 
|
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Journal of the New Zealand Medical Association,
21-September-2007, Vol 120 No 1262
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Glycaemic control and antibody status among Waikato,
New Zealand patients with newly diagnosed Type 1 diabetes
Doron Hickey, Grace Joshy, Peter Dunn, David Simmons, Ross
Lawrenson
The classification of children and young adults presenting
with symptoms of diabetes mellitus has become more difficult with the increase
in Type 2 diabetes in young patients1
(including in New Zealand).2
Type 2 diabetes among children and young adults has
increased due to their increasing obesity and associated insulin
resistance.1 Type 1 diabetes is categorised as
either being positive or negative for various auto-antibodies related to
pancreatic function.3 Although most patients
with Type 1 diabetes are autoantibody positive, ethnicity confers notable
differences and may make confirmation of Type 1 diabetes more difficult.
Studies have shown that up to 90% of those of Northern
European origin have raised levels of at least one antibody at diagnosis 4
whilst they are less frequently found in black Africans or African
Americans.5–8 In the last 10 years it has
become routine to measure anti-GAD and anti-IA2 antibodies to establish the type
of diabetes in a given patient.
While such measurements remain imperfect diagnostic tools,
the results are of use in the management of individual patients. It has not been
established whether the actual titres of anti-GAD or anti-IA2 antibodies at
diagnosis have prognostic implications although the presence of anti-GAD is
believed to be indicative of beta-cell
destruction.9
Admission to hospital with DKA is a serious and potentially
life-threatening situation. Whilst it is rare in children, it is an important
cause of premature death in young adults with
diabetes.10 If we can identify those most at
risk and through more intensive management prevent admission with DKA then this
is a worthwhile goal.
The WDHB (Waikato District Health Board) catchment area
includes 339,100 people (8.3% of New Zealand’s population), of which
74,110 (22%) are Māori and 7,300 (2%) are Pacific people (mostly of Samoan,
Tongan, Niuean, or Cook Islands origin).
Previous studies have shown that Māori in New Zealand
have a prevalence and incidence of Type 1 diabetes lower than that in the New
Zealand European population11,12 whilst Type 2
diabetes is more common in Māori.2 The use
of antibodies is becoming increasingly important to differentiate between Type 1
and Type 2 diabetes in the province, particularly among Māori.
The primary aim of the study was to observe the relationship
between antibody status in newly diagnosed patients with Type 1 diabetes and the
incidence of hospital admission. A secondary aim was to compare antibody status
with long term glycaemic control as measured by HbA1c.
MethodStudy design and subjects—An
inception cohort of newly diagnosed patients with diabetes between 1997 and
2002, under the age of 25 at diagnosis and resident in the WDHB area was
identified from the Waikato Diabetes Service diabetes database. A starting year
of 1997 was chosen because this is when anti-GAD and anti-IA2 measurements began
to be used. Including patients diagnosed up until 2002 assured a minimum of 3
years of follow-up.
Patients were identified as Type 1, Type 2 and other.
Type 1 patients were differentiated from Type 2 patients based upon their
symptoms, insulin dependence, and autoantibody and glucose test results.
Patients with gestational diabetes without diabetes postnatally, drug-induced
diabetes, and diabetes related to surgery were excluded from the study.
Where the type of diabetes was not clear two
independent clinicians reviewed the patient file and provided a clear diagnosis.
Where there was disagreement between clinicians, the case was discussed until a
final diagnosis was reached.
Other patient information extracted from the diabetes
database included: gender, date of birth, ethnicity, year of diagnosis, age at
diagnosis, initial and current treatment, and HbA1c data. Where information and
data was missing from the database, the WDHB electronic database and local
pathology laboratory’s database was used to find these details. If this
did not provide the needed information the patient file was requested from
Waikato Hospital.
Anti-GAD and anti-IA2 data were obtained from the WDHB
and local pathology laboratory’s electronic databases and paper records.
Because two different laboratories were used to measure the antibody levels in
different patients, there were two different units and reference ranges in our
database: units/ml (measured by a radioimmunioassay by the Waikato Hospital
laboratory) and Units (used by Diatranz laboratory).
Consequently results were categorised into normal (as
indicated in the relevant reference range for each laboratory), weakly positive,
positive, or strongly positive as categorised in tertiles of the positive tests.
The primary outcomes of interest identified were: admission to hospital
admission for diabetes related conditions; DKA subsequent to being diagnosed
with diabetes; and most recent HbA1c levels. Death was initially considered as a
possible outcome of interest, but no patients in the cohort died during the
study period.
Information on hospital admissions for diabetes-related
complications (DKA, hypoglycaemia, and infections) were obtained from the WDHB
electronic patient database. This records the date of all admissions and
discharges and the reason for the admission. The results of all laboratory tests
are also available. Follow-up of all those Type 1 patients included in the final
database was through accessing the hospital electronic patient records, through
the laboratory database or by accessing the paper notes.
Statistics and analysis—The
incidence of diabetes by type and ethnicity was calculated. Population figures
were obtained from the WDHB and were based on estimates from the 1996 and 2001
censuses. Ethnic categories for these populations were given as Māori,
Pacific people, and Others (approximately 5% of the non-Māori, non-Pacific
population are of Asian descent. Each outcome of interest (number of hospital
admissions, number of hospitalised episodes of DKA and latest HbA1c level) was
plotted against a number of independent variables, including: age, gender,
ethnicity, age at diagnosis, body mass index (BMI), total cholesterol,
triglyceride, and autoantibody status to investigate any correlation. Logistic
Regression analysis using backward elimination was used with a significance
level of 0.05. Analysis was performed using STATA version 8 software (STATA
Corp., College Station, TX, USA).
ResultsA total of 164 people (84 males and 80 females) under the
age of 25 residing in the WDHB area were diagnosed with diabetes between 1997
and 2002. Of the 164, 133 (81%) were diagnosed with Type 1 diabetes and 27 (16%)
with Type 2 diabetes. Four (2%) had diabetes due to other causes, including
maturity onset diabetes of the young (MODY), cystic fibrosis, and pancreatitis.
Mean age at diagnosis in those with Type 1 diabetes was 13.0
years (14.8 in Māori and 12.6 in European), 67/133 (50.4%) were female and
the mean BMI was 22.9 kg/m2. The average
incidence of Type 1 and Type 2 diabetes in the Waikato region are shown in Table
1 and indicate the incidence of Type 1 diabetes was as expected lower in
Māori than non-Māori whilst the reverse was true in Type 2
diabetes.
Table 1. Incidence (%) of people with diabetes
(per 100,000 per year) in the Waikato DHB area, by diabetes type as well as
ethnicities and ages of people affected
Data are crude incidence rates (95% confidence
interval) based on average Waikato population for the period 1997–2002;
*The 2 patients of Pacific origin are included in the total population but are
not included in the column headed European/others.
Of the 133 patients with Type 1 diabetes, 85/133 (64%) had
anti-GAD results available and 68/133 (51%) had an anti-IA2 result. Of those
that had both tests, 59/68 (87%) had either anti-GAD, anti-IA2, or both
positive. Seventy-six percent were anti-GAD positive and 65% were anti-IA2
positive. The antibody status of patients by ethnicity is shown in Table 2 and
Table 3.
Fifty-nine (44%) of the 133 Type 1 patients had been
admitted to hospital for a diabetes-related complication since their diagnosis.
The main reason for admissions were hypoglycaemia, infections, or DKA. There
were 51 episodes of DKA among 24 patients. Of the 20 patients who had a negative
anti-GAD result, 1 (5%) had an episode of DKA—this compares to the 10
(15%) patients who had an episode of DKA from the group of 65 patients who had a
positive anti-GAD result (OR=1.46, p=0.226). Of the 48 patients who did not have
anti-GAD measured, 13 (27%) had an episode of DKA.
Table 2. Anti-GAD measurements by ethnic
group
Table 3. Anti-IA2 measurements by ethnic
group
*The 2 patients of Pacific origin are included in the
total population but are not included in the column headed
European/others.
Almost all the patients (97%) had HbA1c measurements
available. The mean HbA1c at diagnosis was 10.3% and after 3 or more years of
treatment the mean was 9.4%.
The current patient characteristics that are influencing
DKA, number of hospital admissions, and recent HbA1c >10% were investigated
using logistic regression analysis. The dependent variables used included most
recent HbA1c, current age, gender, duration of diabetes, Māori (Y/N),
anti-GAD positivity, and anti-IA2 positivity in the initial models.
Non-significant variables (p<0.10) were excluded using
backward stepwise regression. The most recent HbA1c was found to be a
significant predictor of DKA (OR=1.5 [1.23–2.01], p=0.001) and hospital
admission (OR=1.24 [1.01–1.57], p=0.039). Male gender and anti-IA2
positivity were significant predictors of most recent HbA1c >10 (OR=4.34
[1.39–13.54], p=0.012 and OR=0.278 [0.09–0.88], p=0.029
respectively).
Table 4. Anti-GAD levels of patients who had an
episode of diabetic ketoacidosis
DiscussionOur study shows that admission to hospital with DKA was a
relatively rare event and only occurred in 24/133 (18%) of patients. Other
common reasons for admission included hypoglycaemia and infections such as
candidiasis, infected pilonidal sinuses, and urinary tract infections.
We did not show any statistically significant associations
between antibody status and subsequent admission for DKA, but there was a
negative association between anti-IA2 positivity and poor glycaemic control as
indicated by a HbA1c > 10%. Unsurprisingly, a current HbA1c >10% was also
associated with risk of hospital admission or of admission with DKA.
Zanone et al also found an inverse relationship between
autoantibody levels and HbA1c.13 They
hypothesised that patients with higher anti-IA2 levels had more functioning
islet cells, which led to more endogenous insulin synthesis and hence less
dependence on exogenous insulin. Our findings support this in that those
patients with positive anti-IA2 were less likely to have a HbA1c >10%.
We did not show any association between anti-GAD levels and
HbA1c, hospital admission or DKA. This is in contrast to the small study of 35
patients by Hoeltke et al who showed an association between positive anti-GAD
status and risk of poor glycaemic control.9
The incidence rates of under-25 year olds with Type 1 and
Type 2 diabetes residing in the WDHB catchment area averaged 17.4 and 3.5 per
100,000 people per year, respectively.
The incidence for the under 15 age group was 17.9 per
100,000, which is the same as that found by Campbell-Stokes et al in their
study11 which covered all of New Zealand during
the 1999–2000 period. Rates for the Māori and non-Māori
0–14 subpopulations (5.3 and 24.3 per 100,000, respectively) are also
similar to that found by Campbell-Stokes et al. (5.6 and 21.7 per 100,000,
respectively).
It should be noted that the classification of ethnicity in
hospital records is not entirely consistent with self-identified ethnicity or
that used in the New Zealand census.14 Whilst
this may have introduced a bias into the estimate of the incidence of Type 1
diabetes in Māori, the consistency with the rate found by Campbell-Stokes
is reassuring and suggests substantial misclassification is unlikely to have
occurred.
A significant difference in incidence by ethnicity was noted
for Type 2 diabetes, with Māori having a much higher incidence rate. These
results are further evidence that the incidence of Type 1 diabetes has increased
since the 1980s.15,16 It also supports the
belief that Type 2 diabetes is becoming a more significant health issue among
younger people, especially Māori.17
Of those Type 1 patients that had two autoantibody
measurements done, 87% tested positive for one or more autoantibody. Of the 85
patients that had an anti-GAD measurement done, 65 (76%) had a positive result.
This is a similar result to that found in other
studies.9,18,19
Some studies have ignored those with idiopathic Type 1
diabetes and only include those insulin-dependent patients who are antibody
positive. However in all other respects these idiopathic Type 1 patients are
similar to those that have positive anti-GAD or anti-IA2. Of those that had both
an anti-GAD and anti-IA2 59/68 (87%) tested positive to one or other. This
result is slightly below the 94% who tested positive in the study by
Campbell-Stokes et al.
Because our study covered an initial period when antibody
levels for anti-GAD and anti-IA2 were not always done, the proportion of
patients that had their antibody status tested was lower in the earlier years of
the study. This may have introduced a bias to our findings. However in those
that did have their antibody status tested the proportions of New Zealand
Europeans and Māori with positive autoantibody results were very similar.
Thus the findings from the USA and South Africa where a higher proportion of
African/African Americans with Type 1 diabetes are antibody negative does not
seem to be true for Māori.
We believe this is the first time this finding has been
reported. Whilst a larger study is needed to confirm this finding it does
suggest that there maybe aetiological differences in the development of Type 1
diabetes in Africans and African Americans compared with other ethnic groups
including Māori.
Some methodological problems included having two
laboratories that used different auto-antibody measurements. This meant that we
had to categorise the level of antibodies rather than treat them as a continual
variable. Despite their benefits, the immunoassay techniques used are not
perfect and so can still quantify auto-antibody levels
incorrectly20 and so our categorisation maybe a
reasonable approach.
Another potential problem is that there may be
under-reporting of hospital admissions due to patients moving out of the WDHB
area or being out of the area when medical assistance was needed. Bias may have
come from clinicians only having auto-antibody measurements done on patients
where they were not confident of their diagnosis. Such a bias may be one
explanation why 27% (13/48) of those Type 1 patients who did not have their
anti-GAD levels measured had a subsequent episode of DKA compared to 15% (10/65)
of those who had their anti-GAD measured and had a positive result.
This study provided an overview of diabetes in children and
young adults in the Waikato Province. It has shown that the most important
predictor of subsequent admission to hospital for newly diagnosed patients with
Type 1 diabetes is poor glycaemic control. If there is evidence of antibodies to
IA2 present then this is a predictor of better glycaemic control and it maybe
that these patients will have less complications than those who are anti-IA2
negative.
Whilst anti-GAD is an important marker indicating the likely
subsequent need for long-term insulin therapy in
adults,21 it does not help predict risk in
newly diagnosed Type 1 patients.
As previously found by Scott et al, good glycaemic control
is hard to achieve in adolescents and young adults with Type 1
diabetes22 but concentrating on improving
glycaemic control in all newly diagnosed patients with Type 1 diabetes would
seem to be the most important factor in reducing hospital admissions and other
complications.
Competing interests: None.
Author information: Doron Hickey, Medical
Student, University of Auckland, Auckland; Grace Joshy, Research Fellow, Waikato
Clinical School, University of Auckland, Hamilton; Peter Dunn, Consultant
Endocrinologist, Regional Diabetes Service, Waikato Hospital, Hamilton; David
Simmons, Professor of Medicine, Waikato Clinical School, University of Auckland,
Hamilton; Ross Lawrenson, Professor of Primary Care, Waikato Clinical School,
University of Auckland, Hamilton
Acknowledgment: Doron Hickey was awarded a
Waikato Clinical School Summer Studentship to undertake this study. The
Studentship was kindly provided by the Waikato District Health Board.
Correspondence: Ross Lawrenson, Waikato
Clinical School, Peter Rothwell Academic Centre, Waikato Hospital, Private Bag
3200, Hamilton. Fax: (07) 839 8712; email: LawrensR@waikatodhb.govt.nz
References:
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| | Current
issue | Search journal |
Archived issues | Classifieds
| Hotline (free ads) Subscribe | Contribute | Advertise | Contact Us | Copyright | Other Journals |
|