Doi:10.1080/00365520510015593

Scandinavian Journal of Gastroenterology, 2005; 40: 800 Á/807 Food-specific IgG4 antibody-guided exclusion diet improvessymptoms and rectal compliance in irritable bowel syndrome SAMEER ZAR1, LYNNE MINCHER2, MARTIN J. BENSON3 & DEVINDER KUMAR4 1OGEM Department, St George’s Hospital Medical School, London, UK, 2Dietetic Department, 3Department ofGastroenterology, St. Helier Hospital, Carshalton, Surrey, UK, and 4Department of Gastrointestinal Surgery, St George’sHospital Medical School, London, UK AbstractObjective. Dietary modification improves symptoms in irritable bowel syndrome (IBS). Identification of offending foodsby dietary elimination/re-challenge is cumbersome. IgG4 antibodies to common food antigens are elevated in IBS. The aimof this article was to evaluate the effect of exclusion diet based on IgG4 titres on IBS symptoms and rectal sensitivity andcompliance. Material and methods. The study comprised 25 patients with IBS (3 M, 22 F, mean age 43 years, Rome IIcriteria). IgG4 titres to 16 foods (milk, eggs, cheese, wheat, rice, potatoes, chicken, beef, pork, lamb, soya bean, fish,shrimps, yeast, tomatoes and peanuts) were measured. Foods with titres /250 mg/l were excluded for 6 months. Symptomseverity was assessed with a previously validated questionnaire at baseline, at 3 months and at 6 months. Rectal complianceand sensitivity were measured in 12 patients at baseline and at 6 months. Results. IgG4 antibodies to milk, eggs, wheat,beef, pork and lamb were commonly elevated. Significant improvement was reported in pain severity (p B/0.001), painfrequency (p 0/0.034), bloating severity (p 0/0.001), satisfaction with bowel habits (p 0/0.004) and effect of IBS on life ingeneral (p 0/0.008) at 3 months. Symptom improvement was maintained at 6 months. Rectal compliance was significantlyincreased (p 0/0.011) at 6 months but the thresholds for urge to defecate/discomfort were unchanged. Conclusions. Food-specific IgG4 antibody-guided exclusion diet improves symptoms in IBS and is associated with an improvement in rectalcompliance.
Key Words: Barostat, exclusion diet, food-specific IgG4 antibody, irritable bowel syndrome, rectal compliance, rectal sensorythreshold demonstrated symptomatic response to a 1-weekelimination diet, several studies have shown a The aetiopathogenesis of irritable bowel syndrome response rate of up to 67% in IBS patients [1,7,8].
(IBS) is thought to be multifactorial involving an Most studies have used either a standard exclusion interaction between diet [1], infection [2], antibio- diet for all the subjects or excluded foods from the tics [3] and psychosocial factors [4]. This causes diet based on the history of intolerance as reported perturbation of the enteric nervous system and by the patients. This approach has been adopted sensorimotor dysfunction. Rectal hypersensitivity is because of a lack of objective testing for food a well-recognized characteristic in IBS patients and hypersensitivity. Recently, IgG4 antibodies to var- increased sensory perception and reduced compli- ious food antigens have been incriminated in the ance, in response to rectal distension with increasing pathogenesis of food hypersensitivity-induced atopic balloon volume or pressure, have been reported conditions [9 Á/11]. A similar finding has been reported in IBS patients, although the exact role of A significant proportion of IBS patients attribute food-specific IgG4 antibodies in the pathogenesis of their symptoms to adverse food reactions. Since the IBS has not been evaluated [12]. It has been initial study by Jones published in 1982, which suggested that, similar to other atopic conditions, Correspondence: Devinder Kumar, PhD, FRCS, Department of Gastrointestinal Surgery, St George’s Hospital Medical School, Cranmer Terrace, LondonSW17 0QT, UK. Tel: '/44 020 8725 1302. Fax: '/44 020 8725 3611. E-mail: [email protected] (Received 17 August 2004; accepted 6 December 2004) ISSN 0036-5521 print/ISSN 1502-7708 online # 2005 Taylor & FrancisDOI: 10.1080/00365520510015593 food-specific IgG4 antibodies may be involved in the cant gastrointestinal disorder were excluded from pathogenesis of a subgroup of IBS patients. Elevated the study. In addition, ‘‘advanced’’ cardiac, respira- titres to specific food antigens may prove useful for tory, renal or hepatic diseases; concurrent malig- targeted dietary exclusion in a subgroup of IBS nancy; major psychiatric disorders or a history of patients. This may obviate the need for excluding a drug/alcohol abuse were also considered as exclusion large number of foods from the diet which are poorly criteria. Patients who had undergone previous ab- tolerated and can be nutritionally inadequate. The dominal surgery, with the exception of uncompli- success of this approach will implicate food-specific cated appendicectomy, were also excluded. Use of IgG4 antibodies in the pathophysiology of IBS.
any medication which could perturb gastrointestinal Abnormalities in sensorimotor physiology of the sensorimotor function was disallowed. Patients who upper and lower gut have been reported in IBS. The were pregnant or became pregnant during the study barostat has increasingly been used as a research tool were also excluded. The protocol was approved by to determine visceral sensitivity thresholds and the local ethics committee and full, written, in- compliance and the standardization of this technique formed consent was obtained from the subjects.
has been described in detail [13 Á/15]. It has beenshown that IBS patients have a lower sensory threshold to volume distension compared to con-trols, in both the rectum and colon [5,16 Á All patients were asked to complete two question- Compliance is considered to be an important factor naires at baseline. The first questionnaire assessed in the reservoir function of the rectum and is the symptoms based on the Rome II criteria [21] reduced in IBS patients [19,20]. These abnormal- including the site and frequency of abdominal pain, ities in the gut sensorimotor function have been used relief with defecation, stool frequency and stool as markers of response to therapeutic interventions.
form. The second questionnaire was used to mea- The aim of this study was, first, to evaluate the sure the severity of symptoms using a previously response to food-specific IgG4 antibody-guided exclusion diet on patients’ symptoms over a period prompted to score their symptoms on a visualanalogue scale (range 0 of 6 months, and secondly, to compare the changes prior to the interview. The first two questions in the rectal visceral sensitivity and compliance as an recorded the severity of pain and bloating. Pain objective measure of the response to such a diet.
was also scored for frequency, thereby giving it agreater influence on the total score. In addition,patients scored the degree of dissatisfaction with their bowel habits and the effect of IBS on their life in general. A composite severity score was thencalculated by adding the response to all the above Twenty-five IBS patients (3 M, 22 F, mean age 42.6 questions. The questionnaire was used to monitor years, SD9/14) participated in the dietary exclusion progress in response to the exclusion diet at 3 and study. The group comprised 13 patients with diar- rhoea-predominant symptoms (D-IBS), 10 patientswith constipation-predominant symptoms (C-IBS)and 2 with alternating symptoms (Alt-IBS). In a subgroup of 12 patients, rectal compliance and The level of anxiety and depression experienced by sensory thresholds for the urge to defecate and the subjects was scored using the Hospital Anxiety discomfort were measured using barostat at baseline and Depression scale (HAD scale) [23]. The ques- and repeated after 6 months on the exclusion diet.
tionnaire prompted the subjects to choose the most The diagnosis of IBS was based on the Rome II appropriate response to 14 questions (7 for anxiety criteria [21], defined as abdominal pain/discomfort and 7 for depression), each scored from 0 to 3 with of at least 12 weeks’ duration in the previous one the higher scores indicating a greater level of anxiety year, which need not be continuous, associated with or depression. The questionnaire was repeated again two of following three criteria; altered stool fre- at 3- and 6-month intervals after initiating the quency, altered stool form and/or relief with defeca- tion. A thorough history, physical examination androutine blood tests were carried out in all subjects.
Colonoscopy or flexible sigmoidoscopy with bariumenema was carried out to exclude bowel pathology.
Serum IgG4 antibody titres to 16 common foods Patients with inflammatory bowel disease, coeliac were measured (milk, eggs, wheat, cheddar cheese, disease, known lactose intolerance or other signifi- rice, yeast, potato, peanut, fish, chicken, lamb, beef, pork, tomatoes and soya bean). The samples The thresholds for urge to defecate and discomfort/ were processed in a central laboratory (Allergy pain were determined using this technique. A Diagnostic Laboratory, Oxfordshire, UK) using a pressure-volume relationship was determined for commercially available radioimmuno-fluorescence each patient by plotting the corresponding pressures technique (Pharmacia Unicap Autolyser). The tech- and volumes for each inflation step and rectal nique involves incubating patients’ serum with compliance (ml/mmHg) was calculated from the a cellulose cap containing the food antigen of slope of the pressure-volume curve using simple interest. After washing away the excess antibodies, the cap is incubated with fluorescence-labelledmouse anti-IgG antibody, which is then measured using a fluorometer after enzyme activation. Themeasured range of antibodies is between 1.5 mcg/l The baseline symptom severity scores (range 0 Á/100) and 30,000 mcg/l. A cut-off value of 250 mcg/l was obtained on the visual analogue scales were com- pared with the scores at 3 and 6 months. Similarly,the baseline rectal compliance and sensory thresh-olds for urge to defecate and discomfort were compared with the data obtained on repeat testing This was done under the supervision of a qualified at 6 months. A paired t -test was used for statistical dietician. Patients had a detailed interview with the dietician where a complete dietary history wasobtained. In addition, they were also asked to maintain diary cards with details of their dietaryintake for a 2-week period prior to the initial inter- IgG4 antibody titres and the exclusion diet view. This information helped in assessing how The food articles against which IgG4 antibody titres much of the suspected foods were being consumed were most commonly elevated ( /250 mg/l) were by the patient so that appropriate alternatives could milk, cheddar cheese, eggs, beef, pork, lamb, wheat be suggested to replenish the intake once the patient and tomatoes (Table I). These were elevated in more started the exclusion diet. This was especially than 50% of patients. Antibody titres to chicken, important for the patients who were excluding fish, rice, yeast, potatoes, soya bean, peanuts and several types of foods, in order to ensure they shrimps were elevated in fewer patients and for these continued to eat a nutritionally balanced diet. The foods the observed titres were B/500 mg/l in most results of the IgG4 antibody tests were then dis- cussed with the patients and the foods that induced On average each patient excluded 8 (range 3 Á/13) antibody titres greater than 250 mcg/l were excluded foods from the diet based on the IgG4 titres from the diet. The patients were also advised to /250 mg/l. The dietary advice was reinforced by check the contents of the pre-prepared foods and a the dietician at the 1-month, 3-month and 6-month list of the suspected foods was supplied for their interviews. The patients reported full compliance to reference. Patients were reviewed at one month from the exclusion diet for the duration of the study. The the start of the exclusion diet to assess their progress, results of the blood tests were shown to the subjects, reinforce the dietary advice and answer any queries which helped in motivating them. In addition, advice they might have. Patients were reviewed at 3 and 6 from the dietician ensured that patients were able to month when their symptoms were assessed as choose alternatives to the excluded foods, thereby described above while reinforcing the dietary advice.
A barostat (Synectics Medical, Copenhagen, Den- The baseline symptom profile was compared with mark) coupled with a disposable, 500 ml polyethy- the response at 3 and 6 months. The symptoms lene bag secured to the distal 5 cm of a multilumen were scored between 0 and 100 using a visual polyvinyl catheter was used to measure rectal com- analogue scale, where 0 signified no adverse symp- pliance and the pressure thresholds for the urge to toms. Data were available from 21/25 patients at 3 defecate and pain/discomfort. This was carried out months and showed significant improvement as using a phasic pressure sensitivity procedure in a compared to baseline in pain severity (0 IQR9/39 double random staircase sequence. Incremental versus 61 IQR9/39, p B/0.001), pain frequency (10 pressure steps (multiples of 4 mmHg) of 20-s IQR9/45 versus 50 IQR9/50, p 0/0.034), bloating duration separated by a 30-s rest period were used.
Table I. Food specific serum IgG4 titres in IBS patients.
Food antigen tested Titres /250 mcg/l Titres /500 mcg/l Titres /1000 mcg/l Abbreviation: IBS 0/irritable bowel syndrome.
p 0/0.001), satisfaction with bowel habits (35 IQR9/ able with the 3-month data showing that improve- 70 versus 70 IQR9/24, p 0/0.004) and effect of IBS ment in all the parameters had been maintained. A on life in general (34 IQR9/49 versus 70 IQR9/27, significant improvement in pain severity (35 IQR9/ p 0/0.008) (Table II, Figure 1). The total score was 43, p 0/0.005), pain frequency (50 IQR9/60, p 0/ also significantly improved (123 IQR9/215 versus 0.027), bloating severity (50 IQR9/42, p 0/0.024), satisfaction with bowel habits (66 IQR9/58, p 0/ Symptom severity score data at 6 months were 0.016) and effect of IBS on life in general (51 available for 15 patients. Of the remaining 10 IQR9/55, p 0/0.001) was observed (Table II, Figure patients, 1 discontinued the diet because of family 1). The total score was also significantly improved problems although having a good initial response, 1 (207 IQR9/220, p 0/0.002) at 6 months.
patient was diagnosed with a gynaecological condi-tion requiring surgery, 1 patient moved away from the area and 1 was working in the entertainmentindustry and found the dietary restrictions incompa- The baseline anxiety and depression scores calcu- tible with lifestyle. The remaining 6 patients were lated from the HAD questionnaire were not signifi- lost to follow-up. For the patients who completed the cantly changed at 3 months. However, at 6 months, study, the results of the 6-month data were compar- a significant improvement in the mean scores for Table II. Effect of exclusion diet on symptom severity scores.
Values are represented as the median and interquartile range (IQR).
$p -values are versus baseline scores.
Figure 1. At 3 months and 6 months, there was significant improvement in pain severity (p B/0.001, p0/0.005), pain frequency (p0/0.034,p 0/0.027), bloating severity (p 0/0.001, p 0/0.024), satisfaction with bowel habits (p 0/0.004, p 0/0.016) and the effect of irritable bowelsyndrome (IBS) on life in general (p 0/0.008, p 0/0.002) versus baseline.
both anxiety and depression was seen compared to demonstrates that IBS patients experience a signifi- the baseline scores (4 IQR9/4 versus 2 IQR9/4, p 0/ cant improvement in symptoms in response to a 0.007 for depression score; 10 IQR9/5 versus 7 ‘‘food-specific IgG4’’-guided exclusion diet. The IQR9/5, p 0/0.015 for the anxiety score) (Table III).
diet was based on elevated IgG4 antibody titres to16 common articles of food. The most commonfoods associated with elevated IgG4 titres were beef, pork, wheat and dairy products. This subjective Rectal compliance and the thresholds for discomfort symptomatic improvement reported by the patients and the urge to defecate were measured using a was mirrored by objective evidence of physiological barostat at baseline and after 6 months on the change in the sensorimotor function (i.e. increase in exclusion diet. A significant increase in rectal com- pliance was seen at 6 months as compared to Symptom severity was assessed by a previously baseline in response to the exclusion diet (5.42 ml/ validated questionnaire in which the patients ranked bloating and pain severity, pain frequency, satisfac- tion with bowel habits and effect/interference of IBS /4.22 at 6 months; p 0/0.011) (Figure 2). No significant change was observed in the pressure on life in general on a visual analogue scale. At 3 thresholds for discomfort/pain or the urge to defe- months, a significant improvement was seen in IBS cate after 6 months of being on the exclusion diet symptoms, i.e. pain severity, pain frequency, bloat- ing severity and satisfaction with bowel habits. Inaddition, the IBS symptoms interfered significantlyless with their lives in general since the introduction of the exclusion diet. The composite scores were also The management of IBS has traditionally been based significantly improved confirming that improvement on reassurance and symptomatic treatment. Evi- in one parameter is not offset by deterioration in dence from dietary elimination and re-challenge another. The improvement observed at 3 months studies supports the role of diet in IBS. This study was maintained at the 6-month follow-up for all the Table III. Effect of exclusion diet on hospital anxiety and depression scores.
Values are represented as the median and interquartile range (IQR).
$p0/0.018 (baseline versus 6 months); %p0/0.002 (baseline versus 6 months).
Figure 2. The exclusion diet resulted in an increase in rectal compliance (ml/mmHg) at 6 months versus baseline (7.75 versus 5.42,p 0/0.011).
parameters and patients continued to benefit from patients can be blinded to the food in such a manner the exclusion diet. Specific data on the effect of the over a prolonged period of time is impractical.
exclusion diet on bowel frequency was not collected Unfortunately, this means that placebo response and would have been difficult to interpret, as the cannot be excluded with certainty. However, the number of subjects in each IBS subgroup was small.
persistence of improvement in symptoms at 6 However, the overall improvement in bowel habits months and the fact that it was associated with reported by the subjects suggests that the interven- objective physiological improvement in rectal com- tion may have had a positive effect on bowel pliance suggests that the response is likely to be real.
frequency. However, larger studies in the future Currently, two empirical methods are used in will be needed specifically to answer this question.
clinical practice for prescribing exclusion diets. The Previous exclusion diet studies have shown a first method involves a hypoallergenic diet consisting similar response rate to exclusion diets. It is techni- of a small number of ‘‘allowed’’ foods. Once the cally difficult to perform double-blind exclusion diet symptoms are controlled, excluded foods are rein- studies. Most of the studies have used open dietary troduced one by one while monitoring for symptom elimination and challenge design and therefore recurrence. This approach is very restrictive and placebo response cannot be excluded. Blinding the cumbersome, although it is likely to benefit patients patient would require changing the appearance of who have hypersensitivity to a large number of foods.
the food, e.g. liquidized form and administering food The second approach is to exclude foods based on through nasogastric tubes to blind the patient to the history and/or to exclude a small number of foods taste and smell of the food. Organizing studies where commonly associated with the disorder and then Table IV. Effect of exclusion diet on rectal sensory thresholds and compliance.
Compliance (ml/mmHg)$ Discomfort (mmHg) Mean urge to defecate (mmHg) exclude additional foods if the symptoms persist.
duration of follow-up might have shown significant This method may take a long time to be effective, especially if the ‘‘offending’’ foods are atypical.
It was also noted that anxiety and depression These techniques are employed primarily because scores improved significantly at 6 months but not at there is no objective way of testing food hypersensi- 3 months. This delayed improvement suggests that the heightened level of anxiety and depression Unlike the two approaches discussed above, the experienced by IBS patients may be secondary to exclusion diets employed in this study were tailored the impact of the symptoms on their life rather than to the individual patient based on the serum IgG4 a primary trigger. Once the symptoms resolve, the antibody titres. This ‘‘targeted’’ approach has many levels of anxiety and depression return to baseline.
advantages. First, it provides objectivity to the In summary, exclusion diet based on food-specific process by excluding only those articles of food serum IgG4 antibody testing may provide a useful that are associated with increased antibody titres.
adjunct in IBS patients with persistent symptoms or Patients are more likely to be compliant as it is more those with a clear history of adverse food reactions.
‘‘scientific’’. Secondly, the diet is individualized to a Serum IgG4 antibody testing may provide an given patient thereby obviating the need for exclud- objective and quick method for selecting an exclu-sion diet for treating such patients.
ing a large number of foods from the diet. Thirdly,this approach is likely to reduce the number ofconsultations needed. Lastly, the clinicians are likelyto be more confident in prescribing exclusion diets This study also raises many questions. It should be [1] Jones VA, McLaughlan P, Shorthouse M, Workman E, Hunter JO. Food intolerance: a major factor in the patho- pointed out that antibody titres to only 16 common genesis of irritable bowel syndrome. Lancet 1982;2(8308): foods were tested and this panel may not be enough for all the patients. In addition, an arbitrary cut-off [2] McKendrick MW, Read NW. Irritable bowel syndrome: post level of 250 mg/l was used. Whether the precise cut- salmonella infection [see comments]. J Infect 1994;29:1 Á/3.
[3] Mendall MA, Kumar D. Antibiotic use, childhood affluence off level can be increased or reduced is not known and irritable bowel syndrome (IBS). Eur J Gastroenterol and whether different cut-off levels should be used for different articles of food remains to be deter- [4] Drossman DA, McKee DC, Sandler RS, Mitchell CM, mined. It is also not known if this approach would Cramer EM, Lowman BC, et al. Psychosocial factors in the lead to reduction in the antibody titres, which may irritable bowel syndrome. A multivariate study of patientsand nonpatients with irritable bowel syndrome. Gastroenter- be useful for monitoring progress. Further studies are needed to determine whether the excluded foods [5] Mertz H, Naliboff B, Munakata J, Niazi N, Mayer EA.
could be reintroduced into the diet once the Altered rectal perception is a biological marker of patients symptoms have been resolved, and, if so, in what with irritable bowel syndrome [published erratum appears inGastroenterology 1997;113:1054]. Gastroenterology 1995; Previous exclusion diet studies have used symp- [6] Distrutti E, Salvioli B, Azpiroz F, Malagelada JR. Rectal tom assessment alone in determining the efficacy of function and bowel habit in irritable bowel syndrome. Am J dietary interventions in IBS patients. This is the first [7] Bentley SJ, Pearson DJ, Rix KJ. Food hypersensitivity in study which demonstrates that symptom improve- irritable bowel syndrome. Lancet 1983;2(8345):295 Á/7.
ment is associated with changes in objective physio- [8] Petitpierre M, Gumowski P, Girard JP. Irritable bowel logical data, i.e. an increase in rectal compliance as syndrome and hypersensitivity to food. Ann Allergy 1985;54: measured with a barostat. Demonstration of a change in the gut sensorimotor function in response [9] Awazuhara H, Kawai H, Maruchi N. Major allergens in soybean and clinical significance of IgG4 antibodies investi- to the exclusion diet supports a cause and effect gated by IgE- and IgG4-immunoblotting with sera from relationship. However, the exclusion diet did not soybean-sensitive patients. Clin Exp Allergy 1997;27:325 Á/ show any significant change in the sensory thresh- olds for urge to defecate and discomfort. This may [10] el Rafei A, Peters SM, Harris N, Bellanti JA. Diagnostic value of IgG4 measurements in patients with food allergy.
partly be due to the fact that the study population included a mixture of diarrhoea, constipation and [11] Nakagawa T, Okano Y, Iwasaki E, Akimoto K, Nagakura T, alternating symptom IBS patients and these may Iikura Y, et al. Clinical significance of IgG4 antibody have different response patterns to sensory testing.
determination in children against egg white, milk, soybeanand Dermatophagoides farinae . Arerugi 1992;41:1694 Á In addition, once established, mucosal hypersensi- [12] Zar S, Benson MJ, Kumar D. Serum IgG4 antibodies to tivity might be irreversible or take a longer time to common food antigens are elevated in irritable bowel revert to normality and repeat testing at a longer syndrome [abstract no. 092]. Gut 2002;50 Suppl 11:A25.
[13] Whitehead WE, Delvaux M. Standardization of barostat [18] Hammer J, Phillips SF, Talley NJ, Camilleri M. Effect of a procedures for testing smooth muscle tone and sensory 5HT3-antagonist (ondansetron) on rectal sensitivity and thresholds in the gastrointestinal tract. The Working Team of compliance in health and the irritable bowel syndrome.
Glaxo-Wellcome Research, UK [see comments]. Dig Dis Sci Aliment Pharmacol Ther 1993;7:543 Á/51.
[19] Salvioli B, Bharucha AE, Rath-Harvey D, Pemberton JH, [14] van der Schaar PJ, Lamers CB, Masclee AA. The role of the Phillips SF. Rectal compliance, capacity, and rectoanal barostat in human research and clinical practice [In Process Citation]. Scand J Gastroenterol Suppl 1999;230:52 Á [15] Azpiroz F, Malagelada JR. Physiological variations in canine [20] Kwan CL, Davis KD, Mikula K, Diamant NE. Abnormal gastric tone measured by an electronic barostat. Am J Physiol rectal motor physiology in patients with irritable bowelsyndrome. Neurogastroenterol Motil 2004;16:251 Á [21] Thompson WG, Longstreth GF, Drossman DA, Heaton [16] Bradette M, Delvaux M, Staumont G, Fioramonti J, Bueno KW, Irvine EJ, Muller-Lissner SA. Functional bowel dis- L, Frexinos J. Evaluation of colonic sensory thresholds in orders and functional abdominal pain. Gut 1999;45 Suppl IBS patients using a barostat. Definition of optimal condi- tions and comparison with healthy subjects. Dig Dis Sci [22] Francis CY, Morris J, Whorwell PJ. The irritable bowel severity scoring system: a simple method of monitoring [17] Whitehead WE, Holtkotter B, Enck P, Hoelzl R, Holmes irritable bowel syndrome and its progress. Aliment Pharma- KD, Anthony J, et al. Tolerance for rectosigmoid distension in irritable bowel syndrome. Gastroenterology 1990;98: [23] Zigmond AS, Snaith RP. The hospital anxiety and depres- sion scale. Acta Psychiatr Scand 1983;67:361 Á/70.

Source: http://www.bio-food.it/intolleranze/Food-specific%20IgG4%20antibody%20guided%20exclusion%20diet%20improves.pdf