Diabetic Ketoacidosis and Hyperglycaemic Hyperosmolar State
The hallmark of diabetes is a raised plasma glucose resulting from an absolute or relative lack of insulin action. Untreated, this can lead to two distinct yet overlapping life-threatening emergencies. Near-complete lack of insulin will result in diabetic ketoacidosis, which is therefore more characteristic of type 1 diabetes, whereas partial insulin deficiency will suppress hepatic ketogenesis but not hepatic glucose output, resulting in hyperglycaemia and dehydration, and culminating in the hyperglycaemic hyperosmolar state. Hyperglycaemia is characteristic of diabetic ketoacidosis, particularly in the previously undiagnosed, but it is the acidosis and the associated electrolyte disorders that make this a life-threatening condition. Hyperglycaemia is the dominant feature of the hyperglycaemic hyperosmolar state, causing severe polyuria and fluid loss and leading to cellular dehydration. Progression from uncontrolled diabetes to a metabolic emergency may result from unrecognised diabetes, sometimes aggravated by glucose containing drinks, or metabolic stress due to infection or intercurrent illness and associated with increased levels of counter-regulatory hormones. Since diabetic ketoacidosis and the hyperglycaemic hyperosmolar state have a similar underlying pathophysiology the principles of treatment are similar (but not identical), and the conditions may be considered two extremes of a spectrum of disease, with individual patients often showing aspects of both.
Pathogenesis of DKA and HHS
Insulin is a powerful anabolic hormone which helps nutrients to enter the cells, where these nutrients can be used either as fuel or as building blocks for cell growth and expansion. The complementary action of insulin is to antagonise the breakdown of fuel stores. Thus, the release of free fatty acids from adipose tissue depots (lipolysis) is normally restrained by the action of insulin. When a person is fasting for a long time, insulin levels will fall and lipolysis will occur; uncontrolled diabetes has often been compared to a state of accelerated fasting.
The resulting flux of free fatty-acids is then either metabolized by 'energy-hungry' tissues, e.g. muscle,
Figure 1. In Diabetic Ketoacidosis, the (complete) absence of insulin leads to unrestrained lipolysis. The resulting high levels of free fatty acids (FFAs) are no longer used for gluconeogenesis, but are oxidized to ketones instead. In the Hyperglycaemic Hyperosmolar State, insulin levels are still sufficient to restrain lipolysis and to prevent ketogenesis, but severe hyperglycaemia ensues.or used by the liver to make glucose (gluconeogenesis) which in its turn will be used as fuel for certain tissues, most notably the brain.
In diabetic ketoacidosis, the insulin level is inappropriately low in relation to the presence of fuel substrates, leading to inappropriately high rates of lipolysis (see figure 1).
The excess of free fatty acids is oxidized to ketoacids (acetoacetate and beta-hydroxybutyrate), which are the hallmark of DKA. Since lipolysis is easily inhibited at low levels of insulin, ketoacidosis only develops in the presence of severe insulin deficiency, and is therefore characteristic of type 1 diabetes.
Partial insulin deficiency may be sufficient to inhibit ketogenesis but insufficient to inhibit hepatic gluconeogenesis, which has a higher threshold for suppression by insulin. This situation is more characteristic of type 2 diabetes.
It should be appreciated that many people experience uncontrolled diabetes without progressing to a life-threatening metabolic emergency. This progression may be due to prolonged insulin deficiency, as in undiagnosed diabetes, sometime aggravated by the attempt to quench thirst with glucose-containing drinks. Alternatively, the transition to a metabolic emergency may be driven by stress due to intercurrent infection or sepsis - which in itself may cause (lactic) acidosis. The release of stress hormones such as cortisol and catecholamines and several vicious cycles (see figure 1) will further aggravate the metabolic dysregulation and acidosis.
Thus, for example, acidosis promotes vomiting, leading to progressive dehydration. Progressive dehydration leads to renal insufficiency, thus impairing renal compensation for the metabolic acidosis. This negative spiral inevitably resulted in death before the advent of treatment with insulin and fluids.
Symptoms of DKA and HHS
As can be deduced from the pathophysiology, hyperglycaemia associated with DKA can be deceptively mild in treated diabetes, sometimes as low as 12-14 mmol/l, whereas in HHS it is invariably high (figure 2). Hyperglycaemia will lead to polyuria, dehydration, weight loss, polydipsia and thirst.
Electrolyte disturbances result from loss of water usually in excess of salt loss; hypovolaemia and severe intravascular dehydration will be accompanied by tachycardia and may give rise to thromboembolic complications (such as stroke or myocardial infarction), whereas cellular dehydration may ultimately cause the hyperosmolar coma.
Figure 2. In the presence of (a little) insulin, glucose levels can become as high as 25 or 30 mmol/l with the patient experiencing relatively mild symptoms; with higher glucose levels the patient will progress to the Hyperglycaemic Hyperosmolar State. In the setting of Diabetic Ketoacidosis, patients will already become seriously symptomatic because of the acidosis at relatively low glucose levels.
The acidosis will also lead to the classical 'Kussmaul' breathing pattern, in which the patients attempts respiratory compensation for the metabolic acidosis by hyperventilation, taking deep sighing breaths.
Many electrolytes are lost to the body by polyuria and vomiting, resulting in whole-body deficiency. Salt depletion is severe but responds well to fluid replacement. Potassium deficiency is a feature of ketoacidosis due to the exchange of intracellular potassium and the intravascular hydrogen ion. Potassium is then lost in the urine, resulting in depletion of whole body potassium - although it is important to appreciate that plasma levels may be raised or normal at the time of presentation. Hypokalaemia is common in the treatment phase as potassium re-enters cells under the influence of insulin and can result in cardiac dysrythmia. Timely potassium replacement is therefore a key element of management. Phosphate deficiency is also a common problem, and may give rise to muscle weakness.
Ketones have a paralytic effect on smooth muscle cells, which may lead to gastric retention (with a gastric splash on physical examination) and profuse vomiting as well as a distended bladder. Acidosis may rarely cause abdominal pain and simulate a surgical acute abdomen. The typical smell of acetone can be detected in the patient's breath, although some people are unable to detect this smell.
Clinical differences between DKA and HHS
Diabetic ketoacidosis is the characteristic metabolic emergency of type 1 diabetes. Those affected therefore tend to be young, with either undiagnosed or insulin-treated diabetes. They manifest the features of acidosis, namely Kussmaul respiration and vomiting, and have the characteristic odour of acetone on their breath. The mental state ranges from fully alert to drowsy.
The hyperosmolar state is more commonly seen in the middle aged or elderly, and is more often associated with intercurrent illness or sepsis. Features of acidosis are lacking, but drowsiness or coma are more frequent due to the extreme dehydration, and renal dysfunction is more marked.
Classical findings in Diabetic Ketoacidosis versus the Hyperglycaemic Hyperosmolar State
|Glucose (mmol/l)||>= 14||>33|
|Ketones (urine stick)||++||+/-|
|Mental Status||Alert to coma||Drowsy to coma|
|Water Deficit (l/kg)||0.1||0.1-0.2|
Please note that many patients may have features of both DKA and HHS
Epidemiology of Diabetic Ketoacidosis
The annual incidence of DKA varies from population to population, thus reflecting both the prevalence of type 1 diabetes and the quality of primary health care. DKA used to be a characteristic presenting symptom of new cases of type 1 diabetes, but increased awareness of the condition and ready access to glucose measurement devices means that the majority of childhood cases are now detected and treated before ketoacidosis occurs.
Risk factors are onset of diabetes at a young age (< 5 years, possibly due to poor recognition of symptoms), social disadvantage, a lower body mass index and a preceding infection, whereas having a family member with type 1 diabetes has a protective effect, once again likely due to earlier detection.
This means that ketoacidosis is nowadays more frequently seen in those with established diabetes, usually in the setting of intercurrent illness or poor compliance. Poor education is also a factor; for example, patients may decide to omit insulin because they are vomiting if not specifically warned otherwise. Behavioural factors may also be involved, for example some younger patients may omit insulin deliberately to promote weight loss or to escape from home abuse. This may lead to recurrent episodes of ketoacidosis.
It is increasingly recognised that some of those with type 2 diabetes may present with ketoacidosis as well, particularly those with Afro-American or Hispanic- American ancestry. This is often referred to as Ketosis-prone type 2 diabetes or Flatbush diabetes and after the initial episode of DKA these patients can generally be transferred to treatment without insulin.
Due to improved supportive care and ready availability of insulin, mortality from DKA has fallen to less than 5% in most countries. This is still too high for an entirely preventable condition.
Epidemiology of the Hyperglycaemic Hyperosmolar State
HHS principally affects those with type 2 diabetes, either as the first presentation of the disease or in those with poor compliance to medication. However, in many developed countries HHS is becoming rare as a result of targeted screening for diabetes, increased physician awareness and the ubiquity of home glucose measuring devices. It is estimated that less than 1% of all hospitalisations for diabetes are related to HHS. However, since it often affects elderly people with comorbidities, mortality rates are still around 10%.
There is frequently a contributing factor such as a concomitant infection or other illness. Previously undiagnosed patients who experience thirst may attempt to quench it with glucose-containing drinks, thus aggravating the hyperglycaemia. Risk factors for mortality are associated illness, increasing age, hypotension, higher level of glucose, high blood urea and low pH.
Management of DKA and HHS
Tthe following is a summary of the principles of management, not a practical guide. Almost all hospitals have their own management protocols for these emergencies, and these must be followed.
As the root cause of DKA and HHS is lack of insulin effect, so the first key aim of treatment is insulin. While subcutaneous insulin may suffice in less severe cases, intravenous administration is to be preferred in more severe cases because severe dehydration and hypovolemia may interfere with the absorption of subcutaneous insulin. Use of insulin pumps must be carefully monitored by trained staff.
Commonly used regimens suggest a bolus of 0.1 IU/kg followed by a continuous infusion of 0.05-1.0 IU/kg/hr, with frequent monitoring of glucose change and adjustment of the dose when necessary. It should be borne in mind that to stop ongoing lipolysis and ketogenesis, a little insulin will already suffice, although the metabolic clearance of ketones may take longer.
Dehydration is the second key problem to address. Given the large fluid deficit (often estimated as between 7-11 litres in an adult) most treatment regimens advocate a rapid start with 1 to 1.5 liters of saline (NaCl 0.9%) in the first hour, followed by about 0.5-1.0 l/hour with careful monitoring of volume status, blood pressure and urine output. Glucose and potassium levels require regular monitoring, and potassium replacement should be instituted at a very early stage. Glucose should not be lowered too rapidly and glucose infusion is generally started to cover onging insulin infusion as the patient approaches normoglycaemia.
The third key problem, and the one most easily overlooked, is identification and treatment of the possible underlying causes of the metabolic disturbance. This includes checking lactate levels, because sometimes a mixed acidosis may occur.
As a general principle, the use of bicarbonate to correct the acidosis of DKA or of hypotonic saline to correct the hyperosmolar state are discouraged, except in exceptional circumstances and when expert advice is available. Most treatment protocols restrict the administration of bicarbonate to those with pH< 7.0. The great majority of patients will recover metabolic homeostasis with the help of standard fluid regimens.
Above all, careful monitoring is essential both clinically and by laboratory investigations (glucose, potassium, sodium, bicarbonate or arterial blood gas analysis).
Complications of treatment of DKA and HHS
Hypoglycaemia, hypokalaemia and hypophospataemia may result from over-aggressive insulin treatment.
Over-aggressive fluid replacement may precipitate fluid overload and cardiac failure in elderly or frail patients, especially in the presence of pre-existing renal insufficiency.
In some patients with DKA (particularly children with newly diagnosed diabetes) and HHS cerebral oedema may develop. This may progress rapidly and be fatal once brain stem herniation occurs. Patients should therefore be carefully monitored for changes in mental status. While the pathogenesis of cerebral oedema is unclear, a too rapid lowering of glucose and the resulting large changes in osmolarity have been implicated as a causative factor. This is another reason for careful monitoring of plasma glucose and osmolarity during treatment, and one should aim for a gradual lowering of both.
Thrombotic complications may arise in severely dehydrated individuals, and most protocols advise thrombo-prophylactic treatment in this situation.
It should never be forgotten that diabetic metabolic emergencies are both avoidable and treatable. No-one who reaches hospital in time should ever be allowed to die, and every such death should prompt careful review.
Treatment errors often arise after the initial period of intensive management as a result of relaxed vigilance.
Finally, the management of metabolic emergencies is not complete until the causes of the episode are fully understood and the appropriate patient education and other arrangements are in place to ensure that it does not recur.