| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
The Juvenile Diabetes Research Foundation Center for Type 1 Diabetes Prevention in Finland (T.K., P.K., K.S., A.K., S.K., T.S., J.I., O.S., M.K.), Department of Pediatrics, Medical School, University of Tampere (T.K., M.K.), and Tampere University Hospital, FIN-33014 Tampere; Department of Pediatrics (P.K., K.S., S.K.), University of Oulu, FIN-90014 Oulu; Departments of Pediatrics (A.K., T.S., O.S.) and Virology (J.I.), University of Turku, FIN-20520 Turku; and Hospital for Children and Adolescents (M.K.), University of Helsinki, FIN-00029 Helsinki, Finland
Address all correspondence and requests for reprints to: Mikael Knip, M.D., Hospital for Children and Adolescents, University of Helsinki, P.O. Box 281, FIN-00029 HUCH, Finland. E-mail: mikael.knip{at}hus.fi.
Abstract
The aim of this study was to evaluate the frequency and predictive value of diabetes-associated autoantibodies, such as islet cell antibodies (ICA) and autoantibodies to insulin (IAA), GAD65 (GADA), and the IA-2 molecule (IA-2A) in genetically susceptible children from the general population during the first 2 yr of life. Of 12,170 newborn infants, 1,005 with increased genetic risk of type 1 diabetes (high risk, human leukocyte antigen DQB1*02/*0302; moderate risk, DQB1*0302/x, where x = other than *02, *0301, or *0602) were monitored for ICA, IAA, GADA, and IA-2A at 3- to 6-month intervals from birth up to a minimum age of 2 yr. In addition, all 15 genetically susceptible children from the general population who had participated in regular immunological follow-up and developed clinical type 1 diabetes by the end of April 2000 were analyzed for the development of autoantibodies. Among 1,005 children, 63 (6.3%) tested positive for at least one autoantibody, 31 for ICA (3.1%), 48 for IAA (4.8%), 23 for GADA (2.3%), and 13 for IA-2A (1.3%) at least once by the age of 2 yr. Both ICA and IAA identified 95% [95% confidence interval (CI), 77.2–99.9%] of those who tested persistently positive for multiple (
2) antibodies at the age of 2 yr, GADA identified 86% (CI, 65.1–97.1%), and IA-2A identified 55% (CI, 32.2–75.6%). Close to half of the antibody-positive children (29 of 63) reverted back to antibody negativity. Autoantibodies disappeared more often among those who tested positive for IAA than among those who tested positive for other autoantibodies (P 
0.021). Among the 15 children who developed type 1 diabetes, the disease sensitivity of ICA was 80% (CI, 51.9–95.7%), that of IAA was 93% (CI, 68.0–99.8%), that of GADA was 60% (CI, 32.3–83.7%), and that of IA-2A was 40% (CI, 16.3–67.7%). These results suggest that IAA are characterized by high sensitivity, early appearance, and high frequency of transient antibody positivity, whereas ICA detected with a thoroughly standardized assay appear to be more specific for the screening of ß-cell autoimmunity in young children with increased genetic susceptibility to type 1 diabetes in the Finnish population, which has the highest incidence of type 1 diabetes in the world.
IT HAS BEEN SUGGESTED that the pathogenetic process of type 1 diabetes is initiated early in life (1). Various triggers have been implicated, but the natural course of the prediabetic process has remained poorly defined. One major question has been whether ß-cell destruction is progressive and persistent once initiated. A lack of knowledge of the emergence of signs of ß-cell autoimmunity at an early age has hampered the identification of predisposing exogenous factors.
Screening strategies for type 1 diabetes aim optimally at high sensitivity to identify as early as possible all those individuals who will eventually develop clinical disease and those who are developing permanent positivity for autoantibodies and have signs of biologically significant autoimmunity and/or ongoing progressive ß-cell destruction. In addition, such strategies should have a high specificity to minimize the frequency of false-positive results and avoid unnecessary concern. Most of the ongoing secondary prevention studies use positivity for islet cell antibodies (ICA) as an inclusion criterion (2, 3). This is based on the fact that the predictive characteristics of ICA have been more closely defined than those of insulin autoantibodies (IAA), GAD65 antibodies (GADA), or IA-2 antibodies (IA-2A). The validity of ICA for primary screening has been challenged, however, because the analysis is labor-intensive and difficult to standardize. A number of studies have reported that combined testing for GADA and IA-2A has a high sensitivity and/or specificity for type 1 diabetes (4, 5, 6, 7, 8). The predictive value of that or any other antibody or antibody combination has not yet been assessed in relation to that of ICA in prospective studies from birth in the general population, where the overall risk of the disease is considerably lower than among first-degree relatives of affected patients.
In the Finnish Type 1 Diabetes Prediction and Prevention (DIPP) Study, ICA are used as the primary screening tool for ß-cell autoimmunity in children with increased genetic risk of type 1 diabetes recruited from the general population (9). This decision was based on the observation that ICA were more sensitive in our hands than GADA or IAA, with more than 84% of children with newly diagnosed type 1 diabetes testing positive (8). The prevalence of ICA has been observed to be 3–4% among Finnish schoolchildren (10, 11). Analyzing the sera that are initially positive for ICA only may result in some subjects who test positive for autoantibodies other than ICA remaining unidentified. In the German BABYDIAB study comprising offspring of parents with type 1 diabetes, about 30% of the subjects with multiple autoantibodies would have remained unidentified if ICA had been used in the primary screening (12). Experiences with risk markers used for the prediction of type 1 diabetes in affected families cannot necessarily be directly transferred to the general population, however (13). Thus, the present work was aimed at evaluating the natural history of early ß-cell autoimmunity and the optimal screening strategy in genetically susceptible children from the general population during the first years of life to identify children with a high risk of type 1 diabetes as early as possible during the prediabetic period. In our recent report in which ICA were used as a primary screening tool for ß-cell autoimmunity, 694 children were observed from birth up to the age of 2 yr, and only those who tested positive for ICA were analyzed for IAA, GADA, and IA-2A (14). To identify those children who test positive for autoantibodies other than ICA, we measured in the present study ICA, IAA, GADA, and IA-2A in all samples available in 1,005 genetically susceptible children from the general population at least up to the age of 2 yr.
Subjects and Methods
Subjects
The DIPP Study was established to assess feasible strategies for predicting type 1 diabetes in the general population and to develop effective tools for preventing or delaying progression to clinical disease (9). It aims at identifying newborn infants carrying an increased genetic risk of type 1 diabetes, recognizing the appearance of signs of ß-cell autoimmunity at an early stage, and delaying the onset of clinical disease in those at genetic risk who have signs of such autoimmunity in a double-blind randomized intervention trial in which the efficacy of daily nasal insulin administration is evaluated in the prevention of type 1 diabetes. By July 1997, written informed parental consent for genetic screening had been obtained for 12,170 newborn infants, all of whom had been analyzed for human leukocyte antigen (HLA)-DQB1 alleles. Of these, 1,596 (13.1%) carried the high-risk genotype (DQB1*02/*0302) or the moderate-risk genotype (DQB1*0302/x, where x = other than *02, *0301, or *0602). The present analysis covers the first 1,005 children (530 boys, 52.7%) carrying the high or the moderate-risk genotype and analyzed for ICA, IAA, GADA, and IA-2A in all samples at least up to the age of 2 yr, comprising altogether 28,328 autoantibody measurements. The initial blood sample was taken at the age of 3 months, the next at the age of 6 months, and subsequent samples at intervals of 3–6 months during the initial 2 yr and at an interval of 6–12 months thereafter. Inclusion in the study required that a sample had been taken at the age of 2 yr. The median observation period was 3.1 yr (range, 2.0–5.0 yr). Autoantibodies with decreasing titers in infants who had also had autoantibodies in their cord blood were excluded from the analysis, because they always disappeared by the age of 15 months at the latest, suggesting that these antibodies had been transferred transplacentally from the mother (15). A total of 254 children (25.3%) had the high-risk genotype, and 751 (74.7%) had the moderate-risk genotype. In addition, 15 children who had participated in the immunological surveillance and developed type 1 diabetes by the end of April 2000 were analyzed for the development of autoantibodies. The DIPP study protocol had been approved by the ethical committees of the three participating hospitals.
Genotyping
HLA-DQB1 alleles were analyzed as described (16). In short, a part of the second exon of the HLA-DQB1 gene was amplified using a primer pair with a biotinylated 3' primer. The biotinylated PCR products were then transferred to streptavidin-coated microtitration plates, denatured, and hybridized with sequence-specific probes labeled with lanthanide chelates: europium, terbium, or samarium. Two hybridization mixtures were used, one containing probes hybridizing with DQB1*0602 and *0603, DQB1*0603 and *0604, and a consensus sequence; and the other containing probes specific to the DQB1*02, *0301, and *0302 alleles. After appropriate incubations and washings, the specific hybridization products were detected using three-color time-resolved fluorescence of the lanthanide chelates.
Autoantibody assays
Diabetes-associated autoantibodies were analyzed in the Research Laboratory, Department of Pediatrics, University of Oulu (Oulu, Finland). Except for ICA, all of the samples from the same child were measured in the same assay to exclude the effect of interassay variation. ICA were quantified by a standard indirect immunofluorescence method on sections of frozen human pancreas from a blood group O donor (17). Altogether, 7078 serum samples were analyzed for ICA. The same pancreas was used for all analyses. The end-point dilution titer of ICA-positive samples was recorded, and the results were expressed in Juvenile Diabetes Foundation units (JDFU). The detection limit was 2.5 JDFU. All samples that were initially positive for ICA were retested to confirm positivity. The sensitivity of the ICA assay in our laboratory was 100%, and the specificity was 96% in the most recent proficiency testing round. The intra-assay coefficient of variation was less than 25%. The interassay coefficient of variation was observed to be 22.4% for samples with low titers (10 JDFU) and 26% for samples with high titers (128 JDFU; Ref. 18).
Serum levels of IAA were quantified in 7083 samples with a microassay (19) modified from that described by Williams et al. (20). The IAA levels representing the specific binding were expressed in relative units (RU) based on a standard curve run on each plate using the MultiCalc software program (PerkinElmer Life Sciences Wallac, Inc., Turku, Finland). A subject was considered to be positive for IAA when the specific binding exceeded 1.55 RU (the 99th percentile in 371 nondiabetic Finnish subjects). The disease sensitivity of our microassay was 14%, and the specificity was 100% in the 2001 Diabetes Autoantibody Standardization Program (DASP) Workshop (21). Subsequently, our IAA assay has been optimized, resulting in a disease sensitivity of 30%, the specificity remaining at 100%, based on a blinded reanalysis of the 100 DASP Workshop samples. The intra-assay coefficient of variation was 7%, and the interassay coefficient of variation was less than 9% in the IAA assay.
GADA were measured in 7083 serum samples with a radiobinding assay as described (22). The results were expressed in RU based on a standard curve constructed from a dilution of positive and negative samples. The cut-off limit for antibody positivity (5.35 RU) was set at the 99th percentile in 373 nondiabetic Finnish children and adolescents. The disease sensitivity of the GADA assay was 76%, and its specificity was 96%, based on the 2001 DASP workshop (21). The intra-assay coefficient of variation was less than 10%, and the interassay coefficient of variation was 18% at a GADA level of 14.6 RU and 12% at a GADA level exceeding 100 RU (8).
IA-2A were quantified in 7084 serum samples with a radiobinding assay as described (8). Antibody levels were expressed in RU based on a standard curve, as for GADA. The limit for IA-2A positivity was set at 0.429 RU, which represents the 99th percentile in 374 nondiabetic Finnish children and adolescents. The disease sensitivity of this assay was 58%, and the specificity was 100%, based on the 2001 DASP Workshop (21). The intra-assay coefficient of variation was less than 10%, and the interassay coefficient of variation was 12% at an IA-2A level of 0.63 RU, 10% at a level of 21.3 RU, and 8% at a level of 82.6 RU. All samples from the same individual were quantified in the same assay run for IAA, GADA, and IA-2A, respectively. Samples with IAA, GADA, or IA-2A levels between the 95th and 99.5th percentiles were reanalyzed to confirm the antibody status.
Definitions and statistical analysis
Persistent positivity was defined as antibody positivity in at least two consecutive samples, the latter being the last available sample. A fluctuating antibody pattern was defined as seroconversion to autoantibody positivity followed by autoantibody negativity and again conversion to autoantibody positivity. Inverse seroconversions included those antibody-positive children who seroconverted to negativity for all autoantibodies. Blood samples obtained during the randomized intervention trial were not taken into account in these definitions.
The distributions of autoantibodies between the two genotypes were evaluated by cross-tabulation and 
2 statistics with Yates’ correction, unless 20% of the cells had an expected value less 5, when the Fisher exact test was used. Sensitivity, specificity, and positive predictive values (PPVs) were calculated as described previously (23). 95% Confidence intervals (CI) were calculated with the exact method. When comparing age at seroconversion, the time point for seroconversion was taken to be in the middle of the interval between the last negative sample and the first positive one. The Mann-Whitney U test was used to compare the antibody levels between genotypes. Parametric analysis of variances, the paired t test, and the unpaired t test were used when comparing the age at seroconversion between various autoantibodies.
Results
Appearance of diabetes-associated autoantibodies by the age of 2 yr
Sixty-three of 1,005 children (6.3%), of whom 32 were boys, tested positive for a minimum of one autoantibody (31 for ICA, 48 for IAA, 23 for GADA, and 13 for IA-2A) at least once by the age of 2 yr. Twenty-one of these 63 index cases (33%) carried the high-risk genotype. Only five (8%) had no ICA or IAA detectable, whereas 37 children (59%) tested negative for GADA and IA-2A. The first children tested positive for IAA and GADA at the age of 3 months (Fig. 1
), after which the proportion of IAA-positive children increased sharply, IAA achieving the frequency of 2.9% (29 of 1,005) and GADA that of 1.7% (17 of 1,005) by the age of 2 yr. The first children tested positive for ICA at the age of 6 months and for IA-2A at the age of 12 months. The proportion of ICA-positive children increased conspicuously after the age of 12 months, reaching a frequency of 2.7% (27 of 1,005) and IA-2A that of 1.2% (12 of 1,005) by the age of 2 yr.
|
The proportion of children who tested positive for ICA was significantly higher among those with the high-risk genotype (13 of 254; 5.1%) than among those with the moderate-risk genotype (14 of 751; 1.9%; P = 0.016; Fig. 2A), but no significant difference was observed in the frequency of IAA at the age of 2 yr (10 of 254, 3.9%; and 19 of 751, 2.5%, respectively; P = 0.34; Fig. 2B). The proportion of children who tested positive for GADA was more than 4-fold among those with the high-risk genotype (10 of 254, 3.9%; vs. 7 of 751, 0.9%; P = 0.006; Fig. 2C), and IA-2A were also detected more frequently in this group (7 of 254, 2.8%; vs. 5 of 751, 0.7%; P = 0.029; Fig. 2D).
|
Altogether, 29 children tested persistently positive for at least one autoantibody, among whom 27 were persistently positive for ICA, 21 for IAA, 17 for GADA, and 11 for IA-2A. ICA screening identified 93% of those with persistent positivity for at least one autoantibody (27 of 29), IAA 79% (23 of 29), GADA 66% (19 of 29), and IA-2A 41% (12 of 29). All of these children were identified with combined screening for ICA and IAA, whereas 8 of the 29 (28%) remained unidentified with combined screening for GADA and IA-2A.
Positivity for multiple autoantibodies by the age of 2 yr
Twenty-three of 1,005 children (2.2%) had multiple (2) autoantibodies at least once (4 children had two antibodies, 9 had three, and 10 had all four), and 22 of the 63 antibody-positive children (35%) tested persistently positive for multiple antibodies. The median duration of persistent positivity for at least two autoantibodies was 0.7 yr (range, 0.1–2.2 yr). Both ICA and IAA screening identified 96% (21 of 22) of those who tested persistently positive for multiple antibodies, GADA 86% (19 of 22), and IA-2A 55% (12 of 22; Table 1). Among the single antibody specificities, positivity for IA-2A had the highest specificity (99.9%) and PPV (92.3%) for this surrogate marker of clinical diabetes. Combined positivity for both ICA and IAA had a sensitivity of 90.9%, a specificity of 99.9%, and a PPV of 95.2% for persistent positivity for multiple antibodies by the age of 2 yr. All 22 children with persistent positivity for multiple antibodies by the age of 2 yr were identified by combined screening for ICA and IAA.
|
The mean age at seroconversion to positivity for IAA was 1.0 yr (range, 0.1–1.9 yr; median, 0.9 yr), that for GADA was 1.2 yr (range, 0.1–1.9 yr; median, 1.2 yr), that for ICA was 1.3 yr (range, 0.4–1.9 yr; median, 1.3 yr), and that for IA-2A was 1.3 yr (range, 0.7–1.9 yr; median, 1.2 yr). Thus, no significant difference in age at seroconversion was observed between these autoantibodies. If the child tested positive for both autoantibodies compared, IAA appeared at a younger age than ICA (P = 0.001; paired t test), GADA (P = 0.016), or IA-2A (P = 0.001), whereas ICA appeared at a younger age than IA-2A (P = 0.004). Among the 23 children who tested positive for at least two autoantibodies by the age of 2 yr, the mean time period between the appearance of the first and second autoantibodies was 0.2 yr (range, 0–1.1 yr).
Inverse seroconversions
Close to half of the antibody-positive children (29 of 63, 46%) reverted to antibody negativity during the follow-up, after one (19 of 29) to six positive samples. All of these 29 children tested positive for only one autoantibody by the age of 2 yr, only three of them having ICA (maximum level, 5 JDFU, 5 JDFU, and 8 JDFU), 22 IAA (median, 4.34 RU; range, 1.93–63.7 RU), three GADA (maximum level, 10.6 RU, 13.3 RU, and 13.9 RU) and one IA-2A (maximum level, 0.61 RU). IAA disappeared more frequently than the other autoantibodies (IAA vs. ICA, P = 0.001; IAA vs. GADA, P = 0.012; and IAA vs. IA-2A, P = 0.021). No significant difference in the frequency of inverse seroconversions was observed between the high and moderate-risk genotypes.
In nine children, the autoantibodies fluctuated between positivity and negativity. In one child ICA appeared, disappeared, and reappeared (maximum titer, 10 JDFU); in six IAA they fluctuated (maximum titer, 20.4 RU); in one GADA they fluctuated (maximum titer, 8.9 RU); and in another IA-2A they fluctuated (maximum titer, 0.43 RU). Two of these children tested antibody-negative in the last sample taken during the follow-up.
Autoantibodies in children who progressed to clinical diabetes in the cohort of 1,005 children
Five of the 1,005 children (0.5%) monitored for appearance of any of the four autoantibodies developed clinical diabetes during the observation period, at the ages of 2.1, 2.2, 2.3, 2.5, and 4.3 yr, and three of them carried the high-risk genotype (Fig. 3). They all had at least two autoantibodies 1.0, 1.2, 0.1, 1.0, and 3.3 yr before the diagnosis of clinical diabetes. All five children who developed diabetes had IAA as their first autoantibody, one also having ICA, and one ICA, IAA, and GADA in that sample.
|
The whole DIPP cohort included 15 children who had taken part in the immunological follow-up and had progressed to clinical type 1 diabetes by the end of April 2000, at a median age of 1.9 yr (range, 1.0–4.3 yr; Table 2). Fourteen of them (93%; CI, 68.0–99.8%) had IAA before the diagnosis. One child had no antibodies in the samples taken before diagnosis, for the last time at the age of 1 yr, but had developed ICA and IAA by the time of diagnosis 8 months later (Table 2, case 4). Twelve children (80%; CI, 51.9–95.7%) tested positive for ICA before the diagnosis, and 14 did so at the time of diagnosis. One of the 15 children had no ICA before and at the time of diagnosis, but had IAA 0.5 yr before the diagnosis and tested in addition positive for GADA at the time of diagnosis (Table 2, case 14). Nine of the progressors (60%; CI, 32.3–83.7%) had GADA, and six had IA-2A (40%; CI, 16.3–67.7%) before the time of diagnosis of type 1 diabetes. IAA were the first or among the first antibodies to appear in the 15 cases in whom type 1 diabetes developed.
|
The incidence of type 1 diabetes among Finnish children under the age of 15 yr was 49 of 100,000 in year 2001, whereas the disease prevalence was 369 of 100,000 children (Reunanen, A., personal communication). On the basis of recent incidence rates, the absolute risk of type 1 diabetes by the age of 15 yr in the Finnish population can be estimated to be close to 0.7%. The risk conferred by the high-risk HLA-DQB1*02/*0302 genotype is around 7%, whereas that associated with the moderate-risk genotypes is close to 2.5% (24). In this cohort of 1,005 genetically susceptible children, 6.3% (CI, 4.8–7.8%) tested positive for at least one autoantibody by the age of 2 yr. This represents a high prevalence, even when considering that the present study population is derived from that portion of the general population that carries increased genetic susceptibility to type 1 diabetes. In the German BABYDIAB study comprising genetically susceptible offspring (monitored from birth) of affected parents at the ages of 9 months, 2, 5, and 8 yr, 20% (CI, 9.4–30.6%) tested positive for at least one autoantibody by the age of 2 yr (12). The proportion of antibody-positive children among those carrying the high-risk genotype DQB1*02/*0302 in our DIPP study (8.3%; CI, 5.2–12.4) is considerably higher than in the American Diabetes Autoimmunity Study in the Young, in which only 4 of 208 (1.9%; CI, 0.5–4.9%) children from the general population with the high-risk genotype DR3/4 (DQ2/8) had autoantibodies at the median age of 2 yr (25). The prevalence of children testing positive for IAA was significantly higher in the present study compared with children from the general population in the American Diabetes Autoimmunity Study in the Young (48 of 1,005, 4.8%, vs. 11 of 793, 1.4%; P = 0.0001), whereas there was no significant difference in the prevalence of GADA and IA-2A.
In the present series, like in BABYDIAB, IAA were the most common antibodies to appear at a young age. The proportion of children testing positive for IAA increased earlier than the proportions of those becoming positive for other autoantibodies. However, IAA disappeared clearly more often than the other antibodies. If ICA only had been analyzed in the children of this study, 51% (32 of 63) of the children who developed other autoantibodies would have remained unidentified, together with 90% (26 of 29) of those who seroconverted to autoantibody negativity during the follow-up. Nevertheless, ICA screening identified 93% (27 of 29) of those who developed persistent positivity for at least one antibody and IAA 79% (23 of 29), and by combining these two tests, 95% (21 of 22) of those with persistent positivity for at least two antibodies would have been detected. Thus, the vast majority of those developing signs of persistent ß-cell autoimmunity during the first few years of life can be recognized with ICA or IAA screening when genetically susceptible children are first selected from the general population. We have previously shown that the presence of multiple autoantibodies is a relatively good surrogate marker of clinical type 1 diabetes among young siblings of affected children (26). The present data indicate that positivity for at least two diabetes-associated autoantibodies represents in most cases a point of no return, because inverse seroconversions from multiple autoantibody positivity to antibody negativity or positivity for a single antibody are rare. In fact, there was only one child who tested positive for two autoantibodies at some point but seroconverted later to positivity for only one autoantibody. No such child seroconverted to autoantibody negativity. It is possible that the number of inverse seroconverters among subjects positive for multiple autoantibodies might have been higher, if there would have been more children in the latter group.
The BABYDIAB Study indicates that the detection of autoantibodies after 9 months of age predicts future diabetes (12). In the present series, 2 of the 15 children who developed clinical diabetes during the follow-up had IAA in their 6-month sample, and 1 of them had IAA already in the 3-month sample, whereas no detectable antibodies were found in cord blood. These data suggest that the first signs of biologically significant ß-cell autoimmunity may at least in some cases appear already during the first few months of life. Thirteen of the 15 children (87%) who progressed to type 1 diabetes had multiple (2) antibodies before the diagnosis, whereas all 15 children had multiple antibodies at the time of diagnosis. The increased risk of clinical type 1 diabetes associated with the presence of multiple autoantibodies is well in line with previous findings in surveys among young first-degree relatives of patients with type 1 diabetes (12, 26).
IAA have been reported to commonly be the first detectable autoantibody (12, 27, 28), and we similarly found that they were most often the first autoantibody to appear and were actually the first or among the first antibodies to emerge in all 15 children who progressed to type 1 diabetes. On the other hand, a substantial proportion of the children testing positive for IAA lost their IAA during follow-up, and the majority of these (90%) never developed ICA. Transient autoantibody positivity has been reported to be a fairly uncommon phenomenon among children in the general population; it mainly occurs in individuals testing positive for one autoantibody only, and then the antibody titer is usually low (25, 29). In an Australian report comprising infants with a family member with type 1 diabetes, transient ß-cell autoimmunity was observed to be a relatively common phenomenon, because 14% of the infants tested positive for autoantibodies on only one occasion. Inverse seroconversions occurred mainly in individuals testing positive for IAA or IA-2A in that survey (30). Similarly, all inverse seroconverters in the present study had initially only one autoantibody. In contrast to the Australian study, however, inverse seroconversions were rare in children testing positive for IA-2A. Yu et al. (25) reported no differences in the frequency of transient autoantibody positivity between high- and moderate-risk genotypes. In our study, in which all of the children carried high- or moderate-risk genotypes, no differences were seen in transient antibody positivity between the two genotype groups. The significance of transient IAA positivity at a young age remains to be defined, however. Transient IAA positivity may be harmless, but one might also speculate that it could be an early marker preceding the appearance of other autoantibodies later in life.
One might argue that the high frequency of transient IAA positivity observed in the present study could be due to methodological limitations. This is, however, not very likely, because all samples from the same individual were analyzed for IAA in the same assay to eliminate the interassay variation, and the IAA method had the lowest intra-assay variation among the four antibody assays applied. The IAA microassay in the present study is based on the use of Protein A for the separation of antigen-antibody complexes and is more specific than the conventional liquid phase RIA using polyethylene glycol for the precipitation of bound insulin (31). Accordingly, transient antibody positivity most likely reflects true disappearance of IAA rather than assay variation due to methodological limitations.
The high frequency of transient IAA in young children complicates optimal screening strategies for intervention trials. The psychosocial effects of the implementation of a screening and intervention program and the safety of the measures aimed at preventing the disease need to be carefully considered. Information given to the parents of a healthy child that signs of ß-cell autoimmunity have been detected will almost certainly induce at least transient anxiety in the family. Accordingly, IAA alone do not offer an optimal approach for screening of the general population early in life, unless the intervention planned is perfectly safe. Repeated sampling is obviously needed in screening programs to monitor the progression of the autoimmune processes and to determine the appropriate timing of intervention. Repeated sampling is also necessary because the first autoantibodies may appear at any age, and until the significance of the transiently appearing autoantibodies has been defined.
The present data confirm that the first signs of preclinical type 1 diabetes may appear during the first months of life, and IAA are in most cases the first or among the first autoantibodies to appear. High sensitivity, early appearance, and high frequency of transient antibody positivity are characteristic of IAA, whereas ICA detected with a thoroughly standardized assay have a higher specificity and PPV than IAA. When young genetically predisposed subjects are studied, IAA combined with ICA are sensitive markers for the identification of individuals at high risk of developing type 1 diabetes, and their combination might provide an effective tool for screening of young children for secondary intervention trials when the use of this combination is linked to genetic screening of the general population.
Acknowledgments
We thank Paula Arvilommi, Reija Hakala, Anu-Maaria Hämäläinen, Päivi Keskinen, Birgitta Nurmi, Hilkka Pohjola, Kaija-Leena Rasimus, Helena Savolainen, Riikka Sihvo, Aino Stenius, Aila Suutari, and Maija Törmä for their commitment to the study. We are also grateful to Sirpa Anttila, Susanna Heikkilä, Terttu Lauren, Tuovi Mehtälä, Riitta Päkkilä, Sirpa Pohjola, Päivi Salmijärvi, and Ritva Suominen for their skillful technical assistance.
Footnotes
This work was supported by the Foundation for Pediatric Research, Finland; Medical Research Funds; Tampere, Oulu, and Turku University Hospitals; Medical Research Council; Academy of Finland; the Diabetes Research Foundation in Finland; the Juvenile Diabetes Research Foundation International (Grants 197032 and 4-1998-274); Novo Nordisk Foundation; and EU Biomed 2 Program (BMH4-CT98-3314).
Abbreviations: CI, 95% Confidence interval; DASP, Diabetes Autoantibody Standardization Program; DIPP, Type 1 Diabetes Prediction and Prevention Project; GADA, antibodies to GAD65; HLA, human leukocyte antigen; IAA, insulin autoantibodies; IA-2A, IA-2 antibodies; ICA, islet cell antibodies; JDFU, Juvenile Diabetes Foundation units; PPV, positive predictive value; RU, relative units.
Received January 10, 2002.
Accepted June 23, 2002.
References
This article has been cited by other articles:
 |
|
 |
|
  M. Oresic, S. Simell, M. Sysi-Aho, K. Nanto-Salonen, T. Seppanen-Laakso, V. Parikka, M. Katajamaa, A. Hekkala, I. Mattila, P. Keskinen, et al. Dysregulation of lipid and amino acid metabolism precedes islet autoimmunity in children who later progress to type 1 diabetes J. Exp. Med., December 22, 2008; 205(13): 2975 - 2984. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Makinen, T. Harkonen, J. Ilonen, M. Knip, and the Finnish Pediatric Diabetes Register Characterization of the humoral immune response to islet antigen 2 in children with newly diagnosed type 1 diabetes. Eur. J. Endocrinol., July 1, 2008; 159(1): 19 - 26. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. L.P. Caforio, N. G. Mahon, M. K. Baig, F. Tona, R. T. Murphy, P. M. Elliott, and W. J. McKenna Prospective Familial Assessment in Dilated Cardiomyopathy: Cardiac Autoantibodies Predict Disease Development in Asymptomatic Relatives Circulation, January 2, 2007; 115(1): 76 - 83. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Kondrashova, H. Viskari, P. Kulmala, A. Romanov, J. Ilonen, H. Hyoty, and M. Knip Signs of {beta}-Cell Autoimmunity in Nondiabetic Schoolchildren: A comparison between Russian Karelia with a low incidence of type 1 diabetes and Finland with a high incidence rate Diabetes Care, January 1, 2007; 30(1): 95 - 100. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Mrena, S. M. Virtanen, P. Laippala, P. Kulmala, M.-L. Hannila, H. K. Akerblom, M. Knip, and the Childhood Diabetes in Finland Study Group Models for predicting type 1 diabetes in siblings of affected children. Diabetes Care, March 1, 2006; 29(3): 662 - 667. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Achenbach, E. Bonifacio, K. Koczwara, and A.-G. Ziegler Natural History of Type 1 Diabetes Diabetes, December 1, 2005; 54(suppl_2): S25 - S31. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Pihoker, L. K. Gilliam, C. S. Hampe, and A. Lernmark Autoantibodies in Diabetes Diabetes, December 1, 2005; 54(suppl_2): S52 - S61. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Kukko, T. Kimpimaki, S. Korhonen, A. Kupila, S. Simell, R. Veijola, T. Simell, J. Ilonen, O. Simell, and M. Knip Dynamics of Diabetes-Associated Autoantibodies in Young Children with Human Leukocyte Antigen-Conferred Risk of Type 1 Diabetes Recruited from the General Population J. Clin. Endocrinol. Metab., May 1, 2005; 90(5): 2712 - 2717. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Pundziute-Lycka, L.-A. Persson, G. Cedermark, A. Jansson-Roth, U. Nilsson, V. Westin, and G. Dahlquist Diet, Growth, and the Risk for Type 1 Diabetes in Childhood: A matched case-referent study Diabetes Care, December 1, 2004; 27(12): 2784 - 2789. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Kukko, S. M. Virtanen, A. Toivonen, S. Simell, S. Korhonen, J. Ilonen, O. Simel, and M. Knip Geographical Variation in Risk HLA-DQB1 Genotypes for Type 1 Diabetes and Signs of {beta}-Cell Autoimmunity in a High-Incidence Country Diabetes Care, March 1, 2004; 27(3): 676 - 681. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Achenbach, K. Warncke, J. Reiter, H. E. Naserke, A. J.K. Williams, P. J. Bingley, E. Bonifacio, and A.-G. Ziegler Stratification of Type 1 Diabetes Risk on the Basis of Islet Autoantibody Characteristics Diabetes, February 1, 2004; 53(2): 384 - 392. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Atkinson and E. A. M. Gale Infant Diets and Type 1 Diabetes: Too Early, Too Late, or Just Too Complicated? JAMA, October 1, 2003; 290(13): 1771 - 1772. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
), IAA (
), GADA (
), and IA-2A (•) from birth to the age of 2 yr in children with increased genetic susceptibility to type 1 diabetes identified from the general population.
