OUP user menu

Complications and cardiovascular risk factors in South Asians and Europeans with early‐onset type 2 diabetes

T.A. Chowdhury, S.S. Lasker
DOI: http://dx.doi.org/10.1093/qjmed/95.4.241 241-246 First published online: 1 April 2002


Background: Type 2 diabetes is a major cardiovascular risk factor, and early‐onset (<40 years) type 2 diabetes is becoming more common.

Aim: To determine the prevalence of complications, and cardiovascular risk factors at diagnosis, in early‐onset type 2 diabetes, and to compare these between South Asians and Europeans.

Design: Prospective study of newly‐diagnosed type 2 diabetes patients aged <40 years, attending hospital and primary care clinics 1999–2001.

Methods: Patients were assessed for signs of macrovascular disease, retinopathy, neuropathy and nephropathy. Cardiovascular risk factors were also determined.

Results: Overall, 292 patients were enrolled (165 South Asians). Macrovascular disease was more prevalent in South Asians (15.7% vs. 9.4%, p<0.001), as was microvascular disease (27.3% vs. 16.5%, p<0.001), including retinopathy (17.5% vs. 7.9%, p<0.001), and nephropathy (18.1% vs. 7.8%, p<0.001). South Asians had trends towards greater waist:hip ratio (0.95 vs. 0.90), and higher blood pressure (127/80 vs. 123/76 mmHg). HDL cholesterol was lower (1.0 vs. 1.3 mmol/l, p<0.001) and fasting triglycerides higher (1.9 vs. 1.5 mmol/l, p<0.001) in South Asians. Absolute CHD risk was significantly higher in South Asians (16.9% vs. 13.7%, p<0.001).

Discussion: Complications were common at diagnosis, with a quarter of all patients having evidence of at least one diabetic complication. South Asians had a higher prevalence of established macrovascular and microvascular disease, compared to Europeans, and a higher risk of CHD, predominantly because of lower HDL cholesterol and higher blood pressure.


The prevalence of type 2 diabetes is increasing world‐wide.1 In the UK, a significant proportion of the diabetic population is of South Asian descent (Indian, Pakistani, Bangladeshi and Sri Lankan), residing in the UK as a result of large‐scale immigration in the late 1960s. South Asians have a 40% higher mortality from coronary heart disease (CHD) than Europeans,2 and diabetes is a major risk factor for CHD in South Asians.3 Diabetes is common amongst UK and indigenous South Asians,4,,5 affecting around 20% of South Asian subjects over the age of 50 years. Diabetes also appears to present at a younger age in South Asians compared to Europeans.4,,6 There is considerable evidence that complications are commoner amongst South Asian diabetic patients, particularly with respect to cardiovascular and renal disease.7,,8 Mortality from CHD appears to be four times higher in South Asians living in the UK compared to White Europeans.9 Many of the risk factors for diabetic complications, such as poor glycaemic control, smoking, hypertension and dyslipidaemia, are amenable to intervention. Longer duration of diabetes is also a major risk factor for complications of diabetes. Patients with early‐onset type 2 diabetes are at particular risk of complications, due to longer exposure to hyperglycaemia, and are a group of patients in whom risk factor intervention may significantly attenuate the risk of complications. The United Kingdom Prospective Diabetes Study (UKPDS) previously examined differences between Asian, Afro‐Caribbean and White patients at diagnosis of diabetes, but the mean age at diagnosis in this study was over 50 years, and patients with established cardiovascular disease were excluded from the study.10 The aim of this study was to determine the prevalence of complications, and cardiovascular risk factors at diagnosis, in early‐onset (age <40 years) type 2 diabetes patients, and to compare the prevalence of complications and risk factors between Europeans and South Asians.


Ethical approval was obtained for the study, and full informed consent was obtained from all patients. All newly‐diagnosed diabetic patients attending a large hospital and three primary care diabetic clinics were prospectively screened over the period 1999–2001, and were followed up for at least 6 months. Diabetes was diagnosed on the basis of at least two fasting plasma glucoses >7.0 mmol/l or random plasma glucoses >11.1 mmol/l. No active screening policy was in place for screening for type 2 diabetes. Opportunistic screening took place in the primary care clinics, particularly in patients deemed at high risk, such as patients with a strong family history of diabetes, obese patients, those with previous gestational diabetes or South Asians.

Patients included in the study had type 2 diabetes diagnosed within the preceding 6 months, and were aged 20–40 years. Type 2 diabetes was defined by the lack of absolute requirement for insulin, absence of ketonuria, presence of a fasting C‐peptide >0.6 nM, and treatment without insulin, with adequate control, for at least 6 months from diagnosis.

All patients were fully assessed by one physician (TAC). The assessment included questions on their previous history of vascular disease (presence of chest pains, previous myocardial infarction (MI), angina, stroke or peripheral vascular disease (PVD) or symptoms of intermittent claudication). Women were asked about a previous history of gestational diabetes mellitus. All patients were questioned on their family history of diabetes or cardiovascular disease. All patients underwent a standard 12‐lead electrocardiogram (ECG), which was graded using a Minnesota code,11 and if abnormal, or if there was any history of chest pains, the patients underwent exercise electrocardiography. The ECG was also assessed for signs of left ventricular hypertrophy. Peripheral pulses were assessed by palpation of femoral, popliteal, posterior tibial and dorsalis pedis pulses. Peripheral vascular disease was present if one or more pulses were impalpable. Carotids were auscultated for presence of a bruit, to signify carotid disease. Presence of vascular disease was recorded if there was: (i) a history of MI, angina, stroke or PVD; (ii) presence of pathological Q waves on ECG (Minnesota codes 1.1, 1.2, 7.1) or ST segment depression (Minnesota codes 4.1, 4.2); (iii) abnormal exercise ECG; (iv) loss of peripheral pulse in either foot; or (v) presence of carotid bruit.

Signs of microvascular complications were sought. Neuropathy was assessed using a Semmes‐Weinstein 10 g filament over 10 areas of the feet, ankle reflexes, and vibration perception over the great toe and ankle. A standard neuropathy disability score (NDS) was determined, and deemed to show significant neuropathy if the score was >6.12 Neuropathy was present if there were three or more areas over the each foot with loss of 10 g fine touch sensation, or loss of vibration sense over the great toe or ankle, or a presence of painful neuropathy as determined by a neuropathy symptom score (NSS).12 Retinopathy was assessed by dilated direct and indirect ophthalmoscopy. Presence of background retinopathy was defined as mild non‐proliferative retinopathy with no maculopathy. Sight‐threatening retinopathy was defined as the presence of severe non‐proliferative or proliferative retinopathy, or maculopathy. Nephropathy was assessed by first voided urine dipstick testing for albuminuria, and albumin:creatinine ratio (ACR) on the first voided urine. If dipstick testing or ACR was abnormal, the test was repeated 4 weeks later, and if still abnormal, culture of urine, full blood count, erythrocyte sedimentation rate, full biochemical profile and renal ultrasound were done to exclude non‐diabetic causes of proteinuria. Three overnight timed urine collections were undertaken in all patients with albuminuria for estimation of albumin excretion rate (AER). Macroalbuminuria was diagnosed when two out of the three AERs were >200 μg/min, in the absence of other causes of albuminuria. Microalbuminuria was defined when two out of the three ACRs were 20–200 μg/min in the absence of other causes.

Blood pressure was recorded as the mean of three supine readings taken 5 min apart. Diastolic blood pressure was taken as Korotkov phase V. Blood was taken for urea and electrolytes, fasting total cholesterol, LDL cholesterol, HDL cholesterol and triglycerides, glycated haemoglobin, fasting plasma glucose, fasting insulin and C‐peptide. Absolute Coronary Heart Disease (CHD) risk was calculated using the Framingham Equation, with the Joint British Societies computer programme.13 Insulin sensitivity and β‐cell function was calculated using the homeostasis model assessment (HOMA).14

Data are presented as mean±SD unless otherwise stated. To examine differences in variables between the two groups, Student's t‐test was used. To compare the prevalence of complications or risk factors between the two groups, a χ2‐test was performed for discrete variables, and Mann‐Whitney test for continuous variables. Statistical significance was set at p=0.05. Statistical tests used the statistical package Minitab (Minitab Inc).


Of 1220 newly‐diagnosed patients seen, a total of 292 were suitable for the study, of whom 165 were South Asian. Demographic characteristics are shown in Table 1. The groups were equivalent in age and sex distribution. The South Asian cohort was made up of Indian (n=76, 46.1%), Sri Lankan (n=10, 6.0%), Pakistani (n=64, 38.8%) and Bangladeshi (n=15, 9.1%) ethnic groups. South Asians had more frequent family history of diabetes and vascular disease, and South Asian women had a more frequent past history of gestational diabetes than Europeans. Fasting plasma glucose and glycated haemoglobin at diagnosis were not significantly different between the two groups. β‐cell function, as calculated by the HOMA, was slightly higher in South Asians, but insulin sensitivity was lower. The presence of significant osmotic symptoms at diagnosis (polyuria, polydipsia, weight loss, thirst) was not significantly different between Europeans and South Asians.

The prevalence of macrovascular disease was higher in South Asians compared to Europeans (Table 2). Overall, 26 (15.7%) South Asian patients had evidence of macrovascular disease at diagnosis of diabetes, compared to 12 (9.4%) European patients (p<0.001). Microvascular complications were also more common amongst South Asian patients at diagnosis (Table 3). Overall, 45 (27.3%) South Asian patients had evidence of microvascular disease at diagnosis of diabetes, compared to 21 (16.5%) European patients (p<0.001). South Asians had significantly more retinopathy at diagnosis (17.5% vs. 7.9%), including sight‐threatening retinopathy (3.6% vs. 1.6%). South Asians also had a significantly higher prevalence of microalbuminuria and macroalbuminuria (18.1% vs. 7.8%), although serum creatinine was similar between the two groups. There was a trend towards more neuropathy in the European patients, although this was not significant.

Assessment of cardiovascular risk factors at diagnosis of diabetes is shown in Table 4. Similar numbers in the two groups were current smokers. South Asians had a slightly lower BMI, but significantly greater waist:hip ratio. South Asians showed a trend towards higher systolic and diastolic blood pressures, possibly due to more patients with nephropathy, but this was not significantly different. A total of 42 (25.5%) South Asian patients had a systolic blood pressure >140 mmHg or diastolic blood pressure >90 mmHg, compared to 29 (22.8%) European patients. Total cholesterol and LDL cholesterol levels were very similar between the two groups, but HDL cholesterol was significantly higher (1.3 vs. 1.0) and fasting triglycerides significantly lower (1.5 vs. 1.9) in European patients. Mean absolute 10‐year CHD risk was significantly higher in the South Asian group compared to the European group (16.9% vs. 13.7%), as was proportion of patients with CHD risk >30% (7.2% vs. 4.7%) and >15% (17.5% vs. 11.8%).

Analysing of the South Asian cohort separately by Indian, Sri Lankan, Pakistani and Bangladeshi ethnic groups showed no significant difference from the combined South Asian cohort in all clinical parameters.

View this table:
Table 1 

Demographic characteristics of European and South Asian cohorts at diagnosis of early‐onset type 2 diabetes

EuropeanSouth Asianp
Male (n (%)) 72 (56.7) 96 (58.2)  0.23
Age (years) 35.3 (2.3) 33.5 (2.4)  0.24
Family history of diabetes (n (%)) 41 (32.3) 86 (52.1)<0.001
Family history of early vascular disease (n (%)) 27 (21.2) 68 (41.2)<0.001
Previous gestational diabetes mellitus in females (n (%))  9 (16.3) 25 (36.2)<0.001
Fasting plasma glucose (mmol/l)  9.5 (2.6)  9.4 (2.4)  0.45
Glycated haemoglobin (%)  8.4 (1.9)  8.6 (1.4)  0.17
β‐cell function (%) 41.7 (10.2) 46.2 (9.6)<0.001
Insulin sensitivity (%) 32.8 (7.5) 28.7 (6.7)  0.05
Osmotic symptoms at diagnosis (n (%)) 53 (41.7) 67 (40.6)  0.69
View this table:
Table 2 

Prevalence of macrovascular disease at diagnosis of early‐onset type 2 diabetes

EuropeanSouth Asianp
History of previous vascular disease (n (%))  3 (3.1)  8 (4.8)   0.02
History of chest pain (n (%))  2 (1.6)  4 (2.4)  0.09
Ischaemia on ECG/Ex. ECG (n (%))  6 (4.7)  8 (4.8)  0.12
Previous vascular disease or chest pain/ischaemia on ECG/Ex. ECG (n (%)) 11 (8.7) 20 (12.1)<0.001
Intermittent claudication/lost peripheral pulse (n (%))  5 (3.9)  9 (5.4)  0.32
Carotid bruit (n (%))  1 (0.8)  1 (0.6)  0.76
Evidence of macrovascular disease at diagnosis (n (%)) 12 (9.4%) 26 (15.7%)<0.001
  • ECG, electrocardiogram; Ex., exercise.

View this table:
Table 3 

Prevalence of microvascular complications at diagnosis of early‐onset type 2 diabetes

EuropeanSouth Asianp
Neuropathy (n (%))  8 (6.3)  5 (3.0)  0.08
Microalbuminuria (n (%))  7 (5.5) 22 (13.3)<0.001
Macroalbuminuria (n (%))  3 (2.3)  8 (4.8)<0.001
Creatinine (μmol/l) 78 (10) 81 (12)  0.26
Creatinine >150 μmol/l (n (%))  2 (1.6)  4 (2.4)  0.08
Background retinopathy (n (%))  8 (6.3) 23 (13.9)<0.001
Sight‐threatening retinopathy (n (%))  2 (1.6)  6 (3.6)<0.001
View this table:
Table 4 

Cardiovascular risk factors at diagnosis of early‐onset type 2 diabetes

EuropeanSouth Asianp
Current smokers (n (%)) 28 (22.0) 39 (23.6)  0.46
Body mass index (kg/m2) 27.2 (5.3) 26.0 (5.4)  0.076
Waist:hip ratio  0.90 (0.2)  0.95 (0.2)  0.05
Systolic BP (mmHg)123 (12)127 (10)  0.12
Diastolic BP mean (mmHg) 76 (7) 80 (5)  0.09
LVH on ECG (n (%))  2 (1.6)  2 (1.2)  0.53
Total cholesterol (mmol/l)  5.4 (1.2)  5.2 (1.3)  0.38
LDL cholesterol (mmol/l)  3.5 (0.7)  3.3 (0.8)  0.19
HDL cholesterol (mmol/l)  1.3 (0.2)  1.0 (0.3)<0.001
Fasting triglycerides (mmol/l)  1.5 (0.4)  1.9 (0.5)<0.001
Absolute 10‐year CHD risk (%) 13.7 (4.6) 16.9 (5.4)<0.001
CHD risk >30% (n (%))  6 (4.7) 12 (7.2)  0.01
CHD risk >15% (n (%)) 18 (14.1) 35 (21.2)<0.001


A significant proportion of our patients with early‐onset type 2 diabetes diagnosed at age <40 years had established complications, with about a quarter showing some sign of a diabetic complication at diagnosis. The study also demonstrates that South Asians with early‐onset type 2 diabetes have significantly more complications at diagnosis compared to Europeans, including established macrovascular disease, retinopathy and nephropathy. Indeed, this study used simple clinical criteria for the definition of macrovascular disease, and there may have been a degree of under‐ascertainment of macrovascular disease.

Many of our patients had risk factors for complications that were amenable to treatment. A significant proportion of patients smoked, had blood pressure above treatment thresholds, or had poor diabetic control and dyslipidaemia. South Asians appear to have a more atherogenic lipid profile (low HDL cholesterol, raised triglycerides), predisposing to atherogenesis, and intervention at an early stage is likely to have a significant impact on prevention and progression of complications, particularly CHD.

Patients with type 2 diabetes are at high risk of CHD.15 One of the strongest factors for the development of CHD is age. In this study, despite the comparatively young age of the patients, around one patient in ten had evidence of established CHD. In addition, around one in five patients had an absolute 10‐year CHD risk >15%, the threshold suggested as requiring intervention in most guidelines.13 The CHD risk was also significantly higher amongst South Asians, predominantly due to their lower HDL cholesterol, and higher blood pressure.

There are a number of studies suggesting that South Asians have a significantly higher risk of diabetic complications compared to White Europeans.4,,5 In particular, renal disease and CHD appear to be more common amongst South Asians with diabetes, as demonstrated by follow up of the Southall diabetes survey.7 The UKPDS, however, showed results at slight variance with these studies and our study results.10 In a survey of the patients entering the UKPDS, clinical characteristics of 4177 White European patients and 534 South Asian patients were compared. Mean age at diagnosis for White Europeans was 52.3 years, and for South Asians, 47.0 years, around 14 years older than our cohorts. Similarly to our findings, South Asian patients had a greater waist:hip ratio and lower insulin sensitivity. In the UKPDS, however, Asians did not appear to have more complications at diagnosis of diabetes compared to Europeans. Of note, however, is the fact that the UKPDS excluded patients with cardiovascular disease at the outset, and hence the cohort were to some extent subject to selection bias. In a survey of 597 Sri Lankan newly‐diagnosed diabetic subjects, very high levels of complications at diagnosis were observed. CHD was present in 21%, nephropathy in 29% and retinopathy in 15%.16 Indeed, the data from our study may have underestimated the degree of macrovascular disease, as only clinical examination was used to assess carotid and peripheral arterial disease, which may not be as sensitive as Doppler ultrasonography in detecting significant stenosis. A further drawback of the study is the small numbers of patients with the complications under study. As age and duration of disease are amongst the strongest risk factors for diabetes complications, this is unsurprising. This fact, however, should suggest care in interpretation of our results, although grouping complications into micro‐ and macrovascular appears to give similar results.

Diabetes amongst younger patients is an important problem, as these patients are likely to be at higher risk of complications due to longer hyperglycaemic exposure. As screening for diabetes becomes more widespread, it is likely that the age of diagnosis of diabetes is going to decline. Early screening for complications is important, as are effective interventions to reduce complications, such as improving glycaemic control,17 early treatment of hypertension,18 and correction of dyslipidaemia.13 This appears to be particularly important in South Asians with early‐onset type 2 diabetes. Our results suggest that South Asians have a higher prevalence of established diabetic complications, and also a higher prevalence of risk factors for the development of complications, including higher blood pressure and dyslipidaemia. Thus, early intervention to attenuate this risk is of particular importance in South Asian patients.


  • Address correspondence to Dr T.A. Chowdhury, Jeffrey Kelson Diabetic Centre, Central Middlesex Hospital, Acton Lane, London NW10 7NS. e‐mail: tahseen.chowdhurynwlh.nhs.uk


View Abstract