Early-onset type 2 diabetes : higher burden of atherogenic apolipoprotein particles during statin treatment

Aims: To determine the burden of atherogenic apolipoprotein particles in early-onset type 2 diabetes (T2D) compared to those with later-onset disease during statin treatment. Methods: Early and later-onset T2D was defined as current age below and above 40 years respectively. Conventional lipid profile, LDL, non-HDL cholesterol, apolipoprotein B and A1 were determined in those without cardiovascular disease treated with simvastatin to achieve LDL cholesterol <2 mmol/l. Results: Fifty subjects were recruited (early-onset n = 24 and later-onset n = 26). The mean age was 34.5 and 59.6 years and mean age of diagnosis was 29.1 and 49.1 years for early and later-onset T2D respectively. Obesity, dyslipidaemia, microalbuminuria, glycaemic control and diabetes complication burden were similar in both cohorts. Early-onset subjects received non-significantly higher simvastatin dose (37.5 vs. 31.9 mg daily, p = NS). On-treatment LDL cholesterol was similar in both cohorts (early vs. later-onset; 2.12 vs. 1.97 mmol/l, p = NS). Fasting triglyceride, nonHDL, apo B and B/A1 ratio were significantly higher in early-onset cohort. There was no difference in apo A1, HDL and total cholesterol/HDL ratio. Apo B level remained significantly higher among early-onset subjects after adjustment for insulin treatment. Lower current age and age of diagnosis were significant predictors of higher apo B level. Conclusions: The burden of atherogenic apolipoprotein particles was greater in early-onset T2D despite adequate statin treatment indicating an adverse phenotype for vascular disease.


Introduction
Early-onset type 2 diabetes (T2D) is a condition with an adverse cardiovascular risk profile.Despite their young age, these subjects are severely insulin resistant, 1 characterized by a high prevalence of obesity and dyslipidaemia 2 with elevated levels of inflammatory and prothrombotic markers for cardiovascular disease 3,4 and a higher burden of atherogenic particles carrying apolipoprotein B (apoB). 4,5ompared to controls without diabetes, subjects with early-onset T2D had a significantly higher risk of myocardial infarction than a later-onset cohort. 6o improve cardiovascular health and reduce other diabetes-related complications in young people with diabetes, the National Institute for Health and Clinical Excellence (NICE) guideline recommends intensive risk factor management which includes lowering LDL cholesterol with statins to achieve a treatment target of <2 mmol/l. 7LDL cholesterol concentration is an independent risk factor for the development of atherosclerosis and has been !The Author 2012.Published by Oxford University Press on behalf of the Association of Physicians.All rights reserved.For Permissions, please email: journals.permissions@oup.comwidely accepted as the primary target for lipidlowering therapy.Although the benefit of statin treatment in the prevention of cardiovascular disease has been demonstrated in T2D, 8 nevertheless, the residual risk for future cardiovascular events still remains high even with adequate LDL cholesterol lowering. 9,10polipoprotein markers, such as apo B, apo A1 and the apo B/A1 ratio, have been shown to be better predictors of future cardiovascular events than LDL cholesterol during statin therapy. 11,12hese parameters are better indicators of the total mass of lipoprotein particles directly involved in the atherosclerotic process, therefore, capture the risk for cardiovascular disease more accurately than the measurement of the cholesterol content of the LDL lipoproteins as these small, dense particles are enriched in apo B relative to its cholesterol content in T2D.Also, LDL cholesterol is calculated using the Friedewald equation 13 rather than direct measurement for practical convenience and the accuracy of this equation in diabetic dyslipidaemia with inherent lipoprotein abnormalities has been questioned. 14o date, studies have investigated the comparative burden of atherogenic apo B particles between statin-naı ¨ve early-onset T2D cohort with type 1 diabetes 5 and controls without diabetes 4 of similar age.No studies have examined the burden of apolipoprotein markers as an indicator of residual risk during statin treatment in early-onset T2D compared to those with later-onset disease, in whom statin treatment is the norm in current clinical practice.This is an important concept to establish as it has clinical relevance pertaining to the intensiveness of statin treatment in subjects with early-onset disease.We have previously shown that early-onset T2D has high risk of developing cardiovascular disease in their lifetime with dyslipidaemia being one of the principal risk factors. 15In this study, we hypothesized that the burden of the atherogenic lipoprotein particles would be higher in early-onset T2D despite adequate statin treatment.

Study population
The study population comprised of early and later onset subjects with T2D who attended the diabetes clinics at Sheffield Teaching Hospitals for their clinical care.The relevant clinical data were routinely entered into this database each time these subjects were reviewed in these clinics.Following identification from the hospital diabetes register, suitable subjects were recruited into this study.

Inclusion criteria
Type 2 diabetes is defined as those whose diabetes is controlled with oral hypoglycaemic agents without the need for insulin therapy within the first year of diagnosis and/or with negative antibody screen (glutamic acid decarboxylase, islet cell antibody and islet antigen-2).Early onset T2D was defined as current age <40 years while later-onset T2D was defined as current age and age diagnosis >40 years.Only those without history of cardiovascular disease and had received a stable dose of simvastatin for at least 6 months duration aiming to achieve adequate lipid-lowering defined by on-treatment LDL cholesterol concentration of <2 mmol/l as recommended by NICE, 7 were included.Those with type 1 diabetes, gestational diabetes, secondary diabetes, monogenic diabetes and latent autoimmune diabetes in adulthood were excluded.
To minimize selection bias and to achieve comparable cardiovascular risk characteristics, a nested case control approach was adopted.The later-onset subjects were selected to match the early-onset cohort for the on-treatment LDL cholesterol level as well as the clinical and study characteristics such as sex, obesity, glycaemic control, presence of microalbuminuria and treatment with one statin type (simvastatin).To further explore the comparability of the study groups, the severity of diabetes complications between the two cohorts was assessed using the validated Diabetes Complications Severity Index (DCSI) model as described by Young et al. 16 In brief, this method assigns a severity index score based on the number and severity of vascular and metabolic complications present at any one point in time.
Consent was obtained from each subject before being recruited into the study.This study obtained research ethics approval from the Leicestershire, Northamptonshire and Rutland Research Ethics Committee, UK.

Data collection
For each subject, a blood sample was obtained for the measurement of total cholesterol, high density lipoprotein (HDL) cholesterol, triglyceride, apolipoprotein A1 (apo A1) and apolipoprotein B levels after an overnight fast.Apo A1 and B were assayed by immunonephelometry on a Siemens BN2 nephelometer.The LDL cholesterol was calculated using the Friedewald's formula 13 while non-HDL cholesterol was calculated by subtracting HDL cholesterol from total cholesterol and is a measure of cholesterol content in the atherogenic lipoprotein particles.Other relevant clinical data were obtained from the electronic diabetes register.
Hypertension was defined as blood pressure >140/ 80 mmHg or >130/80 mmHg with diabetic complications or on specific treatment for this condition while dyslipidaemia was defined as fasting triglyceride >1.7 mmol/l and/or HDL <1.03 mmol/l for male or <1.29 mmol/l for female or on specific treatment for this condition.Overweight and/or obesity were defined as body mass index (BMI) >25 kg/m 2 .

Statistical analysis
All statistical analyses were performed using the Statistical Package for Social Science (SPSS) for Windows (version 16.0; SPSS Inc, Chicago, IL).Continuous data were expressed as mean and standard deviation (SD) while categorical data were expressed as percentages.The normality of the continuous data distribution was evaluated using Shapiro-Wilk test (P > 0.05 as normal) to determine the appropriate method for parametric or nonparametric statistical analysis.Student t-test and Mann-Whitney U test were used to compare the mean between two groups while Chi-squared ( 2) test was used to compare categorical variables between two groups.Correlation between apo B and apo A1 with the other lipid parameters was assessed using Pearson's correlation coefficient.Simple linear regression analysis was used to determine the strength and significance of the association between the various lipid parameters with current age, age of diagnosis, BMI, diabetes duration and glycaemic control (HbA1c).P-value of <0.05 was considered significant.

Results
The clinical characteristics of the study subjects are shown in Table 1.A total of 24 and 26 subjects with early-and later-onset T2D were recruited respectively.
The early-onset cohort was approximately 25 years younger, had significantly shorter diabetes duration with similar degree of obesity, dyslipidaemia, glycaemic control and microalbuminuria.Despite the young age, a substantial proportion of the early-onset cohort had hypertension.Both cohorts were comparable in the magnitude and severity of diabetes complications as demonstrated by the DCSI score.The younger subjects received a higher dose of simvastatin although the difference was not statistically significant.Similar proportion of subjects was on insulin treatment in both cohorts.
The lipid profiles during statin treatment are shown in Table 2.The apo B and non-HDL cholesterol concentrations were significantly higher in the early-onset cohort indicating a greater burden of atherogenic lipoprotein particles in these young subjects.The apo A1 level was similar in the two cohorts but due to the differential apo B concentration, the apo B/A1 ratio was significantly higher in the early-onset cohort.The total cholesterol and triglyceride levels were significantly higher in early-onset subjects while HDL cholesterol and the total cholesterol/HDL ratio did not differ significantly.The on-treatment LDL cholesterol level was similar in both cohorts at approximately 2 mmol/l.
To adjust for the effect of insulin treatment which is known to lower the apo B level, 17 the lipid profile was analysed among subjects with and without Early-onset type 2 diabetes insulin in early and later-onset cohort as shown in Table 3.The apo B level was significantly higher in insulin-naı ¨ve early-onset T2D subjects.Among those on insulin treatment, the early-onset cohort had greater burden of apo B particles but this was not statistically significant.In both groups, the on-treatment LDL cholesterol achieved the treatment target and the early-onset subjects received non-significantly higher simvastatin dose (insulin-naı ¨ve: early vs. later-onset; 38.7 (16.0) vs. 33.3(9.7) mg, p = NS and insulin treatment: early vs. later-onset: 35.6 (8.8) vs. 30.0(11.8) mg, p = NS).The triglyceride levels were higher in the early-onset cohort with and without insulin treatment, however, the significance was borderline among those not on insulin.Overall, there is a consistent trend of higher apo B and triglyceride levels among early-onset subjects irrespective of insulin treatment.negating the effect of insulin treatment, apo B remained significantly higher in subjects with early-onset T2D.
As expected, significant positive correlations were observed between apo B with total cholesterol (r = 0.86, P < 0.001), LDL cholesterol (r = 0.76, P < 0.001), non-HDL cholesterol (r = 0.90, P < 0.001) and triglyceride (r = 0.43, P = 0.002), while apo A1 correlated significantly with HDL cholesterol (r = 0.76, P < 0.001).The relationship between the conventional lipid parameters and apolipoproteins with current age, age of diagnosis, BMI, diabetes duration and glycaemic control is shown in Table 4. Lower current age and age of diagnosis were consistently the significant predictors of higher apo B level and apo B/ A1 ratio.There was no such observation for apo A1 and HDL cholesterol.There was a consistent trend of an relationship between higher non-HDL cholesterol and triglycerides levels with lower current age and age of diagnosis.There was no significant association between BMI, diabetes duration and HbA1c with all the lipid parameters.Figure 1 illustrates the significant inverse association between higher apo B level with lower current age.

Discussion
The process of atherosclerosis is intimately linked with apo B-containing lipoprotein particles.Being a measure of the total atherogenic particle number, apo B concentration is a better predictor of future cardiovascular events than LDL cholesterol in statin-naı ¨ve and statin-treated populations.In further support for its prognostic role, apo B is also associated with coronary artery calcification, 18 a marker of subclinical atherosclerosis and predictor  Early-onset type 2 diabetes of cardiovascular disease in diabetes. 19In earlyonset T2D, however, relatively few studies have investigated the contributory role of these atherogenic particles to the cardiovascular risk, the study populations being statin-naı ¨ve subjects.
Our study findings add to the existing literature by showing that statin-treated subjects with early-onset T2D have higher residual cardiovascular risk compared to the later-onset cohort despite receiving a non-significantly higher simvastatin dose.The mass of atherogenic apo B particles were significantly greater in these young subjects.Younger age and lower age of diagnosis were consistently significant predictors of a higher burden of these atherogenic particles indicating that early-onset T2D could be an inherently more aggressive phenotype to develop cardiovascular complications, despite lipidlowering therapy.
There are potential explanations for this observation.First, sub-optimal compliance with statin treatment could result in higher apo B level in the young cohort.However, this is unlikely given the comparable concentration of on-treatment LDL cholesterol to the recommended treatment target in early and later-onset cohorts.Secondly, the burden of apo B particles increases with hyperglycaemia. 5Since both cohorts had similar glucose control, the glycaemic factor is an unlikely contributory cause.Thirdly, the severity of insulin resistance could be greater in the early-onset cohort.Insulin resistance can induce hepatic over-production of triglyceriderich VLDL particles, resulting in the over-production of apo B-containing small, dense LDL particles. 20ertainly, this is a plausible explanation for the higher concentrations of fasting hypertriglyceridaemia and apo B consistently observed in the early-onset cohort.In support of this hypothesis, young subjects with T2D have been shown to be severely insulin resistant 1 and this adverse metabolic condition was less likely to respond to aerobic exercise 21 compared to the later-onset cohort 22 despite comparable BMI (as observed in our study cohort).Since the severity of insulin resistance can vary widely within similar BMI range in T2D, 23 it can be postulated that early-onset T2D is a more insulin resistant phenotype than the later-onset cohort despite having similar BMI.
Our observations have potential ramifications for clinical practice.First, it supports the hypothesis that early-onset T2D is a high risk cardiometabolic phenotype for vascular disease and merits aggressive risk factor intervention.We have previously shown that these young subjects experienced significant vascular complications in their lifetime. 15This has been supported by the findings of a recent study which utilized mathematical modelling to estimate lifetime cumulative incidence of diabetes-related complications in this young cohort. 24Various studies have demonstrated the presence of traditional (obesity, dyslipidaemia and hypertension) 2 and non-traditional (raised inflammatory markers, higher apo B and increased carotid media thickness) 3,4,25 cardiovascular risk markers in these young subjects and given that the likelihood of vascular events increases with diabetes duration, 26 early intensive management at an age when the complication burden is low and the opportunity to prevent the onset of complications is greatest, seems like a logical strategy.This adverse risk is present within 5 years of diagnosis 27 (similar to the diabetes duration of the early-onset cohort) and intensive management of the atherogenic risk factors can reduce early mortality 28 and cardiovascular complication later in the course of disease. 29Intensive statin treatment has been shown to confer cardiovascular benefit in the long term. 30econdly, LDL cholesterol does not give an accurate reflection of the residual cardiovascular risk during statin treatment in young subjects with T2D.Despite adequate lipid-lowering as determined by the LDL cholesterol concentration, the burden of the atherogenic apo B-containing lipoprotein particles was greater in the early-onset cohort.This is attributable to the increased heterogeneity in the cholesterol content of the LDL lipoprotein particles with a shift towards the small, dense cholesteroldepleted variety as a consequence of insulin resistance, 31 particularly when fasting triglyceride is >1.7 mmol/l 32 as observed in our early-onset cohort ($2.4 vs. 1.4 mmol/l in later-onset cohort).Since each atherogenic lipoprotein particle contains one apo B molecule, quantification of the apo B mass is a better index to gauge the reduction in atherogenic particle burden thus reducing the risk of vascular disease in early-onset T2D.
Thirdly, a lower LDL cholesterol treatment target may be required for early-onset T2D.Despite achieving similar on-treatment LDL cholesterol concentration, the later-onset cohort had mean apo B level <0.8 g/l, the treatment target recommended by the American Diabetes Association 33 while this was not the case for the early-onset subjects.Our observation suggests that more aggressive statin treatment may be needed to achieve parity in apo B level for the later-onset cohort.
Fourthly, apo B has the advantage of a direct measurement of the total burden of particles intimately involved in atherosclerosis using a validated standardized assay, 34 while LDL cholesterol is estimated from other measurements.Hence, it is only an approximation of the cholesterol content in this lipoprotein.
The strengths of this study include similarities in the study subject clinical characteristics (obesity, dyslipidaemia, microalbuminuria, glycaemic control) and diabetes complication status, statin dose, on-statin treatment LDL level and using only one type of statin treatment to avoid the varying impact of different statins on apo B concentration. 35owever, there are some limitations to this study.Firstly, the cohort of early-onset T2D is small.We found that very few young subjects were given lipid-lowering therapy.This experience is not unique to our centre 2 and this is attributable to both patient and healthcare professional factors. 36lthough the small sample of the early-onset cohort precludes a definitive conclusion, the novel findings from this study has highlighted an important clinical sphere that requires further attention and raises more hypothesis-generating questions.Secondly, there is a risk of selection bias given the great disparity in the prevalence of early and later-onset T2D, particularly in the recruitment of the later-onset subjects, which may lead to disparate level of cardiovascular risk between the study cohorts and confound the interpretation of the atherogenic apolipoprotein results.We, however, minimized this bias by adopting the nested case control approach in subject recruitment and successfully matched the two study cohorts to comparable cardiovascular risk characteristics and diabetes complication burden.Thirdly, our findings may not be extrapolated to the majority of the T2D population whose care is delivered in the community as those subjects with more adverse cardiovascular risk may be preferentially referred to hospital clinics.
Future research directions include a larger community-based population study to confirm the observation of this study and to determine the impact of statin treatment on vascular outcome in early-onset cohort, ascertainment of the predictive value of the apolipoproteins and its ratio in cardiovascular risk assessment in early-onset T2D, defining the thresholds of traditional lipid profile and apolipoproteins for adequate lipid-lowering that will confer clinical benefit and investigation into the pathophysiological mechanism of insulin resistance and its severity in young subjects with T2D.
To conclude, the prevention of vascular complications is an important goal in diabetes management.Lipid-lowering is an essential component in this objective.Early-onset T2D appears to be an inherently aggressive phenotype which may merit lipid-lowering treatment with an intensity greater than currently recommended.More research is needed to further explore this hypothesis.

Figure 1 .
Figure 1.Relationship between current age and apo B level with 95% confidence interval, P < 0.05 based on linear regression analysis.

Table 1
Clinical characteristics of study subjects

Table 3
Lipid profile with and without insulin treatment

Table 2
Lipid profile during statin treatment

Table 4
Linear regression analysis