Types of Diabetes
Keith, a competitive cyclist from South Africa, developed type 2 diabetes at age 55, despite biking 150 miles (250 km) every week.
Types of Diabetes
The main types of diabetes are type 2 diabetes, type 1 diabetes, and gestational diabetes. There are other less common types as well, including Maturity Onset Diabetes of the Young (MODY). In general, type 2 is the most common, and associated with insulin resistance and obesity. Type 1 is an autoimmune disease. Gestational diabetes occurs during pregnancy. In practice, however, it is not always easy to distinguish between diabetes types. This page discusses the details of the main types of diabetes and how to distinguish them from each other, including areas of overlap. Redondo et al. 2020 provide a useful overview of the various types of diabetes, their overlap and heterogeneity, and how to try to distinguish them. There are a number of different proposals for classifying diabetes, e.g., using genetics as a basis, or phenotype (e.g., BMI, insulin production, autoantibodies) (Deutsch et al. 2022).
"Diabetes mellitus" of any type develops when the body cannot produce as much insulin as it needs, resulting in high blood glucose levels (Gale 2005). Diabetes used to be considered one disease. Then it was divided into two main types called "juvenile" and "adult onset." These were later renamed "insulin dependent diabetes mellitus (IDDM)" and "non-insulin dependent diabetes mellitus (NIDDM)." Now they are called "type 1" and "type 2." Complications from any type of diabetes may include neuropathy (nerve damage), retinopathy (eye damage), nephropathy (kidney damage), and premature death. These may occur even with good glucose control (Narendran et al. 2005). Taking insulin can treat the disease, but is not a cure. The American Diabetes Association (as well as associations in other countries) releases diagnosis and classification guidelines on a regular basis (e.g., American Diabetes Association 2018).
What Is Type 2 Diabetes?
Type 2 is much more common than type 1 diabetes. Type 2 is generally associated with increased insulin resistance and often obesity, and is associated with certain susceptibility genes that tend to be different from the genes associated with type 1 diabetes (Raj et al. 2009). While type 2 diabetes is historically a disease seen in adults, children, even younger than 10 years old, are now developing it (Pettitt et al. 2014). The health risks of type 2 diabetes are higher when diagnosed during childhood than in adulthood (Al-Saeed et al. 2016).
While obesity increases the risk of developing type 2 diabetes, obesity alone cannot account for all cases of type 2 diabetes. A large U.S. study has found that of adults with diabetes, 80% were overweight, and 49% had obesity. It is interesting to note, however, that 20% of the adults with diabetes in this study did not have overweight or obesity (Nguyen et al. 2011). A subset of the U.S. minority population with diabetes are actually lean and have rapid failure of the insulin-producing beta cells; it is not clear why (Coleman et al. 2014). A portion of people with type 2 diabetes don't have typically related health problems like high cholesterol or high blood pressure either (Rottenkolber et al. 2021). In Uganda, in a group of people with adult-onset diabetes presumed to be type 2, 6.4% had autoantibodies (so therefore type 1), and 32% were lean (Kibirige et al. 2022).
In 2018, Ahlqvist et al. proposed regrouping adult-onset diabetes into 5 clusters based on six clinical features (GAD autoimmunity, beta cell function, BMI, insulin resistance, HbA1c, and age of diagnosis). The clusters are:
severe autoimmune diabetes (SAID),
severe insulin-deficient diabetes (SIDD),
severe insulin-resistance diabetes (SIRD),
mild obesity-related diabetes (MOD), and
mild age-related diabetes (MARD).
The first two clusters would generally be considered type 1 diabetes, and the others type 2. These subgroups have been used in various different countries with some success (Li et al. 2020; Wang et al. 2020).
What Is Type 1 Diabetes?
Type 1 diabetes, formerly called juvenile diabetes, is an autoimmune disease where the body's own immune system destroys the insulin-producing beta cells in the pancreas. People with type 1 usually have certain autoantibodies that may appear years before the disease develops, even in the womb. These autoantibodies are used as markers of the disease, but do not necessarily cause the beta cell destruction. Some people, in fact, have these autoantibodies but never develop diabetes (Narendran et al. 2005). (The immune cells that cause the beta cell destruction are thought to be largely T-cells (CD4+ and CD8+), although other immune cells may also play a role.)
A number of genes have been identified that are associated with the risk of developing type 1 diabetes. Some people, then, have a higher genetic risk than others, in other words, are more genetically susceptible. The genetic component of type 1 diabetes, however, is "neither sufficient nor necessary" (Vehik et al. 2008). That is, there is some environmental component to the disease: someone with high genetic risk might never develop it, while someone with low genetic risk might. More than 85% of the people who do develop type 1 diabetes do not have a parent or sibling with the disease (Larsson et al. 2004). (See the section at the bottom of this page for more on genetics).
I was first diagnosed with gestational diabetes, then pre-type 2, then finally type 1...
While type 1 diabetes is autoimmune, a surprisingly high percentage (16%) of children and young adults newly diagnosed with type 1 diabetes in a Colorado diabetes center test negative for the antibodies associated with the disease. The higher the age of diagnosis, the lower the chance of autoantibodies being positive. Those who tested negative for antibodies had a higher body mass index, and may have a non-immune form of diabetes (perhaps different from either type 1 or type 2) (Wang et al. 2011). Meanwhile, a study from India looked at young, lean adults with diabetes, and measured for antibodies. It found that most people tested negative for antibodies (despite having normal body weight), and were determined to have type 2 (Kotwal et al. 2017). So while body weight is often a distinguishing feature of type 1 vs type 2 in adults, it's not always.
Sweden is one country where diabetes diagnosis is done right, at least in children. Since 2005, data on all children and adolescents with newly diagnosed diabetes in Sweden are collected and include family history, clinical symptoms, c-peptide levels (a measure of beta cell function), genetic analyses, and detection of numerous autoantibodies. Overall, 95% of the youth with diabetes had type 1, 2% had type 2, 1% Maturity Onset Diabetes of the Young (MODY), and the rest secondary/unspecified type. Of those with type 1, 7% were autoantibody negative (Persson et al. 2018). (For more on autoantibody negative type 1 diabetes, see Patel et al. 2021). In the U.S., the government doesn't really track type 1 diabetes; the CDC tracks diabetes overall but it's an estimate.
Type 1 Diabetes in Adults/LADA
There is a good reason the name of "juvenile diabetes" was changed to "type 1" diabetes-- it's a disease that can be diagnosed at any age-- including in babies under 6 months old (Johnson et al. 2020). The oldest I have seen is 95 years of age (Ng et al. 2021), and a couple of 93 year olds as well (Oriot et al. 2018; Ahmad et al. 2022)! A missed diagnosis can have serious consequences, as was seen in this case study of a 51 year old man (Bao et al. 2019).
A recent study looked at all new cases of diabetes diagnosed during a three year period in an area of Sweden, and tested people of all ages for GAD and/or islet antibodies (if positive, they were classified as type 1). It found that almost 60% of newly diagnosed type 1 patients were over age 40, with incidence peaks in ages 0-9 and 50-80. Among adults, 6.9% of people with diabetes had type 1, which is similar to or slightly lower than other studies have found (Thunander 2008). In a group of UK adults who developed diabetes after age 30 and were being treated with insulin, 21% met the criteria for type 1 diabetes-- but 47% of them thought they had type 2 (Thomas et al. 2019). In Iceland, population-wide, the number of adults diagnosed with type 1 is about half the number of children (Halldora Thordardottir et al. 2019). In California, about 8% of the people who developed diabetes as younger adults (age 20-45) tested positive for autoantibodies and were determined to have type 1 (Lawrence et al. 2020). So while type 1 in adults may be less common than in children, it's still pretty common.
Type 1 Diabetes Incidence by Age
Type 1 diabetes incidence by age of diagnosis, U.S. The incidence is highest at puberty. Figure from Giwa et al. 2020.
LADA, Latent Autoimmune Diabetes in Adults, is defined by an adult age of onset, the presence of islet autoantibodies (another marker of type 1), and the lack of a need for insulin until at least 6 months after diagnosis. LADA may actually be the same as adult-onset type 1 diabetes-- the only difference is whether or not insulin is required at diagnosis, and that can depend on when and how the diagnosis is made, as well as who makes it (Fourlanos et al. 2005). These authors suggest that we use the term "autoimmune diabetes" instead of LADA, until we determine whether LADA is actually different than adult-onset type 1 or not. The World Health Organization considers LADA to be a slowly progressing form of type 1 diabetes. It does appear that type 1 onset tends to be more severe in children than in adults (Leete et al. 2018), but is it really a different disease? I don't know. Research shows that it may be at the "genetic intersection" of type 1 and type 2 diabetes (Grant et al. 2010; Ramu et al. 2018), perhaps genetically slightly closer to type 1 than type 2 (Carlsson 2019; Mishra et al. 2017; Cousminer et al. 2018), although also distinct from type 1 (Heneberg et al. 2018; Cousminer et al. 2018; Ji et al. 2019) and with its own genetic marker (Andersen 2020). Insulin secretion in LADA is also mid-way between type 1 and type 2 diabetes (Hernandez et al. 2015), as are the immune system changes (Singh et al. 2019). Lifestyle factors generally linked to type 2 diabetes also are linked to LADA (Carlsson 2019). Scientists are trying to find markers that can be used to help diagnosis and help distinguish LADA from type 2 diabetes (e.g., Yu et al. 2019). Some autoantibodies may help to distinguish LADA from type 1 diabetes as well (Baumann et al. 2020).
Some researchers propose that the term LADA be retired altogether (while acknowledging that the term helped "put the existence of autoimmune diabetes in adults on the scientific map") (Rolandsson and Palmer 2010). They also point out that considering LADA as a form of type 1 diabetes means that more adults than children are affected by autoimmune diabetes. Other researchers are trying to figure out how to distinguish LADA from "classic adult-onset type 1 diabetes" (Castelblanco et al. 2018; Hernandez et al. 2022) and honestly I'm not really clear on what the difference is.
The Immunology of Diabetes Society proposed the these criteria for diagnosing LADA:
age of onset over 30 years;
presence of any islet cell autoantibody; and
absence of insulin dependency for at least six months after diagnosis
However, these criteria have been criticized, since they seem inadequate to detect the full spectrum of the LADA. Even the existence of LADA as a distinct clinical entity has been debated, who consider LADA and type 1 diabetes as "two opposing ends of the continuum of autoimmune diabetes," and many think it is time for a new definition of LADA, as well as diabetes overall (Koufakis et al. 2019).
A review of the literature on type 1 diabetes incidence in adults (age 15 and over) finds that the geographic incidence parallels that in children, but is higher in males than females. Overall, they found that incidence decreased after age 14 (that is, fewer were diagnosed at older ages, although as above, more adults than children could still be affected by the disease, as children do grow into adults. Usually.) (Diaz-Valencia et al. 2015).
The journal Frontiers in Endocrinology published an entire research topic on LADA; see the introduction and summary article. The editor notes, no matter what we call it, "Either way the question is why does it differ and what should we do about it?" (Leslie 2022). For starters, some suggest just testing for autoantibodies in people diagnosed with type 2 diabetes (Ludvigsson, 2023). Makes sense to me!
A Heterogeneous Disease
New research from human pancreatic tissue shows that type 1 diabetes is complex, and that many factors may contribute to its development. They find that type 1 can look very different in different people; that even in one person's pancreas, parts may have functional beta cells while other parts don't; that beta cell dysfunction may be a cause of high blood sugar (not just beta cell death)-- in sum, that we have much to still learn (Pugliese et al. 2014). Interestingly, a significant number of people with type 1 still secrete a small amount of insulin-- even after having long-term diabetes-- especially if they were diagnosed in adulthood (Davis et al. 2015; Oram et al. 2015; Williams et al. 2016). A new marker has been found that helps distinguish type 1 diagnosed at early and later ages. Those diagnosed before age 7 have a more aggressive form of beta cell destruction, and those diagnosed after 13 have a less aggressive form (Leete et al. 2016); there are other differences as well (Parviainen et al. 2022).
Insulitis, which is inflammation of the islets of the pancreas, is thought to contribute to type 1 diabetes. Yet samples from the pancreatic tissue of people with diabetes have found that 28% of people with type 2 had insulitis, and 31% of those with type 1 (the authors therefore propose a new definition of insulitis, as it now does not distinguish between type 1 and type 2 diabetes) (Lundberg et al. 2017).
LADA, for example, is heterogeneous. Levels of various antibodies differ among individuals, and these antibodies come and go over time. Some who lost antibodies over time resembled type 2 patients-- although with lower beta cell function-- and some who gained antibodies more closely resembled people with "classic" type 1 (Sørgjerd et al. 2012). The pancreases of both rats and humans with LADA are similar, and show that LADA may be a more mild form of type 1 diabetes in older people than classic type 1 in children (Jörns et al. 2020).
The Network for Pancreatic Organ Donors with Diabetes (nPOD), a project that sends the pancreas tissue of people with diabetes (primarily type 1) who have died to researchers for examination, has found that the disease is highly heterogeneous (Kaddis et al. 2015).
Ethnicity is also a factor in the heterogeneous nature of type 1 diabetes. Certainly incidence rates vary be ethnicity (see the Incidence page), but also numerous other factors. Unfortunately, most studies have been conducted in non-Hispanic whites so far, so there is much we do not know yet (Tosur and Redondo 2018). A genetic risk score has been developed to identify people of African ancestry who are at increased risk of type 1 diabetes (Onengut-Gumuscu et al. 2019); we need more studies like this. In Africa, at least sub-Saharan Africa, type 1 diabetes appears to be a bit different than in Europe, with lower rates of autoimmunity, more insulin secretion, older age of onset, and different genetic background (Katte et al. 2023).
Genes likely play a role in the heterogeneity of type 1 diabetes. One study found that people with type 1 who had type 2 related genes had more minor immune-related characteristics at diagnosis (e.g., fewer autoantibodies, more beta cell function), and thus type 2 related mechanisms may play a larger role in their disease (Redondo et al. 2018).
Heterogeneity is also seen in the development of microvascular complications from type 1 diabetes. People with type 1 who have these complications are genetically different from those without those complications (Mukhopadhyay et al. 2018).
Some type 1 researchers have recently proposed the concept of using "endotypes" to address the heterogeneity of type 1 diabetes (Battaglia et al. 2019). This approach could better target treatments or prevention measures based on which subgroup of type 1 a person has. It's basically applying a precision medicine approach to type 1 diabetes.
The term "brittle diabetes" used to be used for people who had wild swings in blood glucose levels (which is also somewhat normal in type 1 diabetes), so it wasn't quite clear what it meant. Some propose a modern version that attempts to identify the reason for those wild swings, such as eating disorders, additional health issues like gastroparesis or dementia, or those who have social barriers such as food insecurity (Hirsch and Gaudiani, 2021).
Before Diabetes Develops: The Stages of Disease Development
In 2015, new stages of type 1 diabetes development were proposed, to encourage risk screening and prevention studies (Insel et al. 2015). The three stages are:
Stage 1: pancreatic beta cell autoimmunity (2 or more autoantibodies) with normal blood sugar levels (presymptomatic)
Stage 2: pancreatic beta cell autoimmunity with higher blood sugar levels or glucose intolerance (presymptomatic)
Stage 3: clinical onset of type 1 diabetes with high blood sugar levels and symptoms
In 2017, a new name for Stage 1 type 1 diabetes was proposed, Autoimmune Beta Cell Disorder (ABCD) (Bonifacio et al. 2017). There are various controversies surrounding this proposed name (e.g., ABCD is already used as a term to describe obesity, Adiposity-Based Chronic Disease) but it seems like general support of the use of stages to better characterize the early development of type 1 diabetes (Becker and Insel, 2018).
Insulin Resistance and Obesity in Type 1 Diabetes
People with type 1 diabetes have higher insulin resistance than those without diabetes, even in lean patients (e.g., Donga et al. 2013). A Colorado study found that insulin resistance was a significant feature in adults with type 1. This insulin resistance was not explained by body mass index, percentage fat, visceral fat, lipids/cholesterol, exercise, or high blood glucose (Bergman et al. 2012).
Curiously, in a study of Malaysian people, those with LADA had higher insulin resistance levels than antibody-negative adults with diabetes, despite a lower body mass index (Salem et al. 2014).
There are people with type 1 diabetes who also have obesity, and that can also cause misdiagnosis. Misdiagnosis can have serious consequences, even in adults. For example, a person who underwent bariatric surgery for obesity ended up with ketoacidosis, because they had type 1, not type 2 diabetes (Pilla et al. 2018).
It seems pretty clear that non-immune processes like insulin resistance or obesity can stress beta cells, as well as amplify autoimmunity and contribute to the development of type 1 diabetes, especially at older ages (Redondo et al. 2019). See the Insulin Resistance and Height and Weight pages for more on this topic.
Fulminant Type 1 Diabetes
Fulminant type 1 diabetes is a subtype of type 1 diabetes characterized by extremely rapid onset and complete deficiency of insulin due to the destruction of pancreatic beta cells, more common in Asians. Autoimmunity may (or may not) play a role in fulminant type 1 diabetes, although this is under investigation (Oikawa and Shimada 2020; Wang et al. 2016). One study defines it as, "characterized by remarkably rapid and complete β-cell destruction. The established diagnostic criteria include the occurrence of diabetic ketosis soon after the onset of hyperglycemic symptoms, elevated plasma glucose with relatively low HbA1c at the first visit, and extremely low C-peptide." This study, however, describes a case where the patient recovered some beta cell function over time, taking a small amount of insulin. The authors point out that fulminant type 1 diabetes is heterogeneous (Kaneko et al. 2017), as is type 1 diabetes throughout Asia (Park et al. 2017). Glucagon levels are also lower in those with fulminant type 1 diabetes as compared to people with type 1 diabetes (Komada et al. 2018). Liu et al. (2018) present the diagnostic criteria for fulminant type 1 diabetes as agreed on by the Japanese Diabetes Association. Another study finds a new marker that could help distinguish it from type 1 (Yin et al. 2023). The development of fulminant type 1 is so rapid that it may take only a few days, seldom longer than a week-- which could explain why some people have a relatively normal HbA1c but very high blood sugar levels (Keskin et al. 2018). Since it is so fast-acting, there is a high risk of death and proper diagnosis is important (You et al. 2019).
For more on fulminant type 1 diabetes, see the Oct. 2020 issue of Diabetology International (vol. 11 issue 4) which contains numerous articles, as well as additional articles (e.g., Ying et al. 2021).
Interestingly, fulminant type 1 diabetes often appears in pregnant women (Yang et al. 2020). So I am wondering if there is a link between this and "autoimmune gestational diabetes"? (See below under Gestational Diabetes). Does pregnancy somehow trigger autoimmunity and a fast-onset type 1? (I personally developed type 1 during pregnancy and know a lot of other women who have as well).
Double Diabetes in Children
More and more children, meanwhile, have characteristics of both type 1 and 2 diabetes, including autoantibodies to beta cells, as well as signs of increased insulin resistance or obesity. Some call this "double" or "hybrid" diabetes (Khawandanah 2019; Popovic and Papanas 2021; Pozzilli et al. 2007; Pozzilli et al. 2011). In fact, a large number of people with diabetes may have both type 1 and type 2 associated processes contributing to their diabetes (Tuomi 2005). Interestingly, one research group took another look at the diagnosis of children with diabetes (mostly type 1), after an average of 7 years. Fully 20% of the patients had a different diagnosis that what was originally given, after reevaluation. And, 10% of the children were determined to have both type 1 and type 2, a "mixed diabetes phenotype" Lipton et al. 2011).
One case study, for example, of an adolescent with unclear diabetes type. He had obesity and insulin production, which are signs of type 2, plus multiple islet autoantibody positivity and severe hyperglycemia at diagnosis, which are signs of type 1. He was successfully treated with liraglutide, without insulin. The authors note the "difficulty of diabetes classification and subsequent need for personalized medicine with regard to diabetes management" (Ginnard et al. 2022).
A German study of children diagnosed with type 2 diabetes found that 15% tested positive for islet autoantibodies (Awa et al. 2013). A similar UK study found that 30% of children with type 2 tested positive for autoantibodies, while 26% of those with type 1 tested negative (Perchard et al. 2015). A U.S. study found that of 145 youth diagnosed with obesity and type 2 diabetes, 70 tested positive for autoantibodies-- that's 48% (Rivera-Vega et al. 2015). A study of children from India found that 7% had "double diabetes," 51% had type 1, 13% had type 2, and fully 29% had an "unknown" type of diabetes (Mishra et al. 2018). In Polish children with obesity, 17% tested positive for islet autoantibodies (Chylińska-Frątczak et al. 2022).
Other authors have described a "bidimensional spectrum" in diabetes. They analyzed US children diagnosed with diabetes, and tested their autoantibody levels and levels of insulin resistance. They found that 55% of the children could be classified as autoimmune and insulin sensitive (traditional type 1), and 16% as non-autoimmune and insulin resistant (traditional type 2). Almost 20% had autoimmunity and insulin resistance, signs of both type 1 and type 2, perhaps including those who are overweight with type 1. Then 10% were non-autoimmune but insulin sensitive, showing signs of neither type 1 or type 2, perhaps indicating a different type altogether (such as MODY) (Dabelea et al. 2011).
Immune Dysfunction in Type 2 Diabetes
There is growing evidence that immune system abnormalities may play a role in type 2 diabetes. While type 2 diabetes is not considered an autoimmune disease, 15-35% of type 2 patients diagnosed before age 45 test positive for antibodies to GAD, as do 7-9% of patients diagnosed at an older age (Tuomi 2005). (These antibodies are also a marker of type 1 diabetes; see the Autoimmunity page for more information on the antibodies associated with type 1 diabetes). Here are examples of studies where people diagnosed with "type 2" were tested for autoimmunity:
In the U.K., about 10% of adults who had been diagnosed with type 2 diabetes tested positive for islet or GAD autoantibodies (Fourlanos et al. 2005).
In the U.S., 6.4% of a group of adults diagnosed with overweight or obesity and type 2 diabetes tested positive for islet autoantibodies, including 12.8% of those with severe obesity (Pilla et al. 2018).
In the U.S., 41.3% of a group of people with type 2 diabetes tested positive for "cellular islet autoimmunity" (Brooks-Worrell et al. 2022).
In the Czech Republic, 13% of people diagnosed with type 2 tested positive for GAD autoantibodies (Martinka et al. 2016).
In Bulgaria, 10% of adults diagnosed with type 2 tested positive for autoantibodies (Zaharieva et al. 2017).
In China, about 6% of adults with diabetes tested positive for autoantibodies (Xiang et al. 2018).
Another Chinese study found much higher rates, 28.5% (Gao et al. 2019).
And another from China found 5.5% in males and 6% in females (Tang et al. 2019).
A study of Chinese youth/young adults (age 15-29) with type 2 diabetes found that 11.7% tested positive for autoantibodies (Xiang et al. 2021).
In Italy, 25% of people diagnosed with type 2 tested positive for autoantibodies (Tiberti et al. 2018).
In India, about 1.5% of adults diagnosed with type 2 diabetes tested positive for autoantibodies (Sachan et al. 2015).
In Iran, about 20% of adults diagnosed with type 2 diabetes tested positive for autoantibodies (Moosaie et al. 2020).
Including elderly adults:
In China, over 35% of elderly adults diagnosed with type 2 diabetes tested positive for islet autoimmunity (Li et al. 2018).
In Argentina, 16% of adults without obesity but diagnosed with diabetes after age 65 tested positive for autoantibodies (Yohena et al. 2019).
And children as well:
In Turkey, almost 15% of children diagnosed with type 2 tested positive for autoantibodies (Haliloğlu et al. 2018)
In the U.S., almost 10% of children diagnosed with type 2 tested positive for autoantibodies (Klingensmith et al. 2010).
In China, about 9% of adolescents and young adults with what looked like type 2 were positive for autoantibodies (Cheng et al. 2022).
In Vietnam, about 28% of adolescents tested positive for GAD autoantibodies (Huynh et al. 2023).
In both adults and children, then, a percentage of people with type 2 diabetes show signs of autoimmunity. It may be that they were simply misdiagnosed, and really have type 1 diabetes.
However, chronic inflammation (a sign of immune dysfunction), is also involved in type 2 diabetes, insulin resistance, and obesity. Some authors are questioning whether or not type 2 diabetes may also involve some sort of autoimmune response (Itariu and Stulnig, 2014;Velloso et al. 2013). Immune cells have been found in the islets of people with type 2 diabetes, showing islet inflammation likely plays a role in the beta cell failure found in type 2 diabetes (Westwell-Roper and Ehses, 2014). Obesity can also play a role. The levels of antibodies can vary by body mass index (BMI); certain antibodies are higher in people who have a higher BMI, while others are lower (Buzzetti et al. 2015).
There is evidence that islet autoimmunity plays a role in the development of type 2 diabetes. Some type 2 patients may even test negative for islet autoantibodies, but positive for islet reactive T-cells-- these authors refer to this characteristic as "autoantibody negative autoimmune type 2" (Brooks-Worrell et al. 2011). For example, certain inflammatory autoreactive T-cells are found in people with type 1 and type 2 diabetes (CD-4+ T-cells), while other autoreactive cells are only found in people with type 1 (CD-8+ T-cells) (Sarikonda et al. 2014).
Other authors have found that additional autoantibodies are found in people with type 1 diabetes, and a small percentage (5%) of people with type 2. These are "IA-2ec autoantibodies," which are "autoantibodies directed to the extracellular domain of the neuroendocrine autoantigen IA-2." I do not know what that means, but the authors suggest that, "islet autoimmunity may be more common in clinically diagnosed type 2 diabetes than previously observed (Acevedo-Calado et al. 2017). Researchers have also found "transient" autoimmunity in people who would classify as type 2 (Sørgjerd et al. 2018).
And then obesity as well is also linked to an increased risk of beta cell autoimmunity (Chylińska-Frątczak et al. 2022).
The Diabetes Spectrum
I like the idea that diabetes is a spectrum of disease, with type 1 on one end, and type 2 on the other.
While type 1 and type 2 diabetes are now classified as separate diseases, there is some overlap between them (Tuomi 2005). For example:
In type 1 diabetes, the beta cells in the pancreas do not produce any (or at least enough) insulin. Yet there is evidence of decreased beta cell mass and impaired beta cell function in type 2 diabetes as well (Portha 2005).
There are differences (and some similarities) in the mechanisms that lead to beta cell death in type 1 and type 2. Beta cell apoptosis, or programmed cell death, is probably involved in both types (Cnop et al. 2005). Beta cell death in type 1 is thought to be largely due to an autoimmune attack on the beta cells (Narendran et al. 2005). The cause of beta cell loss in type 2 diabetes is not known, but perhaps due to high blood glucose levels, or certain types of immune system cells called cytokines (Rhodes 2005).
Both metabolic (beta cell dysfunction and insulin resistance) and immune processes (autoimmunity and inflammation) may contribute to different extents to the development of different types of diabetes, and play a role in beta cell destruction. Some authors call for a revision of the classification system for diabetes altogether (Brezar et al. 2011; Schwartz et al. 2016).
Some researchers have proposed that diabetes should be considered more as a continuum, since clinical characteristics change gradually from type 2 to LADA to type 1 (van Deutekom et al. 2008). Genetic background also seems to support this idea of a continuous spectrum of diabetes (Lin et al. 2008). People with a family history that includes both type 1 and type 2 probably have a more intermediate type of diabetes (Tuomi 2005).
The distinctions between type 1 and 2 are most clear when comparing the extreme ends of the diabetes spectrum: a young child diagnosed with diabetes whose body produces no or very little insulin, who requires insulin immediately to stay alive, who is sensitive to insulin, and who tests positive for islet autoantibodies (type 1); compared to an adult whose pancreas still produces insulin, who is insulin resistant, who never requires insulin treatment, and who tests negative for certain antibodies (type 2). Things get more fuzzy when dealing with those diagnosed as adults who may or may not require insulin either immediately or after a few years, may or may not test positive for various autoantibodies, and who have varying amounts of pancreatic beta cell function remaining. The onset of type 1 diabetes tends to be slower at an older age of diagnosis, and faster at a younger age. However, age is not necessarily a great determinant either. A Danish study found that c-peptide (a marker of beta cell function), and GAD antibody status, better distinguished different groups of patients from each other than age did (Wod et al. 2017). A German study used a variety of markers like c-peptide and autoimmunity and BMI and identified 10 subgroups of type 1 diabetes, just in children (Achenbach et al. 2022).
Clinicians tend to just use clinical features (like BMI) to distinguish type 1 from type 2, but this results in a high rate of misclassification (7-15%). A review finds that age of diagnosis and time to insulin were the best clinical characteristics to use (and not BMI) (Shields et al. 2015). I personally think these are not adequate and that other measurements should be used.
Autoimmunity remains the main distinction between type 1 and type 2 diabetes, although even then there are limitations in measuring autoimmunity (Gale 2005). There is good evidence that type 1 diabetes involves immune processes, although we are not really sure if these immune processes are a cause, preexisting condition, or result of the disease. Actually, both immune and non-immune processes may be involved in the development of type 1 diabetes, especially in adults. Also, immune processes may play a role in the development of type 2 diabetes (Gale 2006). Confused yet?
Gestational diabetes develops during pregnancy, and disappears after the pregnancy ends. Yet women who have had gestational diabetes are at increased risk of developing type 2 diabetes later in life. Gestational diabetes is actually quite heterogeneous (e.g., Huvinen et al. 2018)-- not everyone who is overweight develops it, and plenty of women who do not have overweight or obesity do develop it. In one Japanese center, for example, fully 40% of women who developed gestational diabetes were lean but had impaired insulin secretion for some reason (Furukawa and Kobayashi, 2019).
Interestingly, some women (myself included), from populations around the world, develop type 1 diabetes following gestational diabetes (e.g., see Fujishima et al. 2023; Unnikrishnan et al. 2016). One study found that about 10% of a group of Finnish women with gestational diabetes developed diabetes over the subsequent six years, and that nearly half of them developed type 1. While most of the women who later developed type 1 initially tested positive for autoantibodies (during pregnancy), two of the 23 women who were determined to have developed type 2 also tested positive for autoantibodies (Järvelä et al. 2006). In a study from Egypt, the number of women with gestational diabetes found to be autoantibody positive was 44%! Half of them had developed permanent diabetes one year after childbirth (Amer et al. 2018). In Finland, 5.7% of the women with gestational diabetes developed type 1, after 7 years (Auvinen et al. 2020). There are apparently some genetic links between gestational diabetes and type 1 diabetes as well (Guo et al. 2016).
Wucher et al. (2010) describe an "autoimmune gestational diabetes," where women with gestational diabetes have a high risk of developing type 1 diabetes. These women tend to have a low body mass index, and require insulin during pregnancy (although usually not afterwards for a time). According to these authors, 2-17% of pregnant women develop gestational diabetes, depending on the population studied. Perhaps 15-30% of these women have "pregestational" diabetes, that is, unrecognized type 2 diabetes. About 10% of women with gestational diabetes have "autoimmune gestational diabetes" and may go on to develop type 1. That is what happened to me. At the time, the doctors just looked at me as if I was some kind of anomaly; no one had ever heard of a patient developing type 1 after gestational diabetes. But it is important for doctors to be more aware of this possibility; at least one case study describes a 28 year old woman who died a month after giving birth. She had had prior gestational diabetes-- but she was also antibody positive and her pancreas showed signs of type 1 diabetes. Thus undiagnosed type 1 diabetes probably played a role in the health problems that led to her death (Jackson et al. 2017). It is interesting that doctors are so unaware of this happening that individual case studies are still being published (e.g., Ikeoka et al. 2018). Some doctors propose testing all women with gestational diabetes for type 1-related autoantibodies (Incani et al. 2019), although others find that it may not be worth the trouble (Beunen et al. 2022).
There are also reports of fulminant type 1 diabetes appearing during or just after pregnancy (Farrant et al. 2016; Furukawa et al. 2016; Liu et al. 2018; Ying et al. 2019).
And, some women (mis)diagnosed with gestational diabetes may actually have Maturity Onset Diabetes of the Young (MODY) (see more on MODY below) (Majewska et al. 2022).
Other Types of Diabetes
Type 1, type 2, and gestational diabetes are the most common types of diabetes. Yet looking at the list provided by the American Diabetes Association, I count 48 other types of diabetes, plus "others." These include those that are drug or chemical induced, those caused by a genetic defect of the beta cells. and more (American Diabetes Association, 2018). One interesting type is found in people with a low BMI in developing countries, that used to be called "malnutrition-related diabetes." Researchers are trying to figure out how to classify, describe, and diagnose it (Lontchi-Yimagou et al. 2022).
Diabetes diagnosed shortly after birth and mainly before six months of age could be congenital diabetes (also called neonatal diabetes), which is rare. About half of these cases are transient and resolve, and about half are permanent. The treatment does not always have to involve insulin (Beltrand et al. 2020; Iafusco et al. 2020).
Maturity Onset Diabetes of the Young
One type of diabetes caused by a genetic defect is Maturity Onset Diabetes of the Young (MODY), a type of monogenic diabetes. MODY is characterized by genetic inheritance, onset before 25 years of age, the absence of beta-cell autoimmunity, and sustained pancreatic beta-cell function (Urakami et al. 2019). There are even 13 subtypes of MODY (Heuvel-Borsboom et al. 2016) and mutations in at least 14 genes associated with MODY (Oliveira et al. 2019). MODY is often confused for type 1 diabetes, but MODY can sometimes be treated without insulin. A subset of children with diabetes (usually assumed to be type 1) who are tested for MODY find a small percentage actually have the genes linked to MODY:
In a nation-wide screening of all Norwegian children with diabetes, fully 6.5% of those who were antibody-negative had MODY (Johansson et al. 2017).
A U.K. study found that 2.5% of children with diabetes had monogenic diabetes (Shepherd et al. 2016), and another found 3.6% of those diagnosed under 30 (Shields et al. 2017).
An Australian study found 2% of children had genetic signs of MODY (Johnson et al. 2018).
An Italian study found that 6.3% of children with diabetes had monogenic diabetes (while 1% had type 2 diabetes) (Delvecchio et al. 2017).
A Finnish study found that when diagnosed before age 6 months of age, autoimmunity is less common, and monogenic diabetes more common. When diagnosed from 7-12 months of age, autoimmunity is more common (Huopio et al. 2016).
A Finnish study found that over 10% of children who tested negative for autoantibodies when newly diagnosed with diabetes actually had MODY (Harsunen et al. 2022).
A Swedish study determined that testing those who have diabetes diagnosed in childhood, without all islet autoantibodies, and with only modest hyperglycemia, leads to a pretty good MODY detection rate (Carlsson et al. 2019).
A Turkish study found classifying monogenic diabetes is more complicated than previously thought (the authors note that "four patients died during follow-up" which seems to me to be appalling; why did they have to die? It is not clear to me how that would happen, even if they didn't know what type of diabetes they had) (Şıklar et al. 2017).
A Chinese study found that 19.5% of Chinese Han people diagnosed with type 1 but antibody negative had mutations associated with MODY. Based on this, they estimate that about 6% of Chinese Han diagnosed with type 1 actually have MODY (Li et al. 2019).
A Thai study found that 7 of 73 adults diagnosed with type 1 diabetes probably had MODY (Thewjitcharoen et al. 2023).
In adults, a Joslin Diabetes Center tested Joslin "medalists" who have had type 1 diabetes for over 50 years. Fully 8% of those people had the genes linked to MODY, and of those, about half did not have a genetic risk for type 1. So at least 4% have MODY and the researchers aren't even sure what type the other 4% have. They suggest that everyone under 18 diagnosed with type 1 diabetes be screened for MODY (Yu et al. 2019; Joslin Diabetes Center News Release, July 2, 2019).
In adults, diabetes following pancreatic disease is also relatively common, and can be misdiagnosed (Woodmansey et al. 2017). Diabetes can also be a sign of pancreatic cancer (Riaz and Dolan, 2018).
And then there are reports of people having more than one type of diabetes-- in this case study, a woman had gestational, then LADA, then MODY... (Garrahy et al. 2019). I don't know why it wouldn't just be MODY and misdiagnosed as LADA. Another report describes a child diagnosed with MODY and then "Latent Autoimmune Diabetes in Youth (LADY)"-- what? They propose that the beta cell stress of MODY led to autoimmunity over time (Yoshida et al. 2020).
The classification of diabetes is more complicated than I had assumed, and distinguishing one from another can be difficult.
People with type 1 have higher rates of some other autoimmune diseases than the general population, especially thyroid disease, celiac disease, autoimmune gastritis, and Addison's disease (Krzewska and Ben-Skowronek, 2016). Of patients with type 1 diabetes, perhaps 25% will have thyroid disease (the most common other autoimmune disease associated with type 1), around 7% will have celiac disease, up to 20% will have markers of pernicious anemia (associated with autoimmune gastritis), and about 2% will have markers of Addison's disease (Narendran et al. 2005). Having both type 1 diabetes and celiac disease leads to an increased risk of autoimmune thyroid disease (Kurien et al. 2016).
Autoimmune diseases may show up early, or late. One clinic screened patients for thyroid, celiac, and Addison's disease association antibodies when they were diagnosed with type 1, and found that one-third of new type 1s had autoantibodies for these other autoimmune diseases, already at diagnosis (although not necessarily the disease yet) (Triolo et al. 2011). A study looking at celiac disease in people with type 1 found that it tends to appear in people who have had long-standing type 1 (over 15 years duration). Curiously, celiac disease was more common in people with adult-onset type 1 than in those who had been diagnosed with type 1 during childhood (Tiberti et al. 2014).
About 90% of people who develop type 1 diabetes do not have an immediate family member with diabetes. Since I had type 1, my son had a slightly increased risk, which would have been higher if his father had had diabetes instead of his mother (The TEDDY Study Group 2007).
There are two main categories of autoimmune diseases: systemic autoimmune diseases, that target multiple organs or tissues; and organ-specific autoimmune diseases, that target one organ (such as the pancreas in type 1 diabetes). Some researchers propose that many autoimmune diseases may share a common mechanism, especially organ-specific autoimmune diseases. Under this hypothesis, a single environmental factor may lead to the occurrence of different autoimmune diseases, frequently more than one, in a single individual. Also, different environmental factors might lead to a single type of autoimmune disease in individuals genetically susceptible to that disease (Sakaguchi 2004). Thus, the role that environmental exposures may play in other autoimmune diseases may be relevant for type 1 diabetes.
Various differing autoimmune diseases have also been found to have common susceptibility genes (e.g., see Maiti et al. 2010). Type 1 diabetes and celiac disease, for example, are both associated with some of the same susceptibility genes (Smyth et al. 2008).
The role of environmental factors in various autoimmune diseases is a growing area of research. The current state of this research is summarized nicely for the public in this article, Questions Persist: Environmental Factors in Autoimmune Disease, published in the peer-reviewed journal, Environmental Health Perspectives (Schmidt 2011).
Non-Alcoholic Fatty Liver Disease (NAFLD)
Adults (but not children) with type 1 diabetes are at a higher risk of NAFLD than the general population (Kummer et al. 2017).
Metabolic syndrome is a cluster of medical conditions that increases the risk of developing cardiovascular disease (and sometimes diabetes). The WHO defines the syndrome as the presence of either diabetes, impaired fasting glucose or insulin resistance, AND, two of the following: high blood pressure, high "bad" cholesterol or triglyerides, central obesity, or microalbuminuria. (Different organizations have differing definitions; a standard definition has yet to be determined).
While type 2 diabetes is often associated with the metabolic syndrome, it is interesting that people with type 1 diabetes also have higher risk of the metabolic syndrome as well, even sometimes without the obesity. A worldwide review and meta-analysis found that about 24% of people with type 1 diabetes had metabolic syndrome (Belete et al. 2021). One study from Brazil found that 45% of the young women with type 1 diabetes they studied had the metabolic syndrome, despite the fact that none had obesity (Momesso et al. 2011). This number, surprisingly, is consistent with other, larger studies. A Finnish study found in that 38% of men and 40% of women with type 1 diabetes also had the metabolic syndrome. These numbers are three times higher than those seen in Finns without diabetes, but lower than in people with type 2 diabetes (78-84%) (Thorn et al. 2005). An Australian study found that 15% of a group of people with type 1 diabetes had metabolic syndrome (McGill et al. 2008). These studies used different definitions of metabolic syndrome, and it remains to be determined whether this syndrome is really the same in people with different types of diabetes.
Various metabolic changes are seen in both humans and animals during the period leading up to type 1 diabetes. Together they involve increased insulin secretion and increased stress on the beta cells (Sysi-Aho et al. 2011).
Like diabetes incidence, metabolic syndrome incidence is increasing. In the U.S., incidence has increased over the past 20 years in adults, and especially in young women (Mozumdar and Liguori 2011).
Genes and Environment
A large amount of research has focused on identifying gene variants that affect the risk of a person developing diabetes, as well as other diseases. These Genome-Wide Association (GWA) studies have identified a large number of genetic variants associated with various diseases. Yet most variants so far yield only small changes in risk levels. Only 6% of type 2 diabetes can be explained by heritable factors, for example (Manolio et al. 2009). GWAS studies are showing that "the magnitude of genetic effects is uniformly very small" (Dermitzakis and Clark, 2009).
Studies of twins are also useful for looking at the role of genes. For type 1 diabetes, if one member of an identical twin pair has type 1, the majority of their twins do not develop the disease, although the younger the diagnosis, the higher the risk . To assess the role of genetics, however, most twin studies compare the rates of disease between identical versus non-identical twins. Twin studies of type 1 show that disease risk is higher among identical twins than non-identical twins, showing a role for genetic background (Hyttinen et al. 2003). For type 2, a study from Denmark has found that the risk of type 2 diabetes does not differ between identical and non-identical twins (Petersen et al. 2011). Note that the environment, including the prenatal environment, is shared by both twins, if they are raised together, except that they may be exposed to different environmental factors during life, e.g., one might get a virus but the other does not.
An interesting study adapted the techniques used in GWA studies, and instead conducted a pilot Environment-Wide Association study to consider 266 separate environmental factors with diabetes, using a large U.S. dataset. It found that the factors most associated with diabetes include the pesticide heptachlor epoxide, PCBs, and a form of vitamin E. Protective factors included beta carotenes. The size of the effects that these factors have on type 2 diabetes are comparable to the highest risk gene loci found in GWA studies. This study is only the first EWA study ever done; there have been at least 16 GWAS studies on type 2 diabetes alone. Future studies might benefit from combining GWA and EWA data and methodologies, to consider the combined effects of genes and environment (Patel et al. 2010).
To download or see a list of all the references cited on this page, as well as additional references on LADA, double diabetes, and other issues in diabetes classification, see the collection Classification and types of diabetes in PubMed.