  |
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Journal of the New Zealand Medical Association,
17-December-2004, Vol 117 No 1207
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Insulin resistance in a rural Maori community
David Tipene-Leach, Helen Pahau, Nathan Joseph, Kirsten
Coppell, Kirsten McAuley, Chris Booker, Sheila Williams, Jim Mann
The prevalence of diabetes is increasing
worldwide.1 In New Zealand, only limited
prevalence data are available, but evidence suggests that this increase is also
occuring.2–9 The most recent New Zealand
Health Survey (NZHS) found the prevalence of self-reported diabetes for people
aged over 45 years to be 8.1% for females and 10.0% for
males.6 Prevalence surveys have consistently
shown diabetes (both known and newly diagnosed) to be more common among Maori
compared with New Zealanders of European descent. Most recently, in the NZHS
self-reported diabetes prevalence among Maori aged over 45 years was 21.4% and
13.0% for males and females, respectively—compared with 8.6% and 7.5% for
non-Maori males and females, respectively.6
Previously, Simmons et al8
found the prevalence of known diabetes mellitus in South Auckland was 6.9% among
Maori compared with 2.8% among Europeans—and in the New Zealand
Multiracial Workforce Survey, the prevalence of known diabetes mellitus was 5.3%
among Maori compared with 1.1% among
Europeans.7
In New Zealand, information about the prevalence of impaired
glucose tolerance (IGT) and impaired fasting glycaemia (IFG) is
limited2,3,7,9.
Even less is known for any population regarding the prevalence of insulin
resistance, a condition generally present prior to the development of IGT and
IFG, and a major risk factor for cardiovascular disease, but it has been
estimated that approximately 25% of people of European descent have insulin
resistance.10
While there is no information on prevalence of insulin
resistance in New Zealand, Simmons et al9 found
that, compared with Europeans, Maori and Pacific people have poorer insulin
sensitivity—when applying the Homeostasis Model Assessment (HOMA) of
fasting glucose and insulin as a proxy measure of insulin resistance. This
finding was attributed to the high rates of obesity among Maori and Pacific
people, rather than of inherent insulin resistance among people of Polynesian
descent.
Trials have demonstrated that progression from IGT to type 2
diabetes can be halted through lifestyle
changes,11–13 but approximately 40% of
those people still develop type 2 diabetes despite lifestyle intervention. This
may be due to lack of compliance, but also likely to be due to the considerable
beta-cell dysfunction already present in those with IGT. Thus, lifestyle
intervention among those with insulin resistance, but normal glucose tolerance,
may be a more effective approach to preventing or delaying the onset of type 2
diabetes, as well as reducing cardiovascular risk.
However, people with poor insulin sensitivity need to be
easily identified, and this study examines different methods to achieve this
goal in the clinical setting. A Maori Primary Health Organisation (PHO), Ngati
Porou Hauora, has initiated a 2-year community lifestyle intervention programme
(the Ngati and Healthy Programme) aimed at reducing the prevalence of insulin
resistance among a predominantly Maori community living along the rural East
Coast area, north of Gisborne, in the North Island of New Zealand. The outcome
of the intervention will be assessed by pre- and post-intervention prevalence
surveys of insulin resistance, as well as IFG, IGT, and diabetes. This paper
presents results of the pre-intervention prevalence survey.
MethodsThis study was based on the
sparsely populated East Coast of New Zealand, among communities from Tolaga Bay
(50 km north of Gisborne) to Potaka (near Te Araroa). (See the map of the East
Coast region; Figure 1.)
The East Coast region has a population of approximately
6000 people. The main types of employment are forestry and farming. Ngati Porou
Hauora has over 13,500 enrolled patients in Gisborne and the East Coast, and
provides comprehensive services to the East Coast communities through six
community clinics and is the only primary care provider in this rural region.
The Ngati Porou Hauora East Coast Enrolled Patient
Register was used to obtain a random sample
stratified by sex, age group, and ethnicity. Ethical approval was
obtained from the Tairawhiti Ethics Committee in March 2003 and the study took
place throughout May to December 2003.
The Project Co-ordinator, and Ngati Porou Hauora rural
health nurses and kaiawhina, (community health workers) invited selected
individuals to participate in the survey by letter, and if necessary by phone
and home visit (up to three visits in some cases).
Figure 1: The survey was based on the East Coast region
of New Zealand’s North Island—from Tolaga Bay (50 km north of
Gisborne) to Potaka (near Te Araroa)
Of the 741 individuals selected to participate in the
study, two individuals were excluded because they had a terminal illness or
died, and 150 were unable to be contacted because they had moved away from the
East Coast study area. Thus, 589 individuals received an invitation to
participate in the study.
At clinics undertaken at four different sites in the
East Coast region, a questionnaire including demographic information, relevant
medical history, and exercise and dietary history was administered. Height,
weight, and waist circumference (midpoint between the anterior superior iliac
crest and the lowest rib) were measured, body mass index (BMI) calculated, and
blood pressure recorded (after 10 minutes of rest using random zero
sphygmomanometers).
Duplicate measures were taken for each of the
anthropometric measures, and the average of the two measures used in the
analysis. A 75 g oral glucose tolerance test (OGTT) was performed with glucose
and insulin measured at 0 and 120 minutes post-glucose load. Participants with
documented diabetes did not have an OGTT. Blood was also taken for fasting
lipids. All samples were spun and separated after collection. The plasma insulin
samples were frozen and transported to Gisborne Laboratory in a mobile freezer
(approximately -15°C) either the same day of collection or the following
day.
Blood samples were packaged with Bio-freeze Blue Ice
bottles to keep the samples at approximately -15°C, and sent immediately to
Canterbury Health Laboratory (Christchurch, New Zealand), where they were
processed. All other samples were stored in polystyrene boxes for transportation
the same or the following day, and were processed on arrival at Gisborne
Hospital’s Laboratory (New Zealand).
Plasma glucose, total cholesterol, and triglycerides
were measured using an enzymatic colorimetric method (Ortho-Clinical Diagnostic
reagents). HDL cholesterol was measured using the direct magnetic method. In
accordance with the Royal College of Pathologists of Australasia Quality
Assurance Programme, coefficients of variation were 2.2% for glucose, 3.7% for
total cholesterol, 6% for HDL, and 3.7% for triglycerides.
Canterbury Health Laboratory, using a Roche Elecsys
2010 automated analyzer with polyethylene glycol to remove antibodies, measured
plasma insulin after extraction. The assay detection limit is 0.4 mIU/L. The
intra-assay coefficient of variation was 6%.
IFG, IGT, and diabetes were defined according to WHO
diagnostic criteria.14 Insulin resistance was
predicted using the McAuley formula based on fasting insulin, triglycerides and
BMI,15—where predicted insulin
sensitivity was expressed as exponent (3.29-0.25ln[fasting insulin]-0.22ln[body
mass index]-0.28 ln[fasting triglycerides]). Normoglycaemic individuals with
calculated values ≤ 6.3 M• mU-1• l-1 were defined as insulin
resistant. As there is no internationally agreed simple method for predicting
insulin sensitivity, insulin resistance was also estimated using three other
methods: the Homeostasis Model Assessment, HOMA Calculator computer model
(version 2.1, 2004), 16 based on fasting
insulin and fasting glucose, the National Education Program (NCEP) Adult
Treatment Panel (ATP III) definition,17 which
uses a set of clinical criteria (based on blood pressure, waist circumference,
triglycerides, HDL, and fasting glucose) and an insulin sensitivity index (based
on the average of a fasting and 2-hour glucose, and the average of fasting and
2-hour insulin.)18
The ATP III criteria were applied to all study
participants, while HOMA 2.1 and ISI0,120 calculations excluded known diabetics
taking oral hypoglycaemic medications or insulin.
Data were entered into a Microsoft Access-based
software program. Regression analysis was used to estimate differences between
groups after adjustment for sex.
Results289 people agreed to participate in
the study, giving an overall response rate of 48.7%. Males aged 25–29
years had the lowest response rate (24.0%)—whereas males aged 60 years and
over, and females aged 30 years and over, had response rates higher than 50%,
the highest being 76% for females aged 50–54 years. The female:male ratio
was 1.5, and 249 (86%) respondents self-identified as Maori. The following
results are for Maori participants only.
Table 1 shows the demographic and clinical characteristics
of Maori respondents. The mean BMI was 33.4
kg/m2, and more than 90% of both females and
males were either overweight or obese, defined as a BMI of 25
kg/m2 or more.
Table 2 shows the age-standardised prevalence of insulin
resistance, IFG or IGT, and diabetes (estimated using our equation based on
fasting insulin, triglycerides, and BMI). Overall, the age-standardised
prevalence of diabetes, both known and newly diagnosed, was 10.6%. The
age-standardised prevalence of known diabetes was about twice that for newly
diagnosed diabetes. IGT or IFG was relatively uncommon, whereas the
age-standardised prevalence of insulin resistance was 40.3% for females and
36.0% for males.
Figure 2 shows the overall age-specific prevalence rates for
diabetes (both known and newly diagnosed) and insulin resistance with normal
glucose tolerance. Insulin resistance was more common among the young age
groups—with the 30–39 year age group having the highest age-specific
rate (44.3%), whereas the prevalence of diabetes increased with age, peaking in
the 60–69 year age group at 34.1%.
The characteristics of the group identified as being insulin
resistant were compared with the group that did not have any disorders of
glucose metabolism (Table 3). The mean age of these groups was similar, as was
the proportion who smoked. A history of gout and a family history of diabetes
were more common among the insulin-resistant group. Also, individuals in this
insulin-resistant group were more likely to be overweight or obese and have an
elevated blood pressure, and an elevated triglyceride level. Total cholesterol
and LDL levels were similar.
Table 1. The demographic and clinical characteristics
of study participants by sex
Mean values and standard
deviations are presented unless otherwise stated; BP=blood pressure;
HDL=high-density lipoprotein; BMI=body mass index.
Table 2. Age-standardised prevalence of insulin
resistance, IFG or IGT, and diabetes in adults aged 25 years and over
Age-standardised to the WHO
world population; IFG=impaired fasting glycaemia; IGT= impaired glucose
tolerance;*Insulin
resistance calculated using the McAuley formula
among those with normal glucose
tolerance.
Table 4A shows different estimates of age-specific
prevalence of glucose metabolism disorders in three age categories using our
prediction equation and the ATP III criteria. Comparable age trends are evident
with the two approaches. As no cut-offs to define insulin resistance have been
applied to the HOMA2.1 and the ISI0,120 method,
means and standard deviations are presented in Table 4B rather than prevalence
rates. Those persons taking diabetes medications were excluded for the
calculation of the HOMA2.1 as fasting insulin levels are less meaningful in this
setting, and as this group did not have an OGTT, they could not be included in
the ISI0,120 calculation.
Figure 2. Age-specific prevalence of diabetes (known
and newly diagnosed) and insulin resistance with normal glucose
tolerance
 Table 3. Characteristics of the insulin resistance but
normal glucose tolerance group and the ‘healthy’ group
Data
presented are mean values and standard deviations unless otherwise stated;
Differences or odds ratios are adjusted for sex; IHD=ischaemic heart disease;
BP=blood pressure; HDL=high-density lipoprotein; BMI=body mass index; *Odds
ratio (95% CI); †Geometric means (95% CI) and their ratio (95% CI) based
on a log transformation.
Table 4A. Comparison of age-specific prevalence rates
of insulin resistance using our formula and the ATP III criteria
*McAuley
formula = exp[3.29 – 0.25ln(insulin) – 0.22ln(BMI) –
0.28ln(TAG)]. (The formula was developed to predict insulin sensitivity, values
≤ 6.3 M• mU-1• l-1 define those who are insulin resistant.);
IFG=impaired fasting glycaemia; IGT=impaired glucose tolerance.
Table 4B Comparison of means and standard deviations by
age groups for HOMA 2.1 and ISI0,120
*
%S is derived from the HOMA 2.1 computer model, and is a measure of insulin
sensitivity with 100% defined as normal, and higher numbers signifying greater
sensitivity; ISI0,120 is a calculated measure of
sensitivity with higher numbers signifying greater
sensitivity;
† HOMA 2.1 and
ISI0,120 calculations exclude those on diabetes
medications.
DiscussionSlightly less than half the
eligible participants (49%) completed all components of the survey. While a
higher response rate would have been desirable, this rate was comparable with
that of similar surveys such as the recent AUSDIAB
Study.19 The low overall response rate in our
study can to a considerable extent be explained by the poor response among the
younger age group, especially males. The forestry industry employs a high
proportion of the young men, and because their working days begin early, many
young men were unable to obtain time away from work.
The population of the East Coast is spread over a large
geographic area, and the distances required to travel to the survey centres was
a major disincentive to participation in all age groups. However, the
coordinated efforts of the study coordinator, kaiawhina, and rural health nurses
to provide frequent reminders and to arrange transport to the survey centres
resulted in a much higher response rate among middle aged and older individuals.
Among women aged 40 years and over, the response rate was 62%, which compares
favourably with the 68% response rate among 40–79 year old Maori women who
were invited to complete a questionnaire and have anthropometric measures and a
random blood glucose in a 1995/96 South Auckland
survey.9 Comparison of the response rate for
males is less favourable (43% vs 63%). The high response rate (93% of
households) in an earlier South Auckland survey involved only the completion of
a questionnaire and suggests that the oral glucose tolerance test (OGTT) and
physical examination may be disincentives.8
However, an OGTT is an essential component of any study which aims to assess the
prevalence of disturbances of carbohydrate metabolism.
The limited number of prevalence studies among Maori
populations in New Zealand are not directly comparable, and it was not possible
to assess changes in diabetes prevalence over time. However, the comparable
prevalence of known diabetes among women aged over 45 years in the East Coast
population studied in the present survey (14%) and of self-reported diabetes in
the 2002/03 NZHS (13%)6 provides strong
confirmation of the overall high prevalence. Of interest, are the appreciably
higher rates of self reported diabetes among males aged over 45 years in the
NZHS (21%) than in the present study (10%), where the diagnosis was confirmed.
The difference may be partly explained by the low response rate among East Coast
men, but we cannot explain this difference with certainty.
Interestingly, in our data, newly diagnosed diabetes rates
were only half that of known diabetes whereas most other surveys have reported
comparable rates of known and newly diagnosed
cases.3,7,9 This may well be due to the high
level of awareness of diabetes in the study area. This in turn has resulted in
more frequent screening of high-risk individuals.
A key purpose of the present study was to assess the
prevalence of insulin resistance. Individuals with insulin resistance in the
general population urgently need to be targeted for diabetes prevention and
cardiovascular risk reduction strategies, but no universally accepted method for
predicting insulin sensitivity exists. Euglycaemic clamps and intravenous
glucose tolerance tests (IVGTTs) are limited to research settings. Various
surrogate methods for predicting insulin sensitivity have been used in studies
and a number of newer approaches have been
suggested.20
The most widely published method for predicting insulin
sensitivity is the Homeostasis Model Assessment based on fasting glucose and
insulin,21 which should ideally be based on
three separate blood measurements taken 5 minutes apart and calculated using the
model programme, but in most cases is based on a single measure and is estimated
using a simplified formula.16 Only the
HOMA-model has been shown to correlate well with the euglycaemic
clamp.16,18
Our study and others have found the HOMA formula to be no
better than a fasting insulin in this
regard.15,18 The HOMA formula is generally
applied to those with normal glucose tolerance (NGT), IFG, IGT, and diabetes
with several caveats for use in those on sulponylureas and exogenous
insulin.16
No cut-off has been proposed to identify a group with poor
insulin action. Furthermore, there has been criticism of choosing surrogates to
predict insulin sensitivity that correlate well with a euglycaemic clamp, a
dynamic test under non physiological conditions. Thus, we have selected a
further method for predicting insulin resistance, developed by Gutt et al, based
on the average of the fasting and 2-hour glucose and insulin levels. This has
been shown in prospective studies to be the best method for predicting the
development of type 2 diabetes.20
The failure to show a deterioration in insulin sensitivity
with age (Table 4B) with both these measures reflects the fact that those with
diabetes on medication have been excluded because of the difficulty in
interpreting their insulin sensitivity data using this approach. It is clear
that HOMA 2.1 and ISI0,120
formulae are currently inappropriate to
determine prevalence of insulin resistance. However they are likely to be of
value in assessing response to intervention programmes aimed at improving
insulin sensitivity, and will be used for this purpose in the Ngati and Healthy
programme.
To date, the most frequently used approach for determining
frequency of insulin resistance or identifying insulin resistant individuals has
been to use a set of surrogate clinical and laboratory criteria. We have
compared our equation15 which has been
independently validated22 with the ATP III
criteria17 for the definition of those with the
metabolic syndrome. Our equation combines fasting insulin, triglycerides and BMI
as continuous variables, so that those who would have fallen just outside a
particular cut off can still be included depending on the other variables.
An arbitrary cut off (of less than or equal to 6.3
M•mU-1•l-1) is applied to select those with poor insulin
sensitivity, based on the lowest quartile for a lean population. Inevitably, the
cut-off point is somewhat arbitrary but a similar difficulty applies to the ATP
III criteria which might be expected to miss an even greater number of insulin
resistant individuals since arbitrary cut offs are applied to several clinical
and metabolic variables.
Table 4A shows that using our approach, more than half the
population have insulin resistance, and this increases with increasing age, when
those with IFG, IGT, and type 2 diabetes are included. Not surprisingly, the ATP
III criteria gives rates substantially lower than this, but the same pattern of
increasing rates across age groups is observed.
It has been estimated that as many as 25% of adults of
European descent may be insulin resistant.10
The appreciably higher rates observed here (40% among women, 36% among men)
represent considerable cause for concern given that this condition is believed
to be the underlying cause of most cases of type 2 diabetes mellitus as well as
being an important contributor to cardiovascular risk. Of especially great
concern are the high rates among young individuals (Figure 2). This suggests
that the future burden of diabetes and other diseases associated with insulin
resistance and the metabolic syndrome is likely to escalate in the near future
unless effective intervention programmes are in place.
The Ngati Porou Hauora Ngati & Healthy Programme is one
such pioneering programme, which will be formally evaluated using well
established methods.23 A national diabetes
prevalence survey, which would include estimates of IFG, IGT, and insulin
resistance as well as associated clinical, anthropometric, and metabolic
variables and assessment of nutritional status, is imperative since no such
national data exist. Such information is essential for health care planning for
what is arguably the most important epidemic disease in New Zealand and for
assessing the effects of national strategies aimed at reducing obesity and
diabetes rates.
Author information:
David Tipene-Leach, General Practitioner; Helen Pahau, Research Coordinator;
Nathan Joseph, General Practitioner, for Ngati Porou Hauora, Gisborne; Kirsten
Coppell, Senior Research Fellow; Kirsten McAuley, Senior Research Fellow; Chris
Booker, Senior Research Technician; Jim Mann, Director, for Edgar National
Centre for Diabetes Research, University of Otago, Dunedin;
Sheila Williams, Biostatistician, Department of Preventive and Social
Medicine, University of Otago, Dunedin.
Acknowledgements: We
received funding for the project from the Bristol Meyer Squibb Mead Johnson
Unrestricted Research Grant. We also thank Sally Abel and Terry Ehau from Ngati
Porou Hauora, Marina Ngatai, Ginny Dawn Reedy, Hera Sykes, Mereana
Northover, Kura Forrester, TeoArani Wilson (Kaiawhina), BJ Taare, Ariana
Roberts, JJ Hitchener, Claudia White, Gina Chaffey-Aupouri, and Maryanne Barton
(Rural Health Nurses) for their invaluable input and assistance; and the staff
of Tairawhiti Laboratory in Gisborne for assistance with venepuncture, storage,
and transportation of blood samples.
Correspondence: Dr
David Tipene-Leach, Ngati Porou Hauora, Puhi Kaiti Hauora, Kaiti Mall, PO Box
3028, Gisborne. Fax: (06) 867 3260. email: davidt@nph.co.nz
References:
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| | Current
issue | Search journal |
Archived issues | Classifieds
| Hotline (free ads) Subscribe | Contribute | Advertise | Contact Us | Copyright | Other Journals |
|