OBJECTIVE

Latent autoimmune diabetes in adults (LADA) is a heterogenous, slowly progressing autoimmune diabetes. We aim to contribute new knowledge on the long-term prognosis of LADA with varying degrees of autoimmunity by comparing it to type 2 diabetes and adult-onset type 1 diabetes.

RESEARCH DESIGN AND METHODS

This Swedish population-based study included newly diagnosed LADA (n = 550, stratified into LADAlow and LADAhigh by median autoimmunity level), type 2 diabetes (n = 2,001), adult-onset type 1 diabetes (n = 1,573), and control subjects without diabetes (n = 2,355) in 2007–2019. Register linkages provided information on all-cause mortality, cardiovascular diseases (CVDs), diabetic retinopathy, nephropathy, and clinical characteristics during follow-up.

RESULTS

Mortality was higher in LADA (hazard ratio [HR] 1.44; 95% CI 1.03, 2.02), type 1 (2.31 [1.75, 3.05]), and type 2 diabetes (1.31 [1.03, 1.67]) than in control subjects. CVD incidence was elevated in LADAhigh (HR 1.67; 95% CI 1.04, 2.69) and type 2 diabetes (1.53 [1.17, 2.00]), but not in LADAlow or type 1 diabetes. Incidence of retinopathy but not nephropathy was higher in LADA (HR 2.25; 95% CI 1.64, 3.09), including LADAhigh and LADAlow than in type 2 diabetes (unavailable in type 1 diabetes). More favorable blood pressure and lipid profiles, but higher HbA1c levels, were seen in LADA than type 2 diabetes at baseline and throughout follow-up, especially in LADAhigh, which resembled type 1 diabetes in this respect.

CONCLUSIONS

Despite having fewer metabolic risk factors than type 2 diabetes, LADA has equal to higher risks of death, CVD, and retinopathy. Poorer glycemic control, particularly in LADAhigh, highlights the need for improved LADA management.

Diabetes increases the risks of cardiovascular disease (CVD) and microvascular diseases, which may lead to kidney failure, blindness, and ultimately, premature death (1). The magnitude of the risk increase, as well as the nature of complications, varies by types of diabetes and by clinical, genetic, and lifestyle characteristics. Precision medicine holds the hope that better patient classification and more accurate assessment of prognostic factors may eventually lead to individualized and more efficient disease management and prevention (2).

Latent autoimmune diabetes in adults (LADA) is a hybrid form of diabetes. Its pathophysiology involves both autoimmunity and type 2 diabetes-like characteristics such as insulin resistance (3). Glycemic control is worse in LADA than in type 2 diabetes (4,5), possibly because of insufficient endogenous insulin production (3) and a lack of established treatment guidelines. This may put LADA patients at high risk of vascular complications since glycemic control is one of the main drivers of adverse outcomes in diabetes (6,7). LADA is common. An estimated 4.3–13.2% of people diagnosed with type 2 diabetes have autoantibodies indicative of LADA (3), and it accounts for 3–12% of all adult-onset diabetes (8). Despite this, only a few small prospective studies (4,916) have explored its prognosis. The largest prospective study to date is based on ∼600 participants with LADA from the U.K. Prospective Diabetes Study (UKPDS), a randomized clinical trial of type 2 diabetes conducted in 1991–2007 (4,9). It found that LADA conferred the same risk of CVD as type 2 diabetes (9) but higher long-term risks of microvascular complications (4). Prospective studies of LADA compared with type 1 diabetes are scarce (11). Further investigations into the prognosis of people with LADA are warranted, especially when considering the lack of contemporary, large-scale studies based on real-world data.

LADA is heterogeneous and encompasses a type 2 diabetes-like phenotype with less autoimmunity, more insulin resistance and obesity, and a more type 1 diabetes-like phenotype with more marked severe autoimmunity and insulin deficiency together with fewer metabolic risk factors (17). Whether such endotype heterogeneity will influence prognosis has only been addressed in relation to microvascular complications (4,16).

Our aim was to contribute with new knowledge on the prognosis of LADA by investigating all-cause mortality and vascular diseases in LADA with varying degrees of autoimmunity and describing treatment, glycated hemoglobin A1c (HbA1c), and metabolic trajectories. We used Swedish, prospective, population-based data and compared newly diagnosed LADA to adult-onset type 1 diabetes, type 2 diabetes, and people without diabetes.

Study Population

We used data from ESTRID (Epidemiological Study of Risk Factors for LADA and Type 2 Diabetes), a Swedish case-control study nested within the ANDIS (All New Diabetics in Scania) register, which aims to characterize all new cases of diabetes in the county of Scania (18). The ESTRID design has been described in detail (19). Briefly, we invited all individuals with LADA and a random sample of individuals with type 2 diabetes (4:1) to ESTRID, together with matched control subjects without diabetes (6:1) randomly sampled from the population (19). Participants provide detailed information on demographic and lifestyle factors (19). Eligible for the current study were all individuals with LADA (n = 550), type 2 diabetes (n = 2,001), and control subjects without diabetes (n = 2,355) included from 2007 to 2019 (Fig. 1). They were followed in the National Diabetes (NDR; 2009–2020), Patient (NPR; 1963–2019), Prescribed Drug (NPDR; 2005–2020), and Cause of Death Registers (CDR; 2009–2020). We supplemented the ESTRID study population with individuals with newly diagnosed type 1 diabetes identified in NDR (n = 1,573) (Fig. 1). ESTRID participants provided informed consent. Approval was obtained by the Ethical Review Board in Stockholm (2010/336-31/1, 2018/1036-32, 2022-02549-02).

Figure 1

Flowchart of included participants and register linkages. *The sample sizes for different outcomes differed since we excluded participants with CVD at baseline for the analysis of incident CVD and excluded those with microvascular diseases at baseline for the analysis of microvascular diseases. †Measurements from anyone in the full cohort in each diabetes duration were used, whenever possible.

Figure 1

Flowchart of included participants and register linkages. *The sample sizes for different outcomes differed since we excluded participants with CVD at baseline for the analysis of incident CVD and excluded those with microvascular diseases at baseline for the analysis of microvascular diseases. †Measurements from anyone in the full cohort in each diabetes duration were used, whenever possible.

Close modal

Diabetes Classification and Laboratory Analyses

In ANDIS, blood samples are collected at diagnosis and analyzed for GAD antibodies (GADA) and C-peptide (18). LADA was defined as age ≥35 years (3), GADA positivity (≥10 IU/mL), and C-peptide >0.3 nmol/L (20). This is in line with proposed criteria (21), except for C-peptide, which served as an indicator of remaining insulin production to separate LADA from type 1 diabetes, instead of time until insulin initiation. We categorized LADA into LADAlow and LADAhigh by median GADA level (250 IU/mL). Type 2 diabetes was defined as age ≥35 years, GADA negativity, and C-peptide >0.72 nmol/L (the definition of “classical” type 2 diabetes used in ANDIS [22]). From NDR we included all individuals diagnosed with type 1 diabetes at age ≥35 years (to increase the comparability with LADA) from 2007 to 2015, with exclusive insulin use within 6 months of diagnosis.

HOMA of insulin resistance (HOMA-IR), insulin sensitivity, and β-cell function (HOMA-B) were calculated based on fasting plasma glucose and C-peptide (23). DNA was analyzed to determine HLA genotypes (Supplementary Method 1).

Clinical Characteristics

NDR was created in 1996 to monitor the health of people with diabetes (24). From NDR, data were obtained on HbA1c, blood pressure, lipid profiles, and estimated glomerular filtration rate (eGFR) from multiple time points following diabetes diagnosis (Supplementary Method 2). Information on drug use was retrieved from NPDR, which records all prescribed drugs dispensed at Swedish pharmacies since 2005 (25) (Supplementary Method 2 and Supplementary Table 1).

Survival Outcomes

The primary outcomes were all-cause mortality, incidence of CVD, severe diabetic retinopathy (referred to as “retinopathy”), and severe diabetic nephropathy (referred to as “nephropathy”). We defined incident CVD as the first inpatient record of ischemic heart disease (ICD-10 codes: I20-I25), stroke (I60-I64), or heart failure (I50) in NPR (Supplementary Method 3). We also analyzed CVD rehospitalization among those with a history of CVD at baseline (Supplementary Method 3).

Retinopathy was defined as a diagnosis of severe nonproliferative retinopathy, preproliferative diabetic retinopathy or proliferative diabetic retinopathy in NDR, or a diagnosis of preproliferative diabetic retinopathy (E10.3B, E11.3B, E12.3B, E13.3B, E14.3B), proliferative diabetic retinopathy (H36.0B, E10.3C, E11.3C, E12.3C, E13.3C, E14.3C), diabetes with advanced eye disease (E10.3D-W, E11.3D-W, E12.3D-W, E13.3D-W, E14.3D-W), other proliferative retinopathy (H35.2), diabetic cataract (H28.0), retinal hemorrhage (H35.6), visual impairment (H54), or vitreous hemorrhage (H43.1) recorded in NPR, or death from diabetic retinopathy (E10.3, E11.3, E12.3, E14.3, H28.0, H36.0). Nephropathy included renal failure (N17-N19), renal transplant (Z94.0), dialysis (Z49, Z99.2), and death from any diabetic nephropathy (E10.2, E11.2, E12.2, E13.2, E14.2, N08.3, N17-N19, Z49, Z94.0, Z99.2). We could not analyze nephropathy or retinopathy in type 1 diabetes because we did not have ICD codes at the necessary level of detail.

Statistical Analysis

Characteristics

Basic characteristics are presented as means or medians for continuous variables and as proportions for categorical variables. Differences across groups were tested using the Student t test, Kruskal-Wallis test, or χ2 test.

Survival Analysis

We calculated person-years from baseline to the date of first event, death, or the end of follow-up (31 December 2019 for vascular outcomes and 31 December 2020 for mortality, based on the availability of the different registers), or the occurrence of diabetes (for control subjects), whichever came first. Baseline was date of diabetes diagnosis/enrollment for diabetes case subjects/control subjects (date of enrollment for everyone when the outcome was mortality, to minimize immortal time bias).

Curves for cumulative probabilities were estimated with Cox proportional hazards regression models for all-cause mortality and with competing-risks regression models for vascular outcomes (26), with time since baseline as the time scale and with age, sex, and calendar year as covariates.

Cox models were also used to estimate the hazard ratios (HRs) of mortality and vascular diseases by diabetes and GADA status, with attained age as the time scale. The main model for LADA and type 2 diabetes (vs. control subjects) was adjusted for sex, calendar year at diagnosis/enrollment, education, smoking, alcohol intake, physical activity, BMI, and history of CVD (only in analyses of mortality). The models for type 1 diabetes did not include education or lifestyle since such information was unavailable for these individuals. When comparing LADA or type 1 diabetes to type 2 diabetes, we additionally adjusted for diabetes duration. We also performed sensitivity analyses with further adjustments for potential mediators such as HbA1c, blood pressure, lipids, and eGFR (Supplementary Table 3). This had minor influence on the results, and these variables were therefore not retained in the main analyses.

Trajectories

Trajectories of HbA1c, blood pressure, lipid profiles, eGFR, and treatments (glucose-lowering drugs, antihypertensive drugs, or statins) were estimated using generalized linear models with cluster robust SEs to account for interdependence among measurements in the same individual (27) (Supplementary Method 4). The models (except for treatments) were adjusted for sex, age, and calendar year.

We used SAS 9.4 or Stata 17.0 software. All tests were two-sided, with P < 0.05 indicating statistical significance.

Data and Resource Availability

National register data are available from Statistics Sweden and the Swedish National Board of Health and Welfare, but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. ANDIS and ESTRID data were not publicly available.

Baseline Characteristics

Mean age at diagnosis/recruitment was 59.6 years in LADA, 63.0 years in type 2 diabetes, 47.7 years in type 1 diabetes, and 58.8 years in control subjects (Table 1). Compared with people with type 2 diabetes, those with LADA had a higher frequency of high-risk HLA genotypes, lower prevalence of obesity, lower HOMA-IR, and better blood pressure and lipid profiles but lower HOMA-B and worse glycemic control. LADAhigh was characterized by lower HOMA-B and HOMA-IR than LADAlow. Individuals with type 2 diabetes or LADAlow were more likely to have a history of CVD at diagnosis than those with LADAhigh. People with type 1 diabetes had lower CVD prevalence and worse glycemic control than those with LADA or type 2 diabetes. During a median follow-up of 5.9 years, we identified 509 deaths (LADA 49), 386 first CVD (LADA 35), 190 retinopathy (LADA 74), and 77 nephropathy (LADA 10) events.

Table 1

Basic characteristics in LADA and other adult-onset diabetes

ESTRID participantsType 1 diabetes in the NDR (n = 1,573)††
Population control subjects (n = 2,355)Type 2 diabetes (n = 2,001)LADA, overall (n = 550)P value#LADAlow (n = 264)LADAhigh (n = 275)P value**
Men, % 47.2 60.5 52.7 0.002 58.0 47.3 0.013 60.7 
Age at diagnosis, years* 58.8 (13.8) 63.0 (10.4) 59.6 (12.1) <0.001 60.0 (12.0) 59.0 (12.2) 0.323 47.7 (10.9) 
Intermediate education or above, % 77.3 64.6 72.5 <0.001 68.6 77.5 0.020 — 
Current smokers, % 18.9 19.9 23.4 0.072 18.7 26.9 0.024 — 
Daily alcohol intake ≥15g, % 23.3 24.2 23.0 0.571 20.3 25.4 0.164 — 
Physically inactive, % 14.3 24.8 19.0 0.005 21.5 17.2 0.219 — 
Obesity, % 15.3 52.2 33.8 <0.001 36.4 31.6 0.247 — 
History of CVD, % 8.1 18.1 11.6 <0.001 15.5 7.3 0.002 2.5 
Prevalent nephropathy, % — 2.1 2.0 0.830 2.2 1.3 0.447 — 
Prevalent retinopathy, % — 0.2 0.4 0.483 0.4 0.3 0.957 — 
With high-risk HLA genotypes, % — 31.2 59.1 <0.001 52.9 65.9 0.012 — 
GADA in IU/mL — — 250 (30–250) — 29 (17–77) 250 (250–250) <0.001 – 
C-peptide in nmol/L — 1.2 (1.0–1.6) 0.8 (0.5–1.2) <0.001 1.0 (0.6–1.4) 0.7 (0.5–1.0) <0.001 — 
HOMA-B — 71 (45–96) 43 (16–70) <0.001 50 (21–82) 32 (14–60) <0.001 — 
HOMA-insulin sensitivity — 28 (21–36) 34 (22–52) <0.001 31 (21–43) 39 (23–56) 0.004 — 
HOMA-IR — 3.6 (2.8–4.8) 3.0 (1.9–4.6) <0.001 3.2 (2.3–4.8) 2.6 (1.8–4.4) 0.004 — 
HbA1c in mmol/mol§ — 50 (44–65) 56 (47–78) <0.001 54 (46–73) 58 (48–85) 0.019 67 (55–89) 
HbA1c within glycemic target, %§ — 58.4 41.0 <0.001 43.2 39.0 0.449 22.0 
Blood pressure within control target, %§ — 47.9 55.0 0.029 53.8 56.5 0.647 76.9 
With favorable HDL cholesterol level, %§ — 43.2 52.7 0.006 46.3 58.0 0.063 75.1 
With low-risk LDL cholesterol level, %§ — 32.1 25.2 0.037 26.3 25.2 0.850 30.7 
With low-risk triglyceride level, %§ — 44.0 62.1 <0.001 59.3 65.8 0.304 80.1 
With normal eGFR, % — 89.2 92.6 0.086 90.6 94.1 0.262 96.1 
Use of noninsulin glucose-lowering drugs, % — 66.3 63.6 0.241 59.1 68.0 0.032 
Insulin use, % — 5.2 34.7 <0.001 28.0 42.2 <0.001 97.1 
Use of antihypertensive drugs, % — 70.1 51.1 <0.001 54.5 46.9 0.076 16.1 
Statins use, % — 47.6 40.4 <0.003 41.7 38.9 0.514 17.3 
ESTRID participantsType 1 diabetes in the NDR (n = 1,573)††
Population control subjects (n = 2,355)Type 2 diabetes (n = 2,001)LADA, overall (n = 550)P value#LADAlow (n = 264)LADAhigh (n = 275)P value**
Men, % 47.2 60.5 52.7 0.002 58.0 47.3 0.013 60.7 
Age at diagnosis, years* 58.8 (13.8) 63.0 (10.4) 59.6 (12.1) <0.001 60.0 (12.0) 59.0 (12.2) 0.323 47.7 (10.9) 
Intermediate education or above, % 77.3 64.6 72.5 <0.001 68.6 77.5 0.020 — 
Current smokers, % 18.9 19.9 23.4 0.072 18.7 26.9 0.024 — 
Daily alcohol intake ≥15g, % 23.3 24.2 23.0 0.571 20.3 25.4 0.164 — 
Physically inactive, % 14.3 24.8 19.0 0.005 21.5 17.2 0.219 — 
Obesity, % 15.3 52.2 33.8 <0.001 36.4 31.6 0.247 — 
History of CVD, % 8.1 18.1 11.6 <0.001 15.5 7.3 0.002 2.5 
Prevalent nephropathy, % — 2.1 2.0 0.830 2.2 1.3 0.447 — 
Prevalent retinopathy, % — 0.2 0.4 0.483 0.4 0.3 0.957 — 
With high-risk HLA genotypes, % — 31.2 59.1 <0.001 52.9 65.9 0.012 — 
GADA in IU/mL — — 250 (30–250) — 29 (17–77) 250 (250–250) <0.001 – 
C-peptide in nmol/L — 1.2 (1.0–1.6) 0.8 (0.5–1.2) <0.001 1.0 (0.6–1.4) 0.7 (0.5–1.0) <0.001 — 
HOMA-B — 71 (45–96) 43 (16–70) <0.001 50 (21–82) 32 (14–60) <0.001 — 
HOMA-insulin sensitivity — 28 (21–36) 34 (22–52) <0.001 31 (21–43) 39 (23–56) 0.004 — 
HOMA-IR — 3.6 (2.8–4.8) 3.0 (1.9–4.6) <0.001 3.2 (2.3–4.8) 2.6 (1.8–4.4) 0.004 — 
HbA1c in mmol/mol§ — 50 (44–65) 56 (47–78) <0.001 54 (46–73) 58 (48–85) 0.019 67 (55–89) 
HbA1c within glycemic target, %§ — 58.4 41.0 <0.001 43.2 39.0 0.449 22.0 
Blood pressure within control target, %§ — 47.9 55.0 0.029 53.8 56.5 0.647 76.9 
With favorable HDL cholesterol level, %§ — 43.2 52.7 0.006 46.3 58.0 0.063 75.1 
With low-risk LDL cholesterol level, %§ — 32.1 25.2 0.037 26.3 25.2 0.850 30.7 
With low-risk triglyceride level, %§ — 44.0 62.1 <0.001 59.3 65.8 0.304 80.1 
With normal eGFR, % — 89.2 92.6 0.086 90.6 94.1 0.262 96.1 
Use of noninsulin glucose-lowering drugs, % — 66.3 63.6 0.241 59.1 68.0 0.032 
Insulin use, % — 5.2 34.7 <0.001 28.0 42.2 <0.001 97.1 
Use of antihypertensive drugs, % — 70.1 51.1 <0.001 54.5 46.9 0.076 16.1 
Statins use, % — 47.6 40.4 <0.003 41.7 38.9 0.514 17.3 

Data are presented as mean (SD) or median (IQR), unless indicated otherwise. We presented basic characteristics of participants as means (normal distribution) or medians (nonnormal distribution) for continuous variables and as proportions for categorical variables. Differences between diabetes status were tested using the Student t test (continuous variables, normal distribution), Kruskal-Wallis test (continuous variables, nonnormal distribution), or χ2 test (categorical variables).

*

Age at enrollment for population controls.

At the time of diabetes diagnosis.

Only calculated for participants with C-peptide ranging from 0.2 to 3.5 nmol/L and fasting glucose ranging from 3.0 to 25.0 mmol/L.

§

Measured within 3 months after diabetes diagnosis. According to the American Diabetes Association, HbA1c within target was defined as HbA1c<53 mmol/mol; blood pressure within control target was defined as systolic blood pressure <140 mmHg, and diastolic blood pressure <90 mmHg; low-risk HDL level was defined as HDL >1.0 mmol/L in men and >1.3 mmol/L in women; low-risk LDL profile was defined as LDL <2.6 mmol/L; low-risk triglyceride level was defined as triglyceride <1.7 mmol/L.

At 6 months after diabetes diagnosis.

#

Comparison between LADA and type 2 diabetes.

**

Comparison between LADAlow and LADAhigh.

††

We did not have data on lifestyle factors, HLA, GADA, C-peptide, or HOMA for type 1 diabetes since these variables were only available for ESTRID participants.

All-Cause Mortality

People with LADA (HR 1.44; 95% CI 1.03, 2.02), type 1 diabetes (HR 2.31; 95% CI 1.75, 3.05), or type 2 diabetes (HR 1.31; 95% CI 1.03, 1.67) had higher mortality rates than control subjects (Supplementary Fig. 1 and Supplementary Table 2). Stratifying LADA by degrees of autoimmunity revealed excess mortality only in LADAlow (HR 1.74; 95% CI 1.18, 2.57) and not in LADAhigh (Fig. 2 and Supplementary Table 2). Mortality rates were similar in LADA and type 2 diabetes (HR 1.13; 95% CI 0.81, 1.57 for LADA vs. type 2 diabetes) but higher in type 1 diabetes (HR 1.79; 95% CI 1.33, 2.41) (Supplementary Table 3). Compared with type 1 diabetes, the HR was 0.49 (95% CI 0.33, 0.72) for LADA overall, 0.59 (95% CI 0.38, 0.90) for LADAlow, and 0.33 (95% CI 0.18, 0.61) for LADAhigh. Inspection of the causes of death (Supplementary Table 4) revealed that diabetes was listed as the underlying cause of death more frequently in type 1 diabetes (16%) than in LADA (10%) or type 2 diabetes (2%).

Figure 2

Cumulative probability of all-cause mortality (A), incident CVD (B), diabetic retinopathy (C) and diabetic nephropathy (D) over time in LADA by GADA levels and other types of diabetes. Curves were plotted based on Cox proportional hazards regression models for all-cause mortality and competing-risks regression models for other outcomes, with time since baseline as the time scale, and with adjustment for sex, age, and calendar year. HRs (95% CIs) were estimated in Cox models with attained age as the time scale. Models for LADA or type 2 diabetes vs. control subjects were adjusted for sex, calendar year, education, lifestyle, and history of CVD at baseline (only in the analysis of mortality). *Models for type 1 diabetes did not include education or lifestyle as covariates due to unavailability of data. When comparing LADA or type 1 diabetes to type 2 diabetes, we adjusted for diabetes duration in addition to the abovementioned covariates.

Figure 2

Cumulative probability of all-cause mortality (A), incident CVD (B), diabetic retinopathy (C) and diabetic nephropathy (D) over time in LADA by GADA levels and other types of diabetes. Curves were plotted based on Cox proportional hazards regression models for all-cause mortality and competing-risks regression models for other outcomes, with time since baseline as the time scale, and with adjustment for sex, age, and calendar year. HRs (95% CIs) were estimated in Cox models with attained age as the time scale. Models for LADA or type 2 diabetes vs. control subjects were adjusted for sex, calendar year, education, lifestyle, and history of CVD at baseline (only in the analysis of mortality). *Models for type 1 diabetes did not include education or lifestyle as covariates due to unavailability of data. When comparing LADA or type 1 diabetes to type 2 diabetes, we adjusted for diabetes duration in addition to the abovementioned covariates.

Close modal

CVD

In people free of CVD at baseline, we observed higher CVD incidence in LADAhigh (HR 1.67; 95% CI 1.04, 2.69) and type 2 diabetes (HR 1.53; 95% CI 1.17, 2.00) but not in LADAlow or type 1 diabetes compared with control subjects (Fig. 2 and Supplementary Table 2). Compared with type 2 diabetes, CVD incidence was similar in LADAhigh (HR 1.03; 95% CI 0.64, 1.65) but lower in LADAlow (HR 0.52; 95% CI 0.30, 0.91) and type 1 diabetes (HR 0.63; 95% CI 0.44. 0.89). The HR in LADAlow and LADAhigh versus type 1 diabetes was 0.76 (95% CI 0.41, 1.41) and 1.32 (95% CI 0.75, 2.31), respectively.

Among those with CVD at baseline, the incidence of CVD rehospitalization was higher in LADAlow (HR 2.78; 95% CI 1.48, 5.25), but not in other types of diabetes, compared with control subjects (Supplementary Table 5).

Microvascular Diseases

The incidence of retinopathy was higher in LADA than in type 2 diabetes, and the excess risk was observed in both LADAlow (HR 2.12; 95% CI 1.43, 3.14) and LADAhigh (HR 2.41; 95% CI 1.62, 3.59) (Fig. 2 and Supplementary Table 3). Adjustment for HbA1c attenuated these associations (Supplementary Table 3). The absolute risk of developing retinopathy within 10 years of diagnosis was ∼20% in LADA and 10% in type 2 diabetes (Fig. 2). For nephropathy, the tendency was toward lower incidence in LADA than in type 2 diabetes, but the CIs were wide (Fig. 2 and Supplementary Table 3).

Trajectories

The proportion of individuals reaching the HbA1c target (<53 mmol/mol [7.0%]) declined with diabetes duration irrespective of diabetes type (Fig. 3A). People with LADAhigh or type 1 diabetes fared worst: 10 years after diagnosis, the HbA1c target was met by 31% and 32% of individuals with LADAhigh or type 1 diabetes compared with 43% of LADAlow and 58% of type 2 diabetes. Within 6 months of diagnosis, 34.7% of people with LADA (42.2% of LADAhigh and 28.0% of LADAlow) and 5.2% of people with type 2 diabetes used insulin (Table 1), and the proportion increased (Fig. 3B) to ∼80% in LADAhigh and 60% in LADAlow 10 years after diagnosis (Supplementary Fig. 2). Approximately 15% of people with LADA with C-peptide ≤0.7 nmol/L or LADAhigh and 25% of people with LADAlow or type 2 diabetes did not retrieve any glucose-lowering drugs during follow-up (Supplementary Figs. 3 and 4). Better control of blood pressure and lipids was observed in LADA (especially LADAhigh) and type 1 diabetes than in type 2 diabetes throughout the follow-up (Fig. 3 and Supplementary Figs. 5 and 6). Use of antihypertensive drugs was more frequent in type 2 diabetes than in LADA or type 1 diabetes (Supplementary Fig. 3). Treatment patterns in LADA with HbA1c, blood pressure, or lipid levels out of range at 3 and 5 years after diagnosis are summarized in Supplementary Table 6. The results revealed that 18.6% of LADAhigh with HbA1c above target were not on insulin treatment, >60% of LADA individuals with LDL cholesterol above target were not on statins treatment, and >20% of LADA individuals who failed to reach blood pressure control targets were not treated with antihypertensive drugs 5 years after the LADA diagnosis. The eGFR levels did not vary across diabetes types (Supplementary Figs. 5 and 6).

Figure 3

Trajectories of HbA1c (A), insulin use (B), and triglycerides (C) within control target in LADA and other types of diabetes. HbA1c within target was defined as HbA1c <53 mmol/mol (7.0%) according to the American Diabetes Association; low-risk triglyceride level was defined as triglyceride <1.7 mmol/L. Generalized linear models with binominal distribution and logit link were fitted with clustered-adjusted SEs to account for nonindependence among measures from the same individual. The trajectories of HbA1c and triglycerides were based on 98.3% and 91.8% of case subjects with diabetes with any record of corresponding biomarkers, respectively. The models for HbA1c and triglycerides were further adjusted for sex, age at diabetes diagnosis, and year at diabetes diagnosis.

Figure 3

Trajectories of HbA1c (A), insulin use (B), and triglycerides (C) within control target in LADA and other types of diabetes. HbA1c within target was defined as HbA1c <53 mmol/mol (7.0%) according to the American Diabetes Association; low-risk triglyceride level was defined as triglyceride <1.7 mmol/L. Generalized linear models with binominal distribution and logit link were fitted with clustered-adjusted SEs to account for nonindependence among measures from the same individual. The trajectories of HbA1c and triglycerides were based on 98.3% and 91.8% of case subjects with diabetes with any record of corresponding biomarkers, respectively. The models for HbA1c and triglycerides were further adjusted for sex, age at diabetes diagnosis, and year at diabetes diagnosis.

Close modal

In this population-based Swedish study that followed newly diagnosed LADA for up to 10 years, we found similar all-cause mortality and CVD incidence but higher retinopathy incidence in LADA compared with type 2 diabetes. LADA was characterized by a better metabolic risk profile in blood pressure and lipid levels but by a lower probability of achieving adequate glycemic control than type 2 diabetes. This was particularly evident in individuals with LADAhigh, who together with people with adult-onset type 1 diabetes were least likely to have good glycemic control both at diagnosis and throughout the follow-up. A significant proportion of LADA individuals were not on any glucose-lowering treatment. These findings point to the need for improved glucose management of this common but largely unrecognized patient group.

We observed similar rates of CVD and mortality in LADA and type 2 diabetes, in line with results based on UKPDS (9) and most, but not all, previous studies (912,14). Interestingly, stratification of LADA by GADA levels at diagnosis revealed two groups with different disease trajectories: mortality was higher in LADAlow, while elevated rates of first CVD events were only observed in LADAhigh. The higher mortality in LADAlow can be explained by the fact that these individuals were more likely to have experienced CVD already at the diabetes diagnosis and had excess risk of recurrent CVD during the follow-up. Their phenotype was more type 2-like, with more insulin resistance and higher lipid levels, which confirms previous observations (28). These metabolic risk factors together with hyperglycemia could be present years before diabetes manifests and may explain their higher CVD prevalence at diagnosis. People who did not develop CVD under such high-risk circumstances may be less susceptible to CVD and therefore not be at excess risk once diabetes was diagnosed and the glycemic control was improved. This could explain why we did not observe an increased risk of a first CVD event after diabetes diagnosis in LADAlow. In LADAhigh, we found excess risk of experiencing a first CVD event after diagnosis, coupled with fewer metabolic risk factors but poorer glycemic control than both LADAlow and type 2 diabetes. The risk elevated was of similar magnitude in LADAhigh and type 2 diabetes. This suggests that the advantages of an overall healthier metabolic profile in LADAhigh are outweighed by worse glycemic control. Our study is the first to examine CVD risks and mortality in LADA with different degrees of autoimmunity. Notably, there were few events in the LADA subgroup analyses, and confirmation and replication of these findings are clearly warranted.

The incidence of retinopathy was higher in LADA than in type 2 diabetes. Among the few previous prospective studies, some did not observe any differences between LADA and type 2 diabetes (16), whereas the UKPDS reports a lower risk of retinopathy in LADA during the first 9 years following diagnosis and a higher risk beyond that (4). In contrast, we noted higher incidence of retinopathy in LADA throughout the follow-up. Excess risk of retinopathy was observed in LADAlow as well as in LADAhigh, both in our study and UKPDS. Hyperglycemia is the main cause of diabetic retinopathy, and the results may be explained by worse glycemic control in LADA. The excess risk attenuated but remained after adjustment for HbA1c. However, HbA1c is an indicator of mean glucose levels and does not capture glycemic variability that may also affect vascular health (29,30). Our analysis of nephropathy was hampered by small numbers, but the tendency was toward lower risks in LADA than in type 2 diabetes. In contrast, the trajectories of kidney function, indicated by eGFR levels, appeared similar across the different diabetes types (including type 1 diabetes). Previous studies investigating nephropathy in LADA were mostly cross-sectional, while sample sizes in the three prospective studies (4,11,16) were insufficient. Unfortunately, we could not compare microvascular diseases between LADA and type 1 diabetes. Previous studies on that topic are few, mostly cross-sectional, and the results inconclusive (11,3135).

We confirm previous observations of higher HbA1c levels in LADA than in type 2 diabetes (4,5). In addition, we show that glycemic control was especially poor in individuals with LADAhigh, who resembled people with type 1 diabetes in this respect. One explanation is insulin deficiency caused by the underlying autoimmune reactivity that characterizes type 1 diabetes and LADA (especially LADAhigh) but not type 2 diabetes. Lack of established treatment guidelines for LADA may also contribute. In 2020, an expert committee proposed treatment algorithms for LADA (17), suggesting LADA with C-peptide <0.3 nmol/L should be treated as type 1 diabetes, while insulin in combination with oral glucose-lowering drugs should be considered for LADA with C-peptide between 0.3 and 0.7 nmol/L. We observed that 15% of LADA patients with C-peptide ≤0.7 nmol/L at diagnosis did not use any glucose-lowering drugs during the 10-year follow-up and that only 70% were treated with insulin. In addition, ∼15% of those with LADAhigh and 25% of those with LADAlow did not have any recording of glucose-lowering drugs. These findings, together with the observation of poor HbA1c trajectories, point at a need for measuring GADA and C-peptide for better phenotyping and improved management of LADA. The significant number of LADA patients with unfavorable lipids levels or high blood pressure, who are not receiving statins or antihypertensive treatment, underscores an additional treatment gap.

An important finding was the high mortality rate observed in type 1 diabetes that exceeded that observed in LADA and type 2 diabetes. The proportion of deaths with diabetes listed as the main cause was higher in type 1 diabetes. Such causes include acute diabetes-related complications, such as diabetic coma and ketoacidosis, that are rarely seen in people with LADA or type 2 diabetes. This finding adds to the limited knowledge on the prognosis of people with adult-onset type 1 diabetes (36) and warrants further investigations.

Strengths include the use of contemporary, population-based data and linkage to nationwide registers with virtually no loss of follow-up. The validity of most diagnoses recorded in these registers, including myocardial infarction and stroke, is high (37,38). Furthermore, diabetic retinopathy is diagnosed by ophthalmologists in Sweden, ensuring high sensitivity and specificity (39). Through the registers we could provide real-world data on treatment and clinical characteristics across the different diabetes types from the time of diagnosis over a 10-year follow-up period. LADA is not recorded in any national Swedish register. This study was possible because of the ANDIS study, which enrolls all individuals with incident diabetes in a large region in the south of Sweden and carefully characterizes them. This allowed us to separate LADA from type 1 and 2 diabetes through assessment of GADA and C-peptide. ANDIS did not have enough individuals with adult-onset type 1 diabetes for viable analyses. Therefore, we identified these individuals from the NDR. Consequently, we lacked information on some clinical characteristics, education, and lifestyle factors for these individuals. Furthermore, we could not assess their risk of microvascular disease. The study was conducted in Sweden where health care is available to everyone at a very limited cost. Whether our findings are generalizable to other countries, health care systems, or more diverse populations with lower LADA prevalence, is uncertain. The few studies on the prognosis of LADA are primarily conducted in European and Chinese populations (4,916,3135,40), and there is a clear need for further investigations.

In conclusion, the risk of death and CVD in LADA was comparable to that in type 2 diabetes, while the risk of diabetic retinopathy was higher and glycemic control poorer. Furthermore, stratification by GADA levels reveals two subgroups with distinct disease trajectories: the LADA group with lower GADA were more likely to have CVD already at diabetes diagnosis and had increased risk of recurrent CVD and mortality following diagnosis. People with higher GADA had fewer metabolic risk factors but poor glycemic control and increased risk of a first CVD event after diagnosis. LADA is characterized by worse glycemic control than type 2 diabetes, particularly LADAhigh. This indicates that it is clinically informative to consider both the presence of autoantibodies and levels of antibody titers in individuals with adult-onset diabetes for a more precise diabetes management. Our findings are of clinical importance since improper LADA diagnosis and management may increase the risk of complications and subsequently raise mortality among these patients.

This article contains supplementary material online at https://doi.org/10.2337/figshare.23804568.

Funding. The ESTRID study was supported by the Swedish Research Council (2018-03035, 2022-00811), Research Council for Health, Working Life and Welfare (Forte, 2018-00337), and the Swedish Diabetes foundation (DIA2022-735). The ANDIS study was financed by ALF (Swedish governmental funding of clinical research), the Swedish Research Council project grant nos. 2020-02191 and 2015-2558, infrastructure grant nos. 2010-5983, 2012-5538, and 2014-6395, Linnaeus grant no. 349-2006-237, and strategic research grant nos. 2009-1039 (Excellence of Diabetes Research in Sweden [EXODIAB]) and IRC15-0067 (Lund University Diabetes Centre [LUDC]-Industrial Research Centre [IRC]). Y.W. received a scholarship from the China Scholarship Council (student no. 202006010041).

The sponsors had no role in the study design, data collection, data analysis and interpretation, writing of the report, or the decision to submit the article for publication.

Duality of Interest. K.H. has a position in Novo Nordisk, but the employer was not involved in any aspect of the work. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. Y.W. wrote the first draft of the manuscript. Y.W. and K.H. analyzed data. E.A., T.N., T.T., and S.C. acquired and managed data. T.A. and Y.Z. contributed to methodological issues. S.C. conceived and designed the study. All authors critically revised the manuscript for important intellectual content and made substantial contributions to the interpretation of data. All authors reviewed and approved the final manuscript. Y.W. and S.C. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the 17th Symposium of the International Diabetes Epidemiology Group (IDEG), Lisbon, Portugal, 2–5 December 2022, and at the 58th Annual Meeting of the European Association for the Study of Diabetes (EASD), Stockholm, Sweden, 19–23 September 2022. A non–peer reviewed version of this article was submitted to the SSRN preprint server (https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4287478) on 1 December 2022.

1.
Sarwar
N
,
Gao
P
,
Seshasai
SR
, et al.;
Emerging Risk Factors Collaboration
.
Diabetes mellitus, fasting blood glucose concentration, and risk of vascular disease: a collaborative meta-analysis of 102 prospective studies
.
Lancet
2010
;
375
:
2215
2222
2.
Nolan
JJ
,
Kahkoska
AR
,
Semnani-Azad
Z
, et al
.
ADA/EASD precision medicine in diabetes initiative: an international perspective and future vision for precision medicine in diabetes
.
Diabetes Care
2022
;
45
:
261
266
3.
Tuomi
T
,
Santoro
N
,
Caprio
S
,
Cai
M
,
Weng
J
,
Groop
L
.
The many faces of diabetes: a disease with increasing heterogeneity
.
Lancet
2014
;
383
:
1084
1094
4.
Maddaloni
E
,
Coleman
RL
,
Agbaje
O
,
Buzzetti
R
,
Holman
RR
.
Time-varying risk of microvascular complications in latent autoimmune diabetes of adulthood compared with type 2 diabetes in adults: a post-hoc analysis of the UK Prospective Diabetes Study 30-year follow-up data (UKPDS 86)
.
Lancet Diabetes Endocrinol
2020
;
8
:
206
215
5.
Tuomi
T
,
Carlsson
A
,
Li
H
, et al
.
Clinical and genetic characteristics of type 2 diabetes with and without GAD antibodies
.
Diabetes
1999
;
48
:
150
157
6.
Lind
M
,
Pivodic
A
,
Svensson
AM
,
Ólafsdóttir
AF
,
Wedel
H
,
Ludvigsson
J
.
HbA1c level as a risk factor for retinopathy and nephropathy in children and adults with type 1 diabetes: Swedish population based cohort study
.
BMJ
2019
;
366
:
l4894
7.
Rawshani
A
,
Rawshani
A
,
Franzén
S
, et al
.
Risk factors, mortality, and cardiovascular outcomes in patients with type 2 diabetes
.
N Engl J Med
2018
;
379
:
633
644
8.
Mishra
R
,
Hodge
KM
,
Cousminer
DL
,
Leslie
RD
,
Grant
SFA
.
A global perspective of latent autoimmune diabetes in adults
.
Trends Endocrinol Metab
2018
;
29
:
638
650
9.
Maddaloni
E
,
Coleman
RL
,
Pozzilli
P
,
Holman
RR
.
Long-term risk of cardiovascular disease in individuals with latent autoimmune diabetes in adults (UKPDS 85)
.
Diabetes Obes Metab
2019
;
21
:
2115
2122
10.
Wod
M
,
Thomsen
RW
,
Pedersen
L
,
Yderstraede
KB
,
Beck-Nielsen
H
,
Højlund
K
.
Lower mortality and cardiovascular event rates in patients with Latent Autoimmune Diabetes In Adults (LADA) as compared with type 2 diabetes and insulin deficient diabetes: a cohort study of 4368 patients
.
Diabetes Res Clin Pract
2018
;
139
:
107
113
11.
Luk
AOY
,
Lau
ESH
,
Lim
C
, et al
.
Diabetes-related complications and mortality in patients with young-onset latent autoimmune diabetes: a 14-year analysis of the prospective Hong Kong Diabetes Register
.
Diabetes Care
2019
;
42
:
1042
1050
12.
Laugsand
LE
,
Janszky
I
,
Vatten
LJ
, et al
.
Autoimmune diabetes in adults and risk of myocardial infarction: the HUNT study in Norway
.
J Intern Med
2016
;
280
:
518
531
13.
Olsson
L
,
Grill
V
,
Midthjell
K
,
Ahlbom
A
,
Andersson
T
,
Carlsson
S
.
Mortality in adult-onset autoimmune diabetes is associated with poor glycemic control: results from the HUNT Study
.
Diabetes Care
2013
;
36
:
3971
3978
14.
Myhill
P
,
Davis
WA
,
Bruce
DG
,
Mackay
IR
,
Zimmet
P
,
Davis
TM
.
Chronic complications and mortality in community-based patients with latent autoimmune diabetes in adults: the Fremantle Diabetes Study
.
Diabet Med
2008
;
25
:
1245
1250
15.
Isomaa
B
,
Almgren
P
,
Henricsson
M
, et al
.
Chronic complications in patients with slowly progressing autoimmune type 1 diabetes (LADA)
.
Diabetes Care
1999
;
22
:
1347
1353
16.
Hawa
MI
,
Buchan
AP
,
Ola
T
, et al
.
LADA and CARDS: a prospective study of clinical outcome in established adult-onset autoimmune diabetes
.
Diabetes Care
2014
;
37
:
1643
1649
17.
Buzzetti
R
,
Tuomi
T
,
Mauricio
D
, et al
.
Management of latent autoimmune diabetes in adults: a consensus statement from an international expert panel
.
Diabetes
2020
;
69
:
2037
2047
18.
Ahlqvist
E
,
Storm
P
,
Käräjämäki
A
, et al
.
Novel subgroups of adult-onset diabetes and their association with outcomes: a data-driven cluster analysis of six variables
.
Lancet Diabetes Endocrinol
2018
;
6
:
361
369
19.
Rasouli
B
,
Andersson
T
,
Carlsson
PO
, et al
.
Smoking and the risk of LADA: results from a Swedish population-based case-control study
.
Diabetes Care
2016
;
39
:
794
800
20.
Löfvenborg
JE
,
Andersson
T
,
Carlsson
PO
, et al
.
Coffee consumption and the risk of latent autoimmune diabetes in adults--results from a Swedish case-control study
.
Diabet Med
2014
;
31
:
799
805
21.
Fourlanos
S
,
Dotta
F
,
Greenbaum
CJ
, et al
.
Latent autoimmune diabetes in adults (LADA) should be less latent
.
Diabetologia
2005
;
48
:
2206
2212
22.
Groop
L
,
Pociot
F
.
Genetics of diabetes--are we missing the genes or the disease?
Mol Cell Endocrinol
2014
;
382
:
726
739
23.
The Oxford Centre for Diabetes
.
The Oxford Centre for Diabetes, Endocrinology and Metabolism
.
HOMA calculator. Accessed 22 September 2022. Available from http://www.dtu.ox.ac.uk/homacalculator/index.php/
24.
Eeg-Olofsson
K
,
Cederholm
J
,
Nilsson
PM
, et al
.
Glycemic control and cardiovascular disease in 7,454 patients with type 1 diabetes: an observational study from the Swedish National Diabetes Register (NDR)
.
Diabetes Care
2010
;
33
:
1640
1646
25.
Wettermark
B
,
Hammar
N
,
Fored
CM
, et al
.
The new Swedish Prescribed Drug Register–opportunities for pharmacoepidemiological research and experience from the first six months
.
Pharmacoepidemiol Drug Saf
2007
;
16
:
726
735
26.
Fine
JP
,
Gray
RJ
.
A proportional hazards model for the subdistribution of a competing risk
.
J Am Stat Assoc
1999
;
94
:
496
509
27.
Levy
M
,
Chen
Y
,
Clarke
R
, et al.;
China Kadoorie Biobank Collaborative Group
.
Socioeconomic differences in health-care use and outcomes for stroke and ischaemic heart disease in China during 2009-16: a prospective cohort study of 0·5 million adults
.
Lancet Glob Health
2020
;
8
:
e591
e602
28.
Østergaard
JA
,
Laugesen
E
,
Leslie
RD
.
Should there be concern about autoimmune diabetes in adults? Current evidence and controversies
.
Curr Diab Rep
2016
;
16
:
82
29.
Zhou
Z
,
Sun
B
,
Huang
S
,
Zhu
C
,
Bian
M
.
Glycemic variability: adverse clinical outcomes and how to improve it?
Cardiovasc Diabetol
2020
;
19
:
102
30.
Holt
RIG
,
DeVries
JH
,
Hess-Fischl
A
, et al
.
The management of type 1 diabetes in adults. A consensus report by the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD)
.
Diabetes Care
2021
;
44
:
2589
2625
31.
Maddaloni
E
,
Moretti
C
,
Del Toro
R
, et al
.
Risk of cardiac autonomic neuropathy in latent autoimmune diabetes in adults is similar to type 1 diabetes and lower compared to type 2 diabetes: a cross-sectional study
.
Diabet Med
2021
;
38
:
e14455
32.
Li
W
,
Cheng
Z
,
Song
Y
,
Fang
Y
,
Yang
M
,
Zhang
M
.
Is diabetic retinopathy affected by diabetes type? A retrospective study using electronic medical record data from patients with latent autoimmune diabetes in adults, type 1 diabetes, and type 2 diabetes
.
Acta Diabetol
2021
;
58
:
1503
1511
33.
Fadiga
L
,
Saraiva
J
,
Catarino
D
,
Frade
J
,
Melo
M
,
Paiva
I
.
Adult-onset autoimmune diabetes: comparative analysis of classical and latent presentation
.
Diabetol Metab Syndr
2020
;
12
:
107
34.
Hernández
M
,
López
C
,
Real
J
, et al
.
Preclinical carotid atherosclerosis in patients with latent autoimmune diabetes in adults (LADA), type 2 diabetes and classical type 1 diabetes
.
Cardiovasc Diabetol
2017
;
16
:
94
35.
Roh
MO
,
Jung
CH
,
Kim
BY
,
Mok
JO
,
Kim
CH
.
The prevalence and characteristics of latent autoimmune diabetes in adults (LADA) and its relation with chronic complications in a clinical department of a university hospital in Korea
.
Acta Diabetol
2013
;
50
:
129
134
36.
Holman
N
,
Wild
SH
,
Gregg
EW
,
Valabhji
J
,
Sattar
N
;
National Diabetes Audit Research Group
.
Comparison of mortality in people with type 1 and type 2 diabetes by age of diagnosis: an incident population-based study in England and Wales
.
Lancet Diabetes Endocrinol
2022
;
10
:
95
97
37.
Ludvigsson
JF
,
Andersson
E
,
Ekbom
A
, et al
.
External review and validation of the Swedish national inpatient register
.
BMC Public Health
2011
;
11
:
450
38.
Eriksson
A
,
Stenlund
H
,
Ahlm
K
, et al
.
Accuracy of death certificates of cardiovascular disease in a community intervention in Sweden
.
Scand J Public Health
2013
;
41
:
883
889
39.
Adamsson Eryd
S
,
Svensson
AM
,
Franzén
S
,
Eliasson
B
,
Nilsson
PM
,
Gudbjörnsdottir
S
.
Risk of future microvascular and macrovascular disease in people with Type 1 diabetes of very long duration: a national study with 10-year follow-up
.
Diabet Med
2017
;
34
:
411
418
40.
Li
J
,
Zhang
Y
,
Zhang
J
,
Dong
R
,
Guo
J
,
Zhang
Q
.
Oxidative stress and its related factors in latent autoimmune diabetes in adults
.
BioMed Res Int
2021
;
2021
:
5676363
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at https://www.diabetesjournals.org/journals/pages/license.