Rising Rates of All Types of Diabetes in South Asian and Non-South Asian Children and Young People Aged 0–29 Years in West Yorkshire, U.K., 1991–2006

Data on individuals aged 0–29 years diagnosed with any type of diabetes (excluding gestational) from 1 January 1991 to 31 December 2006 were extracted from clinical notes by a dedicated Register Manager and entered onto the population-based Yorkshire Register of Diabetes in Children and Young People for residents of West Yorkshire, which has a 97.6% level of completeness (1).

Type of diabetes was defined in accordance with World Health Organization guidelines (10) as type 1, type 2, maturity-onset diabetes of the young (MODY), “J”-type/other, and uncertain/unclassified. Original diagnoses of uncertain/unclassified were reclassified as type 1 for 18% (n = 32) and type 2 for 11% (n = 20). Analyses are based on the final diagnosis type and first diagnosis date.

Ethnicity was assigned as south Asian (Pakistani, Indian, Bangladeshi) or non-south Asian (all other ethnicities) using the South Asian Names Analysis Program (Nam Pehchan) and the South Asian Name and Group Recognition Algorithm (SANGRA), two independent name recognition programs that are validated and sensitive tools for identifying south Asians (11,12). Where the programs disagreed (<1%), a local expert examined the names and assigned ethnicity.

Incidence was defined as the number of newly diagnosed cases per calendar year and 100,000 population at risk. Age-sex standardized rates were derived using indirect standardization. Populations for West Yorkshire were provided by the Office for National Statistics, supplemented by estimated midyear population denominators by age, sex, and ethnic group. Three-year moving average rates were calculated to assess trends in incidence.

Effects on incidence of age at diagnosis (5-year groups), sex, ethnicity (south Asian or not), and year of diagnosis on all types, type 1, type 2 and uncertain/unclassified diabetes were analyzed using Poisson regression. Average annual percentage change (AAPC) and 95% CIs were derived using the coefficient for year in the Poisson model. All analyses were done in Stata 11 software (StataCorp LP, College Station, TX).

Diabetes was diagnosed in 2,889 children and young people, with 83% (n = 2,410) type 1, 12% (n = 337) type 2, 0.7% (n = 19) MODY, 0.1% (n = 1 of each) “J”-type or other, 0.1% (n = 4) not recorded, and 4% (n = 117) uncertain/unclassified.

The incidence (95% CI) of all diabetes (aged 0–29 years) was 21.0 (20.2–21.7) per 100,000, with no statistically significant differences between south Asians and non-south Asians. The incidence of type 1 diabetes was 17.6 (16.9–18.3), with significantly higher rates in non-south Asians (18.3 [18.3–19.0]) than in south Asians (13.1 [11.4–14.8]). Type 2 incidence was 2.5 (2.2–2.7), which was significantly lower for non-south Asians (1.8 [1.6–2.1]) than south Asians (6.9 [5.6–8.1]). Incidence of uncertain/unclassified diabetes was 0.8 (0.67–0.98) and was lower for non-south Asians.

AAPC was 4.0% (95% CI 3.2–4.8) overall, and significantly greater in south Asians (7.0% [4.8–9.2]) compared with non-south Asians (3.6% [2.7–4.4]) (Fig. 1) and females (5.3% [4.1–6.4]) compared with males (3.0% [1.9–4.0]). For type 1 diabetes, AAPC was 1.3% (0.4–2.2) overall, with a significant increasing trend seen in non-south Asians (1.4% [0.5–2.3]) compared with no trend in south Asians (0.0% [−2.8 to 2.8]). Annual incidence increased significantly for children (aged 0–14 years) at 2.9% (1.7–4.0), but was stable for young people aged 15–29 years (0.6% [−1.9 to 0.7]). For type 2, the incidence in those aged younger than 30 years rose from 0.1 to 4.9 per 100,000 between 1991 and 2006 and AAPC was 19.4% (16.5–22.2) overall, with incidence increasing at a similar rate for both ethnic groups. Incidence increased at a faster rate for children aged 0–14 (37.4% [21.5–53.2]) compared with those aged 15–29 (18.1% [15.3–21.0]).

The proportion of uncertain/unclassified cases per year rose from 0% in 1991 to 8% in 2006. AAPC was 26.4% (95% CI 20.7–32.1) overall, with similar increases for both ethnic groups.

An increasing burden of diabetes was observed in non-south Asian and south Asian children and young people, with a rising incidence of type 1 diabetes confirmed in children (3,4) and stable rates for young adults (1,2).

Although the proportion of south Asians in the diabetes population (13%) was similar to the background population, south Asians had a lower incidence of type 1 diabetes but an excess of type 2, with an apparent shift from diagnosis of type 1 to type 2 from 2000 onwards. The previously reported steep rise in incidence of type 1 diabetes in south Asian children up to 1999 (5) was not sustained, possibly as a result of a change in the recognition that type 2 and other forms of diabetes are emerging in younger age groups. All south Asian children diagnosed with type 2 diabetes between 1991 and 2006 were 10 years or older, and a fall in type 1 in those 15 to 29 years old in the second period is unlikely to be attributable to changes in diagnostic classification.

There are clinical challenges in the differential diagnosis of type 1 and type 2 diabetes (13,14). Treatment can be based on symptoms with no clear diagnosis, reflected in rising numbers of uncertain/unclassified diagnoses (two diagnoses in the first 5 years, with 91 thereafter). We acknowledge that we may have underestimated the incidence of type 2 especially in earlier years, a problem identified in an earlier survey (15), but these observations confirm the importance of including all forms of diabetes in our analyses.

This work was supported by funding from NHS Diabetes administered by the Yorkshire and Humber Strategic Health Authority.

No potential conflicts of interest relevant to this article were reported.

K.L.H. analyzed data and wrote the manuscript. R.G.F. reviewed the manuscript. P.A.M. reviewed and edited the manuscript. H.J.B. and F.M.C. contributed to discussion. R.C.P. proposed analysis, reviewed and edited the manuscript, and contributed to discussion.

All the pediatricians, physicians, and general practitioners in Yorkshire are thanked for their continuing collaboration, and the authors are grateful for the active support of all the Diabetes Specialist Nurses. The authors thank Mark Hannigan, National Paediatric Workstream Lead, NHS Diabetes, and Irfan Mohammed, Portfolio Manager, North of England, NHS Diabetes for their support, and Gurjit Chhokar for his help with assigning ethnicity.