Prevalence of celiac disease in argentina: screening of an adult population in the La Plata area


of 6
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Prevalence of celiac disease in argentina: screening of an adult population in the La Plata area
  Value of a Screening Algorithm for Celiac DiseaseUsing Tissue Transglutaminase Antibodies as FirstLevel in a Population-Based Study Juan C. Gomez, M.D., Gisella Selvaggio, M.D., Bibiana Pizarro, M.D., Martin J. Viola, M.D.,Graciela La Motta, M.D., Edgardo Smecuol, M.D., Roberto Castelletto, M.D., Rau´l Echeverrı´a, M.D.,Horacio Vazquez, M.D., Roberto Mazure, M.D., Adriana Crivelli, M.D., Emilia Sugai, M.D.,Eduardo Maurin˜o, M.D., and Julio C. Bai, M.D.  Malabsorption and Nutritional Unit, San Martin Hospital, La Plata, Argentina; and Small Intestinal Section,Clinical Service, Department of Medicine, “Dr. Carlos Bonorino Udaondo” Gastroenterology Hospital, Buenos Aires, Argentina OBJECTIVE:  Serological screening for celiac disease (CD)can detect a large number of otherwise undiagnosed patientsbased on the sequential evaluation of serological tests andintestinal biopsy. The aim of this study was to compare thescreening value for CD of two different protocols for thesame community-based population. METHODS:  We screened 1000 consecutive subjects (497women, age range 16–71 yr) attending a centralized labo-ratory for obligatory prenuptial blood tests. Serum samplesobtained from all subjects were processed using two differ-ent protocols: 1) a three-level classic screening consisting of the parallel use of IgG and IgA antigliadin antibodies as firstlevel, followed by endomysial antibodies and total serumIgA for positive patients, and finally, intestinal biopsy of positive patients; and 2) a study screening protocol consist-ing of the parallel use of a commercial guinea pig antitissuetransglutaminase antibody and total serum IgA as first line,endomysial antibodies (type IgA and/or IgG) for positivepatients, and finally, intestinal biopsy. RESULTS:  The classic screening protocol identified five sub- jects who were eligible for intestinal biopsy, which con-firmed the presence of CD in all (prevalence 5.0    1000,95% CI    1.6–11.6). Using the study algorithm, we de-tected seven new patients including the five patients de-tected by the first protocol (prevalence 7.0    1000, 95%CI    2.8–14.4). The two additional patients diagnosedusing the proposed algorithm had positive IgG antigliadinantibodies and normal total serum IgA and were not de-tected by the classic protocol. Both patients were endomy-sial antibodies positive. The comparative analysis showedthat the classic approach was more expensive (U.S. $4687per new patient detected) compared with the proposed studyalgorithm (U.S. $3006). CONCLUSIONS:  Our data showed that a new screening pro-tocol using antitissue transglutaminase as first line followedby endomysial antibodies is a cost-effective screening andyielded more realistic figures of prevalence for CD in acommunity setting than the classic three-level sequentialevaluation using antigliadin antibodies. (Am J Gastroenterol2002;97:2785–2790. © 2002 by Am. Coll. of Gastroenter-ology) INTRODUCTION Celiac disease (CD) is a very common life-long intoleranceto gluten occurring in genetically predisposed individuals.Very recent studies have shown that CD is increasinglyrecognized in every part of the world where it has beeninvestigated with prevalence values in Europe or America of 4.0  1000 individuals of the general population or higher(1–6). Interestingly, the reason for the lack of recognitionand diagnosis based on clinical consultation is that most of these patients are clinically silent (5). This high prevalencehas suggested the necessity of screening for CD becausemany of the health complications of CD appear to be relatedto the enteropathy rather than to the severity of symptoms(6, 7).Screening is the systematic application of tests to identifysubjects from the general population at sufficient risk of aspecific disorder to benefit from further investigation. Casefinding is the searching process to identify silent and/orsubclinical cases in a clinical setting. During the past 10 yr,both types of interventions have modified epidemiologicalcriteria on CD, and thus have allowed identification of otherwise undiagnosed cases (1).Most recent screening studies for CD have employed asimilar serological protocol consisting of a three-level se-quential evaluation based on the reliability of antigliadin(AGA) or endomysial (EmA) antibodies with the definitivediagnosis established by intestinal biopsy (8, 9). Interest-ingly, the recent advent of tissue transglutaminase as theautoantigen eliciting EmA have allowed to perform a newELISA-based test, which has been incorporated as a tool for T HE  A MERICAN  J OURNAL OF  G ASTROENTEROLOGY  Vol. 97, No. 11, 2002© 2002 by Am. Coll. of Gastroenterology ISSN 0002-9270/02/$22.00Published by Elsevier Science Inc. PII S0002-9270(02)05440-0  screening and diagnosis of CD (10). Although the sensitivityand specificity of these tests is critical for mass screening,they have not been assessed in population-based studies, andthere is no exact knowledge about the proportion of casesthat could be missed by the current serological screening.Our objective in the present study was to compare thescreening value for detecting CD using two different algo-rithms applied in a population-based study. MATERIALS AND METHODS Subjects Between March 1999 and May 2000, 503 couples attendingthe centralized laboratory of the San Martı´n Hospital (in LaPlata, Buenos Aires Province) for an obligatory prenuptialexamination were offered participation in a screening pro-gram for CD. They were part of a larger population screenedfor CD aiming to establish the prevalence of the disease inan adult urban population of Argentina (3). The screeningincluded 1000 consecutive subjects (497 women, age range16–71 yr). After a written informed consent, all the indi-viduals completed a clinical questionnaire at the time thatserum samples were obtained. The questionnaire inquiredabout previous diagnosis or suspicion of CD, diagnosis inrelatives, presence of diarrhea, anemia, diabetes, thyroiddisorders, malignancies and hepatic diseases, and gyneco-logical and obstetric history. Study Design The study compared the classic three-level screening pro-tocol with a proposed simplified algorithm.CLASSIC ALGORITHM. The first-level evaluation con-sisted of determination of serum levels of IgA and IgG AGAin samples obtained from all those subjects who agreed to beincluded in the study. After that, subjects negative for bothtests were excluded from further assessment. All those sam-ples positive to any or both antibodies went onto the secondlevel of the screening. Samples positive for either, bothantibodies, or for the IgA subtype exclusively were testedfor the IgA EmA. Sera positive only for IgG AGA weretested for serum levels of total IgA. Those with normal totalIgA concentration (  200 mg/ml) were excluded from fur-ther evaluation. Those samples with low total IgA serumlevels were tested for IgG EmA. Finally, those subjects withserum samples positive for EmA (either IgA or IgG) wereeligible for a small intestinal biopsy.PROPOSED ALGORITHM. All samples were tested inparallel for antitissue transglutaminase (antitTG) and totalserum IgA. All those subjects positive for antitTG or evi-dencing IgA deficiency were tested for EmA types IgAand/or IgG (only IgA-deficient samples). Those positivesubjects underwent intestinal biopsy. Diagnosis of CD wasbiopsy proven in all those patients included. CD-Related Serology Determinations performed on frozen serum samples col-lected at the time of the clinical interview included: 1) typesIgA and IgG AGA by ELISA as previously described usingcommercial kits (INOVA Diagnostics, San Diego, CA) (cut-off values for both tests were 20 AU/ml) (3); 2) EmA (IgAand IgG) by an immunofluorescence method on primateesophagus substrate (INOVA Diagnostics) (3) (duplicateserum samples were tested at 1:10 dilution); 3) antitTGperformed using a commercial guinea pig-based ELISA test(INOVA Diagnostics) (cutoff value 30 AU/ml) (11); and 4)total serum IgA (radial immunodiffusion test (Diffu-Plate,Biocientı´fica S.A., Buenos Aires, Argentina).All samples and decisions were made based on the resultsof the local laboratory. All positive samples and a randomlyselected group of negative ones were rechecked in a referrallaboratory by a trained observer who was unaware of theformer results. The agreement between results from bothlaboratories was absolute in terms of a bimodal result (pos-itive or negative), both in samples from new patients andthose from normal individuals. Small Bowel Histology Biopsy samples were obtained from the distal duodenum byduodenoscopy. Histological and quantitative morphometricevaluations (intraepithelial lymphocyte counts) were per-formed by an independent experienced observer (R.C.) whowas unaware of the clinical and laboratory findings of sub- jects. Intestinal biopsies were categorized according toMarsh’s criteria.  Ethical Issues and Statistics All individuals included in the study gave written informedconsent before the clinical interview according to the LisbonDeclaration of the World Medical Association. They wereclearly informed about the objectives of the study and theeventual necessity of small intestinal biopsy samples. Theprevalence and 95% CIs were calculated by conventionalformulas using the EpiInfo96 software. Direct cost analysisof both interventions was determined based on those esti-mated costs arising from expenditures of screening tests anddiagnostic procedures. Because both screening protocolsexhibited different effectiveness, a cost-effectiveness anal-ysis was performed and expressed as cost of screening perdiagnosed patient. Costs were expressed in U.S. currency.Tests were valued according to the current procedural ter-minology Code Book (“Nomenclador Nacional”) for bio-chemical and medical procedures available for the SocialSecurity System and for public hospitals of Argentina. Thecosts of EmA and antitTG, which are not included in theCode, were estimated based on the cost of equivalent bio-chemical determinations and commercial kits.The unitary costs for tests were as follow: 1) for AGA, thevalue was U.S. $10 for each subtype; 2) U.S. $12 for eachEmA subtype; 3) U.S. $12 for antitTG; 4) U.S. $52 forendoscopic biopsies; and 5) U.S. $24 for histological exam- 2786 Gomez  et al.  AJG – Vol. 97, No. 11, 2002  inations. We also estimated that 10% of samples were re-evaluated for first-level serological tests based on the pres-ence of borderline results. The cost of tests performed in thereferral laboratory was not included in the cost analysis. RESULTS  Results of the Classic Screening Algorithm Figure 1 shows the outcome of the classic screening in thestudied population. At the end of the first-level screening,86.3% of individuals were excluded from further studiesbased on the presence of negative IgA and IgG AGA tests.In contrast, 13.7% of serum samples were assessed with asecond-level evaluation based on the presence of positiveresults (sole IgA AGA positive 0.9%, sole IgG positive11.4%, and both IgA and IgG AGA tests positive 1.4%).After the second-level evaluation, only five subjects wereeligible for intestinal biopsy. Although four patients werepositive to both AGA tests and to EmA, one individual witha positive IgG AGA had IgA deficiency and exhibited apositive IgG EmA. The histological assessment of intestinalbiopsies showed a classic celiac flat mucosa in all fivepatients. The prevalence estimated according to this screen-ing algorithm is 5.0  1000 (95% CI  1.6–11.6).  Results of the Proposed Screening Algorithm Figure 2 shows in a schematic flow chart the proposedscreening protocol using antitTG and total serum IgA as thefirst-level approach. At the end of the first-step screeningwith antitTG and total IgA, we observed normal titers of both tests in 98.6% of samples. According to the protocol,all these subjects were excluded from further evaluation. Onthe other hand, 12 samples were IgA antitTG positive, andtwo others had serum IgA levels below the lower end of thenormal range. Although six of the former samples wereseropositive for EmA IgA, one sample with IgA deficit wasIgG EmA positive. The rest of the patients were seronega-tive for EmA. All EmA-positive patients had a characteristicceliac mucosal atrophy. Up to now, only two of the sixantitTG positive EmA negative samples underwent intesti-nal biopsy. Both histological samples showed normal mu-cosa. The prevalence estimated according to the proposedscreening algorithm is 7.0  1000 (95% CI  2.8–14.4). Characteristics of the New CD Patients Table 1 shows some clinical and epidemiological charac-teristics of the new CD patients. The gender ratio of thispopulation was as expected 2.5:1 (F:M), and the body massindex was normal in all patients. Although two patientspresented with subclinical CD (both with chronic anemia asthe sole marker), the other five patients were completelysilent from a clinical point of view. No patients had beendetected or suspected as CD before diagnosis. Biopsies fromall newly diagnosed CD patients had a flat duodenal mucosa(IIIc of Marsh’s criteria). Cost Evaluations of Screening Algorithms Table 2 shows the comparative analysis of costs of bothalgorithms discriminated according to their components.Although the overall expenditure for the classic protocolwas U.S. $23,437, such cost for the simplified algorithm wasU.S. $21,044. As the sensitivity of both protocols wasdifferent, we performed a cost-effectiveness analysis wherethe cost of the classic algorithm was U.S. $4,687 per newpatient diagnosed and for the proposed antitTG-based algo-rithm was U.S. $3,006. Figure 1.  Results of the screening algorithm using the classic protocol. 2787 AJG – November, 2002  Screening Algorithm for Celiac Disease  DISCUSSION The diversity of CD and its potential for complications areleading to noninvasive screening as a necessity to identifyand treat CD patients hitherto unrecognized (12–16). CD-related serology (AGA and EmA) is basically suitable forscreening the general population, but many problems stillremain unsolved, in particular the lack of estimation of thenumber of missed new CD cases and the cost-effectiveanalysis, among others. Our present study addressed bothaspects of screening to clarify future recommendations forthe application of screening algorithms. Thus, we accom-plished these goals by comparing two different screeningalgorithms on the same population of adult individuals at-tending a laboratory for an obligatory prenuptial examina-tion. One of them was the classic algorithm, and the otherwas the so-called by us “proposed algorithm,” which wasbased on the parallel use of a commercial guinea pig antitTGtest and total serum IgA as the first-level assessment. Themain finding of our study was that in comparison with thewidely used three-level system, the new algorithm increaseddetection of CD patients. Thus, although prevalence of CDdetected by the classic algorithm was 5.0    1000 individ-uals, the proposed algorithm estimated the prevalence in7.0    1000 subjects. Interestingly, although five patientswere diagnosed by both methodologies, the two other pa-tients detected by the study algorithm were IgG AGA pos-itive but negative for the IgA subtype in the classic screen-ing. Thus, according to this approach, both patients wereunderestimated and excluded from further investigation. Bycontrast, both subjects were positive for antitTG and EmAfor the new algorithm, and intestinal biopsy confirmed CD.The screening using a commercial guinea pig antitTG an-  Table 2.  Comparative Cost of Both Screening Algorithms in theSame Studied PopulationParametersClassic Algorithm(U.S.$)Proposed Algorithm(U.S.$)First level 22,000 19,800Second level 1,067 180Third level 370 1,064*Total value 23,437 21,044 * Includes biopsies from all positive IgA antitTG patients. Figure 2.  Results of the screening using the proposed protocol.  Table 1.  Epidemiological Data and Clinical Characteristics of the Newly Diagnosed CD PatientsCharacteristicsNumber (F:M) 7 (5:2)Age, median (yr) (range) 28 (23–31)Body weight, median (kg) (range) 59 (49–88)Height, median (cm) (range) 165 (157–186)Body mass index, median (kg/m 2 ) (range) 22.0 (18.5–25.5) 2788 Gomez  et al.  AJG – Vol. 97, No. 11, 2002  tibody also detected seropositivity in six individuals inwhom no further evidence of gluten sensitivity was demon-strated. As it was expected and in agreement with the classicalgorithm, the discrimination of new patients by the newapproach was much improved by performing IgA and IgGEmA as a second system and more specific mean to betterselect candidates for intestinal biopsy. Although we still didnot perform intestinal biopsy on all those subjects withpositive antitTG tests but negative EmA, current evidenceappears to suggest that the addition of EmA to the seropos-itive antitTG patients might have a key role in the simplifiedscreening avoiding unnecessary biopsies.The comparative assessment of both algorithms showsimpressive differences in favor of the proposed approachand puts a question mark on the reliability of former esti-mations of the prevalence of CD in different studies (3, 9,14–16). Our study adds substantial support to a similarexperience recently reported on the comparative value of serological screening in a population-based study (with alow prevalence of the disease) where sensitivity, specificity,and predictive values should be different (17). Both studiesagree to suggest a sequential testing approach with antitTGfollowed by EmA tests and biopsy as the most appropriatescreening algorithm for the general adult population.Interestingly, EmA, a highly sensitive and 100% specificmarker of CD, has been suggested as a potential candidateto be the only serological test for screening of CD in thegeneral population. However, its value was not yet accom-plished by epidemiological studies, and several reasons havebeen argued such as the cost of determinations, the limitedavailability of monkey esophagus substrates, etc. Despite itsabsolute and positive predictive value, what remains prob-lematic is the relatively moderate sensitivity specificity(presence of false-negative cases), as it was recently shownin patients with mild enteropathy or patchy mucosal atrophy(18). However, from a theoretical point of view and basedon the results of the present study, a screening using IgAEmA and total serum IgA findings would cost U.S. $2904per newly diagnosed patient. In our opinion, the combina-tion of highly sensitive tests at the first step and a highlyspecific one at the second step appears to be a more reliablescreening algorithm. Furthermore, this algorithm appears tobe able to reduce the potential underestimation of preva-lence by more specific markers. However, we believe thatthe studied algorithm proposed here still requires evaluationto screen for CD among children younger than 2 yr of agewhere the utility of IgA serological markers appears lower(19).Another important aspect of screening procedures thathas not been addressed previously is related to the costanalysis which, in this study, was evaluated for both algo-rithms. This assessment demonstrated a cost/benefit profilefavorable to the proposed algorithm using antitTG at the firstlevel. This srcinal observation clearly shows that advan-tages of the study algorithm are related to the fact that thecombination of an increased sensitivity of first-level testsand the absolute specificity of EmA allowed detection of anincreased number of new patients.In conclusion, the determination of CD-related serologyis a useful part of the screening for CD. Our present studyperformed in a community-based population suggests that ascreening algorithm based on the combination of the highlysensitive antitTG antibodies as first level and the specificEmA is more appropriate than the commonly used three-level screening because of cost-effective reasons. Firstly, ahigher prevalence of cases was observed and, secondly, thecost of screening for each new diagnosed case was 42% lessthan that of the classic screening.  ACKNOWLEDGMENTS This study was funded by the Comisio´n de InvestigacionesCientı´ficas of Buenos Aires Province, Argentina. The au-thors thank Dr. Christina Surawicz from the University of Washington for her generous help in preparing the manu-script. Reprint requests and correspondence:  Julio C. Bai, M.D., Gas-troenterology Hospital, Av. Caseros 2061 (1264), Buenos Aires,Argentina.  Received Sep. 24, 2001; accepted Apr. 11, 2002. REFERENCES 1. Catassi C. Screening of coeliac disease. In: Maki M, Collin P,Visakorpi JK, eds. Coeliac disease. Proceedings of the 7thInternational Symposium on coeliac disease. Tampere: CeliacDisease Study Group, 1997:23–33.2. Logan RFA. Screening for coeliac disease: Has the timecome for mass screening? Acta Paediatr 1996;85(suppl412):15–9.3. Gomez JC, Selvaggio G, Viola M, et al. Prevalence of celiacdisease in Argentina: Screening of an adult population in theLa Plata area. Am J Gastroenterol 2001;96:2700–4.4. Collin P, Kaukinen K. Clinical news: The occurrence andtreatment of coeliac disease. In: Auricchio S, Greco L, MaiuriL, Troncone R, eds. Coeliac disease. Proceedings of the EightInternational Symposium on coeliac disease. Naples: JCGEditions, 2000:119–23.5. Ascher H. The role of quantity and quality of gluten-contain-ing cereals in the epidemiology of coeliac disease. In: Maki M,Collin P, Visakorpi JK, eds. Coeliac disease. Proceedings of the 7th International Symposium on coeliac disease. Tampere:Celiac Disease Study Group, 1997:15–22.6. Fasano A, Catassi C. Current approaches to diagnosis andtreatment of celiac disease: An evolving spectrum. Gastroen-terology 2001;120:636–51.7. Ciclitira PJ. AGA technical review on celiac sprue. Gastroen-terology 2001;120:1526–40.8. Stern M, Adenauer M, Keilmann C, et al. Standardization of screening tests. In: Maki M, Collin P, Visakorpi JK, eds.Coeliac disease. Proceedings of the 7th International Sympo-sium on coeliac disease. Tampere: Celiac Disease StudyGroup, 1997:35–46.9. Catassi C, Ratsch IM, Fabiani E, et al. Celiac disease in theyear 2000: Exploring the iceberg. Lancet 1994;343:200–3. 2789 AJG – November, 2002  Screening Algorithm for Celiac Disease
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks