Clinical characteristics and disease progression of typical adult‐onset type 1 diabetes and insulin‐dependent latent autoimmune diabetes in adults

ABSTRACT

Aims

We investigated the characteristics of typical adult‐onset type 1 diabetes (T1D) and insulin‐dependent latent autoimmune diabetes in adults (LADA) in Thailand.

Methods

This retrospective chart review enrolled typical T1D and insulin‐dependent LADA patients diagnosed at age ≥30 years, and who were followed up at Siriraj Hospital, Thailand, during 2014–August 2024. Potential patients were searched from the Type 1 Diabetes Registry of Siriraj Diabetes Center and electronic medical record.

Results

Among 84 patients, 32 were typical T1D and 52 LADA. At diagnosis, both groups had normal body mass index (20.9 ± 4.5 kg/m²), significant weight loss (13.3 ± 6.4%), short duration of hyperglycemic symptoms (1 [1–2] months), and high fasting blood glucose (259 [174–345] mg/dL). Diabetic ketoacidosis (DKA) was the first presentation in 76.7% of typical T1D and in 2.6% of LADA. Anti‐glutamic acid decarboxylase (anti‐GAD) was tested in 91.7% of patients (positive: T1D 72%, LADA 100%). Almost 70% of patients with LADA were initially misdiagnosed as T2D and were ultimately diagnosed at 7.6 [2.6–17.5] years later. Patients with LADA had a higher incidence of recurrent DKA. At latest follow‐up, patients with LADA had significantly higher systolic blood pressure (SBP), glycated hemoglobin (HbA_1c), and insulin resistance compared to typical T1D.

Conclusions

We found a higher proportion of LADA than typical T1D, and 70% of LADA were initially misdiagnosed as T2D. Both groups were characterized by a normal BMI, significant weight loss, and a short duration of hyperglycemic symptoms. However, LADA rarely presented with DKA and had higher HbA_1c and insulin resistance during follow‐up.

Abbreviations

ADA

American Diabetes Association

anti‐GAD

anti‐glutamic acid decarboxylase

anti‐IA2

anti‐insulinoma‐associated antigen 2

anti‐ZnT8

anti‐zinc transporter 8

BMI

body mass index

DBP

diastolic blood pressure

DCCT

Diabetes Control and Complications Trial

DKA

diabetic ketoacidosis

DM

diabetes mellitus

EASD

European Association for the Study of Diabetes

eGDR

estimate glucose disposal rate

eGFR

estimated glomerular filtration rate

ELISA

enzyme‐linked immunosorbent assay

HbA_1c

glycated hemoglobin

HDL‐cholesterol

high‐density lipoprotein cholesterol

HOMA‐IR

homeostatic model assessment of insulin resistance

HT

hypertension

ICD‐10

international statistical classification of diseases and related health problems 10th revision

LADA

insulin‐dependent latent autoimmune diabetes in adults

LDL‐cholesterol

low‐density lipoprotein cholesterol

NIRAD

non‐insulin requiring autoimmune diabetes

SBP

systolic blood pressure

SGLT2

sodium‐glucose co‐transporter 2

SIRB

Siriraj Institutional Review Board

SPIDDM

slowly progressive insulin dependent diabetes mellitus

SPSS

Statistical Package for the Social Sciences

T1D

type 1 diabetes

T2D

type 2 diabetes

typical T1D

typical adult‐onset type 1 diabetes

WHO

World Health Organization

INTRODUCTION

Type 1 diabetes (T1D) is caused by autoimmune beta cell destruction that normally results in absolute insulin deficiency and required insulin therapy ¹ . T1D has long been viewed as a condition that is mainly diagnosed during childhood or adolescence; however, recent epidemiological data indicate a higher incidence of adult‐onset T1D compared to childhood‐onset T1D in many parts of the world ² , ³ , ⁴ , ⁵ . T1D is a heterogeneous disease with a diverse palette and gradient of risk factors (e.g., genetic risks, environmental exposures, and immune features), disease progression, and metabolic outcomes ⁶ . Adult‐onset T1D tends to have a lower genetic load, different pattern of islet autoantibodies, and slower disease progression compared to childhood‐onset T1D ⁷ . Adult‐onset type 1 diabetes (T1D) can present either as a rapid‐onset classical form or as a slowly progressive form, commonly referred to as latent autoimmune diabetes in adults (LADA) ⁸ , ⁹ .

LADA is generally diagnosed in people with diabetes onset after age 30 years, presence of islet autoantibodies, and absence of insulin requirement for at least 6 months after diagnosis ¹⁰ . In 2019, the World Health Organization (WHO) reclassified diabetes into six types, replacing the term “LADA” with “slowly evolving immune‐mediated diabetes,” categorized as a hybrid form of diabetes ¹¹ . However, the Standard of Care in Diabetes 2026 published by the American Diabetes Association (ADA) accepts and includes the term LADA as a form of T1D due to its practical value in raising awareness of slowly progressive beta‐cell destruction in adults ¹ . In Japan, T1D is classified into three subtypes including acute‐onset T1D, slowly progressive insulin dependent diabetes mellitus (SPIDDM), and fulminant T1D ⁹ . In 2023, the diagnostic criteria of SPIDDM were updated. Definite SPIDDM is defined by three criteria: (1) presence of islet‐associated autoantibodies at some point in time during the disease course; (2) absence of ketosis or ketoacidosis at the diagnosis of diabetes with no requirement for insulin treatment immediately after diagnosis in principle; and (3) gradual decrease of insulin secretion over time, with insulin treatment required at more than 3 months after diagnosis, and the presence of severe endogenous insulin deficiency ¹² . The definition of LADA and definite SPIDDM overlap substantially. LADA is characterized by adult‐onset diabetes, positive islet autoantibodies, and initial preserved insulin secretion, with variable progression to insulin dependence. In contrast, definite SPIDDM can occur at any age and is characterized by a gradual decline in insulin secretion leading to progressive insulin dependence.

Over 40% of individuals diagnosed with T1D after the age of 30 years are initially misclassified as having type 2 diabetes. This misclassification can result in prolonged hyperglycemia and glucose toxicity due to delayed initiation of insulin therapy or the use of inappropriate insulin regimens, thereby increasing the risk of diabetic ketoacidosis and the development of chronic diabetes‐related complications.

To address the scarcity of current data specific to adult‐onset T1D in Thailand, the aim of this study was to investigate the prevalence, clinical and laboratory characteristics, glycemic management, diabetic complications, and disease progression of typical adult‐onset T1D and insulin‐dependent LADA in Thailand.

METHODS

Study setting, patients, and ethical approval

This retrospective study was conducted at the Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand. Siriraj Hospital is a 2,300‐bed national tertiary referral center. Potential patients were searched from the Type 1 Diabetes Registry of the Faculty of Medicine Siriraj Hospital and electronic medical record. Patients satisfying all of the following criteria were considered eligible for enrollment in this study:

(1) diagnosed with diabetes at age 30 years or older; (2) was receiving insulin therapy; (3) having an International Statistical Classification of Diseases and Related Health Problems 10th Revision (ICD‐10) disease code of T1D (E10) or having at least one islet‐associated autoantibody positive, including anti‐glutamic acid decarboxylase (anti‐GAD), anti‐insulinoma‐associated antigen 2 (anti‐IA2), or anti‐zinc transporter 8 (anti‐ZnT8) or having C‐peptide <0.6 ng/dL; and (4) was being followed‐up at Siriraj Hospital during 2014–August 2024. The exclusion criteria were as follows: (1) positive islet‐associated autoantibodies without a requirement for insulin therapy; (2) absence of available medical data from the time of diabetes diagnosis; and (3) non‐Thai ethnicity. The study was conducted in accordance with the principles of the Declaration of Helsinki and was approved by the Siriraj Institutional Review Board (SIRB) (certificate of approval no. Si 737/2024). Written informed consent to participate was not obtained from study subjects due to the retrospective design of our study.

Definitions

Typical adult‐onset T1D was defined as (1) typical presentations of T1D, such as diabetic ketoacidosis (DKA), short duration of hyperglycemic symptoms; (2) one or more islet‐associated autoantibodies positive or C‐peptide <0.6 ng/dL; and (3) requiring insulin therapy within 6 months of diabetes diagnosis. Insulin‐dependent latent autoimmune diabetes in adults (LADA) ¹⁰ was defined as (1) diabetes diagnosed at age 30 years or older; (2) one or more islet‐associated autoantibodies positive; and (3) requiring insulin therapy after 6 months of diabetes diagnosis. Clinical characteristics, laboratory findings, diabetic complications, and estimated glucose disposal rate (eGDR) were compared between the typical adult‐onset T1D and insulin dependent LADA groups.

Data collection

Patient data were collected from the Siriraj Hospital patient medical database via the assistance of the Siriraj Informatics and Data Innovation Center. Patient demographic and clinical data, laboratory characteristics, glycemic management, diabetic complications, and disease progression data were collected at both the initial diagnosis of diabetes (Table 1) and at the last follow‐up at our center (Table 2). Since most patients enrolled in this study were referred from other centers where the diabetes diagnosis was first made, the patient data in many instances were discovered to be incomplete. In contrast, all data from the last follow‐up at our center were complete.

Table 1.

Baseline patient demographic and clinical characteristics at initial diabetes diagnosis for all patients, and compared between the typical adult‐onset T1D and insulin‐dependent LADA groups

Characteristics	All patients (n = 84)	Typical adult‐onset T1D (n = 32)	Insulin‐dependent LADA (n = 52)	P‐value*
Mean age at DM diagnosis (years)	45.4 ± 13.2	44.0 ± 14.9	46.2 ± 12.1	0.486
Age at DM diagnosis (years)				0.074
30.0–39.9	36 (42.9%)	19 (59.4%)	17 (32.7%)
40.0–49.9	19 (22.6%)	4 (12.5%)	15 (28.8%)
50.0–59.9	16 (19.0%)	5 (15.6%)	11 (21.2%)
60.0–69.9	8 (9.5%)	1 (3.1%)	7 (13.5%)
≥70	5 (6.0%)	3 (9.3%)	2 (3.8%)
Male sex	28 (33.3%)	9 (28.1%)	19 (36.5%)	0.631
Family history of DM	40 (71.4%) (n = 56)	18 (72.0%) (n = 25)	22 (71.0%) (n = 31)	0.216
Underlying disease
Hypertension	18 (24.3%)	7 (22.6%)	11 (25.6%)	0.143
Dyslipidemia	28 (37.8%)	7 (22.6%)	21 (48.8%)	0.010
Coronary artery disease	1 (1.4%)	0 (0.0%)	1 (2.3%)	0.103
Cerebrovascular disease	2 (2.7%)	0 (0.0%)	2 (4.7%)	0.070
Autoimmune disease	13 (17.6%)	6 (19.4%)	7 (16.3%)	0.141
Graves' disease	8 (10.8%)	4 (12.9%)	4 (9.3%)	0.132
Hashimoto thyroiditis	3 (4.1%)	1 (3.2%)	2 (4.7%)	0.143
Vitiligo	3 (4.1%)	0 (0.0%)	3 (7.0%)	0.047
Other autoimmune disease	2 (2.7%) (n = 74)	2 (6.5%) (n = 31)	0 (0.0%) (n = 43)	0.034
Hyperglycemic symptoms	44 (81.5%) (n = 54)	15 (71.4%) (n = 21)	29 (87.9%) (n = 33)	0.307
Duration of hyperglycemic symptoms (months)	1 [1–2] (n = 27)	1 [0.3–2.3] (n = 14)	1 [1–3] (n = 13)	0.294
Diabetic ketoacidosis at initial diagnosis	24 (35.3%) (n = 68)	23 (76.7%) (n = 30)	1 (2.6%) (n = 38)	<0.001
Percent weight loss (%)	13.3 ± 6.4 (n = 33)	13.0 ± 5.4 (n = 14)	13.6 ± 7.1 (n = 19)	0.784
Body weight (kg)	54.6 ± 13.3 (n = 38)	53.9 ± 15.7 (n = 16)	55.0 ± 11.7 (n = 22)	0.803
Body mass index (kg/m2)	20.9 ± 4.5 (n = 38)	21.3 ± 5.6 (n = 16)	20.6 ± 3.7 (n = 22)	0.674
Random blood glucose (mg/dL)	542 [369–817] (n = 19)	615 [512–850] (n = 13)	320 [262–593] (n = 6)	0.029
Fasting plasma glucose (mg/dL)	259 [174–345] (n = 34)	325 [216–369] (n = 15)	246 [150–322] (n = 19)	0.145
HbA1c (%)	10.2 [7.7–14.0] (n = 41)	9.8 [6.7–12.9] (n = 22)	11.1 [7.7–14.7] (n = 19)	0.170
Systolic blood pressure (mmHg)	120 ± 14 (n = 25)	121 ± 17 (n = 10)	121 ± 13 (n = 15)	0.978
Diastolic blood pressure (mmHg)	73 ± 13 (n = 25)	67 ± 15 (n = 10)	77 ± 10 (n = 15)	0.049
DM classification at diagnosis				<0.001
T1D	18 (21.4%)	18 (56.2%)	0 (0.0%)
T2D	42 (50.0%)	6 (18.8%)	36 (69.2%)
Unclassified	24 (28.6%)	8 (25.0%)	16 (30.8%)
Insulin therapy at first diagnosis	27 (48.2%) (n = 56)	23 (92.0%) (n = 25)	4 (12.9%) (n = 31)	<0.001
Time to insulin initiation after DM diagnosis (years)	1.5 [0.5–4.0] (n = 37)	0.2 [0.1–0.4] (n = 4)	1.9 [0.6–4.7] (n = 33)	0.006

Data presented as number and (percentage), mean ± standard deviation, or median and [interquartile range]. A P‐value <0.05 indicates statistical significance (chi‐square test, Fisher's exact test, Student's t‐test, or Mann–Whitney U test).

^*Represents a comparison between the T1D and LADA groups.

Table 2.

Patient demographic and clinical characteristics at the latest follow‐up at Siriraj Hospital for all patients, and compared between the typical adult‐onset T1D and insulin‐dependent LADA groups

Characteristics	All patients (n = 84)	Typical adult‐onset T1D (n = 32)	Insulin‐dependent LADA (n = 52)	P‐value*
Current age (years)	58.4 ± 13.5	56.7 ± 14.6	60.9 ± 12.7	0.170
Diabetes duration (years)	13.7 [5.4–18.6]	9.7 [3.8,17.8]	13.7 [6.8,19.0]	0.242
Body weight (kg)	58.9 ± 10.6	58.8 ± 13.0	59.0 ± 9.1	0.917
Body mass index (kg/m2)	22.7 ± 3.2	22.6 ± 3.9	22.7 ± 2.7	0.851
Fasting plasma glucose (mg/dL)	166 [79–100]	147 [98–193]	157 [100–217]	0.996
HbA1c (%)	8.3 [7.2–9.4]	7.8 [6.7–9.1]	8.5 [7.4–9.6]	0.022
HbA1c <7%	16 (19%)	11 (34.4%)	5 (9.6%)	0.005
HbA1c <7.5%	26 (31%)	13 (40.6%)	13 (25%)	0.135
Insulin regimen				0.155
Basal bolus regimen	62 (73.8%)	26 (81.2%)	36 (69.2%)
Basal insulin	4 (4.8%)	0 (0.0%)	4 (7.7%)
Other regimen	18 (21.4%)	6 (18.8%)	12 (23.1%)
Total Daily Insulin Dose (unit/kg/day)	0.7 ± 0.3	0.8 ± 0.3	0.7 ± 0.3	0.219
Oral antihyperglycemic agents	17 (20.2%)	4 (12.5%)	13 (25.0%)	0.134
SBP (mmHg)	131 ± 19	125 ± 17	135 ± 19	0.030
DBP (mmHg)	72 ± 12	69 ± 12	74 ± 12	0.085
BP <130/80 mmHg	33 (40.7%)	17 (54.8%)	16 (32%)	0.040
Hypertensive therapy	35 (41.7%)	8 (25.0%)	27 (51.9%)	0.022
Cholesterol (mg/dL)	164 [139–183]	169 [155–193]	155 [132–174]	0.027
Triglyceride (mg/dL)	71 [54–81]	62 [47–92]	77 [54–103]	0.200
HDL‐cholesterol (mg/dL)	64 [52–81]	73 [54–88]	62 [51–72]	0.055
LDL‐cholesterol (mg/dL)	78 [63–96]	84 [67–101]	75 [59–94]	0.156
LDL‐C <100 mg/dL	63 (75.9%)	21 (65.6%)	42 (80.8%)	0.085
Lipid‐lowering therapy	71 (84.5%)	23 (71.9%)	48 (92.3%)	0.026
Statin therapy	69 (82.1%)	23 (71.9%)	46 (88.5%)	0.078
eGFR (mL/min/1.73m2)	89 [71–103]	94 [74–107]	87 [71–103]	0.287
eGDR (mg/kg/min)	7.8 [6.1–9.3]	8.9 [7.0–9.8]	7.4 [5.7–8.6]	0.011
Diabetic retinopathy a	12 (14.3%)	5 (19.2%)	7 (17.9%)	1.000
Diabetic nephropathy b	16 (19.0%)	6 (22.2%)	10 (21.7%)	1.000
Diabetic neuropathy c	5 (6.0%)	3 (12.5%)	2 (6.9%)	0.649
Coronary artery disease	9 (10.7%)	3 (9.4%)	6 (11.5%)	1.000
Cerebrovascular disease	7 (8.3%)	2 (6.3%)	5 (9.6%)	0.704

Data presented as number and (percentage), mean ± standard deviation, or median and [interquartile range]. A P‐value <0.05 indicates statistical significance (chi‐square test, Fisher's exact test, Students t‐test, or Mann–Whitney U test).

^*Represents a comparison between the T1D and LADA groups.

^a At least moderate non‐proliferative diabetic retinopathy.

^b Urine protein creatinine ratio >30 mg/gCr.

^c Abnormal monofilament.

Islet autoantibodies and eGDR data collection

Autoantibodies to GAD, IA2, and ZnT8 have been measured at our center using the enzyme‐linked immunosorbent assay (ELISA) method (RSR, Cardiff, UK) since 2018. The sensitivities of the GAD, IA2, and ZnT8 assays were 92%, 76%, and 72%, respectively, while the specificities were 98%, 98%, and 99%, respectively. The manufacturer's cutoff values for determining positivity had been established as follows: ≥5 U/mL for GAD; ≥7.5 U/mL for IA2; and ≥15 U/mL for ZnT8. Data specific to GAD level measurements measured by radioimmunoassay prior to 2018 were excluded from quantitative analysis. There was no uniform protocol for islet autoantibodies testing in our institution. Typically, anti‐GAD was measured as the primary antibody, either concurrently with anti‐IA2 or followed by sequential anti‐IA2 testing. anti‐ZnT8 testing was only available at our institution in 2018, resulting in a limited number of anti‐ZnT8 measurements in this study. Estimated glucose disposal rate (eGDR), ¹³ which is a laboratory parameter used to measure insulin resistance, was calculated using the following formula eGDR = 19.02–(0.22*BMI[kg/m ² ])−(3.26*HT)−(0.61*HbA _1c [%]), whereby HT is hypertension, and the entry for that variable is either 1 (presence of HT) or 0 (no presence of HT).

Statistical analysis

All statistical analyses were performed using the SPSS Statistics version 18.0 software (SPSS, Inc.; IBM Corp., Armonk, NY, USA). Baseline patient demographic and clinical characteristics were compared between the typical adult‐onset T1D and insulin dependent LADA groups, and patient demographic and clinical characteristics at the last follow‐up were also compared between the same two groups. Data are shown as number and percentage for categorical data, as mean plus/minus (±) standard deviation (SD) for normally distributed continuous data, and as median and interquartile range (IQR) for non‐normally distributed continuous data. Comparisons of categorical data were performed using chi‐square test or Fisher's exact test depending on the size of the sample being tested. Comparisons of continuous data were performed using Student's t‐test and Mann–Whitney U test for normally and non‐normally distributed continuous data, respectively. A P‐value of less than 0.05 was considered statistically significant for all tests.

RESULTS

A total of 501 patients were reviewed; 84 patients were eligible for this study (Study flow diagram shown in Figure 1). Baseline patient demographic and clinical characteristics at initial diabetes diagnosis for all patients, and compared between the typical adult‐onset T1D and insulin‐dependent LADA groups are shown in Table 1.

Figure 1.

Study flow diagram.

Baseline characteristics

There were 84 patients with T1D diagnosed after the age 30 years who were retrospectively enrolled in this study. Of those, 32 patients had typical, rapid‐onset presentation of adult‐onset T1D, whereas the other 52 patients had a clinical presentation consistent with insulin‐dependent LADA. The mean age at diagnosis did not differ significantly between groups (mean age: 45.4 years). Most patients with typical adult‐onset T1D were diagnosed during their 30th to 39th year of age. After age 40, the prevalence of insulin‐dependent LADA was higher than that of typical adult‐onset T1D; however, both types of diabetes were observed across all age groups (Figure 2). At diagnosis, both groups had a normal mean body mass index (BMI) of 20.9 ± 4.5 kg/m², a substantial mean percentage of body weight loss (13.3 ± 6.4%), a high median HbA_1c (10.2 [IQR: 7.7–14.0]%), and a short median duration of hyperglycemic symptoms (1 [IQR: 1–2] months).

Figure 2.

Distribution of age at diagnosis of typical adult‐onset type 1 diabetes (T1D) or insulin‐dependent latent autoimmune diabetes in adults (LADA) (N = 84).

Clinical presentations

Compared with insulin‐dependent LADA, typical adult‐onset T1D had a significantly higher prevalence of DKA, and a significantly higher random blood glucose level at first presentation. Among adult‐onset T1D, 50% of patients (70% insulin‐dependent LADA and 19% typical adult‐onset T1D) were initially diagnosed with T2D. Insulin treatment was initiated at a median 1.9 [IQR: 0.6–4.7] years after diagnosis in the insulin‐dependent LADA group, and LADA was ultimately diagnosed at 7.6 [IQR: 2.6–17.5] years after the initial diabetes diagnosis. Patients with insulin‐dependent LADA experienced more episodes of recurrent DKA (Figure 3). Specifically, 21.2% (11/52) of insulin‐dependent LADA patients had at least two DKA events (excluding the initial presentation). Among those with at least two DKA events, 64.9% (24/37) of episodes occurred before the diagnosis of LADA. Data specific to the distribution of DKA events can be found in Figure 3.

Figure 3.

The percentage of patients stratified by the number of diabetic ketoacidosis (DKA) episodes from 1 to ≥5 (excluding first presentation) and compared between those diagnosed with typical adult‐onset diabetes (T1D) and those diagnosed with insulin‐dependent latent autoimmune diabetes in adults (LADA).

Islet autoantibodies and C‐peptide

In the typical adult‐onset T1D group, the proportion of patients tested for anti‐GAD, anti‐IA2, and anti‐ZnT8 were 78.0%, 37.5% and 21.9%, respectively, compared with 100%, 42.3%, and 5.8% in the insulin‐dependent LADA group. Among those tested, anti‐GAD was positive in 72% (18/25) of patients, anti‐IA2 in 16.7% (2/12), and anti‐ZnT8 in 14.3% (1/7) in the typical adult‐onset T1D group. In the insulin‐dependent LADA group, anti‐GAD was positive in 100% (52/52), anti‐IA2 in 27.3% (6/22), and anti‐ZnT8 in 33.3% (1/3). The median time from diabetes diagnosis to autoantibody measurement was 0.15 [IQR: 0.01–4.52] years in typical adult‐onset T1D, and 6.57 [IQR: 1.65–16.87] years in insulin‐dependent LADA. The level of islet autoantibodies using ELISA method including both positive and negative results were reported in supplementary S1. No bimodal distribution of anti‐GAD levels was observed among insulin‐dependent patients with LADA, with a median (IQR) level of 329 (89–1,295) U/mL (Appendix S1). The median C‐peptide level was 0.02 ng/mL (IQR, 0.01–0.40) in the typical adult‐onset T1D group (n = 21) and 0.17 ng/mL (IQR, 0.02–0.60) in the insulin‐dependent LADA group (n = 39), with no significant difference between the groups (P = 0.071). However, the timing of C‐peptide measurement differed substantially between the groups. C‐peptide was measured shortly after diabetes diagnosis in the typical adult‐onset T1D group, at a median of 0.50 (IQR, 0.10–7.2) years, compared with 8.9 (IQR, 3.7–17.0) years after diagnosis in the insulin‐dependent LADA group.

Clinical presentations during follow‐up periods

Patient demographic and clinical characteristics at the latest follow‐up at Siriraj Hospital for all patients, and compared between the typical adult‐onset T1D and insulin‐dependent LADA groups are shown in Table 2.

At the most recent follow‐up, patients with insulin‐dependent LADA were older (60.9 versus [vs.] 56.7 years, respectively) and had a slightly longer duration of diabetes (13.7 vs. 9.7 years, respectively) than typical adult‐onset T1D (both P > 0.05). Both groups maintained a normal body mass index. However, the insulin‐dependent LADA group had a significantly higher glycated hemoglobin (HbA_1c) level, systolic blood pressure (SBP), and lower estimated glucose disposal rate (eGDR) (indicating higher insulin resistance). Furthermore, the insulin‐dependent LADA group achieved glycemic, blood pressure, and cholesterol targets less than the adult‐onset T1D group. The prevalence of microvascular and macrovascular complications was found not to be statistically significantly different between groups.

DISCUSSION

Our study reported the demographic and clinical presentations, laboratory features, glycemic management, diabetes complications, and disease progression of typical adult‐onset T1D and insulin‐dependent LADA in Thailand. In our study, 42.9% of adult‐onset T1D presented during the age of 30 to 39 years, and there was comparable prevalence between typical adult‐onset T1D and insulin‐dependent LADA. However, after the age of 40, insulin‐dependent LADA was more common than typical adult‐onset T1D, which is consistent with reports from Europe ¹⁴ and Japan ⁹ .

Apart from the higher prevalence of DKA at first presentation in typical adult‐onset T1D (76.7%), other clinical presentations of typical adult‐onset T1D and insulin‐dependent LADA were comparable. Both groups had rapid unintentional weight loss (13.3%), normal BMI (20.9 kg/m²), and high random blood glucose (596 mg/dL). These presentations are consistent with the most discriminative features of T1D suggested by the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD) 2021, including lower BMI (<25 kg/m²), unintentional weight loss, ketoacidosis, and glucose >360 mg/dL ¹⁵ . Although some distinguishing features of T1D were observed in both groups in the present study, 50% of adult‐onset T1D (70% of insulin‐dependent LADA and 19% of typical adult‐onset T1D) were initially diagnosed with T2D. LADA was ultimately diagnosed at a median 7.6 years later. The issue of misclassification has been previously documented in studies from England (47% misclassification) ¹⁶ and the United States (46–55% misclassification) ¹⁷ . This misclassification might reflect a lack of awareness that the onset of T1D can occur in adulthood, the overwhelming prevalence of T2D patients in adulthood, and limited access to T1D diagnostic biomarkers ² . Delayed initiation of insulin therapy in individuals with LADA can result in prolonged hyperglycemia during the early stage of the disease and may contribute to an increased risk of chronic diabetes‐related complications through the phenomenon of “metabolic memory,” as demonstrated in the Diabetes Control and Complications Trial (DCCT) ¹⁸ . An accurate T1D diagnosis confers impact beyond insulin treatment alone. Insulin regimen, the content of diabetes self‐management education, and access to glucose monitoring and newer diabetes technologies heavily depend on accurate diagnosis ¹⁵ . In Thailand, the Universal Health Coverage scheme allows for reimbursement of the costs of continuous glucose monitoring only in patients with T1D or pregnant women with pre‐existing diabetes mellitus. One potential approach to reducing diabetes misclassification is to temporarily use the term “unclassified diabetes”, as recommended by the WHO classification, ¹¹ while further investigations are being conducted to establish the appropriate diabetes subtype.

Interestingly, there were more episodes of recurrent DKA in the insulin‐dependent LADA group than in the typical adult‐onset T1D in our study, and the majority of events occurred before LADA diagnosis. A case series from Malaysia also reported recurrent DKA episodes that led to diagnoses of LADA ¹⁹ . Recurrent DKA in patients with LADA might be explained by interruption of insulin therapy, initiation of sodium‐glucose co‐transporter 2 (SGLT2) inhibitor for glycemic control or organ protection, and/or a lack of knowledge about sick day management.

Anti‐GAD is the most common and the most sensitive marker in adult‐onset T1D ¹⁴ , ²⁰ , ²¹ . Anti‐GAD was positive in 72% of typical adult‐onset T1D, and 100% of insulin‐dependent LADA in our study. The TrialNet Pathway to Prevention study screened autoantibodies of relatives of patients with T1D followed by monitoring for disease progression. They found that adults with a single autoantibody positive had a higher likelihood of progressing to stage 3 T1D compared to children with a single autoantibody positive (40 vs. 15%, respectively), and the majority (90.6%) of single autoantibody positive in adults was anti‐GAD. However, at clinical diagnosis, the percentage of single and multiple autoantibodies positive in adults was 40.2% and 58%, respectively ²⁰ . In contrast, our study had no data specific to single or multiple autoantibodies since anti‐IA2 and anti‐ZnT8 were not measured in every patient. Bimodal distribution of anti‐GAD level in LADA was highlighted in previous studies, such as the Non‐Insulin Requiring Autoimmune Diabetes (NIRAD) study, ²² the Action LADA 7 study, ¹⁴ and the LADA China study ²³ . Patients with a high GAD level (higher than 200–300 WHO units) usually had more features of autoimmunity, including lower BMI, lower prevalence of metabolic syndrome, and higher HbA_1c compared to those with a low GAD level. However, there was no bimodal distribution of anti‐GAD level in our study. This might be explained by differences in study populations among studies. For instance, our study included only patients with insulin‐dependent LADA, whereas previous studies included only insulin‐independent LADA ²² or a blend of insulin‐dependent and insulin‐independent LADA ¹⁴ , ²³ . Therefore, patients with a low anti‐GAD level may have been excluded from our study.

At the latest follow‐up at our center, we found a higher HbA_1c level in insulin‐dependent LADA compared to typical adult‐onset T1D. Data specific to glycemic control among the subtypes of adult‐onset T1D in Thailand are limited. We propose the causes of higher glycemic control in patients with insulin‐dependent LADA in our study to be (1) insulin resistance as evidenced by low eGDR, and (2) a lower proportion of intensive insulin regimen usage. The prevalence of metabolic syndrome in patients with LADA was reported to be higher than in patients with T1D, but lower than in patients with T2D ²⁴ . In China, Li et al. found that patients with LADA had a higher Homeostatic Model Assessment of Insulin Resistance (HOMA‐IR) level than patients with T1D, but a lower HOMA‐IR level than patients with T2D ²⁵ . In the United States, Chiu, et al. found that patients with LADA had a HOMA‐IR level comparable to that of patients with T2D ²⁶ . Those two studies demonstrated insulin resistance in patients with LADA who were predominantly not treated with insulin, whereas our study demonstrates the presence of insulin resistance in patients who were receiving insulin therapy. Despite residual beta‐cell function in the early stages of LADA, the disease gradually progresses to insulin dependence and the need for insulin therapy. A basal‐bolus insulin regimen is recommended for patients with T1D because it more closely mimics physiological insulin secretion ¹⁰ . Therefore, the lower proportion of intensive insulin regimen usage observed in the insulin‐dependent LADA group may have influenced the higher HbA_1c level in this group.

There are several practical approaches that may help identify patients with LADA in Thailand. First, islet autoantibody testing, particularly anti‐GAD, should be considered in patients presenting with clinical features suggestive of T1D, including a BMI <25 kg/m², unintentional weight loss, ketoacidosis, or marked hyperglycemia (blood glucose >360 mg/dL), irrespective of age at diabetes onset. Islet autoantibody testing should also be considered in patients who require insulin early after diagnosis or who develop recurrent episodes of DKA.

This study also has some mentionable limitations. First, its retrospective design introduces potential selection bias, incomplete data, and an inability to establish causality. Second, because our facility is a national tertiary referral center managing highly complex cases, the findings may not be fully generalizable to other care settings. Third, our study had a relatively small sample size, and many patients were referred from outside hospitals, resulting in missing data on some initial clinical presentations, such as percentage of weight loss and hyperglycemic symptoms at diabetes diagnosis. These missing data and small sample size may have reduced our statistical power to detect significant differences between groups. Fourth, HOMA‐IR and waist circumference data were not available in our study. Therefore, insulin resistance could not be defined using these parameters. Finally, not all patients with T2D who had discriminative features of T1D were screened for islet autoantibodies. Thus, some patients with type 1 diabetes or LADA may not have been included in this study. However, and importantly, these data represent the best available information on the clinical presentation of adult‐onset T1D in Thailand. Having acknowledged these potential shortcomings, further study is needed to shore up our findings and to further develop our knowledge of adult‐onset T1D in Thailand.

CONCLUSIONS

The results of this study revealed important features of typical adult‐onset T1D and insulin‐dependent LADA in Thailand. Among those, the proportion of LADA is higher than the proportion of typical adult‐onset T1D; 70% of LADA patients were initially misdiagnosed as T2D, resulting in a diagnostic delay of 7.6 years. Both adult‐onset T1D and insulin‐dependent LADA were characterized by a normal body mass index, significant weight loss, and a short duration of hyperglycemic symptoms. However, insulin‐dependent LADA rarely presented with DKA. During follow‐up, the LADA group had higher HbA1c and blood pressure levels compared to the adult‐onset T1D group.

DISCLOSURE

The authors declare no conflict of interest.

Approval of the research protocol: Approved by the Siriraj Institutional Review Board (SIRB) (COA no. Si 737/2024).

Informed consent: Was waived due to the retrospective study.

Registry and the registration no. of the study/trial: N/A.

Animal studies: N/A.

Supporting information

Appendix S1. Figure Anti‐GAD level distribution (ELISA method) among patients with insulin‐dependent latent autoimmune diabetes in adult. Table Islet autoantibodies level using ELISA method in patients with typical adult‐onset T1D and insulin‐dependent LADA including only positive result.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.