SGLT-2 inhibitor associated euglycaemic diabetic ketoacidosis in an orthopaedic trauma patient
- 1 Department of Trauma and Orthopaedic Surgery, NHS Grampian, Aberdeen, UK
- 2 School of Medicine Medical Sciences and Nutrition, University of Aberdeen, Aberdeen, UK
- Correspondence to Dr Duncan Taylor Ritchie; duncan.ritchie@nhs.scot
Abstract
Euglycaemic diabetic ketoacidosis is a serious but rare adverse effect of treatment with sodium-glucose cotransporter-2 (SGLT-2) inhibitors. A man in his 60s with type 2 diabetes mellitus underwent total hip replacement for an intracapsular neck of femur fracture. His SGLT-2 inhibitor was continued perioperatively and blood glucose levels were normal throughout the admission. A diagnosis of severe euglycaemic diabetic ketoacidosis was made in the operating theatre which required treatment in a critical care unit. This resulted in increased morbidity due to decreased postoperative mobilisation and a new requirement for subcutaneous insulin. This case highlights the need for withholding SGLT-2 inhibitors in patients admitted for emergency surgery and a need for regular ketone monitoring in these patients, even in the context of normoglycaemia.
Background
Diabetic ketoacidosis (DKA) is a serious complication of diabetes mellitus. Admission to hospital for DKA has an incidence of 35.84 per 1000 person-years.1 The diagnostic criteria for DKA according to the Joint British Diabetes Societies for Inpatient Care (JBDS-IC) is a triad of blood glucose >11 mmol/L or known diabetes, blood ketone concentration >3 mmol/L and bicarbonate concentration of <15 mmol/L and/or venous pH <7.3.2 The classification of acidosis or the calculation of the anion-gap does not feature as part of the formal diagnosis of DKA. However, DKA would result in metabolic acidosis with raised anion-gap.3 In patients with metabolic acidosis with raised anion-gap, consideration should be given to DKA as a possible cause. Inpatient diabetic protocols, such as those at our own centre, typically do not include regular ketone monitoring. In our centre, diabetic patients have regular monitoring of capillary blood glucose levels only, with hyperglycaemia (>14 mmol/L) triggering a measurement of ketones. This is generally appropriate for type 1 diabetic patients for whom DKA is associated with insulin deficiency and therefore marked hyperglycaemia. In some patients, however, DKA can occur in the context of normal blood glucose levels, known as euglycaemic DKA (EuDKA). This is a rare feature associated with treatment with sodium-glucose cotransporter-2 (SGLT-2) inhibitors. SGLT-2 inhibitors decrease insulin secretion and increase glucagon levels. This increases the glucagon-insulin ratio which leads to increased gluconeogenesis, lipolysis and ketogenesis. SGLT-2 inhibitors reduce renal reabsorption of glucose via inhibition of SGLT-2 in the proximal tubules leading to glycosuria and reduction in blood glucose level. This combination of pharmacodynamics can lead to the potential for ketosis (and subsequent ketoacidosis) in the absence of hyperglycaemia; however, the exact mechanism has not been fully established.4 5 Certain triggering events can increase the risk of development of DKA, namely conditions leading to restricted food intake or severe dehydration, sudden reduction in insulin, increased insulin requirements due to acute illness, alcohol abuse and surgery.6 In this report, we describe a case of a patient with type 2 diabetes mellitus managed with an SGLT-2 inhibitor who during an emergency orthopaedic procedure was found to be in severe EuDKA. In this case, a traditional procedure for monitoring diabetic patients led to delayed diagnosis and subsequent morbidity from which there are important learning points.
Case presentation
A man in his early 60s with a history of type 2 diabetes mellitus, knee osteoarthritis and cervicalgia presented with left hip pain following a fall tripping over a kerb while walking his dog. He was unable to weight-bear postinjury and was assisted at home by his wife. He was positioned in a recliner chair at home where he stayed for 30 hours before phoning an ambulance. On examination, he was tender over the left hip and his left leg was shortened and externally rotated. Plain radiographs taken on admission demonstrated a left intracapsular neck of femur fracture.
His preadmission regular medications are displayed in table 1. These medications were continued during the perioperative period, with the exception of glimepiride which was withheld on the morning of surgery in accordance with local guidance in use at the time.
Preadmission regular medications
| Drug | Dose | Frequency |
| Empagliflozin | 10 mg | Once a day |
| Metformin | 1 g | With breakfast and evening meal |
| Glimepiride | 3 mg | Once a day |
| Sertraline | 50 mg | Once a day |
Despite the patient’s comorbidities, his preinjury function was high, mobilising with no walking aids and being active outdoors regularly. Therefore, it was planned to perform a total hip replacement. On day 2 of admission, he was fasted in anticipation of going to the emergency operating theatre; however, the procedure was cancelled due to theatre pressures, and he was given oral food and drink early evening. His procedure was cancelled again on day 3 after fasting and he was operated on day 4. Over this period, he slept more than usual. He also had a new oxygen requirement of 3 L/min and tachycardia of 110 beats per minute. Chest radiograph and infection markers did not indicate an infectious cause. On admission, the patient was euglycaemic, with capillary blood glucose levels 4–11 throughout. On admission, his HbA1c was 103 mmol/mol indicating poor diabetic control. HbA1c 2 years prior to admission was 132 mmol/mol. The patient had no pre-existing diabetic retinopathy, neuropathy or chronic kidney disease. In theatre, a severe high anion-gap metabolic acidosis was discovered. A summary of admission and intraoperative blood results is displayed in table 2.
Relevant blood results on admission and intraoperative
| On admission | Intraoperative | |
| Sodium (mmol/L) | 139 | 139 |
| Potassium (mmol/L) | 4.3 | 4.7 |
| Chloride (mmol/L) | 102 | 109 |
| Bicarbonate (mmol/L) | 17 | 9.2 |
| pH | 6.96 | |
| pCO2 (kPa) | 5.45 | |
| pO2 (kPa) | 13.8 | |
| Base excess | −22 | |
| Blood glucose (mmol/L) | 8.3 | 11.3 |
| Lactate (mmol/L) | 0.5 | |
| Ketones (mmol/L) | 4.9 |
Differential diagnosis
EuDKA: most likely diagnosis as the JBDS-IC criteria are met. He is (1) a known diabetic with (2) elevated ketones and (3) severe high anion-gap metabolic acidosis.
Starvation ketoacidosis: this is possible given repeated fasting for theatre; however, food was consumed each day. Starvation ketoacidosis usually causes mild ketoacidosis which makes this diagnosis less likely.7
Alcoholic ketoacidosis: unlikely given no reported history of alcohol excess.
Lactic acidosis: excluded by normal lactate level.
Treatment
DKA is a medical emergency and treatment was initiated as soon as possible after the initial diagnosis was made in the anaesthetic room. This followed the local DKA management pathway involving intravenous insulin and 0.9% sodium chloride. As this patient was euglycaemic, 10% glucose + 20 mmol potassium chloride infusion was also started to prevent iatrogenic hypoglycaemia and hypokalaemia.
Repeat blood gases showed an improvement in pH from 6.9 to 7.01 after initial fluid administration. A multiprofessional decision was taken to continue with the operation as the patient was already anaesthetised and had suffered significant delays to the theatre. He underwent total hip arthroplasty without any surgical complications. Following the operation, the patient was admitted to the intensive care unit (ICU) for critical care management of his severe DKA. During the first 4 days postoperative he remained ketotic but eventually stabilised on variable-rate insulin. He was stepped down to the endocrine ward. He did not receive physiotherapy until postoperative day 4. He developed a urinary tract infection (UTI) and was found to have glucosuria, this was managed with oral antibiotics.
Outcome and follow-up
From an endocrine perspective, preoperatively his HbA1c was 103 mmol/mol. On discharge, he is now taking metformin 1 g two times per day and insulin (Levemir 16 units at night and Novorapid 8–12 units before meals) resultantly his glycaemic control prior to discharge was excellently controlled. At 3 months, the diabetologist described a ‘massive overhaul of diet and lifestyle’. The HbA1c had fallen to 64 mmol/mol from 103 mmol/mol and Body Mass Index 34.2 from 37.3. Overall, in future this new regimen may lead to improved diabetic control. From a quality-of-life perspective, he now must inform the Driver and Vehicle Licensing Agency and modify his life around his insulin use.
From an orthopaedic perspective, at the time of discharge, he required a Zimmer frame to mobilise. Prior to his fall, he required no walking aids apart from occasionally one stick for long distances. At 4 months, he was mobilising with one walking stick indoors and outdoors, managing short distances only.
Discussion
The case literature describes two similar previous cases in orthopaedic trauma patients. One in a patient with L1 vertebral, left ulnar and right radius fractures8 and another in a patient with a femoral diaphyseal fracture.9 To our knowledge, this is the first case describing EuDKA associated with SGLT-2 inhibitor use in a patient undergoing surgical fixation of a hip fracture.
After hip fracture, significantly more patients require walking aids compared with prefracture state10 11 and they are more likely to be living in an institution after previously living independently.10 11 The requirement for walking aids is increased by lengthening postoperative immobility12–14 and a delay of >24 hours between the time of injury and surgical management.13–15 Comorbid diabetes mellitus is associated with a longer hospital stay, cardiac postoperative complications and the development of pressure ulcers16 but not independently associated with poorer long-term mobility.16 Diabetic patients undergoing hip fracture surgery have higher mortality.17 National guidelines dictate diabetic patients should be operated first on the list, and a maximum of 6 hours fasting for theatre.18 In this case, the patient was delayed extensively and had prolonged periods of fasting. This demonstrates the importance of adhering to established guidelines and ensuring diabetic patients undergoing hip fracture surgery are prioritised to reduce the risk of increased morbidity and mortality.
Current guidance from the Scottish Standard of Care for Hip Fracture Patients and British Orthopaedic Association Standards recommend surgery for hip fractures take place within 36 hours,19 20 while the American Academy of Orthopaedic Surgeons recommend surgery within 24–48 hours.21 The National Institute for Health and Care Excellence recommend surgery take place either on the day or the day after injury.22 This is summarised in figure 1. There are multiple recognised factors that contribute to delays to surgical fixation. These include delayed diagnosis,23 delayed consent,23 24 lack of theatre time,23 25 26 delays to facilitate total hip replacement23 and the patient being medically unfit for surgery.23–25 27 Delays to surgery for hip fracture fixation are associated with both increased mortality28 and morbidity29 as summarised in table 3.
Summary of guidance of timing for surgical repair of hip fractures. Times shown are the maximum times recommended between admission and surgical repair.
Outcomes affected by delay to surgical fixation of hip fractures beyond 48 hours
| Outcome | Effect of delay to surgery |
| Mortality | 41% increased mortality at 30 days.28 |
| Development of pressure ulcers | Increased risk of development of pressure ulcers. Absolute risk of 11% for delayed surgery and 6% for non-delayed.29 |
| Medical complications (including thromboembolism, myocardial infarction and pneumonia) | Increased risk. Absolute risk of complications of 17% for delayed surgery compared with 8% for non-delayed.29 |
In the present case, a delay to operative management from initial injury increased the risk of poorer long-term postoperative mobility. This has been further exacerbated by the EuDKA requiring ICU admission which delayed postoperative mobilisation. The Scottish Standard of Care for Hip Fracture Patients aims for all patients to have begun mobilisation by the end of the first day after surgery.19 In our case, the patient did not mobilise or receive physiotherapy until postoperative day 4. This has likely resulted in a poorer functional outcome for the patient who now mobilises with the aid of a Zimmer frame. This represents a significant drop in function from previously mobilising independently, only using a single stick for long distances. Diabetes mellitus is associated with an increased risk of the development of respiratory, urinary tract, gastrointestinal and soft tissue infections.30 In patients undergoing arthroplasty, comorbid diabetes mellitus is associated with an increased incidence of postoperative periprosthetic joint infection.31 32 SGLT-2 inhibitors are themselves associated with an increased risk of genital and UTIs.33 34 This includes the rare risk of severe infection such as Fournier’s gangrene.35 The mechanism for this increased risk of infection is believed to be due to glycosuria caused by SGLT-2 inhibitor action providing a substrate for bacterial growth.36 It is known that a distant focus of infection, such as the urinary tract, can lead to periprosthetic joint infection via haematogenous spread.37 However, it is not clear if the increased risk of infections associated with SGLT-2 inhibitor use leads to an increase in periprosthetic infection after surgical management of hip fracture and this warrants further study. During admission, our patient developed a UTI; however, it is not possible in this single case to determine a causal relationship with SGLT-2 inhibitor use. It is known that reduced endogenous insulin secretion is associated with an increased risk in the development of EuDKA,4 unfortunately insulin secretion was not evaluated in this case.
In this case, the identification of EuDKA was delayed. This was in part due to clinical guidelines in use at our centre that at the time did not recommend ketone testing in the context of normal or low blood capillary blood glucose readings. As introduced previously, data regarding this rare but potentially serious adverse effect of SGLT-2 inhibitors is becoming more apparent with the Medicines and Healthcare products Regulatory Agency releasing a Drug Safety Update in 2016.6 National guidance now includes information on the management of patients taking SGLT-2 inhibitors in the perioperative period. Namely, we have reviewed the Handbook of Perioperative Medicines38 and a diabetes mellitus specific guideline from the Centre for Perioperative Care.18 For elective procedures, SGLT-2 inhibitors should be omitted the day before and on the day of surgery.18 38 For emergency surgical admissions, SGLT-2 inhibitors should be stopped on admission and blood ketones checked daily.18 38 Some sources suggest a longer withdrawal period for some SGLT-2 inhibitors,39 given the nature of trauma admissions longer withdrawal periods prior to surgery are seldom possible. Delays to theatre for the neck of femur fractures have been proven to increase morbidity and mortality.28 The benefits of the decision to delay theatre for optimisation of medications must be carefully balanced against the risks. For example, extensive research has gone into anticoagulant medications perioperatively.40 As SGLT-2 inhibitor associated EuDKA with regard to the neck of femur fractures is not widely discussed in the literature, the timing of surgery in relation to stopping SGLT-2 inhibitors is an area for further research. Centre for Perioperative Care guidance recommends all diabetic patients who are admitted in an emergency should have blood ketones checked regardless of if SGLT-2 inhibitors are used.18 In the postoperative period, in all patients admitted with SGLT-2 inhibitors, blood ketones should continue to be monitored, even in the context of normal blood glucose.18 38 The Endocrine Society recommends all patients with type 2 diabetes mellitus who miss more than one meal should be started on a variable-rate insulin infusion.41 Treatment with SGLT-2 inhibitors can be restarted once the patient is medically stable,18 38 eating and drinking normally,18 38 has normal ketones18 38 and had any volume depletion corrected.38 SGLT-2 inhibitors should not be restarted until variable-rate insulin infusion is stopped.38 A summary of this guidance is displayed in figure 2.
Summary of guidance on use of sodium-glucose cotransporter-2 (SGLT-2) inhibitors in the perioperative period. Source: authors’ own work.
Patient’s perspective
The whole experience did cause me significant trauma and I was very surprised to learn postoperative just what had happened. On reflection, I was not fully aware of my condition and controlling my diabetes. Postoperation that has been addressed and my levels have improved through medication and now my insulin intake.
Just to note that the care and attention I received throughout my 2 week hospital stay from all concerned was first class and they are a credit to the NHS.
Learning points
-
Patients taking sodium-glucose cotransporter-2 (SGLT-2) inhibitors requiring emergency surgery should have their SGLT-2 inhibitor withheld on admission.
-
Patients taking SGLT-2 inhibitors at presentation require ketone monitoring, even in the absence of hyperglycaemia.
-
Active measures should be taken to minimise the perioperative risk for diabetic patients, these include minimising prolonged fasting for theatre and prioritisation for the first slot on the operating list.
Ethics statements
Patient consent for publication
Footnotes
-
Twitter @duncanrit
-
Contributors DTR: planning, literature review, manuscript writing and editing. JD: planning and manuscript editing.
-
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
-
Case reports provide a valuable learning resource for the scientific community and can indicate areas of interest for future research. They should not be used in isolation to guide treatment choices or public health policy.
-
Competing interests None declared.
-
Provenance and peer review Not commissioned; externally peer reviewed.
- © BMJ Publishing Group Limited 2022. No commercial re-use. See rights and permissions. Published by BMJ.
References
Use of this content is subject to our disclaimer