ORIGINAL ARTICLES

Poor Beta-cell Function After theClinical Manifestation of Type 1Diabetes in Children Initially Positivefor Islet Cell Specific AutoantibodiesJ. Komulainena, M. Knipb, R. Lounamaac, P. Vahasalob, J. Karjalainenb, E. Sabbahb,H.K. Åkerblomd, and the Childhood Diabetes in Finland Study Group*

aDepartment of Pediatrics, Kuopio University Hospital, KuopiobDepartment of Pediatrics, University of Oulu, OulucDepartment of Epidemiology, National Public Health Institute,HelsinkidThe Children’s Hospital, II Department of Pediatrics, University ofHelsinki, Helsinki, Finland

The prognostic significance of islet cell specific autoantibodies at the diagnosis of Type 1(insulin-dependent) diabetes mellitus for the persistence of residual beta-cell function overthe first 2 years of clinical disease was evaluated in a prospective population-based study.Seven hundred and eighty probands, aged 0.8–14.9 years, were examined for islet cellantibodies (ICA) and insulin autoantibodies (IAA), while 769 probands were studied forantibodies to glutamic acid decarboxylase (GAD65A). They were subsequently observedfor 2 years. Lower serum C-peptide concentrations and higher requirement of exogenousinsulin during the follow-up period were observed in the group of probands positive forat least one of the antibodies, especially for ICA or IAA. We conclude that the residualbeta-cell function after the presentation of Type 1 diabetes is less in children initiallypositive for islet cell specific autoantibodies than in those testing negative at diagnosis.This might reflect possible heterogeneity in the pathogenesis of childhood diabetes. It alsodemonstrates that ICA and IAA negativity at the diagnosis of Type 1 diabetes is notassociated with a smaller amount of functioning beta-cell mass, but the absence ofantibodies probably reflects a slower beta-cell destructive process and a longer durationof preclinical disease. 1997 by John Wiley & Sons, Ltd.

Diabet. Med. 14: 532–537

No of Figures: 2. No of Tables: 3. No of Refs: 26

KEY WORDS Islet cell antibodies Insulin autoantibodies Glutamic acid decarboxylaseantibodies Serum C-peptide Insulin dose Metabolic control Type 1diabetes

Received 24 June 1996; revised 14 February 1997; accepted 9 March 1997

IntroductionAbbreviations: ICA islet cell antibodies, IAA insulin autoantibodies,

Islet cell specific autoantibodies are found in the majorityGAD65A glutamic acid decarboxylase antibodies, DKA diabetic ketoac-idosis. of children with newly diagnosed Type 1 (insulin-* Correspondence to: Dr Jorma Komulainen, Department of Pediatrics,Kuopio University Hospital, P.O. Box 1777, FIN-70211 Kuopio, Finland.*The Childhood Diabetes in Finalnd (DiMe) Study Group comprises T. Otonkoski, K. Pihlajamaki, S. Pontynen, J. Rajantie, J. Sankala, J.

Schumacher, M. Sillanpaa, M.-R. Stahlberg, C.-H. Strahlmann, T.the following members:Contact grants Nos: 188517 (Juvenile Diabetes Foundation, USA), DK- Uotila, M. Vare, P. Varimo, G. Wetterwstrand. Special Investigators:

A. Aro (National Public Health Institute, Helsinki), M. Hiltunen (Medical37957 (National Institutes of Health, USA).Sponsors: Orion Corporation Research Foundation, Helsinki; Kyllikki School, University of Tampere), H. Hurme (Abo Akademi, Pedagogic

Faculty in Vaasa), H. Hyoty (Medical School, Universtiy of Tampere),and Uolevi Lehikoinen Foundation, Helsinki; Finnish Cultural Foun-dation; North-Savo Foundation, Kuopio, Finland; Alma and K.A. J. Ilonen (Department of Virology, Universtiy of Turku), J. Karjalainen

(Department of Pediatrics, Universtiy of Oulu), M. Knip (DepartmentSnellman Foundation, Oulu, Finland,; Association of Finnish LifeInsurance Companies; Sigrid Juselius Foundation; University of Helsinki; of Pediatrics, University of Oulu), P. Lienikki (National Public Health

Institute, Helsinki), A. Miettinen (Department of Bacteriology andNordisk Insulin Foundation; Novo-Nordisk A/S, Denmark.The Childhood Diabetes in Finland (DiMe) Study Group comprises the Immunology, University of Helsinki), T. Petays (Department of Bacteri-

olgy and Immunology, University of Helsinki), L. Rasanen (Departmentfollowing members:Principal investigators: H.K. Akerblom, J. Tuomilehto. Co-ordinators: of Nutrition, University of Helsinki), H. Reijonen (Department of

Virology, University of Turku) A. Reunanen (Social Insurance Institute,R. Lounamaa, L. Toivanen. Data management: E. Virtala, J. Pitkaniemi.Local Investigators: A. Fagerlund, M. Flittner, B. Gustafsson, C. Helsinki), T. Saukkonen (Children’s Hospital, University of Helsinki),

E. Savilahti (Children’s Hospital, University of Helsinki), E. Tuomilehto-Haggquist, A. Hakulinen, L. Herva, P. Hiltunen, T. Huhtamaki, N.-P.Huttunen, T. Huupponen, M. Hyttinen, T. Joki, R. Jokisalo, M.-L. Kaar, Wolf (National Public Health Institute, Helsinki), P. Vahasalo

(Department of Pediatrics, University of Oulu), S.M. VirtanenS. Kallio, E.A. Kaprio, U. Kaski, M. Knip, L. Laine, J. Lappalainen, J.Maenpaa, A.-L. Makela, K. Niemi, A. Niiranen, A. Nuuja, P. Ojajarvi, (Department of Nutrition, University of Helsinki).

532 CCC 0742–3071/97/070532–06$17.50 1997 by John Wiley & Sons, Ltd. DIABETIC MEDICINE, 1997; 14: 532–537

ORIGINAL ARTICLESdependent) diabetes mellitus, the prevalence of islet and negative groups for both ICA, IAA, and GAD65A.

In the longitudinal study the groups were compared atcell antibodies (ICA) being 60–86 %,1–5 while insulinautoantibodies (IAA) are found in 18–69 %,6–11 and the time of diagnosis and at 6, 12, 18, and 24 months

after diagnosis.antibodies to the 65-kDa isoform of glutamic aciddecarboxylase (GAD65A) in 70–80 % of patients at thetime of diagnosis.12–15 Definitions and Methods

A series of studies considering the possible relationbetween ICA and residual beta-cell function after the Diabetic Ketoacidosisclinical manifestation of the disease have been pub-lished,3,5,16–20 whereas there are only a few reports on DKA was defined as a capillary or venous blood pH of

less than 7.30.the role of IAA in relation to the preservation of theendogenous insulin secretion after diagnosis.9,21 Theresults of earlier studies have been somewhat contradic- Daily Insulin Dose and Remissiontory. Most of the surveys have reported an inversecorrelation between ICA positivity and the recovery of The recommended amount of insulin and the weight of

the proband were registered at 3 weeks after diagnosisbeta-cell function,5,19,20 while no relation has been foundin other studies,16,17 and even a positive correlation and 6-monthly thereafter, and the daily insulin dose

(IU kg−1day−1) was calculated.between persistent ICA and endogenous insulin secretionhas been observed by one group.3,18 Previous reportshave shown either no effect9,21 or a negative effect20 of Endogenous Insulin SecretionIAA on residual beta-cell function after the diagnosis.

To study the effect of initial ICA, IAA, and GAD65A Random serum C-peptide concentrations were analysedwith a radioimmunoassay, as described earlier,22 bypositivity on the severity of diabetic ketoacidosis (DKA)

at diagnosis and on residual beta-cell function in a large using antiserum K6 (Novo Research Institute, Bagsvaerd,Denmark). The intra-assay coefficient of variation waspopulation-based series of subjects, 780 Finnish children

and adolescents, aged 0.8–14.9 years and participating 1.8 % and that of inter-assay variation 10 %. We havepreviously shown that there is a strong correlationin a nationwide multicentre study of childhood diabetes

were evaluated at the time of diagnosis of Type 1 between random serum C-peptide levels and serumC-peptide concentrations measured 120 min after adiabetes for the presence of ICA, IAA, and GAD65A,

and observed thereafter for endogenous insulin secretion, standardized meal and 24-h urinary C-peptide secretion.9

exogenous insulin requirement and metabolic controlfor 24 months. Metabolic Control

Standard methods for blood HbA1 and HbA1C analysisSubjects and Methodswere used in the various hospitals participating in thestudy. To make the results comparable, they wereSubjectsexpressed as standard deviation scores (SDS) above themean for non-diabetic subjects. Clinical remission wasAs part of the Finnish nationwide ‘Childhood Diabetes

in Finland’ study, 801 probands less than 15 yesrs of defined as a period in which daily insulin dose was lessthan 0.5 IU kg−1day−1 with glycated haemoglobin (HbA1age, diagnosed to have diabetes during the recruitment

period from 1 September 1986 to 30 April 1989, were or HbA1C) not exceeding the mean laboratory specificreference value by 4 SD or more.invited to participate in the study. Seven hundred and

fifty-nine of them were analysed for ICA and IAA at thediagnosis of their disease, and twenty-one 2 to 3 weeks Autoantibody Testsafter that. At the time of diagnosis, capillary or venousblood pH, blood glycated haemoglobin and serum C- Autoantibody tests were performed in the Research

Laboratory, Department of Pediatrics, University of Oulu.peptide levels were measured. The probands were thenfollowed in their own outpatient clinics (n = 31), and ICA were determined by a standard indirect immunofluo-

rescence assay performed on sections of frozen humandata were collected at 6-month intervals for 2 years. Ateach 6-monthly follow-up visit, the recommended blood group O pancreas.23 Rabbit anti-human IgG or

rabbit anti-human C3c (Behringwerke, Marburg,amount of exogenous insulin was registered, blood forglycated haemoglobin (GHb) and serum C-peptide was Germany) were used to detect ICA. The results were

expressed in Juvenile Diabetes Foundation (JDF)-units.24taken, and a clinical examination including height andweight was performed. The data were subsequently The detection limit for ICA was 2.5 JDF-units. The

laboratory has participaated in the international work-transferred to the Central Data Office of the study.The mean age of the probands was 8.5 (range 0.8– shops on the standardization of the ICA assay, with

sensitivity, specificity, validity, and consistency of 100 %,14.9) years. The majority of them were male (427;54.7 %). The subjects were divided into antibody positive 98 %, 98 %, and 98 %, respectively, in the fourth round.

533RESIDUAL BETA-CELL FUNCTION AND AUTOANTIBODIES

ORIGINAL ARTICLESIAA was measured by a modification of the liquid subjects at the time of hospitalization. Approximately

one-fifth of them (157; 21.8 %) had DKA at the diagnosisphase radioimmunoassay originally described by Palmeret al.6 In this assay the specific insulin binding in 105 of IDDM. Eighty-six (11.9 %) had a pH value between

7.20–7.29, 38 (5.3 %) between 7.10–7.19, and 33 (4.6 %)non-diabetic children was 23.3 + 11.2 nU ml−1. The cut-off limit for antibody positivity was 57 nU ml−1 (mean less than 7.10. There was no difference in the frequency of

DKA between autoantibody positive or negative subjects.+3 SDS in 105 non-diabetic subjects). The laboratoryhas participated in the international workshops on Age and C-peptide concentrations showed a positive

correlation (dg: r = 0.23; p , 0.001, 6 months: r = 0.42;standardization of insulin autoantibody assays and in theproficiency testing programme for several years, and p , 0.001, 12 months: r = 0.46; p , 0.001, 18 months;

r = 0.45; , 0.001, 24 months: r = 0.37; p , 0.001). Afterachieved a sensitivity of 78 %, a specificity of 100 %,validity of 92 %, and a consistency of 100 % in the first age adjustment, the mean serum C-peptide concentrations

at the diagnosis and during the follow-up period wereproficiency testing programme.GAD65A was measured with the radioligand assay as lower in children positive for at least one antibody than

in children negative for all antibodies (Figure 1). Thedescribed by Petersen et al.15 The results were expressedin relative units (RU), representing the specific binding same held true for ICA positive children compared to

ICA negative children at the diagnosis and during theas a percentage of that obtained with a positive standardserum. The cut-off limit for antibody positivity was set follow-up period as well as for IAA positive children

when compared to IAA negative ones during the follow-at the 99th percentile of 372 non-diabetic children andadolescents, i.e. 6.6 RU. The intra-assay coefficient of up period (Table 2). Children positive for GAD65A had

lower serum C-peptide concentrations than GAD65Avariation was less than 5 %, and the inter-assay coefficientof variation was less than 10 %. The disease sensitivity negative children 24 months after the diagnosis.

A weak but statistically significant inverse correlationof the present assay was 80 %, and the specificity was95 % based on the 101 samples included in the Second was found between ICA levels and serum C-peptide

levels (dg: r = −0.03; NS, 6 months: r = −0.14; p , 0.001,International GAD Antibody Workshop.25

12 months: r = −0.15; p , 0.001, 18 months: r = −0.12;p , 0.01, 24 months: r = −0.12; p , 0.01) as well asStatistical Analysisbetween IAA levels and serum C-peptide levels (dg:r = −0.07; p = 0.06, 6 months: r = −0.12; p , 0.01, 12The data were subjected to contingency table analysis

with chi-square statistics and the Mantel-Haenzel testfor linear association as well as linear regression analysis,multiple linear regression analysis, and analysis ofvariance with convariance analysis, when appropriate.

Results

At the time of the diagnosis, 732 (95.2 %) subjects werepositive for at least one of the antibodies. Six hundredand fifty-six (84.1 %) subjects were found to be ICApositive, 327 (47.7 %) IAA positive, and 567 (73.7 %)GAD65A positive. There was no statistically significantdifference between girls and boys in ICA (85.3 % vs83.2 %) or IAA positivity (49.3 % vs 46.4 %), whereasgirls were positive for GAD65A more often than boys

Figure 1. Mean (SE) serum C-peptide levels (nmol−1) in children(77.5 % vs 70.6 %; p , 0.05, chi-square test). Young ageinitially positive for at least one antibody (shadowed columns)at diagnosis was associated with a higher frequency ofand in children initially negative for all antibodies (open

ICA and IAA, while the oldest probands were most often columns) at diagnosis and during the follow-up period. (1)positive for GAD65A (Table 1). p , 0.05, (2) p , 0.001; multiple linear regression analysis

with age includedCapillary or venous blood pH was registered in 721

Table 1. Prevalence of islet cell specific autoantibodies in different age groups at the time of diagnosis of diabetes. Mantel–Haenzeltest for linear association was used for statistical analysis

0.8–4.9 yr 5.0–9.9 yr 10.0–14.9 yr p

At least one antibody + 96.5 % (164/170) 96.6 % (287/297) 93.0 % (281/302) 0.057ICA+ 86.0 % (148/172) 89.7 % (271/302) 77.5 % (237/306) ,0.01IAA+ 71.7 % (124/172) 43.2 % (130/302) 38.6 % (118/306) ,0.001GAD65A 70.6 % (120/170) 69.7 % (207/297) 79.5 % (240/302) ,0.05

534 J. KOMULAINEN ET AL.

ORIGINAL ARTICLESTable 2. Mean (95 % confidence interval) serum C-peptide concentrations (nmol−1) at the diagnosis and during the follow-up period in probands positive or negative for the different autoantibodies. The multiple linear regression analysis withage included was used for calculation of statistical significance

dg 6 months 12 months 18 months 24 months

ICA+ 0.19 (0.18–0.20) 0.24 (0.23–0.26) 0.14 (0.13–0.15) 0.08 (0.08–0.09) 0.06 (0.05–0.07)ICA− 0.24 (0.20–0.27) 0.28 (0.25–0.32) 0.24 (0.20–0.28) 0.15 (0.12–0.18) 0.12 (0.09–0.16)p ,0.05 NS ,0.001 ,0.001 ,0.001IAA+ 0.19 (0.18–0.20) 0.24 (0.23–0.26) 0.14 (0.13–0.15) 0.08 (0.08–0.09) 0.06 (0.05–0.07)IAA− 0.24 (0.20–0.27) 0.28 (0.25–0.32) 0.24 (0.20–0.27) 0.15 (0.12–0.18) 0.12 (0.09–0.16)p ,0.05 NS ,0.001 ,0.001 ,0.001GAD65A+ 0.19 (0.18–0.20) 0.24 (0.23–0.26) 0.16 (0.14–0.17) 0.09 (0.08–0.10) 0.07 (0.06–0.08)GAD65A− 0.20 (0.18–0.23) 0.26 (0.23–0.28) 0.16 (0.14–0.18) 0.10 (0.08–0.11) 0.08 (0.06–0.09)p NS NS NS NS ,0.05

months: r = −0.13; p , 0.01, 18 months: r = −0.13; p , 0.01; 24 months: 22.2 % vs 3.1 %, p , 0.0001).This was also seen in ICA negative vs ICA positivep , 0.01, 24 months: r = −0.06, NS). No correlation

was found between GAD65A levels and serum C- subjects (6 months: 36.3 % vs 29.9 %, NS; 12 months:27.7 % vs 17.4 %, p , 0.05; 18 months: 18.1 % vspeptide levels.

When adjusted for age, the daily insulin doses at 6, 2.8 %, p , 0.001; 24 months: 7.0 % vs 5.1 %, NS). IAAnegative children also experienced partial remission12, 18, and 24 months were higher in children positive

for at least one antibody than in antibody negative more often than IAA positive children, although afterage adjustment, the difference was statistically significantprobands (Figure 2). This was true also in ICA positive

children when compared to ICA negative ones, and in only at 12 months (6 months: 33.5 % vs 26.1 %, NS;12 months: 22.4 % vs 14.4 %, p , 0.05; 18 months:IAA positive children when compared to IAA negative

ones (Table 3). No differences in insulin doses were 7.9 % vs 2.7 %, NS; 24 months: 5.5 % vs 5.4 %, NS).For GAD65A negative vs positive patients, clinicalobserved between GAD65A positive and negative pro-

bands. remission was significantly more common only at 18months (6 months: 34.3 % vs 34.1 %, NS; 12 months:When adjusted for age, the glycaemic control assessed

with GHb values did not differ between the autoantibody 15.0 % vs 14.1 %, NS; 18 months: 17.6 % vs 9.1 %,p , 0.05; 24 months: 6.1 % vs 3.2 %, p = 0.091).positive and negative groups (data not shown).

Clinical remission as defined in the methods sectionwas more common in patients who tested negative for

Discussionall antibodies than in those who tested positive for atleast one. After adjustment for age, the differences

We have prospectively evaluated the possible prognosticremained statistically significant at 18 and 24 months (6significance of islet cell specific autoantibodies detectedmonths: 35.0 % vs 34.1 %, NS; 12 months: 27.3 %at the time of diagnosis of Type 1 diabetes in childrenvs 13.8 %, p = 0.095; 18 months: 33.3 % vs 10.1 %,and adolescents with regard to the persistence of residualbeta-cell function over the first 2 years of clinical disease.The present study is a nationwide population-basedsurvey in the country having the highest incidence ofchildhood diabetes in the world.26 Owing to the largenumber of probands it is possible to detect in variousgroups of diabetic subjects even small differences inclinical outcome, which might indicate heterogeneity inthe aetiology of the disease.

Due to the high number of participating hospitals,there was some variability in the treatment of the patientswhich might bias the results. This is, however, not likely,since, when divided into five geographical regions, theonly differences found between the regions were a lowerdaily insulin dose at the 6-month follow-up visit in one

Figure 2. Mean (SE) daily insulin dose (IU kg−1day−1) in children of the regions, and a lower frequency of insulin treatmentinitially positive for at least one antibody (shadowed columns) based on multiple injections in another region (data notand in children initially negative for all antibodies (open

shown). The islet cell specific autoantibodies and C-columns 3 weeks after hospitalization (dg) and during thepeptide levels were analysed in a central laboratory tofollow-up period. (1) p , 0.01, (2) p , 0.001; multiple linear

regression analysis with age included avoid possible interlaboratory differences. The prevalence

535RESIDUAL BETA-CELL FUNCTION AND AUTOANTIBODIES

ORIGINAL ARTICLESTable 3. Mean (95 % confidence interval) endogenous insulin doses (IU kg−1day−1) at the diagnosis and during the follow-up period in probands positive or negative for the different autoantibodies. The multiple linear regression analysis withage included was used for calculation of statistical significance

dg 6 months 12 months 18 months 24 months

ICA+ 0.46 (0.44–0.48) 0.46 (0.44–0.47) 0.59 (0.57–0.61) 0.68 (0.66–0.71) 0.73 (0.72–0.75)ICA− 0.45 (0.40–0.50) 0.39 (0.35–0.43) 0.50 (0.45–0.54) 0.59 (0.53–0.65) 0.64 (0.60–0.68)p NS ,0.01 ,0.001 ,0.01 ,0.001IAA+ 0.48 (0.45–0.50) 0.48 (0.45–0.50) 0.60 (0.58–0.62) 0.70 (0.67–0.74) 0.74 (0.71–0.76)IAA− 0.45 (0.42–0.47) 0.42 (0.40–0.44) 0.55 (0.52–0.58) 0.64 (0.60–0.67) 0.70 (0.68–0.72)p ,0.05 ,0.05 ,0.05 ,0.05 ,0.05GAD65A+ 0.47 (0.45–0.49) 0.45 (0.43–0.47) 0.58 (0.56–0.60) 0.69 (0.66–0.71) 0.73 (0.71–0.75)GAD65A− 0.45 (0.41–0.49) 0.44 (0.41–0.48) 0.56 (0.53–0.60) 0.65 (0.60–0.70) 0.70 (0.66–0.73)p NS NS NS NS 0.09

of islet cell specific autoantibodies in the present study after diagnosis, but the difference was not significant.is a good agreement with previous observations.1,3,4 Again, the number of patients was limited (n = 46).

Our data showed poorer beta-cell function after the We found no clear difference in endogenous insulindiagnosis of Type 1 diabetes in children with ICA secretion capacity during the early course of clinicaldetected at the diagnosis of the disease. This is not diabetes between GAD65A positive and negative patients.explained only by different frequencies of ICA positivity Thus it seems that especially ICA and IAA are useful inin various age groups. The observations on residual beta- predicting the clinical course of the disease over the firstcell function and daily insulin dose are in accordance 2 years after the diagnosis. However, the small minoritywith each other. These findings are also consistent with (4.8 %) of our diabetic children and adolescents whoearlier reports by Manna et al.19 and Peig et al.,20 both were negative for all studied antibodies (ICA, IAA, andof whom studied patients who were older than ours. In GAD65A) at the time of diagnosis of IDDM had thecontrast to our observations, Crossley et al.17 found no highest serum C-peptide levels and lowest exogenouseffect of ICA positivity at diagnosis on beta-cell recovery insulin doses.in children with diabetes, while Mustonen et al.3 reported We found no difference in the prevalence of DKA ata positive correlation between persistent ICA after the the diagnosis of Type 1 diabetes in children positive ordiagnosis and endogenous insulin secretion in diabetic negative for islet cell specific autoantibodies. This is inchildren. The number of patients in the study of Crossley contrast with previous observations in older diabeticet al. (n = 21), in which urinary C-peptide was measured, patients by Peig et al.20 In that study, 48 % of the ICAwas too small for detecting minor differences. In the positive and 10 % of the ICA negative patients had DKAstudy of Mustonen et al., children with persistently at the diagnosis of the disease. Elamin et al.5 alsopositive ICA after diagnosis had higher serum C-peptide reported a higher prevalence of DKA, lower serum C-concentrations at 12, 15, and 24 months after diagnosis

peptide concentrations and higher HbA1C levels in ICAthan those subjects who turned negative during the

positive children in Sudan at the diagnosis of IDDM.follow-up. However, they did not find any difference

DKA was seen in 79.3 % of ICA positive and 64.7 % ofbetween ICA positive and negative children at diagnosis.ICA negative children. The prevalence of DKA in ourThe number of children in that study was also smallstudy was much lower than in the two above mentioned(n = 44).studies. This may be a consequence of possible differencesIn the present study, initally IAA positive children hadin the definition of DKA; the criteria for DKA were notlower serum C-peptide levels and a higher daily insulinpresented in the previous studies. It may also reflectdose than IAA negative probands. Young children aredifferences in hospital admission and in timing of clinicalmore often IAA positive than older ones at the diagnosis.diagnosis in the various study populations.They also lose their endogenous insulin secretory capacity

We conclude that children and adolescents with Typefaster. However, the differences in serum C-peptide1 diabetes positive for islet cell specific autoantibodieslevels and daily insulin dose remained even whenat the diagnosis of the disease have a poorer residualadjusted for age. This indicates that all differencesbeta-cell function than antibody negative children andbetween IAA positive and negative subjects cannot beadolescents. This may reflect a pathogenetic variation inexplained by age. Our findings are in accordance withType 1 diabetes presenting in young individuals. Thetwo earlier surveys. In the work by Arslanian et al.,21

observations of a higher C-peptide at diagnosis inthere was a tendency towards higher serum C-peptideantibody negative subjects also suggest that ICA, IAAlevels at diagnosis in IAA negative children, but theand GAD65A negativity at the diagnosis of diabetes isdifference did not reach significance. Also in the studynot associated with a smaller functioning beta-cell massof Karjalainen et al.,9 IAA negative children were liable

to have higher serum C-peptide concentrations 12 months and the absence of antibodies probably reflects a slower

536 J. KOMULAINEN ET AL.

ORIGINAL ARTICLES10. Vardi P, Fiegler AG, Mathews JH, et al. Concentration ofbeta-cell destructive process and a longer duration of

insulin autoantibodies at onset of Type 1 diabetes, inversepreclinical disease.log-linear correlation with age. Diabetes Care 1988; 11:736–739.

11. Ziegler AG, Standl E, Albert E, Mehnert H. HLA-associatedinsulin autoantibody formation in newly diagnosed TypeAcknowledgements1 diabetic patients. Diabetes 1991; 40: 1146–1149.

12. Baekkeskov S, Nielsen JH, Marner B, Blide T, LudvighssonWe are grateful to all participants and their families as J, Lernmark A. Autoantibodies in newly diagnosed diabeticwell as to the respective hospitals and their personnel. children immunoprecipitate human pancreatic islet cell

proteins. Nature 1982; 298: 167–169.We thank L. Toivanen, J. Pitkaniemi, and E. Virtala for13. Christie MR, Landin-Olsson M, Sundkvist G, Dahlquist Gtheir skilful assistance in the collection and arrangement

Lernmark A, Baekkeskov S. Antibodies to a Mr-64,000of the data, and S. Anttila, T. Halonen, R. Pkkila, M.islet cell protein in Swedish children with newly diagnosed

Rissanen, and P. Ronkainen for their valuable technical type 1 (insulin-dependent) diabetes. Diabetologia 1988;assistance. This work was supported by the Orion 31: 597–602.Corporation Research Foundations, Helsinki, Finland, 14. Atkinson MA, Maclaren NK, Sharp DW, Lacy PE, Riley

WJ. 64,000 Mr autoantibodies as predictors of insulinKyllikki and Uolevi Lehikoinen Foundation, Helsinki,dependent diabetes. Lancet 1990; 335: 1357–1360.Finland, Finnish Cultural Foundation, North-Savo Foun-

15. Petersen JS, Moody HA, Karlsen AE, Heinaes KR, Marshalldation, Kuopio, Finland and Alma and K.A. SnellmanMO, Hoier-Madsen M, et al. Detection of GAD65

Foundation, Oulu, Finland. The Childhood Diabetes in antibodies in diabetes and other autoimmune diseasesFinland project has been supported by the Juvenile using a simple radioligand assay. Diabetes 1994; 43:Diabetes Foundation, USA (grant 188517), the Associ- 459–467.

16. Madsbad S, Bottazzo GF, Cudworth AG, Dean B, Faberation of Finnish Life Insurance Companies, the SigridOK, Binder C. Islet-cell antibodies and beta-cell functionJuselius Foundation, the National Institutes of Health,in insulin-dependent diabetics. Diabetologia 1980; 18:USA (DK-37957), the University of Helsinki, the Nordisk45–47.

Insulin Foundation and the Novo-Nordisk A/S, Denmark. 17. Crossley JR, James AG, Elliott RB, Berryman CC, EdgarDW. Residual beta-cell function and islet cell antibodiesin diabetic children. Pediatr Res 1981; 15: 62–65.

18. Mustonen A, Knip M, Akerblom HK. An associationReferences between complement-fixing cytoplasmic islet cell anti-

bodies and endogenous insulin secretion in children with1. Lendrum R, Walker G, Gamble DR. Islet cell antibodies insulin-dependent diabetes mellitus. Diabetes 1983; 32:

in juvenile diabetes mellitus of recent onset. Lancet 1975; 743–747.I: 880–883. 19. Manna R, Salvatore M, Scueri F, et al. Negative correlation

2. Betterle C, Caretto A, Tiengo A, Trevisan A. Complement- between ICA persistence and beta-cell restoration afterfixing islet-cell antibodies in Type I diabetes in a IDDM diagnosis. Diabetes Res 1988; 9: 101–103.susceptible patients with autoimmune diseases. Lancet 20. Peig M, Gomis R, Ercilla G, Casamitjana R, Bottazzo GF,1980; I: 1418–1419. Pujol-Borrell R. Correlation between residual beta-cell

3. Mustonen A, Knip M, Huttunen N-P, Puukka R, Kaar M- function and islet cell antibodies in newly diagnosedL, Akerblom HK. Evidence of delayed beta-cell destruction Type I diabetes. Diabetes 1989; 38: 1396–1401.in Type 1 (insulin-dependent) diabetic patients with 21. Arslanian SA, Becker DJ, Atchison R, Rabin B, Eberhardtpersisting complement-fixing cytoplasmic islet cell anti- M, Cavender D, et al. Correlate of insulin antibodies inbodies. Diabetologia 1984; 27: 421–426. newly diagnosed children with insulin-dependent diabetes

4. Kolb H, Dannehl K, Grueneklee D, et al. Prospective before insulin therapy. Diabetes 1985; 34: 926–930.analysis of islet cell antibodies in children with Type 22. Heding LG. Radioimmunological determination of human1 (insulin-dependent) diabetes. Biabetologia 1988; 31 C-peptide in serum. Diabetologia 1975; 11: 541–548.189–194. 23. Bottazzo GF, Florin-Christiansen A, Doniach D. Islet5. Elamin A, Omer MIA, Tuvemo T. Islet-cell antibodies cell antibodies in diabetes mellitus with autoimuneand endogenous insulin secretion in Sudanese diabetic

polyendocrine deficiences. Lancet 1974; 11: 1279–1282.children. Diabetes Res Clin Pract 1992; 16: 91–96.24. Lernmark A, Molenaar JL, van Beers WM, Yamaguchi Y,6. Palmer JP, Asplin CM, Clemons P, et al. Insulin antibodies

Nagataki S, Ludvigsson J, Maclaren NK. The fourthin insulin-dependent diabetics before insulin treatment.international serum exchange workshop to standardizeScience 1983; 222: 1337–1339.cytoplasmic islet cell antibodies. Diabetologia 1992; 35:7. Wilkin T, Armitage M, Casey C, et al. Value of insulin70–76.autoantibodies as serum markers for insulin-dependent

25. Schmidli RS, Colman PG, Bonifacio E. Disease sensitivitydiabetes mellitus. Lancet 1985; I: 480–482.and specificity of 52 assays for glutamic acid decarboxy-8. Atkinson MA, Maclaren NK, Riley WJ, Winter WE, Fisklase antibodies. The second International GADab Work-DD, Spillar RP. Are insulin autoantibodies markers forshop. Diabetes 1995; 44: 636–640.insulin-dependent diabetes mellitus? Diabetes 1986; 35:

26. Tuomilehto J, Lounamaa R, Tuomilehto-Wolf E, et al.894–898.Epidemiology of childhood diabetes mellitus in Finland —9. Karjalainen J, Knip M, Mustonen A, Akerblom HK. Insulinbackground of a nationwide study of Type 1 (insulin-autoantibodies at the clinical manifestation of Type 1dependent) diabetes mellitus. Diabetologia 1992; 35:(insulin-dependent) diabetes — a poor predictor of clinical70–76.course and antibody response to exogenous insulin.

Diabetologia 1988; 31: 129–133.

537RESIDUAL BETA-CELL FUNCTION AND AUTOANTIBODIES