IgG Food Intolerance Support Papers

Genesis Diagnostics Ltd, Eden Research Park, 8 Henry Crabb Road, Littleport, CB61SE

[email protected]

Tel: +44(0)1353 862220

Fax: +44(0)1353 863330

Genesis Diagnostics Ltd is a subsidiary of Omega Diagnostics Group PLC February 2009

Paper 01 Dietary advice based on food specific IgG Results Geoffrey Hardman, Gillian Hart, University of York, Heslington, York, UK Nutrition and food science Vol 37 No 1 2007 pp 16-23

Paper 02 Food elimination based on IgG antibodies in irritable bowel syndrome: a randomised controlled trial. W Atkinson, T A Sheldon, N Shaath, PJ Whorwell Gut 2004:53 1459-1464 doi: 10.1136

Paper 05 Food allergy in irritable bowel syndrome: new facts and old fallacies. E Isolauri, S.Rautava, M.Kalliomaki Gut 2004; 53 1391-1393 10.1136

Paper 06 A prospective audit of Food Intolerance among Migraine patients in primary care clinical practise. Trevor Rees, David Watson, Susan Lipscombe, Helen Speight, Peter Cousins, Geoffrey Hardman and Andrew J. Dowson. Headache care Vol.2 No 2 2005 105-110

Paper 07 Celiac Disease. Peter H.R Green M.D. and Christopher collier, M.D. PhD The New England Journal of Medicine 2007; 357:1731-43

Paper 08 Alterations of food antigen-specific serum immunoglobulins G and E in patients with irritable bowel syndrome and functional dyspepsia. X.L.Zuo, Y.Q. Li, W.J.Li, Y.T. Guo, X.F. Lu, J.M. Li and P.V. Desmond Clinical and Experimental Allergy, 37, 823-830

Paper 10 IgG Antibodies against Food Antigens are Correlated with Inflammation and Intima Media Thickness in Obese Juveniles M. Wilders-Truschnig, H.Mangge, C.Lieners, H.J.Gruber, C Mayer, W Marz Exp Clin Endocrinol Diabetes 2008; 116:241-245

Paper 12 A Vegan diet free of gluten improves the signs and symptoms of Rheumatoid Arthritis.. I Hafstöm, B.Ringertz, A. Spångberg, L. Von Zweigbergk, S. Brannemark, I. Nylander, J.Rönnelid, L.Laasonen, L.Klareskog. British Society of Rheumatology, 2001 pp 1175-1179

Paper 13 The gut-joint axis: cross reactive food antibodies in rheumatoid arthritis. M Hvatum, L Kanerud, R Hällgren, P Brandtzaeg Gut 2006:55 1240-1247 originally published online 16 feb 2006

Paper 14 Toward an understanding of Allergy and In-Vitro Testing By Mary James N.D Great Smokies Diagnostic Laboratory

1459

IRRITABLE BOWEL SYNDROME Food elimination based on IgG antibodies in irritable bowel syndrome: a randomised controlled trial W Atkinson, T A Sheldon, N Shaath, P J Whorwell ......

Gut 2004;53:1459–1464. doi: 10.1136/gut.2003.037697

Background: Patients with irritable bowel syndrome (IBS) often feel they have some form of dietary intolerance and frequently try exclusion diets. Tests attempting to predict food sensitivity in IBS have been disappointing but none has utilised IgG antibodies. Aims: To assess the therapeutic potential of dietary elimination based on the presence of IgG antibodies to food. See end of article for Patients: A total of 150 outpatients with IBS were randomised to receive, for three months, either a diet authors’ affiliations ...... excluding all foods to which they had raised IgG antibodies (enzyme linked immunosorbant assay test) or a sham diet excluding the same number of foods but not those to which they had antibodies. Correspondence to: Methods: Primary outcome measures were change in IBS symptom severity and global rating scores. Non- Dr P J Whorwell, Department of Medicine, colonic symptomatology, quality of life, and anxiety/depression were secondary outcomes. Intention to University Hospital of treat analysis was undertaken using a generalised linear model. South Manchester, Results: After 12 weeks, the true diet resulted in a 10% greater reduction in symptom score than the sham Manchester M20 2LR, UK; peter.whorwell@ diet (mean difference 39 (95% confidence intervals (CI) 5–72); p = 0.024) with this value increasing to smuht.nwest.nhs.uk 26% in fully compliant patients (difference 98 (95% CI 52–144); p,0.001). Global rating also significantly improved in the true diet group as a whole (p = 0.048, NNT = 9) and even more in compliant patients Revised version received (p = 0.006, NNT = 2.5). All other outcomes showed trends favouring the true diet. Relaxing the diet led to a 13 April 2004 Accepted for publication 24% greater deterioration in symptoms in those on the true diet (difference 52 (95% CI 18–88); p = 0.003). 13 April 2004 Conclusion: Food elimination based on IgG antibodies may be effective in reducing IBS symptoms and is ...... worthy of further biomedical research.

rritable bowel syndrome (IBS) is a common disorder which physiological15–17 especially as IgG food antibodies can be causes abdominal pain, abdominal distension, and bowel present in apparently healthy individuals.18–20 It has pre- dysfunction, characterised by loose bowels, constipation, or viously been suggested that IgG food antibodies may have a I 1 21 a fluctuation between these two extremes. This condition role in IBS and it was therefore the purpose of this study to significantly impairs quality of life and places a large burden formally evaluate, in a randomised controlled trial, the on health care resources.2 Treatment of IBS is largely based therapeutic potential of an elimination diet based on the on the use of antispasmodics, antidepressants, and medica- presence of IgG antibodies to food in patients with IBS. tions that modify bowel habit, depending on whether constipation or diarrhoea is the predominant problem.1 The PATIENTS AND METHODS notorious inadequacies of current drug therapy lead to much Patients patient dissatisfaction and a tendency for patients to seek a All patients with uncomplicated IBS (all bowel habit variety of alternative remedies, especially of a dietary nature. subtypes) attending the Gastroenterology Department at IBS is likely to be a multifactorial condition involving a the University Hospital of South Manchester were considered number of different mechanisms although the prominence of eligible for the study, and those aged between 18 and any particular factor may vary from patient to patient.13 75 years, who satisfied the Rome II criteria,22 were invited to However, patients often strongly believe that dietary intoler- participate. Tertiary care patients were excluded from the ance significantly contributes to their symptomatology and study. All patients had normal haematology, biochemistry, some sufferers seem to benefit from eliminating certain foods and endoscopic examination when indicated. Coeliac disease from their diet. Detection of food intolerance is often difficult was excluded using the tissue transglutaminase test and a due to its uncertain aetiology, non-specific symptomatology, hydrogen breath test was used for excluding lactose intoler- and relative inaccessibility of the affected organ. Thus most ance. Patients were also excluded from participating in the previous studies have relied on the use of exclusion diets, study if they had any significant coexisting disease or a which are extremely labour intensive and time consuming.45 history of gastrointestinal surgery, excluding appendicect- Attempts to ‘‘test’’ for food intolerance in IBS have largely omy, cholecystectomy, and hiatus hernia repair. The study focused on ‘‘classic’’ food allergy based on the presence of IgE was approved by the local ethics committee and all patients mediated antibody responses, although it appears that these provided written informed consent. ‘‘immediate type’’ reactions are probably quite rare in this 6–10 condition. It is therefore possible that adverse reactions to Methods food in patients with IBS might be due to some other form of The study used a double blind, randomised, controlled, immunological mechanism, rather than dietary allergy. Such parallel design in which patients were randomised to either a reactions could be mediated by IgG antibodies, which ‘‘true’’ diet or a ‘‘sham’’ diet control group. At screening, characteristically give a more delayed response following 11 exposure to a particular antigen and have been implicated Abbreviations: IBS, irritable bowel syndrome; ELISA, enzyme linked 12–14 in some cases of food hypersensitivity. However, this immunosorbant assay; AU, arbitrary unit; HAD, hospital anxiety and mechanism is controversial and is considered by some to be depression scale; QOL, quality of life; NNT, number needed to treat

www.gutjnl.com 1460 Atkinson, Sheldon, Shaath, et al blood was taken and sent, with only a numerical identifier, to slightly worse, no change, slightly better, better, or excel- YorkTest Laboratories Ltd (York, UK) where an enzyme lent?’’ The atopic status of all patients entering the study was linked immunosorbant assay (ELISA) test was performed to also assessed. detect the presence of IgG antibodies specific to a panel of 29 During the treatment phase, patients were allowed to take different food antigens. This test has been described in detail concomitant medication provided it had been constant for six elsewhere23 and involves specimens being diluted 1/50, 1/150, months prior to the start of the study. They were encouraged and 1/450 with each dilution applied to an allergen panel. not to alter medication use during the course of the trial but Each test was calibrated using 0 arbitrary unit (AU) and any changes were recorded. Any patient withdrawing from 25 AU standards prepared from a serum with a high IgG titre the study was encouraged to complete a final symptom to a cow’s milk allergen extract. A positive control serum at questionnaire at week 12 and their reasons for withdrawal 45 AU was applied to each test. The test results were obtained were recorded. At the end of 12 weeks, patients were asked to from the 1/150 dilution of the specimen. Where a high resume consumption of the foods they had been advised to specimen background was observed, the test results were eliminate in order to assess the effect of their reintroduction. obtained from the 1/450 dilution. The threshold for a positive Patients were then reassessed after four weeks using the (reactive) result was selected as three times the background same measures and the result compared with their scores at signal obtained by the same sample against a no food the end of the elimination phase. allergen coated control well equivalent to 3 AU. Test results were scored as positive or negative only, relative to this cut Data analysis off. Questionnaires were scored by an assessor blinded to the Staff based at the YorkTest Laboratories produced a true randomisation. The primary outcome measures were changes and sham diet sheet for each patient. The sham diet in IBS symptom severity score and global impact score at eliminated the same number of foods to which a patient 12 weeks. Changes in non-colonic symptoms, QOL, and HAD exhibited IgG antibodies but not those particular foods. The scores were regarded as secondary outcome measures. Two goal was to try and include in the sham diet an equally sample t tests were used to establish whether there was an difficult to eliminate staple food for every staple food in the overall difference in the change in continuous outcome true diet. Thus cow’s milk was (generally) replaced with measures between the two groups of patients. Patients were potato, wheat with rice, and yeast with whole egg, where this analysed according to the group to which they were was possible. Nut reactivities were replaced with other nuts randomised, independent of their adherence to the diet. in the sham diet, and legumes with other legumes, but this The global impact score, an ordered categorical variable, was was not systematised. analysed using a Wilcoxon Mann-Whitney test to compare The true and sham diet sheets for each patient were sent to the numbers in the active and sham groups showing the University of York, again with only a number for significant improvement (‘‘better’’ or ‘‘excellent’’), no sig- identification. Patients were allocated to one of the two diet nificant change (‘‘slightly worse’’, ‘‘no change’’, or ‘‘slightly sheets based on a randomisation schedule developed using a better’’), and significant deterioration (‘‘worse’’ or ‘‘terri- random computer number generator. Thus patients would ble’’). The number needed to treat (NNT) was calculated receive either an elimination diet based on their true from the global impact score by calculating the reciprocal of sensitivity results (true diet) or a sham diet. All patients the difference in probability of a significant improvement and clinical staff in the Gastroenterology Research between the treatment and control groups. General linear Department and YorkTest Laboratory were blinded to the modelling in SPSS was used to explore whether there was a group assignment of all patients for the duration of the study. Patients were given their allocated diet sheet by staff at the Gastroenterology Research Department and asked to elim- Assessed for inate the indicated foods from their diet for a period of eligibility (n=176) 12 weeks. They also received a booklet with advice on Excluded (n=26): eliminating the different foods and the telephone contact Did not meet inclusion details of a free nutritional advisor whom they were able to criteria (n=19) Refused to participate (n=5) contact for further advice if necessary. Randomised Other reasons (n=2) Symptoms were assessed using a questionnaire scoring (n=150) system validated for use in IBS, including the IBS symptom severity score (range 0–500).24 This is a system for scoring pain, distension, bowel dysfunction, and general well being, with mild, moderate, and severe cases indicated by scores of Allocated to Allocated to 75–175, 175–300, and .300, respectively. A reduction in receive true receive sham score of 50 or over is regarded as a clinically significant diet (n=75) diet (n=75) improvement.24 Non-colonic symptomatology,25 such as lethargy, backache, nausea, and urinary symptoms, was 24 Withdrew: 13 Withdrew: Diet too restrictive (n=11) Diet too restrictive (n=3) assessed and scored using visual analogue scales (range 0– Lack of efficacy (n=1) Lack of efficacy (n=3) 500). Quality of life (QOL) was measured using an instru- Not prepared to follow Notpreparedtofollow ment proven to be sensitive to change in IBS (range 0–500).26–28 diet (n=6) diet (n=4) Anxiety and depression were evaluated using the hospital Other reasons (n=6) Other reasons (n=3) anxiety and depression scale (HAD).29 This instrument scores 10 Lost to follow up 9Losttofollowup anxiety and depression up to a maximum score of 21 for each parameter, with a score above 9 indicating significant psychopathology. Data on these measures were recorded at 65 Included 66 Included baseline and after 4, 8, and 12 weeks of the dietary in the final in the final intention to intention to intervention period. In addition, at 4, 8, and 12 weeks, treat analysis treat analysis patients were asked to give a global rating of their IBS using the question, ‘‘Compared with your IBS before you started the food elimination diet, are you now: terrible, worse, Figure 1 Study flow diagram.

www.gutjnl.com Food elimination based on IgG antibodies in IBS 1461

Table 1 Baseline characteristics of the patients

Group True diet (n = 75) Sham diet (n = 75)

Age (y) (range, SD) 44 (17–72; 12.9) 44 (19–74; 15.2) No of males (%) 7 (9.3%) 13 (17.3%) No of foods to which sensitive 6.65 (3.66) 6.63 (4.1) Symptom duration (y) 11.5 (9.9) 10.1 (7.5) IBS symptom severity score 331.9 (70.8) 309.0 (78.5) Non-colonic features score 459.1 (160.7) 452.6 (170.1) Quality of life score 640.1 (252.6) 639.3 (222.3) HAD anxiety score 9.5 (4.6) 9.5 (4.5) HAD depression score 5.3 (3.4) 6.0 (3.6) No of diarrhoea predominant patients (%) 37 (52.1%) 41 (56.9%) No of constipation predominant patients (%) 19 (26.8%) 16 (22.2%) No of alternating predominant patients (%) 15 (21.1%) 15 (20.8%)

Results are expressed as mean (SD). HAD, hospital anxiety and depression scale. relationship between the change in symptoms from baseline CONSORT statement.31 In summary, between January 2001 and treatment group, patient characteristics (for example, and July 2002, 176 patients were eligible for the study, of IBS subtype, history of atopy, number of foods to which which 26 (15%) were excluded from participation, leaving sensitive, and concomitant medication) and adherence to the 150 patients who were all found to be sensitive to at least one diet.30 food. Seventy five of these were randomised to receive an elimination diet based on their true food sensitivity results Sample size calculation and 75 patients to a sham diet. Data from 131 (87%) patients It was estimated that approximately 40% of the placebo arm who gave 12 week data were available for the intention to would report a significant improvement in symptoms. It was treat analysis: 65 and 66 patients from the true and sham calculated that a sample size of 55 patients would be required groups, respectively. in each group to detect, with 90% power, a difference of 30% points in the proportion reporting such an improvement (that Patient characteristics is, 70% in the treatment arm) as statistically significant at the The patients were typical of those with IBS in secondary care 5% level. Assuming a 20% dropout rate, a minimum of 138 practice, the majority being women. Patients, on average, had patients would need to be entered into the trial. Thus we experienced symptoms of IBS for over a decade and were aimed to recruit a total of 150 patients into the study. found to be sensitive to approximately 6–7 foods (range 1– 19). Baseline demographic and clinical characteristics of the RESULTS two groups, including the use of concomitant medication, Recruitment of patients and their flow through each stage of were found to be similar with the exception of the IBS the study is illustrated in fig 1, as recommended by the symptom severity score which was slightly higher in the treatment group (table 1). Thirty per cent of patients were Table 2 Frequency of foods excluded from the diet (% of found to be atopic. patients) The frequency of foods excluded from the diet is shown in table 2. Adherence was lower in those on the true diet Food Treatment group Sham group although no specific adverse events were recorded in either Barley 26.7 9.3 group. Twenty four patients withdrew from the study in the Corn 22.7 14.7 true diet group (mainly because of difficulty in following the Rice 8 54.7 diet) and 13 from the sham diet group (for a variety of Rye 8 25.3 reasons). However, 12 week data were obtained from 14 of Wheat 49.3 8 Milk 84.3 1.3 those who withdrew in the true diet group and four in the Beef 24 9.3 sham diet group. There were no significant differences Chicken 21.3 13.3 Pork 5.3 36 Cabbage 12 24 Celery 5.3 21.3 0 Haricot bean 17.3 14.7 Pea 38.6 1.3 *** _ Potato 9.3 61.3 50 Sham Soy bean 22.7 10.7 diet Tomato 4 44 (n=66) Apple 1.3 33 _ Orange 6.7 29.3 100 Strawberry 0 20 Almond 28 12 _ Brazil nut 22.7 17.3 150 True diet

Cashew nut 49.3 8 IBS symptom severity (n=65) Peanut 10.7 20 Walnut 2.7 29.3 _ 200 Cocoa bean 1.3 21.3 Low Medium High Shellfish 21.3 10.7 Fish mix 17.3 28 Level of adherence Whole egg 57.3 26.7 Yeast 86.7 0 Figure 2 Mean change in symptom severity scores at 12 weeks according to degree of adherence. Difference between the groups with high adherence: 101 (95% confidence interval 54, 147); ***p,0.001.

www.gutjnl.com 1462 Atkinson, Sheldon, Shaath, et al

A between atopic and non-atopic patients. There was however a 400 statistically significant interaction between treatment group and both adherence to the diet and number of foods to which patients were sensitive. For patients sensitive to the average 300 Sham * diet number of foods who fully adhered to their allocated diet, a (n=66) 26% difference in reduction in symptom severity score was observed in favour of the true diet (a difference in score of 98 200 True diet (n=65) (95% CI 52, 144), p,0.001: a standardised effect size of 1.3). This benefit increased by a further 39 points (12%) (95% CI 7, 100 70; p = 0.016) for each food to which they were sensitive

IBS symptom severity over and above the average number. These results were not materially altered by carrying out an ANCOVA analysis (in 0 which the final score is the dependent variable and the 04812 baseline score is included as a covariate) instead of modelling Time (weeks) change in scores.30 The interaction between treatment group and adherence is demonstrated in fig 2 which shows a B 400 greater reduction in symptoms with full adherence in the true diet but not in the sham diet group. Figure 3A and 3B show the average change in symptom severity score over 300 *** Sham 12 weeks for the group as a whole and for those who fully diet adhered, respectively. This reveals that most improvements in (n=40) symptoms are fully achieved within two months. 200 True diet (n=24) Global impact score The reported global rating of change by treatment group is 100

IBS symptom severity shown in table 3. The difference in mean ranking (70.9 v 60.3) was statistically significant (p = 0.048). When this was 0 repeated including only patients who fully adhered to their 04812 diets (table 3), a greater percentage difference favouring the Time (weeks) true diet was found (p = 0.001). The NNT was 9 in the group as a whole and 2.5 in patients fully adherent to the diet. Figure 3 (A) Average symptom severity scores over time for the group as a whole. Difference in mean change from baseline at 12 weeks: true Secondary outcome measures versus sham 39 (95% confidence interval 5, 72); *p = 0.024. (B) Average As can be seen from fig 4A and 4B, all data show changes symptom severity scores over time for the full adherence group. favouring the true diet group and are consistent with the Difference in mean change from baseline at 12 weeks: true versus sham 98 (95% confidence interval 52, 144); ***p,0.001. results for the primary outcomes. These trends were further strengthened after adjustment for adherence and number of food sensitivities but only reached statistical significance for between baseline characteristics of the 19 who were lost to non-colonic symptomatology (p = 0.05). There were no follow up and those for whom 12 week data were obtained. significant changes in medication use during the course of the trial. Primary outcomes IBS symptom severity Reintroduction of eliminated foods Patients in the true diet group experienced a 10% greater Of the 131 patients who gave 12 week data, 93 (41 in the true reduction in symptom severity than those allocated to the and 52 in the sham diet groups) agreed to attempt sham diet, with change in scores of 100 and 61.5, respectively reintroduction of foods they had been asked to eliminate (mean difference 39 (95% confidence interval (CI) 5.2, 72.3); and provided further follow up data on the primary outcomes p = 0.024): a standardised effect size of 0.52 (see fig 3A). measures. Of these, 62% reported full adherence and 37% There were no differences in the response to the diet in terms moderate adherence to the previous elimination diet. Mean of age, sex, IBS bowel habit subtype, or IBS duration. In IBS symptom severity score increased (that is, worsening of addition, there was no difference in response to the diet symptoms) by 83.3 in the true group and by 31 in the sham

Table 3 Global impact score at 12 weeks

Treatment group

True diet Sham diet (n (%)) (n (%))

All patients Significantly worse 3 (4.7) 8 (12.1) No significant change 44 (67.2) 47 (71.2) Significantly improved 18 (28.1) 11 (16.7) Total 65 66 NNT = 9

Patients fully adhering to the diet Significantly worse 1 (4.2) 5 (12.5) No significant change 10 (41.7) 29 (72.5) Significantly improved 13 (54.1) 6 (15) Total 24 40 NNT = 2.5

www.gutjnl.com Food elimination based on IgG antibodies in IBS 1463

True diet Sham diet Table 4 Global rating following reintroduction of foods relative to the end of the elimination phase All patients Full adherence A Treatment group p=0.14 p=0.05 180 n=24 True diet group Sham diet group (n (%)) (n (%))

Significantly worse 17 (41.5) 13 (25) No significant change 23 (56.1) 35 (67.3) 120 n=65 Significantly improved 1 (2.4) 4 (7.7) n=40 Total 41 (100) 52 (100) n=66

60

Non-colonic symptoms worsening of health compared with the sham diet group (p = 0.047). 0

180 p=0.27 p=0.27 DISCUSSION n=24 A clinically significant improvement in IBS symptomatology was observed in patients eliminating foods to which they n=65 120 were found to exhibit sensitivity, as identified by an ELISA test for the presence of IgG antibodies to these foods. The n=66 n=40 number needed to treat of 9 for the group as a whole and 2.5 for patients closely adhering to the diet are both considerably 60 Qualityof life better than the value of 17 achieved after three months of treatment with tegaserod,32 a drug that has been recently licensed in the USA for use in IBS. IBS symptom severity and 0 global rating scores were chosen as primary outcome measures in this study as they represented the most direct measure of clinical improvement in this condition based on patient self assessment. Rather than using the traditional True diet Sham diet method of classifying global improvement as any value All patients Full adherence exceeding adequate relief of symptoms, we used a much B stricter definition requiring patients to report symptoms as p=0.18 p=0.18 being either ‘‘better’’ or ‘‘excellent’’ compared with pretreat- 2.5 n=24 ment levels. Despite this, the diet still achieved a significant n=65 improvement. However, as might be expected, the placebo 2 response using this end point was somewhat lower than that usually reported in IBS treatment trials which have used less 1.5 n=66 n=40 demanding criteria. The observation that patients on the sham diet also improved, although to a lesser extent, 1 emphasises the importance of conducting double blind randomised controlled trials of such non-drug interventions 0.5 in order to avoid overestimating their potential. Most patients with IBS have attempted at least some form 0 of dietary modification, which in some cases can be very extreme. Conflicting results have been reported using 2.5 p=0.78 p=0.26 exclusion diets4 5 33–36 and this approach also suffers from the limitation that it has to be empirical. Thus potentially 2 n=24 offending foods can only be identified after their elimination and subsequent reintroduction. This time consuming process 1.5 would be much reduced if the offending foods could be identified beforehand. Attempts to do this using IgE 1 n=65 antibodies have been disappointing8–10 but the results of this

Depression Anxiety n=40 n=66 study suggest that measuring IgG antibodies may be much 0.5 more rewarding. The response to the IgG based diet in our trial did not correlate with atopic status, the prevalence of 0 which was found to be no greater than that occurring in the general population.37 Figure 4 (A) Mean change in the secondary outcome measures of non- The observation that adherence to the diet is critical in colonic symptoms and quality of life for the group as a whole and the full determining a good outcome in the ‘‘true’’ diet group but not adherence group. (B) Mean change in the secondary outcome measures the ‘‘sham’’ group is indicative of the fact that the diet is an of anxiety and depression for the group as a whole and the full ‘‘active treatment’’ which if not adhered to, does not seem to adherence group. have an effect. This notion is further supported by the observation that a significantly greater deterioration was group, a statistically significant difference of 52 (24%) (95% observed in subjects in the true diet group compared with CI 18, 86; p = 0.003). The change in global score following those in the sham group when they reintroduced eliminated reintroduction of foods is shown in table 4. This indicates a foods at the end of the diet phase of the trial. Furthermore, reversal of the pattern observed during the active treatment the improvement of 98 in the symptom severity score in those phase, with more patients in the true diet group showing fully adherent in the true diet group is well above the value of

www.gutjnl.com 1464 Atkinson, Sheldon, Shaath, et al

50, which is regarded as being of clinical significance both in 11 Crowe SE, Perdue MH. Gastrointestinal food hypersensitivity: basic 24 26–28 mechanisms of pathophysiology. Gastroenterol 1992;103:1075–95. validation studies and clinical practice. It was interesting 12 el Rafei A, Peters SM, Harris N, et al. Diagnostic value of IgG4 measurements to note that patients exhibiting a greater number of in patients with food allergy. Ann Allergy 1989;62:94–9. sensitivities, as determined by the IgG test, experienced a 13 Host A, Husby S, Gjesing B, et al. Prospective estimation of IgG, IgG subclass greater symptom reduction if they adhered to the true but not and IgE antibodies to dietary proteins in infants with cow’s milk allergy. Levels of antibodies to whole milk protein, BLG and ovalbumin in relation to repeated the sham diet. milk challenge and clinical course of cow’s milk allergy. Allergy There is currently considerable interest in the concept that 1992;47:218–29. at least in some patients, IBS may have an inflammatory 14 Awazuhara H, Kawai H, Maruchi N. Major allergens in soybean and clinical 38–42 significance of IgG4 antibodies investigated by IgE and IgG4 immunoblotting component. Most of the work in this area has centred on with sera from soybean-sensitive patients. Clin Exp Allergy 1997;27:325–32. post dysenteric IBS, with gut pathogens being viewed as the 15 Barnes RMR, Johnson PM, Harvey MM, et al. Human serum antibodies initiators of this process which can be identified by subtle reactive with dietary proteins: IgG subclass distribution. Int Arch Allergy Appl 38 Immunol 1988;87:184–8. changes on histology. However, if, as indicated in this study, 16 Lessof MH, Kemeny DM, Price JF. IgG antibodies to food in health and IgG antibodies to food are important in the pathogenesis of disease. Allergy Proc 1991;12:305–7. IBS in some patients, they too may be of relevance. Not all 17 Husby S, Mestecky J, Moldoveanu Z, et al. Oral tolerance in humans. T cell but not B cell tolerance after antigen feeding. J Immunol 1994;152:4663–70. patients exhibiting histological features consistent with post 18 Haddad ZH, Vetter M, Friedmann J, et al. Detection and kinetics of antigen- dysenteric IBS give a history of a previous dysenteric illness. specific IgE and IgG immune complexes in food allergy. Ann Allergy This is usually assumed to be due to the fact that this has 1983;51:255. 19 Husby S, Oxelius VA, Teisner B, et al. Humoral immunity to dietary antigens in been forgotten by the patient but our results may suggest an healthy adults. Occurrence, isotype and IgG subclass distribution of serum alternative mechanism for immune activation and inflam- antibodies to protein antigens. Int Arch Allergy Appl Immunol mation without the need for prior infection. 1985;77:416–22. It is now well recognised that up to 70% of patients with 20 Kruszewski J, Raczka A, Klos M, et al. High serum levels of allergen specific 43 IgG-4 (asIgG-4) for common food allergens in healthy blood donors. Arch IBS have evidence of hypersensitivity of the rectum, which Immunol Ther Exp 1994;42:259–61. probably extends to involve most of the gut in many 21 Finn R, Smith MA, Youngs GR, et al. Immunological hypersensitivity to individuals.44 It is possible that this hypersensitivity renders environmental antigens in the irritable bowel syndrome. Br J Clin Pract 1987;41:1041–3. patients more reactive to a low grade inflammatory process 22 Drossman DA, Corazziari E, Talley NJ, et al. Rome II: a multinational which would not necessarily cause symptoms in a normal consensus document on functional gastrointestinal disorders. Gut individual. This would explain why excluding foods to which 1999;45:1–81. 23 Foster AP, Knowles TG, Hotston Moore A, et al. Serum IgE and IgG responses patients have IgG antibodies might be particularly beneficial to food antigens in normal and atopic dogs, and dogs with gastrointestinal in IBS despite the fact that these antibodies may also be disease. Vet Immunol Immunopathol 2003;92:113–24. present in the general population. Indeed, if this mechanism 24 Francis CY, Morris J, Whorwell PJ. The irritable bowel scoring system: A simple method of monitoring IBS and its progress. Aliment Pharmacol Therap is particularly important in IBS, it might be anticipated that 1997;11:395–402. IgG food antibodies would be relatively common in this 25 Whorwell PJ, McCallum H, Creed FH, et al. Non-colonic features of irritable condition, as was the case in our study. bowel syndrome. Gut 1986;27:452–6. 26 Houghton LA, Heyman DJ, Whorwell PJ. Symptomatology, quality of life and Many patients with IBS would prefer a dietary solution to economic features of irritable bowel syndrome—the effect of hypnotherapy. their problem rather than having to take medication, and the Aliment Pharmacol Ther 1996;10:91–5. economic benefits of this approach to health services are 27 Gonsalkorale WM, Toner BB, Whorwell PJ. Cognitive change in patients obvious. It is well known that patients expend large sums of undergoing hypnotherapy for irritable bowel syndrome. J Psychosom Res 2004;56:271–8. money on a variety of unsubstantiated tests in a vain attempt 28 Gonsalkorale WM, Houghton LA, Whorwell PJ. Hypnotherapy in irritable to identify dietary intolerances. The results of this study bowel syndrome: a large scale audit of a clinical service with examination of suggest that assay of IgG antibodies to food may have a role factors influencing responsiveness. Am J Gastroenterol 2002;97:954–61. 29 Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta in helping patients identify candidate foods for elimination Psychiatr Scand 1983;67:361–70. and is an approach that is worthy of further biomedical and 30 Everitt BS, Pickles A. Statistical aspects of the design and analysis of clinical clinical research. trials. London: Imperial College Press Publishers, 2003:108–42. 31 Altman DG, Schulz KF, Moher D, et al. The revised CONSORT statement for reporting randomized trials: explanation and elaboration. Ann Intern Med ...... 2001;134:663–94. Authors’ affiliations 32 Novick J, Miner P, Krause R, et al. A randomised, double blind, placebo W Atkinson, N Shaath, P J Whorwell, Department of Medicine, controlled trial of tegaserod in female patients suffering from irritable bowel University Hospital of South Manchester, Manchester, UK syndrome with constipation. Aliment Pharmacol Ther 2002;16:1877–88. 33 Niec AM, Frankum B, Talley NJ. Are adverse reactions to food linked to T A Sheldon, Department of Health Sciences, University of York, York, irritable bowel syndrome? Am J Gastroenterol 1998;93:2184–90. UK 34 Burden S. Dietary treatment of irritable bowel syndrome: current evidence and guidelines for future practice. J Hum Nutr Diet 2001;14:231–41. 35 Bentley SJ, Pearson DJ, Rix KJB. Food hypersensitivity in irritable bowel REFERENCES syndrome. Lancet 1983;2:295–7. 1 Drossman DA, Camilleri M, Mayer EA, et al. American Gastroenterological 36 McKee AM, Prior A, Whorwell PJ. Exclusion diets in irritable bowel syndrome: Association Technical Review on Irritable Bowel Syndrome. Gastroenterology Are they worthwhile? J Clin Gastroenterol 1987;9:526–8. 2002;123:2108–31. 37 Durham SR, Church MK. Principles of allergy diagnosis. In: Holgate ST, 2 Lea R, Whorwell PJ. Quality of life in irritable bowel syndrome. Church MK, Lichtenstein LM, eds. Allergy, 2nd edn. London: Mosby, Pharmacoeconomics 2001;19:643–53. 2001:3–16. 3 Talley NJ, Spiller R. Irritable bowel syndrome: a little understood organic 38 Spiller RC, Jenkins D, Thornley JP, et al. Increased rectal mucosal bowel disease? Lancet 2002;360:555–64. enteroendocrine cells, T lymphocytes, and increased gut permeability 4 Jones VA, McLaughlan P, Shorthouse M, et al. Food intolerance: a major following acute Campylobacter enteritis and in post-dysenteric irritable bowel factor in pathogenesis of irritable bowel syndrome. Lancet 1982;2:1115–17. syndrome. Gut 2000;47:804–11. 5 Nanda R, James R, Smith H, et al. Food intolerance and the irritable bowel 39 Gonsalkorale WM, Perrey C, Pravica V, et al. Interleukin 10 genotypes in syndrome. Gut 1989;30:1099–104. irritable bowel syndrome: evidence for an inflammatory component? Gut 6 Zwetchkenbaum J, Burakoff, R. The irritable bowel syndrome and food 2003;52:91–3. hypersensitivity. Ann Allerg 1988;61:47–9. 40 Collins SM, Piche T, Rampal P. The putative role of inflammation in the irritable 7 Zar S, Kumar D, Benson M. J. Review article: food hypersensitivity and bowel syndrome. Gut 2001;49:743–5. irritable bowel syndrome, Aliment Pharm Ther 2001;15:439–43. 41 Collins SM. A case for an immunological basis for irritable bowel syndrome. 8 Petitpierre M, Gumowski P, Girard JP. Irritable bowel syndrome and Gastroenterology 2002;122:2078–80. hypersensitivity to food. Ann Allergy 1985;54:538–40. 42 Chadwick VS, Chen W, Shu D, et al. Activation of the mucosal immune system 9 Barau E, Dupont C. Modifications of intestinal permeability during food in irritable bowel syndrome. Gastroenterology 2002;122:1778–83. provocation procedures in pediatric irritable bowel syndrome. J Pediatr 43 Mertz H. Review article: visceral hypersensitivity. Aliment Pharmacol Ther Gastroenterol Nutr 1990;11:72–7. 2003;17:623–33. 10 Roussos A, Koursarakos P, Patsopoulos D, et al. Increased prevalence of 44 Francis CY, Houghton LA, Whorwell PJ. Enhanced sensitivity of the whole gut irritable bowel syndrome in patients with bronchial asthma. Respir Med in patients with irritable bowel syndrome. Gastroenterology 2003;97:75–9. 1995;108:601(abstract).

www.gutjnl.com 1391

LEADING ARTICLE Food allergy in irritable bowel syndrome: new facts and old fallacies

E Isolauri, S Rautava, M Kallioma¨ki ......

Gut 2004;53:1391–1393. doi: 10.1136/gut.2004.044990 The notion of food allergy in irritable bowel syndrome (IBS) the presence or absence of disease, but the complex cascade of events determining their use. is not new. However, recent evidence suggests significant The notion of food allergy in irritable bowel reduction in IBS symptom severity in patients on elimination syndrome (IBS) resurfaces in scientific thinking diets, provided that dietary elimination is based on foods in this issue of Gut3 on the basis of a solid randomised placebo controlled trial conducted by against which the individual had raised IgG antibodies. Atkinson and colleagues (see page 1459). These findings should encourage studies dissecting the Determination of serum IgG antibodies against mechanisms responsible for IgG production against dietary foods was used to guide the construction of elimination diets. antigens and their putative role in IBS The presence of specific IgG class antibodies is ...... often accepted as uncommitted or protective ‘‘altered reactivity’’, unlike those of the IgE class. Detection of antigen specific IgE is invariably ringing empirical observations ad fontes taken as an attribute of causality, a condition advances science. In astrophysics, the term called ‘‘IgE mediated disease’’ and, more speci- ‘‘black hole’’ was introduced to describe an B fically, of ‘‘allergy’’.1 However, empirical data are extremely dense star which had collapsed into a accumulating to suggest that transient increases singularity under its own gravity. A black hole in antigen specific IgE antibodies prevail in most radiates nothing; it absorbs all matter and energy healthy asymptomatic children during the first falling within its sphere. The name was coined five years of life.4 Secondly, generation of these only after revisiting the initial theoretical antibodies (sensitisation) on antigen exposure achievements of Karl Schwarzschild, when may not necessarily induce clinical disease observations made outside the earth’s atmos- (atopic disease).5 Thirdly, reducing the risk of phere gave astrophysicists empirical ray data on x atopic disease does not necessitate reduction of a new type of cosmic object. In allergology, in sensitisation6–8 and, finally, resolution or aggra- contrast, adherence to a paradigm whereby vation of clinical disease is not invariably allergy is defined by the presence of specific IgE associated with a corresponding alteration in antibodies has hampered disentanglement. As a antibody concentration. Bearing these limita- result, allergy remains a dubiously defined term tions in mind, however, the clinician may with no unambiguous empirical content or successfully profit from determination of specific explanatory power. The time has come to seize IgE to complete the clinical history in an attempt upon the available empirical data and plunge to identify potential offending antigens in a into the original theory of Clemens von Pirquet. symptomatic patient’s diet for the explicit The term allergy was introduced by von diagnostic elimination-challenge procedure.9 Pirquet to denote a changed immunological This is precisely what Atkinson et al did, with reactivity which manifests itself on second specific IgG antibodies.3 They identified a sig- 1 exposure to an antigen (reviewed by Kay ). This nificant reduction in IBS symptom severity in altered reactivity is uncommitted, giving no patients on elimination diets, provided that indication of the direction of change; equally dietary elimination was based on foods against harmful and protective immune reactivity which the individual had raised IgG antibodies; reflects prior encounter (see fig 1). In modern fully compliant patients showed the best clinical terms, altered reactivity can be seen to evince improvement. The reverse pattern was observed See end of article for either the most common mode of immune after reintroduction of the respective foods. authors’ affiliations response elicited by the intestinal immune ...... system, tolerance, recently defined as any ‘‘IBS appears to result from an interplay Correspondence to: mechanism by which a potentially injurious Dr E Isolauri, Department immune response is prevented, suppressed, or between susceptibility genes and impaired of Paediatrics, Turku shifted to a non-injurious class of immune gut barrier functions, immunological dysre- University Central response,2 or abrogation of such an actively gulation, together with bacterial and viral Hospital, 20520 Turku, Finland; erika.isolauri@ maintained process, which is currently linked infections and other environmental factors’’ utu.fi to immunoinflammatory disease. Reassessment of the original theory of allergy is important as it In common with allergic disease, IBS appears Revised version received would appear that it is not the immunological to result from an interplay between susceptibility 10 June 2004 Accepted for publication resources gained during antigen exposure, mea- 19 June 2004 surable by specific antibodies or specifically Abbreviations: IBS, irritable bowel syndrome; PRR, ...... responding lymphocytes, which are decisive for pattern recognition receptor

www.gutjnl.com 1392 Isolauri, Rautava, Kallioma¨ki

Th2 Th1 Allergic IBD Inflammation disease IBS?

Normal range

Naïve Tolerance uncommitted Antigen Consolidation of Exposure immune reactivity

Figure 1 An attempt to reformulate the original conception of the immunological basis of allergy. On exposure to dietary and microbial antigens, the naı¨ve immune system matures either to establish appropriate immune reactivity and tolerance or, in the case of lack of suppressive and regulatory signals, develops inflammatory disease expressed as Th1 skewed autoimmune reactivity (for example, inflammatory bowel disease (IBD)) or Th2 skewed allergic disease. The exact nature of immune reactivity in irritable bowel syndrome (IBS) on this axis remains elusive. genes and impaired gut barrier functions, immunological addition to its effects on T cell function, transforming growth dysregulation, together with bacterial and viral infections factor b is a key factor in IgA production21 and thus and other environmental factors. It is no easy matter to contributes to maintenance of gut barrier function and to describe succinctly ‘‘gut barrier function’’. In the gastro- immune responses at other mucosal surfaces also. Taken intestinal tract, the external and internal environments are in together, ‘‘gut barrier function’’ strongly depends on antigen close proximity. The dilemma of the mucosal surface of the processing and presentation and the cytokine milieu in the intestine is to fend off the constant challenge from antigens, mucosal immune system, and determines the nature of the such as microorganisms, in mounting a brisk response to immune response (that is, tolerance or inflammation) pathogens, and to enable assimilation of innocuous antigens elicited to a particular antigen. derived from food. In order to perform these opposing functions, the intestine is in a state of continuous immune ‘‘Inflammation can cause profound alterations in the responsiveness, and a delicate balance is generated and function of smooth muscle and enteric nerves as well as maintained between concomitant facilitation and suppres- in deeper neuromuscular layers’’ sion of inflammatory responses. Gut barrier function consists of physiological and immu- In certain circumstances, such as metabolic stress, the nological factors which exclude and degrade antigens and peaceful coexistence across the barrier is disturbed and an restrict their adherence, penetration, and transfer. Antigen inflammatory response ensues.22 Abrogated barrier function presenting cells, and more precisely dendritic cells, are pivotal of the gut mucosa leads to greater antigen transfer when the 10–12 in directing mucosal immune responses. Three dendritic routes of transport are also altered, thereby evoking aberrant cell derived signals are required for an effective T cell immune responses and release of proinflammatory cytokines 11 response. The nature of signal 1 depends on the antigen with further impairment of barrier function. Inflammation in question and its processing; necessary costimulatory can cause profound alterations in the function of smooth molecules create the second signal and the pericellular muscle and enteric nerves as well as in deeper neuromuscular cytokine milieu is the basis of the third. On antigen layers.23 Indeed, a subtle inflammatory response and exag- recognition, maturation of dendritic cells and secretion of gerated sensitivity to that type of response has been cytokines and chemokines occur. These secretions direct the suggested to be causative in IBS. In view of recently reported polarisation of a naı¨ve T helper cell to type 1, type 2, or a alterations in the immunological defence in IBS,24 the regulatory T cell and thus regulate other adaptive immune trigger(s) of the vicious circle can be depicted among the responses, such as B cell derived immunoglobulin produc- intraluminal antigens. tion.11 Tolerance to lumenal dietary and microbial antigens is In this issue of Gut, Atkinson and colleagues3 describe IgG likely to be achieved through those dendritic cells which antibody responses to dietary antigens of clinical significance induce production of regulatory T cells secreting interleukin and an apparent causal relation to symptoms in IBS, in a 10 and transforming growth factor b.13 These cytokines fashion resembling the elimination-challenge procedure in promote gut barrier function by suppressing the production food allergy. To broaden this concept, it is intriguing to of both T helper 1 and 2 cytokines,14 overexpression of which speculate that IBS may perhaps also be associated with IgG is associated with increased gut permeability.15 16 Moreover, antibodies against other intraluminal antigens such as those the anergic T cells induced by interleukin 10 exposed from the indigenous microbiota, partially analogously to loss dendritic cells appear to be able to suppress other T cells in of tolerance to gut microbiota in inflammatory bowel an antigen specific manner.17 Transforming growth factor b disease.25 26 downregulates both T helper 1 and 2 responses directly18 and The human gastrointestinal tract harbours a complex indirectly by modulating the activity of antigen presenting collection of microorganisms which form the individual cells19 and favouring the development of regulatory T cells.20 microbiota typical for each person. Defence is facilitated by After intestinal priming, these cells migrate to the periphery, peristalsis, secretion of mucus and antimicrobial peptides thus mediating peripheral tolerance on reactivation. In such as defensins and cathelicidins, and commensal induced

www.gutjnl.com Food allergy in IBS 1393

IgA.27 28 Intestinal epithelial cells further contribute to the 5 Lau S, Illi S, Sommerfeld C, et al. Early exposure to house-dust mite and cat allergens and development of childhood asthma: a cohort study. Lancet homeostasis of gut barrier function by a scarcity of both 2000;356:1392–7. pattern recognition receptors (PRRs; for example, toll-like 6 Kallioma¨ki M, Salminen S, Kero P, et al. Probiotics in the primary prevention receptors and nucleotide binding oligomerisation domain of atopic disease: a randomised, placebo-controlled trial. Lancet proteins) and their coreceptors, expression of active negative 2001;357:1076–9. 7 Riedler J, Braun-Fahrla¨nder C, Eder W, et al. Exposure to farming in early life regulators of PRR signalling, and secretion of the suppressive and development of asthma and allergy: a cross-sectional survey. Lancet cytokines interleukin 10 and transforming growth factor b.13 2001;358:1129–33. All of these characteristics assist in preventing unnecessary 8 Jones CA, Holloway JA, Popplewell EJ, et al. Reduced soluble CD14 levels in amniotic fluid and breast milk are associated with the subsequent development and even hazardous systemic immunity to commensals while of atopy, eczema, or both. J Allergy Clin Immunol 2002;109:858–66. allowing local protective mucosal immune responses. In 9 Bock SA. Diagnostic evaluation. Pediatrics 2003;111:1638–44. addition, some specific strains of non-pathogenic bacteria 10 Pulendran B, Banchereau J, Maraskovsky E, et al. Modulating the immune response with dendritic cells and their growth factors. Trends Immunol have been shown to attenuate intestinal inflammation by 2001;22:41–7. selective inhibition of intracellular signalling pathways 11 Kapsenberg ML. Dendritic-cell control of pathogen-driven T-cell polarization. elicited by diverse potentially deleterious stimuli.29 30 A Nat Rev Immunol 2003;3:984–93. healthy gut microbiota is thus an indispensable component 12 Stagg AJ, Hart AL, Knight SC, et al. The dendritic cell: its role in intestinal inflammation and relationship with gut bacteria. Gut 2003;52:1522–9. of ‘‘gut barrier function’’. 13 Nagler-Anderson C, Bhan AK, et al. Control freaks: immune regulatory cells. Nat Immunol 2004;5:119–22. 14 Rissoan MC, Soumelis V, Kadowaki N, et al. Reciprocal control of T helper cell ‘‘A healthy gut microbiota is thus an indispensable and dendritic cell differentiation. Science 1999;283:1183–6. component of gut barrier function’’ 15 Heyman M, Darmon N, Dupont C, et al. Mononuclear cells from infants allergic to cow’s milk secrete tumor necrosis factor a, altering intestinal function. Gastroenterology 1994;106:1514–23. The findings of Atkinson and colleagues3 should encourage 16 Berin MC, Yang PC, Ciok L, et al. Role for IL-4 in macromolecular transport studies dissecting the mechanisms responsible for IgG across human intestinal epithelium. Am J Physiol 1999;276:C1046–52. 17 Steinbrink K, Graulich E, Kubsch S, et al. CD4(+) and CD8(+) anergic T cells production against dietary antigens and their putative role induced by interleukin-10-treated human dendritic cells display antigen- in IBS. This may serve not only IBS research but also that into specific suppressor activity. Blood 2002;99:2468–76. allergy and allergic diseases. In the perspectives of both 18 Lu´dviksson BR, Seegers D, Resnick AS, et al. The effect of TGF-beta1 on immune responses of naive versus memory CD4+ Th1/Th2 T cells. normal gut barrier function and the vague findings in a few Eur J Immunol 2000;30:2101–11. 31–33 studies of probiotic supplementation in IBS, we suggest 19 Takeuchi M, Alard P, Streilein JW. TGF-beta promotes immune deviation by that the possible role of the gut microbiota in the pathogen- altering accessory signals of antigen-presenting cells. J Immunol esis of IBS may deserve closer attention. If the host-microbe 1998;160:1589–97. 20 Toms C, Powrie F. Control of intestinal inflammation by regulatory T cells. cross talk is misinterpreted in IBS, a working target for novel Microbes Infect 2001;3:929–35. therapeutic interventions beyond elimination diets could be 21 Stavnezer J. Regulation of antibody production and class switching by TGF- provided in modulating the composition and/or activity of the beta. J Immunol 1995;155:1647–51. 22 Nazli A, Yang PC, Jury J, et al. Epithelia under metabolic stress perceive gut microbiota and promoting gut immune defence. Research commensal bacteria as a threat. Am J Pathol 2004;164:947–57. interest in the science of nutrition is directed towards 23 Collins SM, Piche T, Rampal P. The putative role of inflammation in the irritable improvement of defined physiological functions beyond the bowel syndrome. Gut 2001;49:743–5. 24 Gonsalkorale WM, Perrey C, Pravica V, et al. Interleukin 10 genotypes in nutritional impact of food. The search for active non-nutritive irritable bowel syndrome: evidence for an inflammatory component? Gut compounds is also a focus for research in the treatment and 2003;52:91–3. prevention of allergic diseases. 25 Duchmann R, Kaiser I, Hermann E, et al. Tolerance exists towards resident intestinal flora but is broken in active inflammatory bowel disease (IBD). Clin ...... Exp Immunol 1995;102:448–55. 26 Lodes MJ, Cong Y, Elson CO, et al. Bacterial flagellin is a dominant antigen in Authors’ affiliations Crohn disease. J Clin Ivest 2004;113:1296–306. E Isolauri, S Rautava, Department of Paediatrics, University of Turku and 27 Otte J-M, Kiehne K, Herzig K-H. Antimicrobial peptides in innate immunity of Turku University Central Hospital, Finland the human intestine. J Gastroenterol 2003;38:717–26. M Kallioma¨ki, Massachusetts General Hospital East, Combined Program 28 Macpherson AJ, Uhr T. Induction of protective IgA by intestinal dendritic cells in Pediatric Gastroenterology and Nutrition, Charlestown, carrying commensal bacteria. Science 2004;303:1662–5. 29 Haller D, Jobin C. Interaction between resident luminal bacteria and the host: Massachusetts, USA can a healthy relationship turn sour? J Pediatr Gastroenterol Nutr 2004;38:123–36. 30 Kelly D, Campbell JI, King TP, et al. Commensal anaerobic gut bacteria REFERENCES attenuate inflammation by regulating nuclear-cytoplasmic shuttling of PPAR-c 1 Kay AB. Concepts of allergy and hypersensitivity. In: Kay AB, ed. Allergy and and RelA. Nat Immunol 2004;5:104–12. allergic diseases. Oxford: Blackwell Science Ltd, 1997:23–35. 31 Nobaek S, Johansson M-L, Molin G, et al. Alteration of intestinal microflora is 2 Weiner HL. Oral tolerance: immune mechanisms and the generation of Th3- associated with reduction in abnormal bloating and pain in patients with type TGF-beta-secreting regulatory cells. Microbes Infect 2001;3:947–54. irritable bowel syndrome. Am J Gastroenterol 2000;95:1231–8. 3 Atkinson W, Sheldon TA, Shaath N, et al. Food elimination based on IgG 32 Sen S, Mullan MM, Parker TJ, et al. Effect of Lactobacillus plantarum 299v on antibodies in irritable bowel symdrome: a randomised controlled trial. Gut colonic fermentation and symptoms of irritable bowel syndrome. Dig Dis Sci 2004;53:1459–64. 2002;47:2615–20. 4 Kulig M, Bergmann R, Klettke U, et al. Natural course of sensitization to food 33 Kim HJ, Camilleri M, McKinzie S, et al. A randomized controlled trial of a and inhalant allergens during the first 6 years of life. J Allergy Clin Immunol probiotic, VSL#3, on gut transit and symptoms in diarrhoea-predominant 1999;103:1173–9. irritable bowel syndrome. Aliment Pharmacol Ther 2003;17:895–904.

www.gutjnl.com

HEADACHE CARE 1742–3430 VOL. 2, NO. 2, 2005, 105–110 doi:10.1185/174234305X14962

© 2005 LIBRAPHARM LIMITED

ORIGINAL ARTICLE A Prospective Audit of Food Intolerance among Migraine Patients in Primary Care Clinical Practice Trevor Rees1, David Watson2, Susan Lipscombe3, Helen Speight4, Peter Cousins4, Geoffrey Hardman5 and Andrew J. Dowson6 1Hawthorns Surgery, Sutton Coldfield, West Midlands, UK 2Hamilton Medical Group, Aberdeen, Scotland 3Park Crescent New Surgery, Brighton, East Sussex, UK 4YORKTEST Laboratories Ltd (YTL), Osbaldwick, York, UK 5Centre for Health Economics, University of York, York, UK 6King’s Headache Service, King’s College Hospital, London, UK

Address for correspondence: Dr Andrew J. Dowson, The King’s Headache Service, King’s College Hospital, Denmark Hill, London SE5 9RS, UK. Tel./Fax: +44-1428-712546; email: [email protected] Key words: Diet – Food intolerance – Migraine – Primary care

SUMMARY

This prospective audit was set up to investigate took part in the audit and 39 completed 2 months whether migraine sufferers have evidence of of investigation. The mean number of foods IgG-based food intolerances and whether their identified in the IgG test was 5.3 for all condition can be improved by the withdrawal from participants and 4.7 for those successfully the diet of specific foods identified by intolerance altering their diet. About 90% of patients testing. Migraine patients were recruited from changed their diet to a greater or lesser primary care practices and a blood sample was extent following the identification of possible taken. Enzyme-linked immunosorbent assays food intolerances. A marked proportion of the (ELISA) were conducted on the blood samples to migraine patients benefited from the dietary detect food-specific IgG in the serum. Patients intervention, approximately 30% and 40% identified with food intolerances were encouraged reporting considerable benefit at 1 and 2 to alter their diets to eliminate appropriate foods months, respectively. Also, over 60% of patients and were followed up for a 2-month period. who reintroduced the suspect foods back into Endpoints included identification of the specific their diets reported the return of their migraine foods that the patients were intolerant to, symptoms. This investigation demonstrated assessing the proportion of patients who altered that food intolerances mediated via IgG may their diet and the benefit obtained by these be associated with migraine and that changing patients at 1 and 2 months. Patients reported the the diet to eradicate specific foods may be a level of benefit on a 6-point scale, where 0 = no potentially effective treatment for migraine. benefit and 5 = high benefit. Sixty one patients Further clinical studies are warranted in this area.

Paper 0047 105 Introduction The present audit was set up to investigate whether migraine sufferers have evidence of IgG-based food Dietary components are frequently proposed as precip- intolerances and whether their condition can be itating factors for migraine, particularly in children and improved by the withdrawal from the diet of specific adolescents1,2. Many different foods have been implic- foods identified by intolerance testing. ated as potential triggers for migraine attacks, including chocolate, cheese, red wine and many others3. However, evidence for this interaction is mostly anecdotal and Patients and Methods based on patient reports4. Open studies indicated that 5 6 low-fat and high carbohydrate diets could lead to Patients improvements in migraine frequency and/or severity. In contrast, no controlled study has confirmed the incidence Established adult migraine patients (age > 18 years) of food-evoked migraine attacks. A controlled study were recruited from primary care clinical practices by with chocolate failed to show that it provoked migraine their GPs. Patients were required to have high-impact attacks7. An alternative concept of the relationship of headaches. Patients were diagnosed with episodic food with migraine is that food cravings occur during migraine (≤ 15 days of headache per month) or chronic the prodrome phase; the food intake thus being a migraine (> 15 days of headache per month), according consequence of the attack rather than a cause of it8. to the GP’s usual practices. All patients provided their Migraine may be precipitated by food via chemical or written informed consent to take part in the audit. immunological mechanisms. Dietary components may affect phases of the migraine process by influencing Study Design release of serotonin and noradrenaline, causing vaso- constriction or vasodilatation, or by direct stimulation This prospective audit investigated whether migraine of trigeminal ganglia, brainstem and cortical neuronal patients identified in primary care clinical practice pathways1. Immunological reactions may be mediated exhibited food intolerances measured as elevated IgG by Immunoglobulin E (IgE [classical food allergies levels to specific foods. The audit also investigated the occurring immediately after eating]) or, more controv- effect of withdrawing foods associated with high IgG ersially, by Immunoglobulin G (IgG [food intolerance levels on patients’ migraine attacks over a 2-month involving a delayed allergic reaction 2–120 hours period. after eating]). Available evidence indicates that an Primary care physicians were briefed on the rationale IgE mechanism is relatively unimportant in food- and objectives of the audit at a meeting of the UK induced migraine9 and a review of the clinical literature charity Migraine in Primary Care Advisors (MIPCA) established no clear evidence of an immune dysfunction and agreed to participate. Each physician recruited in migraine sufferers10. However, the role of a putative up to 20 migraine patients and provided them with IgG mechanism is presently unknown. information about the audit. Before entering the study, The usual way to treat food intolerance is by food all patients completed a baseline questionnaire to elimination and re-challenge procedures, which record demography and allergy history and a Headache are imprecise, lengthy and inefficient. As a more Impact Test (HIT-613) questionnaire to record headache efficient alternative to this approach, an enzyme-linked severity. immunosorbent assay (ELISA) test to a panel of 113 Patients who completed the initial questionnaires food allergen extracts has been developed (YORKTEST were sent a validated blood testing kit by YTL. Patients Laboratories Ltd [YTL], York, UK). This detects raised took a blood sample by skin prick as detailed in the food-specific IgG in the serum of people with one or leaflet enclosed with the testing kit and returned the more, usually chronic, conditions. Patients with raised kit to YTL by mail. The blood samples were processed IgG levels to specific foods are advised to remove these by YTL on receipt of the questionnaires and blood kit. from their diets and their progress is monitored with ELISA tests on blood samples were used to detect food- a series of questionnaires. An independent audit of specific IgG in the serum of the blood samples. Results patients treated in this way between February 1998 and of the ELISA tests were sent directly to the patients by August 1999 showed that approximately 50% of all YTL, together with a guidebook on food intolerances patients reported a high or relatively high response and their treatment14. to dietary therapy, based on their levels of food- Patients were free to change their diets to eliminate specific IgGs11. A randomised, controlled clinical trial specific foods identified by the ELISA tests as possibly has demonstrated beneficial effects of this form of causing intolerance, either on their own initiative or dietary therapy on symptom relief for irritable bowel after consultation with their GP or other healthcare syndrome12. professional. Patients had access by telephone to a

106 A Prospective Audit of Food Intolerance among Migraine Patients in Primary Care Clinical Practice © 2005 LIBRAPHARM LTD – Headache Care 2005; 2(2) professional dietitian to help them with any dietary The majority of patients (78.0%) were in full-time alterations that they wished to implement. Follow-up education or employment. In examining the allergy questionnaires were sent to patients after 1 and 2 months history, 15 patients (24.6%) were aware of foods they to monitor their progress (investigation period). felt they were allergic to, 42 (68.9%) were in contact with pets, 12 (20.0%) were in contact with chemicals Study Endpoints and Statistical Analyses or occupational dust, 52 (86.7%) were currently taking medication and 21 (34.4%) knew about medications The main study endpoints were: they felt they were allergic to. Fifteen patients (24.6%) were current smokers and 18 (29.5%) had given up • Demographic data on the patient population, and smoking. Forty three patients (72.9%) drank alcohol but details of their allergy and headache histories, analysed only eight drank over seven units per week and only one as descriptive statistics. drank more than 14 units per week. • Identification of the specific foods to which the Most patients had suffered from headache for a patients could be intolerant, identified from the considerable time; 64% for ≥ 10 years, 20% for 5–10 ELISA tests of IgG levels and analysed as descriptive years and 16% for < 5 years. Patients were severely statistics. affected by their headaches (Table 2). Eighty two per • The proportion of patients who altered their diet due cent ‘very often’ or ‘always’ had severe pain, while to their ELISA test results, analysed as descriptive 67% were ‘very often’ or ‘always’ limited in their usual statistics. activities during their headaches. Between 87% and 90% • The benefit obtained at 1 and 2 months by the of patients were too tired to work, felt irritation and patients who altered their diet compared with the suffered from lack of concentration at least sometimes situation before diet alteration, analysed as descriptive during their attacks. Patients reported a mean of 10.1 statistics. Patients reported their level of benefit on a symptoms (range 1–24) associated with their headaches. 6-point scale, where 0 = no benefit and 5 = high Over 80% of patients reported that their headaches benefit. interfered with sleep, leisure and overall comfort. The mean weighted HIT score at baseline was 64.9 (range 48–78), corresponding to severe impact13. Results Table 1. Baseline demography (n = 61) Patient Disposition Gender Male 12 20.0% Female 48 80.0% Sixty-one patients from six UK GP practices (range Age group Under 30 10 16.7% 1–17 per centre) were recruited into the audit and 30 to 39 9 15.0% completed baseline assessments. In the investigation 40 to 49 14 23.3% period, 46 patients (75.4%) continued in the study to 1 50 to 59 21 35.0% 60 and over 6 10.0% month and 39 (63.9%) to 2 months. Employment status Retired 3 5.1% Sick/disabled 4 6.8% Housewife 3 5.1% Baseline Demography and Headache Part time 2 3.4% Severity Full time skilled 25 42.4% Full time semi-skilled 17 28.8% Table 1 shows the demography of the patients who Full time unskilled 2 3.4% took part in the study. The average age was 45.2 years Student 2 3.4% Unemployed 1 1.7% (range 21–68) and most patients (80%) were women.

Table 2. Severity of patients’ headaches: pain intensity, impact on daily activities and mood alterations

Headache severity Proportion of patients (%) Never Rarely Sometimes Very often Always Severe pain 2 0 16 61 21 Limit to usual activities 0 2 31 51 16 Desire to lie down 2 0 18 41 39 Too tired to work 5 8 51 31 5 Irritation 5 7 33 39 16 Lack of concentration 3 7 38 40 12

© 2005 LIBRAPHARM LTD – Headache Care 2005; 2(2) A Prospective Audit of Food Intolerance among Migraine Patients in Primary Care Clinical Practice Rees et al. 107 Identification of Food Intolerances had reactions to a total of 36 different foods, with an average of 4.7 (range 0–17) reactions per patient. Table 3 Food intolerances identified by IgG testing were analysed shows the distribution of food intolerances in these two for the 61 patients who took part in the study and for populations. The most frequently reported intolerances the 39 who completed the 2 months of investigation. In (in over 10% of patients in either population) were to the total study population, 60 of 61 patients (98.4%) had cow’s milk, yeast, egg white, egg yolk, wheat, gluten reactions to a total of 48 different foods, with an average (gliadin), corn, cashew nuts, mollusc mix, brazil nut, of 5.3 (range 0–17) reactions per patient. In the patients cranberry and garlic (Table 3), and were similar in who completed 2 months, 38 of 39 patients (97.4%) prevalence in the two populations.

Table 3. Food intolerances in the audit population: number and proportion of patients with a positive ELISA test to IgG from various foodstuffs

Food Positive ELISA test ( n [%]) Whole study population Patients completing 2 months ( n = 61) ( n = 39) Cow’s milk 52 (85.2%) 34 (87.2%) Yeast 37 (60.7%) 22 (56.4%) Egg white 34 (55.7%) 23 (59.0%) Egg yolk 20 (32.8%) 13 (33.3%) Wheat 19 (31.1%) 12 (30.8%) Gliadin 16 (26.2%) 10 (25.6%) Corn 15 (24.6%) 8 (20.5%) Cashew 12 (19.7%) 7 (17.9%) Mollusc mix 10 (16.4%) 3 (7.7%) Brazil nut 9 (14.8%) 6 (15.4%) Cranberry 7 (11.5%) 5 (12.8%) Garlic 5 (8.2%) 4 (10.3%) Beef 3 (4.9%) 2 (5.1%) Pork 3 (4.9%) 1 (2.6%) Ginger 3 (4.9%) 2 (5.1%) Buckwheat 4 (6.6%) 1 (2.6%) Crustacean mix 5 (8.2%) 1 (2.6%) Rye 2 (3.3%) 2 (5.1%) Millet 3 (4.9%) 2 (5.1%) Rice 1 (1.6%) 1 (2.6%) Soya bean 5 (8.2%) 3 (7.7%) Hazelnut 4 (6.6%) 3 (7.7%) Mustard seed 1 (1.6%) 1 (2.6%) Salmon/trout 2 (3.3%) 1 (2.6%) Plaice/sole 3 (4.9%) 1 (2.6%) Peanut 3 (4.9%) 2 (5.1%) Chicken 3 (4.9%) 1 (2.6%) Lentils 3 (4.9%) 1 (2.6%) Pea 2 (3.3%) 1 (2.6%) Almond 5 (8.2%) 3 (7.7%) Cola nut 3 (4.9%) 1 (2.6%) Duck 1 (1.6%) 0 Lamb 3 (4.9%) 1 (2.6%) Turkey 2 (3.3%) 0 White fish 3 (4.9%) 1 (2.6%) Kiwi 4 (6.6%) 2 (5.1%) Pineapple 2 (3.3%) 0 Sunflower seed 2 (3.3%) 0 Oat 2 (3.3%) 0 Haricot bean 3 (4.9%) 2 (5.1%) Coconut 1 (1.6%) 1 (2.6%) Tea 1 (1.6%) 0 Carrot 1 (1.6%) 0 Barley 1 (1.6%) 0 Tuna 1 (1.6%) 0 Sesame seed 1 (1.6%) 0 Coffee 1 (1.6%) 0 Avocado 1 (1.6%) 0

108 A Prospective Audit of Food Intolerance among Migraine Patients in Primary Care Clinical Practice © 2005 LIBRAPHARM LTD – Headache Care 2005; 2(2) Proportion of Patients who Altered their Diets migraine, reporting high levels of pain and impact on their everyday activities. This is a group of patients who are Of the 46 patients who returned the questionnaire after typically poorly managed in primary care15 and for whom 1 month of investigation, 41 (89.1%) patients changed new management initiatives would be welcome. their diets to eliminate foods identified by the IgG Almost all patients had multiple food intolerances testing and 5 (10.9%) did not. Of those who changed in this investigation, identified as positive food-specific their diet, 19 (46.3%) reported that they altered their IgG test results. Typically, individuals were positive to at diets a lot and 22 (53.7%) reported they had made a least one of cow’s milk, egg and yeast, together with a ‘reasonable attempt’ to avoid the specified foods. small number of more individual reactions. These results Of the 39 patients who returned the questionnaire are similar to those reported for other conditions11,12. Of after 2 months of investigation, 22 (56.4%) reported the patients who took part in the investigation, about that they altered their diets a lot and 13 (33.3%) 90% changed their diet to a greater or lesser extent at reported they had made a ‘reasonable attempt’ to avoid both 1 and 2 months. the specified foods. Two patients reported that they did A marked proportion of the migraine patients not change their diet at all. benefited from dietary intervention by cutting out foods for which they had an elevated IgG level. Approx- Benefit Obtained from Changing Diets imately 30% and 40% reported considerable benefit at 1 and 2 months, respectively. Reinforcing this is the fact Figures 1 and 2 show the level of benefit reported by that over 60% of patients who re-introduced the suspect patients after 1 and 2 months, respectively, using the 6- foods back into their diets reported the return of their point scale (0 = no benefit and 5 = high benefit). After migraine symptoms. These results are encouraging and 1 month, 27.5% of patients reported considerable benefit indicate that changing diet to counteract food intoler- (scoring 4 or 5), while 30.0% reported little or no benefit (scoring 0 or 1). Of 18 patients who had retried foods they had stopped taking, five (27.8%) reported a strong return of migraine symptoms and seven (38.9%) a slight return. After 25 22.5 2 months, 38.2% of patients reported considerable benefit 20.0 (scoring 4 or 5), while 32.4% reported little or no benefit 20 (scoring 0 or 1). Of 26 patients who had retried foods they 15.0 15 13.1 12.5 had stopped taking, seven (26.9%) reported a strong return 10.0 10 of migraine symptoms and 11 (42.3%) a slight return. % patients A limited post hoc analysis was conducted to investigate 5 the factors possibly associated with benefit. Of the 13 patients who reported considerable benefit from dieting 0 0 1 2 3 4 5 after 2 months, nine (69.2%) said they had dieted strictly No High after 1 month and 12 (92.3%) after 2 months. Of the 11 benefit benefit patients who reported little or no benefit after 2 months, Figure 1. Benefit of the diet after 1 month of the investigation: only two (18.2%) had dieted strictly after 1 month and proportion of patients reporting their level of benefit on a five (45.5%) after 2 months. Compared to those who did 6-point scale, where 0 = no benefit and 5 = high benefit not benefit, the patients who benefited were more likely to have suffered from bloating and sleep deprivation and to have never smoked (although all patients had given up at 25 least 10 years previously). Those who reported no benefit 20.6 20.6 20 17.6 17.6 from dieting were more likely to be trying other remedies 15 as well, including avoiding chocolate and taking sumatriptan 11.8 11.8 and homeopathic remedies. However, none of the above 10 differences was testable for statistical significance due to % patients the small number of patients involved. 5

0 0 1 2 3 4 5 Discussion No High benefit benefit

To our knowledge, this is the first investigation of possible Figure 2. Benefit of the diet after 2 months of the investigation: IgG-mediated food intolerances in migraine patients. The proportion of patients reporting their level of benefit on a 6- patients who took part were all severely affected by their point scale, where 0 = no benefit and 5 = high benefit

© 2005 LIBRAPHARM LTD – Headache Care 2005; 2(2) A Prospective Audit of Food Intolerance among Migraine Patients in Primary Care Clinical Practice Rees et al. 109 ances may be an effective treatment for at least some Acknowledgements migraine sufferers. However, it is not yet possible to recommend this The authors are pleased to acknowledge the help of Dr approach for general clinical use. This investigation was Frances Carter during the setting up and running of this a small audit to establish a possible relationship between audit. The audit was conducted by MIPCA, with help food intolerances and migraine. In this it was successful, and sponsorship from the Migraine Action Association although benefits experienced by patients may have and YORKTest Laboratories Ltd. been due (in part or in whole) to a placebo effect. There remains a series of questions that need to be answered before we have proof of this concept: References • Do migraine sufferers differ from unaffected people 1. Millichap JG, Yee MM. The diet factor in pediatric and adolescent migraine. Pediatr Neurol 2003;28:9-15 or people with other disorders in the pattern of IgG 2. Leira R, Rodriguez R. Diet and migraine. Rev Neurol 1996; that circulates? 24:534-8 • Do symptomatic reports of food intolerance correlate 3. Breslau N, Rasmussen BK. The impact of migraine: Epidemiology, risk factors, and co-morbidities. Neurology 2001;56(Suppl. 1): with the IgG data? 4-12 • Are migraine sufferers able to self-identify food 4. Peatfield RC. Relationships between food, wine, and beer- intolerances? precipitated migrainous headache. Headache 1995;35: 355-7 • Does allergen avoidance lead to an improvement 5. Bic Z, Blix GG, Hopp HP et al. The influence of a low-fat diet on in migraine and can this be confirmed by re- incidence and severity of migraine headaches. J Womens Health Gend Based Med 1999;8:623-30 challenge? 6. Hasselmark L, Malmgren R, Hannerz J. Effect of a carbohydrate- rich diet, low in protein-tryptophan, in classic and common We suggest two follow-up studies that may answer migraine. Cephalalgia 1987;7:87-92 7. Marcus DA, Scharff L, Turk D et al. A double-blind provocative these questions. Whether migraine patients differ from study of chocolate as a trigger of headache. Cephalalgia the general population and whether self-reported allergies 1997;17:855-62 8. Bedell AW, Cady RK, Diamond ML et al. Patient-centered correlate with food intolerances in migraine sufferers can strategies for effective management of migraine. Springfield, be examined in a blinded study investigating the pattern of Missouri: Primary Care Network, 2000 IgG-related food intolerances in migraine patients (with and 9. Pradalier A, Launay JM. Immunological aspects of migraine. Biomed Pharmacother 1996;50:64-70 without a history of allergy) and matched healthy controls 10. Kemper RH, Meijler WJ, Korf J et al. Migraine and function without migraine. A small placebo-controlled study can then of the immune system: a meta-analysis of clinical literature be used to study the effect of diet alteration on migraine published between 1966 and 1999. Cephalalgia 2001; 21:549-57 symptoms. The study requires a re-challenge phase, and 11. Sheldon TA. Independent audit of IgG food intolerance tested robust, validated endpoints, over a 3-month evaluation patient survey. British Allergy Foundation, 2000 time. 12. Atkinson W, Sheldon TA, Shaath N et al. Food elimination based on IgG antibodies in irritable bowel syndrome: a randomised In conclusion, this pilot audit demonstrated that controlled trial. Gut 2004;53:1459-64 migraine attacks may be related to food intolerances 13. Kosinski M, Bayliss MS, Bjorner JB et al. A six-item short-form survey for measuring headache impact: the HIT-6. Qual Life Res mediated via IgG and that changing the diet to eradicate 2003;12:963-74 specific foods may be a potentially effective treatment 14. YORKTEST guidebook on food intolerances and their treatment. for migraine. Further clinical studies are required York, UK: YORKTEST Laboratories Ltd, 2002 15. Lipton RB, Goadsby PJ, Sawyer JPC et al. Migraine: diagnosis to confirm these findings and examine the clinical and assessment of disability. Rev Contemp Pharmacother importance of this treatment approach. 2000;11:63-73

CrossRef links are available in the online published version of this paper: http://www.cmrojournal.com Paper HC-0047_2, Accepted for publication: 27 January 2005 Published Online: 18 February 2005 doi:10.1185/174234305X14962

110 A Prospective Audit of Food Intolerance among Migraine Patients in Primary Care Clinical Practice © 2005 LIBRAPHARM LTD – Headache Care 2005; 2(2) T h e new england journal o f medicine

review article

Medical Progress Celiac Disease

Peter H.R. Green, M.D., and Christophe Cellier, M.D., Ph.D.

eliac disease is a unique autoimmune disorder, unique because From the Department of Medicine, Colum- the environmental precipitant is known. The disorder was previously called bia University College of Physicians and Surgeons, New York (P.H.R.G.); and the celiac sprue, based on the Dutch word sprue, which was used to describe a dis- Department of Gastroenterology, Euro- C pean Georges Pompidou Hospital, René ease similar to tropical sprue that is characterized by diarrhea, emaciation, aphthous stomatitis, and malabsorption.1,2 Celiac disease is precipitated, in genetically pre- Descartes Paris V University, Assistance Publique–Hôpitaux de Paris, INSERM disposed persons, by the ingestion of gluten, the major storage protein of wheat U793, Paris (C.C.). Address reprint re- and similar grains.3 Originally considered a rare malabsorption syndrome of child- quests to Dr. Green at the Department of hood, celiac disease is now recognized as a common condition that may be diag- Medicine, Columbia University College of Physicians and Surgeons, 180 Fort nosed at any age and that affects many organ systems. The therapy for the disease is Washington Ave., Rm. 956, New York, NY a gluten-free diet; however, the response to therapy is poor in up to 30% of patients, 10032, or at [email protected]. and dietary nonadherence is the chief cause of persistent or recurrent symptoms. N Engl J Med 2007;357:1731-43. Small intestinal adenocarcinoma, refractory sprue, and enteropathy-associated T-cell Copyright © 2007 Massachusetts Medical Society. lymphoma are complications of celiac disease that must be ruled out when alarm- ing symptoms such as abdominal pain, diarrhea, and weight loss develop despite a strict gluten-free diet.

Pathogenesis

Celiac disease results from the interaction between gluten and immune, genetic, and environmental factors (Fig. 1).

The Role of Gluten Celiac disease is induced by the ingestion of gluten, which is derived from wheat, barley, and rye. The gluten protein is enriched in glutamine and proline and is poor- ly digested in the human upper gastrointestinal tract. The term “gluten” refers to the entire protein component of wheat; gliadin is the alcohol-soluble fraction of gluten that contains the bulk of the toxic components. Undigested molecules of gliadin, such as a peptide from an α-gliadin fraction made up of 33 amino acids, are resis- tant to degradation by gastric, pancreatic, and intestinal brush-border membrane proteases in the human intestine and thus remain in the intestinal lumen after gluten ingestion.4 These peptides pass through the epithelial barrier of the intes- tine, possibly during intestinal infections or when there is an increase in intestinal permeability, and interact with antigen-presenting cells in the lamina propria.

Mucosal Immune Responses In patients with celiac disease, immune responses to gliadin fractions promote an inflammatory reaction, primarily in the upper small intestine, characterized by infil- tration of the lamina propria and the epithelium with chronic inflammatory cells and villous atrophy (Fig. 1). This response is mediated by both the innate and the adaptive immune systems. The adaptive response is mediated by gliadin-reactive CD4+ T cells in the lamina propria that recognize gliadin peptides, which are bound

n engl j med 357;17 www.nejm.org october 25, 2007 1731

Downloaded from www.nejm.org by JUAN SABATER MD on November 14, 2007 . Copyright © 2007 Massachusetts Medical Society. All rights reserved. T h e new england journal o f medicine

Normal duodenal Duodenal mucosa Figure 1. Interaction of Gluten with Environmental, mucosa in celiac disease Immune,­ and Genetic Factors in Celiac Disease. Gluten is digested by luminal and brush-border enzymes into amino acids and peptides. The gliadin peptides in- duce changes in the epithelium through the innate im- mune system and, in the lamina propria, through the adaptive immune system. In the epithelium, gliadin damages epithelial cells, resulting in increased expres- sion of interleukin-15, which in turn activates intraepi- thelial lymphocytes. These lymphocytes become cyto- toxic and kill enterocytes that express MIC-A (a stress protein) on their surface. During infections or as the result of permeability changes, gliadin enters the lamina propria, where it is deamidated by tissue transglutamin- ase, allowing interaction with HLA-DQ2 (or HLA-DQ8) on the surface of antigen-presenting cells. Gliadin is presented to gliadin-reactive CD4+ T cells through a T-cell receptor, resulting in the production of cytokines that cause tissue damage. This leads to villous atrophy Gluten Other environmental and crypt hyperplasia, as well as the activation and ex- factors pansion of B cells that produce antibodies. Images of mucosa courtesy of Govind Bhagat, M.D. Gliadin Villous atrophy Increased intestinal Crypt hyperplasia Epithelial permeability Intraepithelial lymphocytosis absorptive proteinases and other tissue-damaging media- cells tors that induce crypt hyperplasia and villous in- jury.8 Gliadin peptides also activate an innate immune response in the intestinal epithelium that Overexpression of interleukin-15 is characterized by increased expression of inter- leukin-15 by enterocytes, resulting in the activa- tion of intraepithelial lymphocytes expressing the activating receptor NK-G2D, a natural-killer-cell Tissue marker.9 These activated cells become cytotoxic transglutaminase Intraepithelial and kill enterocytes with surface expression of lymphocytes major-histocompatibility-complex class I chain- related A (MIC-A), a cell-surface antigen induced Deamidated 10,11 gliadin by stress, such as an infection. The mecha- TCR CD4 Cytokine release T cell nism of the interaction between the processes in Lamina the epithelium and lamina propria has not been propria elucidated. HLA-DQ2 B cell Genetic Factors Antigen- Genetic The genetic influence in the pathogenesis of celiac factors presenting cell disease is indicated by its familial occurrence.12 Antibodies (antigliadin, antiendomysial, Celiac disease does not develop unless a person and tissue transglutaminase) has alleles that encode for HLA-DQ2 or HLA-DQ8 proteins, products of two of the HLA genes.13 However, many people, most of whom do not have to HLA class II molecules DQ2 or DQ8 on anti- celiac disease, carry these alleles; thus, their pres- gen-presenting cells; the T cells subsequently ence is necessary but not sufficient for the devel- 5 produce proinflammatory cytokines, particular­ opment of the disease.COLOR FIGURE Studies in siblings and 6 ly interferon-γ. Tissue transglutaminase is an en- identical twinsVersion 8suggest10/04/07 that the contribution of zyme in the intestine that deamidates gliadin HLA genesAuthor to theGreen genetic component of celiac dis- 7 Fig # 1 14 peptides, increasing their immunogenicity. The ease is lessTitle thanCeliac 50%. Disease Several non-HLA genes ensuing inflammatory cascade releases metallo- that mayME influenceIK susceptibility to the disease DE JRI Artist LAM AUTHOR PLEASE NOTE: Figure has been redrawn and type has been reset 1732 n engl j med 357;17 www.nejm.org october Please25, check2007 carefully Issue date 10/25/07

Downloaded from www.nejm.org by JUAN SABATER MD on November 14, 2007 . Copyright © 2007 Massachusetts Medical Society. All rights reserved. Medical Progress have been identified, but their influence has not more often diagnosed in women. The predomi- been confirmed. nance of the disease in women diminishes some- what after the age of 65 years.32 The classic pre- Environmental Factors sentation in adults is diarrhea, which may be Environmental factors that have an important role accompanied by abdominal pain or discomfort. in the development of celiac disease have been However, diarrhea has been the main presenting suggested by epidemiologic studies. These include symptom in less than 50% of cases in the past a protective effect of breast-feeding15 and the intro­ decade.33 Other, silent presentations in adults in- duction of gluten in relation to weaning.16-18 The clude iron-deficiency anemia, osteoporosis, and initial administration of gluten before 4 months incidental recognition at endoscopy performed of age is associated with an increased risk of dis- for other reasons, such as symptoms of gastro- ease development,18 and the introduction of glu- esophageal reflux.34 Less common presentations ten after 7 months is associated with a marginal include abdominal pain, constipation, weight risk.18 However, the overlap of gluten introduction loss, neurologic symptoms, dermatitis herpeti- with breast-feeding may be a more important pro- formis, hypoproteinemia, hypocalcemia, and ele- tective factor in minimizing the risk of celiac dis- vated liver enzyme levels.35 Substantial proportions ease.16 The occurrence of certain gastrointestinal of patients have received a previous diagnosis of infections, such as rotaviral infection, also increas- the irritable bowel syndrome32,36 and are over- es the risk of celiac disease in infancy.19 Further weight.37 Patients often have symptoms for a study of environmental factors might facilitate long time and undergo multiple hospitalizations the development of strategies for primary pre- and surgical procedures before celiac disease is vention of celiac disease.20 diagnosed.32,38,39 Some cases are diagnosed because of increased Epidemiology surveillance for celiac disease among people who have a family history of the disease24 and among Celiac disease occurs in adults and children at those with Down syndrome, Turner’s syndrome,40,41 rates approaching 1% of the population.21-25 The or type 1 diabetes, all of which are associated disease is recognized not only throughout Europe with celiac disease.42 Persons with celiac disease and in countries populated by persons of Euro- have an increased risk of autoimmune disorders pean ancestry but also in the Middle East,23,26 as compared with the general population.43-45 Asia,27 South America.28 and North Africa.29 In A case-finding study performed in multiple most affected people, celiac disease remains un- primary care practices in North America reported diagnosed,21 although the rate of diagnosis is a 43-fold increase in the rate of diagnosis of ce- increasing.30 liac disease over the 2 years of the study.46 The indications for screening in those who received Clinical Manifestations the diagnosis included bloating, the irritable bowel syndrome, thyroid disease, chronic unexplained Clinical manifestations of celiac disease vary diarrhea, chronic fatigue, and constipation. The greatly according to age group. Infants and young case-finding study shows that many patients with children generally present with diarrhea, abdom- celiac disease are seeking health care for a great inal distention, and failure to thrive. However, variety of common symptoms and that the more vomiting, irritability, anorexia, and even constipa- frequent use of screening is uncovering more tion are also common. Older children and adoles- cases of the disease. cents often present with extraintestinal manifesta­ tions, such as short stature, neurologic symptoms, Diagnosis or anemia.31 Among adults, two to three times as many The diagnosis of celiac disease requires both a duo­ women have the disease as men, for unknown denal biopsy that shows the characteristic find- reasons. In general, the prevalence of autoimmune ings of intraepithelial lymphocytosis, crypt hyper- diseases is higher in women than in men, and plasia, and villous atrophy and a positive response iron deficiency and osteoporosis, each of which to a gluten-free diet. The diagnostic criteria devel- prompts an assessment for celiac disease, are oped by the European Society for Pediatric Gastro-

n engl j med 357;17 www.nejm.org october 25, 2007 1733

Downloaded from www.nejm.org by JUAN SABATER MD on November 14, 2007 . Copyright © 2007 Massachusetts Medical Society. All rights reserved. T h e new england journal o f medicine

enterology and Nutrition require only clinical im- sive than the endomysial antibody test. Overall, provement with the diet,47 although histologic the sensitivity of the tests for both endomysial improvement on a gluten-free diet is frequently antibodies and anti–tissue transglutaminase anti­ sought and is recommended in adults because vil- bodies is greater than 90%,48 and a test for ei- lous atrophy may persist despite a clinical response ther marker is considered the best means of to the diet. In most patients, the diagnosis is eas- screening for celiac disease.48 The titers of endo- ily established. However, roughly 10% of cases are mysial antibodies and anti–tissue transglutamin- difficult to diagnose because of a lack of concor- ase antibodies correlate with the degree of mu­ dance among serologic, clinical, and histologic cosal damage51,52; as a result, the sensitivity of findings. these antibody tests declines when a greater number of patients with lesser degrees of vil- Serologic Testing lous atrophy are included in studies.53,54 The various commercially available assays for anti– Typical indications for serologic testing include tissue transglutaminase antibodies have differ- unexplained bloating or abdominal distress; chron­ ent characteristics and resultant sensitivities and ic diarrhea, with or without malabsorption or the specificities.55 irritable bowel syndrome; abnormalities on labo- Selective IgA deficiency is more common in ratory tests that might be caused by malabsorp- patients with celiac disease than in the general tion (e.g., folate deficiency and iron-deficiency population — 1 case in 40 as compared with 1 in anemia); first-degree relatives with celiac disease; 400. Consequently, patients with celiac disease and autoimmune diseases and other conditions and selective IgA deficiency lack IgA endomysial known to be associated with celiac disease (for antibodies and IgA antitissue antibodies against more information on indications for serologic test­ tissue transglutaminase. It is recommended that ing, see the table in the Supplementary Appendix, the test for anti–tissue transglutaminase anti- available with the full text of this article at www. bodies be used as a single screening test for ce- nejm.org). liac disease.48,56 If the levels of this marker are The most sensitive antibody tests for the diag- within the normal range (or if it is absent) and nosis of celiac disease are of the IgA class. The there is a high suspicion of celiac disease, selec- available tests include those for antigliadin anti- tive IgA deficiency needs to be ruled by measur- bodies, connective-tissue antibodies (antireticulin ing total IgA levels. In such cases, a test for IgG and antiendomysial antibodies), and antibodies antibodies against tissue transglutaminase should directed against tissue transglutaminase, the en- be performed.57 zyme responsible for the deamidation of gliadin These antibody tests fare less well in the clini- in the lamina propria. The antigliadin antibodies cal-practice setting than in the research setting.58,59 are no longer considered sensitive enough or spe- A recently developed rapid test for anti–tissue cific enough to be used for the detection of ce- transglutaminase antibodies that uses a sample liac disease, except in children younger than 18 of fingertip blood may be a convenient point-of- months of age,48 although new-generation anti- care test for the purpose of both case finding bodies to deamidated gliadin peptides appear to and dietary monitoring.60 be promising.49 Antireticulin antibodies are also rarely measured, having been surpassed in use by The Role of HL A-DQ2 endomysial and anti–tissue transglutaminase anti­ and HL A-DQ8 Assessment bodies. The diagnostic standard in celiac serologies re­ The HLA-DQ2 allele is identified in 90 to 95% of mains the endomysial IgA antibodies; they are patients with celiac disease, and HLA-DQ8 is iden- highly specific markers for celiac disease, ap- tified in most of the remaining patients.61 Be- proaching 100% accuracy. The recognition that cause these alleles occur in 30 to 40% of the gen- the enzyme tissue transglutaminase is the auto- eral population (with HLA-DQ2 more common than antigen for the development of endomysial anti- HLA-DQ8), the absence of these alleles is impor- bodies50 allowed development of automated tant for its high negative predictive value.62 Thus, enzyme-linked immunoassays that are less expen- the presence or absence of HLA-DQ2 and HLA-DQ8

1734 n engl j med 357;17 www.nejm.org october 25, 2007

Downloaded from www.nejm.org by JUAN SABATER MD on November 14, 2007 . Copyright © 2007 Massachusetts Medical Society. All rights reserved. Medical Progress is important for determining which family mem- Table 1. Causes of Villous Atrophy Other Than Celiac Disease. bers should be screened with serologic testing and is useful for ruling out the disease in patients Giardiasis already on a gluten-free diet or for patients in Collagenous sprue whom the diagnosis is unclear. Common-variable immunodeficiency Autoimmune enteropathy Radiation enteritis Biopsy and Histologic Assessment Whipple’s disease Tuberculosis Biopsy of the small intestine remains the stan- Tropical sprue dard for diagnosing celiac disease, and it should Eosinophilic gastroenteritis always be performed when clinical suspicion is Human immunodeficiency virus enteropathy high, irrespective of the results of serologic test- Intestinal lymphoma ing. Biopsy confirmation is crucial, given the life- Zollinger–Ellison syndrome long nature of the disease and the attendant need Crohn’s disease for an expensive and socially inconvenient diet. Intolerance of foods other than gluten (e.g., milk, soy, chicken, tuna) Although no studies have examined the number of biopsies required for diagnosis, we believe that at least four to six endoscopic-biopsy speci- mens should be obtained from the duodenum, Table 2. Fundamentals of the Gluten-free Diet. given the patchy nature of the disease and the Grains that should be avoided difficulty of orienting the small pieces of tissue Wheat (includes spelt, kamut, semolina, triticale), rye, barley (including malt) taken during biopsy for assessment of villous Safe grains (gluten-free) 63,64 morphology. Rice, amaranth, buckwheat, corn, millet, quinoa, sorghum, teff (an Ethiopian Who should undergo endoscopic biopsy? In cereal grain), oats addition to patients whose serologic tests are Sources of gluten-free starches that can be used as flour alternatives positive, any patient who has chronic diarrhea, Cereal grains: amaranth, buckwheat, corn (polenta), millet, quinoa, sorghum, iron deficiency, or weight loss should undergo teff, rice (white, brown, wild, basmati, jasmine), montina (Indian rice grass) duodenal biopsy, irrespective of whether serolog­ Tubers: arrowroot, jicama, taro, potato, tapioca (cassava, manioc, yucca) ic testing for celiac disease has been performed. Legumes: chickpeas, lentils, kidney beans, navy beans, pea beans, peanuts, The recognition of endoscopic signs of villous soybeans atrophy, such as scalloping of mucosal folds, Nuts: almonds, walnuts, chestnuts, hazelnuts, cashews Seeds: sunflower, flax, pumpkin absent or reduced duodenal folds, or a mosaic pattern of the mucosa, should prompt biopsy.65 However, because these abnormalities are not sensitive markers of the presence of celiac dis- nosis is confirmed when there is a favorable re- ease,66 biopsy should be performed even if they sponse to the diet. are absent. The spectrum of pathologic changes in celiac Treatment disease ranges from near-normal villous architec- ture with a prominent intraepithelial lymphocy- Nutritional therapy, the only accepted treatment tosis to total villous atrophy.67 Pitfalls in the for celiac disease, involves the lifelong elimination pathological diagnosis include overinterpretation of wheat, rye, and barley from the diet. Clinical of villous atrophy in poorly oriented biopsy speci- studies suggest that oats are tolerated by most mens and inadequate biopsy sampling in patients patients with celiac disease and may improve the with patchy villous atrophy.63,64 The histologic nutritional content of the diet and overall quality findings in celiac disease are characteristic but of life.69 However, oats are not uniformly recom- not specific68; their presence permits a presump- mended, because most commercially available tive diagnosis of celiac disease and initiation of oats are contaminated with gluten-containing a gluten-free diet. Indeed, celiac disease is not the grains during the growing, transportation, and only cause of villous atrophy (Table 1). The diag- milling processes.70

n engl j med 357;17 www.nejm.org october 25, 2007 1735

Downloaded from www.nejm.org by JUAN SABATER MD on November 14, 2007 . Copyright © 2007 Massachusetts Medical Society. All rights reserved. T h e new england journal o f medicine

Although wheat, rye, and barley should be terfere with the immune response — by blocking avoided, there are other grains that can serve as the binding of deaminated gliadin to HLA-DQ2 substitutes as well as other sources of starch that or HLA-DQ8, for example, or by blocking the ac- can provide flours for cooking and baking (Ta- tion of tissue transglutaminase — are unlikely to ble 2). Because the substitute flours are not forti- be without side effects. fied with B vitamins, vitamin deficiencies may occur; they have been detected in patients who are Assessment of Cases with on the diet for a long time (more than 10 years).71 a Poor Response to Therapy Therefore, vitamin supplementation is advised. Meats, dairy products, and fruits and vegetables A gluten-free diet fails to induce clinical or histo- are naturally gluten-free and help to make for a logic improvement in 7 to 30% of patients,78 and more nutritious and varied diet. such lack of response should trigger a systematic After the diagnosis of celiac disease has been evaluation (Fig. 2). The first step is to reassess the established, the patient should be assessed for initial diagnosis, since villous atrophy with asso- deficiencies of vitamins and minerals, including ciated crypt hyperplasia is not exclusive to celiac folic acid, B12, fat-soluble vitamins, iron, and cal- disease (Table 1). Rare causes should at least be cium, and any such deficiencies should be treated. considered in people who do not have the expected All patients with celiac disease should undergo response to the diet. In patients with a question- screening for osteoporosis, which has a high able diagnosis, HLA-DQ2 or HLA-DQ8 typing may prevalence in this population.72,73 The health care be useful, since the negative predictive value of team should include a skilled dietitian who mon­ this test is almost 100%.62 itors the patient’s nutritional status and dietary The second step is to address the likelihood adherence on a regular basis. In children, ongo- of dietary nonadherence, the most common cause ing evaluation includes monitoring of growth and of unresponsive celiac disease. An experienced development. dietitian is required to assess the degree of ad- The elimination of gluten usually induces herence and possible reasons for nonadherence clinical improvement within days or weeks, though (Table 3). The highest rates of adherence are re- histologic recovery takes months or even years, ported among patients with a diagnosis in child- especially in adults, in whom mucosal recovery hood and those with severe symptoms at presen- may be incomplete.74 In rare cases, children toler- tation. In France and Belgium, less than half of ate the reintroduction of a normal diet after a adults with celiac disease who were studied ad- long-term clinical and histologic response.75 hered strictly to the diet for more than a year after Patient-support organizations are a valuable diagnosis.79 In a study in the United Kingdom, source of information about the disease and the the rate of adherence was low for both teenagers diet. Most countries have national support groups and adults,80 and in a study in Italy, adolescents in that are easily accessed on the Internet. The cost whom the diagnosis was established on the basis of the gluten-free products varies by country, but of mass serologic screening had poor adherence.81 the diet is usually expensive, making dietary treat- In another study, many people in whom the dis- ment problematic for patients with limited finan- ease was diagnosed in childhood became non- cial resources. Gluten-free products are particu- adherent to a strict gluten-free diet as adults.82 larly expensive and hard to find in developing The persistence of endomysial antibodies or anti– countries, whereas in other countries (including tissue transglutaminase antibodies in patients on the Netherlands, the United Kingdom, New Zea- a gluten-free diet for a year or more is suggestive land, Italy, Sweden, and Finland), the government of poor dietary adherence.79 Other causes of per- subsidizes these products. sistent symptoms in patients on a strict gluten- There is considerable interest in the develop- free diet are listed in Table 3. ment of nondietary therapies that might either replace or supplement the rigorous gluten-free Complications of Celiac Disease diet. Currently, the most attractive alternative in- volves the use of recombinant enzymes that di- Although the majority of patients who have re- gest the toxic gliadin fractions in the stomach or current or new symptoms when on a supposedly the upper small intestine.76,77 Therapies that in- gluten-free diet are in fact ingesting gluten, either

1736 n engl j med 357;17 www.nejm.org october 25, 2007

Downloaded from www.nejm.org by JUAN SABATER MD on November 14, 2007 . Copyright © 2007 Massachusetts Medical Society. All rights reserved. Medical Progress

Poorly responsive celiac disease

Exclusion of other causes of villous atrophy

Assessment of adherence to gluten-free diet

Good adherence Poor adherence

Evaluation Dietary counseling Exclusion of bacterial overgrowth Referral to support group and pancreatic insufficiency Regular follow-up by experienced Upper gastrointestinal endoscopy dietitian or colonoscopy and biopsy Intraepithelial-lymphocyte pheno- type on duodenal biopsy Enteroscopy (push or push–pull) Abdominal radiology (small bowel studies and CT scan) Video-capsule endoscopy

Enteropathy-associated Refractory celiac Adenocarcinoma T-cell lymphoma disease

Normal Abnormal clonal intraepithelial-lymphocyte intraepithelial-lymphocyte phenotype (type 1) phenotype (type 2)

Figure 2. An Assessment Plan for a Patient with Poorly Responsive Celiac Disease.

RETAKE 1st ICM AUTHOR: Green 2nd REG F FIGURE: 2 of 3 3rd intentionally or unintentionally,CASE a substantial pro- either intestinalRevised or extraintestinal; oropharyngeal portion may have a serious complicationEMail of celiac Lineand esophageal4-C SIZE adenocarcinoma; and cancers of ARTIST: ts H/T H/T 33p9 disease: intestinal adenocarcinoma,Enon enteropathy- Combothe small and large intestine, hepatobiliary sys- 83,84 86,87 associated T-cell lymphoma, or refractory sprue.AUTHOR, PLEASEtem, NOTE: and pancreas. The risk of breast cancer, Figure has been redrawn andhowever, type has beenappears reset. to be reduced.85,86 Adenocarcinoma of the Small IntestinePlease check carefully.In patients with celiac disease, the risk of ad- Patients with celiac diseaseJOB: have35717 an overall risk enocarcinomaISSUE: 10-25-07of the small intestine, generally a of cancer that is almost twice that in the general rare cancer, is increased manyfold as compared population. New population-based studies dem- with the risk in the general population88; still, onstrate that the risk is not as great as once con- the overall risk is very low given the rarity of this sidered.85 Reported cancers include both T-cell cancer. These carcinomas are most often located and B-cell non-Hodgkin’s lymphoma that may be in the jejunum and are more likely to develop as an

n engl j med 357;17 www.nejm.org october 25, 2007 1737

Downloaded from www.nejm.org by JUAN SABATER MD on November 14, 2007 . Copyright © 2007 Massachusetts Medical Society. All rights reserved. T h e new england journal o f medicine

toms of celiac disease after a period of good Table 3. Problems of Dietary Adherence and Poor Response in Celiac Disease. response to gluten withdrawal. Enteropathy-asso- Reasons for poor adherence to a gluten-free diet ciated T-cell lymphoma usually develops in the High cost jejunum but may also be found in the ileum or in Poor availability of gluten-free products (in developing countries) extraintestinal sites (e.g., liver, brain, chest, and Poor palatability bone) and is often multifocal. The prognosis is Absence of symptoms when dietary restrictions not observed poor; less than 20% of patients survive for 30 Inadequate information on gluten content of food or drugs months.91 The phenotype of enteropathy-asso­ Inadequate dietary counseling ciated T-cell lymphoma is consistent with a tu- Inadequate initial information supplied by diagnosing physician mor that derives from a clonal proliferation of in­ Inadequate medical or nutritional follow-up traepithelial lymphocytes. Immunohistochemical Lack of participation in a support group phenotyping indicates that this lesion is most Inaccurate information from physicians, dietitians, support groups, or Internet commonly CD3+, CD4–, CD8–, CD30+, and Dining out of the home CD103+.92 The treatment of this tumor is chemo- Social, cultural, or peer pressures therapy based, although there is a role for sur- Transition to adolescence gery in the treatment of localized or complicated Inadequate medical follow-up after childhood tumors.93 Successful autologous stem-cell trans- Causes of poorly responsive celiac disease plantation has been reported.94 Incorrect diagnosis Gluten ingestion (intentional or unintentional) Refractory Celiac Disease Microscopical colitis Approximately 5% of patients may have refractory Lactose intolerance celiac disease, defined as persistent symptoms Pancreatic insufficiency and villous atrophy despite scrupulous adherence Bacterial overgrowth to a gluten-free diet.95 The symptoms that usually Intolerance of foods other than gluten (e.g., fructose, milk, soy) develop in these patients include diarrhea, weight Inflammatory bowel disease loss, recurrence of malabsorption, abdominal pain, Irritable bowel syndrome bleeding, and anemia, and ulcerative jejunitis often Anal incontinence arises as well. This syndrome is also known as Collagenous sprue refractory sprue. The term was originally coined Autoimmune enteropathy because it was unclear whether patients with diar- Refractory celiac disease (with or without clonal T cells) rhea and villous atrophy whose condition did not Enteropathy-associated T-cell lymphoma improve with a gluten-free diet actually had celiac disease.2 adenoma–carcinoma sequence than as dysplasia Refractory celiac disease may be classified as in flat mucosa.89 Intuitively, video-capsule endos- type 1, in which there is a normal intraepithelial copy, which allows for visualization of the entire lymphocyte phenotype, or type 2, in which there small intestinal mucosal surface, would seem to is a clonal expansion of an aberrant intraepithe- be ideal in screening for cancers, but as yet no lial lymphocyte population. The identification of data support this approach. At present, video- an aberrant clonal population is primarily prog- capsule endoscopy plays a role in the evaluation nostic, since it is associated with a high risk of of celiac disease that is complicated by the devel- ulcerative jejunitis and frank enteropathy-associ- opment of abdominal pain or occult bleeding.90 ated T-cell lymphoma; the risk is so high that this condition has been described as a cryptic T-cell Enteropathy-Associated T-Cell Lymphoma lymphoma.83,96,97 The intraepithelial lymphocyte Enteropathy-associated T-cell lymphoma occurs expansion may be driven by overexpression of in adults, with the incidence peaking in the sixth interleukin-15 by the epithelium.9-11 Specific im- decade of life, and is usually at an advanced stage munophenotypic changes in the intraepithelial at diagnosis. Symptoms may include malaise, an- lymphocytes are seen in type 2 refractory celiac orexia, weight loss, diarrhea, abdominal pain, and disease. Intraepithelial lymphocytes in active ce- unexplained fever. The development of lympho- liac disease exhibit surface expression of CD3, ma is usually indicated by clinical relapse of symp- CD8, T-cell receptor (TCR) αβ, and TCRγδ, where­

1738 n engl j med 357;17 www.nejm.org october 25, 2007

Downloaded from www.nejm.org by JUAN SABATER MD on November 14, 2007 . Copyright © 2007 Massachusetts Medical Society. All rights reserved. Medical Progress

A

CD3+ +

Active Celiac Disease CD8+ +

B

CD3+ +

Refractory CD8– Celiac Disease -

Figure 3. Phenotyping Intraepithelial Lymphocytes in Active versus Refractory Celiac Disease.

In active celiac disease (Panel A, hematoxylinICM AUTHOR and eosin),green most intraepithelialRETAKE lymphocytes1st express CD3 and CD8 (CD3+,CD8+) receptors, whereas inREG refractory F FIGURE celiac f3disease a_b (Panel B, hematoxylin 2ndand eosin), most of these lym- phocytes express CD3 but not CD8 CASE(CD3+,CD8−). In insets, brown color denotes positive3rd immunostaining. Images TITLE Revised provided by Dr. Diane Damotte, DepartmentEMail of Pathology, EuropeanLine 4-C Georges Pompidou Hospital. SIZE Enon ARTIST: mleahy H/T H/T FILL Combo 33p9 AUTHOR, PLEASE NOTE: Figure has been redrawn and type has been reset. Please check carefully.

n englJOB: j med 35717 357;17 www.nejm.orgISSUE: october10-25-07 25, 2007 1739

Downloaded from www.nejm.org by JUAN SABATER MD on November 14, 2007 . Copyright © 2007 Massachusetts Medical Society. All rights reserved. T h e new england journal o f medicine

as intraepithelial lymphocytes in type 2 (clonal) intraepithelial lymphocytes and progression to an refractory sprue continue to express CD3 in the overt lymphoma have been reported in some treat­ cytoplasm but lack surface expression of CD8, ed patients, indicating that these drugs do not CD3, TCRαβ, and TCRγδ.92,96,98,99 Immunohis- cure the disease.106,107 Autologous hematopoietic tochemical studies demonstrating the presence stem-cell transplantation has been successful.108 of these cell-surface markers can be performed New therapeutic strategies, such as blocking on formalin-fixed biopsy specimens in most clin­ interleukin-15, should be aggressively investigat­ ical pathology laboratories. They allow for differ- ed, since the prognosis for patients with clonal entiation of type 2 refractory celiac disease, in refractory disease is poor, with a 5-year survival which an abnormal phenotype of the intraepi- rate of less than 50%.83 thelial lymphocytes (CD3+,CD8–), from type 1, in which there is a normal phenotype (CD3+,CD8+). Summary Patients who are not adherent to the diet will also express this normal phenotype (Fig. 3).100 Celiac disease occurs in nearly 1% of the popula- The development of new symptoms (e.g., weight tion in many countries. The diagnosis, which is loss, abdominal pain, or fever) or the recurrence straightforward in most cases, is usually estab- of diarrhea in patients who are on a strict gluten- lished on the basis of serologic testing, duodenal free diet often requires extensive investigation, biopsy, and observation of the response to a gluten- including the use of contrast radiology or com- free diet. A poor response to the diet is common puted tomographic (CT) enteroclysis (in which and requires extensive evaluation to rule out in- CT imaging is performed after infusion of a large testinal lymphoma and refractory sprue, compli- volume of contrast material — often 2 liters — cations that arise as the result of clonal expansion into the small intestine), video-capsule endoscopy, of intraepithelial lymphocytes. positron-emission tomographic scanning, extend­ Increasing awareness of the epidemiology and ed upper endoscopy (so-called push or push–pull diverse manifestations of the disease, as well as enteroscopy), and laparoscopy.90,101 the availability of sensitive and specific serologic Treatment of refractory celiac disease involves tests, especially among primary care physicians, nutritional support and repletion of vitamins and will lead to more widespread screening and diag- minerals, together with a strict gluten-free diet. nosis, which in turn will lead to greater availabil- In most cases, corticosteroids induce clinical im- ity of gluten-free foods and efforts to develop provement.83 Immunosuppressive drugs may be drug therapies that relieve patients of the burden beneficial102,103 but should be used with caution, of a gluten-free diet. In addition, earlier diagno- since they may promote the progression to lym- sis may lead to a reduction in the complications phoma.104 of the disease. The successful use of infliximab,105 an anti- CD52 monoclonal antibody, and cladribine (2-clo- Dr. Green reports receiving a consulting fee from Alvine Pharmaceuticals and lecture fees from Prometheus Laboratories. rodeoxyadenosine) has been reported, although No other potential conflict of interest relevant to this article was the persistence of clonally expanded, aberrant reported.

References 1. Collins JR, Isselbacher KJ. Treatment al. Gluten induces an intestinal cytokine 9. Mention JJ, Ben Ahmed M, Bègue B, et of adult celiac disease (nontropical sprue). response strongly dominated by interferon al. Interleukin 15: a key to disrupted intra­ N Engl J Med 1964;271:1153-6. gamma in patients with celiac disease. epithelial lymphocyte homeostasis and 2. Trier JS, Falchuk ZM, Carey MC, Schrei­ Gastroenterology 1998;115:551-63. lymphomagenesis in celiac disease. Gas- ber DS. Celiac sprue and refractory sprue. 7. Molberg O, McAdam SN, Körner R, et troenterology 2003;125:730-45. Gastroenterology 1978;75:307-16. al. Tissue transglutaminase selectively 10. Meresse B, Chen Z, Ciszewski C, et al. 3. Green PH, Jabri B. Coeliac disease. modifies gliadin peptides that are recog- Coordinated induction by IL15 of a TCR- Lancet 2003;362:383-91. nized by gut-derived T cells in celiac dis- independent NKG2D signaling pathway 4. Shan L, Molberg Ø, Parrot I, et al. ease. Nat Med 1998;4:713-7. [Erratum, Nat converts CTL into lymphokine-activated Structural basis for gluten intolerance in Med 1998;4:974.] killer cells in celiac disease. Immunity celiac sprue. Science 2002;297:2275-9. 8. Mohamed BM, Feighery C, Kelly J, et 2004;21:357-66. 5. Sollid LM. Coeliac disease: dissecting al. Increased protein expression of matrix 11. Hüe S, Mention JJ, Monteiro RC, et al. a complex inflammatory disorder. Nat Rev metalloproteinases -1, -3, and -9 and A direct role for NKG2D/MICA interaction Immunol 2002;2:647-55. TIMP-1 in patients with gluten-sensitive in villous atrophy during celiac disease. 6. Nilsen EM, Jahnsen FL, Lundin KE, et enteropathy. Dig Dis Sci 2006;51:1862-8. Immunity 2004;21:367-77.

1740 n engl j med 357;17 www.nejm.org october 25, 2007

Downloaded from www.nejm.org by JUAN SABATER MD on November 14, 2007 . Copyright © 2007 Massachusetts Medical Society. All rights reserved. Medical Progress

12. Bevan S, Popat S, Braegger CP, et al. 28. Gomez JC, Selvaggio GS, Viola M, et al. based population. Postgrad Med J 2007; Contribution of the MHC region to the Prevalence of celiac disease in Argentina: 83:132-6. familial risk of coeliac disease. J Med screening of an adult population in the La 43. Ventura A, Magazzù G, Greco L. Dura- Genet 1999;36:687-90. Plata area. Am J Gastroenterol 2001;96: tion of exposure to gluten and risk for 13. Sollid LM, Lie BA. Celiac disease ge- 2700-4. autoimmune disorders in patients with netics: current concepts and practical 29. Catassi C, Rätsch IM, Gandolfi L, et al. celiac disease. Gastroenterology 1999;117: applications. Clin Gastroenterol Hepatol Why is coeliac disease endemic in the peo- 297-303. 2005;3:843-51. ple of the Sahara? Lancet 1999;354:647-8. 44. Bai D, Brar P, Holleran S, Ramakrish- 14. Greco L, Romino R, Coto I, et al. The 30. Murray JA, Van Dyke C, Plevak MF, nan R, Green PH. Effect of gender on the first large population based twin study of Dierkhising RA, Zinsmeister AR, Melton manifestations of celiac disease: evidence coeliac disease. Gut 2002;50:624-8. LJ III. Trends in the identification and for greater malabsorption in men. Scand 15. Persson LA, Ivarsson A, Hernell O. clinical features of celiac disease in a J Gastroenterol 2005;40:183-7. Breast-feeding protects against celiac dis- North American community, 1950-2001. 45. Viljamaa M, Kaukinen K, Huhtala H, ease in childhood — epidemiological evi- Clin Gastroenterol Hepatol 2003;1:19-27. Kyronpalo S, Rasmussen M, Collin P. Co- dence. Adv Exp Med Biol 2002;503:115-23. 31. D’Amico MA, Holmes J, Stavropoulos eliac disease, autoimmune diseases and 16. Ivarsson A, Hernell O, Stenlund H, SN, et al. Presentation of pediatric celiac gluten exposure. Scand J Gastroenterol Persson LA. Breast-feeding protects against disease in the United States: prominent ef- 2005;40:437-43. celiac disease. Am J Clin Nutr 2002;75: fect of breastfeeding. Clin Pediatr (Phila) 46. Catassi C, Kryszak D, Louis-Jacques O, 914-21. 2005;44:249-58. et al. Detection of celiac disease in primary 17. Carlsson A, Agardh D, Borulf S, Grod­ 32. Green PHR, Stavropoulos SN, Panagi care: a multicenter case-finding study in zinsky E, Axelsson I, Ivarsson SA. Preva- SG, et al. Characteristics of adult celiac North America. Am J Gastroenterol 2007; lence of celiac disease: before and after a disease in the USA: results of a national 102:1454-60. national change in feeding recommen- survey. Am J Gastroenterol 2001;96:126- 47. Revised criteria for diagnosis of coe- dations. Scand J Gastroenterol 2006;41: 31. liac disease: report of Working Group of 553-8. 33. Rampertab SD, Pooran N, Brar P, European Society of Paediatric Gastroen- 18. Norris JM, Barriga K, Hoffenberg EJ, Singh P, Green PH. Trends in the presen- terology and Nutrition. Arch Dis Child et al. Risk of celiac disease autoimmunity tation of celiac disease. Am J Med 2006; 1990;65:909-11. and timing of gluten introduction in the 119(4):355.e9-355.e14. 48. Rostom A, Dubé C, Cranney A, et al. diet of infants at increased risk of disease. 34. Green PH, Shane E, Rotterdam H, The diagnostic accuracy of serologic tests JAMA 2005;293:2343-51. Forde KA, Grossbard L. Significance of un­ for celiac disease: a systematic review. Gas­ 19. Stene LC, Honeyman MC, Hoffenberg suspected celiac disease detected at endos- troenterology 2005;128:Suppl 1:S38-S46. EJ, et al. Rotavirus infection frequency and copy. Gastrointest Endosc 2000;51:60-5. 49. Sugai E, Vázquez H, Nachman F, et al. risk of celiac disease autoimmunity in 35. Green PH. The many faces of celiac Accuracy of testing for antibodies to syn- early childhood: a longitudinal study. Am disease: clinical presentation of celiac dis- thetic gliadin-related peptides in celiac dis­ J Gastroenterol 2006;101:2333-40. ease in the adult population. Gastroenter- ease. Clin Gastroenterol Hepatol 2006;4: 20. Ivarsson A. The Swedish epidemic of ology 2005;128:Suppl 1:S74-S78. 1112-7. coeliac disease explored using an epide- 36. Sanders DS, Carter MJ, Hurlstone DP, 50. Dieterich W, Ehnis T, Bauer M, et al. miological approach — some lessons to et al. Association of adult coeliac disease Identification of tissue transglutaminase be learnt. Best Pract Res Clin Gastroen- with irritable bowel syndrome: a case- as the autoantigen of celiac disease. Nat terol 2005;19:425-40. control study in patients fulfilling ROME Med 1997;3:797-801. 21. West J, Logan RF, Hill PG, et al. Sero- II criteria referred to secondary care. Lan- 51. Sategna-Guidetti C, Pulitano R, Gros- prevalence, correlates, and characteristics cet 2001;358:1504-8. so S, Ferfoglia G. Serum IgA antiendomy- of undetected coeliac disease in England. 37. Dickey W, Kearney N. Overweight in sium antibody titers as a marker of intes- Gut 2003;52:960-5. celiac disease: prevalence, clinical charac- tinal involvement and diet compliance in 22. Mäki M, Mustalahti K, Kokkonen J, et teristics, and effect of a gluten-free diet. adult celiac sprue. J Clin Gastroenterol al. Prevalence of celiac disease among chil- Am J Gastroenterol 2006;101:2356-9. 1993;17:123-7. dren in Finland. N Engl J Med 2003;348: 38. Dickey W, McConnell JB. How many 52. Tursi A, Brandimarte G, Giorgetti GM. 2517-24. hospital visits does it take before celiac Prevalence of antitissue transglutaminase 23. Tatar G, Elsurer R, Simsek H, et al. sprue is diagnosed? J Clin Gastroenterol antibodies in different degrees of intesti- Screening of tissue transglutaminase anti- 1996;23:21-3. nal damage in celiac disease. J Clin Gastro­ body in healthy blood donors for celiac 39. Ciacci C, Cavallaro R, Romano R, et al. enterol 2003;36:219-21. disease screening in the Turkish popula- Increased risk of surgery in undiagnosed 53. Dickey W, Hughes D. Disappointing tion. Dig Dis Sci 2004;49:1479-84. celiac disease. Dig Dis Sci 2001;46:2206- sensitivity of endoscopic markers for vil- 24. Fasano A, Berti I, Gerarduzzi T, et al. 8. lous atrophy in a high-risk population: Prevalence of celiac disease in at-risk and 40. Shamaly H, Hartman C, Pollack S, et implications for celiac disease diagnosis not-at-risk groups in the United States: al. Tissue transglutaminase antibodies are during routine endoscopy. Am J Gastro- a large multicenter study. Arch Intern Med a useful serological marker for the diag- enterol 2001;96:2126-8. 2003;163:286-92. nosis of celiac disease in patients with 54. Abrams JA, Diamond B, Rotterdam H, 25. Bingley PJ, Williams AJ, Norcross AJ, Down syndrome. J Pediatr Gastroenterol Green PHR. Seronegative celiac disease: et al. Undiagnosed coeliac disease at age Nutr 2007;44:583-6. increased prevalence with lesser degrees of seven: population based prospective birth 41. Bettendorf M, Doerr HG, Hauffa BP, villous atrophy. Dig Dis Sci 2004;49:546- cohort study. BMJ 2004;328:322-3. et al. Prevalence of autoantibodies associ- 50. 26. Shahbazkhani B, Malekzadeh R, Sotou­ ated with thyroid and celiac disease in 55. Wong RC, Wilson RJ, Steele RH, Rad- deh M, et al. High prevalence of coeliac Ullrich-Turner syndrome in relation to ford-Smith G, Adelstein S. A comparison disease in apparently healthy Iranian blood adult height after growth hormone treat- of 13 guinea pig and human anti-tissue donors. Eur J Gastroenterol Hepatol 2003; ment. J Pediatr Endocrinol Metab 2006; transglutaminase antibody ELISA kits. 15:475-8. 19:149-54. J Clin Pathol 2002;55:488-94. 27. Sood A, Midha V, Sood N, Malhotra V. 42. Goh C, Banerjee K. Prevalence of co- 56. Rostom A, Murray JA, Kagnoff MF. Adult celiac disease in northern India. In- eliac disease in children and adolescents American Gastroenterological Associa- dian J Gastroenterol 2003;22:124-6. with type 1 diabetes mellitus in a clinic tion (AGA) Institute technical review on

n engl j med 357;17 www.nejm.org october 25, 2007 1741

Downloaded from www.nejm.org by JUAN SABATER MD on November 14, 2007 . Copyright © 2007 Massachusetts Medical Society. All rights reserved. T h e new england journal o f medicine

the diagnosis and management of celiac 70. Thompson T. Oats and the gluten-free 86. Askling J, Linet M, Gridley G, Hal- disease. Gastroenterology 2006;131:1981- diet. J Am Diet Assoc 2003;103:376-9. stensen TS, Ekström K, Ekbom A. Cancer 2002. 71. Hallert C, Grant C, Grehn S, et al. incidence in a population-based cohort of 57. Lenhardt A, Plebani A, Marchetti F, et Evidence of poor vitamin status in coeliac individuals hospitalized with celiac dis- al. Role of human-tissue transglutaminase patients on a gluten-free diet for 10 years. ease or dermatitis herpetiformis. Gastro- IgG and anti-gliadin IgG antibodies in the Aliment Pharmacol Ther 2002;16:1333-9. enterology 2002;123:1428-35. diagnosis of coeliac disease in patients 72. Meyer D, Stavropolous S, Diamond B, 87. Smedby KE, Akerman M, Hildebrand with selective immunoglobulin A deficien- Shane E, Green PH. Osteoporosis in a H, Glimelius B, Ekbom A, Askling J. Ma- cy. Dig Liver Dis 2004;36:730-4. North American adult population with lignant lymphomas in coeliac disease: evi- 58. Dickey W, Hughes DF, McMillan SA. celiac disease. Am J Gastroenterol 2001; dence of increased risks for lymphoma Reliance on serum endomysial antibody 96:112-9. types other than enteropathy-type T cell testing underestimates the true prevalence 73. Cellier C, Flobert C, Cormier C, Roux C, lymphoma. Gut 2005;54:54-9. of coeliac disease by one fifth. Scand J Schmitz J. Severe osteopenia in symptom- 88. Green PH, Fleischauer AT, Bhagat G, Gastroenterol 2000;35:181-3. free adults with a childhood diagnosis of Goyal R, Jabri B, Neugut AI. Risk of ma- 59. Abrams JA, Brar P, Diamond B, Rotter­ coeliac disease. Lancet 2000;355:806. lignancy in patients with celiac disease. dam H, Green PH. Utility in clinical prac- 74. Lee SK, Lo W, Memeo L, Rotterdam H, Am J Med 2003;115:191-5. tice of immunoglobulin A anti-tissue Green PH. Duodenal histology in patients 89. Green PH, Rampertab SD. Small bowel transglutaminase antibody for the diag- with celiac disease after treatment with a carcinoma and coeliac disease. Gut 2004; nosis of celiac disease. Clin Gastroenterol gluten-free diet. Gastrointest Endosc 2003; 53:774. Hepatol 2006;4:726-30. 57:187-91. 90. Culliford A, Daly J, Diamond B, Rubin 60. Raivio T, Kaukinen K, Nemes E, et al. 75. Matysiak-Butnik T, Malamut G, Patey- M, Green PH. The value of wireless cap- Self transglutaminase-based rapid coeliac Mariaud de Serre N, et al. Long-term fol- sule endoscopy in patients with compli- disease antibody detection by a lateral low-up of 61 celiac patients diagnosed in cated celiac disease. Gastrointest Endosc flow method. Aliment Pharmacol Ther childhood: evolution toward latency is pos­ 2005;62:55-61. 2006;24:147-54. sible on a normal diet. Gut 2007;56:1379- 91. Howdle PD, Jalal PK, Holmes GK, 61. Johnson TC, Diamond B, Memeo L, et 86. Houlston RS. Primary small-bowel malig- al. Relationship of HLA-DQ8 and severity 76. Siegel M, Bethune MT, Gass J, et al. nancy in the UK and its association with of celiac disease: comparison of New York Rational design of combination enzyme coeliac disease. QJM 2003;96:345-53. and Parisian cohorts. Clin Gastroenterol therapy for celiac sprue. Chem Biol 2006; 92. Bagdi E, Diss TC, Munson P, Isaacson Hepatol 2004;2:888-94. 13:649-58. PG. Mucosal intra-epithelial lymphocytes 62. Kaukinen K, Partanen J, Mäki M, Col- 77. Stepniak D, Spaenij-Dekking L, Mitea in enteropathy-associated T-cell lympho- lin P. HLA-DQ typing in the diagnosis of C, et al. Highly efficient gluten degrada- ma, ulcerative jejunitis, and refractory celiac disease. Am J Gastroenterol 2002; tion with a newly identified prolyl endo- celiac disease constitute a neoplastic pop- 97:695-9. protease: implications for celiac disease. ulation. Blood 1999;94:260-4. 63. Bonamico M, Mariani P, Thanasi E, et Am J Physiol Gastrointest Liver Physiol 93. Gale J, Simmonds PD, Mead GM, al. Patchy villous atrophy of the duode- 2006;291:G621-G629. Sweetenham JW, Wright DH. Enteropa- num in childhood celiac disease. J Pediatr 78. O’Mahony S, Howdle PD, Losowsky thy-type intestinal T-cell lymphoma: clin- Gastroenterol Nutr 2004;38:204-7. MS. Management of patients with non- ical features and treatment of 31 patients 64. Ravelli A, Bolognini S, Gambarotti M, responsive coeliac disease. Aliment Phar- in a single center. J Clin Oncol 2000;18: Villanacci V. Variability of histologic le- macol Ther 1996;10:671-80. 795-803. sions in relation to biopsy site in gluten- 79. Vahedi K, Mascart F, Mary JY, et al. 94. Rongey C, Micallef I, Smyrk T, Murray J. sensitive enteropathy. Am J Gastroenterol Reliability of antitransglutaminase anti- Successful treatment of enteropathy-asso- 2005;100:177-85. bodies as predictors of gluten-free diet ciated T cell lymphoma with autologous 65. Lee SK, Green PH. Endoscopy in celiac compliance in adult celiac disease. Am J stem cell transplant. Dig Dis Sci 2006;51: disease. Curr Opin Gastroenterol 2005;21: Gastroenterol 2003;98:1079-87. 1082-6. 589-94. 80. Kumar PJ, Walker-Smith J, Milla P, 95. Trier JS. Celiac sprue. N Engl J Med 66. Oxentenko AS, Grisolano SW, Murray Harris G, Colyer J, Halliday R. The teen- 1991;325:1709-19. JA, Burgart LJ, Dierkhising RA, Alexander age coeliac: follow up study of 102 patients. 96. Cellier C, Patey N, Mauvieux L, et al. JA. The insensitivity of endoscopic mark- Arch Dis Child 1988;63:916-20. Abnormal intestinal intraepithelial lym- ers in celiac disease. Am J Gastroenterol 81. Fabiani E, Taccari LM, Rätsch IM, Di phocytes in refractory sprue. Gastroenter- 2002;97:933-8. Giuseppe S, Coppa GV, Catassi C. Compli- ology 1998;114:471-81. 67. Marsh MN. Gluten, major histocom- ance with gluten-free diet in adolescents 97. Verkarre V, Asnafi V, Lecomte T, et al. patibility complex, and the small intes- with screening-detected celiac disease: Refractory coeliac sprue is a diffuse gastro­ tine: a molecular and immunobiologic a 5-year follow-up study. J Pediatr 2000; intestinal disease. Gut 2003;52:205-11. approach to the spectrum of gluten sensi- 136:841-3. 98. Carbonnel F, Grollet-Bioul L, Brouet tivity (‘celiac sprue’). Gastroenterology 82. Bardella MT, Molteni N, Prampolini L, JC, et al. Are complicated forms of celiac 1992;102:330-54. et al. Need for follow up in coeliac dis- disease cryptic T-cell lymphomas? Blood 68. Memeo L, Jhang J, Hibshoosh H, Green ease. Arch Dis Child 1994;70:211-3. 1998;92:3879-86. PH, Rotterdam H, Bhagat G. Duodenal 83. Cellier C, Delabesse E, Helmer C, et al. 99. Daum S, Weiss D, Hummel M, et al. intraepithelial lymphocytosis with normal Refractory sprue, coeliac disease, and Frequency of clonal intraepithelial T lym- villous architecture: common occurrence enteropathy-associated T-cell lymphoma. phocyte proliferations in enteropathy-type in H. pylori gastritis. Mod Pathol 2005;18: Lancet 2000;356:203-8. intestinal T cell lymphoma, coeliac dis- 1134-44. 84. Rampertab SD, Forde KA, Green PH. ease, and refractory sprue. Gut 2001;49: 69. Peräaho M, Kaukinen K, Mustalahti Small bowel neoplasia in coeliac disease. 804-12. K, et al. Effect of an oats-containing glu- Gut 2003;52:1211-4. 100. Patey-Mariaud De Serre N, Cellier C, ten-free diet on symptoms and quality of 85. West J, Logan RF, Smith CJ, Hubbard Jabri B, et al. Distinction between coeliac life in coeliac disease: a randomized RB, Card TR. Malignancy and mortality disease and refractory sprue: a simple im- study. Scand J Gastroenterol 2004;39:27- in people with coeliac disease: population munohistochemical method. Histopatho- 31. based cohort study. BMJ 2004;329:716-9. logy 2000;37:70-7.

1742 n engl j med 357;17 www.nejm.org october 25, 2007

Downloaded from www.nejm.org by JUAN SABATER MD on November 14, 2007 . Copyright © 2007 Massachusetts Medical Society. All rights reserved. Medical Progress

101. Cellier C, Cuillerier E, Patey-Mariaud 104. Goerres MS, Meijer JW, Wahab PJ, et ma. N Engl J Med 2006;354:2514-5. [Erra- de Serre N, et al. Push enteroscopy in ce- al. Azathioprine and prednisone combi- tum, N Engl J Med 2006;355:856.] liac sprue and refractory sprue. Gastroin- nation therapy in refractory coeliac dis- 107. Al-Toma A, Goerres MS, Meijer JW, test Endosc 1999;50:613-7. ease. Aliment Pharmacol Ther 2003;18: et al. Cladribine therapy in refractory ce- 102. Mauriño E, Niveloni S, Cherñavsky A, 487-94. liac disease with aberrant T cells. Clin et al. Azathioprine in refractory sprue: 105. Gillett HR, Arnott ID, McIntyre M, Gastroenterol Hepatol 2006;4:1322-7. results from a prospective, open-label et al. Successful infliximab treatment for 108. Al-toma A, Visser OJ, van Roessel study. Am J Gastroenterol 2002;97:2595- steroid-refractory celiac disease: a case HM, et al. Autologous hematopoietic stem 602. report. Gastroenterology 2002;122:800- cell transplantation in refractory celiac 103. Wahab PJ, Crusius JB, Meijer JW, Uil 5. disease with aberrant T cells. Blood 2007; JJ, Mulder CJ. Cyclosporin in the treat- 106. Vivas S, Ruiz de Morales JM, Ramos 109:2243-9. ment of adults with refractory coeliac dis- F, Suarez-Vilela D. Alemtuzumab for re- Copyright © 2007 Massachusetts Medical Society. ease — an open pilot study. Aliment Phar- fractory celiac disease in a patient at risk macol Ther 2000;14:767-74. for enteropathy-associated T-cell lympho-

n engl j med 357;17 www.nejm.org october 25, 2007 1743

Downloaded from www.nejm.org by JUAN SABATER MD on November 14, 2007 . Copyright © 2007 Massachusetts Medical Society. All rights reserved.

Clinical and Experimental Allergy, 37, 823–830 c 2007 The Authors ORIGINAL PAPER Journal compilation c 2007 Blackwell Publishing Ltd

Alterations of food antigen-specific serum immunoglobulins G and E antibodies in patients with irritable bowel syndrome and functional dyspepsia w X. L. ZuoÃ,Y.Q.LiÃ,W.J.LiÃ,Y.T.GuoÃ,X.F.LuÃ,J.M.Lià and P. V. Desmond ÃDepartment of Gastroenterology, Qilu Hospital, Shandong University, Jinan, China and wDepartment of Gastroenterology, St Vincent’s Hospital, Fitzroy Vic., Australia

Clinical and Summary Background Post-prandial worsening of symptoms as well as adverse reactions to one or Experimental more foods are common in the patients with functional gastrointestinal diseases, such as irritable bowel syndrome (IBS) and functional dyspepsia (FD). However, the role played by Allergy true food allergy in the pathogenesis of these diseases is still controversial and there are no well-established tests to identify food allergy in this condition. Objective To investigate serum food antigen-specific IgG, IgE antibody and total IgE antibody titres in controls and patients with IBS and FD, and to correlate symptoms with the food antigen-specific IgG titres in IBS and FD patients. Methods Thirty-seven IBS patients, 28 FD patients and 20 healthy controls participated in this study. Serum IgG and IgE antibody titres to 14 common foods including beef, chicken, codfish, corn, crab, eggs, mushroom, milk, pork, rice, shrimp, soybean, tomatoes and wheat were analysed by ELISA. Serum total IgE titres were also measured. Last, symptomatology was assessed in the study. Results IBS patients had significantly higher titres of IgG antibody to crab (P = 0.000), egg (P = 0.000), shrimp (P = 0.000), soybean (P = 0.017) and wheat (P = 0.004) than controls. FD patients had significantly higher titres of IgG antibody to egg (P = 0.000) and soybean (P = 0.017) than controls. The percentage of individuals with detectable positive food antigen- specific IgE antibodies of the three groups did not show any significant differences (P = 0.971). There were no significant differences between IBS patients, FD patients and controls in the serum total IgE antibody titres (P = 0.978). Lastly, no significant correlation was seen between symptom severity and serum food antigen-specific IgG antibody titres both in IBS and FD patients. Correspondence: Conclusion Serum IgG antibody titres to some common foods increased in IBS and FD Prof. YanQing Li, Department of patients compared to controls. But there is no significant correlation between symptom Gastroenterology, Qilu Hospital of severity and elevated serum food antigen-specific IgG antibodies in these patients. Shandong University, Jinan 250012, China. Keywords food allergy, functional dyspepsia, IgE, IgG, irritable bowel syndrome E-mail: [email protected] Submitted 25 November 2006; revised 6 March 2007; accepted 16 March 2007

ing their symptoms to food allergy [2–4]. However, the Introduction role played by true food allergy in the pathogenesis of IBS Food allergy is a complex area of medicine. Up to 20% of and other functional diseases of the GI tract are still the population have adverse reactions to food and claim controversial and there are no well-established tests to to be of food allergy or food intolerance [1]. The percen- identify food allergy in this condition [5–7]. tage of food allergy is much higher in some functional Previously, food allergy was believed to be associated diseases of the gastrointestinal (GI) tract such as irritable with an IgE-mediated immune response to a particular bowel syndrome (IBS), with 20–65% of patients attribut- allergen in the diet. Therefore, the standardized skin prick 824 X. L. Zuo et al testing and RAST testing were frequently used to diagnose Controls food allergy [8, 9]. However, there is no evidence that The control group consisted of 20 healthy subjects (six men shows IgE does indeed play an important role in hyper- and 14 women, mean age 36.5 years) recruited from the sensitive reactions to food in IBS patients [10–14]. The community. All control subjects were free of GI symptoms gold standard in this condition is the double blind and had no evidence of acute or chronic illnesses. placebo-controlled food challenge test under careful None of the patients or healthy individuals was on any supervision in a hospital. But this test is cumbersome, medication at the time of the study. The patient groups time consuming and of poor patient compliance, which and the control group did not differ significantly with limits its use in clinical practise. Fortunately, accumulat- respect to mean age and sex ratio. ing data in recent years have indicated that IgG-mediated Informed, preferably written, consent was obtained immune response, which characteristically gives a more from each subject. The study has been performed accord- delayed response following exposure to a particular anti- ing to the Declaration of Helsinki and the procedures have gen, is of great importance in food allergy [15–17] and the been approved by Qilu hospital ethics committee. measurement of serum IgG titres opens up a new avenue for diagnosing food allergy in patients suffering adverse reactions to foods. Measurement of serum food antigen-specific immunoglo- Post-prandial worsening of symptoms as well as ad- bulins G and E antibodies. Serum samples were collected 1 verse reactions to one or more foods are common [18, 19] from all subjects and stored at À20 C for subsequent and dietary elimination can lead to symptomatic improve- analysis. Serum IgG and IgE antibody titres to 14 common ments in patients with IBS [3, 20–22]. Recently, Zar et al. foods including beef, chicken, codfish, corn, crab, eggs, [23] have shown elevated IgG titres in IBS patients while mushroom, milk, pork, rice, shrimp, soybean, tomatoes Atkinson et al. [24] have demonstrated that food elimina- and wheat were analysed by ELISA. tion based on IgG antibodies may be effective in reducing IBS symptoms. Other studies have shown food allergy is Serum food antigen-specific immunoglobulin G enzyme- also present in functional dyspepsia (FD), which has over- linked immunosorbent assay. Serum food antigen-specific lapping symptoms with IBS [25, 26]. However, fewer IgG ELISA was performed with Allerquant Food Allergy studies focus on the role of food allergy in the pathogen- Screening ELISA Kit (Biomerica, Inc. Newport Beach, CA, esis of FD and the level of serum food antigen-specific IgG USA). antibodies in FD has not been investigated. According to the instruction of the ELISA Kit, 50, 100, The aim of this study is to compare serum food antigen- 200 and 400 U/mL of Food IgG Calibrator were prepared specific IgG and IgE antibodies in controls and patients and added to the microplate together with blanks and with IBS and FD, and to correlate symptoms with the food positive controls. This is the calibration curve to be used in antigen-specific IgG and IgE titres in IBS and FD patients. the assay. Serum samples were diluted to 1/100 and 25 mL of each serum sample were taken and added to 2.5 mL of serum dilute regents. Then 100 mL of the diluted patient Materials and methods serum were placed into the microwells coated with 14 food antigens for 60 min at room temperature. And then, 100 mL Patients of food IgG–HRP conjugate were added to each well after Thirty-seven IBS patients (12 men and 25 women; mean washing and the plates were incubated for 30 min at room age 36 years) and 28 FD patients (nine men and 19 temperature. Another 100 mL of working substrate mix women; mean age 35 years) participated in this study. All (TMB and H2O2) were applied to each well and the plates patients were recruited from the Department of Gastro- were covered and incubated for 10 min at room tempera- enterology of Qilu Hospital, Shandong University and ture. The reaction was stopped using 50 mL per well of 1 N were diagnosed by using the Rome II criteria [27] for IBS sulphuric acid. At last, the plates were read using a Dynex (pain or abdominal discomfort accompanied by two or ELISA plate reader (Dynex Technologies, Inc., Chantilly, three symptoms, such as relief with defecation and/or with VA, USA) at 450 nm. The IgG concentration to each food alterations in the frequency of evacuations or in the shape antigen was expressed as units per millitre. of the feces for at least 12 weeks, which need not be Plates with microwells were washed three times with consecutive, in the preceding 12 months; in the absence of washing buffer between steps, and were incubated at room organic GI diseases) and for FD (persistent or recurrent temperature for each stage of the assay. symptoms, such as pain or discomfort in the upper abdo- men at least 12 weeks earlier, not necessarily serial, during Serum food antigen-specific immunoglobulin E and total the preceding 12 months). Organic GI disorders were immunoglobulin E enzyme-linked immunosorbent assay. excluded by routine laboratory tests and endoscopies with Serum food antigen-specific IgE and total IgE ELISA were biopsies. performed with the IVT Allergy Profile kit (In vitro

c 2007 The Authors Journal compilation c 2007 Blackwell Publishing Ltd, Clinical and Experimental Allergy, 37 : 823–830 Food allergy in patients with IBS and FD 825

Technologies Inc. Arlington, TX, USA). It is a qualitative individuals with positive IgE antibodies between groups. test designed to identify IgE levels in the test samples. A Comparison between groups for interval data (age, positive result is identified visually by a gradual yellow to weight) was carried out with t-test. The correlation be- purple colour change around a reactive segment within tween the individual symptom scores and the IgG titres the capillary. This kit is also semiquantitative in that the was analysed by Pearson’s correlation test. Significance rate of conversion from yellow to purple and the intensity was accepted at 5% level (P o 0.05). The statistics package of the purple colour is proportional to the level of IgE SPSS v 13.0 running on Windows XP Professional was antibody in the patient sample. All components were used for the analyses. allowed to come to the room temperature before use. The test procedures are as follows. Firstly, each serum sample was added to a properly Results labelled multiple immunoassay device and allowed to react for 100 min at room temperature. Then was the Serum food antigen-specific immunoglobulin G sample expelled from the device and the device washed antibody titres with 1 mL wash solution. Secondly, the conjugate reagent was added to the device and allowed to react for 100 min. The serum IgG antibody titres of each of the IBS, FD and Then, was the conjugate reagent expelled from the device control groups to each food antigen are shown in this and the device washed with wash solution. Thirdly, the study together with the results of the ANOVA (Table 1). There substrate indicator was added to the device and the were significant differences between the three groups in characteristic yellow to purple colour change was within their IgG responses to crab, egg, shrimp, soybean and the observed next 30 min period. wheat. IBS patients had significantly higher titres of IgG Highly allergic reactions to a particular allergen caused antibody to crab, egg, shrimp, soybean and wheat than a rapid purple colouration around that specific segment controls and higher titres of IgG antibody to crab, egg, within 5 min. Weaker allergen took progressively longer shrimp and wheat than FD patients. FD patients had to effect the colour change. Negative reactions showed significantly higher IgG antibody titres to egg and soy- equivalent colour to the inert spacers. On the total IgE bean than controls and lower IgG antibody titres than IBS segment, low total IgE levels (430 IU/mL) were evident by patients to crab, egg, shrimp and wheat (Fig. 1). There a yellow to grey colouration, distinguishable from normal were no significant differences between the three groups IgE levels (60 Æ 30 IU/mL), which yield moderately purple in their IgG responses to beef, chicken, codfish, corn, colour throughout, and elevated IgE levels (>90 IU/mL), mushroom, milk, pork, rice and tomatoes. which yield intense purple colouration after 30 min in- cubation with the substrate indicator.

Symptom questionnaire. Symptomatology was assessed Table 1. Serum IgG antibody titres to food antigens showing the means in IBS, FD and control groups, the standard error of the mean (SEM) and in the study. All the patients completed a symptom P-value from the ANOVA assessment questionnaire based on Rome II criteria to evaluate the severity of symptoms for the 1 week period Mean Æ SEM (U/mL) before the interview. Questions targeting the presence and IBS (n = 37) FD (n = 28) Control (n = 20) P-value severity of abdominal pain/discomfort, bloating, bowel Beef 32.86 Æ 0.46a 32.43 Æ 0.44a 32.10 Æ 0.60a 0.556 urgency/diarrhoea, constipation, early satiety, nausea and a a a belching were asked, for example ‘Have you experienced Chicken 28.65 Æ 0.51 27.68 Æ 0.58 27.20 Æ 0.55 0.171 Codfish 32.86 Æ 0.49a 32.89 Æ 0.52a 32.55 Æ 0.59a 0.902 abdominal pain/discomfort during this week?’ If the a a a Corn 31.87 Æ 0.43 31.82 Æ 0.46 31.29 Æ 0.60 0.617 subjects reported ‘yes’, they were then asked to grade the Crab 50.27 Æ 0.89a 37.53 Æ 0.95b 37.90 Æ 0.78b 0.000 severity of that symptom using the scale 1 = mild, Eggs 119.3 Æ 11.8a 69.71 Æ 4.63b 51.93 Æ 3.74c 0.000 2 = moderate, 3 = intense and 4 = severe. If the subjects Mushroom 27.78 Æ 0.51a 26.93 Æ 0.63a 27.25 Æ 0.49a 0.515 reported ‘no’, the score was recorded as 0. The total score Milk 33.14 Æ 0.51a 32.36 Æ 0.64a 32.41 Æ 0.98a 0.615 of each patient is the sum of each symptom score. Pork 31.54 Æ 0.45a 31.82 Æ 0.53a 30.80 Æ 0.59a 0.437 Rice 28.68 Æ 0.52a 29.54 Æ 0.75a 28.41 Æ 0.83a 0.493 Statistical analysis. Linear models were fitted to the log Shrimp 65.05 Æ 3.70a 45.39 Æ 1.62b 43.25 Æ 1.73b 0.000 a a b transformation of the variables, such as the IgG and total Soybean 55.83 Æ 3.35 53.29 Æ 1.89 43.60 Æ 1.95 0.017 Tomatoes 34.65 Æ 0.78a 33.79 Æ 0.98a 34.05 Æ 1.23a 0.782 IgE titres. A one-way ANOVA was carried out to test for Wheat 60.59 3.4a 49.39 2.05b 48.10 2.01b 0.004 differences in levels of antibody titres between the groups. Æ Æ Æ Tamhane’s T2 test was used to test for pairwise differences Means with different superscripts across rows were significantly different 2 between IBS, FD and control groups. The w tests were (P o 0.05). used to test for the differences of the percentage of IBS, irritable bowel syndrome; FD, functional dyspepsia.

c 2007 The Authors Journal compilation c 2007 Blackwell Publishing Ltd, Clinical and Experimental Allergy, 37 : 823–830 826 X. L. Zuo et al

Fig. 1. Comparisons of serum IgG antibodies to crab, egg, shrimp, Fig. 2. Comparision of total IgE antibody titres to food antigens between soybean and wheat between irritable bowel syndrome (IBS), functional irritable bowel syndrome (IBS), functional dyspepsia (FD) and control dyspepsia (FD) and control groups. IBS patients have significantly higher groups. The data were expressed with the individual values in the three IgG antibodies to crab, egg, shrimp, soybean and wheat than controls. FD groups. Horizontal bars represent the means for every group. There were patients have significantly higher IgG antibodies to egg and soybean no significant differences between IBS patients, FD patients and controls than controls. IBS patients have significantly higher IgG antibodies to in the serum total IgE antibody titres (P = 0.978). crab, egg, shrimp and wheat than FD patients. The data were expressed à with the geometric means and the standard errors of the means. P o 0.05 compared with controls; #P o 0.05 compared with FD patients. against eggs, one against mushroom and four against wheat. Of 28 FD patients, one had positive IgE antibodies Serum food antigen-specific immunoglobulin E against beef, one against codfish, two against eggs, one antibody titres against soybean and three against wheat. In addition, one control had positive IgE antibodies against codfish, two In terms of serum food antigen-specific IgE antibodies, 10 against eggs, one against mushroom and two against of 37 IBS patients, eight of 28 FD patients and six of 20 wheat. The number of patients and controls with positive controls had detectable positive IgE ELISA results to beef, IgE antibodies against each individual food antigen was codfish, eggs, mushroom, soybean and wheat (Table 2). too small to apply a statistical test. The percentage of individuals with detectable positive IgE antibodies of the three groups did not show any signifi- cant differences (27.03% for IBS, 28.57% for FD and 30% Serum total immunoglobulin E antibody titres for controls, w2 = 0.059, P = 0.971). In IBS patients, only The serum total IgE antibody titres were also measured in two had positive IgE antibodies against codfish, three this study. There were no significant differences between IBS patients, FD patients and controls (means Æ SEM Table 2. Summary of the individuals with positive IgE ELISA results 66.87 Æ 7.65 for controls, 68.69 Æ 5.9 for FD patients and in IBS, FD and control groups 67.30 Æ 5.49 for IBS patients, P = 0.978, Fig. 2). IBS (n = 37) FD (n = 28) Control (n = 20) Beef 0 1 0 Correlation with symptoms Chicken 0 0 0 Codfish 2 1 1 The symptom severity of each FD and IBS patient was Corn 0 0 0 scored according to the symptom questionnaire. The Crab 0 0 0 relation between the symptoms and elevated IgG re- Eggs 3 2 2 sponses to some food antigens was observed and no Mushroom 1 0 1 significant correlation was seen between symptom sever- Milk 0 0 0 ity and serum food antigen-specific IgG antibody titres Pork 0 0 0 both in IBS and FD patients (Fig. 3). Rice 0 0 0 Shrimp 0 0 0 Soybean 0 1 0 Discussion Tomatoes 0 0 0 Wheat 4 3 2 This study demonstrated a significant increase in IgG antibody titres to several common foods in patients with IBS, irritable bowel syndrome; FD, functional dyspepsia. functional diseases of the GI tract compared with healthy

c 2007 The Authors Journal compilation c 2007 Blackwell Publishing Ltd, Clinical and Experimental Allergy, 37 : 823–830 Food allergy in patients with IBS and FD 827

Fig. 3. The correlation between abdominal symptom scores and IgG titres to food antigen in irritable bowel syndrome (IBS) patients and functional dyspepsia (FD) patients. (a–e) The individual plots of values for food antigen-specific IgG titres and symptom scores in 37 IBS patients. No significant correlations are seen between symptom scores and IgG titres to crab (a, P = 0.426), eggs (b, P = 0.657), shrimp (c, P = 0.800), soybean (d, P = 0.854) and wheat (e, P = 0.794). (f) The individual plots of values for food antigen-specific IgG titres and symptom scores in 28 FD patients. No significant correlations were seen between symptom scores and IgG titres to eggs (P = 0.460) and soybean (P = 0.986).

subjects. In IBS patients, elevated food antigen-specific in China and they are used to consume less beef and more IgG antibodies to crab, egg, shrimp, soybean and wheat seafood. were observed while elevated food antigen-specific IgG Food allergy can involve different organs and systems antibodies to egg and soybean were observed in FD such as the digestive tract, the skin, the respiratory tract patients. It should be mentioned that, consistent with the and the cardiovascular system. While dermatologic, re- previous study [23], the antibodies against wheat were spiratory and systemic manifestations of food allergy are significantly elevated in IBS patients. However, we also well recognized, the reactions manifested primarily in the got some raised food antigen-specific antibodies different digestive tract can be difficult to recognize, diagnose and from the previous study [23], such as elevated antibody treat. This is due to the protean ways food can cause GI titres to crab and shrimp instead of beef. This might be due symptoms, the relatively poorly understood pathophysio- to the difference between a Chinese diet and a western logic mechanisms and the limited diagnostic methods diet. More specifically, most of the subjects in this study available to objectively identify afflicted individuals. are from a coastal area (where this study was performed) These deficiencies are, in part, a consequence of the

c 2007 The Authors Journal compilation c 2007 Blackwell Publishing Ltd, Clinical and Experimental Allergy, 37 : 823–830 828 X. L. Zuo et al difficulty accessing the GI tract to establish mechanisms serum IgE measurements do not correlate well with the of disease and develop methods to diagnose and treat food mucosal allergic response in the intestine [28, 37]. This allergy [28, 29]. Often, patients of this nature are classified suggests that an IgE-mediated hypersensitivity response as being psychosomatic or being functional without to food is unlikely to play an important role in most of the defining the real problem. It has been recognized for some IBS and FD patients. time now that several ‘functional diseases’ might be Both IBS and FD patients reported that GI symptoms associated with food allergy [30–34]. often occur after certain food intake. There was also a The mechanisms that underlie these increases in IgG great overlap in the post-prandial dyspeptic symptoms in antibody responses to some common foods remain spec- the two groups of patients, such as gas problems, pain, ulative. The role of IgG and IgA antibodies in the coeliac nausea and upper abdominal discomfort. This came as no disease is well studied [35]. Coeliac disease occurs due to a surprise, as a high prevalence of overlap between FD and delayed immune reaction to gluten in wheat. This causes IBS has been universally reported and some shared intestinal membrane damage, IgG/IgA change and the common pathophysiological disturbances might exist in resultant diarrhoea, abdominal bloating and anaemia. these GI functional diseases, such as delayed gastric Raised IgG antibodies to food antigens have also been emptying, visceral hypersensitivity including food hyper- reported in patients with asthma caused by milk allergy sensitivity [38–40]. But, interestingly, we were unable to and patients with atopic dermatitis and/or bronchial correlate the level of food-specific IgG antibodies with the asthma caused by soybean allergy [17, 36]. Exclusions of severity of symptoms both in IBS and FD patients in this the offending foods from diet have shown to improve study. The underlying mechanisms of the pathogenesis symptoms in these diseases. Raised food-specific IgG have not yet been fully defined. Maybe some patients with antibodies may play a similar pathophysiological role in food sensitivities have non-allergic food reactions and the IBS and FD patients. The allergic food antigens trans- elevated IgG antibodies to food may be secondary to ported by way of M cells into the lamina propria activate ‘inflammation’ and therefore be more of an epi-phenom- T helper cells and B cells, increase the production of IgG enon. In these patients, there are no real food allergies or and cytokines. Then the increased IgG antibodies and immunity responses while experiencing symptoms. Psy- cytokines lead to the inflammation response of the gut, chological factors have also been suggested to be of great which is now believed to play an important role in the importance for the reported food intolerance in these pathogenesis of IBS through inducing alterations in GI patients [41, 42]. In addition, IBS and FD symptoms may peristalsis, abdominal discomfort and bowel dysfunction. also be related to abnormal intestinal bacteria, caffeine, In addition to the immune inflammation reaction ex- alcohol, low dietary fibre, overgrowth of intestinal yeasts planation, another possible mechanism that may account and excessive dietary sugars. for the elevated IgG antibodies is the alteration in the In conclusion, increased antigen-specific IgG antibody permeability of gut mucosa. Theoretically, any increases titres for some foods were found in IBS and FD patients in the gut mucosal permeability in IBS and FD patients compared with controls but there is no evidence that these might increase the uptake of undigested protein and findings contribute to the pathogenesis of these functional increase antigenic load presented to the mucosal immune GI diseases. Future studies along these lines are expected system. This may lead to the increased IgG antibody titres to lead to a better understanding about the role of elevated even with a normal physiological response of the gut food antigen-specific IgG antibodies in these functional immune system. If the above hypothesis is correct, then a GI diseases. generalized increase in the IgG antibodies to all 14 food antigens should have been observed. However, in this study we found that the food-specific IgG antibodies Acknowledgements increased to only some rather than to all food antigens. The authors appreciate the considerable assistance from The possible explanation is that the patients might have the Gastroenterology kinetic laboratory and the Central selective gut permeability to food antigens and the in- Laboratory of Immunity in Qilu Hospital, Shandong Uni- crease of food-specific IgG titres is a specific reaction versity. This study was founded by a research grant (NSFC, rather than a non-specific response to increased gut 30570831) from National Natural Science Foundation of mucosal permeability. China and a grant (Y2005C02) from the Department of In terms of serum food antigen-specific IgE antibody Science and Technology of Shandong Province of China. and total IgE antibody titres, no significant difference was found in IBS, FD patients and controls. Furthermore, there were fewer individuals who had positive food antigen- specific IgE responses compared with food antigen-specific References IgG responses in the studied population. This came as no 1 Young E, Stoneham MD, Petruckevitch A, Barton J, Rona R. A surprise, as several other studies also demonstrate that population study of food intolerance. Lancet 1994; 343:1127–30.

c 2007 The Authors Journal compilation c 2007 Blackwell Publishing Ltd, Clinical and Experimental Allergy, 37 : 823–830 Food allergy in patients with IBS and FD 829

2 Bischoff SC, Crowe SE. Gastrointestinal food allergy: new in- 20 Whorwell P, Lea R. Dietary treatment of the irritable bowel sights into pathophysiology and clinical perspectives. Gastro- syndrome. Curr Treat Options Gastroenterol 2004; 7:307–16. enterology 2005; 128:1089–113. 21 Bentley SJ, Pearson DJ, Rix KJ. Food hypersensitivity in irritable 3 Nanda R, James R, Smith H, Dudley CR, Jewell DP. Food intoler- bowel syndrome. Lancet 1983; 2:295–7. ance and the irritable bowel syndrome. Gut 1989; 30:1099–104. 22 Farah DA, Calder I, Benson L, MacKenzie JF. Specific food 4 Dainese R, Galliani EA, De Lazzari F et al. Discrepancies between intolerance: its place as a cause of gastrointestinal symptoms. reported food intolerance and sensitization test findings in Gut 1985; 26:164–8. irritable bowel syndrome patients. Am J Gastroenterol 1999; 23 Zar S, Benson MJ, Kumar D. Food-specific serum IgG4 and IgE 94:1892–7. titers to common food antigens in irritable bowel syndrome. Am 5 Niec AM, Frankum B, Talley NJ. Are adverse food reactions J Gastroenterol 2005; 100:1550–7. linked to irritable bowel syndrome? Am J Gastroenterol 1998; 24 Atkinson W, Sheldon TA, Shaath N, Whorwell PJ. Food 93:2184–90. elimination based on IgG antibodies in irritable bowel 6 Jones VA, McLaughlan P, Shorthouse M, Workman E, Hunter JO. syndrome: a randomised controlled trial. Gut 2004; 53: Food intolerance: a major factor in the pathogenesis of irritable 1459–64. bowel syndrome. Lancet 1982; 2:1115–7. 25 Soares RL, Figueiredo HN, Maneschy CP, Rocha VR, Santos JM. 7 McKee AM, Prior A, Whorwell PJ. Exclusion diets in irritable Correlation between symptoms of the irritable bowel syndrome bowel syndrome: are they worthwhile? J Clin Gastroenterol and the response to the food extract skin prick test. Braz J Med 1987; 9:526–8. Biol Res 2004; 37:659–62. 8 Roudebush P. Diagnosis and management of adverse food reac- 26 Park MI, Camilleri M. Is there a role of food allergy in irritable tions. In: Bonagura JD, ed. Kirk’s current veterinary therapy XII. bowel syndrome and functional dyspepsia? A systematic review. Philadelphia: W.B. Saunders Co, 1995; 59–64. Neurogastroenterol Motil 2006; 18:595–607. 9 Spergel JM, Brown-Whitehorn T. The use of patch testing in 27 Thompson WG, Longstreth GF, Drossman DA et al. Functional the diagnosis of food allergy. Curr Allergy Asthma Rep 2005; bowel disorders and functional abdominal pain. Gut 1999; 45 5:86–90. (Suppl. 2):II43–7. 10 Zwetchkenbaum J, Burakoff R. The irritable bowel syndrome and 28 Bischoff SC, Mayer JH, Manns MP. Allergy and the gut. Int Arch food hypersensitivity. Ann Allerg 1988; 61:47–49. Allergy Immunol 2000; 121:270–83. 11 Zar S, Kumar D, Benson M.J. Review article: food hypersensitiv- 29 Crowe SE. Gastrointestinal food allergies: do they exist? Curr ity and irritable bowel syndrome. Aliment Pharm Ther 2001; Gastroenterol Rep 2001; 3:351–7. 15:439–43. 30 Stefanini GF, Saggioro A, Alvisi V et al. Oral cromolyn sodium 12 Petitpierre M, Gumowski P, Girard JP. Irritable bowel syndrome in comparison with elimination diet in the irritable bowel syn- and hypersensitivity to food. Ann Allergy 1985; 54:538–40. drome, diarrheic type. Multicenter study of 428 patients. Scand 13 Barau E, Dupont C. Modifications of intestinal permeability J Gastroenterol 1995; 30:535–41. during food provocation procedures in pediatric irritable bowel 31 Niec AM, Frankum B, Talley NJ. Are adverse food reactions syndrome. J Pediatr Gastroenterol Nutr 1990; 11:72–77. linked to irritable bowel syndrome? Am J Gastroenterol 1998; 14 Roussos A, Koursarakos P, Patsopoulos D, Gerogianni I, Philip- 93:2184–90. pou N. Increased prevalence of irritable bowel syndrome in 32 Iacono G, Cavataio F, Montalto G et al. Intolerance of cow’s milk patients with bronchial asthma. Respir Med 2003; 97:75–79. and chronic constipation in children. N Engl J Med 1998; 15 Crowe SE, Perdue MH. Gastrointestinal food hypersensitivity: 339:1100–4. basic mechanisms of pathophysiology. Gastroenterology 1992; 33 Read NW. Food and hypersensitivity in functional dyspepsia. Gut 103:1075–95. 2002; 51 (Suppl. 1):i50–3. 16 Host A, Husby S, Gjesing B, Larsen JN, Lowenstein H. Prospective 34 Spanier JA, Howden CW, Jones MP. A systematic review of estimation of IgG, IgG subclass and IgE antibodies to dietary alternative therapies in the irritable bowel syndrome. Arch Intern proteins in infants with cow’s milk allergy. Levels of antibodies Med 2003; 163:265–74. to whole milk protein, BLG and ovalbumin in relation to repeated 35 O’Farrelly C, Kelly J, Hekkens W et al. Alpha gliadin antibody milk challenge and clinical course of cow’s milk allergy. Allergy levels: a serological test for coeliac disease. Br Med J (Clin Res 1992; 47:218–29. Ed) 1983; 286:2007–10. 17 Awazuhara H, Kawai H, Maruchi N. Major allergens in soybean 36 Shakib F, Brown HM, Phelps A, Redhead R. Study of IgG sub- and clinical significance of IgG4 antibodies investigated by IgE class antibodies in patients with milk intolerance. Clin Allergy and IgG4 immunoblotting with sera from soybean-sensitive 1986; 16:451–8. patients. Clin Exp Allergy 1997; 27:325–32. 37 Schwab D, Raithel M, Klein P et al. Immunoglobulin E and 18 Ragnarsson G, Bodemar G. Pain is temporally related to eating eosinophilic cationic protein in segmental lavage fluid of the but not to defaecation in the irritable bowel syndrome (IBS). small and large bowel identify patients with food allergy. Am Patients’ description of diarrhea, constipation and symptom J Gastroenterol 2001; 96:508–14. variation during a prospective 6-week study. Eur J Gastroenterol 38 Gwee KA, Chua AS. Functional dyspepsia and irritable bowel Hepatol 1998; 10:415–21. syndrome, are they different entities and does it matter? World J 19 Dainese R, Galliani EA, De Lazzari F, Di Leo V, Naccarato R. Gastroenterol 2006; 12:2708–12. Discrepancies between reported food intolerance and sensitiza- 39 Stanghellini V, Tosetti C, Barbara G et al. Dyspeptic symptoms tion test findings in irritable bowel syndrome patients. Am J and gastric emptying in the irritable bowel syndrome. Am Gastroenterol 1999; 94:1892–7. J Gastroenterol 2002; 97:2738–43.

c 2007 The Authors Journal compilation c 2007 Blackwell Publishing Ltd, Clinical and Experimental Allergy, 37 : 823–830 830 X. L. Zuo et al

40 Haag S, Talley NJ, Holtmann G. Symptom patterns in functional gastrointestinal disorders. Neurogastroenterol Motil 2005; dyspepsia and irritable bowel syndrome: relationship to distur- 17:500–11. bances in gastric emptying and response to a nutrient challenge 42 Addolorato G, Marsigli L, Capristo E, Caputo F, Dall’Aglio C, in consulters and non-consulters. Gut 2004; 53:1445–51. Baudanza P. Anxiety and depression: a common feature of 41 Houghton LA, Whorwell PJ. Towards a better understand- health care seeking patients with irritable bowel syndrome and ing of abdominal bloating and distension in functional food allergy. Hepatogastroenterology 1998; 45:1559–64.

c 2007 The Authors Journal compilation c 2007 Blackwell Publishing Ltd, Clinical and Experimental Allergy, 37 : 823–830

Downloaded from gut.bmj.com on 17 October 2008 1240

GUT IMMUNOLOGY The gut–joint axis: cross reactive food antibodies in rheumatoid arthritis M Hvatum, L Kanerud, R Ha¨llgren, P Brandtzaeg ......

Gut 2006;55:1240–1247. doi: 10.1136/gut.2005.076901

Background and aims: Patients with rheumatoid arthritis (RA) often feel there is an association between food intake and rheumatoid disease severity. To investigate a putative immunological link between gut immunity and RA, food antibodies were measured in serum and perfusion fluid from the jejunum of RA patients and healthy controls to determine the systemic and mucosal immune response. Methods: IgG, IgA, and IgM antibodies to dietary antigens were measured in serum and jejunal perfusion See end of article for fluid from 14 RA patients and 20 healthy subjects. The antigens originated from cow’s milk (a-lactalbumin, authors’ affiliations b-lactoglobulin, casein), cereals, hen’s egg (ovalbumin), cod fish, and pork meat...... Results: In intestinal fluid of many RA patients, all three immunoglobulin classes showed increased food Correspondence to: specific activities. Except for IgM activity against b-lactoglobulin, all other IgM activities were significantly Professor P Brandtzaeg, increased irrespective of the total IgM level. The RA associated serum IgM antibody responses were Institute of Pathology, Rikshospitalet, N-0027 relatively much less pronounced. Compared with IgM, the intestinal IgA activities were less consistently Oslo, Norway; per. raised, with no significant increase against gliadin and casein. Considerable cross reactivity of IgM and [email protected]. IgA antibodies was documented by absorption tests. Although intestinal IgG activity to food was quite low, no it was nevertheless significantly increased against many antigens in RA patients. Three of the five RA Revised version received patients treated with sulfasalazine for 16 weeks had initially raised levels of intestinal food antibodies; 24 November 2005 these became normalised after treatment, but clinical improvement was better reflected in a reduced Accepted for publication erythrocyte sedimentation rate. 1 February 2006 Conclusions: The production of cross reactive antibodies is strikingly increased in the gut of many RA Deceased Published online first patients. Their food related problems might reflect an adverse additive effect of multiple modest 16 February 2006 hypersensitivity reactions mediated, for instance, by immune complexes promoting autoimmune reactions ...... in the joints.

atients with rheumatoid arthritis (RA) often feel that reported in RA patients, but the levels were low compared there is an association between food intake and their with IgG and IgM directed against the same antigen in disease activity, but evidence to support such a connec- patients as well as controls.8 Raised IgG activity to gliadin P 1 tion has been contradictory. Reports are usually based on was found in 47% of 93 RA patients, and 41% concurrently diet experiments with quite different protocols, followed by had IgA rheumatoid factor (RF) and there was some some sort of food challenge. Food hypersensitivity in RA does association with duodenal villous atrophy.9 However, a not reflect IgE mediated allergy, and most studies have subsequent study was contradictory.10 concluded that food is unlikely to have a pathogenic effect in Overall, serum antibodies do not appear to provide an RA. Thus, Panush2 carried out blinded encapsulated chal- immunological link between diet and RA. The reason might lenges, and no more than 5% of the RA patients were deemed be that activation of the intestinal immune system is not to show immunological food sensitivity. Nevertheless, a reliably reflected in the serum, and the fact that circulating recent large European epidemiological study, determining IgA RF is predominantly polymeric supports a mucosal odds ratios (ORs) after adjusting for possible confounding origin.11 Moreover, RA patients may have occult small variables, suggested that there is a significant association intestinal inflammation and increased mucosal permeability between inflammatory polyarthritis and a high intake of red independent of the use of non-steroidal anti-inflammatory meat (OR = 1.9), meat and meat products combined drugs (NSAIDs).12 13 We therefore measured antibodies in (OR = 2.3), and total proteins (OR = 2.9).3 perfusion fluid from the jejunum of RA patients and healthy Few previous reports have considered that a pathogenic controls to investigate directly the mucosal immune response dietary effect on RA could depend on a persistent intake of to a variety of food proteins. food; a brief test challenge with a relatively small dose might not precipitate clinical symptoms. Studies considering the METHODS quantitative variable have in fact tended to suggest that food Study subjects does have pathogenic importance, at least in a significant 45 We studied 17 patients with seropositive RA diagnosed fraction (20–40%) of the patients. according to the criteria of the American College of Attempts to identify food sensitive RA patients by Rheumatology.14 Their mean age was 50 years (range 26 to measuring food specific antibodies or immune complexes in 70) and the mean duration of disease 10 years (range 6 serum have failed. Our previous investigation likewise 9 months to 30 years). Twenty healthy subjects served as concluded that serum antibody measurements seldom pre- dict or confirm food hypersensitivity in RA patients. Abbreviations: RA, rheumatoid arthritis; RF, rheumatoid factor; NSAID, Although increased IgM activities were found, there was no non-steroidal anti-inflammatory drug; ELISA, enzyme linked convincing association with clinical variables or dietetic immunosorbent assay; BSA, bovine serum albumin; OD, optical density; benefits.7 Raised serum IgA activity to gliadin has been SIgA, secretory IgA.

www.gutjnl.com Downloaded from gut.bmj.com on 17 October 2008 Intestinal food antibodies in rheumatoid arthritis 1241 controls; their sex distribution was similar to the RA patients solubilised in 0.02 M ammonium acetate, pH 7.0, while crude but because of ethical constraints we were unable to match antigens were extracted with ammonium acetate from the controls for age, although the ranges were highly wheat, oatS, soy, pork meat, or codfish. The antigen overlapping (mean age 29 years; range 23 to 39). RA patients preparations were frozen and thawed several times during and controls were all white and were taking a normal diet the extraction procedure to increase the protein output. with no restrictions; none complained of gastrointestinal symptoms. Antibody reagents for immunoassays The RA patients suffered from active disease as defined by Isotype specific rabbit antisera to human IgG and IgA were the presence of at least two of the following three criteria: prepared in our own laboratory (LIIPAT Nos 38 and 252). duration of morning stiffness >60 minutes, tenderness or Rabbit antibody to human IgM (Code No A0425) was swelling or both of six or more joints, and an erythrocyte obtained from Dako (Glostrup, Denmark). Alkaline phos- sedimentation rate (ESR) of .30 mm in the first hour. None phatase conjugated swine anti-rabbit IgG was obtained from of the patients had received treatment with gold, penicilla- Orion Diagnostica (Espoo, Finland). Total immunoglobulin mine, chloroquine, sulfasalazine, corticosteroids, or immuno- levels in serum and jejunal fluid were determined by enzyme suppressants for the three months before study inclusion. linked immunosorbent assay (ELISA), as previously Most of them were treated with NSAIDs, but in all except two 18 19 this treatment was withdrawn for at least three days before described. Albumin was measured by a commercial intestinal perfusion was undertaken. radioimmunoassay (Pharmacia, Uppsala, Sweden). Five of the patients were reinvestigated after treatment with sulfasalazine for 16 weeks; three of these had not Measurements of antibody activities received NSAIDs for at least one month. The dose of Relative levels of IgG and IgA antibody activities against sulfasalazine was increased by 0.5 g weekly from 1 g/day different food antigens were determined with an ELISA up to 2–3 g/day. slightly modified from our previous method.17 Briefly, The patient protocols as well as the sampling of serum and antigens in 0.02 M ammonium acetate were coated onto jejunal perfusion fluid were based on informed consent and Costar microtitre plates, No 3590 (Cambridge, were approved by the local ethics committee. Most of the Massachusetts, USA) and washed. Activities were deter- subjects were included in a previously reported study of mined in triplicates of perfusion samples, diluted 1/5 (IgG) or microbial antibodies.15 1/10 (IgA) in 0.02 M Tris buffer (pH 7.4) containing 0.05% Tween 20 and 0.5% (wt/vol) bovine serum albumin (BSA) for Jejunal perfusion fluid gliadin, wheat, oatE, oatS, soy, casein, b-lactoglobulin, and A segment of the jejunum was perfused, as detailed codfish, or 0.5% (wt/vol) gelatin for a-lactalbumin, ovalbu- elsewhere, by a small diameter tube 175 cm long and min, and pork meat, in addition to NaCl (29 g/l) and KCl containing six channels.16 The original report described the (0.2 g/l) for both procedures. The reactions with secondary insertion of the tube, gastric drainage, inflation of the (rabbit antibody to human IgG or IgA) and tertiary (alkaline balloons, and rinsing of the closed intestinal segment. C- phosphatase conjugated swine anti-rabbit IgG) immuno- labelled polyethylene glycol was used as a volume marker reagents took place in the same buffers as those used for the and phenolsulphonphthalein (phenol red) as a marker of the respective primary steps. The final step with alkaline patency of the proximal balloon. The recovery of the volume phosphatase substrate took place in diethanolamine buffer marker was (mean (SD)) 86 (10)% in RA patients and 89 (pH 9.6) before reading of optical densities (OD) at 405 nm. (6)% in controls. The fluid was collected on ice, centrifuged in After coating, the ELISA plates were treated with 0.5% BSA cold conditions at 2500 6g, and frozen in aliquots of 2 ml in ammonium acetate for three hours; also the successive until analysis. Successful sampling was achieved from 14 RA reaction steps took place in the presence of BSA, which made patients and all controls. it possible to store coated plates beneath a moist filter paper in the refrigerator for a couple of weeks. With gelatine, Antigens however, the plates were preferably used immediately after Antigen solutions were prepared as described.17 Briefly, coating. Deionised water was used for washing after the gliadin (Karl Roth, Karlsruhe, Germany) and oatE represent coating and the blocking steps. For washings between ethanol (70%) extractable prolamins from wheat and oat sequential antibody reaction steps, Tween 20 (500 ml/l) was flour, respectively, while wheat and oatS represent the added to the water to inhibit non-specific binding of proteins comparable water soluble antigens. Purified antigens such to the plates. It was ensured that the substances added to the as a-lactalbumin, b-lactoglobulin, casein, and ovalbumin perfusion fluid did not interfere with the measurements. (Sigma Chemical Company, St Louis, Missouri, USA) were IgM antibody activities to the same antigens were measured both in serum (1/400) and intestinal fluid (1/10) as described above for IgG and IgA, with the following Table 1 Total levels (mg/l) of immunglobulins, including exceptions: The first step (test sample) was reduced from secretory IgA (SIgA) in jejunal perfusion fluid overnight to two hours, while the concentration of BSA was increased to 2.5% (wt/vol) and gelatine to 1% (wt/vol) Subjects IgA SIgA IgM IgG because non-specific binding is a problem when IgM RA patients antibodies are determined in serum. Blocking with BSA Median 33 25 5.4 5.2 was omitted after coating with soy and both of the oats Range 11 to 70 13 to 50 0.3 to 9.0 1.9 to 14 Quartile 14; 42 13; 36 3.0; 7.8 4.1; 7.1 preparations. With codfish, the gelatine buffer was used instead of the BSA buffer. Healthy controls The results are expressed in units per ml (U/ml) related to a Median 24 20 1.6 5.1 serum pool from patients with untreated coeliac disease,17 Range 5.0 to 56 2.8 to 44 0.3 to 7.3 1.1 to 10.8 Quartile 14.6; 31 11.7; 28 0.8; 2.3 2.5; 7.5 arbitrarily taken to contain 1000 U/ml of IgG and IgA antibodies and 250 U/ml of IgM antibodies, thus providing Probability p,0.33 p,0.20 p,0.0003 p,0.51 incomparable isotype and specificity values. Standard curves were constructed from serial dilutions of the reference serum, and sample readings were carried out using an ELISACalc

www.gutjnl.com Downloaded from gut.bmj.com on 17 October 2008 1242 Hvatum, Kanerud, Ha¨llgren, et al

A Jejunal fluid (IgA) B Jejunal fluid (IgG) 30 150 Patients Patients

100 20

50 10

30 150 Controls Codfish Controls Codfish Pork meat Pork meat Ovalbumin Ovalbumin

lgA against food (units/ml) Casein Casein lgG against food (units/ml) Σ α-Lactalbumin Σ α-Lactalbumin β β 100 -Lactoglobulin 20 -Lactoglobulin Soy Soy OatS OatS OatE OatE Gliadin Gliadin

50 10

Figure 1 Enzyme linked immunosorbent assay (ELISA) determinations of IgA (A) and IgG (B) antibody activities against various food antigens (see key) in jejunal perfusion fluid obtained from patients with rheumatoid arthritis (patients) (n = 14) and healthy controls (n = 20). Columns represent accumulative antibody levels, arranged in decreasing order for individual subjects. Note different scales on vertical axes. data program developed in our laboratory to provide even after full Bonferroni correction (table 2). Exceptions mathematical curve fitting.17 were IgM and IgG activities against b-1actoglobulin, and IgG activities against a-lactalbumin, ovalbumin, and soy. Statistical evaluation Likewise, intestinal IgA activity against gliadin and casein The Mann–Whitney two tailed non-parametric test was used showed no convincing increase (table 2). for statistical comparisons, with p,0.05 as the significance Interestingly, the two latter food proteins were the only level. Because the positive results obtained generally antigens against which the serum IgM activity was sig- appeared to be mutually correlated, as shown by Spearman nificantly raised in RA (fig 2B and table 3), perhaps reflecting correlation analyses (see later), full Bonferroni adjustment abundant antigen uptake owing to lack of a substantial for multiple tests would be too conservative (http://home.- mucosal IgA response. Also, serum and intestinal IgM clara.net/sisa/bonhlp.htm). The probability (p) values should activities to these two antigens were not correlated probably be adjusted somewhere between no correction and (r = 0.02); only IgM directed against b-1actoglobulin full; both these extremes are given in the tables, because it is (r = 0.75, p = 0.0005), oatS (r = 0.55, p = 0.023) and a- not possible to know exactly the extent to which the data lactalbumin (r = 0.54, p = 0.025) showed some relation in must be corrected. Notably, however, full Bonferroni correc- the two body fluids (uncorrected p values). Notably, the tion did not alter the main conclusions of the study. serum IgM activities to casein and ovalbumin tended to be correlated with the intestinal albumin level (r = 0.67, p,0025 and r = 0.59, p,0.05, respectively), possibly supporting the RESULTS idea that systemic immune activation depends on an Total immunoglobulin levels and food antibodies inadequate mucosal barrier function. Conversely, the intest- The total level of IgM in jejunal fluid of RA patients was inal IgM activities to all antigens were statistically unrelated significantly increased, whereas that of IgA and secretory IgA to any excessive leakage of albumin into the lumen (data not (SIgA) only showed a trend towards an increase (table 1). shown); and, as mentioned above, there was no indication of The median intestinal level of IgG was the same in RA a generally increased intestinal permeability for proteins in patients as in healthy control subjects, supporting the view the RA patients, because the average jejunal IgG level was that NSAID treatment had not caused any general increase in normal (table 1). mucosal permeability for intact proteins (see below). Jejunal IgA, IgG, and IgM activities to nearly all test antigens were highly or moderately increased in RA patients Antibody activities to food antigens are both related when compared with controls by ranking of accumulated and unrelated antibody levels (figs 1 and 2A). In particular, the IgM Because food intake and disease severity show an apparent activities were strikingly raised, and this elevation was connection in some RA patients only, we identified those unrelated to the total IgM levels (fig 2A). The antibody with increased intestinal antibody activities. In the control increases were generally significant or highly significant, group, IgA activities to most food antigens correlated with

www.gutjnl.com Downloaded from gut.bmj.com on 17 October 2008 Intestinal food antibodies in rheumatoid arthritis 1243

A Jejunal fluid (IgM) B Serum (IgM) 10 Patients Patients 300 2400 8

6 200 1600

4

100 800 2

Controls Wheat Controls Cod 300 Cod 2400 Pork Total IgM (mg/l) Total Pork 8 Ovalbumin Ovalbumin Casein

SlgM against food (units/ml) Casein SlgM against food (units/ml) α-Lactalbumin α β -Lactalbumin 6 -Lactoglobulin 200 β-Lactoglobulin 1600 Soy Soy OatS OatS Gliadin Gliadin 4

100 800 2

Figure 2 Enzyme linked immunosorbent assay (ELISA) determinations of IgM antibody activities against various food antigens (see key) in jejunal perfusion fluid (A) obtained from patients with rheumatoid arthritis (patients) (n = 13) and healthy controls (n = 20), and in serum (B) from patients (n = 17) and controls (n = 14). Columns represent accumulative antibody levels, arranged in decreasing order for individual subjects. Thin vertical bars in (A) represent total intestinal IgM concentrations for comparison with the individual IgM antibody levels. Note different scales on the three vertical axes. the total intestinal IgA levels. We therefore drew a linear (r = 0.61), soy (r = 0.67), cod fish (r = 0.74), and ovalbumin regression line for total IgA and IgA activity in controls and (r = 0.90); a-lactalbumin v ovalbumin (r = 0.55), casein considered RA activities to be increased when they were (r = 0.57), and oatE (r = 0.73); and oatE v gliadin above the limiting control lines, as suggested by Karol et al20 (r = 0.74). Conversely, IgA activities to most antigens were for serum IgE antibodies. However, IgG and IgM activities, as significantly correlated with each other in the controls. well as IgA antibodies to flour antigens (including soy), did Intestinal antibodies to one and the same antigen generally not show regression with the total isotype levels; for these did not correlate so well among the three Ig classes as did the activities the mean values for controls plus two standard activities of a specific isotype against different antigens, with deviations were considered the upper limit when recording the exception of antibodies to gliadin (r = 0.85). Otherwise an antibody increase in RA patients. the relations among different isotypes varied considerably The proportion of patients showing antibody increase (r = 0.06 to r = 0.61). depended both on the Ig class and on the type of food antigen. Thus patients deemed to have increased intestinal Effect of sulfasalazine IgA activity varied for flours from 7% (soy) to 21% (gliadin, Treatment of five RA patients with sulfasalazine for 16 weeks OatS), for cow’s milk proteins from none (casein) to 50% (a- reduced the increased food antibody levels seen initially in lactalbumin) or 57% (b-lactoglobulin), and for other types of three of them; this effect was striking in the two with the protein from 28% (ovalbumin) to 71% (cod fish) or 100% highest levels (fig 3), suggesting that sulfasalazine had an (pork meat). Increased IgG activity was less common and immunosuppressive effect on intestinal immune responses. It varied from 14–21% (a-lactalbumin, b-lactoglobulin, casein, has also been proposed that this drug can diminish mucosal ovalbumin, soy, oats) to 36–57% (pork meat, cod fish, permeability, but luminal albumin levels were not consis- gliadin). IgM activities were often increased, varying from tently decreased after the treatment, regardless of whether 36–50% (pork meat, cod fish, gliadin, b-1actoglobulin) to 57– the patients had received NSAIDs recently or not (fig 4); 86% (soy, OatS, a-lactalbumin, casein, ovalbumin). neither was the antibody decrease accompanied by any Individual intestinal IgM activities to different antigens apparent reduction of total Ig levels in jejunal fluid (data not were positively correlated in the RA patients (r = 0.68 to 0.99) shown)—in contrast to our observations after sulfasalazine and also in the controls, although at lower magnitude treatment in a contemporary study of patients with ankylos- (r = 0.33 to 0.93). IgG activities showed positive correlations ing spondylitis.19 In that disease, the suppressive effect on the for most antigens both in RA patients (r = 0.68 to 0.97) and intestinal IgM (p,0.01) and SIgA (p,0.002) levels was controls (r = 0.51 to 0.99), but usually of lower magnitude accompanied by clinical improvement and reduced ESR than for IgM activities in the patients. IgA activities did not (p,0.004).19 In the few RA patients subjected to sulfasalazine correlate in the RA patients except against some antigens: treatment, however, the clinical effect was associated with pork meat v a-lactalbumin (r = 0.53), b-lactoglobulin reduced ESR but not necessarily with decreased jejunal

www.gutjnl.com Downloaded from gut.bmj.com on 17 October 2008 1244 Hvatum, Kanerud, Ha¨llgren, et al

Table 2 Relative IgA, IgG, and IgM activity levels (ELISA units/ml, median and range) to various food antigens in jejunal perfusion fluid from patients with rheumatoid arthritis and healthy control subjects

Antigens Patients Controls Probability*

IgA Gliadin 5.3 (1.3 to 75) 2.8 (0.5 to 20.0) p = 0.068 (p = 0.68) OatE 7.0 (2.5 to 25.0) 3.8 (0.4 to 9.7) p = 0.029 (p = 0.29) OatS 2.1 (0.9 to 9.3) 0.9 (0.3 to 3.8) p = 0.0026 (p = 0.026) Soy 2.2 (0.7 to 6.2) 1.0 (0.1 to 5.8) p = 0.027 (p = 0.27) b-lactoglobulin 2.1 (0.9 to 6.3) 0.8 (0.1 to 2.3) p,,0.0004 (p,0.004) a-lactalbumin 4.6 (1.5 to 21.0) 1.8 (0.8 to 5.5) p = 0.0086 (p = 0.086) Casein 2.0 (0.3 to 5.7) 1.5 (0.1 to 7.5) p = 0.301 (p = 1.0) Ovalbumin 2.7 (0.9 to 11.0) 1.7 (0.1 to 4.8) p = 0.016 (p = 0.16) Pork meat 8.8 (3.3 to 25.0) 2.4 (0.3 to 7.5) p,,0.0004 (p = ,0.004) Codfish 5.3 (1.7 to 10.0) 1.9 (0.3 to 7.5) p = 0.0004 (p = 0.004)

IgG Gliadin 1.1 (0.3 to 4.5) 0.09 (0.03 to 1.1) p,0.0004 (p,0.004) OatE 1.0 (0.3 to 3.0) 0.20 (0.03 to 0.9) p,0.0004 (p,0.004) OatS 0.3 (0.03 to 1.1) 0.04 (0.03 to 0.8) p = 0.0086 (p = 0.086) Soy 0.4 (0.03 to 1.4) 0.08 (0.03 to 0.7) p = 0.0147 (p = 0.147) b-lactoglobulin 0.3 (0.03 to 1.9) 0.16 (0.03 to 0.6) p = 0.149 (p = 1.0) a-lactalbumin 0.7 (0.03 to 2.8) 0.15 (0.04 to 1.7) p = 0.086 (p = 0.86) Casein 0.5 (0.04 to 1.9) 0.04 (0.03 to 0.9) p = 0.0028 (p = 0.028) Ovalbumin 0.7 (0.04 to 4.3) 0.16 (0.03 to 1.8) p = 0.023 (p = 0.23) Pork meat 1.5 (0.1 to 5.8) 0.27 (0.04 to 0.4) p = 0.0006 (p = 0.006) Codfish 0.9 (0.07 to 3.5) 0.18 (0.03 to 0.6) p = 0.0004 (p = 0.004)

IgM Gliadin 10.9 (3.6 to 45.2) 3.0 (0.7 to 12.8) p,0.0004 (p,0.0036) OatE Not done Not done OatS 7.8 (3.5 to 33.7) 2.7 (1.2 to 6.8) p,,0.0004 (p,0.0036) Soy 5.3 (1.8 to 21.8) 1.7 (1.3 to 3.1) p,0.0004 (p,0.0036) b-lactoglobulin 3.8 (0.5 to 24.8) 1.2 (0.5 to 4.1) p = 0.031 (p = 0.279) a-lactalbumin 5.3 (2.3 to 37.1) 1.4 (0.5 to 2.3) p,,0.0004 (p,0.0036) Casein 4.3 (2.3 to 30.5) 0.6 (0 to 2.1) p,,0.0004 (p,0.0036) Ovalbumin 3.5 (1.6 to 26.8) 0.5 (0 to 1.3) p,,0.0004 (p,0.0036) Pork meat 9.8 (3.1 to 34.9) 2.9 (0.4 to 20.0) p = 0.0014 (p = 0.0126) Codfish 8.8 (3.0 to 40.0) 2.8 (0.1 to 22.0) p = 0.0016 (p = 0.0144)

*The probability that food specific IgA, IgG, or IgM activity is increased compared with matched controls (Mann–Whitney two tailed test). p Values in parentheses were subjected to full Bonferroni adjustment for multiple tests. ELISA, enzyme linked immunosorbent assay; OatE. oat flour extracted with 70% ethanol; OatS, oat flour extracted with ammonium acetate. antibody production (fig 3). Because a high protein intake had no effect on the readings, or could even increase them— gives an OR of only 2.9 for polyarthritis,3 a much larger study probably because of the formation of soluble immune group would clearly be required to demonstrate convincingly complexes which remained able to react with the coat. We an association between intestinal food antibodies and the therefore carried out antibody absorption with a large excess severity of RA. of gliadin, which is insoluble in aqueous solution. Thus intestinal fluid (fig 5A) and serum samples (fig 5B) were Cross reactivity of food antibodies revealed by tested in ELISA for remaining IgA and IgM antibody absorption test activities after being mixed with gliadin (0.1 g/ml) overnight Varying amounts of the coating antigen (1.56 to 10 mg/ml) on a shaker at room temperature. Residual IgM activity added to intestinal fluid the day before the ELISA generally against gliadin was only 0%–18% and against unrelated

Table 3 Relative IgM activity levels (ELISA units/ml, median and range) against various food antigens in serum samples from patients with rheumatoid arthritis and healthy control subjects

Antigens Patients Controls Probability*

Gliadin 597 (104 to 1087) 239 (651 to 1396) p,0.0004 (p,0.0036) OatS 270 (36 to 627) 136 (25 to 575) p = 0.057 (p = 0.513) Soy 159 (34 to 403) 131 (39 to 774) p = 0.575 (p = 1.0) b-lactoglobulin 61 (10 to 344) 32 (5 to 101) p = 0.074 (p = 0.666) a-lactalbumin 46 (2 to 103) 52 (18 to 158) p = 0.189 (p = 1.0) Casein 191 (43 to 523) 46 (1 to 204) p,,0.0004 (p,0.0036) Ovalbumin 71 (12 to 121) 49 (1 to 130) p = 0.18 (p = 1.0) Pork meat 54 (31 to 98) 109 (34 to 159) p = 1.0 Codfish 84 (35 to 200) 123 (39 to 381) p = 1.0

*The probability that food specific IgM activity is increased compared with matched controls (Mann–Whitney two tailed test). p Values in parentheses were subjected to full Bonferroni adjustment for multiple tests. ELISA, enzyme linked immunosorbent assay; OatS, oat flour extracted with ammonium acetate.

www.gutjnl.com Downloaded from gut.bmj.com on 17 October 2008 Intestinal food antibodies in rheumatoid arthritis 1245

Figure 3 Enzyme linked P1 P2 P3 P4 P5 immunosorbent assay determinations of IgM, IgA, and IgG antibody activities IgM against various food antigens as Gliadin 40 indicated (see key) in jejunal perfusion OatE fluid obtained from five patients with OatS rheumatoid arthritis (P1–P5) before (Be) Wheat and after (Af) treatment with Soy 30 β-Lactoglobin sulfasalazine for 16 weeks. P1, P3, and Casein P4 had not received NSAIDs for at least α-Lactalbumin one month before sampling (see fig 4). Ovalbumin The effect of sulfasalazine on the 20 Pork erythrocyte sedimentation rate (ESR) Codfish and clinical improvement (+ to ++)or worsening (2) following the treatment is indicated at the bottom. Note different scales on vertical axes. NSAID, non- 10 steroidal anti-inflammatory drug.

20 IgA

15

Specific antibody activity (units/ml) 10

5

6 IgG

4

2

Be Af Be Af Be Af Be Af Be Af

ESR57 119 70 34 35 7 48 16 38 18 Clinical _ +++++++ effect

80 Figure 4 Radioimmunoassay antigens 50–85%. Remaining IgA activity was measured only determinations of albumin in in one intestinal fluid (fig 5A); with most antigens the jejunal perfusion fluid obtained proportion of residual antibody activity was lower than that from the same patients with for IgM. 70 rheumatoid arthritis (P1–P5) reported in fig 3; those encircled had not received NSAIDs for at DISCUSSION 60 least one month. There was no P4 consistent difference in the This is the first extensive study of intestinal food antibodies albumin levels before (Be) or after in RA patients. Despite considerable variability—which can (Af) sulfasalazine treatment for be expected for antibody levels even in healthy adults 50 16 weeks (medians indicated by regardless of age and sex21—our results were remarkably horizontal bars). NSAID, non- positive in RA, particularly for the IgM class, and included P1 steroidal anti-inflammatory drug. most test antigens in a surprising manner. Absorption with 40 insoluble gliadin revealed a substantial level of cross mg/l P3 reactivity for both IgM and IgA antibodies. Notably, even in the normal state, SIgA in various human secretions has been 30 reported to show a relatively high level of cross reactivity, recognising both self and microbial antigens,22 but apparently P2 without involving peritoneally derived B1 cells, in contrast to 23 20 P5 the situation in mice. Rather, it reflects a substantial innate drive of the intestinal immune system.23 The IgM reactivity in jejunal fluid was only marginally 10 related to that in serum of the same patient, and it was neither related to total IgM levels nor to mucosal protein permeability as deemed by mucosal leakage of albumin or 0 IgG. Therefore, a truly RA related mucosal production of Be Af antibodies was strongly suggested, rather than excessive

www.gutjnl.com Downloaded from gut.bmj.com on 17 October 2008 1246 Hvatum, Kanerud, Ha¨llgren, et al

36 100 antibodies used in the assay. A panel of food antigens to A IgM Perf 455 IgM Perf 451 IgA Perf 451 document gut antibody cross reactivity in RA has apparently not been used before, although it should be pointed out that

n 80 we have not demonstrated polyreactivity in the formal sense. Intestinal IgM and IgA with RF activity have been observed 60 in patients with untreated coeliac disease, and the IgM RF level was quite high in coeliac patients with IgA deficiency.37 Mucosal RF synthesis is apparently linked to the gluten 40 response because RF in serum from patients with coeliac disease or dermatitis herpetiformis was carried only by IgA.38 20 Furthermore, SIgA RF has been detected in serum from RA patients, so some intestinal RF synthesis may take place also in RA.11 However, we found that intestinal fluid from RA 100 patients contained only low levels of IgA and IgM RF, some B IgM Ser 457 IgM Ser 451 IgM Ser 453 1000 times less than in serum.15 The IgA, but notably not the 80 IgM, RF activities were generally well correlated with the food antibody levels of all the three Ig classes (r = 0.65 to 0.94; p = 0.01 to 0.0001). 60 Multispecific antibodies may exist in antigen complexes.39 In the gut, such complex formation depends on antigen 40 stability and on pH dependent pepsin hydrolysis. Thus infants are prone to develop cow’s milk allergy while their gastric acidity is pH 3–4 (compared with pH 2 in adults); at

Remaining antibody activity (%) to food after gliadin inhibitio 20 pH 4 the degradation of a-lactalbumin, BSA, and bovine IgG is markedly reduced in contrast to b lactoglobulin.40 Some 80% of untreated RA patients have been shown to have 41

Gil reduced maximum gastric acid output, which could -La -La Soy Cas Cod Pork Ova β α OatS OatE contribute to enhanced food immunoreactivity. Wheat Antigen A germ-free state prevents the development of gut and joint inflammation in HLA-B27 transgenic rats, thereby Figure 5 Enzyme linked immunosorbent assay determinations of IgA giving strong support to a connection between mucosal and IgM antibody activities (per cent of original levels) against various immunity and arthritis.42 Also, reactive arthritis in humans food antigens as indicated (see key), remaining in two different perfusion fluids (Perf 451 and Perf 455) obtained from the jejunum (A) and in three appears to be caused by a combination of a mucosa 43 serum samples (B) of patients with rheumatoid arthritis after absorption associated microbial impact and genetic predisposition. of the samples with an excess of insoluble gliadin. Interestingly, some 90% of patients with reactive arthritis or ankylosing spondylitis express HLA-B27, and these disorders can be associated with Crohn’s disease, ulcerative colitis, and immunostimulation after potentially NSAID induced absorp- jejuno-ileal bypass surgery43—again emphasising the putative tion of dietary antigens.12 Also, notably, two patients who had gut–joint axis which is also supported by shared homing not received NSAIDs for at least one month showed increased properties of activated intestinal immune cells.44 intestinal IgM and IgA food reactivity as well. Increased Moreover, animal experiments have demonstrated a wide- levels of circulating SIgA associated with IgA RF complexes spread tissue distribution of food antigens shortly after have been observed in RA patients,11 likewise suggesting that feeding,45 which could predispose to synovial immune intestinal immunity is overactivated in this disease. complex formation and thereby autoimmune joint reac- RA sera are known to contain increased amounts of so tions.27 We have previously reported that intestinal levels of called ‘‘natural antibodies’’, which are encoded by germ line IgM and IgA are increased in patients with ankylosing Ig variable genes with only few or no somatic mutations. spondylitis related to disease activity.19 Antigens from the gut Such antibodies display a broad array of mostly autoimmune microbiota rather than food are apparently involved in that activity, perhaps enabling them to clear waste products.24 In disease,43 44 because the IgM reactivity to dietary antigens was serum, this activity has been thought to represent low avidity not different from normal control levels (our unpublished IgM antibodies but is, instead, mainly of the IgG class.24 RF is observations), in striking contrast to the data presented here mostly of the IgM, IgG, or IgA class25 and appears to be for RA. Disparate antigenic or mitogenic stimulation in the subjected to antigen driven mutation in RA, but still showing gut might explain the different response to sulfasalazine substantial cross reactivity, like other autoantibodies.26 treatment noted in the two disorders with regard to reduction Therefore, it cannot be excluded that food antigens are of total intestinal immunoglobulin levels as mentioned in involved in the rheumatoid factor induction process, accord- Results. ing to the ‘‘multiple hit model’’ for RA.27 The underlying immunoregulatory defect may involve both poor activity of Conclusions regulatory T cells28 and impaired early B cell tolerance.29 Both systemic and intestinal humoral immunity was found to Polyreactive antibodies most probably have a flexible be aberrant in many RA patients, with a particularly striking antigen binding ‘‘pocket’’ that can accommodate different elevation of cross reactive food antibodies in proximal gut antigens.30 Polyreactivity of RF could hence explain the secretions. IgM reactivity against some food items was overall tendency to increased IgM activities in RA serum increased also in serum, but relatively much less so. irrespective of test antigen, for instance Proteus mirabilis, Measurements of serum antibodies (except for RF) appears staphylococcal enterotoxin B, b2 microglobulin, and cyto- to be of little informative value in RA. Conversely, measure- kines.31–34 It should also be noted that investigation of food ments of intestinal antibodies provide more striking results, antibodies in RA sera is difficult because RF of different Ig suggesting a connection between mucosal immune activation classes may cause assay interference.35 Another complication and the pathogenesis of RA, at least in some patients. Their is that IgM RF binds IgG from other species, including the food related problems may reflect the additive effect of

www.gutjnl.com Downloaded from gut.bmj.com on 17 October 2008 Intestinal food antibodies in rheumatoid arthritis 1247 multiple modest hypersensitivity reactions mediated—for 17 Hvatum M, Scott H, Brandtzaeg P. Serum IgG subclass antibodies to a variety of antigens in patients with coeliac disease. Gut 1992;33:632–8. instance, by immune complexes—which could predispose 18 Kanerud L, Engstro¨m GN, Tarkowski A. Evidence for differential effects of 27 the joints for autoimmune tissue destructive reactions. sulphasalazine on systemic and mucosal immunity in rheumatoid arthritis. Ann Rheum Dis 1995;54:256–62. 19 Feltelius N, Hvatum M, Brandtzaeg P, et al. Increased jejunal secretory lgA ACKNOWLEDGEMENTS and IgM in ankylosing spondylitis: normalization after treatment with Supported by the Research Council of Norway, the Norwegian sulfasalazine. J Rheumatol 1994;21:2076–81. Rheumatological Society, the Swedish Research Council, the Swedish 20 Karol MH, Kramarik JA, Ferguson J. Methods to assess RAST results in Association Against Rheumatism, the King Gustaf V’s 80 year Fund, patients exposed to chemical allergens. Allergy 1995;50:48–54. and Professor Nanna Svartz’ Foundation. Kathrine Hagelsteen is 21 Cummings JH, Antoine JM, Azpiroz F, et al. PASSCLAIM—gut health and thanked for excellent assistance with ELISA, and Hege Eliassen for immunity. Eur J Nutr 2004;43(suppl 2):II118–73. assistance with the manuscript. 22 Quan CP, Berneman A, Pires R, et al. Natural polyreactive secretory immunoglobulin A autoantibodies as a possible barrier to infection in humans. Infect Immun 1997;65:3997–4004...... 23 Brandtzaeg P, Johansen F-E. Mucosal B cells: phenotypic characteristics, Authors’ affiliations transcriptional regulation, and homing properties. Immunol Rev M Hvatum, P Brandtzaeg, Laboratory for Immunohistochemistry and 2005;206:32–63. Immunopathology (LIIPAT), Institute of Pathology, University of Oslo, 24 Avrameas S. Natural autoantibodies: from ‘‘horror autotoxicus’’ to ‘‘gnothi seauton’’. Immunol Today 1991;12:154–9. Rikshospitalet University Hospital, Oslo, Norway 25 Jefferis R. Rheumatoid factors, B cells and immunoglobulin genes. Br Med Bull L Kanerud, Department of Rheumatology, Karolinska Institute, Stockholm 1995;51:312–31. So¨der Hospital, Stockholm (current address: Rheumatology Unit, 26 Levinson SS. Humoral mechanisms in autoimmune disease. J Clin Karolinska Institute, Farsta La¨karhus, Farsta), Sweden Immunoassay 1994;17:72–84. RHa¨llgren, Department of Medical Sciences, University Hospital, 27 van Gaalen F, Ioan-Facsinay A, Huizinga TW, et al. The devil in the details: the emerging role of anticitrulline autoimmunity in rheumatoid arthritis. Uppsala, Sweden J Immunol 2005;175:5575–80. Conflict of interest: None declared. 28 Ehrenstein MR, Evans JG, Singh A, et al. Compromised function of regulatory T cells in rheumatoid arthritis and reversal by anti-TNFalpha therapy. J Exp Med 2004;200:277–85. 29 Samuels J, Ng YS, Coupillaud C, et al. Impaired early B cell tolerance in REFERENCES patients with rheumatoid arthritis. J Exp Med 2005;201:1659–67. 1 Haugen M, Fraser D, Førre Ø. Diet therapy for the patient with rheumatoid 30 Notkins AL. Polyreactivity of antibody molecules. Trends Immunol arthritis? Rheumatology (Oxford) 1999;38:1039–44. 2004;25:174–9. 2 Panush RS. Food induced (‘‘allergic’’) arthritis: clinical and serologic studies. 31 Senior BW, McBride PD, Morley KD, et al. The detection of raised levels of IgM J Rheumatol 1990;17:291–94. to Proteus mirabilis in sera from patients with rheumatoid arthritis. J Med 3 Pattison DJ, Symmons DP, Lunt M, et al. Dietary risk factors for the Microbiol 1995;43:176–84. development of inflammatory polyarthritis: evidence for a role of high level of 32 Origuchi T, Eguchi K, Kawabe Y, et al. Increased levels of serum IgM antibody red meat consumption. Arthritis Rheum 2004;50:3804–12. to staphylococcal enterotoxin B in patients with rheumatoid arthritis. Ann 4 van de Laar MA, van der Korst JK. Food intolerance in rheumatoid arthritis. I. Rheum Dis 1995;54:713–20. A double blind, controlled trial of the clinical effects of elimination of milk 33 Williams RC, Malone CC, Tsuchiya N. Rheumatoid factors from patients with allergens and azo dyes. Ann Rheum Dis 1992;51:298–302. rheumatoid arthritis react with b2–microglobulin. J Immunol 5 Kjeldsen-Kragh J, Haugen M, Borchgrevink CF, et al. Controlled trial of 1992;149:1104–13. fasting and one-year vegetarian diet in rheumatoid arthritis. Lancet 34 Hansen MB, Andersen V, Rohde K, et al. Cytokine autoantibodies in 1991;338:899–902. rheumatoid arthritis. Scand J Rheumatol 1995;24:197–203. 6 Darlington LG, Jump A, Ramsey NW, et al. A prospective study of clinical and 35 Levinson SS. Antibody multispecificity in immunoassay interference. Clin serologic responses to single or double food challenges in patients with Biochem 1992;25:77–87. rheumatoid arthritis. Br J Rheumatol 1989;(suppl 2):116. 36 Hamako J, Ozeki Y, Matsui T, et al. Binding of human IgM from a rheumatoid 7 Kjeldsen-Kragh J, Hvatum M, Haugen M, et al. Antibodies against dietary factor to IgG of 12 animal species. Comp Biochem Physiol B Biochem Mol Biol antigens in rheumatoid arthritis patients treated with fasting and a one-year 1995;112:683–88. vegetarian diet. Clin Exp Rheumatol 1995;13:167–72. 37 Ha¨llgren J, Knutson F, Lavo¨B,et al. Increased mucosal synthesis of 8 Koot VC, van Straaten M, Hekkens WT, et al. Elevated level of IgA gliadin rheumatoid factor (RF) in coeliac disease. Clin Exp Immunol 1996;103:94–8. antibodies in patients with rheumatoid arthritis. Clin Exp Rheumatol 38 So¨kjer M,Jo¨nsson T, Bo¨dvarsson S, et al. Selective increase of lgA rheumatoid 1989;7:623–26. factor in patients with gluten sensitivity. Acta Derm Venereol 9 O’Farrelly C, Marten D, Melcher D, et al. Association between villous atrophy 1995;75:130–32. in rheumatoid arthritis and a rheumatoid factor and gliadin—specific IgG. 39 Sigounas G, Kolaitis N, Monell-Torrens E, et al. Polyreactive IgM antibodies in Lancet, 1988;ii, 819–22. the circulation are masked by antigen binding. J Clin Immunol 10 MacGillivray AJ, Fernandes L, Waiters M, et al. Antibodies to wheat gliadin in 1994;14:375–81. rheumatoid arthritis. Br J Rheumatol 1994;33(suppl 1):146. 40 Schmidt DG, Meijer RJ, Slangen CJ, et al. Raising the pH of the pepsin- 11 Jo¨rgensen C, Moynier M, Bologna C, et al. Rheumatoid factor associated with catalysed hydrolysis of bovine whey proteins increases the antigenicity of the a secretory component in rheumatoid arthritis. Br J Rheumatol hydrolysates. Clin Exp Allergy 1995;25:1007–17. 1995;34:236–40. 41 Kanerud L, Hafstro¨m I, Berg A. Effects of antirheumatic treatment on gastric 12 Bjarnason I, Peters TJ. Influence of anti-rheumatic drugs on gut permeability secretory function and salivary flow in patients with rheumatoid arthritis. Clin and on the gut associated lymphoid tissue. Baillieres Clin Rheumatol Exp Rheumatol 1991;9:595–601. 1996;10:165–76. 42 Taurog JD, Richardson JA, Croft JT, et al. The germfree state prevents 13 Katz KD, Hollander D. Intestinal mucosal permeability and rheumatological development of gut and joint inflammatory disease in HLA–B27 transgenic diseases. Baillieres Clin Rheumatol 1989;3:271–84. rats. J Exp Med 1994;180:2359–64. 14 Arnett FC, Edworthy SM, Bloch DA, et al. The American Rheumatism 43 Lichtman SN. Role of endogenous enteric organisms in the reactivation of Association 1987 revised criteria for the classification of rheumatoid arthritis. arthritis. Mol Med Today 1995;1:385–91. Arthritis Rheum 1988;31:315–24. 44 Brandtzaeg P. Review article: Homing of mucosal immune cells—a possible 15 Ma¨ki-Ikola O,Ha¨llgren R, Kanerud L, et al. Enhanced jejunal production of connection between intestinal and articular inflammation. Aliment Pharmacol antibodies to Klebsiella and other Enterobacteria in patients with ankylosing Ther 1997;11(suppl 3):24–39. spondylitis and rheumatoid arthritis. Ann Rheum Dis 1997;56:421–5. 45 Gu¨tgemann I, Fahrer AM, Altman JD, et al. Induction of rapid T cell activation 16 Knutson L, Odlind B, Ha¨llgren R. A new technique for segmental jejunal and tolerance by systemic presentation of an orally administered antigen. perfusion in man. Am J Gastroenterol 1989;84:1278–84. Immunity 1998;8:667–73.

www.gutjnl.com Toward An Understanding of Allergy and In-Vitro Testing By Mary James, N.D.

Learning to recognize Food represents the largest anti- don’t will be discussed, or why a genic challenge facing the immune specific food elicits symptoms in and manage food system. Assuming complete diges- an individual at one point in time, allergies can go a long tion, an intact intestine, a sturdy yet appears to be well tolerated at constitution, and minimal anti- other times. We will examine why way in achieving better genic exposure such that the an individual sometimes feels bet- immune system is not over- ter from eating an allergenic food, clinical results whelmed, all goes well. Weaknesses but feels worse from eating it fol- with patients. in one or more of these areas, lowing a period of elimination. We however, can result in immune will look at why an allergy test attacks upon foods as if they were might be normal in an individual foreign invaders. A long list of who knows he experiences symp- conditions have been associated toms when ingesting certain foods, with food reactions, including or why an in-vitro test may show fatigue, migraines, irritable bowel elevated antibodies to foods in an syndrome, inflammatory bowel asymptomatic person. Finally, we disease, gallbladder disease, arthri- will discuss how to effectively man- tis, asthma, rhinitis, Attention age allergenic food elimination, Deficit Hyperactivity Disorder reintroduction and rotation. While (ADHD), enuresis, epilepsy, there are admittedly many useful eczema, psoriasis, aphthous ulcers, approaches to the diagnosis and and recurrent sinusitis, otitis media treatment of allergic disorders, the and other infections.(1) A patient following discussion will be limited with numerous and seemingly primarily to in-vitro assessment unrelated symptoms often moves and dietary management. from doctor to doctor in search of a diagnosis. When the inflammato- The Immune System ry response to an allergen takes Before embarking on a discussion hours or days to develop, the of allergy, let’s start with a few relationship between foods and basics of the immune response. symptoms is often hard to pin The following overview will provide down. Learning to recognize and a foundation for our discussion. manage food allergies can go a long way in achieving better clinical The principle cells of the immune results with these patients. The system are lymphocytes, plasma intent of this paper is to provide cells and macrophages, collectively such an understanding. organized into lymphoid tissue. Lymphocytes can be further divid- In this paper, we will focus on IgE- ed into B-cells and T-cells. and IgG-mediated immune reac- Interestingly, these cell names were tions. Reasons why some individu- based on parts of a chicken, as 2 als develop allergies while others studies of the thymus (T) and the bursa of Fabricius (B) (a lymphoid reactions to foods and other sub- organ near the cloaca) brought stances. Examples include lactose about the first understanding of intolerance, pharmacological their respective immunological responses to alkaloids in foods functions. The chief role for B-cells such as solanine (potato family), is the provision of humoral immu- salicylate sensitivity, and lectin nity. Upon exposure to an antigen, reactions, in which dietary lectins B-cells proliferate and evolve to interact with surface antigens on antibody-synthesizing plasma cells cells, causing them to agglutinate. that then produce antigen-specific Bacteria and bacterial toxins may immunoglobulins of different isotypes, called IgM, IgG, IgE, IgD and IgA. T-cells, on the Non-immune Mediated Reactions to Foods other hand, provide cell-mediat- ed immunity. • Lactase deficiency (dairy) ➔ bloating, flatulence, diarrhea, abdominal pain • Spoilage or contamination of food by bacteria (Proteus) or heat-stable toxins There is considerable interaction (tuna, bonita, mackerel) ➔ itching, rash, vomiting, diarrhea among components of the • Gallbladder disease ➔ abdominal pain (RUQ), nausea, flatulence, aggravation by fats immune system. Subsets of T- cells, function as: helper T-cells, • Vasoactive amines -phenylethylamine (chocolate, aged cheese, red wine) ➔ migraine which stimulate B-cell activity; -tyramine (cheddar cheese, French cheeses, brewer’s yeast, chianti, canned fish) ➔ or as suppressor T-cells which migraine, erythema, urticaria, hypertensive crises in patients on MAO inhibitors suppress both humoral and cell -histamine (fermented cheese, fermented foods (e.g. sauerkraut), pork sausage, mediated immune responses. canned tuna, anchovies, sardines) ➔ erythema, headache, hypotension Macrophages are released from -histamine-releasing agents (shellfish, chocolate, strawberries, tomatoes, peanuts, ➔ the bone marrow as monocytes, pork, wine, pineapple) urticaria, eczema, pruritis and develop into macrophages • Food additives (e.g. tartrazine, FD&C Yellow No. 5, sodium benzoate) ➔ hives, upon entering tissue. rash, asthma Macrophages serve to present • Sulfites (salad bar lettuce, shrimp, dried fruits and vegetables, wine, beer) ➔ asthma antigen to both T- and B-cells, or anaphylaxis, loss of consciousness as well as to clear antigen/anti- • Monosodium glutamate (Chinese and Japanese dishes) ➔ headache, facial tension, body complexes from the circu- sweating, chest pain, dizziness lation. The efficiency of this • “Nightshade” alkaloids (potatoes, tomatoes, eggplant, peppers, tobacco) ➔ joint function is critical in controlling pain food-induced hypersensitivity ➔ reactions, as we will discuss. • Hypoglycemia Fatigue, palpitations, shakiness, cognitive impairment, mood swings, poor memory, blurred vision, anxiety, dizziness, headache • Lectin reactions (wide variety of foods) ➔ wide variety of symptoms, depending on Getting Our Terms blood type compatibility. (Refer to: Eat Right 4 Your Type, by Peter D’Adamo, N.D.) Straight The terms "food intolerance," "allergy" and "hypersensitivity" are cause gastrointestinal and systemic often used interchangeably. For reactions, such as in scombroid clarification, it’s useful to poisoning from the ingestion differentiate among them… of contaminated tuna. Vasoactive amines (epinephrine, norepineph- Intolerance rine, tyramine, dopamine, hista- mine and 5-hydroxytryptamine) The term "intolerance" generally are found in bananas, tomatoes, applies to non-immune mediated avocados, cheeses, pineapples and 3 wines, and can contribute to •Type II immune reactions involve symptoms such as migraine antibody-mediated destruction of headache. Sulfite, used as a tissue following adherence of preservative in foods such as let- foreign material. This reaction is tuce, shrimp, dried fruit and wine, often referred to as a "cytotoxic can cause asthma and urticaria.(2) reaction." Examples of Type II Some compounds, such as alco- reactions include penicillin reac- Any food eliciting an hol, may even induce histamine tions and those resulting in red responses, although the reaction is cell or platelet destruction. adverse reaction should not immune mediated. •Type III reactions are mediated ideally be avoided, no Because of the conspicuous by mixed immunoglobulins, but matter what the relationship between ingestion primarily IgG. Complexes com- of the food and the onset of posed of antigen and antibody mechanisms, and symptoms, such reactions are fre- activate complement and cytokines quently mistaken for allergy, yet an in the body, resulting in an inflam- careful investigation allergy test may be negative for that matory response. Type III reactions aimed toward the food. Any food eliciting an adverse constitute the basis of "delayed- reaction should ideally be avoided, onset" food allergies. Symptoms causative factors. no matter what the mechanism, are delayed because of the time and careful investigation aimed required for the formation of toward the causative factors. complexes. These reactions will also be discussed in more detail.

Hypersensitivity •Type IV refers to cell-mediated Although the use of the term immune reactions, where T-cells "hypersensitivity" is sometimes act as the primary players. T-cells reserved only for Gell and Coombs’ become cytotoxic cells when acti- classification of Type III, IgG- vated by antigen, capable of killing mediated reactions, traditionally viruses, bacteria, tumor cells or the term is applied to all four types other target cells. Type IV reactions of tissue injury. Types I through IV play a significant role in tuberculo- all depend upon the interaction of sis, mycotic and viral infections, antigen with humoral antibody, contact dermatitis and allograft and result from an excessive rejection. These reactions may also immune reaction to antigen, be involved in some food allergies, leading to gross tissue changes such as protein-losing entero- and symptoms. pathies and celiac disease.

•Type I reactions are mediated by Allergy IgE antibodies, and are character- The definition of the term "allergy" ized by the release of histamine is much debated. Although many and other chemical mediators traditional allergists strictly reserve upon exposure to an allergen. the use of the term for Type I IgE- Type I reactions are responsible mediated reactions such as hay for "immediate-onset" allergies, fever, a more general definition of such as hay fever or anaphylaxis. "allergy" refers to any acquired They will be discussed in more hypersensitivity to an antigen that detail below. results in harmful immunologic 4 consequences. For our purposes, bronchial and cutaneous allergies we will apply this broader defini- show a familial tendency. This is par- tion to both Type I and Type III ticularly true for asthma, hay fever, immune reactions. recurrent rhinitis and bronchitis, and eczema; when present in a parent, Immediate-Onset IgE there is an increased prevalence of Reactions the same disorder in the child.(3) If neither parent is allergic, a patient’s When an antigen attaches to IgE IgE antibody levels may symptoms are less likely to reflect an antibodies already stationed on a IgE-mediated reaction. IgE allergies drop with avoidance mast cell or basophil, pre-formed are in place for life. Antibody levels bundles of histamine are released. may drop with avoidance of expo- of exposure, but Inhalation of antigens usually sure, but re-exposure will quickly leads to the symptoms we com- re-exposure will quickly result in the mobilization of IgE anti- monly associate with "allergy" bodies and the subsequent release of result in the such as sneezing, itching of the histamine. palate or ears, runny nose, itching mobilization of and tearing in the eyes, and fatigue. As mentioned above, IgE-mediated Ingestion of an antigen may lead to antibodies and the food allergies are usually easy to symptoms such as asthma, an itchy spot because of the immediate subsequent release rash, angioedema or gastrointesti- appearance of symptoms. Common nal symptoms such as abdominal culprits include peanuts and shell- of histamine. cramps and diarrhea. fish. Sensitivity to these substances is so extreme in some cases that Chronic IgE allergies may manifest anaphylaxis may result from the as sinusitis, recurrent ear or upper mere inhalation of vapors from respiratory infections, mouth food or contact with the skin. breathing and post-nasal drip. Needless to say, management of "Allergic shiners" under the eyes these allergies requires strict avoid- and a white line across the nose, ance of the offending substance. from repeated upward wipes of the IgE reactions triggered by inhalants nose with one’s hand, are common often show a seasonal pattern, with signs of IgE-mediated allergies in reactions to tree pollens children, although these signs have typically occurring in the spring, also been observed in individuals grass pollens in late spring and early with IgG allergies. Severe Type I summer, and weed pollens in late reactions may include anaphylaxis. summer and early fall. Reactions to The allergic reaction occurs imme- dust mites often appear in the win- diately upon exposure to the anti- ter, with the onset of home heating. gen (usually less than two hours); Allergy testing during a sympto- consequently, the person experi- matic period can help to identify encing the reaction usually easily the responsible antigens. Allergies recognizes the link between aller- to pet dander can occur at any gen and symptom. time, upon exposure. Specific IgE-mediated allergies are not inherited, although Type I aller- Mold and Fungi Reactions gic diseases in general tend to run Mold allergies tend to coincide in families, and end-organ sensitivi- with wet months that feature ties associated with nasal, above-freezing temperatures, e.g. 5 December through March in typically are Type III and IV and California, summer in the mid- involve IgG antibodies. Symptoms west, or year-round in Florida. Very may include flu-like illness with dry states, such as Arizona or fatigue, rash, muscle and joint Nevada, are unfortunately not pain, headache, fever and exempt from inhalant mold aller- nightsweats. Fungal colonization A food allergy test can gies, as the moisture in air condi- may be life-threatening in immuno- be most clinically tioning units invites the growth of compromised individuals. mold which then becomes dissemi- useful for measuring nated throughout the living space. Avoidance of allergen, in the case Delayed-Onset IgG IgG antibodies involved of molds and fungi, can be Reactions in delayed-onset extremely challenging since these Type III reactions involve the substances are so ubiquitous in the formation and deposition of Type III reactions. environment. Common food antigen/antibody (Ag/Ab) com- sources include fermented cheese, plexes, mostly involving IgG. In wine, beer and bread. Other com- contrast to the immediate IgE mon, but less suspected sources histamine-mediated reactions, include tea, processed foods, these reactions are delayed, since dough conditioner, commercial they involve the gradual formation fruit juices, citric acid, malt flavor- of immune complexes. Because ings, chocolate, soy sauce, tomato these reactions are delayed by products (crushed and left to sit hours or even days following the for better flavor), Lactaid and B exposure, the relationship between vitamins. Fungal colonization of food and symptoms is much more the skin or mucous membranes difficult to spot. It is these reac- represents a third source of expo- tions for which a food allergy test sure which may serve to amplify measuring IgG antibodies can be reactions to other fungi in the envi- most clinically useful. ronment. Since different fungi from different sources share common IgG-mediated reactions typically surface proteins, the immune result from exposure to an excess response to fungi may cross-react. of antigen over an extended period As a result, correcting conditions of time. In the case of food allergy, such as dandruff, athlete’s foot increased intestinal permeability and Candida overgrowth often coupled with repetitious ingestion helps to lessen reactions to the of particular foods causes excessive other sources. antigen to be presented to the immune system. Formation of Such reactions are all primarily insoluble antigen/antibody com- IgE-mediated and feature the plexes results in the activation of symptoms commonly associated complement and the subsequent with histamine, such as sneezing, respiratory burst in neutrophils, runny nose and asthma. More the release of proteolytic enzymes, severe illness may result when an mast cell mediators and vasoactive individual is exposed to a much peptides, and the aggregation of larger number of fungal particles, platelets. Although complement such as in occupational exposures. stimulates inflammation, it also Due to the massive amount of functions to prevent the progres- 6 antigen exposure, these reactions sion from small complexes to larger ones, a factor that helps may NOT elicit symptoms, despite minimize the severity of symptoms. the fact that an immune reaction is Macrophage activity triggers the occurring. An overload of antigen, release of inflammatory mediators however, will saturate the such as interleukin-1, tumor necro- macrophages’ capacity, resulting in sis factor, reactive oxygen species the circulation of complexes and and nitric oxide. their deposition in tissue. Immune compromise may lead to the same Symptoms are typically delayed end, resulting in symptoms. (This in onset, by hours or days, and process is quite different from the vary, not only according to the "loaded gun" IgE response which is specific nature of the immune elicited with every exposure.) This complex, but also according to "overload" phenomenon may help the tissue in which the complexes explain why reducing exposure to are deposited. Headache, vasculitis one allergen may result in an indi- or hypertension may result from vidual being better able to tolerate deposition in the vascular space; other allergens. Unlike Ag/Ab inter- asthma, alveolitis or recurrent actions, macrophage clearance is infection may result from deposi- NON-specific. This means that tion in respiratory tissue, dermato- ANY reduction in demand on logic changes from deposition in macrophages (including non-aller- the skin, and joint pain from gic conditions such as infection or deposition of complexes in the joint xenobiotic exposure) may serve to space. Symptoms such as rhinitis or reduce symptoms and allow other angioedema may also occur, since reactive foods to be eaten. Antigen “overload” two elements in the complement cascade (C3 and C5) are capable This concept of "total load" can- may help explain why of inducing histamine release. Any not be overemphasized. It was not system may be affected and any until the middle of the last century reducing exposure to one symptom is possible, depending on and the growth of the industrial allergen may result in an an individual’s susceptibilities. revolution that the diseases which Reactions may last for days. we now call atopic allergic diseases, individual being better able were recognized as entities. (4) to tolerate other allergens. Hypersensitivities involving Immune Competency and bronchial symptoms and asthma "Total Load" nearly doubled during the 1980s Although an immune interaction and early 1990s.(5) Air pollution, between antibody and antigen including the contribution by diesel occurs every time an individual is exhaust particle emissions, has exposed to an allergenic food, the been shown to enhance both nasal presence and the degree of symp- IgE production and the expression tomatology depends upon the sol- of Th2 cytokines (6), and may ubility of complexes and the reticu- serve as a carrier for pollen and loendothelial system’s ability to other compounds. clear them. Macrophages pick up Ag/Ab complexes immediately, but Food allergies can develop at any have a finite capacity. With an effi- point in one’s life, but an individ- cient immune response, the half- ual may also be born with them. life of a complex may only be a few Such allergies are usually to foods minutes, and exposure to allergens that the mother consumed fre- 7 quently during her pregnancy, symptoms, yet allergies exist, a test although the antibodies are the can help to identify those foods baby’s own. Since maternal dietary which to some degree are stressing proteins are capable of reaching the immune system. If a person is amniotic fluid, and since the fetus is allergic to numerous foods, the capable of mounting antibody and test can also help provide a start- other immune responses as early as ing place, in terms of elimination. the tenth week of gestation, it is pos- sible that fetal sensitization to these In-Vivo Skin Testing proteins begins as early as the first Type I allergies are often trimester of pregnancy.(7) Maternal diagnosed with skin prick testing, IgG antibodies traverse the placenta or with intradermal testing, often during pregnancy, but levels of these employed as a follow-up to a in the infant are typically down by 3- negative skin prick test. Advantages 6 months after birth.(8) Dietary of skin testing include rapid results, proteins from the mother’s diet good sensitivity, and the ability to are also transferred to breast milk. test any antigen. Disadvantages While a certain amount of antigen include the discomfort inherent in passage into the breast milk is prob- the procedure, ably important for the development possible danger of anaphylaxis,(9) of the infant’s tolerance the contraindicating effects of An allergy test to foods, excessive exposure can medications such as anti-hista- result in hypersensitivity. Breast-fed mines, decongestants, beta should always be infants whose mothers take dietary blockers, bronchodilators and theo- precautions during lactation are assessed in conjunction phylline, and occasional interfer- observed to have a markedly reduced ence from skin disease. Since skin with a patient’s incidence of atopic eczema.(7) testing only reflects IgE-mediated reactions, it also cannot inform clin- clinical picture. Unlike IgE allergies, those icians as to the potential for the mediated by IgG may be cured, delayed hypersensitivity reactions following a period of avoidance responsible for such a wide range of and attention to underlying con- symptoms seen clinically.(10) tributing factors. Although a food may be tolerated at some point on In-vitro Antibody a limited basis, the immune system Measurement "holds a grudge," in a sense. The measurement of antigen-spe- Because the hypersensitivity is cific antibodies is a useful tool for recorded in the body’s "memory assessment of allergies, particularly cells" (antigen-stimulated lympho- to foods. One study of young chil- cytes), the response may be reacti- dren found that 62.5% of children vated if exposure again becomes with symptoms had specific IgG excessive or too frequent. antibodies and 22.9% had specific IgE antibodies, while the children DIAGNOSIS without symptoms of food allergy An allergy test should always be had neither.(11) assessed in conjunction with a patient’s clinical picture. If an indi- Three widely used methods for vidual’s immune clearance mecha- measuring specific antibodies 8 nisms are effective in averting include ELISA, MAST and RAST/RASP. ELISA (enzyme-linked ing can provide a starting place for immunosorbent assay) can detect trial elimination programs. As a either IgG or IgE antibodies. MAST screen for an asymptomatic individ- and RAST (radioallergosorbent ual, in-vitro testing may reveal food procedure) measure IgE antibod- allergies which are currently being ies, although MAST testing for IgG effectively managed by the immune antibodies is currently being devel- system (hence, no symptoms), but oped. Some laboratories measure which could manifest symptomati- In-vitro allergy tests only IgG4 for foods; however, cally in the future, in the event of measurement of total IgG increased immune burden. offer the advantages of is recommended. Occasionally, a test will exhibit "across-the-board" low-level IgG convenience, safety (no Although all IgG subclasses are reactivities for an individual. Clinical danger of anaphylaxis), involved in the immune response, observation has suggested the pos- IgG1 is thought to be the main sibility of a chronically "leaky gut" lack of interference by instigator of inflammation. The in such situations, and the con- role of IgG4 in food allergy has comitant immune reaction to a medications or skin been debated. IgG4 is unable to large number of absorbed antigens. condition, and good activate complement, so does not contribute to a true inflammatory Disadvantages of in-vitro testing reproducibility. reaction. It is, however, able to include the requirement of serum precipitate the release of histamine collection, the sometimes lengthy from basophils (12), which might incubation and the possibility of partially explain the elevated levels some false negatives for IgE, which observed in atopic individuals and may result from a number of fac- amelioration of symptoms upon tors. Antibodies that are directed removal of the "positive" foods toward altered forms of antigen not from the diet.(13) At the same time, used in the test, e.g. cooked, it has been suggested that IgG4 spoiled or processed foods, might may function as a "blocking" anti- go undetected. False negatives may body to allergic reactions, particu- also follow immunotherapy, a larly since levels tend to increase result of IgG "blocking" antibody dramatically following successful production. (This is one of the immunotherapy for IgE-mediated mechanisms behind immunothera- pollen allergies.(14) Although IgG4 py’s effectiveness for IgE-mediated can induce histamine release, its allergies; antigen now preferentially release appears to be more delayed binds to IgG, rather than IgE, so and the reaction less acute than in that the immediate and acute hist- the Type I IgE reactions. amine reaction is prevented.) Finally, with the possible exception In-vitro tests offer the advantages of anaphylaxis-inducing allergens, of convenience, safety (no danger specific immunoglobulins tend to of anaphylaxis), lack of interfer- gradually diminish in response to ence by antihistimines or skin con- antigen elimination. Although the dition, and good reproducibility. half-life of IgE antibodies is only 3 In -vitro also allows the use of par- days and the half-life of IgG 23 allel controls with each run. A pos- days, absorbed antigens which have itive reaction on an in-vitro test sig- been sequestered by the liver may, nifies allergy in that individual. For in some cases, be slowly released the "global reactor," in-vitro test- over several months, resulting in 9 some persistent antibody ticular food may not, in fact, corre- production. The levels, however, late with improvement in, or aggra- will still decline over time, barring vation of, an individual’s level of any new exposure. Because foods reactivity. Even the broad cate- such as wheat, dairy and corn are gories of 0-3+ should be evaluated widely used as additives in against the patient’s clinical pic- processed foods or cosmetics, IgG ture. A 3+ IgG reaction will gener- levels are more likely to persist in ally imply a stronger immune reac- an individual who mistakenly tion (or a more severe allergy) than presumes that he has completely a 1+ reaction. However, that 3+ eliminated the foods. IgE antibod- reaction may never manifest as ies to seasonal allergens may be symptoms in a person whose sys- undetectable if measured during an tem is effectively neutralizing asymptomatic period. immune complexes, while a 1+ reaction may result in debilitating Interpreting the In-Vitro symptoms in a person whose retic- Test ulendothelial system is over- It should be noted that in-vitro whelmed. In other words, the best antibody tests are semi-quantita- use of the test is to identify reactive tive. All procedures involve the substances which can then be binding of specific anti-food anti- avoided or rotated in a clinical bodies to an antigen that is already trial. bound to a solid phase, e.g. a plate or test tube. Each procedure Elimination Diets includes a tag, or signal, which is Elimination and reintroduction of quantitated. The tag used in the foods is an invaluable means of RAST is a radioactive isotope, establishing a relationship between In-vitro antibody tests while MAST uses a luminescent sig- a symptom and a particular food. nal. In the ELISA test, the tag is an The typical protocol involves the are semi-quantitative. enzyme that induces a color elimination of all possibly aller- change which is then read photo- The best use of genic foods (commonly the routine metrically as "optical densities." foods in one’s diet) for 1-2 weeks the in-vitro allergy The more intense the color change, and, assuming any clinical the higher the concentration of improvement during this time, the test is to identify specific antibodies. Because of gradual reintroduction of one food inherent imprecision in the multi- reactive substances every 2-3 days. If a food in one step processes used in these proce- meal fails to produce symptoms, which can then be dures, there will always be some then a larger amount is eaten in natural "drift" in the quantifica- the next couple of meals. If symp- avoided or rotated in tion. It is for this reason that toms reappear within the 2-3 days reporting of broad categories for a clinical trial. of repeated ingestion of a food, reactivity, e.g. 0-3+, has become an that food is regarded as allergenic, industry standard. and eliminated from the diet. If no symptoms are induced, the food is Literal reliance upon the larger now included in the hypoallergenic optical density numbers (in the diet, and the next food is tested in case of ELISA) can be misleading, the same manner. as a follow-up test featuring higher or lower optical densities for a par- Needless to say, this process may 10 take several weeks to complete, Inhalant allergies are most often depending upon the number of treated with a combination of foods tested, and it requires avoidance, where possible, and more than a modicum of self- immunotherapy, which consists of discipline. The results, however, injections of serially increasing con- prove invaluable. The patient is centrations of antigen. As antigen not only able to ascertain which is introduced into a patient’s sys- foods produce which symptoms, tem, the immune system gradually but the clinical experience of develops a mixed IgG response to improvement on the elimination it, although never enough to cause diet and the aggravation upon tissue damage. The IgG antibodies ingestion of a food, often serves are thought to block the more to provide the needed incentive severe IgE reaction, and symptoms usually abate within a few weeks. for the subsequent long-term Combining a elimination or rotation of offend- Standard immunotherapy is not ing foods. The process assists in used for the desensitization to preliminary in-vitro revealing the NON-immune medi- foods and molds because the large ated reactions as well. As discussed amount of antigen exposure from test with an natural sources is usually already earlier, a relationship between food elimination diet and symptom does not automati- so high that severe reactions are cally mean "allergy." likely to be induced. It should be represents the most noted that, although IgE antibod- There are a few disadvantages to ies to a substance remain in the efficient approach of system, an IgE blood test may fail relying solely on the elimination all, and improves diet for diagnosis. The length of to pick them up following treat- time and discipline required are ment, due to the competitive patient compliance. mentioned above. Sub-clinical inhibition by IgG. This significant hypersensitivities will not be detect- a decline, however, appears to ed. Finally, the possibility exists occur only after about two years that some of the foods consumed of hyposensitization.(15) as components of the "hypoaller- genic" diet are antigenic for a given individual. In this case, ameliora- Enzyme-potentiated tion of symptoms on the diet does desensitization not occur, and the false conclusion Enzyme-potentiated desensitization is reached that foods play no part (EPD) also utilizes the periodic in the patient’s illness. Combining injection of antigen into the a preliminary in-vitro test with an patient’s skin, but the amount of elimination diet represents the antigen compared to standard most efficient approach of all, and immunotherapy is exceedingly improves patient compliance. small. Furthermore, the antigen is accompanied by the enzyme TREATMENT beta-glucuronidase which helps to activate T-suppressor (CD8) cells. Following is a brief discussion of a Unlike immunotherapy, IgG levels few of the most popular methods do NOT rise in response to the for managing allergies. treatment, although tolerance develops. Injections are adminis- Immunotherapy tered every two months and in most cases can eventually be (Hyposensitization) 111 spaced further apart, assuming greater likelihood that sensitivities improvement. EPD is used for food to oyster, scallop and squid (other and mold allergy and is preferred members of the "mollusk" family) by many physicians for treatment might be present. At the same time, of inhaled allergens, as well. such cross-reactions have been observed to be relatively infrequent, Medications and patient compliance is always a Other standard treatments for consideration. Care must also be inhalant allergies include the use taken that suspected foods are not of anti-histamines, decongestants, inadvertently consumed while hid- glucocorticosteroids, and sodium den in other foods, e.g. eggs con- cromoglycate. This last one serves tained in mayonnaise. to block mast cell degranulation, eosinophil and neutrophil chemo- Eliminating the antigenic food taxis and mediator release; howev- provides the immune system a rest. er, it is poorly absorbed orally, Levels of antibodies gradually so must be taken frequently in decline, immune complexes are order to maintain adequate cleared, and symptoms improve. mucosal concentrations. Quercitin Lesser reactive foods are often is a natural bioflavonoid with tolerated in the beginning if eaten similar activity.(16) All of these no more often than every four medications are merely palliative, days or so; however, this tolerance Eliminating the but of course can make a world varies between individuals, and in of difference in a patient who is the more symptomatic patient antigenic food experiencing paroxysmal sneezing even foods with 1+ reactivities or a frightening bout with asthma. may have to be temporarily provides the immune removed from the diet in order to system a rest. Levels of Allergen Elimination improve clinical outcome. The most important component antibodies of any allergy program is the reduc- Antigen/Antibody gradually decline, tion of exposure to the offending Equilibrium agent. As long as exposure is main- One of the most important deter- immune complexes are tained, antibodies will continue to minants of the pathogenicity of be produced and the immune sys- cleared, and the Ag/Ab complex and the clinical tem "primed" to react. IgE-mediat- course of the patient is the equilib- symptoms improve. ed allergies tend to be "fixed," thus rium between antigen and anti- avoidance is usually life-long. In body. At equivalent concentrations contrast, the IgG-mediated allergies (also called the "zone of equiva- may be reversed over time. Reactive lence"), complexes can grow to foods are ideally eliminated from enormous size. The larger and the diet for a minimum of three more numerous the complexes, the months, particularly those that more likely they are to overwhelm evoke the most severe symptoma- the reticuloendothelial system and tology or the highest scores on an deposit in bodily tissues. in-vitro test. Ideally exposure to Individuals with IgG allergies often foods within the same food family have the paradoxical experience of of a reactive food should be feeling immediately better, rather reduced as well. A sensitivity to than worse, after consuming an clam, for example, may indicate a allergenic food. What is happening 121 here is a shift into the zone of Plasma cells retain a memory of "antigen excess."(17) In this zone, that sensitivity, and new antibodies complexes become more soluble could be produced with any expo- and symptoms may diminish, that sure. Only a trial reintroduction is, until more antibody is produced will answer that question. If no and the "zone of equivalence" is symptoms are elicited when the

Antigen Excess Zone of Equivalence Antibody Excess

again reached, producing more food is reintroduced, the food can symptoms. Eating more of the usually be replaced into the diet, food once again brings relief, but ideally on a rotational basis. perpetuating a vicious cycle of Repetitive exposure to a food usu- addiction that is difficult to break. ally contributes to the development Following 7-10 days of strict of hypersensitivity in the first place, elimination of a reactive food, an thus a return to the same frequen- individual moves into a zone of cy of exposure often brings with it One of the "antibody excess." With less anti- an eventual return of symptoms. most important gen on board, complexes reduce in size and number, and symptoms determinants of the diminish. During this period, he Other Considerations pathogenicity of the or she is extremely sensitive (hence In the case of reactivity to molds, the conspicuous and diagnostic yeast or fungi, it is often useful to clinical course of aggravation from reintroductions treat any existing infections. of offending foods at this point), Examples include ringworm, ath- the patient is the and should stay with the elimina- lete’s foot or jock itch, as well as equilibrium between tion diet for three or more months, vaginal or intestinal yeast over- while antibody levels drop. growth. Doing so may help to antigen and antibody. eliminate cross-reactivity reactions After this point, a follow-up in- and lower the overall burden on vitro test might well show lower the immune system. reactivities for those foods that have been avoided. This fact does Because IgG allergies can be not guarantee that the food will be established as a result of the tolerated when reintroduced. translocation of food antigens 13 across a permeable gut wall (18), imbalances will greatly help to attention to gut repair is a prereq- reduce the likelihood of developing uisite for preventing further prob- additional allergies. lems. It is not uncommon for a person who is eating soy foods, Finally, it’s worth noting that e.g. as a substitute for foods such adjunctive nutritional support for as cow’s milk and wheat, to sud- the immune system, as well as Correcting imbalances denly find herself now unable to selective botanicals, homeopathic tolerate soy after a few months. remedies or acupuncture, may con- in the gastrointestinal Evaluation of intestinal permeabili- tribute significantly to the reduc- flora and providing ty as well as a comprehensive stool tion in immune sensitivity and the analysis can serve to establish the inflammatory response. nutritional support for presence of "leaky gut" and con- tributing factors such as maldiges- Acknowledgements: Special thanks the immune system can tion, malabsorption, imbalances in to Vincent Marinkovich, M.D., greatly help reduce flora, secretory IgA deficiency and (Diplomat, American Board of mucosal injury due to infection or Allergy and Clinical Immunology) immune sensitivity. agents such as non-steroidal anti- for his generous contribution of inflammatories. Correcting any information on the subject.

14 REFERENCES 1) Gaby AR. The role of hidden food allergy/intolerance in chronic disease. 10) El Rafei A, Peters SM, Harris N. Bellanti JA. Diagnostic value of IgG4 mea- Alt Med Review 1998;3(2):90-100. surement in patients with food allergy. Ann Allergy 1989;62:94-99.

2) Buckley RH, Metcalfe D. Food allergy. JAMA 1982;248:2627-31. 11) Hofman T. IgE and IgG antibodies in children with food allergy. Rocz Akad Med Bialmyst 1995;40(3):468-473. 3) Gerrard JW, Ko CG, Vickers P. The familial incidence of allergic disease. Ann All 1976;36:10. 12) Vijay HM, Perelmutter L. Inhibition of reagin-mediated PCA reactions in monkeys and histamine release from human leukocytes by human IgG4 sub- 4) Mygind N. History of Allergy. In: Essential Allergy—An Illustrated Text for class. Int Arch Allergy Appl Immunol 1977;53:78-87. Students and Specialists. Boston: Blackwell Scientific Publications, 1986:1-9. 13) Gwynn CM, Ingram J. Bronchial provocation tests in atopic patients with 5) Chandra RK. Food allergy and food intolerance: lessons from the past and allergen specific IgG4 antibodies. Lancet 1982;1:254-256. hopes for the 21st century. In: Somoyogi JC, Muller HR, Ockhuizen T, editors. Food allergy and food intolerance. Nutritional aspects and developments. Bibl 14) Van der Giessen M, Homan WL, van Kernebeek G, et al. Subclass typing Nutr Dieta 1991;48:149-156. of IgG antibodies formed by grass pollen allergic patients during immunother- apy. Int Arch Allergy Applied Immunol 1976;50:625-639. 6) Casillas AM, Nel AE. An update on the immunopathogenesis of asthma as an inflammatory disease enhanced by environmental pollutants. Allergy 15) Szymanski W, Chrek-Borowska S, Obrzut D. IgG, IgA, IgM and IgE serum Asthma Proc 1997 Jul-Aug;18(4):227-33. levels in asthmatic patients sensitive to house dust and mite allergens after two-year specific immunotherapy. Arch Immunol Ther Exp (Warsz) 7) Chandra RK. Food allergy and food intolerance: lessons from the past and 1984;32(4):381-7. hopes for the 21st century. In: Somoyogi JC, Muller HR, Ockhuizen T, editors. Food allergy and food intolerance. Nutritional aspects and developments. Bibl 16) Middleton E Jr , Drzewiecki G. Flavonoid inhibition of human basophil Nutr Dieta 1991;48:149-156. histamine release stimulated by various agents. Biochem Pharmacol 1984 Nov 1;33(21):3333-8. 8) Host A, Husby S, Gjesing B, Larsen JN, Lowenstein H. Prospective estima- tion of IgG, IgG subclass and IgE antibodies to dietary proteins in infants with 17) Randolph TG. Specific adaptation. Ann Allergy 1978;40:333-345. cow milk allergy. Levels of antibodies to whole milk protein, BLG and ovalbu- min in relation to repeated milk challenge and clinical course of cow milk 18) Andre C, Francoise A, Colin L. Effect of allergen ingestion challenge with allergy. Allergy 1992 Jun;47(3):218-29. and without cromoglycate cover on intestinal permeability in atopic dermati- tis, urticaria and other symptoms of food allergy. Allergy 1989;44(9):47-51. 9) Sampson HA, Metcalfe DD. Food allergies. JAMA 1992;268(20):2840- 2844.

15 Great Smokies Diagnostic Laboratory SM 63 Zillicoa St., Asheville, NC 28801-1074 800-522-4762 E-mail: [email protected] www.greatsmokies-lab.com

Published and distributed by Great Smokies Diagnostic Laboratory in accordance with permissions granted by author Mary James, N.D., owner of the copyright to material contained herein, this monograph may not be reproduced in any form without the express permission of Great Smokies Diagnostic Laboratory or Mary James, N.D.

This information is for the sole use of a licensed health care practitioner and is for educational purposes only. It is not meant for use as diagnostic information. All claims submitted to Medicare/Medicaid for GSDL laboratory services must be for tests that are medically necessary. "Medically neces- sary" is defined as a test or procedure that is reasonable and necessary for the diagnosis or treatment of illness or injury or to improve the functioning of a malformed body member. Consequently, tests performed for screening purposes will not be reimbursed by the Medicare program. Physicians may deem it medically necessary to order a single test or a portion of a profile.

©2001 Great Smokies Diagnostic Laboratory

i,ml,Mjallpaper,062701