Preclinical carotid atherosclerosis in patients with latent autoimmune diabetes in adults (LADA), type 2 diabetes and classical type 1 diabetes

Macrovascular disease is the leading cause of morbidity and mortality in patients with both type 1 diabetes and type 2 diabetes [1]. Latent autoimmune diabetes of the adults (LADA), that is patients with adult-onset autoimmune diabetes who do not initially require insulin, represent 4–14% of subjects previously diagnosed with type 2 diabetes [2]. LADA can be estimated to have a higher prevalence than classic type 1 diabetes in both children and adults. Notwithstanding its clinical relevance, data about cardiovascular events and mortality in patients with LADA are limited. It could be anticipated that patients with LADA would be less likely to develop macrovascular disease than patients with type 2 diabetes, as they have a reduced risk of metabolic syndrome [3‐5]. However, most studies to date, despite their limited size, population heterogeneity and variable disease duration, have failed to show such a reduced risk, and indeed many find an increased risk [6‐9].

Preclinical atherosclerosis can be detected and quantified non-invasively by carotid ultrasound, which is a strong predictor of cardiovascular events [10‐16]. Although there is still controversy in this field, established carotid atherosclerosis (carotid plaques), detected by ultrasound imaging, improves risk stratification when added to traditional risk factors. Indeed, individuals with carotid plaques are considered a very-high risk category according to the most recent guidelines, and carotid artery plaque assessment using ultrasonography may be considered to be a valid risk modifier in cardiovascular prediction [17‐19]. Patients with diabetes mellitus are at high risk of cardiovascular events; therefore, the added potential of carotid plaque detection in risk assessment would be different from those cases without diabetes [20‐22]. However, primary prevention targets are often not reached in patients with diabetes, and carotid plaque detection could focus more intensive preventive strategies in these cases, especially in those patients without known organ damage or important associated cardiovascular risk factors [18, 19, 23].

The current study was initially designed to assess the difference in preclinical carotid plaque burden in patients with LADA, as compared to type 2 diabetes. After the results of the initial study, we extended the comparison to age and gender matched patients with classic onset type 1 diabetes.

This was a cross-sectional study to evaluate carotid atherosclerotic disease in patients with LADA in comparison with patients with type 2 diabetes and classic onset type 1 diabetes.

We recruited 71 patients with LADA, 191 patients with type 2 diabetes and 116 patients with type 1 diabetes. The clinical characteristics of the study groups are shown in Table 1. Patients with LADA, classic type 1, and type 2 diabetes were similar in age, gender distribution, current smoking status, use of antihypertensive drugs, and HbA1c concentrations. As patients were matched by age, the diabetes duration was longer in patients with type 1 diabetes than in patients with LADA, but also, in the latter duration was longer than in type 2 diabetes (23.7 ± 12.4, 13.2 ± 9.7 and 8.7 ± 7.9 years, respectively, P < 0.001). While the prevalence of diabetic retinopathy, as expected given the longer disease duration, was higher in type 1 diabetes than in LADA and type 2 diabetes, the frequency of microalbuminuria was similar. About half the patients with LADA were treated with metformin, and the majority were receiving insulin. Patients with type 2 diabetes, as expected, exhibited a more adverse lipid profile and anthropometric measures than the other two cohorts, though fewer were on statin treatment.

For carotid ultrasound findings: mean carotid IMT was lower in patients with type 1 diabetes compared to both LADA and type 2 diabetes (Table 1). As shown in Fig. 1 subclinical carotid atherosclerosis (presence of atherosclerotic plaques) was more frequent in patients with LADA than in patients with type 1 diabetes and type 2 diabetes [73.2% vs. 57.1% (P = 0.026) and 56.9% (P = 0.018), respectively]. Multiple plaques were also more frequent in patients with LADA than in type 1 diabetes and type 2 diabetes [45.1% vs. 33.6% (P = 0.077), and 27.2% (P = 0.019), respectively] (Fig. 1).

Carotid plaques were also related to older age, longer diabetes duration, hypertension, retinopathy and dyslipidemia (Table 2).

A multivariate logistic regression model including age, diabetes duration, gender, type of diabetes, retinopathy, hypertension, smoking and dyslipidemia (Table 3) revealed that age, but not gender, type of diabetes, and the presence of hypertension and smoking habit were associated with carotid atherosclerosis. The adjusted inverse OR for the presence of carotid plaques in patients with LADA as compared with those with type 1 and type 2 diabetes was 2.44 [(1.65–3.6); P = 0.0225] and 2.06 [(1.47–2.9); P = 0.0334], respectively.

To explore the impact of disease duration on carotid plaque frequency in patients with LADA, we limited the analysis to patients with LADA and type 2 diabetes, as few patients with type 1 diabetes had a short diabetes duration. We divided patients with LADA and type 2 diabetes into age-adjusted tertiles of disease duration: ≤5, 6–11, ≥12 years. Table 4 shows the results of this analysis; within 5 years of diabetes duration the prevalence of carotid plaques was lower, though not significantly so, in patients with LADA than in those with type 2 diabetes [26.7% vs. 54.8%, inverse OR 0.38 (0.11–1.39); P = 0.137]. However, with increasing diabetes duration the frequency of carotid atherosclerotic plaques increased in patients with LADA, but not in patients with type 2 diabetes [81.8% vs. 50%, inverse OR 5.72 (1.5–21.8); P = 0.009].

In this study, we unexpectedly found an increased frequency of subclinical carotid atherosclerosis in patients with LADA, as compared to subjects with classic type 1 diabetes and type 2 diabetes. That excess frequency of carotid disease was also evident when assessing multiple plaques in LADA compared with the other cohorts. We did not find the same results for cIMT as for carotid plaque, potentially due to the high variability and low intra-individual reproducibility of the former measurement [33]. Moreover, the pathophysiology underlying the development of carotid plaque and cIMT is distinct and the evidence regarding the value of cIMT in cardiovascular risk prediction is contradictory and of debatable value [34‐36]. Most recent clinical guidelines on cardiovascular risk do not recommend carotid IMT for individual risk prediction [18, 37]. In contrast, the presence of carotid plaques has been shown to be a good predictor of future cardiac events [15, 18].

The high frequency of carotid plaques, despite the increased use of statins in LADA cases, remained after adjusting for the main cardiovascular risk factors, including diabetes duration. The current cross-sectional study does not allow for the identification of potential causes of our unexpected finding of higher atherosclerotic burden in LADA, even though patients with type 2 diabetes tend to exhibit, as in other studies, a worse cardiovascular risk profile including higher blood pressure, more obesity and an adverse lipid profile. In both the Freemantle Study [38] and the Collaborative Atorvastatin Study [7], patients designated LADA had comparable cardiovascular disease as the patients with type 2 diabetes; while in the Botnia study [6] after 13 years of diabetes 56% of LADA patients had ischemic heart disease and 5% had had cerebrovascular events. To date, only one other study has assessed the frequency of carotid atherosclerosis in patients with LADA and type 2 diabetes [8], and none have assessed that frequency in the three cohorts we assessed. That Chinese study was retrospective and not designed to answer the question we posed here, but they did find a comparable frequency of carotid atherosclerosis compared with type 2 diabetes patients. Taken together, these studies indicate that patients with adult-onset diabetes, irrespective of the cause of the diabetes, are at risk of carotid and coronary atherosclerosis. It is difficult to understand why this should be, given that autoimmune diabetes cases are less likely to have the metabolic syndrome and related features [7]. Two possibilities suggest themselves. First, the time of onset of autoimmune diabetes antedates the disease by months, even years, and it remains possible that LADA patients have had low-grade disease many years before clinical presentation, as can be the case with type 2 diabetes. However, the duration of LADA pre-disease is unknown both in this study and in general, while the risk of atherosclerosis in these cases was increased with increasing diabetes duration post-diagnosis. Second, patients with LADA are not well managed and are consistently shown to have higher HbA1c levels than comparable cohorts of patients with type 2 diabetes [4, 39‐41]. Higher HbA1c has been found to be associated with higher coronary and carotid atherosclerotic burden in non-diabetic patients [42]. Additionally, deranged glycemic control has been independently related to coronary heart disease in both LADA patients [6], and diabetes patients in general [40, 43]. Higher HbA1c concentrations have also been related to cardiovascular disease in the non-diabetic population in prospective studies, including the Norfolk cohort of European Prospective Investigation into Cancer and Nutrition study [44] and The Copenhagen City Heart Study [45].

In subjects with diabetes, other potential different pathophysiological mechanisms may be involved in the atherosclerotic process. First, advanced glycation end products (AGEs) which are increased in diabetes are pro-oxidants that induce oxidative stress and are thought to play a major role in atherosclerosis [46, 47]. Although diet is the principal source of AGEs, their concentrations rise with increasing blood glucose. An increased level of advanced AGEs has been associated with cardiovascular disease in type 1 diabetes [48] and type 2 diabetes [49]; no such data is available in LADA. Secondly, patients with systemic autoimmune diseases, such as rheumatoid arthritis or systemic lupus erythematosus, have an increased risk of atherosclerosis, with both cellular and humoral components of the immune system involved in its pathogenesis and some shared genetic risk [50], although the precise mechanisms underlying the high cardiovascular risk in these patients is unclear [51]. It is well known that chronic inflammation plays an important role in atherogenesis [52, 53], and inflammatory markers associated with macrovascular disease are elevated both in patients with type 1 diabetes and type 2 diabetes [54, 55]. Moreover, poor glycemic control and hypoglycaemic episodes are related to pro-inflammatory status [56, 57], and may be additional contributing factors in atherosclerosis in patients with autoimmune diabetes [58]. Unfortunately, in the current study we did not determine any markers of chronic inflammation in the LADA patients, though we have previously demonstrated such changes in LADA patients [59].

Nevertheless, this study has some weaknesses. First of all, the size of the populations studied is relatively small, and some important information is missing, notably the familial history of premature cardiovascular disease. While we found no differences in HbA1c between the three diabetes groups, this is a cross-sectional study, and the relationship between glycemic control and atherosclerosis can only be answered with a prospective design. The situation is made more complex because most of the patients with LADA in this study had already been referred to specialized care, usually because of poor glycemic control. The specialized diabetes care of all the patients with type 1 diabetes and most of the patients with LADA may explain the increased use of statins and antiplatelet agents in these groups compared to that of patients with type 2 diabetes, since the latter are usually under general practitioner’s care. Since LADA patients were recruited from a specialty-based cohort our results may show a bias; however, this group of LADA subjects is likely representative of LADA patients in general in Spain.

Nevertheless, despite its limitations, our study has the strength of comparing for the first time these three groups of diabetes, which come from a homogeneous population. While we use the term LADA throughout the text for clarity, we would prefer that LADA is called adult-onset autoimmune diabetes since there is no firm evidence that LADA can be categorically distinguished from type 1 (autoimmune) diabetes.

Our results indicate that the frequency of preclinical carotid atherosclerosis in patients with adult-onset initially non-insulin requiring autoimmune diabetes, also called LADA, is comparable, even greater, than in adults of similar age with classic type 1 diabetes and type 2 diabetes. Our data should be confirmed in both cross-sectional and prospective studies in other populations. However, the results draw attention to the importance of macrovascular disease in patients with diabetes, irrespective of the cause of the disease, and emphasize the potential value of rigorous treatment of macrovascular risk factors even in patients with autoimmune diabetes.