Definition of reference limits for autoantibodies to thyroid peroxidase and thyroglobulin in a large population of outpatients using an indirect method based on current data.
Abstract: Context.--The reference limits for thyroid antibodies are generally made by measuring thyroid peroxidase and thyroglobulin antibody values in a group of healthy subjects (direct method), as proposed by the National Academy of Clinical Biochemistry.

Objective.--To define the upper reference limits of thyroid peroxidase and thyroglobulin, by using an indirect method to analyze data from a large number of outpatients that were stored in the information system of a general hospital laboratory.

Design.--Thyroid peroxidase and thyroglobulin values from 21 492 patients, who had undergone antithyroid antibody measurements, were retrieved from the laboratory information system; the upper reference limits (in the top 97.5 percentile) were calculated using the indirect Kairisto method, after exclusion of outliers.

Results.--The mean upper reference limits for females and males were 15 kIU/L and 9 kIU/L for thyroid peroxidase, and 21 kIU/L and 19 kIU/L for thyroglobulin, respectively. The upper limits showed minimal or no differences in the different age classes in either females or males.

Conclusions.--Using a vast population of patients, we demonstrated that the upper limits for thyroid antibodies are much lower than the values obtained with classic, direct methods and that they do not vary in relation to age and sex.
Subject: Peroxidase (Methods)
Autoimmunity (Methods)
Medical laboratories (Methods)
Autoantibodies (Methods)
Authors: Tozzoli, Renato
Giavarina, Davide
Villalta, Danilo
Soffiati, Giuliano
Bizzaro, Nicola
Pub Date: 12/01/2008
Publication: Name: Archives of Pathology & Laboratory Medicine Publisher: College of American Pathologists Audience: Academic; Professional Format: Magazine/Journal Subject: Health Copyright: COPYRIGHT 2008 College of American Pathologists ISSN: 1543-2165
Issue: Date: Dec, 2008 Source Volume: 132 Source Issue: 12
Product: Product Code: 1542240 Hospital-Health Facilities; 8000320 Health Construction; 9105240 Health Planning & Construction; 7397030 Biomedical Testing Labs NAICS Code: 23332 Commercial and Institutional Building Construction; 62 Health Care and Social Assistance; 92312 Administration of Public Health Programs; 54138 Testing Laboratories SIC Code: 1542 Nonresidential construction, not elsewhere classified; 8071 Medical laboratories; 8731 Commercial physical research
Accession Number: 230246891
Full Text: Autoantibodies against thyroperoxidase (TPOAb) and thyroglobulin (TgAb) are diagnostic hallmarks and early indicators of autoimmune thyroid diseases (AITD), with an important predictive role in healthy subjects, in pregnant women, and in high-risk patients. (1-4)

The estimation of reference limits for thyroid antibodies is still a matter of debate because of the difficulty associated with correctly defining the reference population. The 2003 proposal of the National Academy of Clinical Biochemistry (NACB) (5) recommends the use of a direct method and a reference group composed of a population of men younger than 30 years, biochemically euthyroid (ie, with serum thyrotropin stimulating hormone [TSH], concentrations between 0.5 mU/L and 2.0 mU/L), and without risk parameters (eg, goiter, family history of AITD, or other autoimmune diseases). This recommendation has been disputed on the basis that the prevalence of AITD is 5 to 10 times higher in women than in men and, therefore, a reference value for the euthyroid female population should also be included. However, the higher frequency of AITD in women raises the likelihood of including sub-clinical conditions or early stage illnesses in the group of reference subjects. Recently, several prospective studies (4,6) have demonstrated that many autoantibodies can be detected in the serum of asymptomatic or paucisymptomatic individuals, who later develop an overt autoimmune disease.

When, as in this case, a priori problems of distribution exist in the reference population, indirect systems for estimating the reference interval can be useful. This approach was proposed almost 50 years ago, (7,8) but only recently has the availability of vast amounts of digitized data in laboratory information systems made it really practicable for clinical laboratories to use indirect methods. In the indirect strategy, the reference subjects are not selected individually; on the contrary, the reference intervals for the health-related subgroup are derived from the total distribution by statistical means. Either to find practical solutions for establishing reference intervals or to make observations in a more objective way, methods based on all the results that a laboratory produces for a given analysis have been proposed. These methods are based on the assumption that most results produced by the laboratory are from subjects who are not ill and include the Bhattacharya procedure (9) and the Martin (10) and Kairisto methods. (11,12) The last method hypothesizes that the mode of the frequency distribution is the same for all the data as it is for the reference population. This method, in which each side of the frequency distribution is taken as half of a Gaussian distribution, with the same mode and the same mode frequency, but with different standard deviations, offers the advantage of being included in a statistical program for Windows.

Using current data and following the Kairisto model, the aims of this study were (1) to calculate the upper reference limits (URLs) of TgAb and TPOAb using current data related to the individuals who had thyroid antibodies measured and stored in the laboratory information system of an Italian hospital laboratory, and (2) to compare the results with those obtained using the method proposed by the NACB.



Approximately 120 000 records related to 21 492 subjects (without repetitions of tests in the same subject), collected during a 5-year period (2001 through 2005) in the laboratory information system of the Clinical Chemistry and Hematology Laboratory of the San Bortolo Hospital (Vicenza, Italy), were retrieved to constitute the original database for this study. The subjects lived in an area of borderline-sufficient iodine intake. Only outpatients with a request for TSH and TPOAb and/or TgAb were selected. Most patients submit to these thyroid function tests because they are suspected of having thyroid diseases, either autoimmune (Hashimoto thyroiditis, Graves disease) or nonautoimmune (goiter, thyroid nodules). The sample group included 30 160 females and 7118 males. Table 1 summarizes age class and sex of the index population.

In addition, to compare the data, we also studied other selected groups of healthy individuals using direct methods: (1) 120 healthy subjects living in the same area (mean age, 44.5 years; range, 12-88 years; ratio of females to males, 9:1), according to International Federation of Clinical Chemistry (IFCC) criteria (13); (2) the same group of healthy subjects, but excluding those with TSH >3 mU/L or ultrasound hypoechogenicity (10/120; 83%); and (3) 100 young males (mean age, 22.5 years; range, 12-30 years), living in the same area, according to NACB criteria. (5)

Assay Methods

A TSH assay was performed on plasma samples obtained from blood collected in Vacutainer PST II tubes (BD Diagnostics, Franklin Lakes, NJ), centrifuged within 2 hours of collection, and assayed within 4 hours. Thyroid-stimulating hormone measurement was performed on the ADVIA Centaur immunoassay system (Siemens Diagnostics Solutions, Tarrytown, NY), using reagents and calibrators from the same manufacturer (TSH-3 assay).

Thyroid peroxidase and TgAb were performed on serum samples obtained from blood collected in BD Vacutainer Serum Plus tubes, clotted, centrifuged within 2 hours of collection, and assayed within 4 hours. Thyroid peroxidase and TgAb measurements were performed using an automated instrument (LIAISON, Diasorin, Stillwater, Minn) and reagents and calibrators from the same manufacturer.

During the analysis of the 5 years, the measurements were made using the same reagents and the same instruments for all patients. Detection limits for TPOAb and TgAb, according to the manufacturer, were 1.0 kIU/L and 5 kIU/L, respectively. The intra-assay imprecision was 3.6% to 7.4% for TPOAb and 4.4% to 8.6% for TgAb; the between-assay imprecision was 7.9% to 12.7% for TPOAb and 6.4% to 12.1% for TgAb, respectively.

Statistical Analysis

An URL (97.5th percentile) for TPOAb and TgAb was calculated using the Kairisto indirect method, (11) after the exclusion of outliers with the iterative [+ or -] 4 SD method.

To allow indirect methods to be applied, it is necessary that the health-related subdistribution be a major part of the total distribution, that the total distribution be unimodal (but it can be skewed to either direction), and that the modes of the total distribution and the health-related subdistribution be the same or quite close to each other.

We adopted the Kairisto procedure to correct and standardize the huge amount of data that can be downloaded from the laboratory information system. The main modifications are splitting the distribution and forcing the mode (rather than the mean) of the hypothesized health-related distribution to be the same as the mode in the original distribution. In this procedure, each side of the health-related distribution is estimated separately, so that the resulting health-related distribution consists of 2 halves of 2 different Gaussian distributions, with the same mode and mode frequency, but with different standard deviations.

Because of the positively skewed distribution and mode equal to 0.0 mU/L, we considered a reciprocal for every datum to obtain a symmetric distribution (eg, for 5.6 mU/L, a value of -5.6 mU/L was added), which allowed the statistical program (GraphROC for Windows, version 2.0) (14) to perform the analysis. This software applies the indirect strategy for deriving reference intervals. It is assumed that most healthy patients, with respect to the particular test and URLs, are derived from the total distribution by statistical means. The program accurately estimates the mode of the unselected data and the underlying health-related distribution by 2 split Gaussian distributions, giving extra weight to the mode, and calculates the reference limits from these distributions by a nonparametric method.

The URL was calculated for both TPOAb and TgAb, in 2 groups of subjects: the whole population (group A) and the population with TSH values within the reference interval (0.1-4.0 mU/L; group B).


Indirect Method

For TPOAb, in the entire population (group A), 21 492 results (with 1859 outliers, 9.0%) were considered, and the mean URL was 15 kIU/L for women and 9 kIU/L for men. Considering only the results obtained from subjects with TSH values within reference limits (group B, 14036 results; 558 outliers, 4.0%), the URL for TPOAb ranged from 7 to 10 kIU/L in the different age classes in females and from 6 to 10 kIU/L in males (Table 2).

For TgAb, when we considered only the subjects that were selected by TSH within 0.1 and 4.0 mU/L (group B, 16046 results; 1619 outliers, 10.1%), the URL for TgAb ranged from 17 to 22 kIU/L for the different age classes in females and showed no differences by age in males (1718 kIU/L). The limits in the group with no selection (group A, 20 661 results, with 1934 outliers, 9.3%) were roughly similar (Table 3).

Direct Methods

In the 3 groups of subjects analyzed, the URLs for TPOAb were 102.5 kIU/L in group 1, 56.0 kIU/L in group 2, and 18 kIU/L in group 3; and for TgAb, they were 116 kIU/L, 62 kIU/L, and 24.5 kIU/L, respectively (Table 4).


Variability of Reference Limits for Thyroid Antibodies

The normal reference limits for TPOAb and TgAb are highly variable and often arbitrarily established. (15,16) The high variability is related to 2 main factors: the type of assay method and the selection of the representative population.

Earlier studies showed that reference limits clearly differed between females and males. However, these results were obtained with the use of first-generation, insensitive methods (eg, indirect immunofluorescence, passive hemoagglutination), now replaced by more sensitive, immunometric assays (eg, enzyme-linked immunosorbent assay, immunoradiometric assay, immunoluminescent assay, multiplex bead array). (4,15,17,18) Unfortunately, like the previous methods, the new ones also demonstrated an unexpectedly high variability and a high degree of dispersion in cut-off levels. (17) This condition largely affects the estimation of reference intervals.

The choice of the reference sample group is another factor of variability, depending on different procedures of selection for sex and age and on rejection criteria. The groups used include healthy subjects without selection or rejection, healthy subjects with selection by sex and age, and healthy subjects but rejecting patients with either self-reported thyroid diseases or with biochemical/subclinical thyroid diseases (as determined by either ultrasound hypoechogenicity or by a TSH value greater than the reference limits).

Estimation of Reference Intervals for TPOAb and TgAb

In the direct strategy, the reference values are often derived without considering a basic question: are the subjects in the reference sample group really representative of the reference population? A good characterization of the reference sample group is often difficult. The selection of reference individuals is based on the clinical diagnosis; the diseases and conditions known to be associated with a change in the level of a particular analyte are listed, and patients corresponding to those diagnostic categories are excluded. The remaining subjects form the health-related reference-sample group. However, a critical factor is the reliability and accuracy of the information recorded in the patient database, especially when subclinical conditions may be ignored. Indeed, sometimes it is impossible to define the reference distributions according to the IFCC recommendations. (13)

In the indirect strategy, the reference intervals for health-related subgroups are derived from the total distribution by statistical means. The rationale for the indirect method is based on the observation that the disease-associated laboratory results are more frequent in the tails of the distribution of unselected patient data. We adopted the procedure proposed by Kairisto et al, (11) also called the mode-method. (19) In this procedure, derived from principles described by Pryce (7) and Hoffmann (8) and available in the software GraphROC for Windows, (13) each side of the health-related distribution was estimated separately, so that the resulting health-related distribution consisted of 2 halves of 2 Gaussian distributions, with the same mode and mode frequency but with different standard deviations.

In this study, we used an indirect method for the definition of reference intervals based not on a control group but on a population of a considerably large number, as recently adopted for other analytes (ie, thyroid hormones, enzymes, and substrates). (20-23)

The estimation of reference limits from patients' results has many advantages, including being easy and cheap. Reference limits are derived from values obtained with the same premetrologic conditions as the patients' data, and patients' results can be easily selected, when it is necessary, according to biological-variation factors. It has been reported (24) that the indirect reference limits made using software based on the Kairisto procedure and those made using the 2 programs based on Bhattacharya and Martin procedures are similar.

The determination of TgAb and TPOAb is the most frequent test performed in the autoimmunology laboratory, and a vast amount of data (often more than 1000 records) can be reached in any laboratory in only a few years, allowing the safe application of indirect methods. (25)

In this study, which is the first evaluation, to our knowledge, of reference values for autoantibodies using a large number of subjects, we found that the URLs of TPOAb and TgAb are very low, compared with those reported in the literature, those in the manufacturers' package inserts, or those obtained from the direct method according to the IFCC procedure.

In a previous report, (16) where we compared several immunometric methods for thyroid antibodies detection, we found much higher URLs (as reported by the manufacturers' package inserts and as obtained with various direct methods, without rejections or exclusions of patients) and an elevated degree of dispersion, with a 4-fold to 5-fold range between the minimum and maximum limits, that is, from 20 to 100 kIU/L for TPOAb and from 80 to 325 kIU/L for TgAb. In the present study, the limits in the entire population ranged, instead, from 10 to 16 kIU/L for TPOAb and from 17 to 22 kIU/L for TgAb. These data are similar to those obtained very recently by La'ulu et al, (26) using a direct method and an automated immunoluminescent assay.

Using selection criteria and a progressive elimination of outlier results (ie, high levels of autoantibodies), which correspond to asymptomatic patients in the preclinical or initial phase of AITD, URLs can be obtained. This approach reduces the values of the upper 97.5 percentile reference limits. Analyzing data obtained with the same method in selected groups of healthy patients (Table 4), we observed that the URLs decreased from 102.5 to 18.5 kIU/L for TPOAb and from 116 to 24.5 kIU/L for TgAb. Similar data were obtained recently by Jensen et al, (21) in a large group of patients with a different TPOAb and TgAb assay method, in which the URLs decreased from 284 to 9.8 kIU/L for TPOAb and from 84 to 19 kIU/L for TgAb, respectively.

The results of the present study show that the URLs obtained with the NACB direct method and the Kairisto indirect method are similar (18 kIU/L vs 15 kIU/L for TPOAb and 24.5 vs 22 kIU/L for TgAb). Consequently, it can be affirmed that the 2 methods are alternative to each other and equally applicable for defining the reference interval. Our opinion, however, is that the direct method, which is based on the data collected from 120 young men, does not completely eliminate the risk of including healthy, biochemically euthyroid patients who have high levels of antithyroid antibodies (>100 kIU/L). If only 3 of 100 subjects have these high autoantibody levels, the upper reference value will be elevated on the order of 2 or 3 times. Compared with direct methods, an indirect method allows a more accurate exclusion of patients with outlier values and a more exact estimation of the true value of the analytes in the population.

We propose that the direct NACB method on 120 young men should be the initial procedure for defining the URLs of thyroid antibodies and that those values be confirmed later, or corrected, if needed, using indirect methods applied to the current data of a few thousand records, which each laboratory can obtain in a short time. As can be deduced from Tables 2 and 3, selection of overall data for referred laboratory patients is not necessary because the results obtained for the population, without rejection, are substantially similar to those obtained with rejection. Patient selection according to normal TSH values does not cause significant variation in the URLs.

In addition, our data highlight that a considerable number of healthy subjects present with low, but detectable, amounts of thyroid antibodies, suggesting that thyroperoxidase and thyroglobulin are antigens for naturally occurring autoantibodies. (27,28) These data agree with the results of early reports that were obtained in both small and large populations. (29,30)

Sex-Dependent Differences in TPOAb and TgAb Reference Limits

Analyzing the data in relationship to sex, we found that the URLs were 15 kIU/L in females and 9 kIU/L in males for TPOAb and 21 kIU/L in females and 19 kIU/L in males for TgAb, suggesting that the gender differences in thyroid antibodies levels are minimal and not statistically significant. The indirect method, eliminating the high levels of autoantibodies most frequently present in asymptomatic females, as demonstrated by several reports1, (2,31-34) on unselected or selected general populations, produces a substantial uniformity of results between sexes. (16,22) Therefore, we conclude that it is not necessary to distinguish between males and females in the reference limits of TPOAb and TgAb. This consideration confirms the validity of the NACB proposal, which uses males as controls to define reference intervals that are valid for females as well.

Age-Dependent Differences in TPOAb and TgAb Reference Limits

The URL in relation to age (Tables 2 and 3) does not show substantial differences among the various age groups, so it does not seem useful to differentiate reference limits on the basis of this parameter. The stability of the limits in relation to age shown in our study is remarkable because it has long been recognized that the percentage of subjects with antithyroid antibodies increases with age, as demonstrated in large population studies. (1,2,34-37)

Our data are similar to those obtained by Eskelinen et al, (19) who evaluated 126 subjects older than 65 years, by Jensen et al, (21) who studied 1441 subjects of various ages, and by Mariotti et al, (35) who studied centenarians. The latter authors (35) showed that thyroid autoantibodies are rare in healthy centenarians and in subjects older than 90 years, whereas they are frequently observed in unselected or hospitalized elderly patients. Taken together, these data suggest that thyroid autoimmune phenomena are not the consequence of the aging process itself, but rather, might be related to age-associated disease, confirming the validity of the Kairisto model in the URL definition.

The Reference Limits of Thyroid Antibodies Affect the Diagnostic Accuracy of the Test

Note that the use of low URL values, such as those obtained in this study, has an important effect on the diagnostic reliability of autoantibody tests in AITD cases. In fact, they cause a significant elevation in the diagnostic sensitivity (which, in the case of TPOAb, reaches nearly 100%), to the detriment of the nosographic specificity. Low autoantibody values that are still higher than the URL are frequently detectable in patients who are either apparently healthy or affected by diseases other than AITD. These levels are the sign of unrecognized focal lymphocytic thyroiditis, and their presence, in some case, can be predictive of the future development of AITD. (3-5)

In conclusion, using the indirect method on current data from a vast population of patients directed to a hospital laboratory service, we demonstrated that the URLs of TPOAb and TgAb are much less than the values classically detected by applying direct methods on samples of healthy subjects and that these values do not vary in relation to age and sex. Such observations confirm the validity of the direct NACB/IFCC method for defining the upper limit. This method could be adopted by every clinical laboratory as the initial approach, and the results obtained could be confirmed or modified in subsequent testing using an indirect method applied on current data from patients.


(1.) Vanderpump MP, Turnbridge WM, French JM, et al. The incidence of thyroid disorders in the community: a twenty-year follow-up of the Whickam Survey. Clin Endocrinol (Oxf). 1995;43:55-68.

(2.) Hollowell J, Staheling NW, Flanders WD, et al. Serum TSH, T(4), and thyroid antibodies in the United States population (1988 to 1994): National Health and Nutrition Examination Survey (NHANES III). J Clin Endocrinol Metab. 2002;87: 489-499.

(3.) Premawardhana LDKE, Parkes AB, John R, Harris B, Lazarus JH. Thyroid peroxidase antibodies in early pregnancy: utility for prediction of postpartum thyroid dysfunction and implications for screening. Thyroid. 2004;8:610-615.

(4.) Prummel MF, Wiersinga WM. Thyroid peroxidase autoantibodies in euthyroid subjects. Best Pract Res Clin Endocrinol Metab. 2005;19:1-15.

(5.) Baloch Z, Carayon P, Conte-Devolx B, et al. Laboratory medicine practice guidelines: laboratory support for the diagnosis and monitoring of thyroid disease. Thyroid. 2003;13:3-126.

(6.) Bizzaro N, Tozzoli R, Shoenfeld Y. Are we at a stage to predict autoimmune rheumatic diseases? Arthritis Rheum. 2007;56:1736-1744.

(7.) Pryce SD. Level of haemoglobin in whole blood and red blood cells, and proposed convention for defining normality. Lancet. 1960;2:333-236.

(8.) Hoffmann RG. Statistics in the practice of medicine. JAMA. 1963;185:964973.

(9.) Bhattacharya CG. A simple method of resolution of a distribution into Gaussian components. Biometrics. 1967;23:115-135.

(10.) Martin HF, Hologgitas JV, Driscoll J, Fanger H, Gudzinowicz BJ, Barnett R. Reference values based on population accessible to hospitals. In: Grasbeck R, Alstrom T, eds. Reference Values in Laboratory Medicine. Chichester, England: John Wiley & Sons Ltd; 1981:233-262.

(11.) Kouri T, Kairisto V, Virtanen A, et al. Reference intervals developed from data for hospitalized patients: computerized method based on combination of laboratory and diagnostic data. Clin Chem. 1994;40:2209-2215.

(12.) Kairisto V, Hanninen KP, Leino A, et al. Generation of reference values for cardiac enzymes from hospital admission laboratory data. Eur J Clin Chem Clin Biochem. 1994;32:789-796.

(13.) International Federation of Clinical Chemistry. Approved recommendation (1987) on the theory of reference values, part 5: statistical treatment of collected reference values--determination of reference limits. J Clin Chem Clin Biochem. 1987;25:645-656.

(14.) Kairisto V, Poola A. Software for illustrative presentation of basic clinical characteristics of laboratory tests--GraphROC for Windows. Scand J Clin Lab Invest. 1995;222(suppl):43-60.

(15.) Feldt-Rasmussen U. Analytical and clinical performance goals to testing autoantibodies to thyroperoxidase, thyroglobulin, and thyrotropin receptor. Clin Chem. 1996;42:160-163.

(16.) Tozzoli R, Villalta D, Bizzaro N, Tonutti E, Manoni F. Laboratory diagnosis of autoimmune thyroid disease. Recenti Prog Med. 2001;92:609-617.

(17.) Tozzoli R, Bizzaro N, Tonutti E, et al. Immunoassay of anti-thyroid auto-antibodies: high analytical variability in second generation methods. Clin Chem Lab Med. 2002;40:568-573.

(18.) Tozzoli R, Villalta D, Kodermaz G, Bagnasco M, Tonutti E, Bizzaro N. Autoantibody profiling of patients with autoimmune thyroid disease using a new multiplexed immunoassay method. Clin Chem Lab Med. 2006;44:837-842.

(19.) Eskelinen S, Suominen P, Vahlberg T, et al. The effect of thyroid antibody positivity on reference intervals for thyroid stimulating hormone (TSH) and free thyroxine ([FT.sub.4]) in an aged population. Clin Chem Lab Med. 2005;43:1380-1385.

(20.) Jensen E, Petersen PH, Blaabjerg O, etal. Establishmentofa serum thyroid stimulating hormone (TSH) reference interval in healthy adults: the importance of environmental factors, including thyroid antibodies. Clin Chem Lab Med. 2004;42:824-832.

(21.) Jensen EA, Petersen PH, Blaabjerg O, Hansen PS, Brix TH, Hegedus L. Establishment of reference distribution and decision values for thyroid antibodies against thyroid peroxidase (TPOAb), thyroglobulin (TgAb) and the thyrotropin receptor (TRAb). Clin Chem Lab Med. 2006;44:991-998.

(22.) Grossi E, Colombo R, Cavuto S, Franzini C. The REALAB Project: a new method for the formulation of reference intervals based on current data. Clin Chem. 2005;51:1232-1240.

(23.) Ilcol YO, Aslan D. Use of total patient data for indirect estimation of reference intervals for 40 clinical chemical analytes in Turkey. Clin Chem Lab Med. 2006;44:867-876.

(24.) Ferre-Masferrer M, Fuentes-Arderiu X, Puchal-Ane R. Indirect reference limits estimated from patients' results by three mathematical procedures. Clin Chim Acta. 1999;279:97-105.

(25.) Oosterhuis WP, Modderman TA, Pronk C. Reference values: Bhattacharya or the method proposed by the IFCC? Ann Clin Biochem. 1990;27:359-365.

(26.) La'ulu SL, Slev PR, Roberts WL. Performance characteristics of 5 automated thyroglobulin autoantibody and thyroid peroxidase autoantibody assays. Clin Chim Acta. 2007;376:88-95.

(27.) O'Leary PC, Feddema PH, Michelangeli VP, et al. Investigation of thyroid hormones and antibodies based on a community health survey: the Busselton thyroid study. Clin Endocrinol (Oxf). 2006;64:97-104.

(28.) Lacroix-Desmazes S, Kaveri SV, Mouthon L, et al. Self-reactive antibodies (natural antibodies) in healthy individuals. J Immunol Methods. 1998;216:117 137.

(29.) Zophel K, Saller B, Wundrlich G, et al. Autoantibodies to thyroperoxidase (TPOAb) in a large population of euthyroid subjects: implication for the definition of TPOAb reference intervals. Clin Lab. 2003;49:591-600.

(30.) Gilmour J, Brownlee Y, Foster P, et al. The quantitative measurement of autoantibodies to thyroglobulin and thyroid peroxidase by automated microparticle based immunoassays in Hashimoto's disease, Graves' disease and a follow-up study on postpartum thyroid disease. Clin Lab. 2000;46:57-61.

(31.) Groves CJ, Howells RD, Williams S, Darke C, Parkes AB. Primary standardization for the ELISA of serum thyroperoxidase and thyroglobulin antibodies and their prevalence in a normal Welsh population. J Clin Lab Immunol. 1990; 32:147-151.

(32.) Bjoro T, Holmen J, Kruger O, et al. Prevalence of thyroid disease, thyroid dysfunction and thyroid peroxidase antibodies in a large, unselected population: the Health Study of Nord-Trondelag (HUNT). Eur J Endocrinol. 2000;143:639-647.

(33.) Pedersen IB, Knudsen N, Jorgensen T, Perrild H, Ovesen L, Laurberg P. Thyroid peroxidase and thyroglobulin autoantibodies in a large survey of population with mild and moderate iodine deficiency. Clin Endocrinol (Oxf). 2003;58: 36-42.

(34.) Hoogendoorn EH, Hermus AR, deVegtF, etal. Thyroid function and prevalence of anti-thyroperoxidase antibodies in a population with sufficient iodine intake: influences of age and sex. Clin Chem. 2006;52:104-111.

(35.) Mariotti S, Sansoni P, Barbesino G, et al. Thyroid and other organ-specific autoantibodies in healthy centenarians. Lancet. 1992;339:1506-1508.

(36.) Konno N, Yuri K, Taguchi H, et al. Screening for thyroid diseases in an iodine sufficient area with sensitive thyrotropin assays, and serum thyroid auto-antibody and urinary iodide determinations. Clin Endocrinol (Oxf). 1993;38:273 281.

(37.) Szabolcs I, Podoba J, Feldkamp J, et al. Comparative screening for thyroid disorders in old age in areas of iodine deficiency, long-term iodine prophylaxis and abundant iodine intake. Clin Endocrinol (Oxf). 1997;47:87-92.

Renato Tozzoli, MD; Davide Giavarina, MD; Danilo Villalta, MD; Giuliano Soffiati, MD; Nicola Bizzaro, MD

Accepted for publication May 8, 2008.

From the Laboratory of Clinical Pathology, Ospedale Civile, Latisana Italy (Dr Tozzoli); the Department of Clinical Pathology, Ospedale San Bortolo, Vicenza, Italy (Drs Giavarina and Soffiati); the Allergology and Clinical Immunology, Azienda Ospedaliera Santa Maria degli Angeli, Pordenone, Italy (Dr Villalta); and the Laboratory of Clinical Pathology, Ospedale Civile,Tolmezzo, Italy (Dr Bizzaro).

The authors have no relevant financial interest in the products or companies described in this article.

Reprints: Nicola Bizzaro, MD, Laboratorio di Patologia Clinica, Ospedale Civile, Via Morgagni 18, 33028 Tolmezzo (UD), Italy (e-mail:
Table 1. Descriptive Data (Age and Sex) of the
Studied Population

                                     Age Classes, y

                 <15    15-30    31-45    46-60      >60     Total

Females, No.     703     3921     9166     8313     8057     30160
Males, No.       513      977     1743     1953     1932      7118

Total           1216     4898    10 909   10 266    9989    37 278

Table 2. Upper Reference Limit (URL) for Antithyroperoxidase
Antibodies *

                   Group A,               Group B,
Age, y/Sex     No. (URL, kIU/L)       No. (URL, kIU/L)

0-98/both        21 492 (12)             14 036 (7)
0-98/F           17 002 (15)             10 932 (7)
<15/F               490 (10)                309 (7)
15-30/F            2381 (10)              1 593 (9)
31-45/F            4969 (16)               3331 (7)
46-60/F            4606 (23)               3045 (9)
>60/F              4556 (19)               2654 (10)
0-98/M              4490 (9)               3104 (6)
<15/M               341 (10)                204 (6)
15-30/M             653 (10)                506 (10)
31-45/M            1071 (11)                798 (9)
46-60/M             1179 (9)                874 (8)
>60/M               1246 (9)                722 (9)

* Group A indicates the whole population; group B, the population
with thyrotropin values between 0.1 and 4.0 mU/L.

Table 3. Upper Reference Limit (URL) for
Antithyroglobulin Antibodies *

                   Group A,               Group B,
Age, y/Sex     No. (URL, kIU/L)       No. (URL, kIU/L)

0-98/both            20 661 (20)            16 046 (22)
0-98/F               16 310 (21)            12 638 (21)
< 15/F                  464 (23)               332 (22)
15-30/F                2318 (19)              1851 (22)
31-45/F                4797 (21)              3892 (18)
46-60/F                4445 (21)              3503 (17)
> 60/F                 4286 (19)              3060 (18)
0-98/M                 4351 (19)              3408 (17)
< 15/M                  332 (19)               237 (18)
15-30/M                 636 (19)               552 (18)
31-45/M                1025 (20)               863 (18)
46-60/M                1157 (19)               958 (17)
> 60/M                 1201 (18)               798 (18)

* Group A indicates the whole population; group B, the population
with thyrotropin values between 0.1 and 4.0 mU/L.

Table 4. Upper Reference Limit (URL) of
Antithyroperoxidase Antibodies (TPOAb) and
Antithyroglobulin Antibodies (TgAb) by Different
Estimation Methods

                                        URLs, kIU/L

Group *              No.      Sex      TPOAb    TgAb

1. IFCC criteria/
no rejection         120    F/M: 9/1   102.5   116.0

2. IFCC criteria/
with rejection       110    F/M: 9/1   56.0     62.0

3. NACB criteria     100       M       18.5     24.5

* IFCC indicates International Federation of Clinical Chemistry;
NACB, National Academy of Clinical Biochemistry.
Gale Copyright: Copyright 2008 Gale, Cengage Learning. All rights reserved.