Comparison of IgA endomysium antibody and IgA tissue transglutaminase antibody in celiac disease

668 Can J Gastroenterol Vol 14 No 8 September 2000 Division of Gastroenterology, Department of Medicine, University of British Columbia, Vancouver, British Columbia Correspondence and reprints: Dr Helen R Gillett, Gastroenterology – Room F 137, University Hospital, 2211 Wesbrook Mall, Vancouver, British Columbia V6T 1Z3. Telephone 604-822-7216, fax 604-822-7236, e-mail pgillett@interchange.ubc.ca Received for publication March 19, 1999. Accepted November 30, 1999 ORIGINAL ARTICLE

I mmunoglobulin (Ig) A endomysium antibody (EmA) has been found to be highly sensitive and specific for celiac disease (1)(2)(3).EmA is usually detected with the use of indirect immunofluorescence against either monkey esophagus or human umbilical cord, making it, at best, only semiquantitative.In general, therefore, IgA EmA is of little use in monitoring response to a gluten-free diet.The identification of the antigen for endomysium as tissue transglutaminase (tTG) (4) has, therefore, opened up the possibility of using an ELISA to detect and quantify tTG, potentially improving our understanding of its role in the pathogenesis of celiac disease.
We have established an ELISA to measure IgA antibody to tTG and have tested a group of patients with untreated celiac disease, a group of patients with treated celiac disease and a disease control group for this antibody and for IgA EmA with the use of human umbilical cord to compare the sensitivities and specificities of the two tests.

PATIENTS AND METHODS
Sera from 21 patients with untreated celiac disease and 128 disease control subjects were tested for IgA EmA and IgA tTG antibodies.The diagnoses of the disease control group are listed in Table 1.In addition, 54 samples from 48 patients with treated celiac disease were tested.Celiac disease was defined by biopsy in all patients (ie, severe flat lesion responding to a gluten-free diet), and biopsy results of 42 disease control subjects were normal.The other 86 control subjects did not undergo small intestinal biopsies.The control group was, therefore, subdivided into 'biopsy' and 'nonbiopsy' control groups.
IgA EmA was detected with the use of indirect immunofluorescence against human umbilical cord according to the method described by Ladinser et al (2); the serum was measured at a 1:5 dilution.Titres of IgA antibody to tTG were measured with the use of an ELISA method devised by Dieterich et al (5).This technique was modified to account for differences in scientific supplies.In brief, high affinity 96well microtitre plates (Costar Corporation, USA) were coated with 1/600 U/well of tTG from guinea pig (Sigma-Aldrich Canada Ltd, Canada) in 100 µL 50 mM tris-hydrochloride, 150 mM sodium chloride and 5 mM calcium chloride (pH 7.5) overnight at 4°C.After washing three times with 50 mM tris-hydrochloride, 150 mM sodium chloride, 10 mM EDTA and 0.1% Tween 20 (pH 7.4), the plate was blocked for 10 min at 37°C using this solution.Samples were then applied to the plates at 1:5 dilution using the same solution as a diluent.The plate was incubated at room temperature for 1 h before being washed three times again.Peroxidase-conjugated rabbit antihuman IgA (Dako Corporation, USA) was applied at 1:3000 dilution for 1 h at room temperature, and the plate was again washed three times.O-phenylenediamine-hydrochloride 0.4 mg/mL in 0.05 mM citric acid and 0.1 mM dibasic sodium phosphate with 0.06% hydrogen peroxide was applied to the plate and incubated for 1 h in the dark at room temperature before reading using 450 nm wavelength light.A positive sample was tested in dou-bling dilutions until the linear portion of the standard curve was obtained.This sample was designated to have 400 arbitrary units (AU)/mL and was used as for the standard curve for each plate.The titre, in AU, of each sample was then calculated from this curve.
Statistical differences were calculated using the Mann-Whitney Test.

RESULTS
IgA EmA: All patients with untreated celiac disease were found to be positive for IgA EmA.One patient from the nonbiopsy control group tested positive for the antibody.All patients in the biopsy control group were negative for IgA EmA.Thirty-nine samples from patients with treated celiac disease were negative for IgA EmA, and 15 were positive.IgA antibody to tTG: The titres of IgA tTG antibody are shown in Figure 1 and Table 2.The titres were significantly higher in the untreated celiac group than in the EmAnegative groups (P<0.0001) and the EmA-positive, treated group (P=0.027).Titres in the EmA-positive, treated group were significantly higher than in all the EmA-negative groups (P<0.0001).The EmA-negative, treated group had higher titres than both the biopsy and the nonbiopsy control groups (P=0.0072 and P=0.0099, respectively).There was no Can J Gastroenterol Vol 14 No 8 September 2000 669 IgA tissue transglutaminase antibody in celiac disease

Hepatic steatosis 1
Pseudomembranous colitis 1 statistically significant difference between the two control groups.
A reference range for IgA tTG antibody was calculated using the biopsy control group.The titres within this group were converted to log 10 to normalize the data.The mean ± 3.1 standard deviations were calculated and antilog 10 determined.The calculated reference range was 1.4 to 138.4 AU/mL.Titres of IgA antibody to tTG in patients with treated celiac disease: In the patients with treated celiac disease, seven of 15 (46.7%) samples positive for IgA EmA had IgA tTG antibody titres within the calculated reference range.Of the treated patients who were negative for IgA EmA, 35 of 39 samples (89.7%) had titres within the reference range; three of these samples had titres above 138.4AU/mL, and one had a titre of less than 1 AU/mL.This patient was found to have selective IgA deficiency (SIgAD).
In five patients, two samples taken 18 to 24 months apart were tested.Three samples were tested from a sixth patient over a two-year period.Three patients were consistently positive for IgA EmA, and three were consistently negative.
The titres of IgA tTG antibody in these samples are shown in Table 3.
Four patients from the group with untreated celiac disease had samples taken five to 18 months after commencing a gluten-free diet.All remained positive for IgA EmA, but titres of IgA tTG antibody fell in each patient.The titres are also shown in Table 3.

Patients from the control groups with positive antibody results:
The one patient from the nonbiopsy control group found to be positive for IgA EmA had an IgA tTG antibody titre of 19,000 AU/mL.He was an elderly man who had been diagnosed many years earlier with ulcerative colitis.He underwent small intestinal biopsy, which demonstrated subtotal villous atrophy and crypt hyperplasia.He, therefore, commenced a gluten-free diet.
No patient from the biopsy control group had elevated titres of IgA tTG antibody.One patient in the nonbiopsy control group -a man with Crohn's disease and a titre of 297 AU/mL -had a titre above the cutoff level.He had a second serum sample tested 10 months later, and the titre had fallen to 17 AU/mL.IgA tTG antibody titres compared with small bowel histology: Biopsies on all 22 patients with untreated celiac disease (21 diagnosed at the start of the study and one diagnosed as a result of the study) and on 26 patients with treated celiac disease were graded as normal, mild, moderate and severe.Titres of IgA tTG antibody in each of these histological grades are shown in Figure 2.
Titres from those with normal mucosae differed from those with moderate (P=0.0001) and severe lesions (P=0.0002).Titres were higher in those with severe lesions than in those with moderate lesions (P=0.0282).Sensitivities, specificities, and positive and negative predictive values: All 21 patients in the group with untreated celiac disease plus one patient with confirmed celiac disease   Because the entire biopsy control group was negative for this antibody, the sensitivity, specificity, and positive and negative predictive values of the assay were all 100%.
One patient with untreated celiac disease had a titre of IgA tTG antibody below 138.4 AU/mL, making the sensitivity of this assay 95.5%.Because everyone in the biopsy control group had titres within the reference range, the specificity was 100%.The positive predictive value, therefore, was 100%, and the negative predictive value 97.7%.

DISCUSSION
Our study shows the excellent correlation between IgA EmA with the use of human umbilical cord and IgA antibody to tTG.Generally, we found markedly elevated titres in the group with untreated celiac disease and lower titres in the treated group.In the treated group, 28% of the patients continued to have elevated titres despite being on a glutenfree diet.The majority of patients (80%) were also positive for IgA EmA.
The sensitivity and specificity of our IgA tTG assay were comparable with those of the IgA EmA assay using monkey esophagus (2,6,7).Studies using human umbilical cord as the substrate have generally quoted better results (1-3), and this is thought to be a result of using human tissue to detect human antibody.The sensitivity of the tTG antibody assay has also been reported to be improved by using cloned human tTG rather than guinea pig tTG ( 8), but this is not yet widely available.
IgA EmA has been used to monitor the response to treatment but can take 12 months to disappear (9), making their use in this clinical setting limited.Nevertheless, the presence of EmA despite treatment can signify continued disease activity.We found that the titres of tTG antibody were higher in the treated group positive for EmA than in the equivalent group negative for EmA, suggesting that serial testing of tTG antibody titres would be useful for noninvasive monitoring of the response to treatment.This is also suggested by our finding that higher titres are generally found in more severe histological lesions.
SIgAD is estimated to occur in one in 400 to one in 700 individuals (10); however, the prevalence of SIgAD in patients with celiac disease has been reported to be as high as 2% to 3% (11)(12)(13).Relying solely on IgA tissue antibodies such as EmA to detect celiac disease, therefore, would lead to missed diagnosis in these patients.The one patient in our study with SIgAD had a very low titre of tTG antibody (less than 1 AU/mL) -below our calculated reference range of 1.4 to 138.4 AU/mL.The use of IgA tTG antibody with the addition of total IgA levels in patients with very low titres would, therefore, minimize the risks of failure to detect a patient with both SIgAD and celiac disease.The use of the IgG class of tTG antibody may also be useful in detecting such patients.
Serological testing does not replace the need for small intestinal biopsy in the diagnosis of celiac disease, and these tests are aimed at screening for the disease in high risk groups or in patients in whom the index of suspicion of the disease is low.The very high negative predictive value indicates that tTG antibody will make an excellent screening tool because only a small number of patients will undergo small intestinal biopsy for a false positive result.Our one false negative result does, however, reinforce the need for small intestinal biopsy in all patients with symptoms suggestive of celiac disease.

Figure 1 )
Figure 1) Titres of immunoglobulin (Ig) A antibody to tissue transglutaminase in patients with untreated celiac disease, patients with treated celiac disease (IgA endomysium antibody [EmA] -positive and -negative), biopsy control subjects and nonbiopsy control subjects.Horizontal lines indicate a reference range of 1.4 to 138.4 arbitrary units/mL

TABLE 3 Serial titres of immunoglobulin (Ig) A tissue transglutaminase (tTG) antibodies in six patients with treated celiac disease and four patients with untreated celiac disease Patient Time interval (months) IgA tTG antibody titres (AU/mL)
Titres of immunoglobulin (Ig) A tissue transglutaminase (tTG) antibody in patients with treated celiac disease and patients with untreated celiac disease by severity of small bowel mucosal lesion.Titres from those with normal mucosae differed from titres from those with moderate (P=0.0001) and severe histology (P=0.0002).Titres were higher in those with severe lesions than in those with moderate lesions (P=0.0282).AU Arbitrary Units from the nonbiopsy control group were positive for IgA EmA.