Advertisement
Research Article

Early Markers of Glycaemic Control in Children with Type 1 Diabetes Mellitus

  • Samuel W. Cutfield,

    Affiliation: Liggins Institute, University of Auckland, Auckland, New Zealand

    X
  • José G. B. Derraik,

    Affiliation: Liggins Institute, University of Auckland, Auckland, New Zealand

    X
  • Peter W. Reed,

    Affiliation: Starship Children's Hospital, Auckland District Health Board, Auckland, New Zealand

    X
  • Paul L. Hofman,

    Affiliations: Liggins Institute, University of Auckland, Auckland, New Zealand, National Research Centre for Growth and Development, University of Auckland, Auckland, New Zealand

    X
  • Craig Jefferies,

    Affiliation: Starship Children's Hospital, Auckland District Health Board, Auckland, New Zealand

    X
  • Wayne S. Cutfield mail

    w.cutfield@auckland.ac.nz

    Affiliations: Liggins Institute, University of Auckland, Auckland, New Zealand, National Research Centre for Growth and Development, University of Auckland, Auckland, New Zealand

    X
  • Published: September 26, 2011
  • DOI: 10.1371/journal.pone.0025251

Abstract

Background

Type 1 diabetes mellitus (T1DM) may lead to severe long-term health consequences. In a longitudinal study, we aimed to identify factors present at diagnosis and 6 months later that were associated with glycosylated haemoglobin (HbA1c) levels at 24 months after T1DM diagnosis, so that diabetic children at risk of poor glycaemic control may be identified.

Methods

229 children <15 years of age diagnosed with T1DM in the Auckland region were studied. Data collected at diagnosis were: age, sex, weight, height, ethnicity, family living arrangement, socio-economic status (SES), T1DM antibody titre, venous pH and bicarbonate. At 6 and 24 months after diagnosis we collected data on weight, height, HbA1c level, and insulin dose.

Results

Factors at diagnosis that were associated with higher HbA1c levels at 6 months: female sex (p<0.05), lower SES (p<0.01), non-European ethnicity (p<0.01) and younger age (p<0.05). At 24 months, higher HbA1c was associated with lower SES (p<0.001), Pacific Island ethnicity (p<0.001), not living with both biological parents (p<0.05), and greater BMI SDS (p<0.05). A regression equation to predict HbA1c at 24 months was consequently developed.

Conclusions

Deterioration in glycaemic control shortly after diagnosis in diabetic children is particularly marked in Pacific Island children and in those not living with both biological parents. Clinicians need to be aware of factors associated with poor glycaemic control beyond the remission phase, so that more effective measures can be implemented shortly after diagnosis to prevent deterioration in diabetes control.

Introduction

Type 1 diabetes mellitus (T1DM) may lead to severe long-term health consequences, such as renal failure, blindness, as well as heart and cerebrovascular disease [1], [2]. Previous data have shown that intensive therapy to maintain blood glucose concentrations near the normal range delays the onset and slows the progression of adverse outcomes associated with T1DM, such as retinopathy, nephropathy and neuropathy [2]. Glycosylated haemoglobin (HbA1c) monitoring has been shown to improve glycaemic control when added to conventional blood glucose testing, and it is now standard practice in the management of patients with T1DM. Thus, the monitoring of HbA1c levels may minimize possible short- and long-term adverse health outcomes. For example, therapy to lower HbA1c from 7.9 to 7.0% reduced the risk of microvascular complications by 25% [3].

To minimize the likelihood of long-term complications, it would be valuable to identify early on those groups where diabetes control is likely to be poor (arbitrarily defined as HbA1c >9%) [4].

By using predictors of metabolic control based on data collected at diagnosis or shortly after, closer monitoring of at-risk subjects would be possible. Early intervention and monitoring in these subjects could potentially reduce the likelihood of severe negative health outcomes associated with sustained hyperglycaemia.

HbA1c remains the simplest indicative measure of diabetes control. Several studies have consequently examined a number of physiological and environmental factors that may affect HbA1c levels in diabetic patients, such as ethnicity [5][10], socio-economic status (SES) [6], [11], [12] and family living arrangements [10], [13], [14]. However, these cross-sectional studies looked at correlations between specific factors and HbA1c, and not their ability to predict HbA1c in the long-term. Further, the remission or ‘honeymoon’ phase occurs in most children shortly after T1DM diagnosis [15][18]. This remission phase is characterized by temporary recovery of residual β-cell function that typically lasts up to 6 months [19], and it is usually defined as requiring an insulin dose <0.5 U/kg/day [15], [17], [18]. During this phase there is a reduction in insulin dose and good glycaemic control is readily achievable [20]. As a result, during the remission phase diabetic children may have risk factors for later poor glycaemic control that are not initially evident. Thus, our retrospective analysis of longitudinal data aimed to identify factors present at diagnosis and 6 months later that are associated with HbA1c levels at 24 months after T1DM diagnosis, so that diabetic children at risk of poor glycaemic control in particular, may be identified.

Methods

Subjects

The Endocrinology Service at Starship Children's Health (Auckland, New Zealand) provides specialist care for all children <15 years of age (yr) diagnosed with T1DM in the Auckland region, with case ascertainment levels of over 95% [21]. These children are reviewed by this service at least once every three months. New Zealand has a social security system that provides medical care free of charge, so that direct costs of T1DM to patients' families are minimal and income is not a direct impediment for take-up of medical care.

Demographic and clinical data were collected from the Starship Children's Hospital Diabetes Database (Starbase), with additional information obtained from hospital records as required. The database includes all children diagnosed with T1DM in the Auckland region, drawing from a total population of approximately 1.3 million [22].

For this study, data on all children <15 yr diagnosed with T1DM between 1 January 2000 and 31 December 2008 were collected. Subjects were included in the study if the following criteria were met: blood tests showed T1DM antibodies glutamic acid decarboxylase (GAD) and/or islet antigen 2 (IA2) at presentation, ongoing requirement for insulin, and regular attendance (≥3 visits per annum) to the Paediatric Diabetes Service at Starship. Subjects who were antibody negative were excluded to remove those with maturity onset diabetes of the young (MODY) or T2DM. Subjects were also excluded if they had been on an insulin pump within 2 years of diagnosis, had data at 6 or 24 months missing, had syndromes associated with impaired cognitive ability, had coeliac disease, or were receiving drugs known to influence insulin sensitivity. Ethics approval to conduct this study was granted by the Auckland District Health Board Research Review Committee (study number ADHB 5013).

Study Parameters

Data on demography, auxology and biochemistry were collected at presentation, at 6±2 months and 24±3 months after diagnosis. Parameters collected at diagnosis were: age, sex, weight, height, ethnicity, family living arrangement, SES, T1DM antibody titre, venous pH and bicarbonate. At 6 and 24 months after presentation at clinic, we collected data on weight, height, HbA1c level, and insulin dose.

Ethnicity was recorded by self-report using a priority system, such that if multiple ethnicities were selected, the patient was assigned to a single ethnicity, following a hierarchical classification of Maori, Pacific Islander, Other, and then European [23]. Those participants described as “Other” were of ethnicities not otherwise stated (e.g. Indian, South-East Asian, African and Middle Eastern). Family living arrangement was defined as either living with both biological parents in the same household, or as living under alternative arrangements such as the care of a single-parent, blended family, foster parents, grandparents, etc.

The New Zealand Deprivation Index 2006 (NZDep2006) [24] was used as the indicator of SES [6]. This index uses household census data reflecting nine aspects of material and social deprivation, giving scores to New Zealand residential addresses from 1–10, where 1 represents the most affluent and 10 the most deprived [25]. Scores are derived from units covering a relatively small area, each reflecting approximately 90 people [24]. These units correspond to “domicile codes” recorded in the New Zealand National Health Index database [26], so that NZDep2006 scores can be assigned to individual patients based on residential address.

Height was measured to the nearest 0.1 cm using a wall-mounted Harpenden stadiometer (Holtain, Crosswell, UK). Weight was measured to the nearest 0.1 kg, with the participant in light clothing, by electronic scales. Body mass index standard deviation scores (BMI SDS) were calculated using The University of Manchester child obesity calculator based on British 1990 growth reference data [27]. HbA1c was measured with a DCA 2000 Analyzer (Bayer, Elkhart, IN, USA), which was standardised to laboratory-tested HbA1c on a monthly basis. DKA occurrence was assessed by blood gases using pH and HCO3-. DKA severity was categorized as mild (venous pH<7.3 or bicarbonate <15 mmol/l), moderate (pH<7.2, bicarbonate <10 mmol/l) or severe (pH<7.1, bicarbonate <5 mmol/l) [28]. Autoantibodies to GAD and IA2 were measured using RSR ELISA kits (RSR Ltd, Cardiff, UK).

Data Analysis

Data were analysed using JMP 8 (SAS Inc. NC, USA). Univariate comparisons were investigated with parametric t-tests or linear regression, as appropriate. Non-parametric tests of the same data made no material difference to the results and are not presented. Multivariate analyses were undertaken using the standard least squares approach. Individual variables associated in univariate tests were combined in a step-wise fashion to identify the multivariate model with the greatest r2, where the evidence for each variable was p<0.05. Statistical significance was defined as p<0.05.

Data are presented as mean ± SEM, unless otherwise stated.

Results

Diagnosis

A total of 450 children diagnosed with T1DM were recorded in Starbase between 2000 and 2008, from which 229 met all the inclusion criteria and were consequently incorporated into this study. Subjects were excluded due to: incomplete data, moving to another region, or diagnosis outside Auckland (162); insulin pump therapy within the first 2 years of the study period (19); coeliac disease (20); absence of T1DM antibodies (15); and syndromes affecting intellectual capacity (5).

There were no differences in any of the demographic characteristics between the total group and the subjects recruited into this study, which are summarised in Table 1. The majority of children were from European families of average affluence, although overall Europeans also had better SES than all other groups (Table 1). Pacific Islanders were considerably heavier than subjects of other ethnicities (p<0.01; Table 1). This trend remained throughout the study period, so that 67% of Pacific Island children were obese (BMI SDS greater than the 95th percentile) throughout the study. Pacific Islanders were of lower SES than all groups but Maori (Table 1). The majority of children (86%) were living with both biological parents (Table 1). Females were more likely than males to be in DKA at diagnosis (30 vs 19%, p<0.05), and females in DKA at diagnosis were 4-fold more likely to be in severe DKA than males (36 vs 9.0%, p<0.05).

thumbnail

Table 1. Demographic characteristics of subjects at diagnosis.

doi:10.1371/journal.pone.0025251.t001

6 months after diagnosis

Several demographic factors at diagnosis were associated with worse glycemic control as reflected in higher HbA1c levels at 6 months (HbA1c 6mo): female sex (p<0.05), lower SES (p<0.01), non-European ethnicity (p<0.01), younger age (p<0.05), and having a first degree relative with T1DM (p<0.05). The remission or ‘honeymoon’ phase (insulin dose <0.5 U/kg/day) was present in 33% of children, and it was strongly associated with ethnicity (17% of Pacific Island vs 37% of European children; p<0.001). Children within the remission phase had a mean HbA1c of 7.1%, and these were predominantly European (75%). Notably, the HbA1c was higher among Pacific Island children in the remission phase (7.5%). Neither the presence nor severity of DKA at diagnosis influenced HbA1c 6mo.

24 months after diagnosis

Although HbA1c 6mo was associated with HbA1c 24mo (p<0.001), it had a weak positive predictive value (r2 = 0.16). However, demographic and clinical characteristics obtained at diagnosis were associated with higher HbA1c 24mo: lower SES (p<0.001), Pacific Island ethnicity (p<0.001), not living with both biological parents (p<0.05), and greater BMI SDS (p<0.05). The effect of ethnicity on HbA1c levels at 6 and 24 months following diagnosis was again marked. Pacific Island children, who had worse glycaemic control at 6 months, displayed HbA1c levels that had deteriorated more markedly than other groups by 24 months (Fig. 1), so that mean HbA1c in Pacific Islanders were 9.9±0.4% compared to 8.1±0.1% (p<0.001) in European children. HbA1c levels were also higher in children with living arrangements other than living with both biological parents (8.9±0.3 vs 8.3±0.1%, p<0.05). At 24 months after diagnosis, there were only three children (1.3%), of whom two were European, still in the remission phase.

thumbnail

Figure 1. Mean changes in HbA1c between 6 and 24 months after diagnosis for Europeans (n = 157; solid black line), Maori (n = 16; dashed black), Pacific Islanders (n = 24; dashed gray), and Other ethnicities (n = 32; solid gray).

Data are mean ± SEM. *p<0.05 and ***p<0.001 for temporal changes within each ethnic group.

doi:10.1371/journal.pone.0025251.g001

Predictive equation

A multivariate regression equation was developed to identify demographic and clinical factors 6 months after diagnosis that predicted glycaemic control at 24 months after diagnosis (r2 = 0.33, p<0.001):

Where, for ethnicity, Pacific Island = +1.06, Other = -0.27, European = -0.37, and Maori = -0.42 for family, living with both biological parents = -0.28, and other living arrangements = +0.28.

The model was not strengthened with the addition of SES, adiposity, or insulin dose at 6 months after diagnosis.

Pacific Island children were markedly heavier than all other ethnicities, but particularly Europeans (BMI SDS of 1.98±0.21 vs 0.64±0.07 at 6 months, and 2.02±0.16 versus 0.75±0.07 at 24 months). However, in the multivariate analysis it was ethnicity and not BMI SDS that influenced HbA1c levels at 6 and 24 months, suggesting that ethnicity itself rather than body composition might have affected glycaemic control across the entire study group. The effect of SES was also lost in multivariate analysis, suggesting that ethnicity rather than SES was the major influence on HbA1c levels.

Discussion

In a regional cohort of children with T1DM, we found that Pacific Islanders and those of low SES had higher HbA1c levels at 6 months from diagnosis, with an even greater deterioration in glycaemic control in Pacific Island children after 24 months. In addition, children not living with both biological parents also displayed poorer glycaemic control.

Cross-sectional analyses have previously shown that both ethnicity and to a lesser degree SES influence HbA1c levels in patients with diabetes [5], [6], [10], [11]. Other studies have observed that children and adolescents from ethnic minorities display worse glycaemic control [6], [10], [29][31]. African American children and adolescents with T1DM showed a similar marked difference in HbA1c values, which were 1.3 to 1.5% higher than European Americans [5], [10]. Less dramatic differences have been seen in other ethnicities such as New Zealand Maori (HbA1c 1.0% higher than Europeans) [6] and American Hispanic children (HbA1c 0.5% higher) [11]. However, even though ethnicity is linked to diabetes control (HbA1c), the key characteristics of minority groups associated with poor glycaemic control have not been well-described. Nonetheless, ethnicity and SES are often associated, and minority ethnic groups tend to be more socially disadvantaged, experiencing higher poverty rates and worse health status [32], [33]. Although SES can be associated with poor glycaemic control in children, this is not always the case [5], [6], [34], [35]. In this study, we observed that SES on its own influenced HbA1c, but when combined with ethnicity in a multivariate regression, ethnicity was the greater influence and there was no independent SES effect.

Family living arrangement was also an important factor determining diabetes management, and children living with both their biological parents were associated with better glycaemic control 24 months after diagnosis. Other studies have shown that living with a single parent is associated with poor glycaemic control in children and adolescents [5], [10], [36][39]. In our study, single-parents were combined with other living arrangements, mainly those living in blended-parent families, for which similar association with poor glycaemic control has been previously observed [40].

With awareness of the risk factors associated with poor glycaemic control, physicians can assist those patients most at risk of inadequate diabetes management in the long-term. This study has identified a number of primary risk factors: Pacific Island ethnicity, lower SES, higher HbA1c 6mo, increased BMI SDS, and living arrangements without both biological parents. Although predictably higher HbA1c 6mo was associated with higher HbA1c 24mo, the relationship was weak.

Recently, a study found that HbA1c at 1 year in children with T1DM would not allow practitioners to predict the trajectory of glycaemic control over the subsequent 3 years [41], but other physiological and environmental factors were not included in their analysis. Thus, to better predict glycaemic control at 24 months from diagnosis (i.e. HbA1c 24mo), we have developed a multivariate regression equation using demographic characteristics at diagnosis and clinical features 6 months afterwards (including HbA1c 6mo). The equation not only identifies primary risk factors associated with inadequate diabetes management, but also enables clinicians to estimate long-term glycaemic control. The regression equation estimating HbA1c 24 months after diagnosis included three factors: HbA1c 6mo, ethnicity, and family living arrangements. So, for example, the model predicts with 95% confidence that a Pacific Island child with an HbA1c 6mo of 7.9% from a single-parent family would have HbA1c 24 months after diagnosis of 10.3±0.6 (mean ± SD). Alternatively, a European child with HbA1c 6mo of 7.3% living with two biological parents would have HbA1c 24 months after diagnosis of 8.1±0.2.

Shortly after diagnosis, children with T1DM have a temporary but appreciable residual insulin secretion (remission phase), which may disguise attributes associated with poor glycaemic control. We observed that 6 months after diagnosis a third of T1DM children were in the remission phase, but by 24 months virtually no children remained in this phase. These data are consistent with values obtained in other studies [15], [16], [42], [43].

In the months after diagnosis clinicians may not adjust for the impact of the remission phase. In this period glycemic control can appear satisfactory, but it is in fact too high for the remission phase. Other studies have not examined changes in HbA1c levels from within and then beyond this remission phase. Thus, in the first few months following diagnosis, most children with T1DM appear to have good glycaemic control, creating an illusion that their diabetes is well managed. Once the remission phase is over, HbA1c levels are more likely to reflect actual diabetes management, rather than endogenous insulin secretion.

The end of this remission phase unmasks non-compliant patients, which show a dramatic increase in HbA1c levels, signalling poor glycaemic control. In our study, although Pacific Island children initially had only a slightly higher HbA1c 6mo than European children, this difference was considerably amplified after the remission phase. As a result, there was a marked deterioration in glycaemic control 24 months after diagnosis among Pacific Island children, whose mean HbA1c 24mo was 9.9% compared to 8.1% among Europeans. Importantly, the remission phase can be prolonged with more intensive insulin treatment and could be of considerable benefit to Pacific Island children [44], [45]. Interestingly, the remission phase among Pacific Island children was very short, and 83% of subjects transitioned out of this phase within 6 months of diagnosis which occurred more than twice as quickly as for European children. Thus, inadequate diabetes management characteristics were likely to have been present among Pacific Island children shortly after diagnosis.

Obesity among Pacific Islanders is extremely high, and a recent representative study of 1011 Pacific Island adults in Auckland showed that 74% of women and 53% of men were obese [46]. Obesity among Pacific Island children is also very common (approximately 24–26%) and rates are considerably higher than in other ethnic groups [47], [48]. The rate of obesity in diabetic Pacific Island children in this study was even higher (67%), and we speculate that it may be a major contributor to their poor glycaemic control and shorter remission phase. However, obesity in European children was found to be only weakly associated with HbA1c.

Pacific Islanders suffer from greater morbidity and mortality than other ethnic groups in New Zealand [49], with greater prevalence of diabetes and metabolic syndrome than Europeans [50]. A study of over 13,000 patients with T1DM and T2DM attending general practices in the South Island showed that Pacific Islanders/Maori also had poorer glycaemic control than Europeans [51]. Although genetics may be a factor [52], cultural and environmental factors other than SES are likely to be at play as observed elsewhere [33], [53], [54], and Pacific Islanders in this study are likely to be representative of minority ethnicities in other countries.

We also found that the prevalence and severity of DKA at diagnosis was considerably greater among girls. Further, girls had worse glycaemic control than boys, a pattern that has been previously observed [29]. Differences in metabolism and/or pubertal status between the sexes may account for these differences, and Hoffman et al. showed that early pubertal girls were less insulin sensitive than boys, but that this difference was compensated by increased insulin secretion [55].

Conclusions

Deterioration in glycaemic control typically occurs from 6 months to 24 months after diagnosis in children with T1DM. This pattern is particularly marked in Pacific Island children and in those not living with both biological parents. Clinicians need to be aware of factors associated with poor glycaemic control beyond the remission phase, so that more effective measures can be implemented shortly after diagnosis to prevent deterioration in diabetes management.

Acknowledgments

The Starbase Type 1 Diabetes Database was designed by C. Hira. We thank the diabetes nurse specialists Rosalie Hornung, Jean Ann Holt, Grace Harris, Rita Dowd and Sheryl Tregurtha, who helped with data collection, as well as Janene Biggs and Deb Rowe for assistance with data entry.

Author Contributions

Conceived and designed the experiments: SC WSC PLH CJ JGBD. Performed the experiments: SC. Analyzed the data: PR JGBD. Wrote the paper: SC WSC JGBD. Revised the manuscript and contributed to discussion: SC WSC PLH JGBD CJ PR.

References

  1. 1. American Diabetes Association (2007) Diagnosis and classification of diabetes mellitus. Diabetes Care 30: S42–S47.
  2. 2. The Diabetes Control and Complications Trial Research Group (1993) The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 329: 977–986.
  3. 3. UK Prospective Diabetes Study Group (1998) Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet 352: 837–853.
  4. 4. Silverstein J, Klingensmith G, Copeland K, Plotnick L, Kaufman F, et al. (2005) Care of children and adolescents with type 1 diabetes. Diabet Care 28: 186–212.
  5. 5. Auslander WF, Thompson S, Dreitzer D, White NH, Santiago JV (1997) Disparity in glycemic control and adherence between African-American and Caucasian youths with diabetes. Family and community contexts. Diabet Care 20: 1569–1575.
  6. 6. Carter PJ, Cutfield WS, Hofman PL, Gunn AJ, Wilson DA, et al. (2008) Ethnicity and social deprivation independently influence metabolic control in children with type 1 diabetes. Diabetologia 51: 1835–1842.
  7. 7. Chalew SA, Gomez R, Butler A, Hempe J, Compton T, et al. (2000) Predictors of glycemic control in children with type 1 diabetes: the importance of race. J Diabet Complications 14: 71–77.
  8. 8. Delamater AM, Albrecht DR, Postellon DC, Gutai JP (1991) Racial differences in metabolic control of children and adolescents with type I diabetes mellitus. Diabet Care 14: 20–25.
  9. 9. Mortensen HB, Swift PG, Holl RW, Hougaard P, Hansen L, et al. (2009) Multinational study in children and adolescents with newly diagnosed type 1 diabetes: association of age, ketoacidosis, HLA status, and autoantibodies on residual beta-cell function and glycemic control 12 months after diagnosis. Pediatr Diabet 11: 218–226.
  10. 10. Frey MA, Templin T, Ellis D, Gutai J, Podolski CL (2007) Predicting metabolic control in the first 5 yr after diagnosis for youths with type 1 diabetes: the role of ethnicity and family structure. Pediatr Diabet 8: 220–227.
  11. 11. Gallegos-Macias AR, Macias SR, Kaufman E, Skipper B, Kalishman N (2003) Relationship between glycemic control, ethnicity and socioeconomic status in Hispanic and white non-Hispanic youths with type 1 diabetes mellitus. Pediatr Diabet 4: 19–23.
  12. 12. Springer D, Dziura J, Tamborlane WV, Steffen AT, Ahern JH, et al. (2006) Optimal control of type 1 diabetes mellitus in youth receiving intensive treatment. J Pediatr 149: 227–232.
  13. 13. Anderson BJ, Vangsness L, Connell A, Butler D, Goebel-Fabbri A, et al. (2002) Family conflict, adherence, and glycaemic control in youth with short duration Type 1 diabetes. Diabet Med 19: 635–642.
  14. 14. Cameron FJ, Skinner TC, de Beaufort CE, Hoey H, Swift PG, et al. (2008) Are family factors universally related to metabolic outcomes in adolescents with Type 1 diabetes? Diabet Med 25: 463–468.
  15. 15. Abdul-Rasoul M, Habib H, Al-Khouly M (2006) ‘The honeymoon phase’ in children with type 1 diabetes mellitus: frequency, duration, and influential factors. Pediatr Diabet 7: 101–107.
  16. 16. Bober E, Dundar B, Buyukgebiz A (2001) Partial remission phase and metabolic control in type 1 diabetes mellitus in children and adolescents. J Pediatr Endocrinol Metab 14: 435–441.
  17. 17. Chase HP, MacKenzie TA, Burdick J, Fiallo-Scharer R, Walravens P, et al. (2004) Redefining the clinical remission period in children with type 1 diabetes. Pediatr Diabet 5: 16–19.
  18. 18. Knip M, Puukka R, Kaar ML, Akerblom HK (1982) Remission phase, endogenous insulin secretion and metabolic control in diabetic children. Acta Diabetol Lat 19: 243–251.
  19. 19. Baker L, Kaye R, Root AW (1967) The early partial remission of juvenile diabetes mellitus: The roles of insulin and growth hormone. J Pediatr 71: 825–831.
  20. 20. Knip M, Puukka R, Käär M-L, Åkerblom H (1982) Remission phase, endogenous insulin secretion and metabolic control in diabetic children. Acta Diabetologica 19: 243–251.
  21. 21. Campbell-Stokes PL, Taylor BJ (2005) Prospective incidence study of diabetes mellitus in New Zealand children aged 0 to 14 years. Diabetologia 48: 643–648.
  22. 22. Statistics New Zealand (2006) 2006 Census Data: QuickStats About Auckland Region. Auckland: Statistics New Zealand.
  23. 23. Douglas NM, Dockerty JD (2007) Survival by ethnicity for children diagnosed with cancer in New Zealand during 1990-1993. J Paediatr Child Health 43: 173–177.
  24. 24. Salmond C, Crampton P (2002) NZDep2001 Index of Deprivation Users' Manual. Wellington: Department of Public Health, University of Otago.
  25. 25. Salmond C, Crampton P, Atkinson J (2007) NZDep2006 Index of Deprivation. Wellington: Department of Public Health, University of Otago.
  26. 26. New Zealand Health Information Service (2010) Domicile code table. Wellington: Ministry of Health.
  27. 27. Delderfield M (2005) SDS Individual Calculator for British 1990 Growth Reference Data v1.1. Public Health Informatics, University of Manchester.
  28. 28. Wolfsdorf J, Craig ME, Daneman D, Dunger D, Edge J, et al. (2009) Diabetic ketoacidosis in children and adolescents with diabetes. Pediatr Diabetes 10: Suppl 12118–133.
  29. 29. Gerstl EM, Rabl W, Rosenbauer J, Grobe H, Hofer SE, et al. (2008) Metabolic control as reflected by HbA1c in children, adolescents and young adults with type-1 diabetes mellitus: combined longitudinal analysis including 27,035 patients from 207 centers in Germany and Austria during the last decade. Eur J Pediatr 167: 447–453.
  30. 30. Povlsen L, Olsen B, Ladelund S (2005) Diabetes in children and adolescents from ethnic minorities: barriers to education, treatment and good metabolic control. J Adv Nurs 50: 576–582.
  31. 31. Hoey H, Aanstoot HJ, Chiarelli F, Daneman D, Danne T, et al. (2001) Good metabolic control is associated with better quality of life in 2,101 adolescents with type 1 diabetes. Diabet Care 24: 1923–1928.
  32. 32. Williams DR (1999) Race, socioeconomic status, and health. The added effects of racism and discrimination. Ann NY Acad Sci 896: 173–188.
  33. 33. Kington RS, Smith JP (1997) Socioeconomic status and racial and ethnic differences in functional status associated with chronic diseases. Am J Public Health 87: 805–810.
  34. 34. Delamater AM, Shaw KH, Applegate EB, Pratt IA, Eidson M, et al. (1999) Risk for metabolic control problems in minority youth with diabetes. Diabet Care 22: 700–705.
  35. 35. Davis CL, Delamater AM, Shaw KH, La Greca AM, Eidson MS, et al. (2001) Parenting styles, regimen adherence, and glycemic control in 4- to 10-year-old children with diabetes. J Pediatr Psychol 26: 123–129.
  36. 36. Auslander WF, Anderson BJ, Bubb J, Jung KC, Santiago JV (1990) Risk factors to health in diabetic children: a prospective study from diagnosis. Health Soc Work 15: 133–142.
  37. 37. Thompson SJ, Auslander WF, White NH (2001) Comparison of single-mother and two-parent families on metabolic control of children with diabetes. Diabet Care 24: 234–238.
  38. 38. Kaufman FR, Halvorson M, Carpenter S (1999) Association between diabetes control and visits to a multidisciplinary pediatric diabetes clinic. Pediatrics 103: 948–951.
  39. 39. Overstreet S, Holmes CS, Dunlap WP, Frentz J (1997) Sociodemographic risk factors to disease control in children with diabetes. Diabet Med 14: 153–157.
  40. 40. Swift EE, Chen R, Hershberger A, Holmes CS (2006) Demographic risk factors, mediators, and moderators in youths' diabetes metabolic control. Ann Behav Med 32: 39–49.
  41. 41. Viswanathan V, Sneeringer MR, Miller A, Eugster EA, DiMeglio LA (2011) The utility of hemoglobin A1c at diagnosis for prediction of future glycemic control in children with type 1 diabetes. Diabet Re Clin Pract 92: 65–68.
  42. 42. Muhammad BJ, Swift PG, Raymond NT, Botha JL (1999) Partial remission phase of diabetes in children younger than age 10 years. Arch Dis Child 80: 367–369.
  43. 43. Dost A, Herbst A, Kintzel K, Haberland H, Roth CL, et al. (2007) Shorter remission period in young versus older children with diabetes mellitus type 1. Exp Clin Endocrinol Diabet 115: 33–37.
  44. 44. Shah SC, Malone JI, Simpson NE (1989) A randomized trial of intensive insulin therapy in newly diagnosed insulin-dependent diabetes mellitus. N Engl J Med 320: 550–554.
  45. 45. Mirouze J, Selam JL, Pham TC, Mendoza E, Orsetti A (1978) Sustained insulin-induced remissions of juvenile diabetes by means of an external artificial pancreas. Diabetologia 14: 223–227.
  46. 46. Sundborn G, Metcalf PA, Gentles D, Scragg R, Dyall L, et al. (2010) Overweight and obesity prevalence among adult Pacific peoples and Europeans in the Diabetes Heart and Health Study (DHAHS) 2002–2003, Auckland New Zealand. NZ Med J 123: 4036.
  47. 47. New Zealand Ministry of Social Development (2010) The Social Report. Wellington: http://www.socialreport.msd.govt.nz/inde​x.html.
  48. 48. Tyrrell V, Richards G, Hofman P, Gillies G, Robinson E, et al. (2001) Obesity in Auckland school children: a comparison of the body mass index and percentage body fat as the diagnostic criterion. Int J Obesity 25: 164–169.
  49. 49. Sopoaga F, Buckingham K, Paul C (2010) Causes of excess hospitalisations among Pacific peoples in New Zealand: implications for primary care. J Prim Health Care 2: 105–110.
  50. 50. Simmons D, Thompson CF (2004) Prevalence of the metabolic syndrome among adult New Zealanders of Polynesian and European descent. Diabet Care 27: 3002–3004.
  51. 51. Tomlin A, Tilyard M, Dawson A, Dovey S (2006) Health status of New Zealand European, Maori, and Pacific patients with diabetes at 242 New Zealand general practices. NZ Med J 119: U2004.
  52. 52. Elliott RB (1992) Epidemiology of diabetes in Polynesia and New Zealand. Pediatr Adolescent Endocrinol 21: 66–71.
  53. 53. Hardy D, Chan W, Liu C-C, Cormier JN, Xia R, et al. (2011) Racial disparities in the use of hospice services according to geographic residence and socioeconomic status in an elderly cohort with nonsmall cell lung cancer. Cancer 117: 1506–1515.
  54. 54. Lee R, Onopa J, Mau MK, Seto TB (2010) Diabetes care in a predominantly Native Hawaiian and Pacific Islander outpatient population. Hawaii Medi J 69: 28–30.
  55. 55. Hoffman RP, Vicini P, Sivitz WI, Cobelli C (2000) Pubertal adolescent male-female differences in insulin sensitivity and glucose effectiveness determined by the one compartment minimal model. Pediatr Res 48: 384–388.