Type 2 diabetes mellitus (T2DM) in children and adolescents (hereinafter, youth) is becoming increasingly common worldwide.1 2 A recent meta-analysis estimated that approximately 41 600 new cases of T2DM were identified among youth in 2021.3 Type 2 DM in youth exhibits relatively rapid clinical progression with a sharp decline in beta-cell function and high risk of complications.4 In a study recently published by the TODAY (Type 2 Diabetes in Adolescents and Youth) Study Group, which analysed 500 young adults with youth-onset T2DM, 60.1% of patients developed at least one microvascular complication (diabetic kidney, nerve, or retinal disease) and 28.4% of patients developed at least two complications.5 In addition to hyperglycaemia, the presence of co-morbidities (eg, hypertension and dyslipidaemia) was associated with an increased risk of complications.5

A similar increase in the incidence of T2DM has been observed in Hong Kong. We previously reported that the crude incidence rate increased from 1.27 per 100 000 person-years in 1997-2007 to 3.42 per 100 000 person-years in 2008-2017.6 However, there have been limited data regarding the outcomes of paediatric T2DM patients in Hong Kong. In this study, we reviewed the glycaemic control findings and microvascular complication rates among recently diagnosed paediatric T2DM patients in Hong Kong, with a focus on outcomes at 2 years after diagnosis; we sought to identify factors associated with poor glycaemic control and the development of microalbuminuria (MA).

Data analysed in this study were retrieved from the Hong Kong Childhood Diabetes Registry, a database established in 2016. The Registry was approved by the Research Ethics Committee of the Hospital Authority of Hong Kong, which includes 11 public hospitals. Investigators retrieved information from medical records and entered relevant data into the Registry. Standardised data entry forms for recording baseline clinical characteristics and annual entry forms were provided for investigators to enter data into the Registry at diagnosis and annually thereafter. Data were cross-checked by at least two investigators.

All patients aged <18 years who had been diagnosed with DM at public hospitals in Hong Kong were recruited. All recruited patients met the diagnostic criteria for DM according to the International Society for Paediatric and Adolescent Diabetes Clinical Practice Consensus Guidelines: patients were symptomatic and had either fasting blood glucose level ≥7 mmol/L, 2-hour blood glucose level ≥11.1 mmol/L during an oral glucose tolerance test, random blood glucose level ≥11.1 mmol/L, or haemoglobin A1c (HbA1c) level ≥6.5%.4 Asymptomatic patients underwent repeat testing with a different test, as suggested in the Guidelines.4 The classification of DM was based on the attending clinician’s assessment of clinical symptoms and laboratory findings, including obesity status, family history, autoimmunity, and clinical course. Patients diagnosed with T2DM from 2014 to 2018 were included in the analysis, including those who had an initial diagnosis of type 1 DM that was subsequently revised to T2DM. Patients who refused Registry recruitment and patients whose diagnosis was revised to type 1 DM or maturity-onset DM of the young were not included in the analysis.

The following data were retrieved from the Registry: patient age, sex, family history of T2DM (in first- or second-degree relatives), symptoms at presentation, anti-islet cell antibody test results, body mass index (BMI), HbA1c level, presence of co-morbidities (non-alcoholic fatty liver disease, dyslipidaemia, hypertension, and obstructive sleep apnoea), presence of complications (MA, retinopathy, and neuropathy), treatments received, and frequency of blood glucose self-monitoring. Overweight and obesity were defined using age- and sex-specific cut-offs established by the International Obesity Task Force, which predicted BMI values at 18 years (25, 30, and 35 kg/m²) by the respective standard deviations to define overweight, obesity and morbid obesity, respectively; standard deviations of BMI were calculated according to age- and sex-specific reference data provided by the International Obesity Task Force.7 In this study, weight loss was defined as any decrease in BMI z-score, and improvement in HbA1c level was defined as any decrease in HbA1c level. Non-alcoholic fatty liver disease was defined as an elevated alanine transferase level (based on age- and sex-specific reference data) or compatible ultrasound findings. Dyslipidaemia was defined as an elevated low-density lipoprotein level of ≥2.6 mmol/L, a triglyceride level ≥1.7 mmol/L, or the receipt of lipid-lowering agents. Hypertension was defined as an elevated systolic blood pressure ≥95th percentile for age, height, and sex—on at least two occasions—or the receipt of anti-hypertensive medication. Obstructive sleep apnoea was defined as the presence of clinical symptoms indicating sleep-disordered breathing, along with polysomnography findings of obstructive apnoeas/hypopneas. Microalbuminuria was defined as an elevated spot urine albumin-creatinine ratio >2.5 mg/mmol for boys and >3.5 mg/mmol for girls in at least two of three samples within a 6-month period, or as the receipt of any treatment for MA. Retinopathy (eg, non-proliferative and proliferative diabetic retinopathy, as well as macula oedema) was identified by digital fundus photography and confirmed via referral to an ophthalmologist. Neuropathy was clinically identified by the presence of symptoms (numbness and paraesthesia) and through clinical examinations including the 10-g monofilament test, vibration sense assessment, and ankle reflex evaluation. Suboptimal glycaemic control was defined as HbA1c level ≥7%, as suggested by the International Society for Paediatric and Adolescent Diabetes Clinical Practice Consensus Guidelines.4

Statistical analyses were performed using SPSS software (Windows version 23; IBM Corp, Armonk [NY], United States). All available data were included in the statistical analysis, and the numbers of available values are listed in the tables. Continuous variables, including age, HbA1c level, and BMI z-score, were tested for normality using the Shapiro–Wilk test. Data with skewed distributions were expressed as medians and interquartile ranges (IQRs), and comparisons were made using the Mann-Whitney U test. Analyses of relationships between two continuous variables were assessed by Spearman rank correlation and expressed using the Spearman correlation coefficient (ρ). Categorical variables were expressed as exact numbers of patients with percentages. Binary logistic regression analysis was used to assess risk factors for suboptimal glycaemic control at 2 years and MA at 2 years. Univariate analyses were performed to determine unadjusted odds ratios and 95% confidence intervals. Multivariate analyses of factors associated with suboptimal glycaemic control at 2 years were performed while including HbA1c level at diagnosis to adjust for initial disease severity. Multivariate analyses of factors associated with MA at 2 years were performed while including HbA1c level at 2 years to eliminate the effect of glycaemic control at 2 years; this approach was intended to independently assess the effects of co-morbidities. Missing data were not included in regression analyses. Statistical tests were two-sided and were performed with a 5% significance threshold (ie, alpha=0.05). The STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) checklist for cohort studies was used when reporting the study findings.

In total, 212 patients diagnosed with T2DM between 2014 and 2018 were recruited into the Registry. Their baseline demographics are summarised in Tables 1 and 2. Of these patients, 71.3% had a family history of T2DM, and 21.7% were symptomatic at diagnosis (Table 1). At 2 years after the diagnosis of DM, 143 patients (67.5%) continued attending follow-up visits. There were no significant differences in baseline characteristics between the groups with and without follow-up at 2 years, except for a higher age at diagnosis in the loss to follow-up group (online supplementary Table).

Haemoglobin A1c levels at diagnosis, 1 year after diagnosis, and 2 years after diagnosis were available for 203, 176, and 135 patients, respectively. The median HbA1c levels at diagnosis, 1 year after diagnosis, and 2 years after diagnosis were 7.5% (IQR=6.5%-10.6%), 6.3% (IQR=5.8%-7.4%), and 6.5% (IQR=5.8%-8.0%), respectively; at these times, 65.5%, 29.0%, and 40.7% of patients had suboptimal glycaemic control (ie, HbA1c level ≥7%), respectively [Table 2].

There was an overall improvement in BMI z-score at 2 years after diagnosis (median BMI z-score decreased from 2.5 at diagnosis to 2.3 at 2 years). Overall, 146 of 191 patients (76.4%) had dyslipidaemia at diagnosis, whereas 62 of 95 patients (65.3%) had dyslipidaemia at 2 years. However, more patients had hypertension at 2 years—the number increased from 45 of 212 patients (21.2%) at diagnosis to 55 of 143 patients (38.5%) at 2 years (Table 2).

Overall, 21 of 113 (18.6%) patients screened at 2 years after diagnosis developed MA, compared with 9.0% at diagnosis. Two patients (1.8%) developed retinopathy, whereas one patient (0.9%) developed neuropathy, at 2 years after diagnosis (Table 2).

At diagnosis, 24.1% of patients were not receiving any pharmacological treatment, 58.0% were receiving anti-diabetic drugs, and 18% required insulin. At 2 years, only 12.6% of patients were not receiving any medications. The proportions of patients requiring insulin were similar at diagnosis and 2 years (2.1%). In total, 64.9% of patients did not perform daily blood glucose self-monitoring (Table 2).

A higher initial HbA1c level was associated with suboptimal glycaemic level at 2 years (correlation coefficient=0.39, P<0.001; n=130). There were no significant correlations of HbA1c level at 2 years with age at diagnosis (correlation coefficient=0.02, P=0.852; n=135), BMI z-score at diagnosis (correlation coefficient=-0.10, P=0.277; n=133), or BMI z-score at 2 years (correlation coefficient=0.04, P=0.638; n=131). Greater decline in BMI z-score was associated with a lower HbA1c level at 2 years (correlation coefficient=-0.22, P=0.011; n=129). However, there was no correlation between the change in BMI z-score and the change in HbA1c level (correlation coefficient <0.01, P=0.973; n=126).

Table 3 shows factors associated with suboptimal glycaemic control at 2 years. The effect of a family history of T2DM was not statistically significant after adjustment for initial HbA1c level (adjusted odds ratio [aOR]=2.29; P=0.075). A similar result was observed regarding the effect of symptomatic disease at diagnosis (aOR=2.01; P=0.174) and weight loss (aOR=0.53; P=0.128). The presence of fatty liver (aOR=2.50; P=0.021) and dyslipidaemia (aOR=3.19; P=0.033) at 2 years were associated with suboptimal glycaemic control at 2 years, even after adjustment for initial HbA1c level.

Patients with MA had higher HbA1c levels at 2 years compared with patients who did not exhibit MA (median HbA1c level: 7.2% vs 6.4%; P=0.037) [Table 4]. Dyslipidaemia at 2 years was associated with MA at 2 years in the univariate analysis (unadjusted odds ratio=5.51; P=0.030), but the effect did not remain statistically significant after adjusting for glycaemic control at 2 years (Table 5). The presence of hypertension at 2 years was a risk factor for MA at 2 years, independent of glycaemic control at 2 years (aOR=4.61; P=0.008) [Table 5].

The results of this study provide insights into the early post-diagnosis clinical course of T2DM among youth in Hong Kong. Nearly 60% of patients (59.3%) in our cohort had optimal glycaemic control with HbA1c level <7% at 2 years after diagnosis. Previous studies regarding glycaemic control among youth with T2DM showed variable results, presumably due to heterogeneity in the study populations, follow-up periods, and glycaemic targets.8 9 10 11 A clinical trial by the TODAY Study Group8 followed up 234 youth with DM, who were put on metformin and lifestyle modifications and with initial HbA1c level <8%, for 3.86 years on average. It showed that 46.6% of youth with DM exhibited loss of glycaemic control, defined by HbA1c level >8%.8 In a study of 301 paediatric T2DM patients with initial HbA1c level ≥7% in the United States, Barr et al9 found that after 1 year, 37% of patients achieved optimal control (HbA1c level ≤6.5%) and 58% achieved durable glycaemic control (HbA1c level ≤8%). However, at 3 years, only 26% of patients achieved HbA1c level ≤6.5%, whereas 59% exhibited HbA1c level ≤8%.9 Candler et al10 followed 100 paediatric T2DM patients in the United Kingdom; the median HbA1c level was 7% after 1 year, and 38.8% of patients exhibited HbA1c level <6.5%. Notably, no data regarding HbA1c levels at diagnosis were provided in the study.10 In a study of 96 patients in Israel, Meyerovitch et al11 found that the median HbA1c level was 7.97% after an average follow-up period of 3.11 years, compared with 7.8% at diagnosis. Additionally, >50% of patients required insulin at the end of the follow-up period.11 Although our cohort appeared to have better glycaemic control compared with the previous studies, our patients might have had lower initial HbA1c levels at diagnosis, considering that most of them were asymptomatic (78.3%). Our study also showed that patients with a higher initial HbA1c level tended to have a persistently high HbA1c level at 2 years. These findings emphasise the importance of early diagnosis and treatment before patients develop clinically significant hyperglycaemia, which makes DM more difficult to control. Most of our patients were overweight or obese (89.8%); many of them also had co-morbidities including hypertension, dyslipidaemia, and fatty liver (Table 2). Previous studies in Hong Kong showed a high risk of metabolic syndrome (OR up to 32.2) in overweight and obese children.12 13 Active screening for metabolic syndrome would enable early diagnosis and treatment of DM and its co-morbidities.

Factors associated with suboptimal glycaemic control were dyslipidaemia and fatty liver at 2 years after diagnosis. A recent study showed that each 1% increase in HbA1c level was associated with 13% and 20% increases in the risks of abnormal triglyceride and low-density lipoprotein levels, respectively.14 The importance of weight loss has been emphasised in various guidelines, for example, The American Diabetes Association recommends weight loss of at least 5% in adult overweight or obese DM patients.15 However, a specific weight loss target cannot be established in growing children. The study by Candler et al10 regarding youth with T2DM showed that each one-unit increase in BMI z-score was associated with a 34.9% increase in HbA1c level. Although the present study indicated that a greater drop in BMI z-score was associated with lower HbA1c level at 2 years, the association between weight loss and prevention of suboptimal glycaemic control at 2 years was not significant after adjustment for initial HbA1c level.

Diabetic retinopathy and neuropathy were rare among youth with T2DM. However, the proportion of patients with MA increased from 9.0% at diagnosis to 18.6% at 2 years (Table 2). Previous studies regarding the prevalence of diabetic complications in youth have shown mixed results, probably due to genetic variation and differences in DM duration. High prevalences have been observed in cohorts with long DM durations.16 The MA prevalence has been approximately 20% to 30% in most studies of youth with a short duration of T2DM. In the SEARCH for Diabetes in Youth study, the MA prevalence in youth with T2DM was 22.2%, and the average duration of disease was 1.9 years.17 In an Australian population, Eppens et al18 found that 28% of patients had MA, with a median disease duration of 1.3 years. Candler et al10 showed that the MA prevalence increased from 4.2% to 16.4% within 1 year after diagnosis. Our cohort showed a similar prevalence compared with other cohorts. Nevertheless, the increasing trend is concerning, particularly because MA has been identified as an independent predictor of mortality risk in adults.19 Thus, we conducted further analysis of risk factors for MA, revealing the associations of higher HbA1c level and hypertension at 2 years, consistent with the SEARCH for Diabetes in Youth study.17 The deleterious effects of hypertension on the kidneys explain the additional increase in MA risk, independent of glycaemic control.

Strengths of this study included its provision of insights regarding the early outcomes of youth with T2DM in Hong Kong, particularly concerning glycaemic control and associated factors. First, our findings supported the implementation of active screening in overweight and obese individuals to allow early diagnosis of DM, considering the high prevalence of overweight or obesity and the relationship of lower initial HbA1c level with better glycaemic control at 2 years. Second, our findings indicated that the presence of co-morbidities at 2 years, rather than baseline, was associated with suboptimal glycaemic control and MA, demonstrating the reversibility of the risk factors and highlighting the importance of co-morbidity management. Third, our study identified challenges in managing youth with T2DM, including a high loss to follow-up rate (n=69, 23.5%) [online supplementary Table], suboptimal glycaemic control in >40% of patients at 2 years, infrequent blood glucose self-monitoring by the patients, and increasing trends in MA and hypertension.

Indeed, the high loss to follow-up rate was a major limitation of our study. Many patients did not return for clinical assessment or were transferred to an adult endocrinology clinic. Although the loss to follow-up group had a higher age at diagnosis (online supplementary Table), this presumably did not have a substantial impact on the results because age was not a significant risk factor for poor glycaemic control or the likelihood of MA onset. Although a high loss to follow-up rate is a common phenomenon in studies of children with T2DM,20 this obstacle hindered the achievement of good glycaemic control and prevention of complications. It also created difficulty in acquiring long-term follow-up data. Another limitation was that our patients were managed by different doctors in different hospitals; there remains no standardised protocol for the management of paediatric T2DM patients in Hong Kong, which may be a confounding factor for multicentre studies such as ours.

Approximately 60% of youth with T2DM in Hong Kong achieved HbA1c level <7% at 2 years after diagnosis. A higher HbA1c level at diagnosis was associated with worse glycaemic control at 2 years. The presence of dyslipidaemia and fatty liver at 2 years were factors associated with suboptimal glycaemic control. Overall, 18.6% of patients developed MA at 2 years; other microvascular complications were rare. These results highlight the importance of early diagnosis and holistic management, including co-morbidity management. The high loss to follow-up rate, high proportion of patients with suboptimal glycaemic control, and increasing number of patients with MA and hypertension are ongoing challenges in the management of youth with DM. The establishment of a standardised protocol may improve outcomes in our patient population. Future research could include studies regarding the effects of insulin resistance and beta-cell function on metabolic outcomes in youth with DM.

Concept or design: All authors.
Acquisition of data: All authors.
Analysis or interpretation of data: WI Yam, SMY Wong.
Drafting of the manuscript: WI Yam.
Critical revision of the manuscript for important intellectual content: All authors.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

All authors have disclosed no conflicts of interest.

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

The research was conducted as a part of the Hong Kong Childhood Diabetes Registry, which was approved by the Kowloon Central / Kowloon East Cluster Research Ethics Committee of Hospital Authority, Hong Kong (Ref No.: KC/KE-16-0087/ER-3). Written informed consent was obtained from parents for patients aged <16 years and from both parents and patients for patients aged between ≥16 and <18 years.

The supplementary material was provided by the authors and some information may not have been peer reviewed. Accepted supplementary material will be published as submitted by the authors, without any editing or formatting. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by the Hong Kong Academy of Medicine and the Hong Kong Medical Association. The Hong Kong Academy of Medicine and the Hong Kong Medical Association disclaim all liability and responsibility arising from any reliance placed on the content.

1. Lawrence JM, Divers J, Isom S, et al. Trends in prevalence of type 1 and type 2 diabetes in children and adolescents in the US, 2001-2017. JAMA 2021;326:717-27. Crossref

2. Haynes A, Kalic R, Cooper M, Hewitt JK, Davis EA. Increasing incidence of type 2 diabetes in indigenous and non-indigenous children in Western Australia, 1990-2012. Med J Aust 2016;204:303. Crossref

3. Wu H, Patterson CC, Zhang X, et al. Worldwide estimates of incidence of type 2 diabetes in children and adolescents in 2021. Diabetes Res Clin Pract 2022;185:109785. Crossref

4. Shah AS, Zeitler PS, Wong J, et al. ISPAD Clinical Practice Consensus Guidelines 2022: type 2 diabetes in children and adolescents. Pediatr Diabetes 2022;23:872-902. Crossref

5. TODAY Study Group; Bjornstad P, Drews KL, et al. Long-term complications in youth-onset type 2 diabetes. N Engl J Med 2021;385:416-26. Crossref

6. Tung JY, Kwan EY, But BW, et al. Incidence and clinical characteristics of pediatric-onset type 2 diabetes in Hong Kong: the Hong Kong Childhood Diabetes Registry 2008 to 2017. Pediatr Diabetes 2022;23:556-61. Crossref

7. Cole TJ, Lobstein T. Extended international (IOTF) body mass index cut-offs for thinness, overweight and obesity. Pediatr Obes 2012;7:284-94. Crossref

8. TODAY Study Group; Zeitler P, Hirst K, et al. A clinical trial to maintain glycemic control in youth with type 2 diabetes. N Engl J Med 2012;366:2247-56. Crossref

9. Barr MM, Aslibekyan S, Ashraf AP. Glycemic control and lipid outcomes in children and adolescents with type 2 diabetes. PLoS One 2019;14:e0219144. Crossref

10. Candler TP, Mahmoud O, Lynn RM, Majbar AA, Barrett TG, Shield JP. Treatment adherence and BMI reduction are key predictors of HbA1c 1 year after diagnosis of childhood type 2 diabetes in the United Kingdom. Paediatr Diabetes 2018;19:1393-9. Crossref

11. Meyerovitch J, Zlotnik M, Yackobovitch-Gavan M, Phillip M, Shalitin S. Real-life glycemic control in children with type 2 diabetes: a population-based study. J Paediatr 2017;188:173-80.e1. Crossref

12. Ozaki R, Qiao Q, Wong GW, et al. Overweight, family history of diabetes and attending schools of lower academic grading are independent predictors for metabolic syndrome in Hong Kong Chinese adolescents. Arch Dis Child 2007;92:224-8. Crossref

13. Kong AP, Ko GT, Ozaki R, Wong GW, Tong PC, Chan JC. Metabolic syndrome by the new IDF criteria in Hong Kong Chinese adolescents and its prediction by using body mass index. Acta Paediatr 2008;97:1738-42. Crossref

14. Brady RP, Shah AS, Jensen ET, et al. Glycemic control is associated with dyslipidemia over time in youth with type 2 diabetes: the SEARCH for Diabetes in Youth study. Pediatr Diabetes 2021;22:951-9. Crossref

15. American Diabetes Association Professional Practice Committee. 8. Obesity and weight management for the prevention and treatment of type 2 diabetes: standards of medical care in diabetes—2022. Diabetes Care 2022;45(Supp 1):S113-24. Crossref

16. Amutha A, Mohan V. Diabetes complications in childhood and adolescent onset type 2 diabetes—a review. J Diabetes Complications 2016;30:951-7. Crossref

17. Maahs DM, Snively BM, Bell RA, et al. Higher prevalence of elevated albumin excretion in youth with type 2 than type 1 diabetes: the SEARCH for Diabetes in Youth study. Diabetes Care 2007;30:2593-8. Crossref

18. Eppens MC, Craig ME, Cusumano J, et al. Prevalence of diabetes complications in adolescents with type 2 compared with type 1 diabetes. Diabetes Care 2006;29:1300-6. Crossref

19. Chronic Kidney Disease Prognosis Consortium; Matsushita K, van der Velde M, et al. Association of estimated glomerular filtration rate and albuminuria with all-cause and cardiovascular mortality in general population cohorts: a collaborative meta-analysis. The Lancet 2010;375:2073-81. Crossref

20. Shoemaker A, Cheng P, Gal RL, et al. Predictors of loss to follow-up among children with type 2 diabetes. Horm Res Paediatr 2017;87:377-84. Crossref