Diabetic ketoacidosis in people with type 2 diabetes, recognising risk factors in addition to sodium-glucose cotransporter 2 inhibitors.

When I completed my formal post graduate qualification in diabetes education nearly thirty years ago, diabetic ketoacidosis (DKA) was presented as a medical emergency that only occurred in people with type 1 diabetes (T1D). Since that time, many of us have seen a number of people admitted to hospital in DKA with either known type 2 diabetes (T2D) or ultimately diagnosed with T2D.  With the rapidly increasing and widespread prescribing of sodium-glucose cotransporter 2 inhibitors (SGLT2i), an increase in people with T2D developing DKA has occurred.  In a recent review of 39 randomised control trials (more than 60,000 participants), DKA was twice as common in people prescribed SGLT2i (0.18% vs. 0.09%).¹ Locally and globally, professional bodies have highlighted the increased risk of DKA in people with T2D prescribed SGLT2i, especially during times of medical instability. A joint position statement from the Australian Diabetes Society (ADS) and New Zealand Society for the Study of Diabetes, outlines the risks of developing euglycaemic DKA in people with T2D prescribed SGLT2i who are undergoing surgery or colonoscopy  procedures.²   The American Diabetes Association (ADA 2021) recommends that use of SGLT2i be avoided for inpatients with T2D with severe illness, ketonaemia , ketonuria or during prolonged fasting.³  

Incidence of DKA in people with T2D

Although there is little Australian data available, internationally an increasing trend in the number of people with T2D and DKA presenting to hospitals has been identified.  Zhong et al (2018) recorded DKA presentations in England between 1998-2013 (prior to the wide use of SGLT2i), finding an annual increase each year of 4.24%.⁴   The authors found that people with T2D represented 20% of all DKA hospital presentations during this time.  Those aged between 50 to 65 years of age, with a duration of T2D more than 10 years and requiring insulin were at higher risk.  In Australia, prior to release of the ADS position statement, data from 11 Melbourne (and Geelong) health services found that among inpatients, those prescribed SGLT2i had a 3-fold higher rate of DKA compared to those not taking these medications.⁵  

Individuals admitted with DKA who do not fit into the usual classification of T1D have previously been described as having  ketosis-prone type 2 diabetes (KPD).⁶   Umpierrez (2006) argues KPD is usually seen in people who are overweight or obese, with a strong family history of diabetes,  with a low prevalence of autoimmune markers, and lacking an HLA genetic association. Researchers have postulated four subgroups of KPD, based on the presence or absence of β–cell autoantibodies and fasting c-peptide level, indicating the recovery or lack of recovery of β–cell functional reserve following the occurrence of DKA.⁷ A detailed clinical history, including autoantibody and fasting c-peptide testing should be performed for all people admitted with DKA and newly diagnosed diabetes. The importance of this is illustrated by a case report of a 23 year-old man with a body mass index (BMI) of 39.1kg/m2 admitted with DKA and discharged on insulin, who was subsequently  found to be antibody negative with a normal c-peptide level.⁸ Following the commencement of  metformin and a GLP-1 receptor agonist, insulin could be successfully stopped.  It is also essential to have blood ketone testing capacity in primary care to assist in identifying the need for escalating to acute emergency services.   In my own clinical practice, this case scenario is becoming increasingly more common.

Regardless of the type of diabetes, diagnosis of DKA and initial acute inpatient treatment is the same.⁹ 

Risk factors for DKA

Several common triggers exist that can increase the risk of developing DKA for people with T2D. 

Ethnic background: Several overseas studies have identified an association between ethnic background (African-American, South African, Korean) and presentation of DKA in individuals with T2D.^10-12   

Infection: Puttanna and Padinjakara (2014) summarise several studies where infection has been identified as a key trigger for DKA in people with T2D, representing between 28.5 – 48% of all hospital presentations.¹³   

Suboptimal adherence to glucose lowering medicines: In a US study including 30 admissions for DKA in people with T2D, discontinuation or non-adherence of insulin was identified as the precipitating cause in 69.2% of all cases.¹⁴  Other studies have observed lower rates, ranging between 19.4 and 50%.¹³

Acute cardiovascular event: In an individual case report, acute myocardial infarction and subsequent cardiac arrest has occurred following the occurrence of euglycaemic ketoacidosis in a person prescribed an SGLT2i.¹⁵  

Glucocorticoid therapy: Glucocorticoids worsen insulin resistance, increasing the risk of ketosis occurring in those who already have insulin insufficiency.¹⁶ DKA in people with T2D treated with glucocorticoids is more common in conjunction with coexisting stressors such as infection, cardio/cerebral vascular events or sudden cessation of insulin.16  However there are some isolated case studies which have attributed the use of glucocorticoids as the sole factor for individuals with T2D developing DKA.¹⁷

Antipsychotic medicines: Being diagnosed with a psychotic illness and its associated treatment is known to increase the risk of developing T2D, and to worsen glycaemia in those people who already have  T2D. ¹⁸ In a review of 60 reports, the occurrence of DKA in people with T2D occurred with almost all atypical antipsychotic medicines.¹⁹   However, clozapine and olanzapine have been associated with the greatest risk of developing DKA.²⁰

Pregnancy: Pregnant women with T2D  are at risk of developing DKA  at lower blood glucose levels  than would be the case in the non-pregnant state.  The risk for women with T2D is lower compared to those with T1D, but is increased at times of intercurrent illness, fasting, inadequate carbohydrate intake or inadequate insulin dosing.³  The recently released ADIPS position paper on pre-existing diabetes in pregnancy (2020) recommends blood ketone testing even if blood glucose levels are only mildly elevated during illness or vomiting.  Consideration of blood ketone testing should also occur if glucocorticoid therapy is used to enhance fetal lung maturation.²¹   ADIPS recommends an individual plan is made for all pregnant with pre-existing diabetes, however the ADA (2021) recommends specific education pertaining to DKA prevention and detection only to women with T1D.³  

Other risk factors:

A variety of rare clinical conditions can produce DKA in people with T2D and includes:

• Pancreatitis as a result of severe hypertriglyceridaemia.^{22, 23}    
• Pancreatic neoplasm (adenocarcinoma).²⁴
• Exogenous insulin allergy.²⁵
• Rhabdomyolysis (breakdown of skeletal muscle fibres with leakage of muscle contents into the circulation caused by a crush injury, prolonged bedrest, alcohol abuse, drug use, toxin exposure).²⁶

COVID-19:

COVID-19 has more recently changed the landscape of diabetes emergency presentations.  A decade ago the SARS-CoV2 virus was observed to have a toxic effect on pancreatic islets, showing an increased expression of ACE2 receptors, leading to a higher β-cell death rate, directly resulting in the reduction of endogenous insulin production, thus increasing the risk of DKA developing.²⁷  In addition, a cytokine storm (amplified immune response causing significant damage to the body) has been observed in response to developing COVID-19 which can increase the incidence of DKA in people prescribed a SGLT2i.²⁸   In the US, a review of five individuals with T2D presenting with DKA and COVID-19 (none treated with a SGLT2i), found suboptimal glycaemic management (HbA1c ranging from 9.5 – 11.9%), rather than gender, age or body mass index as a common feature.²⁹ The authors postulated that having the SAR-COV-2 infection may impact on a person’s ability to rely on compensatory tachypnoea for regulating acidosis. Diabetes UK and the Joint British Diabetes Societies for inpatient care have developed guidance for individuals with COVID-19 who are treated with high-dose glucocorticoids,  highlighting the ‘triple threat’ of high-dose glucocorticoid-induced impaired glucose metabolism, COVID-19 induced insulin resistance and COVID-19 related impaired insulin production, which together could result in significant hyperglycaemia, HHS and DKA in people with and without diabetes.³⁰

Combining triggers and risk

According to the ADA (2021), DKA seldom occurs spontaneously in people with T2D, instead arising in association when a number of clinical issues co-exist.³   The addition of SGLT2i when other co-morbidities and clinical risks concurrently occur, can increase the risk of a person with T2D developing DKA.³¹  In the absence of any tool for risk stratification for people with T2D developing DKA, it is important to highlight the signs and symptoms associated with DKA. It may be useful to identify the clinical circumstances where several triggers may occur simultaneously, for example infection resulting in the prescribing of a glucocorticoid in an individual also prescribed a SGLT2i.  Local inpatient guidelines direct health care professionals when to monitor for blood ketone levels (and may have adapted to the changing nature of DKA presentations by requiring a blood ketone check for all persons with diabetes presenting to the emergency department or for surgery?).  Primary, community, aged care, disability services and correctional services also need to be aware of the need to measure blood ketone levels when a high level of suspicion of ketosis is aroused.³²

Symptoms of DKA in T2D

It is challenging to find information pertaining to the signs and symptoms of DKA in T2D, and importantly the clinical presentation differences compared to T1D. The following is a combination derived from consumer websites. ^33-34

• Lethargy and weakness (can be the primary symptom)

• Abdominal pain (usually milder compared  T1D – unless pancreatitis is suspected)
• Possibly an increase in breathing rate
• Nausea and vomiting 
• Flushed cheeks
• Dizziness and light headedness (more common in T2D)
• Fruity/acetone breath (usually mild or difficult to identify)
• Increased thirst and urination (however in older people the sensation of thirst is often reduced and if an acute kidney injury is also occurring oliguria maybe present especially in older people with chronic kidney disease)
• Confusion and disorientation (more common in T2D)
• Hyperglycaemia (more markedly raised in T2D – unless prescribed SGLT2i)
• Tachycardia and hypotension

It is vital that diabetes healthcare professionals be cognisant of the incidence and risk factors pertaining to clients with T2D developing DKA.  Importantly, being prescribed a SGLT2i is not the only identifiable factor.  The ADA (2021) suggests that any condition leading to a deterioration in glycaemia requires more frequent blood glucose monitoring and recommends testing for ketones in those individuals deemed as ketosis-prone.³ Sick day management education and individualised plans should reflect this risk, especially when other clinical factors (identified above) are present.  As outlined in the ADEA sick day management guidelines, people with T2D at greater risk of developing DKA need to be made aware of the risk and symptoms (Including abdominal pain, nausea, vomiting and fatigue) and when to seek urgent medical care if DKA is suspected.³⁵   Education pertaining to ketone testing is also appropriate.³⁵