The relationship between GAD65 autoantibody and the risk of T1DM onset

Abstract

Objectives

Type 1 diabetes mellitus (T1DM) is a well-known autoimmune disease, characterized by β-cell destruction in pancreas islet cells, which results insulin deficiency and subsequent hyperglycemic sequelae. While there is screening for type 2 DM that leads to better glycemic control and outcome, the majority of T1DM patients are diagnosed when much of the pancreatic cells and their function are disturbed. The aim of this article is to present an overview of the effective factors in the positivity of Glutamic acid decarboxylase antibody )GADA( and identifying the high-risk individuals for T1DM.

Methods

We searched English literature available at National Library of Medicine via PubMed, and Google Scholar through December 2020. Finally, 79 papers have been included in the study. Studies were summarized based on the number of positive autoantibodies and onset of T1DM over time and GADA correlation with different variables.

Conclusions

GADA is an easy marker to measure that can be detected many months prior to the clinical presentation and remains positive even after early childhood.

Introduction

The prevalence of diabetes mellitus (DM) has quadrupled over a span of 30 years from 1980 to 2014 with more than 400 million patients worldwide [1]. Type 1 diabetes mellitus (T1DM) comprises about 5% of all DM cases and is caused by autoimmune destruction of pancreatic β-cells [2]. While there is screening for type 2 DM (T2DM) that leads to better glycemic control and outcome, T1DM is usually diagnosed when much of the pancreatic cells and their function are disturbed. How autoantibodies contribute to T1DM pathogenesis has not been well-understood yet; however, they have remained as mainstay for diagnosis. Among all autoantibodies that have been studied, Glutamic acid decarboxylase (GAD) antibody (GADA) is relatively easy to measure in normal clinical settings and is considered as the first line tool in assessment of T1DM [3]. The aim of this article is to review the literature and evidence for the effective factors in the positivity of GADA and identifying the high-risk individuals for T1DM and monitoring them.

Methodology

We searched English literature available at National Library of Medicine via PubMed, and Google Scholar through December 2020 with the following keywords: “type 1 diabetes mellitus”, “T1DM”, “antibodies to islet cells”, “islet cell antibodies”, “Glutamic acid decarboxylase” and “Glutamic acid decarboxylase antibody”. We identified 3821 papers; 2453 papers were screened after removing duplicates. Articles were peer-reviewed by title; 1164 irrelevant articles were excluded. Peer-reviewed articles then were reviewed through their abstracts and they were selected based on their methodology, novelty, and relevance to the aim of the review. We assessed 142 papers by the full-text review. Finally, 79 papers have been included in the study. Studies were summarized based on the number of positive autoantibodies and onset of T1DM over time and GADA correlation with different variables. Cost of laboratory assays were retrieved from Centers for Medicare and Medicaid Services (CMS) website.

Type 1 diabetes Mellitus background

Epidemiology

The incidence of T1DM among children (age < 19) has increased significantly in recent years. The adjusted annual relative incidence of T1DM has been increased 1.8% during the 2002–2012 period [4]. According to crude estimates for 2018 in the U.S. population, 187,000 children and adolescents younger than age 20 had T1DM [5] and about 90,000 more are diagnosed annually [6]. Global prevalence is estimated to be and over 1 million for children and adolescents [7].

Diagnosis, incidence and prevalence

T1DM have 3 stages; the stage 1 is defined by children with normoglycaemic and asymptomatic that have two or more positive islet autoantibodies. In the stage 2, blood glucose, especially after food, rises, but there are no symptoms. In the stage 3, there are symptoms with a rise in fasting glucose; in this stage, the patient with T1DM needs insulin therapy [8].

Utilizing glycemic indices such as Glycosylated Hemoglobin A1C (HbA1C), fructose amine, 1,5-anhydroglucitol and glycated albumin, clinicians are able to monitor average blood sugar and hence guide the treatment plan. They also can serve as an adjunct to fasting blood sugar for diagnosis of T2DM. Nevertheless, glycemic indices usefulness fades, when it comes to direct information about β-cell destruction as it happens in T1DM. Indeed, relying on fasting glucose, HbA1C, body mass index (BMI) and age at onset of diabetes can lead to misclassification of patients with type 1 and 2 diabetes [9]. Also, it leads to late diagnosis, as more than 20% of children with T1DM develop Diabetic ketoacidosis as their initial presentation, which may be associated with morbidity and mortality [3].

Studies have shown that symptoms of DM would occur when β cell reserve become as low as 20–40% [10, 11], which is followed by persistent hyperglycemia and rise in glycemic indexes [12, 13]. Autopsies in patients who died of diabetic ketoacidosis with recent-onset T1DM showed a β cell mass of about 10% of normal [14]. Hence employing discrete tests that can divulge pancreatic β-cell function or vulnerability should be given precedence with the aim of halting disease progression or perhaps its prevention [15].

Autoantibodies as diagnostic markers for T1DM

Although humoral immune system is not considered as the main contributor for T1DM progression, autoantibodies serve as mainstay of diagnosis for T1DM [16]. In 1974, the first time, antibodies to islet cells (ICA) were discovered [17], followed by insulin autoantibodies (IAAs), GADA, insulinoma-associated protein 2 autoantibody (IA-2), and zinc transporter protein antibody (ZnT8A) [18].

GA, GAD and GADA

Glutamic acid is an acidic amino acid and is well-known for its role in metabolism and serving as an excitatory neurotransmitter. Its decarboxylation by GAD produces gamma amino butyric acid (GABA), an inhibitory neurotransmitter. There are two isoforms of GAD: GAD67 or GAD1, which is more prevalent in the central nervous system (CNS) and GABA-ergic neurons; and GAD65 or GAD2, which is more in pancreatic β-cells [19, 20]. GABA secretion increases in hyperglycemic states and exerts its effect as a paracrine inhibitory hormone on α-cells to suppress glucagon formation [21, 22]. Also, GABA is found in insulin granules and when secreted, it acts as a growth factor on β-cells, helping their survival, and also facilitate α-cell to β-cell conversion [23, 24].

GADA as a marker for T1DM

Prospective studies revealed that GADA is present in blood years prior to T1DM onset. This has been reported in BABYDIAB study in Germany [25], DAISY study in USA [26] and Finnish diabetes prediction and prevention in Finland [27]. In the Finnish study, autoantibody markers were detected in the serum of individuals as early as in the first year of life and on average 1.5 years before the clinical onset of T1DM. Hence, GADA is considered as the most commonly used test, to both screening and progression of T1DM [28].

GADA advantages over other autoantibodies

It is reported that exogenous insulin induces antibody production within 10–14 days after introduction. This can interfere and mask IAA and makes endogenous IAA detection difficult [18]. IA–2 A and ZnT8A autoantibodies disappear quickly after production and it is rare to find them as first autoantibodies [29, 30]. In contrast, GADA remains elevated persistently after T1DM pathogenesis initiates, which is useful for diagnosis. Takashi Murata et al. compared radioimmunoassay (RIA) with enzyme-linked immunosorbent assay (ELISA) [31]. They found that ELISA was better than RIA for detecting GADA and T1DM diagnosis. Sensitivity was 60.8% versus 57.0%, and specificity was 100.0% versus 97.5%. RIA and ELISA methods are compared in Table 1.

Table 1.

The comparison of GADA* measurement by RIA^† and ELISA^ǂ methods

GAA measurement method	RIA	ELISA
Sensitivity (%)	57	60.8
Specificity (%)	97.5	100
Positive predictive value (%)	95.7	100
Negative predictive value (%)	69.4	71.8

*Glutamic acid decarboxylase antibody

^† Radioimmunoassay

^ǂ Enzyme-linked immunosorbent assay

GADA prevalence, seroconversion and correlation with time

Up to 4% of normal population may express single autoantibody without T1DM, but less than 0.3% of normal population might express two or more autoantibodies. However, 74% of patients with T1DM have high GADA and seropositivity rate goes higher by checking more various autoantibodies. However, 2–5% of patients with T1DM will remain seronegative throughout the disease, even if four markers of GADA, IA-2 A, IAA, and ICA or ZnT8A are measured [32–35]. Nevertheless, positive GADA in an individual prompts subsequent autoantibody testing to evaluate T1DM [19]. Positive GADA level at 97.5th and 99.5th centiles is associated and has 6.6% and 23% 10-year risk of progression to insulin dependent DM [32]. A 27-year cohort study in Finland by Knip et al. showed that among children who were seropositive for GADA and IA-A2s, about 26% and 40% developed T1DM within median of 9 years, respectively [36]. It also revealed 33% seronegative conversion rate for GADA and 60% for IA-2As over a period of six years. Also, negative seroconversion over time is reported in the Asian population [37]. It is proposed that as β-cells destruction begins and continues, more GAD is released; hence GADA expands further. In later years as though β-cell mass declines, less GAD is released and GADA trends down [38, 39]. Also, higher titers of antibody can be found at older ages of diabetes onset; this is based on 30-year cohort study of Pittsburgh Epidemiology of Diabetes Complications [40]. This can be due to more robust immune response in older ages to release GAD and it is not necessarily contradicting with other reports that stated GADA declines over time.

Different studies have revealed positive correlation regarding number of different autoantibodies and titer of autoantibodies with T1DM. Table 2 summarizes associated risk with various autoantibody profile.

Table 2.

The number of positive autoantibodies in high-risk individuals and the percentage of onset of T1DM over time

Number of islet autoantibodies positive	Risk of onset T1D	over the course of time (year)	Study
One	15%	-	Mayr et al. [41]
One	< 25%	5	Leonard et al. [42]
One	6.6%	5	Bingley et al. [43]
One (GADA)	26%	27	Knip et al. [36]
One (IA-A2)	40%	27	Knip et al. [36]
Two	25–50%	5	Leonard et al. [42]
Two	11%	5	Katsarou et al. [29]
Two	25%	5	Greenbaum et al. [44]
Two	29%	5	Jacobsen, et al. [45]
Two	100%	27	Knip et al. [36]
Three	40%	5	Greenbaum et al. .[44]
Three	> 50%	5	Leonard et al. [42]
Three	48 to 86%	5	Achenbach et al. [46] Pietropaolo et al. [28]
Three	64 to 86%	10	Achenbach et al. [46] Pietropaolo et al. [28]
Three	36%	5	Katsarou, et al. [29]
Three	approaches 100%	5	Verge et al. [47]
Four	50%	5	Greenbaum et al. [44]
Four	47%	5	Katsarou et al. [29]
More than one	70%	less than 10	Ziegler et al. .[48]
Two or more	24.9%	3	Ziegler et al. [49]
Two or more	39%	5	Tosur et al. [50]
Two or more	50–100%	5 to 10	Knip [51]
Two or more	62%	5	Bingley et al. [52]
More than two	31%	5	Jacobsen et al. [45]

As summarized in Table 2, it is now the general consensus that there is a direct correlation with the increase in the number of autoantibodies and the risk for T1DM over time.

Age

Autoantibody presence sensitivity and power for predicting T1DM has inverse correlation with age [11]. TrialNet PTP cohort study in 1815 individuals with T1DM revealed that 5-year risk of developing T1DM in individuals with multiple autoantibodies, to be 35%, 22% and 15% for age groups < 12, ≥12 and > 18, respectively [45]. This is in congruent with Pittsburgh Epidemiology of Diabetes Complications study [40].

Gender difference

In some populations, the incidence of T1DM, in contrast to most autoimmune diseases, male to female ratio is higher [53]. Paradoxically, GADA prevalence and titers are reported to be higher in females [37, 54–56]. Nevertheless, there are studies that observed no significant gender difference for GADA, IA-2 and ZnT8 antibodies in T1DM individuals [40].

Geographic and Ethnicity

Reports on ethnicity and GADA prevalence are not cohesive. A multiethnic cohort study showed no overall difference in rates of autoantibody positivity (ICA, IAA, and GADA) and GADA levels between white and non-white ethnicity, at the time of presentation with diabetes in UK [57].

In a cohort multiethnic study which revealed association between islet cell autoantibody and poor glycemic control diabetes phenotype, regarding seropositivity no difference between white and non-white was noted [57]. However, a more recent cohort multicenter study in about 1800 children in UK showed white ethnicity to be independently associated with autoantibody positivity with 1.18 risk ratio, comparing non-white ethnicities (Asian, African-Caribbean and other or mixed ethnicity) [55].

The study that analyzed the data from The Type 1 Diabetes Trial Net Pathway to Prevention Study, revealed that conversion from single to two or more (multiple) positive autoantibodies was greater in Non-Hispanic than Hispanics Whites, after adjusting for autoantibody type, age, sex, Diabetes Prevention Trial Type 1 Risk Score and HLA DR3-DQ2/DR4-DQ8 genotype. However, conversion from multiple autoantibody-positive to T1DM was not different by ethnicity [50]. The details on the level of GADA positive (percent) in different regions are demonstrated in Table 3.

Table 3.

The level of GADA positive (percent) in different regions

Study	Year done	Kind of study	Region	Country	Ethnicity	Number of participant	Age(year)	Duration of T1DM	GAA positive (percent)
Ahmadov et al. [58]	2018	cross-sectional	Asia	Azerbaijan	Azeri	106	< 18	9.3 years	62%
Ong et al. [59]	2017	cross-sectional	Asia	Singapore	Asian (Indian)	201	43.8 +/- 0.4	20 to 60 years	11.40%
Ong et al. [59]	2017	cross-sectional	Asia	Singapore	Asian (Chinese)	655	43.8 +/- 0.4	20 to 60 years	6%
Ong et al. [59]	2017	cross-sectional	Asia	Singapore	Asian (Malay)	232	43.8 +/- 0.4	20 to 60 years	6.00%
Marandi et al. [60]	2011	cross-sectional	Asia	Iran	mainly Azeri	136	< 20	within 3 weeks	27.60%
Shivaprasad et al. [61]	2010–2012	cross-sectional	Asia	India	-	88	< 18	< 48 months	65%
Yang et al. [62]	1999–2009	cross-sectional	Asia	China	-	539	24.0 (range 1–70)	2 (range 0–348) months	53.4
Karagüzel et al. [63]	2008	cross-sectional	Asia	Turkey	-	57	11.7	3.4 years	63%
Kawasaki et al. [64]	1982–2008	cross-sectional	Asia	Japan	-	57	< 15	< 6 months	75%
Kawasaki et al. [64]	1982–2008	cross-sectional	Asia	Japan	-	97	≥ 18	< 6 months	78%
Tung et al. [65]	1989 to 2006	cross-sectional	Asia	Taiwan	-	157	< 18	within 3 weeks of diagnosis	73%
Ong et al. [59]	2017	cross-sectional	Europe	Germany	white European	1020	42.0 +/- 1.2	20 to 60 years	14%
Holmberg et al. [66]	2006	cross-sectional	Europe	Lithuania	Lithuanian	96	1–15 (median age 9.0)	-	59.40%
Holmberg et al. [66]	2006	cross-sectional	Europa (Ska°ne)	Sweden	Swedish	96	1–15 ( 9.0 years)	-	76.00%
Jaeger et al. [67]	1992–1996	cross-sectional	Europe	German	-	105	33	21 years	32%
Jaeger et al. [67]	1992–1996	cross- sectional	Europe	German	-	100	21	recent-onset diabetes	79%
Padoa et al. [68]	2007–2012	cross-sectional	Africa	South Africa	black	353	19.7 _ 10.5	5.0 years	60%
Padoa et al. [68]	2007–2013	cross-sectional	Africa	South Africa	white	119	12.7 _ 10.8	8.5 years	66%
Hawa et al. [69]	2006	cross-sectional	Africa	Cameroon	-	47	mean age 30.1 years T 7.6	mean disease duration 3.3 years	34%
Lutale et al. [70]	2003–2004	cross-sectional	Africa	Tanzania	-	94	-	Median (min-max) 3 (0–17)	29.8%
Panz et al. [71]	2000	cross-sectional	Africa	South Africa	black	100	-	varying disease duration	44%
Libman et al. [72]	1983–1997	cross-sectional	America	United States	black	43	< 19	newly diagnosed	58%
Libman et al. [72]	1983–1997	cross-sectional	America	United States	white	394	< 19	newly diagnosed	77%

Practical utility of GADA and autoantibodies

Current practice recommendations for the diagnosis of T1DM from International Society for Pediatrics and Adolescent (ISPAD) 2018 guideline are summarized in Table 4 [53].

Table 4.

Summary of recommendations for the diagnosis of T1DM adapted from ISPAD 2018 guideline

	Criteria / tools
Diagnosis of diabetes mellitus	Classic symptoms of diabetes or hyperglycemic crisis, with plasma glucose concentration ≥ 11.1 mmol/L (200 mg/dl). Or Fasting plasma glucose ≥ 7.0 mmol/L (≥ 126 mg/dL). Or Two-hour post load glucose ≥ 11.1 mmol/L (≥ 200 mg/dL) during an OGTT Or HbA1c ≥ 6.5%
The differentiation between type 1, type 2, monogenic, other forms of diabetes	Diabetes-associated autoantibodies: (GADA, IA2, IAA, ZnT8) 1 Molecular genetic testing can be helpful 2

1: The presence of one or more of these antibodies confirms the diagnosis of type 1 diabetes

2: Treatment of children with suspected monogenic diabetes should be limited to those who on clinical grounds are likely to be positive

Analysis of TrialNet data showed the risk for progression of single autoantibody to multiple autoantibodies positivity within 5 years is 35–37% in children age < 8 years [73]. The risk for progression of multiple islet autoantibodies to T1DM was 9%, in a cohort study done with more than 90 thousand children [49].

According to screening in nearly 12.000 children in Germany, the cumulative risk of T1DM onset with multiple autoantibodies positivity after 10 years and 18 years was 59.7% and 75.1%, respectively, that was similar to individuals with genetically susceptible [74]. Although, it was 84% after 15 years in three pediatric cohorts from Finland, the U.S., and Germany [48]. While according to a cohort study in children with genetically at-risk for T1DM, the cumulative risk of onset, was 70% and 100% after 10 years and lifetime, respectively [73]. Its specificity can be easily augmented by combining with other auto antibodies. Some studies have estimated that the risk of T1DM onset, with triple autoantibody positivity, approaches to 100% after 5 years [75].

On the other hand, based on the screening of 100 thousand children in Germany, multiple autoantibodies were detected in 4 of every 1000 tested individuals [73]. Knip. et al. expressed that screening of the general children population in Finland by GADA and IA-2As can identify 60% of the individuals who will develop T1DM over the next 27 years [36].

To facilitate the screening of the general population of children for autoantibodies, Elgar et al. suggested using capillary samples collected by their family instead of venous sample which is more feasible and has acceptable accuracy [76].

Cost analysis

The GADA radioimmunoassay is a relatively inexpensive laboratory test. Medicare rate for the test in year 2021 is less than 24 USD [77]. The Autoimmunity Screening for Kids (ASK) program, which detects pre-symptomatic T1DM in children and adolescents using a multi-autoantibody kit that includes GADA, reported $47 per screened and $4,700 per case for ASK screening program [78].

The timely diagnosis of T1DM enable us to intervene at early stages which shall decrease complications, mortality and morbidity, and also will translate to reduced cost associated with the disease. Yet, if we had a true cure to slow or even stop β-cell destruction and T1DM progression, cost-effectiveness would be definitely higher.

Conclusions

Identifying the high-risk individuals for T1DM and monitoring them, helps to prevent its progression and also makes the disease control easier, decreases hypoglycemic incidents, delays the long-term complications even avoids some lethal complications such as ketoacidosis with early diagnosis [79].

While T1DM incidence risk is greatest in relatives, however, < 10% susceptible individuals with HLA-conferred diabetes genes develop clinical disease.

GADA is an easy marker to measure that can be detected many months prior to the clinical presentation and remains positive even after early childhood.

Authors’ contributions

EK wrote the primary manuscript, BBN and NR helped in revising and editing the manuscript. All authors have read and confirmed the final version of the manuscript.

Funding

No funds, grants, or other support was received.

Data Availability

given this manuscript is a narrative review, there are no original data or material to be shared.

Code Availability

No codes were written for this manuscript.

Declarations

Conflicts of interest/Competing interests

The authors have no conflicts of interest to declare that are relevant to the content of this article.

Ethics approval

this manuscript did not need ethical approval since there were no animal or human participants involved.

Consent to participate

there were no participants in this study.

Consent for publication

there were no participants in this study.