: 374.