Microsoft Word search phase 3 Title Page Amendment

Temporal Trends in the Incidence of Childhood Diabetes 
4.2.3.1.
 
Type 1 Diabetes 
The majority of epidemiological data on T1D are based on a clinical definition 
including physician diagnosis of diabetes and daily insulin injections 
(36)
.  In addition, 
most studies have limited the age range of populations to < 14 years, to avoid 
misclassification of diabetes type.  It has been assumed, but not confirmed via 
measured diabetes autoantibodies (DA), that “type 1” diabetes is autoimmune 
mediated diabetes.  SEARCH for Diabetes in Youth is unique in that it extends the 
surveillance effort to youth age < 20 years, and it collects data on DA close to the 
time of diagnosis, to validate the clinical assessment of diabetes type.  
The majority of epidemiological data on T1D are based on data on populations of 
European origin.  SEARCH demonstrated that in 2002-2003 the incidence of T1D 
was highest in non-Hispanic whites (NHW), followed by African Americans (AA) 
and Hispanics, and it was lowest in Asian/Pacific Islanders (API) and American 
Indians (AI) 
(5)
.  SEARCH for Diabetes in Youth is the only registry effort to include 
a comprehensive assessment of T1D burden and risk across all major racial/ethnic 
groups. 
4.2.3.1.1.
 
Incidence Trends 
Most 
(10, 37 - 41)
 but not all 
(42 - 46)
 population-based registries showed an increasing 
incidence of T1D over time.  An updated report from the DIAMOND project 
examined the trends in incidence of T1D from 1990-1999 in 114 populations from 
57 countries.  Based on 43,013 cases of T1D from a study population of 84 
million children ≤14 years 
(2)
, the average annual increase in incidence over this 
time period was 2.8% (95% CI 2.4-3.2%).  Similarly, the EURODIAB study, a 
large European survey including 
20 population-based registries in 17 countries
 
showed a 3.2% (95% CI 2.7-3.7) annual increase for the period 1989-1998 
(4)
, and 
a more recent 3.9% (95% CI 3.6-4.2) 
increase
 from 1989-2003 
(8)
.  Interestingly, 
the observed incidence rates confirmed, and in fact exceeded, the incidence 
predicted for 2010 by earlier projections 
(47)
.  In EURODIAB 
(8)
, estimates of the 

---

Section 4A - Study Objectives/Background and Significance (Phase 3 - 11/2010) 
Section 4A - Page 7 
 Registry 
Study
 
 
rates of increase were highest in the youngest age-group [
5.4% (4.8–6.1) 
for 
children age 0-4 years]. 
Recent data from the U.S., where registry efforts have been less coordinated, 
suggest similar trends.  While the U.S. stood apart from other parts of the world in 
reporting a stable incidence of childhood T1D in the 1970’s through the 1990s 
(48)
, 
SEARCH recently reported that the 2002-05 incidence of T1D in NHW youth 
aged ≤ 14 years was 27.5 per 100 000 per year 
(49)
 a rate that exceeds the 
incidence predicted for 2010 from older data from Allegheny County 
(47)
.  Using 
data from the Colorado IDDM registry and the SEARCH-Colorado site the 
incidence of T1D was shown to increase in youth age ≤ 17 years over the past 3 
decades 
(9)
.  During a 26 year period, the incidence of T1D increased by 2.3% 
(95% CI 1.6-3.1) per year and was much higher than predicted from earlier 
Colorado data 
(47)
.  Of note, the increase was significant for both NHW (2.7%; 
95% CI 1.9 - 3.6 per year, P < 0.0001) and Hispanic youth (1.6%; 0.2-3.1 per 
year, P < 0.013).  Similar to the EURODIAB data, in Colorado, the increase in 
incidence was highest among the 0- to 4-year age-group (3.5%; 95% CI 2.1-4.9 
per year).  Additional suggestions of increasing incidence of T1D come from 
registries in Philadelphia 
(50, 51)
, Chicago 
(52)
 and Allegheny County 
(53)
, reporting 
mainly an increase among AA, but also Hawaii, reporting a four-fold overall 
increase (1980 to 1989) 
(54)
.  
4.2.3.2.
 
Trends in Genetic Susceptibility to T1D 
Genetic susceptibility plays a large role in T1D with the human leukocyte antigen 
(HLA) genotypes (DR and DQ genes) explaining approximately 40-50% of T1D risk 
(55)
.  The genetic variation can explain the variation in incidence across racial/ethnic 
groups, but it is unlikely to explain the rapid increase in the incidence of T1D.  
Recent studies have suggested that the contribution of high risk HLA genotype DR3,4 
has been relatively stable or has decreased over time 
(56 – 58)
.  Of note, one of these 
studies was conducted in Colorado, based on the prior Colorado IDDM and the 
current SEARCH data, and found a significant decrease in the proportion of cases of 
T1D with the high risk HLA genotype in the last two decades among both NHW and 
Hispanic participants 
(59)
.  These data suggest that the increase in T1D over the past 
half century is likely not due to increased incidence among those at the highest 
genetic risk in the HLA region, and must rather be explained by an increase in 
environmental factors, increased penetrance of low/moderate HLA genotypes or other 
genetic loci, or interactions between environmental risk factors and non-HLA genes. 

---

Section 4A - Study Objectives/Background and Significance (Phase 3 - 11/2010) 
Section 4A - Page 8 
 Registry 
Study
 
 
4.2.3.3.
 
Trends in Clinical Presentation