ARTICLE IN PRESS

Current Paediatrics (2006) 16, 111 116

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/cupe

Diabetic ketoacidosis
Claire P. Smith
Queens Park Hospital, Haslingden Rd, Blackburn BB2 3HH, UK

KEYWORDS
Diabetic ketoacidosis; Complications; Cerebral oedema; Children; Management; Pathophysiology

Summary Diabetic ketoacidosis (DKA) remains the commonest cause of diabetes-related deaths in children and most deaths occur as a result of cerebral oedema. There are now internationally agreed guidelines for the management of DKA, which aim to optimise treatment and to minimise the risk of complications including cerebral oedema. Much of the guidance is consensus rather than evidence based, but knowledge of the pathophysiology, the evolving clinical presentation and the existing evidence supporting management strategies helps us to understand the rationale behind the consensus guidelines for the management of DKA. & 2006 Elsevier Ltd. All rights reserved.

Denition, frequency and causation

There is no agreed denition for diabetic ketoacidosis (DKA), but in practice it refers to patients who have hyperglycaemia, acidosis and ketosis, the primary cause of which is insulin deciency, either relative or absolute. In newly diagnosed patients, there is wide variation across the world in the proportion of patients presenting with DKA. The best gures for the UK were derived from a prospectively registered population-based study of patients aged less than 21 years with type 1 diabetes mellitus in the Oxford region. Two cohorts of patients who were diagnosed in 19851986 and 1990 were registered. Both groups had an identical proportion of patients (26%) who presented with DKA (pH o7.35, bicarbonate o21 mmol/l).1 This presentation was most frequent in children aged less than 5 years and it became less common with increasing age. A family history of diabetes reduced the risk whereas having parents in social classes 35 appeared to increase the risk of presentation in DKA.1
Tel.: +01254 293750; fax: +01254 293525.

In children with established diabetes, the risk of DKA is 110% per patient per year.2,3 Vulnerable children are those with poor metabolic control, peripubertal girls and those with difcult family and social circumstances. The majority of DKA episodes are thought to be due to insulin omission or treatment error.3,4 Other causes include inadequate insulin therapy during intercurrent illness, alcohol, and technical errors, e.g., a malfunctioning pen injector, a leaking insulin cartridge or insulin pump failure.

Pathogenesis
There are four main pathways that lead to DKA, namely insulin deciency, stress hormone excess, dehydration and fasting. Insulin deciency and stress hormone excess result in increased glucose production and decreased glucose utilisation. Hyperglycaemia leads to increased urinary loss of glucose, water and electrolytes. Dehydration eventually ensues. As relative insulin deciency and stress hormone excess increases, this causes enhanced lipolysis, ketogenesis and decreased ketone clearance, which leads to ketosis and a metabolic acidosis.

E-mail address: claire.smith@mail.bhrv.nwest.nhs.uk. 0957-5839/$ - see front matter & 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.cupe.2005.12.010

ARTICLE IN PRESS
112 C.P. Smith Ketone bodies are weak acids that dissociate completely at physiological pH. In DKA, the large hydrogen ion load rapidly exceeds normal buffering capacity. Kussmaul respirations develop, but despite this effort, a metabolic acidosis ensues. Ketoacidosis can be further complicated by lactic acidosis if the patient becomes shocked.

Insulin deciency (relative or absolute)

The sensitivity of various metabolic processes to insulin varies considerably, resulting in a hierarchy of actions depending upon plasma insulin concentration. Relatively high concentrations are required to stimulate glucose uptake by peripheral tissues, but only low concentrations are needed to inhibit lipolysis, gluconeogenesis and glycogenolysis. Hence, as diabetes develops or when insulin treatment is withdrawn, the rst metabolic consequence is hyperglycaemia alone. Only much later does ketoacidosis develop and this is indicative of marked insulin deciency. The degree of insulin deciency in DKA is rarely absolute,5 but combined with the second essential element, the increased drive from anti-insulin stress hormones, the patient becomes relatively insulin decient, and ketonaemia followed by ketoacidosis ensues.

Mortality and morbidity

Reported mortalities from DKA in national population-based studies in the Western world range from 0.15% to 0.31%.2,6 Cerebral oedema accounts for 5787% of all DKA deaths.2 Using the British Paediatric Surveillance Unit system of reporting, Edge et al.7 investigated the risk and outcome of cerebral oedema in the UK. The calculated risk of developing cerebral oedema was 6.8 per 1000 episodes of DKA (11.9 per 1000 in newly diagnosed patients; 3.8 per 1000 in patients with established diabetes). The mortality was 24%, and 35% of survivors were left with severe neurological sequelae.7 Other potential causes of mortality and morbidity include hypokalaemia, hyperkalaemia, hypoglycaemia, aspiration, pulmonary oedema, acute renal failure, pneumomediastinum, rhabdomyolysis, infection and other neurological complications, e.g., thrombosis.

Stress hormone excess

Increased secretion of stress hormones such as glucagon, catecholamines, cortisol, and growth hormone augment the rise in glucose and ketone bodies during insulin withdrawal. Plasma glucagon rises in parallel with the development of ketoacidosis and plasma catecholamine levels are directly related to the extent of the metabolic derangement. Thus DKA is characterised by relative insulin deciency and stress hormone excess.

Cerebral oedema
Fear of this complication, which is largely conned to children, has had a major inuence on the consensus guidelines that were drawn up to help manage DKA in children and young people.2,8 The cause of cerebral oedema is unknown, although there are many theories. It usually occurs 412 h after the initiation of treatment of DKA,9 but it can develop at any time, including before treatment commences.6,10 Vulnerability seems to be greatest for children with new onset diabetes, younger age and longer duration of symptoms.6,10 It is recognised that these factors may simply reect those who are likely to present with severe DKA.9 Potential risk factors for the development of cerebral oedema, some of which can be avoided, are outlined in Table 1. The level of evidence that supports these risk factors is variable, and overall it is not strong.2 However, considering that this complication is so formidable, it seems logical that any treatment guidelines should avoid potential risk factors where possible.

Dehydration
Once hyperglycaemia exceeds the renal glucose threshold, an osmotic diuresis occurs that causes polyuria and secondary polydipsia. When uid losses exceed intake, particularly when vomiting occurs, dehydration results, which contracts the intravascular volume and eventually causes renal impairment. This results in impaired urinary excretion of glucose, ketones and hydrogen ions.

Fasting
Fasting induces ketosis by several mechanisms including a gradual reduction of insulin levels and a simultaneous rise in counter-regulatory hormone concentrations. Fasting results in decreased peripheral utilisation of ketone bodies, compounding the pre-existing hepatic ketogenesis in the decompensated patient and resulting in severe hyperketonaemia.

Table 1
1. 2. 3. 4. 5.

Potential risk factors for cerebral oedema.

Greater hypocapnia at presentation of DKA, after adjusting for the degree of acidosis.9 Higher serum urea nitrogen at presentation of DKA9, which may simply reect greater dehydration. Severe acidosis13 and treatment with bicarbonate.9 Administration of insulin within the rst hour of intravenous uid therapy.13 An attenuated rise in measured serum sodium concentrations during therapy for DKA.9,11 (Theoretically, a rise by 12 mmol/l sodium would offset a fall of 5.5 mmol/l glucose in terms of plasma osmolality). 6. Overzealous uid administration.911,13 7. Use of hypotonic uids to rehydrate (evidence very weak).2

ARTICLE IN PRESS
Diabetic ketoacidosis 113 cause. It should be noted that a high white cell count is a common feature of DKA and so is unhelpful in detecting infection.

Clinical presentation and evaluation

The cardinal symptoms of diabetes (thirst, polyuria and weight loss) are often not complained about by patients and their families. They may have dismissed polyuria as being a normal consequence of polydipsia, and weight loss may be welcomed by teenagers or families of children who were overweight to start with. Families are much more likely to complain about vulvo-vaginitis or balanitis, secondary enuresis, lethargy, vomiting and abdominal pain. I have even seen a teenage girl present to the Ophthalmology Clinic with loss of vision secondary to cataracts. Sadly, diabetes was not diagnosed until she presented with very severe DKA. Therefore it is always helpful to ask about thirst, polyuria and weight loss directly to try to avoid delay in diagnosis and progression to DKA. The length of history is very important in determining the severity of DKA, particularly with respect to the symptoms of deep breathing, drowsiness and vomiting. Symptoms that suggest a precipitating cause, such as infection, should also be sought. Examination should focus on signs that help to determine the severity of DKA, and those that suggest a precipitating cause. The assessment of the degree of dehydration can be difcult with a tendency to overestimate the severity, possibly because weight loss precedes dehydration. Early Kussmaul respirations can be easily missed by inexperienced staff, and my advice to trainees is that if the patient is breathing more deeply than the examiner, they have air hunger. It is the depth rather than the rate of breathing that is striking, with an absence of increased work of breathing. The level of consciousness needs to be assessed and fundi examined. Evidence of paralytic ileus should be looked for. Infection as a precipitating cause can be difcult to detect clinically, even with the help of investigations. Fever is an unreliable indicator of infection because it is commonly absent in DKA. Its presence, however, is indicative of infection. Overall, it is wise to have a low threshold to treat very sick patients with antibiotics. Lastly, it is important to weigh patients if possible, even if they are very sick. This allows accurate calculations for uid and insulin administration. It also allows the team to check how accurate their assessment of dehydration was by reweighing the patient after recovery.

Management
This section is based on internationally agreed Consensus Guidelines 2000,8 and the level of evidence supporting these guidelines can be found in a review article by Dunger et al., written on behalf of the European Society of Paediatric Endocrinology and the Lawson Wilkins Pediatric Endocrine Society.2

Oxygen therapy
An immediate assessment of airway, breathing and circulation is needed. Oxygen (100%) by face mask should be given to all but the very mildest cases of DKA. It should be continued during the rst hours of treatment, even if the oxygen saturation is 100% in air. This is important because patients with DKA suffer intracellular phosphate depletion and changes in serum phosphate often follow a similar pattern to that of serum potassium. One of the consequences of hypophosphataemia is that it may result in reduced levels of 2,3-diphosphoglycerate in red blood cells. This creates an increased afnity of the haemoglobin molecule for oxygen resulting in a shift of the oxygenhaemoglobin dissociation curve to the left. At the same time as serum phosphate is falling, intravenous uids and insulin will be correcting the metabolic acidosis, which will have had the opposite and protective effect of shifting the oxygen haemoglobin dissociation curve to the right, thereby increasing the availability of oxygen to the brain and other tissues. Some have advocated phosphate replacement because of this problem, but prospective studies have failed to show signicant clinical benet.2 Despite this, an understanding of the pathophysiology emphasises the importance of oxygen therapy.

Resuscitation
Patients with peripheral circulatory failure should be given 0.9% saline 10 ml/kg intravenously over 1030 min. This can be repeated until the peripheral circulation has improved, but extreme caution is required when exceeding 30 ml/kg because of the fear of cerebral oedema. The consultant should always be involved in these decisions.

Investigations
Table 2 outlines the initial investigations in DKA, which help to assess the severity and to establish any precipitating

Table 2

Initial investigations in DKA. Establishing cause Blood and urine cultures Chest X-ray Leukocytosis is a feature of DKA; it does not indicate infection

Assessment of severity Blood glucose Urea and electrolytes, creatinine Capillary, venous or arterial blood gas Haemoglobin and packed cell volume Urinary glucose and ketones

ARTICLE IN PRESS
114 A nasogastric tube should be inserted if the patient is vomiting, particularly if he/she is drowsy. Very sick patients should be catheterised as well to ensure accurate monitoring of urinary output. over the rst 48 h. Deficit mls estimated % dehydration e:g: 10 body weight kg 10. Maintenance (24 h): 100 ml/kg for the rst 10 kg, 50 ml/kg for the second 10 kg and 20 ml/kg for subsequent kilograms. Fluid calculation to replace decit over 48 h: Hourly rate ml=h 2 24 h maintenance deficit resuscitation fluid , 48 C.P. Smith

Clinical observations and monitoring

Careful and frequent clinical monitoring to detect warning signs of complications is of paramount importance. Neurological observations, pulse and respiratory rates, and blood pressure should be monitored at least hourly. An accurate record of uid input and output is essential. Electrocardiogram (ECG) monitoring is required in severe DKA. Changes in the size of the T wave in lead II is a good indicator of the potassium status. It is essential that children with severe DKA are managed in a setting where there are sufcient experienced nurses to provide high-dependency care. There needs to be clear written guidelines and ready access to laboratories. A consultant paediatrician who is experienced at managing DKA should oversee the management.

Type of uid
Saline (0.9%) is recommended at the start of intravenous therapy until the blood glucose falls to 12 mmol/l. Then 0.45% saline with glucose 5% can be used.

Potassium
Total body potassium is always substantially depleted in DKA, and the major loss is from the intracellular pool as a result of hypertonicity, insulin deciency and exchange for hydrogen ions within the cell. Serum potassium concentrations may be low, normal or high at presentation, but intravenous uids and insulin administration will drive potassium back into the cells resulting in a rapid fall in serum potassium. Therefore, potassium replacement should be started as soon as resuscitation is completed provided anuria is not reported by the family, the ECG does not show elevated T waves, and the serum potassium is not elevated (early results may be available using modern blood gas machines). If serum potassium (unhaemolysed) exceeds 5.5 mmol/l, potassium replacement should be withheld for 30 min after which the measurement should be repeated urgently. Potassium chloride 20 mmol should be added to every 500 ml bag of uid. Occasionally more is required to maintain safe serum potassium concentrations, but this is generally seen in patients who are desperately ill and who may already be on paediatric intensive care unit (PICU).

Rehydration
The objectives of rehydration are: (a) restoration of circulating volume, (b) replacement of sodium and water, (c) the restoration of the glomerular ltration rate with enhanced clearance of glucose and ketones from the blood, and (d) the avoidance of cerebral oedema. There is no strong evidence that shows superiority of any uid regimen over another, but there is supportive evidence that suggests that rapid uid replacement with hypotonic uids is associated with an increased risk of cerebral oedema. There is good evidence that a less rapid uid decit correction with isotonic or near isotonic solutions results in earlier reversal of acidosis. However, the use of large amounts of 0.9% saline has been shown to be associated with the development of hyperchloraemic metabolic acidosis.2 Even the contributors to the Consensus Guidelines 20008 could not reach agreement about the best rehydration uid regimen, and they proposed two models for uid correction that could be used. In the more recent consensus statement,2 the evidence was considered again, and subsequently the British Society of Paediatric Endocrinology and Diabetes (BSPE) revised its recommendations about the management of DKA.12,13 Locally, we were unhappy that the UK recommendations calculated maintenance requirements according to age rather than weight. My colleagues and I felt that it was important to have a consistent approach for calculating maintenance requirements for all conditions. Therefore we modied the BSPE guidelines accordingly.

Insulin
Blood glucose falls rapidly with intravenous uids alone in the early stages of treatment. There is now evidence that insulin treatment should be delayed until after the rst hour because earlier treatment was found to be a risk factor for the development of cerebral oedema.14 Insulin therapy is essential thereafter to correct hyperglycaemia, inhibit lipolysis, ketogenesis and glycogenolysis, and to counteract the excessive levels of stress hormones. It is now accepted that low-dose intravenous insulin infusions are the best way to achieve a predictable and gradual decrease in blood glucose. The recommended dose of 0.1 U/kg/h has been shown to achieve supraphysiological plasma insulin concentrations of 100200 mU/l, which are sufcient to correct all the metabolic processes that create DKA.15 The aim of treatment is to achieve a gradual fall in blood glucose, which does not exceed 5 mmol/l/h. If blood glucose is falling too rapidly, the insulin dose can be reduced to

Fluid calculation
When calculating rehydration uid requirements the amount already given as resuscitation uid should be subtracted from the total decit required.13 The decit should be worked out using a maximum of 10% dehydration for initial calculations and this should be replaced at a constant rate

ARTICLE IN PRESS
Diabetic ketoacidosis 0.05 U/kg/h. Once blood glucose reaches 12 mmol/l, the intravenous uids should be changed to 0.45% saline and 5% dextrose. It is very important that insulin is not stopped and that the dose is not reduced below 0.05 U/kg/h because only insulin can switch off ketone production. If blood glucose is falling too rapidly or levels are too low, more glucose should be added to the uids to allow insulin therapy to be continued safely. 115 associated with increased risk of cerebral oedema. Theoretically, serum sodium should rise by 2 mmol for every 5.5 mmol fall in blood glucose. This will result in a slower fall in serum osmolality according to the following equation: Calculated serum osmolalitymOsm 1:86 Na K blood glucose urea 10all in mmol=l.

Bicarbonate
There is no evidence that bicarbonate treatment is either necessary or safe in DKA and it should not be used in the initial resuscitation. The only indication for bicarbonate therapy may be profound acidosis that is causing decreased cardiac contractility and peripheral vasodilatation resulting in further impairment of tissue perfusion. Potential problems associated with bicarbonate therapy are paradoxical central nervous system (CNS) acidosis, hypokalaemia, altered calcium ionisation and excessive osmolar load.

Cerebral oedema
The symptoms and signs of cerebral oedema include headache, changes in the neurological status, e.g. increased irritability, restlessness and drowsiness. Specic features such as cranial nerve palsies, abnormal posture, rising blood pressure and slowing of pulse are usually late signs. If cerebral oedema is suspected, the following measures should be undertaken immediately: Inform senior medical staff. Exclude hypoglycaemia. Intravenous mannitol 0.251.0 g/kg over 20 min can be given but its efcacy is questionable. This needs to be given as soon as warning signs occur. Restrict intravenous uids to 2/3 maintenance and replace over 72 h rather than 48 h. Hypertonic saline (3%) 510 ml/kg over 30 min has been proposed as an alternative to mannitol. Intubation and ventilation may be required whilst awaiting transfer to PICU. A repeat dose of mannitol may be required after 2 h if there is no improvement.

Antibiotics
There should be a low threshold for treating patients with severe DKA with broad-spectrum antibiotics because indicators of infection are often masked by DKA.

Monitoring progress
Frequent clinical assessment and review of uid balance is required until the DKA improves. Capillary blood glucose should be measured hourly and electrolytes, urea, blood glucose and blood gases should be repeated every 24 h until the acidosis corrects. If acidosis persists, it is worth considering whether more resuscitation uids are required or whether hyperchloraemic metabolic acidosis has supervened. As treatment progresses, it is worth considering whether the original assessment of dehydration was correct. If the patient appears to be rehydrating more quickly than expected, it is worth considering whether the infusion rate should be decreased. This is particularly relevant when blood glucose approaches the renal threshold of 810 mmol/ l when one can anticipate that the osmotic diuresis will abate. Despite the depletion of total body sodium in DKA, the elevated osmolality in the hyperglycaemic state results in a dilutional effect on the measured sodium. If there are concerns about abnormal measured sodium levels, the following correction can be calculated: Corrected sodium measured sodium 1:6 blood glucose 5:5=5:5. Physiological calculations suggest that the serum sodium should be lowered by 1.6 mmol/l for every 5.5 mmol/l elevation in serum glucose. Serum sodium usually rises as the blood glucose falls and there should be concern if it does not. There is evidence that an attenuated rise in serum sodium concentrations during therapy of DKA may be

Prevention
Increasing public awareness of the symptoms and signs of diabetes is the most important way to achieve early diagnosis. Patients with established diabetes and their families need clear guidelines for the management of illness and high blood glucose levels. Guidance needs to be customised according to the ability and understanding of individual patients and their families. Missed insulin injections is a temptation for children of all ages and parents need to be reminded regularly about the need to supervise injections. Recurrent DKA is a challenging problem and Edge et al. reported that 4.8% patients accounted for 22.5% of all DKA episodes over a 3-year period.7 A combination of education, psychosocial assessment and treatment, and increased adult supervision is usually required.

References
1. Pinkney JH, Bingley PJ, Sawtell PA, Dunger DB, Gale EAM. Presentation and progress of childhood diabetes mellitus: a prospective population-based study. Diabetologia 1994;37: 704.

ARTICLE IN PRESS
116
2. Dunger DB, Sperling MA, Acerini CL, et al. ESPE/LWPES consensus statement on diabetic ketoacidosis in children and adolescents. Arch Dis Child 2004;89:18894. 3. Smith CP, Firth D, Bennett S, Howard C, Chisholm P. Ketoacidosis occurring in newly diagnosed and established diabetic children. Acta Paediatr 1998;87:53741. 4. Morris AD, Boyle DIR, McMahon AD, Greene SA, MacDonald TM, Newton RW. Adherence to insulin treatment, glycaemic control, and ketoacidosis in insulin-dependent diabetes mellitus. The DARTS/MEMO Collaboration. Lancet 1997;350:150510. 5. Schade DS, Alberti KGMM, Johnston DG. Diabetic coma, ketoacidotic and osmolar. Albuquerque: University of New Mexico Press; 1981. 6. Edge JA, Ford-Adams ME, Dunger DB. Causes of death in children with insulin dependent diabetes 199096. Arch Dis Child 1999;81:31823. 7. Edge JA, Hawkins MM, Winter DL, Dunger DB. The risk and outcome of cerebral oedema developing during diabetic ketoacidosis. Arch Dis Child 2001;85:1622. 8. ISPAD. Consensus guidelines. www.ispad.org, 2000.

C.P. Smith
9. Glaser N, Barnett P, McCaslin I, et al. Risk factors for cerebral edema in children with diabetic ketoacidosis. N Engl J Med 2001;344:2649. 10. Bello FA, Sotos JF. Cerebral oedema in diabetic ketoacidosis in children. Lancet 1990;336:64. 11. Harris GD, Fiordalisi I, Harris WL, Mosovich LL, Finberg L. Minimizing the risk of brain herniation during treatment of diabetic ketoacidemia: a retrospective and prospective study. J Pediatr 1990;117:2231. 12. Diabetes UK. Guidelines for the management of diabetic ketoacidosis in children and adolescents. www.diabetes.org.uk, 2001. 13. BPSED recommended DKA guidelines. April 2004. www.bpsed. org.uk/professional/guidelines 14. Edge JA, Jakes R, Roy Y, Widmer B, et al. The UK prospective study of cerebral oedema complicating diabetic ketoacidosis. Arch Dis Child 2005;90:A2. 15. Schade DS, Eaton RP. Dose response to insulin in man: differential effects on glucose and ketone body regulation. J Clin Endocrinol Metab 1977;44:103853.