Increased cardiovascular risk in people with LADA in comparison to type 1 diabetes and type 2 diabetes: Findings from the DPV registry in Germany and Austria

Abstract

Introduction

We aimed to characterise and compare individuals diagnosed with type 1 diabetes (T1D), latent autoimmune diabetes in adults (LADA) and type 2 diabetes (T2D), in a real‐world setting.

Methods

Anthropometric and clinical data from 36 959 people with diabetes diagnosed at age 30–70 years enrolled in the prospective diabetes patients follow‐up (DPV) registry from 1995 to 2022 were analysed cross‐sectionally at diagnosis and follow‐up (≥6 months after diagnosis). LADA was defined as clinical diagnosis of T2D, positivity of ≥1 islet autoantibody and an insulin‐free interval of ≥6 months upon diabetes diagnosis.

Results

At diagnosis, age, body mass index, waist circumference, C‐peptide and HbA1c in people with LADA (n = 747) fell in between individuals with T1D (n = 940) and T2D (n = 35 272) (all p‐values < 0.01). At follow‐up, after adjusting for age, sex and diabetes duration, the prevalence of dyslipidemia and hypertension was the highest in people with LADA (90.6%, 77.7%) compared to people with T2D (81.8%, 60.4%) and T1D (75.7%, 39.7%) (p < 0.01). The prevalence of diabetic kidney disease (DKD) was higher in LADA (44.2%), than in T1D (19.9%) (p < 0.01). The prevalence of peripheral neuropathy was higher in individuals with LADA (55.1%) than in T2D (43.9%) and T1D (42.1%) (p < 0.05). Coverage of treatment for hypertension and dyslipidemia were 22.4% and 15.0% in T1D, 63.0% and 36.6% in LADA and 29.4% and 18.2% in T2D.

Conclusion

People with LADA had a higher prevalence of cardiovascular risk factors (dyslipidemia, hypertension) and cardiovascular complications (DKD and peripheral neuropathy), suggesting that people with LADA are at need for improved recognition and care.

1. INTRODUCTION

Diabetes is a heterogenous disease and commonly classified into type 1 (T1D) and type 2 diabetes (T2D). ¹ An intermediate form known as latent autoimmune diabetes in adults (LADA) affects about 10% of people with diabetes. ² T1D is caused by an immune‐mediated β‐cell destruction ultimately leading to severe insulin deficiency, whereas T2D is heterogenous and results from insulin resistance and β‐cell failure. ³ , ⁴ , ⁵

LADA is defined as diabetes that manifests in individuals at adult age, exhibiting clinical characteristics resembling T2D, regarding age of onset and metabolic profile at diagnosis, whilst also showing the presence of autoantibodies associated with T1D. ² , ⁶ , ⁷ In current guidelines, LADA is considered a subgroup of T1D. ⁸ , ⁹ Due to the clinical and phenotypical similarity between individuals with LADA and individuals with T2D, and the lack of routine measurements of β‐cell antibodies in individuals with T2D, LADA remains often undetected. ¹

Diabetes is a risk factor for developing macrovascular diseases such as stroke and myocardial infarction, as well as microvascular diseases including retinopathy, neuropathy, and diabetic kidney disease (DKD). ¹⁰ Studies have shown higher values for glycosylated haemoglobin A1c (HbA1c) in people with LADA, compared to people with T2D, both at diagnosis and after a follow‐up period of 8.9 years. ¹¹ , ¹² This was similar for paediatric individuals with latent autoimmune diabetes of the young (LADY), who showed intermediate values of HbA1c, compared to those with T1D and T2D. ¹³ The loss of β‐cell function in LADA is slower than in T1D, ¹⁴ whilst C‐peptide concentration is higher. ¹⁵ This supports the hypothesis that people with LADA have intermediate values for glycemic control and β‐cell function when compared to T1D and T2D. ⁷

Recent work from Swedish registers by Wei et al. compared T1D, LADA and T2D in a cross‐sectional study. It showed that the risk of death, cardiovascular disease (CVD) (including ischemic heart disease, stroke and heart failure) and retinopathy in LADA is similar or even higher compared to T2D. Wei et al. also found that people with LADA had lower blood pressure and lipid profiles compared to those with T2D, but higher levels than those with T1D. ¹¹ Findings from the Action LADA study show that adhesion molecules as risk markers for CVD were highest in T2D, followed by LADA and T1D, but no clinical outcomes were reported here. ¹⁶

Our hypothesis was that people with LADA differ from those with T1D and T2D with respect to cardiovascular risk. This would provide evidence that identification of patients with LADA is prognostically relevant. The goal of our study was to compare T1D, LADA and T2D for clinical and anthropometric measures with a focus on cardiovascular risk factors and diabetes related co‐morbidities. We analysed data from the DPV (diabetes patient follow‐up) registry from people with diabetes at diagnosis and at follow‐up.

2. METHODS

2.1. Data source

The definitions of LADA in our study were according to Hawa et al.: diagnosis of diabetes at 30–70 years, without insulin therapy for the first 6 months upon diagnosis and, the last one, islet autoantibody titre positive. ² Other guidelines (German Diabetes Society [DDG]) define LADA as people over the age of 35 at diagnosis, ¹⁷ but for this analysis, we preferred the international age cutoff over the German one.

People in the T1D group presented with typical symptoms of diabetes, positive β‐cell antibodies and/or insulin antibodies (IAA, only if evaluated within 4 weeks after diagnosis) and treatment with insulin upon diagnosis and during follow‐up. People with T2D had typical symptoms of diabetes, required no treatment with insulin at diagnosis for at least 6 months upon diabetes diagnosis and were β‐cell antibody negative.

For this analysis, data from the DPV registry were used. ¹⁸ Data from 197 medical centres (195 in Germany, two in Austria) were available. Data was analysed at diagnosis of diabetes (±11 days around manifestation) and during follow‐up (most recent documented treatment year for each individual, at least 6 months since diagnosis). Individuals with other forms than T1D, LADA or T2D such as monogenic forms, genetic syndromes or secondary diabetes were excluded. Diabetes diagnosis had to be documented at least to the month.

Age at diagnosis was restricted to 30–70 years in all diabetes groups, to comply with the definition of LADA. ² The analysis included people diagnosed and treated between the years 1995 and 2022. Diabetes manifestation data of 57 702 people were recorded, and people with T1D with incomplete insulin documentation were excluded resulting in 42 825 people. A β‐cell antibody status was not available in some people with suspected LADA or T1D; this further reduced the number of subjects in the analysis from 42 825 to 36 959 individuals at manifestation. At follow‐up, data of 12 792 people were available for analysis. The data of 36 959 individuals was used for this analysis, with measured β‐cell antibody status and documented insulin treatment at follow‐up for all people with T1D and LADA. Out of 35 272 people with T2D, antibody measurements were available in 1307 people (Figure 1). Participating centres transfer anonymised data to the University of Ulm in Germany every 6 months, where data are validated, aggregated and analysed. These methods of data collection are standardised across the DPV registry, using the identical documentation software for more than 20 years. All participating subjects are informed about the registry and approved data collection and anonymous analysis. The ethics committee of the University of Ulm in Germany approved the initiative, approval number 314/21. Local review boards from participating centres granted their consent for the initiative, ensuring compliance with data protection.

FIGURE 1.

Patient acquisition in diabetes patient follow‐u (DPV) during 1995–2022 in patients diagnosed at age 30–70 years with β‐cell antibody positive type 1 diabetes (T1D) and β‐cell antibody positive latent autoimmune diabetes in adults (LADA) and type 2 diabetes (T2D).

2.2. Data analysis

In this cross‐sectional retrospective observational study of the DPV registry, people with T1D, LADA or T2D were assessed at diabetes manifestation and at follow‐up (≥6 months post diagnosis).

Anthropometric, clinical and laboratory data at manifestation (±11 days around diagnosis) included the following: age at diabetes onset, sex, body mass index (BMI) (categorised in intervals), waist circumference, HbA1c, triglycerides, LDL‐cholesterol, liver transaminases (ALT, AST, GGT) and C‐peptide concentration.

For all individuals in the follow‐up cohort, baseline data were available. At follow‐up, duration of diabetes, sex, BMI (categorised in intervals), waist circumference, HbA1c, triglycerides, LDL‐cholesterol, liver transaminases and C‐peptide concentration were assessed. At follow‐up, we also examined cardiometabolic risk factors and diabetes‐associated complications: smoking, dyslipidemia, treatment with lipid‐lowering drugs, arterial hypertension, treatment with antihypertensive drugs, retinopathy, DKD, microalbuminuria, macroalbuminuria, neuropathy (autonomous and peripheral), myocardial infarction, stroke, peripheral artery disease (PAD), cardiovascular revascularisation, severe hypoglycemia with or without coma and diabetic ketoacidosis (DKA). Further comorbidities were analysed: metabolic dysfunction‐associated steatotic liver disease (MASLD), metabolic dysfunction‐associated steatohepatitis (MASH), celiac disease, Hashimoto thyroiditis, Addison disease, atrophic gastritis and vitiligo. Dyslipidemia was defined as serum cholesterol ≥200 mg/dL and/or LDL‐cholesterol ≥100 mg/dL and/or HDL‐cholesterol <40 mg/dL and/or treatment with lipid lowering drugs.

Arterial hypertension was defined as a systolic and/or diastolic blood pressure >140/90 mmHg and/or treatment with antihypertensive drugs. HbA1c values (percent) obtained from various centres were standardised mathematically to fit within the reference range of the Diabetes Control and Complication Trial (DCCT) (4.05%–6.05%). This standardisation was achieved using the ‘multiple of the mean’ transformation method. ¹⁹ The definition of DKD was based on eGFR <60 mL/min/1.73 m² using the MDRD formula and/or kidney transplantation, and/or dialysis and/or microalbuminuria. ²⁰ Microalbuminuria was defined as ≥30 mg albumin in 24‐h urine collection, or an albumin‐creatinine‐ratio ≥25 mg/g creatinine (in males)/≥35 mg/g creatinine (in females).

DKA was defined as pH <7.3 and/or serum bicarbonate <15 mmol/L. Autoimmune thyroiditis was reported in individuals with established autoimmune thyroiditis or in people with a diagnosis of thyroiditis and positive thyroid antibodies or reduced thyroid gland echogenicity. Addison disease was reported in people with established Addison disease or who were diagnosed with antibody‐positive adrenal insufficiency and long‐term hydrocortisol substitution.

2.3. Statistical analysis

In our descriptive analysis, we present continuous outcomes as medians and interquartile range. Dichotomous outcomes are presented as percentages.

Unadjusted group comparisons were implemented using Wilcoxon rank sum test and chi‐squared test. Two‐sided p‐values were adjusted for multiple testing using the Bonferroni–Holm method. The rates of hypoglycemic events during follow‐up are reported as percentages.

Cardiovascular risk factors at follow‐up including triglycerides, liver transaminases, dyslipidemia, arterial hypertension, DKD, microalbuminuria and peripheral neuropathy were compared using linear or logistic regression models, adjusted for age, sex and duration of diabetes. Complete case analysis was used as recommended in the statistical literature for parameters with low or high (c‐peptide values) percentage of missing data.

p‐Values were adjusted for multiple group comparisons using the Tukey–Kramer method.

Two‐sided p‐values <0.05 were considered statistically significant. All analyses were conducted using Statistical analysis System (SAS) version 9.4, build TS1M7 on a Windows Server 2019 mainframe (SAS Institute, Cary, NC, USA).

3. RESULTS

3.1. Anthropometric and clinical characteristics

At diabetes manifestation, the data of 36 959 individuals were analysed. The majority of individuals were diagnosed with T2D (n = 35 272, 61.6% male). A total of 940 people had T1D (57.6% male) and 747 people had LADA (69.3% male) (Table 1).

TABLE 1.

Anthropometric and clinical characteristics of people with latent autoimmune diabetes in adults (LADA), type 1 diabetes (T1D) and type 2 diabetes (T2D) at the time of diagnosis and during follow‐up.

	n	T1D	p‐Value T1 versus LADA	n	LADA	p‐Value LADA versus T2D	n	T2D	p‐Value T1 versus T2D
At diagnosis
Age of onset (years)	940	43.6 (36.0–52.5)	<0.0001	747	50.1 (42.9–57.7)	<0.0001	35 272	56.6 (48.8–63.3)	<0.0001
% male	940	57.6	<0.0001	747	69.3	<0.0001	35 272	61.6	0.024
BMI (kg/m2)	895	23.4 (21.5–27.2)	<0.0001	712	29.8 (26.2–34.6)	<0.0001	29 842	31.2 (27.5–35.9)	<0.0001
% BMI <25 kg/m2	895	60.6	<0.0001	712	17.8	<0.0001	29 842	11.5	<0.0001
% BMI 25–<30 kg/m2	895	27.6	0.019	712	33.6	0.111	29 842	30.6	0.054
% BMI 30–<35 kg/m2	895	7.6	<0.0001	712	25.4	0.106	29 842	29.0	<0.0001
% BMI 35–<40 kg/m2	895	2.5	<0.0001	712	13.3	0.111	29 842	16.0	<0.0001
% BMI ≥ 40 kg/m2	895	1.79	<0.0001	712	9.83	0.067	29 842	12.9	<0.0001
Waist circumfer. (cm)	185	88 (81–95)	<0.0001	307	105 (96–118)	0.048	6067	108 (99–120)	<0.0001
HbA1c (in %)	919	11.3 (9.8–13.0)	<0.0001	719	9.75 (7.5–11.8)	<0.001	30 291	7.8 (6.4–10.4)	<0.0001
HbA1c (mmol/mol)	919	100 (84.0–118)	<0.0001	719	83.0 (58.0–106)	<0.0001	30 291	61.0 (46.0–90.1)	<0.0001
Triglycerides (mg/dL)	747	150 (103–236)	<0.0001	504	197 (137–318)	0.170	17 655	188 (127–290)	<0.0001
LDL‐C (mg/dL)	702	120 (94.0–150)	0.355	485	126 (92.0–151)	0.985	15 735	124 (94.0–154)	0.108
GGT (U/L)	293	25.0 (17.0–51.6)	<0.0001	166	49.5 (28.8–96.0)	0.985	6312	46.0 (29.0–85.0)	<0.0001
AST (U/L)	240	21.0 (16.0–28.0)	<0.0001	121	31.0 (21.0–48.0)	0.096	4317	27.0 (20.0–41.0)	<0.0001
ALT (U/L)	305	26.4 (19.0–41.0)	<0.0001	176	39.0 (25.5–59.7)	0.083	5288	35.0 (23.0–55.0)	<0.0001
C‐Peptide (μg/L)	610	0.74 (0.50–1.30)	<0.0001	366	2.50 (1.50–3.90)	0.026	2727	2.80 (1.80–4.20)	<0.0001
During follow‐up
Diabetes duration (years)	214	3.90 (1.78–7.45)	1.000	276	3.77 (1.74–6.37)	1.000	12 302	3.47 (1.46–6.84)	0.650
Age of onset (years)	214	45.0 (36.4–52.3)	<0.0001	276	51.0 (44.2–58.4)	<0.0001	12 302	56.7 (48.9–63.3)	<0.0001
% male	214	52.3	0.010	276	64.9	0.028	12 302	57.0	0.171
BMI (kg/m2)	193	25.1 (22.7–28.4)	<0.0001	254	29.7 (26.6–35.3)	0.887	9844	31.3 (27.4–35.8)	<0.0001
% BMI <25 kg/m2	193	49.7	<0.0001	254	13.4	1.000	9844	11.5	<0.0001
% BMI 25–<30 kg/m2	193	30.1	1.000	254	37.8	0.294	9844	30.3	1.000
% BMI 30–<35 kg/m2	193	14.0	0.339	254	22.4	0.482	9844	29.2	<0.0001
% BMI 35–<40 kg/m2	193	4.15	0.018	254	13.4	1.000	9844	16.8	<0.0001
% BMI ≥ 40 kg/m2	193	2.07	<0.0001	254	13.0	1.000	9844	12.2	<0.001
Waist circumfer. (cm)	43	101 (88–110)	0.003	162	109 (99–119)	0.384	4400	108 (99–119)	0.003
HbA1c (in %)	203	7.21 (6.48–8.38)	0.034	260	6.82 (6.11–7.86)	1.000	10 864	6.73 (6.01–7.68)	<0.0001
HbA1c (mmol/mol)	203	55.4 (47.3–68.1)	0.034	260	51.1 (43.3–62.4)	1.000	10 864	50.1 (42.2–60.4)	<0.0001
Triglycerides (mg/dL)	146	103 (73.0–152)	<0.0001	187	182 (127–249)	0.711	5729	165 (117–239)	<0.0001
LDL‐C (mg/dL)	140	111 (89.3–135)	0.678	187	116 (89.0–141)	1.000	5407	115 (88.0–141)	0.598
GGT (U/L)	69	19.8 (12.5–35.0)	<0.0001	45	41.0 (27.0–60.0)	1.000	3497	34.0 (22.0–59.0)	<0.0001
AST (U/L)	47	20.0 (16.0–29.0)	0.416	30	24.5 (19.0–36.0)	1.000	2396	25.0 (20.0–32.0)	0.011
ALT (U/L)	65	20.0 (14.0–30.0)	0.007	40	29.3 (23.0–41.3)	1.000	2684	27.0 (20.0–39.0)	<0.001
C‐Peptide (μg/L)	22	0.75 (0.20–1.20)	0.004	17	2.24 (1.50–4.70)	1.000	308	3.20 (2.10–5.00)	<0.0001

Note: Data are in percent or medians (first to third quartiles). n represents the count of individuals for whom data was provided. Significance is established by p‐values <0.05 through the application of both Wilcoxon rank sum test and the chi‐squared test, adjusted with the Bonferroni–Holm method. Significant p‐values are printed in bold.

Abbreviations: ALT, alanine transaminase; AST, aspartate transaminase; BMI, body mass index; GGT, gamma‐glutamyltransferase; HbA1c, glycosylated haemoglobin; LDL‐C, low‐density lipoprotein.

At follow‐up, 276 people with LADA (64.9% male), 214 people with T1D (52.3% male) and 12 302 people with T2D (57.0% male) were analysed. Median duration of diabetes (interquartile range IQR) was 3.9 (1.78–7.45) years for T1D, 3.77 (1.74–6.37) years for LADA and 3.47 (1.46–6.84) years for T2D (Table 1).

At manifestation, individuals with LADA have intermediate values for BMI (p‐value < 0.01), waist circumference (p‐value < 0.05), HbA1c (p‐value < 0.01) and C‐peptide (p‐value < 0.05), compared to people with T2D and T1D. At follow‐up, differences between LADA and T2D are no longer statistically significant. Differences to T1D are still significant, compared to LADA and T2D (p‐values < 0.05) (Table 1; Figure 2).

FIGURE 2.

Anthropometric and laboratory data at manifestation and during follow‐up. Data are medians (first to third quartiles). Cross‐sectional analysis. Not all patients from manifestation were seen at follow‐up. For n‐values, exact p‐values and numbers see Table 1. Significance is established by p‐values <0.05 through the application of both Wilcoxon rank sum test and the chi‐squared test, adjusted with the Bonferroni–Holm method. * Versus type 1 diabetes (T1D) p‐value < 0.05. ** Versus T1D p‐value < 0.01. ^‡Versus type 2 diabetes (T2D) p‐value < 0.05. ^‡‡Versus T2D p‐value < 0.01. BMI, body mass index; LADA, latent autoimmune diabetes in adults.

BMI categories were analysed at clinical manifestation and during follow‐up (Table 1). At clinical manifestation, the majority of individuals with normal weight (BMI < 25 kg/m²) had T1D, followed by LADA and T2D (p‐values < 0.01). Diabetes groups for overweight (BMI 25–30 kg/m²) were similar with approximately one third at clinical diagnosis (p‐value T1D vs. LADA < 0.05, p‐values LADA vs. T2D and T1D vs. T2D ns). Obesity as defined by BMI ≥30 kg/m² was more prevalent in LADA and T2D when compared to T1D (p‐values < 0.01) (Table 1). BMI >30 kg/m² was numerically higher in T2D than LADA. At follow‐up, results were similar (Table 1).

For triglycerides, LADA had the highest median values at manifestation (LADA: 197 mg/dL, T2D: 188 mg/dL, T1D: 150 mg/dL) and at follow‐up (LADA: 182 mg/dL, T2D: 165 mg/dL, T1D: 103 mg/dL), but differences to T2D were not statistically significant either at manifestation or at follow‐up. Both the individuals with LADA and those with T2D had statistically higher triglyceride values than the individuals with T1D (p‐values < 0.01) (Table 1). A regression analysis adjusted for age, sex and duration of diabetes confirmed these results (see Table S1).

3.2. Diabetes associated complications and risk factors (at follow‐up)

Frequency of arterial hypertension was most prevalent in people with LADA (77.79%), followed by people with T2D (60.4%) and T1D (39.7%) (p‐values < 0.01). Treatment of hypertension was reported less in 63.0% of people with LADA, 22.4% of people with T1D and 29.4% of people with T2D (p‐values T1D vs. LADA and LADA vs. T2D <0.01, p‐value T1D vs. T2D ns) (Table 2).

TABLE 2.

Cardiometabolic risk factors and weight related comorbidities at follow‐up.

	n	T1D	p‐Value T1D versus LADA	n	LADA	p‐Value LADA versus T2D	n	T2D	p‐Value T1D versus T2D
Diabetes duration (years)	214	3.90 (1.78–7.45)	1.000	276	3.77 (1.74–6.37)	1.000	12 302	3.47 (1.46–6.84)	0.650
% Dyslipidemia	152	75.7	0.002	213	90.6	0.028	6856	81.8	0.767
% Lipid‐lowering drugs	214	15.0	<0.0001	276	36.6	<0.0001	12 302	18.2	1.000
% Arterial hypertension	204	39.7	<0.0001	255	77.7	<0.0001	10 669	60.4	<0.0001
% Antihypertensive drugs	214	22.4	<0.0001	276	63.0	<0.0001	12 302	29.4	0.455
% Smoking	168	35.7	0.356	199	25.1	1.000	5959	20.8	<0.0001
Weight related comorbidities
% MASH	214	0.47	1.000	276	0.72	1.000	12 302	0.31	1.000
% MASLD	214	0.93	1.000	276	3.26	<0.001	12 302	12.4	<0.0001

Note: Data are in percent or medians (first to third quartiles). n represents the count of individuals for whom data was provided. Significance is established by p‐values <0.05 through the application of both Wilcoxon rank sum test and the chi‐squared test, adjusted with the Bonferroni–Holm method. Significant p‐values are printed in bold.

Abbreviations: LADA, latent autoimmune diabetes in adults; MASH, metabolic dysfunction‐associated steatohepatitis; MASLD, metabolic dysfunction‐associated steatotic liver disease; T1D, type 1 diabetes; T2D, type 2 diabetes.

Dyslipidemia was reported for the majority of people with LADA (90.6%), followed by people with T2D (81.8%) and T1D (75.7%) (p‐values T1D vs. LADA and LADA vs. T2D <0.05, p‐value T1D vs. T2D ns). Similar to the antihypertensive treatment, lipid lowering drugs were recorded less often than dyslipidemia and were prescribed to 36.6% in LADA, 15.0% in T1D and 18.2% in T2D, respectively (p‐values T1D vs. LADA and LADA vs. T2D <0.01, p‐value T1D vs. T2D ns). The differences between T2D and LADA are statistically significant, and people with LADA had the highest prevalence for these parameters (Table 2). These results were confirmed when adjusting for age, sex and diabetes duration (see Table S1).

Whilst smoking in T1D (35.7%) was significantly higher compared to T2D (20.8%) (p‐value < 0.01), LADA (25.1%) did not statistically differ. Contrary to dyslipidemia, metabolic dysfunction‐associated steatotic liver disease (MASLD) showed the highest prevalence in people with T2D (12.4%), followed by people with LADA (3.26%) and T1D (0.93%) (p‐values < 0.01) (Table 2).

3.3. Microvascular complications (at follow‐up)

DKD (as defined by eGFR < 60, microalbuminuria, dialysis or transplantation) was reported in 44.2% of people with LADA at follow‐up, followed by those with T2D (38.5%) and T1D (19.9%). Similarly, the prevalence of microalbuminuria was highest in people with LADA (28.6%) followed by those with T2D (24.7%) and lowest in people with T1D (8.5%) (p‐values T1D vs. LADA and T1D vs. T2D <0.01, p‐value LADA vs. T2D ns) (Table 3). Regression analysis on DKD and microalbuminuria, adjusted for age, sex and duration of diabetes confirmed these findings (see Table S1).

TABLE 3.

Diabetes associated complications, comorbidities and autoimmune disease at follow‐up.

	n	T1D	p‐Value T1D versus LADA	n	LADA	p‐Value LADA versus T2D	n	T2D	p‐Value T1D versus T2D
Microvascular
% Retinopathy	96	5.21	1.000	71	4.23	1.000	2168	6.13	1.000
% Diabetic kidney disease	191	19.9	<0.0001	224	44.2	1.000	7792	38.5	<0.0001
% Microalbuminuria	153	8.5	<0.001	175	28.6	1.000	5189	24.7	<0.0001
% Macroalbuminuria	153	0.00	0.333	175	3.43	1.000	5189	2.22	0.88
% Neuropathy	214	43.5	0.007	276	59.8	1.000	12 302	56.1	<0.0001
% Autonomous neuropathy	214	2.80	1.000	276	2.54	1.000	12 302	3.31	1.000
% Peripheral neuropathy	214	42.1	0.077	276	55.1	0.007	12 302	43.9	1.000
Macrovascular
% Myocardial infarction	214	3.27	1.000	276	4.71	1.000	12 302	5.87	1.000
% Stroke	214	2.34	1.000	276	3.99	1.000	12 302	4.80	1.000
% PAD	214	6.54	0.17	276	13.8	1.000	12 302	10.1	1.000
% Cardiovascular revascularisation	214	3.74	1.000	276	2.54	1.000	12 302	3.02	1.000
Acute diabetes associated complication
% Severe hypoglycemia	214	9.81	<0.0001	276	0.36	1.000	12 213	1.02	<0.0001
% Hypoglycemic coma	214	7.01	<0.001	276	0.36	1.000	12 213	0.29	<0.0001
% DKA	214	0.93	1.000	276	0.00	1.000	12 302	0.08	0.001
Autoimmune disease
% Celiac disease	214	0.47	1.000	276	0.00	1.000	12 302	0.20	1.000
% Hashimoto thyroiditis	214	8.41	0.064	276	2.54	1.000	12 302	2.31	<0.0001
% Addison disease	214	0.00	1.000	276	0.00	1.000	12 302	0.03	1.000
% Atrophic gastritis	214	0.47	1.000	276	0.00	1.000	12 302	0.19	1.000
% Vitiligo	214	1.87	0.339	276	0.00	1.000	12 302	0.20	<0.0001

Note: Data are in percent. n represents the count of individuals for whom data was provided. Significance is established by p‐values <0.05 through the application of both Wilcoxon rank sum test and the chi‐squared test, adjusted with the Bonferroni–Holm method. Significant p‐values are printed in bold.

Abbreviations: DKA, diabetic ketoacidosis; LADA, latent autoimmune diabetes in adults; PAD, peripheral artery disease; T1D, type 1 diabetes; T2D, type 2 diabetes.

Neuropathy was most prevalent in individuals with LADA (59.8%), followed by people with T2D (56.1%) and to a lesser extent in T1D (43.5%) (p‐values T1D vs. LADA and T1D vs. T2D <0.01, p‐value LADA vs. T2D ns). Similar values were found peripheral neuropathy, being the most prevalent in individuals with LADA (55.1%), followed by those with T2D (43.9%) and T1D (42.1%) (p‐value LADA vs. T2D <0.01, p‐values T1D vs. LADA and T1D vs. T2D ns) (Table 3). The regression analysis on peripheral neuropathy, adjusted for age, sex and duration of diabetes showed that only differences between LADA and T2D were statistically significant (see Table S1).

3.4. Macrovascular complications (at follow‐up)

During the follow‐up period, the overall prevalences for macrovascular complications, including myocardial infarction and stroke, were low in our study and differences were not significant (Table 3).

3.5. Acute complications (at follow‐up)

In addition to diabetes associated chronic complications, we assessed the risk for acute complications including hypoglycemia and DKA during follow‐up. 9.81%, 7.01%, and 0.93% of patients with T1D had severe hypoglycemia, hypoglycemic coma, or DKA, which was significantly higher in comparison to T2D (1.02%, 0.29%, 0.08%) and LADA (0.36%, 0.36%, 0.00%) (p‐values T1D vs. LADA and T1D vs. T2D <0.01, p‐value LADA vs. T2D ns) (Table 3).

3.6. Other autoimmune diseases (at follow‐up)

We also examined the prevalence of autoimmune diseases. Whilst overall numbers were small, people with T1D show the highest prevalence with Hashimoto thyroiditis and vitiligo as expected, whilst LADA and T2D did not differ significantly. The prevalence of celiac disease and Addison disease was low and did not significantly differ between diabetes groups (Table 3).

3.7. Comparing different durations of disease

We conducted an additional analysis with different durations of diabetes: one with a duration of disease of less than 5 years and one with a duration of disease of more than 5 years.

As expected, amongst patients with a longer duration of diabetes (>5 years), the prevalences for microvascular complications (DKD [T1D 22.5%, LADA 49.0%, T2D 48.9%], neuropathy [T1D 57.0%, LADA 75.7%, T2D 67.5%]) and macrovascular complications (myocardial infarction [T1D 5.81, LADA 7.21%, T2D 7.38%], stroke [T1D 2.33%, LADA 4.50%, T2D 6.14%] and peripheral artery disease PAD [T1D 11.6%, LADA 22.5%, T2D 11.9%]) were higher than amongst patients with a short duration of diabetes (≤5 years) (DKD [T1D 18.0%, LADA 40.0%, T2D 32.6%], neuropathy [T1D 34.4%, LADA 49.1%, T2D 49.5], myocardial infarction [T1D 1.56%, LADA 3.03%, T2D 4.99%], stroke [T1D 2.34%, LADA 3.46%, T2D 4.03%] and PAD [T1D 3.13%, LADA 7.88%, T2D 9.03%]) (see Tables S2 and S3).

3.8. Gender differences

We also did an additional analysis of diabetes associated complications, separated by sex. Prevalences of neuropathy and myocardial infarction were higher in men (neuropathy [T1D 45.5%, LADA 64.3%, T2D 57.4%], myocardial infarction [T1D 2.68%, LADA 6.70%, T2D 7.64%]) than in women (neuropathy [T1D 41.2%, LADA 51.6%, T2D 54.4%], myocardial infarction [T1D 3.92%, LADA 1.03%, T2D 3.52%]). Only the prevalence of peripheral artery disease (PAD) in people with LADA and T2D was higher in men (LADA 17.9%, T2D 11.5%) than in women (LADA 6.19%, T2D 8.23%) (see Tables S4 and S5).

The overall prevalence of autoimmune diseases in men and in women were low; however, the prevalence of Hashimoto disease was higher in women (T1D 11.8%, LADA 3.09%, T2D 4.46%) than in men (T1D 5.36%, LADA 2.23%, T2D 0.68%) (see Tables S4 and S5).

4. DISCUSSION

LADA has been proposed to have characteristics similar to T1D and also to T2D. ²¹ , ²² The objective of this study was to compare anthropometric and clinical data, determine the prevalence of micro‐ and macrovascular complications in patients with LADA, T1D and T2D reported in the DPV register during 1995–2022. We confirmed that LADA is positioned in between T1D and T2D for anthropometric data, glycemic control and C‐peptide as a marker for residual insulin secretion. At manifestation and follow‐up, people with LADA in the DPV cohort had an intermediate BMI and waist circumference compared to those with T2D and T1D, which is similar to the Fremantle study, a Chinese study, a Swedish cohort, and findings from the German Diabetes Study (GDS). ¹² , ¹³ , ¹⁵ , ²³ , ²⁴ , ²⁵ HbA1c and BMI for the different diabetes forms were similar in these studies. In a separate analysis from the DPV initiative, paediatric individuals with LADY were also in between T1D and T2D. ¹³

Diabetes is a risk factor for both microvascular diseases, such as retinopathy, DKD and neuropathy and macrovascular diseases, including myocardial infarction and stroke.

In our study, the prevalence of dyslipidemia in people diagnosed with LADA (90.7%) was higher compared with T2D (81.8%) and T1D (75.7%). Similarly, a higher prevalence of arterial hypertension in people with LADA (77.7%) compared to those with T2D (60.4%) and T1D (39.7%) was found. Our data are in contrast to the Swedish study by Wei et al., which showed that people with T2D have a higher prevalence of arterial hypertension and dyslipidemia than those with LADA. ¹¹ These Swedish data were in line with another study from China by Fan et al., that also showed higher prevalences of hypertension and dyslipidemia in T2D versus LADA. ²⁶ Similarly, in a study from Australia (Fremantle Diabetes Study), the prevalence of dyslipidemia, but not arterial hypertension was higher in T2D compared to LADA. ²⁴ An explanation, why our findings differed from others may come from different cohorts investigated and diabetes duration at follow‐up. For example, the Swedish cohort assessed arterial hypertension and dyslipidemia after 3 months of diagnosis with diabetes, whereas in our study, these parameters were assessed after a median follow‐up period of 3.5 years in T2D and 3.8 years in LADA. People with LADA in our study therefore had a longer duration of disease, which could possibly contribute to higher prevalence of hypertension and dyslipidemia. The higher prevalence of hypertension and dyslipidemia in people with LADA in our study may also be attributed to a more thorough management of dyslipidemia and hypertension in T2D compared to LADA and T1D.

In the paediatric DPV study, the prevalence of hypertension and dyslipidemia was higher in LADYs classified as having T2D compared to those classified as having T1D, with no notable differences observed compared to those with ‘classical’ T2D. ¹³

In our analysis, only a fraction of individuals received lipid‐lowering or antihypertensive medication. Despite high rates of dyslipidemia and arterial hypertension, treatment rates were low for both conditions amongst individuals with LADA and T2D. Our data from the DPV cohort are similar to the DIAB‐CORE study from Germany in people with T2D, which showed that whilst 82.5% of them had arterial hypertension, only 22.9% of these individuals had controlled blood pressure of <140/90 mmHg. Whilst 54.8% of them had dyslipidemia, only 22.5% of these had controlled lipid levels (TC/HDL < 5). ²⁷ These findings of incomplete treatment coverage for hypertension and dyslipidemia are also in line with recent observations in the United States from Fang et al. ²⁸ These data of insufficient treatment could be due to patient preference, as well as financial restrictions. Overall, data from our study and others indicate that the treatment of hypertension and dyslipidemia requires more attention to reduce cardiovascular risk according to established guidelines not only in people with LADA, but also in those with T1D and T2D.

Clinical outcomes, such as myocardial infarction and stroke, were of low prevalence in our study; therefore, interpretation is limited. Similar results of non‐significant differences in the prevalence of stroke and myocardial infarction between T2D and LADA were found in the Australian study by Myhill et al. ²⁴

When we compared the prevalence of microvascular diseases at follow‐up, individuals with LADA showed a higher frequency of DKD than those with T2D and T1D, similar to our findings with hypertension and dyslipidemia. Due to small case numbers, the prevalence of retinopathy showed no statistical differences and interpretation is limited. These findings are in contrast to the Swedish registry study., which showed a higher prevalence of DKD in T2D than in LADA.

People with LADA in their study had a higher prevalence of diabetic retinopathy than people with T2D. Although our findings differ, both their study and ours show that people with LADA seem to be at increased risk for microvascular complications compared to T2D, and in our study also in comparison to T1D. This should raise awareness to improve glucose control and antihypertensive treatment in those with LADA, to prevent DKD and diabetic retinopathy. ¹¹

Similar to DKD, people with LADA in our study also showed a higher prevalence of peripheral neuropathy compared to T2D and T1D. In our study, 55.1% of those with LADA, 43.9% of those with T2D and 42.1% of those with T1D had peripheral neuropathy. Our data differed from the findings by d'Onofrio et al., who reported that in their cohort, patients with T1D (36%) exhibited the highest level of peripheral neuropathy, followed by those with LADA (34%) and T2D (22%), although the differences were not statistically significant in their study. The overall duration of diabetes in their cohort (mean value T1D: 19.4 years, LADA: 11.6, T2D: 15.1) was longer than in our cohort, which could explain these differences. ²⁹

In addition, we assessed short term complications such as hypoglycemia and DKA in the DPV cohort at follow‐up. As expected, the prevalence of severe hypoglycemia amongst individuals with T1D (9.81%) was higher compared to T2D (1.02%) and LADA (0.36%). Similar results were found in other studies. ²⁶ , ³⁰ We also found the prevalence of DKA to be highest in T1D, which is in line with other research. ¹³ , ²⁶ , ³¹

Our analysis has strengths and limitations. The strength of this analysis is the long observation period from 1995 to 2022 and the large sample size of more than 35 000 individuals with diabetes across 197 medical centres in Germany and Austria which cover an extensive geographic range and represent diverse health‐care environments and the comparison of the three diabetes forms T1D, LADA and T2D, therefore representing the German population. All individuals at follow‐up also had data from diabetes diagnosis. The diagnosis of LADA and T1D was based on positive antibodies.

Limitations are related to the nature of the DPV registry; not every parameter investigated was available for every individual at diagnosis and at follow‐up, and not every parameter was documented in each patient. In addition, islet antibody assessment in people with T2D was done in a small fraction only (in 1307 of 35 272 people with T2D), and thereby, the clinical diagnosis of T2D may include some people with LADA. Although we examined a large sample size of patients, some patient groups might be underrepresented, such as individuals recently diagnosed with T2D who do not receive medication yet. We performed a cross‐sectional analysis, both at diagnosis and at follow‐up. Median time at follow‐up was 3.5–3.9 years and yielded low numbers of myocardial infarction and stroke; a longer interval may be more informative on cardiovascular outcomes. Also, there is potential for variability in data collection across the 197 medical centres in two countries.

In conclusion, amongst people diagnosed between 30 and 70 years, individuals with LADA are phenotypically between those with T1D and T2D in the DPV cohort. But people with LADA exhibit a higher prevalence of dyslipidemia and arterial hypertension than people with T2D and T1D. The prevalence of DKD was higher in LADA than in T2D, and the prevalence of peripheral neuropathy was higher in LADA than in T2D and T1D. People with LADA should therefore be carefully assessed with regard to therapy and cardiovascular complications. Further longitudinal analysis of this cohort is warranted to study the long‐term cardiovascular events such as myocardial infarction, stroke and the progression of coronary artery disease.

FUNDING INFORMATION

This study was supported by the German Centre for Diabetes Research (DZD, 82DZD14E03). Further financial support was received by the German Robert Koch Institute (RKI) and by the German Diabetes Association (DDG).

CONFLICT OF INTEREST STATEMENT

SM has received travel expenses support and honoraria for speaking and consulting activities from Biomarin, Lilly Germany, Novo Nordisk and Sanofi. SM also states that her husband is an employee at Novo Nordisk. JKM is a member in the advisory boards of Abbott Diabetes Care, Becton‐Dickinson/Embecta, Biomea, Eli Lilly, Medtronic, Novo Nordisk, Pharmasens, Roche Diabetes Care, Sanofi and Viatris, received speaker honoraria from Abbott Diabetes Care, A. Menarini Diagnostics, Becton‐Dickinson/Embecta, Boehringer‐Ingelheim, Eli Lilly, MedTrust, Novo Nordisk, Roche Diabetes Care, Sanofi, Servier and Ypsomed and is shareholder of decide Clinical Software GmbH and elyte Diagnostics. NCS is employed at Lilly Deutschland GmbH and a stockholder of Lilly shares. All other authors declare no conflicts of interest.

PEER REVIEW

The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer‐review/10.1111/dom.16048.

Supporting information

Data S1. Supporting tables.

Golomb RC, Tittel SR, Welters A, et al. Increased cardiovascular risk in people with LADA in comparison to type 1 diabetes and type 2 diabetes: Findings from the DPV registry in Germany and Austria. Diabetes Obes Metab. 2025;27(2):563‐573. doi: 10.1111/dom.16048

DATA AVAILABILITY STATEMENT

Due to protection of patient privacy and the available patient approval and the institutional data protection / ethics statements, patient level data are not allowed to be shared with outside investigators. Upon reasonable request, remote data analysis and joint analysis projects are possible options.