506 Gut 2001;49:506–511 Guinea pig immunolinked assay

does not predict in patients with Gut: first published as 10.1136/gut.49.4.506 on 1 October 2001. Downloaded from chronic liver disease

A Carroccio, L Giannitrapani, M Soresi, T Not, G Iacono, C Di Rosa, E Panfili, A Notarbartolo, G Montalto

Abstract patients with atrophy of the intestinal Background—It has been suggested that mucosa were positive for anti-tTG while serological screening for coeliac disease all others were negative, including those (CD) should be performed in patients false positive in the ELISA based on with chronic unexplained hypertransami- guinea pig tTG as the antigen. nasaemia. Conclusions—In patients with elevated Aims—To evaluate the specificity for CD transaminases and chronic liver disease diagnosis of serum IgA antitissue trans- there was a high frequency of false glutaminase (tTG) determination in con- positive anti-tTG results using the ELISA secutive patients with chronic based on tTG from guinea pig as the anti- hypertransaminasaemia using the most gen. Indeed, the presence of anti-tTG did widely utilised ELISA based on tTG from not correlate with the presence of EmA or guinea pig as the antigen. CD. These false positives depend on the Patients and methods—We studied 98 presence of hepatic in the com- patients with chronic hypertransamina- mercial tTG obtained from guinea pig saemia, evaluated for the first time in a liver and disappear when human tTG is hepatology clinic. Serum anti-tTG and used as the antigen in the ELISA system. antiendomysial (EmA) assays were per- We suggest that the commonly used tTG formed. Patients positive for EmA and/or ELISA based on guinea pig antigen should anti-tTG were proposed for intestinal not be used as a screening tool for CD in biopsy. Finally, all sera were reassayed for patients with chronic liver disease. ( 2001; :506–511) anti-tTG using an ELISA based on human Gut 49

recombinant tTG as the antigen. Keywords: liver disease; coeliac disease; antitissue http://gut.bmj.com/ Results—A total of 94/98 hypertransami- transglutaminase ; antiendomysial nasaemic patients were positive for hepa- antibodies; ; intestinal histology titis virus markers, with 82/98 (83%) positive for anti-hepatitis C virus. Liver histology showed that most patients had Hepatic damage is a frequent finding in mild or moderate chronic hepatitis while patients with coeliac disease (CD) on a severe fibrosis or overt liver cirrhosis was containing diet; in fact, hypertransaminasae- found in 20/98. CD screening showed that mia has been reported in 10–54% of patients at on September 29, 2021 by guest. Protected copyright. CD diagnosis.1–4 As a consequence, it has been Internal Medicine, 15/98 (16%) hypertransaminasaemic sub- University Hospital of jects had anti-tTG values in the same suggested that serological screening for CD Palermo, Palermo, should be performed in patients with chronic range as CD patients; however, IgA EmA 5 Italy were positive in only 2/98 (2%). Distal unexplained hypertransaminasaemia. In this A Carroccio respect, a previous study revealed that the duodenal biopsy, performed in nine pa- L Giannitrapani serum antiendomysial (EmA) assay is tients, showed subtotal villous atrophy in M Soresi the optimum test for predicting CD in patients C Di Rosa the two EmA+/anti-tTG+ patients but was with chronic liver disease6 but this indirect A Notarbartolo normal in 7/7 EmA−/anti-tTG+ subjects. immunofluorescence test is not easy to apply in G Montalto The presence of anti-tTG+ values in large scale screening. However, tissue trans- I Pediatric Division, EmA− patients was unrelated to particu- glutaminase (tTG) has recently been identified “Di Cristina lar gastrointestinal symptoms, other asso- as the main (or sole) autoantigen recognised by Hospital”, Palermo, ciated diseases, severity of liver histology, EmA in CD patients7 and this has permitted Italy or distribution of viral hepatitis markers. G Iacono the use of an ELISA test, based on commercial There was a significantly higher frequency guinea pig tissue transglutaminase, to detect Pediatric Division, of positive serum autoantibodies (anti- the presence of anti-tTG autoantibodies in IRCCS Burlo nuclear, antimitochondrial, antismooth Garofolo, Trieste, Italy muscle, and anti-liver- microsomal Abbreviations used in this paper: CD, coeliac T Not antibodies) in anti-tTG+/EmA− patients disease; EmA, antiendomysial antibody; tTG, tissue E Panfili than in the other subjects (9/13 v 10/83; transglutaminase; AST, aspartate aminotransferase; p<0.003). Also, a correlation was found ALT, alanine aminotransferase; HCV, hepatitis C virus; Correspondence to: ANA, antinuclear antibodies; AMA, antimitochondrial Dr A Carroccio, Via CoVaro between serum gamma globulin and anti- 25, 90124 Palermo, Italy. tTG values (p<0.01). When sera were antibodies; ASMA, antismooth muscle antibodies; [email protected] anti-LKM, anti-liver-kidney microsomal antibodies; tested with the ELISA based on human h-tTG, human recombinant tissue transglutaminase; Accepted for publication tTG as the antigen, no false positive PBS, phosphate buVered saline; IEL, intraepithelial 26 February 2001 results were observed: only the two EmA+ lymphocyte.

www.gutjnl.com Anti-tTG autoantibodies in chronic liver diseases 507

serum, specific for the diagnosis of CD.7–11 tTG (h-tTG) as antigen, according to the Previous studies on the clinical utility of methods described below.

anti-tTG determination have reported no or Control sera for evaluation of anti-tTG anti- Gut: first published as 10.1136/gut.49.4.506 on 1 October 2001. Downloaded from very rare false positive results for CD bodies were obtained from two diVerent diagnosis,7–11 and when positive results have groups: the first group included 20 EmA posi- been found in patients with normal intestinal tive coeliac patients on a gluten containing diet histology, the hypothesis of latent CD has been with total or subtotal intestinal villous atrophy, advanced.89However, as in the case of patients diagnosed according to the criteria of the with chronic liver disease (primary biliary European Society of Pediatric Gastroenterol- cirrhosis), the occurrence of false positive anti- ogy and Nutrition15; the second group of tTG antibody results has been found.12 control sera was obtained from 35 EmA nega- In the present study, we evaluated serum tive healthy subjects who were members of the anti-tTG and EmA in consecutive patients medical and laboratory personnel of our clinic. with chronic hypertransaminasaemia due to Subjects in these control groups were sex and various causes. age matched with the hypertransaminasaemic patients. Patients and methods All subjects gave informed consent and the The study included 98 consecutive subjects protocol was approved by the ethics committee (66 males; age range 18–64 years, median 36) of our hospital. with chronic hypertransaminasaemia (serum aspartate aminotransferase (AST) and alanine SERUM ANTIENDOMYSIAL ANTIBODY aminotransferase (ALT) levels >40 IU/l for DETERMINATION more than two months) who attended the out- As previously described,16 IgA class EmA patient clinic for liver disease at the Internal values were determined using a commercially Medicine Division of the University Hospital available indirect immunofluorescence tech- of Palermo for the first time between Septem- nique (Anti-endomisio; Eurospital Pharma, ber 1998 and May 1999. Patients previously Trieste, Italy). hospitalised in our division or examined in our outpatient clinics for liver or gastrointestinal SERUM ANTI-tTG ELISA DETERMINATION USING diseases were excluded; we also excluded sub- tTG FROM GUINEA PIG AS ANTIGEN jects with known CD and those who in the past This assay was performed by an inhouse had undergone complete serological evaluation ELISA in accordance with the method de- for CD diagnosis. scribed by Troncone and colleagues,9 adding

In all patients, alcohol intake, use of drugs, 5 mmol/l CaCl2 to U bottomed microtitre and exposure to potential hepatic toxins were plates according to Sulkanen and colleagues.8 investigated. Laboratory investigations in- Values were expressed as a percentage of posi- cluded routine liver and kidney function tests. tive reference sera, obtained from untreated http://gut.bmj.com/ Immunoglobulin levels were evaluated to coeliac patients diagnosed according to the exclude IgA deficiency. Furthermore, all sub- criteria of the European Society of Pediatric jects underwent serological screening for viral Gastroenterology and Nutrition,15 showing in hepatitis B and C (HCV); anti-HCV immuno- all cases the presence of EmA. Anti-tTG values reactivity was confirmed by a third generation greater than the 95th percentile of the control immunoblot assay (RIBA 3; Chiron Corpora- group, including over 100 healthy controls

tion, Emeryville, California, USA, and Ortho negative for serum EmA, were considered on September 29, 2021 by guest. Protected copyright. Diagnostic Systems). Sera were also tested for positive (8% of the reference serum). The hepatitis B surface antigen using a commercial intra-assay coefficient of variation for the IgA ELISA (Abbott Diagnostic, North Chicago, t-TG autoantibody ELISA was 8.7% (n=22), Illinois, USA). In all subjects the presence of and the inter-assay coeYcient of variation was antinuclear (ANA), antimitochondrial (AMA), 10.3% (n=18). antismooth muscle (ASMA), and anti-liver- kidney microsomal (anti-LKM) antibodies was HUMAN RECOMBINANT TRANSGLUTAMINASE also evaluated by indirect immunofluorescence The h-tTG gene was amplified from the intes- using commercial kits. tinal biopsy of a untreated patient with CD All patients underwent percutaneous liver using primers specific for the coding region of biopsy with a Menghini needle in accordance the gene, as previously described.17 Briefly, with the procedures and precautions previously cDNA was cloned into an expression vector described.13 Liver histological evaluation was (pET28b; Novagen, Madison, Wisconsin, performed according to Desmet and col- USA) expressed in bacteria and purified under leagues.14 non-denaturing conditions using IMAC (Qia- Serological screening for CD was performed gen, Valencia, California, USA). Purity of the in all patients by serum EmA and anti-tTG recombinant was assessed by sodium assays based on tTG from guinea pig as the dodecyl sulphate-polyacrylamide gel electro- antigen, in accordance with the methods phoresis. described below. Subjects with positive serum EmA and/or anti-tTG were asked to undergo SERUM ANTI-tTG ELISA DETERMINATION USING intestinal biopsy and commence a gluten free HUMAN RECOMBINANT tTG AS ANTIGEN diet to confirm the suspected CD diagnosis. This assay was performed in the laboratory of In the second part of the study we re- the paediatric department of the University evaluated all sera for anti-tTG antibodies with Hospital of Trieste, on serum samples which a new ELISA based on human recombinant had been kept frozen at −80°C; a control test

www.gutjnl.com 508 Carroccio, Giannitrapani, Soresi, et al

Table 1 Liver histology findings and viral hepatitis markers in the 98 consecutive 40 hypertransaminasaemia patients included in the study

Anti-HCV+ Anti-HCV− Anti-HCV+ Anti-HCV− Gut: first published as 10.1136/gut.49.4.506 on 1 October 2001. Downloaded from HBsAg− HBsAg+ HBsAg+ HBsAg− 30 Normal histology 1 — — — Steatosis 3 1 — 2 Minimal changes 3 2 — — Minimal chronic hepatitis 15 5 — — 20 Mild chronic hepatitis 30 — 1 — Moderate chronic hepatitis 9 2 1 — Severe chronic hepatitis 3 — — — Hepatitis with severe fibrosis 13 — — — 10 Liver cirrhosis 5 — — 2

Serum anti-tTG values (AU) Serum 5 HCV, hepatitis C virus; HBsAg, hepatitis B surface antigen. performed in that laboratory on 20 serum Group 1 Group 2 Group 3 samples in which anti-tTG antibodies were Figure 1 Serum IgA class tissue transglutaminase first assayed on fresh serum and then eight antibody (anti-tTG) titres determined by ELISA ° performed using Sigma tTG from guinea pig as the months later after freezing at −80 C showed antigen. The results are expressed as AU values. The that the preservation did not significantly alter following groups were studied: group 1, the results (interassay coeYcient of variation hypertransaminasaemia patients (n=98); group 2, coeliac disease patients (EmA positive) on a gluten containing diet was 9.6%). (n=20); group 3, healthy control subjects (EmA negative) Serum IgA anti-h-tTG antibodies were (n=35). An arbitrary cut oV level for positivity (broken determined as recently reported, with slight line) was drawn at an AU of 8. modifications.17 The human recombinant anti- gen was diluted in phosphate buVered saline STATISTICAL ANALYSIS The percentage of anti-tTG and EmA positive (PBS) to yield a protein concentration of results was calculated. Analysis of frequency 10 µg/ml. A 0.1 ml aliquot of this solution was was performed using the ÷2 test. Spearman’s r placed in each well of a flat bottomed plate correlation coeYcient was used to evaluate the (EIA/RIA 2580; Costar, Cambridge, Massa- association of serum anti-tTG results with liver chusetts, USA). After an overnight incubation histology or laboratory test results. Multiple ° at 4 C, the plates were washed three times in linear regression analysis was performed to PBS-0.05% Tween 20 and blocked with PBS- evaluate the association between anti-tTG 0.1% Tween 20 for 20 minutes at room positivity and serum ALT and AST, serum temperature. Serum samples diluted 1:100 in albumin, serum bilirubin and alkaline phos- PBS-0.1% Tween 20 were incubated for one phatase, serum gamma globulin, serum ã hour at room temperature. The plates were glutamyl , and the presence of http://gut.bmj.com/ washed and incubated for one hour at room autoantibodies (ANA, ASMA, AMA, anti- temperature with 1:4000 phosphatase conju- LKM1) in chronic liver disease patients. gated antihuman IgA (Sigma A-3062) diluted in PBS-1% bovine serum albumin-4% polyeth- ylene glycol. The immune reaction was devel- Results Table 1 summarises the liver histology findings oped by adding solution and adsorb- and viral hepatitis marker results in all patients ance was read in a microplate reader at 405 nm

studied. on September 29, 2021 by guest. Protected copyright. until the positive control serum reached an Regarding serological screening for CD, fig 1 optical density value of 1.9. Results were shows anti-tTG values obtained with the expressed as a percentage of the positive ELISA based on tTG from guinea pig as the control serum. Normal values were taken as antigen in subjects with chronic liver disease <16% which represented a value >2 SD above compared with the two control groups; all CD the mean of 500 healthy subjects. The patients with known EmA positivity (CD con- intra-assay coeYcient of variation for the IgA trols, group 2) had anti-tTG values above the h-tTG autoantibody ELISA was 4% (n=10) normal limit (range 10–40 AU) whereas all and the inter-assay coeYcient of variation was healthy EmA negative controls (group 3) had 9% (n=10). This ELISA method has a values within the normal range. In the sensitivity of 97% and a specificity of 99%. hypertransaminasaemic patients (group 1), we found 15/98 (16%) subjects with anti-tTG values above normal; these values were in the INTESTINAL HISTOLOGY same range as those observed in the CD Biopsy specimens were obtained and orien- 18 19 patients. However, when IgA EmA were tated as previously described. Specimens assayed, only 2/98 (2%) patients with chronic were embedded in paraYn. Slides were stained liver disease were positive; in fact, EmA was with haematoxylin and eosin and graded by negative in all anti-tTG negative hyper- conventional histology as normal, partial vil- transaminasaemic patients but also in 13/15 lous atrophy, and subtotal villous atrophy. Fur- patients positive for serum IgA anti-tTG. thermore, a count of the intraepithelial lym- Nine of 15 hypertransaminasaemic patients phocyte (IEL) population was performed and positive for anti-tTG, including two EmA their number was calculated per number of positive patients, consented to intestinal biopsy enterocytes (normal values <35 IEL per 100 for histological study. Intestinal histology enterocytes). Histology was described by an showed subtotal mucosa atrophy in two patients examiner unaware of the laboratory test results. who were positive for both EmA and anti-tTG

www.gutjnl.com Anti-tTG autoantibodies in chronic liver diseases 509

Table 2 Clinical, laboratory, and histological characteristics of the 15 subjects positive for serum IgA anti-tTG antibodies, compared with all other anti-tTG negative hypertransaminasaemia patients

Anti-tTG Natural ALT/AST Intest histology Intestinal Associated Gut: first published as 10.1136/gut.49.4.506 on 1 October 2001. Downloaded from Case No (AU) EmA autoantibodies (UI/L) Liver histology (villi/crypts) histology (IEL) diseases Viral markers 1 17 + ANA+ 40/46 MILH 0.8 63 — HCV+ 2 14 + ASMA+ 60/84 S 1.1 71 IDDM HCV−/HBV− 3 14 − ASMA+ 39/98 MH 3.3 14 — HCV+ 4 10.5 − ANA+ 177/221 SF 3.4 19 — HCV+ 5 8.7 − LKM1+ 67/75 MODH 4.0 30 — HBV+ 6 11 − None 924/698 SF 2.9 27 — HCV+ 7 9.4 − None 84/80 MH 3.6 31 — HCV+ 8 13 − ANA+ 165/117 LC 3.5 28 CG HCV−/HBV− 9 9 − None 47/49 MH Not perf. Not perf. — HCV+ 10 15 − AMA+ LKM1+ 31/53 MH Not perf. Not perf. — HBV+ 11 8.4 − ANA+ 116/162 MH Not perf Not perf. — HCV+ 12 9.7 − ASMA+ ANA+ 121/170 SF Not perf Not perf. Thyroiditis HCV+ 13 8.8 − ANA+ AMA+ 89/95 MH Not perf Not perf. — HCV+ 14 16 − None 42/44 S 2.9 17 — HCV−/HBV− 15 9.9 − ANA+ 84/106 S Not perf Not perf. — HBV+ Anti-tTG negative Range 2–7.2 83/83 neg ASMA+ 2 cases, N=1 Not perf. Not perf. CG=4, HCV+=73 controls AMA+ 2 cases, S=3 IDDM=2, HBV+=7 ANA+ 4 cases, MCHS=5 NIDDM=3, HCV+/ HBV+=2 LKM1 +2 cases MH=14 PU=6, HCV−/ HBV−=1 MILH=30 CVD=3 MODH=11 SH= 3 SF=10 LC=6

EmA, antiendomysial antibody; tTG, tissue transglutaminase; AST, aspartate aminotransferase; ALT, alanine aminotransferase; HCV, hepatitis C virus; HBV, hepa- titis B virus; ANA, antinuclear antibodies; AMA, antimitochondrial antibodies; ASMA, antismooth muscle antibodies; LKM, anti-liver-kidney microsomal antibod- ies; IEL, intraepithelial lymphocyte. Liver histology: N, normal; S, steatosis; MCHS, minimal changes; MH, minimal chronic hepatitis; MILH, mild chronic hepatitis; MODH, moderate chronic hepati- tis; SH, severe chronic hepatitis; SF, hepatitis with severe fibrosis; LC, liver cirrhosis. Associated diseases: CG, congestive gastropathy; IDDM, insulin dependent diabetes mellitus; NIDDM, non-insulin dependent diabetes mellitus; PU, peptic ulcer; CVD, cardiovascular disease. Normal reference values: anti-t-TG <8 AU; EmA: serum dilution <1:5; natural autoantibodies: serum dilution <1:80; ALT and AST <40 UI/l; villi/crypts ratio >2.8; IEL count <35/100 enterocytes.

(villi/crypts ratio 0.8–1.1) with a marked in- values (Spearman’s r correlation coeYcient crease in IELs. However, none of the other seven 0.33; p<0.01). Finally, we performed a multi-

subjects positive for serum IgA anti-tTG (but ple linear regression analysis which describes http://gut.bmj.com/ EmA negative) had intestinal mucosa abnor- the relationship between a dependent variable, malities: all had a normal villi/crypts ratio (range such as positivity of serum anti-tTG in the 2.9–4) and in all the IEL count was below the ELISA based on tTG from guinea pig as the normal limit for our laboratory. antigen, and various independent laboratory Table 2 shows the clinical data, liver and variables; this analysis confirmed the signifi- intestinal histology, and laboratory characteris- cant association between serum anti-tTG posi- tics of the two EmA+/anti-tTG+ and the other

tivity and two of the independent variables on September 29, 2021 by guest. Protected copyright. 13 EmA−/ anti-tTG+ patients. The 13 EmA−/ examined—serum autoantibody positivity anti-tTG+ patients did not diVer from the (beta value 0.566, p<0.0004) and serum other 83 anti-tTG− hypertransaminasaemic gamma globulin (beta value 0.465, p<0.008). subjects for the presence of gastrointestinal There was no correlation between anti-tTG symptoms or associated diseases: none had positivity and the other independent variables overt malabsorption syndrome and no diVer- examined (r2=0.46). ence in body mass index was observed. Severity Eight months after the conclusion of the first of liver disease evaluated from liver histology part of the study, we performed an ELISA for was identical in both groups, and the distribu- IgA anti-tTG antibodies using h-tTG as the tion of viral hepatitis markers was also similar. However, there was a significantly higher antigen, and the sera of all patients included in frequency of patients with positive serum this study were re-evaluated. In the patient autoantibodies (ANA, AMA, ASMA, anti- group with chronic liver disease, only the two LKM) in the anti-tTG+ group than in the subjects with atrophy of the intestinal mucosa other subjects (9/13 v 10/84; ÷2=9.82, and serum EmA+ at the moment of diagnosis p<0.003) and serum ANA titres correlated had elevated anti-tTG antibodies: their values with anti-tTG values (Spearman’s r correlation were in the range of those observed in patients coeYcient 0.78; p<0.0001). There was no cor- with CD (19% and 32.1%, respectively). All relation between serum anti-tTG values and other patients with chronic liver disease had most of the other laboratory parameters evalu- anti-tTG values within the normal limit (range ated (serum albumin, serum ALT and AST, 1–13.1 %), including the 13 subjects with false serum bilirubin and alkaline phosphatase, pro- positive anti-tTG values when evaluated with activity, serum ã glutamyl trans- the ELISA based on tTG from guinea pig as ferase, and platelet count) but a statistically the antigen. Finally, in the serum of these 13 significant positive correlation was found patients we repeated the anti-tTG ELISA between serum gamma globulin and anti-tTG based on tTG from guinea pig as the antigen:

www.gutjnl.com 510 Carroccio, Giannitrapani, Soresi, et al

in all cases a false positive result was con- guinea pig liver; it has been verified that Sigma firmed. For the control groups, all CD patients tTG is composed of at least 14 diVerent com- 23

with known EmA+ were positive for serum IgA ponents and it is logical to hypothesise that in Gut: first published as 10.1136/gut.49.4.506 on 1 October 2001. Downloaded from anti-tTG (range 17–116%) whereas none of the sera of patients with chronic liver disease the healthy controls were positive (range there could be antibodies for protein antigens 1–12.3%) from the liver matrix, probably common (or The two patients with subtotal intestinal cross reacting) to humans and other mammals. mucosa atrophy began a gluten free diet and In fact, in the second part of the study when we after 4–5 months of this diet showed negative were able to test our sera for anti-tTG using a serum IgA EmA and anti-tTG; in one of these pure human recombinant tTG, we did not patients, who was anti-HCV positive, hyper- observe any false positive results: all 13 subjects transaminasaemia persisted on a gluten free with false positive anti-tTG values when evalu- diet. In the other patient, negative for viral ated with the ELISA based on tTG from hepatitis markers, hypertransaminasaemia dis- guinea pig as antigen resulted in negative appeared after three months on a gluten free values with the new ELISA system. This diet. None of the patients positive for anti-tTG confirms that human tTG ELISA has a higher but EmA− commenced a gluten free diet. diagnostic accuracy in CD diagnosis than the commonly used anti-tTG ELISA based on Discussion guinea pig antigen. However, it must be It is known that CD can cause chronic remembered that all gastroenterogical studies hypertransaminasaemia both in adults35and in published to date7–11 which reported a very high children2420 and the relative risk for CD in specificity of anti-tTG determination for CD patients with chronic unexplained hyper- diagnosis were performed with the same transaminasaemia compared with the general ELISA used in the first part of the present population has been estimated as 18.6.21 A study (based on tTG from guinea pig liver) and previous study indicated that in patients with that this method has been suggested for large chronic liver diseases, serum EmA determina- scale CD screening.81011 Our findings appear tion is useful for CD diagnosis.6 However, the to be linked to the presence of chronic liver EmA assay has several limitations: interpret- disease and/or hyperglobulinaemia and we ation of the immunofluorescence pattern is emphasise that such a high frequency of false subjective, it requires monkey oesophagus or positive anti-tTG results cannot be extrapo- human tissues as substrate, and it is not easy to lated to the general population. use in large scale screening. Thus the recent In conclusion, we found a high frequency of identification of the protein tTG as the false positive anti-tTG test results in patients autoantigen of CD7 has made possible large with chronic liver disease using the commonly scale use of an ELISA test, based on tTG anti- used anti-tTG ELISA based on tTG from gen from guinea pig liver, to detect IgA class guinea pig liver as antigen. The presence of http://gut.bmj.com/ anti-tTG antibodies. anti-tTG did not correlate with the presence of Previous studies on IgA anti-tTG determi- EmA or CD. We demonstrated that these false nation have indicated a very high specificity of positive results were due to the “nature” of the this ELISA assay for CD diagnosis, and it has antigen used in the ELISA system and that been suggested that the false positive results they disappeared if a new ELISA, based on could be latent CD cases7–9; hence the use of human recombinant tTG as antigen, was

the ELISA based on this protein has been pro- employed. Thus the anti-human-tTG ELISA, on September 29, 2021 by guest. Protected copyright. moted to screen large populations.11 However, and not the anti-tTG assay based on tTG from none of the previous studies included patients guinea pig as the antigen, must be used as a with chronic liver disease of diVerent causes screening tool for CD in patients with chronic and, before the present study, we had no data liver disease. on the use of the anti-TG antibody assay in these patients. Our data clearly indicate that This work was supported by a grant to G Montalto, 60% from serum IgA anti-TG ELISA determination MURST, and by grant 12/99 from IRCCS “Burlo Garofalo”. based on tTG from guinea pig liver as the antigen is not useful in screening patients with 1 Hagander B, Brandt L, Sjolund K, et al. Hepatic injury in chronic liver disease for CD. In fact, we found celiac disease. Lancet 1977;2:270–2. 2 Bonamico M, Pitzalis G, Culasso F. Hepatic damage during that all patients with positive anti-tTG anti- coeliac disease in childhood. Minerva Pediatr 1986;38:959– bodies and negative EmA, who agreed to 62. 3 Bardella MT, Fraquelli M, Quatrini M, et al. Prevalence of undergo intestinal biopsy (7/13 subjects), had hypertransaminasemia in adult celiac patients and eVect of normal intestinal histology, with normal villi gluten-free diet. Hepatology 1995;22:833–6. 4 Carroccio A, Iannitto E, Cavataio F, et al. Sideropenic ane- and crypts and no increase in IELs, which is mia and celiac disease: one study, two points of view. Dig considered a marker of latent CD.22 Thus it is Dis Sci 1998;43:673–8. 5 Gonzalez-Abraldes J, Sanchez-Fueyo A, Bessa X, et al. Per- evident that the current ELISA based on sistent hypertransaminasemia as the presenting feature of guinea pig tTG is unspecific when used for celiac disease. Am J Gastroenterol 1999;94:1095–7. 6 Sjoberg K, Lindgren S, Eriksson S. Frequent occurrence of CD diagnosis in patients with chronic liver non-specific antibodies in chronic liver disease. disease. Endomysial but not gliadin antibodies predict coeliac disease in patients with chronic liver disease. Scand J Gas- Regarding the cause of these “false positive” troenterol 1997;32:1162–7. anti-tTG tests, we demonstrated that this phe- 7 Dieterich W, Ehnis T, Bauer M, et al. Identification of tissue transglutaminase as the autoantigen of celiac disease. Nat nomenon was clearly linked to the purity of the Med 1997;3:797–801. antigen used in the currently advised anti-tTG 8 Sulkanen S, Halttunen T, Laurila K, et al. Tissue ELISA. In fact, this widely used ELISA is transglutaminase autoantibody -linked immuno- sorbent assay in detecting coeliac disease. Gastroenterology based on tTG which is a crude extract from 1998;115:1322–8.

www.gutjnl.com Anti-tTG autoantibodies in chronic liver diseases 511

9 Troncone R, Maurano F, Rossi M, et al. IgA antibodies to 16 Carroccio A, Cavataio F, Iacono G, et al.IgA antiendomysial tissue transglutaminase: an eVective diagnostic test for antibodies on the umbilical cord in diagnosing celiac celiac disease. J Pediatr 1999;134:166–71. disease. Scand J Gastroenterol 1996;31:759–63. 10 Brusco G, Izzi L, Corazza GR. Tissue transglutaminase 17 Sblattero D, Berti I, Trevisiol C, et al. Human recombinant antibodies for coeliac disease screening. Ital J Gastroenterol tissue transglutaminase ELISA: an innovative diagnostic Gut: first published as 10.1136/gut.49.4.506 on 1 October 2001. Downloaded from Hepatol 1998;30:496–7. assay for coeliac disease. Am J Gastroenterol 2000;95:1253– 11 Dieterich W, Laag E, Schopper H, et al. Autoantibodies to 7. tissue transglutaminase as predictors of celiac disease. Gas- 18 Carroccio A, Iacono G, Lerro P, et al. Role of pancreatic troenterology 1998;115:1317–21. impairment in growth recovery during gluten-free diet in 12 Habior AB, Lewartowska A, Orlowska J, et al. Autoantibod- childhood celiac disease. 1997;112:1839–4 ies to tissue transglutaminase are not a marker of celiac dis- Gastroenterology 19 Perera DR, Weinstein WM, Rubin CE. Small intestinal ease associated with primary biliary cirrhosis. Hepatology 1999;30:474A. biopsy. Hum Pathol 1975;6:157–217. 13 Soresi M, Carroccio A, Bonfissuto G, et al. Ultrasound 20 Vajro P, Fontanella A, Mayer M, et al. Elevated serum detection of abdominal lymphadenomegaly in subjects aminotransferase activity as an early manifestation of with hepatitis C virus infection and persistently normal gluten-sensitive enteropathy. J Pediatr 1993;122:416–19. transaminases. A predictive index of liver histology severity. 21 Bardella MT, Vecchi M, Conte D, et al. Chronic unex- J Hepatol 1998;28:544–9. plained hypertransaminasemia may be caused by occult 14 Desmet VJ, Gerber M, Hoofnagle JH, et al. Classification of celiac disease. Hepatology 1999;29:654–7. chronic hepatitis: diagnosis, grading, staging. Hepatology 22 Marsh MN. The mucosal pathology in gluten sensitivity. In: 1994;19:1513–20. Marsh MN, ed. Coeliac disease. Oxford: Blackwell Scien- 15 Walker-Smith JA, Guandalini S, Schmitz J, et al. Revised tific, 1992:136–91. criteria for diagnosis of coeliac disease. Arch Dis Child 23 Maki M. Tissue transglutaminase as the autoantigen of coe- 1990;65:909–11. liac disease. Gut 1997;41:565–6.

Want full access but don't have a subscription? http://gut.bmj.com/

Pay per access

For just US$25 you can have instant access to the whole website for 30 days. During this time you will be able to access the full on September 29, 2021 by guest. Protected copyright. text for all issues (including supplements) available. You will also be able to download and print any relevant pdf files for personal use, and take advantage of all the special features Gut online has to offer.

www.gutjnl.com

www.gutjnl.com