Cardiovascular and Metabolic Risk

O R I G I N A L A R T I C L E

Models for Predicting Type 1 Diabetes in

Siblings of Affected Children
SAMY MRENA, MD1 HANS K. ÅKERBLOM, MD, DMSC1 that most earlier surveys presenting pre-
SUVI M. VIRTANEN, MD, DMSC2,3,4,5 MIKAEL KNIP, MD, DMSC1,4 dictive models were based on relatively
PEKKA LAIPPALA, PHD3,5† THE CHILDHOOD DIABETES IN FINLAND selected populations (3–5). Accordingly,
PETRI KULMALA, MD, DMSC6 STUDY GROUP assessing predictive strategies in an uns-
MARJA-LEENA HANNILA, MSC2 elected sibling population is important
and clinically relevant.
The contribution of autoantibodies in
OBJECTIVE — To generate predictive models for the assessment of risk of type 1 diabetes assessment of type 1 diabetes risk devel-
and age at diagnosis in siblings of children with newly diagnosed type 1 diabetes. opment is well established at the group
level. It is also well known that HLA-
RESEARCH DESIGN AND METHODS — Cox regression analysis was used to assess
conferred genetic susceptibility and a de-
the risk of progression to type 1 diabetes, and multiple regression analysis was used to estimate
the age at disease presentation in 701 siblings of affected children. Sociodemographic, genetic, creased first-phase insulin response
and immunological variables were included in the analyses. Subanalyses were performed in a (FPIR) to intravenous glucose increase
group of 77 autoantibody-positive siblings with additional metabolic data. risk. Assessment of future risk of type 1
diabetes has two dimensions. First, there
RESULTS — A total of 47 siblings (6.7%) presented with type 1 diabetes during the 15-year is a need to have an estimate of the overall
observation period. Young age, an increasing number of detectable diabetes-associated autoan- risk for subsequent development of clini-
tibodies at initial sampling and of affected first-degree relatives, and HLA DR– conferred disease cal disease. Second, the family would like
susceptibility predicted progression to type 1 diabetes. In the subgroup of 77 autoantibody- to know how soon a high-risk sibling of
positive siblings, young age, HLA DR– conferred susceptibility, an increasing number of auto- the first affected child might progress to
antibodies, a reduced first-phase insulin response, and decreased insulin sensitivity in relation to
type 1 diabetes. We decided to establish a
first-phase insulin response were associated with increased risk of progression to type 1 diabetes.
Age at diagnosis was predicted by age, insulinoma-associated protein 2 antibody levels, and two-step predictive strategy to 1) identify
number of autoantibodies at initial sampling (R2 ⫽ 0.76; P ⬍ 0.001). In the smaller cohort of those siblings at highest risk for clinical
autoantibody-positive subjects, first-phase insulin response and HLA DR– conferred suscepti- disease and 2) assess the time frame
bility were additional predictors of age at diagnosis. within which a high-risk sibling will
likely present with overt type 1 diabetes.
CONCLUSIONS — Information on autoantibody status and levels, HLA-conferred disease Our aim was to generate clinically appli-
susceptibility, and insulin secretion and sensitivity seems to be useful in addition to age and cable predictive models for risk assess-
family history of type 1 diabetes when assessing risk of progression to type 1 diabetes and time ment of clinical diabetes in unaffected
to diagnosis in siblings of children with newly diagnosed type 1 diabetes. siblings of newly diagnosed type 1 dia-
betic children.
Diabetes Care 29:662– 667, 2006

RESEARCH DESIGN AND

S
ince the 1970s, several studies have design predictive models for type 1 dia-
METHODS — The study population
indicated that HLA-conferred dis- betes, integrating sociodemographic, ge-
was derived from the nationwide Child-
ease susceptibility and autoantibod- netic, immunological, and metabolic
hood Diabetes in Finland (DiMe) study
ies are useful in the prediction of type 1 markers, and test their utility in predic-
(6). The observation of the siblings was
diabetes among first-degree relatives of af- tion of type 1 diabetes in siblings of dia-
initiated shortly after the proband was di-
fected patients (1,2). Our purpose was to betic children. This approach is unique in
agnosed with type 1 diabetes. Blood sam-
● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ples were taken at intervals of 3– 6
From the 1Hospital for Children and Adolescents, University of Helsinki, Helsinki, Finland; the 2Department months during the first 2 years and at 6-
of Epidemiology and Health Promotion, National Public Health Institute, Helsinki, Finland; the 3Tampere to 12-month intervals during the follow-
School of Public Health, University of Tampere, Tampere, Finland; the 4Department of Pediatrics, Tampere ing 2 years. Autoantibody-positive sib-
University Hospital, Tampere, Finland; the 5Research Unit, Tampere University Hospital, Tampere, Finland; lings were invited for further testing at an
and the 6Department of Pediatrics, University of Oulu, Oulu, Finland.
Address correspondence and reprint requests to Mikael Knip, MD, DMSc, Hospital for Children and
interval of 6 –12 months to the end of
Adolescents, University of Helsinki, P.O. Box 281, FI-00029 HUCH, Helsinki, Finland. E-mail: 2002, whereas the testing of autoanti-
mikael.knip@hus.fi. body-negative siblings ended after fol-
Received for publication 1 May 2005 and accepted in revised form 28 November 2005. low-up for the first 4 years. Only
†Deceased. autoantibody data from the initial sam-
Abbreviations: FPIR, first-phase insulin response; GADA, GAD antibody; HOMA-IR, homeostasis model
assessment of insulin resistance; IA-2, insulinoma-associated protein 2; IA-2A, IA-2 antibody; IAA, insulin pling were taken into account here. All
autoantibody; ICA, islet cell antibody; IVGTT, intravenous glucose tolerance test; ROC, receiver operating the siblings were observed for progres-
characteristic. sion to type 1 diabetes up to the end of
A table elsewhere in this issue shows conventional and Système International (SI) units and conversion year 2002, i.e., for an average period of
factors for many substances.
© 2006 by the American Diabetes Association.
15.0 years (range 13.7–16.3). Observa-
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby tion of the siblings who progressed to
marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. clinical disease ended at diagnosis, which

662 DIABETES CARE, VOLUME 29, NUMBER 3, MARCH 2006


was based on clinical symptoms and an
increased random blood glucose concen-
tration (⬎10 mmol/l) or elevated fasting
(⬎6.7 mmol/l) or random blood glucose
on two occasions in the absence of symp-
toms (7).
Altogether, at least one blood sample
was available from 758 siblings at the time
of diagnosis in the index case. The present
study cohort included all siblings with at
least one serum sample for autoantibody
analyses and data on HLA class II typing
available. This resulted in a total series of
701 siblings with a mean age of 9.9 years
(range 0.8 –19.7). A total of 217 siblings
were HLA DR3/DR4 heterozygous, 334
carried the DR4/non-DR combination, 97
carried the DR3/non-DR4 combination,
and 53 had neither DR3 nor DR4. A total
of 93 siblings tested positive for at least
one diabetes-associated autoantibody, 49
being positive for a single autoantibody
reactivity and 44 for multiple (two or
more) antibodies. A total of 60 siblings
tested positive for islet cell antibodies
(ICAs), 20 for insulin autoantibodies
(IAAs), 55 for GAD antibodies (GADAs),
and 36 for insulinoma-associated protein
2 (IA-2) antibodies (IA-2As) at initial
sampling. An intravenous glucose toler-
ance test (IVGTT) was performed in 77 of
the 93 antibody-positive children.
Disease-associated autoantibodies
ICAs were determined with conventional
immunofluorescence (8). The sensitivity
of the ICA assay was 100% and the spec-
ificity 98% (9). IAAs, GADAs, and IA-2As
were analyzed with specific radiobinding
assays as described (10). The sensitivity of
the IAA assay was 78% and the specificity
100% in the proficiency-testing program.
The disease sensitivity of the GADA assay
was 79% and the specificity 97% based on
the 1995 Multiple Autoantibody Work-
shop (11). The corresponding character-
istics of the IA-2A assay were 62 and 97%,
respectively.
IVGTT and the homeostasis model
assessment of insulin resistance
The IVGTTs were preformed as described
(12). Blood samples were taken before the
glucose infusion and at 1, 3, 6, 10, 20, 30,
40, 50, and 60 min thereafter. Serum in-
sulin concentrations were measured ra-
dioimmunologically (13), and blood
glucose levels were determined with the
glucose oxidase method (14). The sum of
the insulin concentrations at 1 and 3 min
was defined as the FPIR to glucose. The
glucose disappearance rate (Kg) was ex-
DIABETES CARE, VOLUME 29, NUMBER 3, MARCH 2006
pressed as the percentage decrease in
blood glucose per minute (%/min). FPIR
levels ⬍45 mU/l, a level that represents
the third percentile of FPIR values in
healthy control subjects (12), and Kg val-
ues ⬍1.30%/min were considered abnor-
mal. The homeostasis model assessment
of insulin resistance (HOMA-IR) index
was calculated based on the conventional
formula: HOMA-IR ⫽ fasting glucose
(mmol/l) ⫻ fasting insulin (mU/l)/22.5,
as described previously (15,16). The con-
ventional index correlated strongly (r ⫽
0.99) with the newer HOMA computer
model (17). Insulin resistance was related
to insulin secretion by calculating the
HOMA-IR/FPIR ratio.
Genetics
HLA DR typing was performed by con-
ventional HLA serology as described (18).
The HLA-conferred susceptibility was
graded into four categories: no risk, HLA
non-DR3/non-DR4; low risk, HLA DR
3/non-DR4; moderate risk, HLA DR4/
non-DR3; and high risk, HLA DR3/DR4.
Data handling and statistical
analyses
The data were evaluated statistically using
cross-tabulation and ␹2 statistics for fre-
quencies. Variables with a normal distri-
bution were compared with the t test. The
Mann-Whitney U test and nonparametric
correlation analysis were applied when
analyzing variables with a skewed distri-
bution. The Cox regression analysis was
used to assess factors associated with the
risk of progression to type 1 diabetes,
whereas multiple linear regression analy-
sis was applied for the estimation of vari-
ables related to the age at diagnosis. The
data initially included in the analysis of
the total series of 701 siblings comprised
the following potential predictors: age at
first sampling, sex, HLA-conferred dis-
ease susceptibility (two or four catego-
ries), degree of HLA identity with the
index case, initial autoantibody positivity
and levels (ICAs, IAAs, GADAs, and IA-
2As), age at diagnosis and sex of the index
case, the number of children in the family,
and the number of first-degree relatives
affected by type 1 diabetes. In the smaller
series comprising 77 autoantibody-
positive siblings who had undergone an
IVGTT, FPIR, Kg, HOMA-IR, and the
HOMA-IR/FPIR ratio (natural logarithm
transformed due to skewed distribution),
were also included in the analyses. Cox
regression analyses were performed with
the Stata statistical software package ver-
Mrena and Associates
sion 8.0 (Stata, College Station, TX) and
other statistical tests with the SPSS 11
software (SPSS, Chicago, IL). The propor-
tionality of the hazards was checked by
using log-cumulative hazard plots (Stata).
RESULTS
Progression to clinical diabetes
A total of 47 siblings (6.7%, 95% CI 5.0 –
8.8%) presented with clinical type 1 dia-
betes during the 15-year observation
period. The mean age at the time of diag-
nosis was 13.9 years (range 1.4 –28.4). Of
the 47 progressors, 38 tested initially pos-
itive for at least one diabetes-associated
autoantibody. Seven initially autoanti-
body-negative siblings seroconverted to
antibody positivity before diagnosis. The
risk of developing type 1 diabetes in the
total series was associated with the age at
first sampling, HLA DR– conferred dis-
ease susceptibility, the number of initially
detectable diabetes-associated autoanti-
bodies, and the number of affected family
members (Table 1). Among the 77 au-
toantibody-positive siblings with meta-
bolic data available, the age of the sibling,
HLA DR– conferred susceptibility, the
number of disease-associated autoanti-
bodies, the FPIR, and the HOMA-IR/FPIR
ratio turned out to be significant predic-
tors of progression to type 1 diabetes (Ta-
ble 1).
The individual prognostic risk index
Based on the Cox regression model, we
calculated an individual prognostic risk
index for each subject. We then per-
formed a receiver operating characteristic
(ROC) analysis to define a cutoff index
leading to the best separation between
progressors and nonprogressors. The cut-
off index based on the total series was
judged to be 0.25, resulting in a sensitiv-
ity of 78.7%, a specificity of 95.7%, and a
positive predictive value of 56.9% for
type 1 diabetes (Fig. 1). There were alto-
gether 65 of 701 (9.3%) siblings with a
prognostic index exceeding the cutoff
value. Of these 65 siblings, 37 presented
with clinical type 1 diabetes. The remain-
ing 636 siblings (90.7%) had a prognostic
risk index below the cutoff value of 0.25,
and only 10 of them (1.6%) developed
clinical type 1 diabetes. We compared the
siblings below the cutoff value presenting
with type 1 diabetes with those siblings
who remained unaffected to assess factors
predisposing to overt type 1 diabetes
among these “protected” children. The
progressors had higher GADA and IA-2A
663
Prediction of type 1 diabetes in siblings
levels than the siblings who remained
nondiabetic. In addition, they had ini-
tially more autoantibodies detectable and
tended to be DR3/DR4 heterozygous
more frequently than the unaffected sib-
lings (data not shown). Among those who
presented with type 1 diabetes, the sib-
lings with an index in excess of 0.25 had a
shorter duration of the preclinical period
than those with a lower index (mean
4.9 ⫾ 4.0 vs. 8.8 ⫾ 3.3 years; P ⫽ 0.007).
The prognostic index was inversely re-
lated to the duration of the preclinical pe-
riod (r s ⫽ ⫺0.40; P ⫽ 0.006). The
predictive characteristics of a prognostic
index ⬎0.25 are compared with those of
positivity for multiple (two or more) au-
toantibodies in Table 2.
Table 1—Cox regression analysis for the estimation of risk for progression to clinical disease among 701 siblings of children with recently
diagnosed type 1 diabetes and in a subgroup of 77 siblings with metabolic data available
Baseline characteristics
Age at first sampling (years)
Sex (boys vs. girls)
HLA-DR allele (high and moderate vs. low and
decreased risk)
Antibody positivity (for diabetes-related
autoantibodies ⱖ2 vs. 0–1)
Number of affected first-degree relatives (at the
time of diagnosis in the index case ⱖ1 vs. 0)
FPIR (decreased vs. normal)
Kg (decreased vs. normal)
HOMA-IR
HOMA-IR/FPIR (natural logarithm)
Data are hazard ratios (95% CI). *Adjusted for all the other variables in column 2. †Adjusted for all the other variables in column 4.
664
Figure 1—ROC analysis
of the individual prognos-
tic risk index. The optimal
cutoff index based on the
total series was considered
to be 0.25, resulting in a
sensitivity of 78.7%, a
specificity of 95.7%, and a
positive predictive value of
56.9% for type 1 diabetes.
Prediction of age at diagnosis of type
1 diabetes
The age at disease presentation was most
effectively predicted with a linear regres-
sion model including age, IA-2A levels,
and the number of initially detectable au-
toantibodies. This model explained
⬃76% of the variation in age at diagnosis
(Table 3). When we applied this model on
the 65 siblings with a prognostic index of
⬎0.25, the observed age at clinical pre-
sentation was within the CI of the esti-
mated age in 18 subjects of the 37
progressors (49%), whereas all 28 non-
progressors were predicted to present
with clinical disease before the end of the
follow-up period.
The second model for the estimation
All siblings (n ⫽ 701)
Crude Adjusted*
0.83 (0.77–0.91) 0.76 (0.68–0.84)
0.88 (0.49–1.55)
7.2 (2.8–18.2) 2.9 (1.1–7.6)
55.9 (30.0–104.1) 54.1 (27.8–105.0)
3.3 (1.6–6.6) 3.2 (1.6–6.6)
DIABETES CARE, VOLUME 29, NUMBER 3, MARCH 2006
of age at diagnosis including 77 siblings
with metabolic data were based on the age
of the sibling, the initial IA-2A level, HLA
DR– conferred risk, and the initial FPIR
value. This model explained 83% of the
variation in age at diagnosis (Table 3). The
application of this model on the 33 sib-
lings with a prognostic risk index exceed-
ing the cutoff value showed that the
observed age at diagnosis was within the
CI in all but 1 of the 25 progressors, but
again all nonprogressors were predicted
to present with diabetes before the end of
the observation period.
CONCLUSIONS — Although no ef-
fective modality for preventing or delay-
ing progression to clinical type 1 diabetes
has been recognized so far for clinical use
in subjects at increased disease risk, there
is still a rationale for establishing predic-
tive models capable of identifying those
individuals who are at the highest risk for
developing type 1 diabetes and for esti-
mating disease risk on an individual basis.
Such a model will inevitably be needed as
soon as the first treatment option modu-
lating the pre-diabetic disease process has
evolved. From a family point of view, the
most urgent need is a reliable assessment
of diabetes risk in unaffected siblings of
children with newly diagnosed type 1 di-
abetes. Accordingly, the predictive model
has to be based on information available
or possible to generate within a limited
time period close to the time of diagnosis
in the index case. We have suggested that
positivity for two or more autoantibodies
seems to reflect a progressive irreversible
Siblings positive for autoantibodies at
baseline (n ⫽ 77)
Crude Adjusted†
0.87 (0.79–0.96) 0.78 (0.68–0.89)
0.84 (0.41–1.7)
7.7 (1.8–32.5) 5.7 (1.3–25.2)
6.5 (2.9–14.7) 3.6 (1.5–8.8)
2.0 (0.78–5.3)
10.9 (4.7–25.4) 4.7 (1.9–11.6)
1.8 (0.83–3.9)
1.0 (0.84–1.3)
3.1 (1.8–5.3) 2.4 (1.2–5.0)

Table 2—Predictive characteristics of a prognostic index >0.25 and positivity for multiple
(two or more) autoantibodies
Prognostic index Two or more
⬎0.25 Autoantibodies Difference (95% CI)
Sensitivity (%) 78.7 68.1 10.6 (⫺7.1 to 28.4)
Specificity (%) 95.7 98.2 2.5 (0.6 to 4.3)
Positive predictive value (%) 56.9 72.7 15.8 (⫺2.0 to 33.6)
Negative predictive value (%) 98.4 97.7 0.7 (⫺0.8 to 2.2)
autoimmune process, whereas positivity autoantibody positivity, this difference
for only one type 1 diabetes–associated remained nonsignificant, whereas the lat-
autoantibody appears to reflect harmless ter marker had significantly higher speci-
and even reversible ␤-cell autoimmunity ficity (Table 2). Only when analyzing
(12,19,20). Now we have attempted to individuals who progressed to type 1 di-
further refine the predictive model by in- abetes before the age of 16 years, the sen-
tegrating all data available on siblings of sitivity of the prognostic index (93%) was
affected children close to the time of dis- higher than that of multiple autoantibody
ease presentation in the index case. Based positivity (73%, difference 20%, 95% CI
on our previous experience, we decided 2–38%). The observed predictive charac-
to aim at a two-stage model. The purpose teristics of the ROC cutoff value might,
of the first step was to assess overall risk however, be too optimistic, since they are
for progression to clinical disease; the sec- calculated based on the data on which the
ond step was to estimate the likely age at model was built. The inverse correlation
diagnosis of diabetes. We performed the between the prognostic index and the du-
risk assessment both in the total cohort ration of the preclinical period indicates
including all available siblings and in a that a high index is a marker of a particu-
smaller series including those siblings larly aggressive disease process.
who had additional metabolic markers In the smaller series with metabolic
available. data available, we found that both a re-
The strongest predictive model for duced FPIR and an increased HOMA-IR/
progression to clinical disease in the total FPIR ratio reflecting a reduced insulin
series included the age of the sibling at sensitivity relative to insulin secretion was
first sampling, HLA DR– conferred sus- associated with an enhanced disease risk.
ceptibility, number of initially detectable It is well established that a reduced early
autoantibodies, and the number of first- insulin response is associated with a high
degree relatives with type 1 diabetes. We risk for progression to type 1 diabetes
used the multivariate model to estimate (3,4,12,21), whereas the observation that
the individual risk of a sibling for progres- an increased HOMA-IR/FPIR ratio con-
sion to overt type 1 diabetes by calculat- fers increased risk has been implicated by
ing an individual prognostic risk index in only one recent study (22). Fourlanos et
each sibling based on the Cox regression al. (22) reported that autoantibody-
model. The optimal cutoff point was con- positive first-degree relatives, who pro-
sidered to be 0.25 based on a ROC anal- gressed rapidly to type 1 diabetes, were
ysis. Because the sensitivity was 79%, and characterized by enhanced insulin resis-
the specificity of the model was as high as tance for their level of insulin secretion.
96%, the prognostic risk index may pro- Taken together, these observations sug-
vide a means for estimating individual gest that the manifestation of clinical dis-
risk. Of the 65 siblings with a risk index ease is affected by the balance between the
exceeding the cutoff value of 0.25, 37 insulin secretory capacity and peripheral
(56.9%) developed type 1 diabetes. A to- insulin sensitivity.
tal of 44 siblings tested initially positive When estimating the likely age at di-
for multiple (two or more) diabetes- agnosis in the total study cohort, we ob-
associated autoantibodies in the total se- served that age at initial sampling and the
ries, and 32 of these progressed to type 1 number of autoantibodies initially detect-
diabetes. Accordingly, the sensitivity of able were variables in common with the
this risk marker was 68%, the specificity model predicting risk of progression to
98%, and the positive predictive value type 1 diabetes. The initial IA-2A level
73%. Although the prognostic index was the third parameter included in the
tended to be more sensitive than multiple model predicting age at disease presenta-
DIABETES CARE, VOLUME 29, NUMBER 3, MARCH 2006
Mrena and Associates
tion. IA-2As have been reported to appear
in most cases as the last autoantibody dur-
ing the pre-diabetic disease process
(23,24), and they have also been observed
to be the most predictive autoantibodies
among first-degree relatives (10,25). The
present observation stresses the role of IA-
2As as predictive markers. The model was
able to explain close to 80% of the varia-
tion in the age at diagnosis. The lack of
HLA-conferred disease susceptibility
from the model indicates that the pace of
the pre-diabetic disease process is mainly
regulated by factors other than the HLA
class II genes (20). Approximately half of
the observed ages at diagnosis in the 37
progressors were within the range of the
CI of the estimations by this model among
the 65 siblings with a prognostic index
exceeding the cutoff value of 0.25,
whereas all 28 nonprogressors were pre-
dicted to present with type 1 diabetes be-
fore the end of the observation period.
Accordingly, this model for the prediction
of age at diagnosis did not work precisely
among the high-risk siblings.
The analysis of the series comprising
77 siblings with metabolic data available
resulted in a model by which it was pos-
sible to explain ⬃83% of the variation in
the age at diagnosis. This model included,
in addition to age and IA-2A level at initial
sampling, HLA DR– conferred disease
susceptibility and early insulin response
to intravenous glucose. It is intriguing
that the genetic predisposition defined by
HLA genes is included in this model but
not in the model based on the total study
cohort. One explanation could be that
there is a strong correlation between the
number of autoantibodies present and
HLA-defined disease risk and that the in-
clusion of the former in the first model
resulted in the exclusion of the latter from
that model. The observation that there is
an association between initial insulin se-
cretory capacity assessed in terms of FPIR
and age at diagnosis is consistent with our
previous finding of a relation between the
initial early insulin response and the time
lag from first testing to type 1 diabetes
presentation (12). Neither the fasting
HOMA-IR index nor the HOMA-IR/FPIR
ratio had any significant impact on age at
diagnosis, suggesting that a reduced insu-
lin sensitivity is a stronger determinant of
disease risk than of progression rate to
clinical diabetes. All but one of the ob-
served ages at diagnosis were within the
CIs of the estimated ages at disease pre-
sentation in the smaller series. This was at
least partly a consequence of a substan-
665
Prediction of type 1 diabetes in siblings
Table 3—Multiple regression analysis for the estimation of age at diagnosis in 47 siblings of
affected children who contracted type 1 diabetes and in the 31 siblings with metabolic data
available
Siblings of affected children with
type 1 diabetes
Age at first sampling
IA-2A level at first sampling
Initial number of autoantibodies
Intercept
Fit of the model: R2 ⫽ 0.76; F ⫽ 44.7;
P ⬍ 0.001
Siblings with metabolic data available
Age at first sampling
IA-2A level at first sampling
HLA DR–conferred risk*
FPIR
Intercept
Fit of the model: R2⫽ 0.83; F ⫽ 32.1;
P ⬍ 0.001
*Grading 0 (low risk) to 3 (high risk).
tially greater SD in this model than that in
the model based on the total study cohort.
Again, the eight nonprogressors were pre-
dicted to develop type 1 diabetes before
the end of the observation period, ques-
tioning the utility of this model for pre-
dicting age at diagnosis even when
metabolic data are available.
Our work generated a novel approach
for predicting type 1 diabetes with a mul-
tivariate model including the HOMA-IR/
FPIR ratio as a measure of relative insulin
resistance. The Cox regression model de-
vised seemed to offer a feasible strategy
for the identification of those siblings of
children with newly diagnosed type 1 di-
abetes who will most probably progress to
clinical disease. We think that this kind of
information may be useful when the par-
ents of a child with recently diagnosed
diabetes are informed about the risk of
clinical disease in their other children.
The model for predicting age at diagnosis
appeared to work well or satisfactorily
among the true progressors but poorly
among those who did not present with
type 1 diabetes. Our results suggest that a
short IVGTT providing fasting glucose
and insulin concentrations and an esti-
mate of the early insulin response to in-
travenous glucose provides additional
data that improve the accuracy of both the
risk and time estimates. The HOMA-IR
index appears to be a useful predictive
marker, since a high HOMA-IR/FPIR ratio
was observed to be associated with in-
creased risk for progression to type 1 di-
666
atives: the ICARUS data set: Islet Cell An-
tibody Register Users Study. Diabetes 45:
1720 –1728, 1996
4. Eisenbarth GS, Moriyama H, Robles DT,
Liu E, Yu L, Babu S, Redondo M, Gottlieb
␤ Coefficient SE P P, Wegmann D, Rewers M: Insulin auto-
immunity: prediction/precipitation/pre-
vention type 1A diabetes. Autoimmun Rev
1:139 –145, 2002
1.379 0.123 ⬍0.001 5. Achenbach P, Warncke K, Reiter J, Nas-
⫺0.039 0.012 0.003 erke HE, Williams AJ, Bingley PJ, Bonifa-
⫺0.649 0.415 0.125 cio E, Ziegler AG: Stratification of type 1
5.237 1.349 diabetes risk on the basis of islet autoan-
tibody characteristics. Diabetes 53:384 –
392, 2004
6. Tuomilehto J, Lounamaa R, Tuomilehto-
1.118 0.131 ⬍0.001 Wolf E, Reunanen A, Virtala E, Kaprio EA,
Akerblom HK, the Childhood Diabetes in
⫺0.021 0.011 0.07
Finland Study Group: Epidemiology of
1.972 0.721 0.011 childhood diabetes mellitus in Finland:
0.0658 0.015 ⬍0.001 background of a nationwide study of type
⫺2.816 2.260 1 (insulin-dependent) diabetes mellitus.
Diabetologia 35:70 –76, 1992
7. World Health Organization: Diabetes Mel-
litus: Report of a WHO Study Group. Ge-
neva, World Health Org., 1980 (Tech.
Rep. Ser., no. 727)
8. Bottazzo GF, Florin-Christensen A, Doni-
abetes. These refined predictive models ach D: Islet-cell antibodies in diabetes
may be used for the identification of those mellitus with autoimmune polyendocrine
individuals who would most conspicu- deficiencies. Lancet 2:1279 –1283, 1974
ously benefit from preventive measures 9. Lernmark A, Molenaar JL, van Beers WA,
aimed at stopping the pre-diabetic disease Yamaguchi Y, Nagataki S, Ludvigsson J,
process. Maclaren NK: The Fourth International
Serum Exchange Workshop to standard-
ize cytoplasmic islet cell antibodies: the
Acknowledgments — This study was sup- Immunology and Diabetes Workshops
ported by grants from the Juvenile Diabetes and Participating Laboratories. Diabetolo-
Foundation International (grant 197032), the gia 34:534 –535, 1991
Finnish Diabetes Research Foundation, the 10. Kulmala P, Savola K, Petersen JS, Vä-
Medical Research Council, the Academy of häsalo P, Karjalainen J, Löppönen T, Dyr-
Finland (grant 26109), the Novo Nordisk berg T, Åkerblom HK, Knip M, the
Foundation, and the Maija and Matti Vaskio Childhood Diabetes in Finland Study
Foundation. The other support of the DiMe Group: Prediction of insulin-dependent
study and the composition of the DiMe Study diabetes mellitus in siblings of children
Group are listed in ref. 18. with diabetes. J Clin Invest 101:327–336,
We thank Sirpa Anttila, Susanna Heikkilä, 1998
Erik Mrena, Riitta Päkkilä, and Päivi Salmijärvi 11. Verge CF, Stenger D, Bonifacio E, Colman
for technical assistance. PG, Pilcher C, Bingley PJ, Eisenbarth GS:
Combined use of autoantibodies (IA-2
autoantibody, GAD autoantibody, insulin
References autoantibody, cytoplasmic islet cell anti-
1. Singal DP, Blajchman MA: Histocompati- bodies) in type 1 diabetes: Combinatorial
bility (HL-A) antigens, lymphocytotoxic Islet Autoantibody Workshop. Diabetes
antibodies and tissue antibodies in pa- 47:1857–1866, 1998
tients with diabetes mellitus. Diabetes 22: 12. Mrena S, Savola K, Kulmala P, Akerblom
429 – 432, 1973 HK, Knip M, the Childhood Diabetes in
2. Bonifacio E, Genovese S, Braghi S, Bazzi- Finland Study Group: Staging of preclin-
galuppi E, Lampasona V, Bingley PJ, ical type 1 diabetes in siblings of affected
Rogge L, Pastore MR, Bognetti E, Bottazzo children. Pediatrics 104:925–930, 1999
GF, Gale EAM, Bosi E: Islet autoantibody 13. Herbert V, Lau KS, Gottlieb CW, Bleicher
markers in IDDM: risk assessment strate- SJ: Coated charcoal immunoassay of insu-
gies yielding high sensitivity. Diabetologia lin. J Clin Endocrinol Metab 25:1375–
38:816 – 822, 1995 1384, 1965
3. Bingley PJ: Interactions of age, islet cell 14. Hjelm M: Enzymatic determination of
antibodies, insulin autoantibodies, and hexoses in blood and urine. Scand J Clin
first-phase insulin response in predicting Lab Invest Suppl 18:85–98, 1966
risk of progression to IDDM in ICA⫹ rel- 15. Cutfield WS, Jefferies CA, Jackson WE,
DIABETES CARE, VOLUME 29, NUMBER 3, MARCH 2006

Robinson EM, Hofman PL: Evaluation of
HOMA and QUICKI as measures of insu-
lin sensitivity in prepubertal children. Pe-
diatr Diabetes 4:119 –125, 2003
16. Matthews DR, Hosker JP, Rudenski AS,
Naylor BA, Treacher DF, Turner RC: Ho-
meostasis model assessment: insulin re-
sistance and beta-cell function from
fasting plasma glucose and insulin con-
centrations in man. Diabetologia 28:412–
419, 1985
17. Wallace TM, Levy JC, Matthews DR: Use
and abuse of HOMA modeling. Diabetes
Care 27:1487–1495, 2004
18. Tuomilehto-Wolf E, Tuomilehto J, Ce-
paitis Z, Lounamaa R, the Childhood Di-
abetes in Finland Study Group: New
susceptibility haplotype for type 1 diabe-
tes. Lancet 2:299 –302, 1989
19. Mrena S, Savola K, Kulmala P, Akerblom
HK, Knip M, the Childhood Diabetes in
Finland Study Group: Natural course of
preclinical type 1 diabetes in siblings of
DIABETES CARE, VOLUME 29, NUMBER 3, MARCH 2006
affected children. Acta Paediatr 92:1403–
1410, 2003
20. Mrena S, Savola K, Kulmala P, Reijonen
H, Ilonen J, Akerblom HK, Knip M, the
Childhood Diabetes in Finland Study
Group: Genetic modification of risk as-
sessment based on staging of preclinical
type 1 diabetes in siblings of affected chil-
dren. J Clin Endocrinol Metab 88:2682–
2689, 2003
21. Chase HP, Cuthbertson DD, Dolan LM,
Kaufman F, Krischer JP, Schatz DA,
White NH, Wilson DM, Wolfsdorf J, the
Diabetes Prevention Trial–Type 1 Study
Group: First-phase insulin release during
the intravenous glucose tolerance test as a
risk factor for type 1 diabetes. J Pediatr
138:244 –249, 2001
22. Fourlanos S, Narendran P, Byrnes GB,
Colman PG, Harrison LC: Insulin resis-
tance is a risk factor for progression to
type 1 diabetes. Diabetologia 47:1661–
1667, 2004
Mrena and Associates
23. Ziegler AG, Hummel M, Schenker M,
Bonifacio E: Autoantibody appearance
and risk for development of childhood di-
abetes in offspring of parents with type 1
diabetes: the 2-year analysis of the Ger-
man BABYDIAB study. Diabetes 48:460 –
468, 1999
24. Kimpimaki T, Kulmala P, Savola K,
Kupila A, Korhonen S, Simell T, Ilonen J,
Simell O, Knip M: Natural history of beta-
cell autoimmunity in young children with
increased genetic susceptibility to type 1
diabetes recruited from the general popu-
lation. J Clin Endocrinol Metab 87:4572–
4579, 2002
25. Savola K, Bonifacio E, Sabbah E, Kulmala
K, Vähäsalo P, Karjalainen J, Tuomilehto-
Wolf E, Meriläinen J, Åkerblom HK, Knip
M, the Childhood Diabetes in Finland
Study Group: IA-2 antibodies: a sensitive
marker of IDDM with clinical onset in
childhood and adolescence. Diabetologia
41:424 – 429, 1998
667