Introduction Paediatric hypoglycaemia is a relatively common medical emergency. To allow identification of the underlying cause, investigations need to be performed urgently prior to treatment being given. Careful consideration is needed to ensure correct patient selection, as inadequate investigations have further cost and patient safety implications.
Methods 49 cases of proven or suspected hypoglycaemia (glucose ≤2.6 mmol/L) were identified via the laboratory. Clinical notes, laboratory investigations and results were reviewed.
Results Only 41% of patients (15 neonates, 5 children) required investigation with a ‘Hyposcreen’. Of these 20 patients, 3 had no investigations performed. In the remaining patients the cause for hypoglycaemia was identifiable, but 6 had investigations regardless. In total 23 patients had ‘Hyposcreen’ but only 2 were complete. Intermediary metabolites (96%), lactate (100%), cortisol (100%), insulin (83%) and growth hormone (87%) were taken most commonly with urine samples (52%) and ammonia (30%) taken least often. 40% cortisol, 29% insulin and 56% intermediary metabolite results were abnormal affecting 10 patients, but only 5 had follow-up. A total of £6977 was spent on investigations, of which £1630 has subsequently been found to be unnecessary. If investigations in the 23 children had been complete, this would have totalled £2700 of unnecessary expenditure.
Conclusions Investigations for hypoglycaemia are generally incomplete (91%) or inappropriate (21%). This has major cost implications for both the National Health Service and the individual who is investigated inadequately or incorrectly. We need national evidence-based guidance for investigation thresholds and normal ranges to help avoid inappropriate investigations and delay in diagnosis.
- HEALTH ECONOMICS
Statistics from Altmetric.com
Paediatric hypoglycaemia is a relatively common medical emergency with potentially serious consequences. It has been well recognised that prolonged periods of hypoglycaemia can lead to neurological damage and death.1 ,2 Clearly treatment is imperative but where the underlying cause is not already known, urgent investigation is required to aid diagnosis, and therefore help prevent further episodes. If these investigations are done appropriately, and at the time of hypoglycaemia, this can reduce the need for future tests and hospital admissions.
There are many known causes of paediatric hypoglycaemia3 ,4 (box 1), and investigation is aimed at diagnosing underlying endocrine and metabolic disorders to ensure correct future management. Investigations include blood and urine tests, and collectively are generally termed a ‘Hyposcreen’5 (table 1). There are various guidelines stating what constitutes an appropriate ‘Hyposcreen’, but a comprehensive overview can be found at http://www.metbio.net.3
Growth hormone deficiency
Disorders of amino acid and organic acid metabolism
Disorders of gluconeogenesis
Disorders of carbohydrate metabolism
Disorders of fatty acid oxidation and carnitine transport
Idiopathic ketotic hypoglycaemia
Drug related: insulin, alcohol, aspirin, chemotherapy
Poor feeding, SGA/IUGR
IUGR, intra-uterine growth restriction; SGA, small for gestational age.
Despite the importance, investigation of hypoglycaemia is not always required, and varies in different age groups of children. Neonates are prone to hypoglycaemia due to decreased reserves, therefore neuroprotective measures allow for increased utilisation of ketone bodies and lactate as alternative energy sources.6 The cause of early neonatal hypoglycaemia may be clearly identifiable such as poor feeding or prematurity, and provided the hypoglycaemia responds as expected to treatment, further investigations are unnecessary.
The current laboratory cost of investigating children with hypoglycaemia as per recommendations is approximately £450.5 In these times of austerity, awareness of cost saving is increasing as National Health Service (NHS) services are stretched. One also has to balance individual patient safety in addition to the cost of recurrent investigation.
In reality complete and comprehensive investigation can be difficult to perform in children in a time pressured situation. The results of investigations for hypoglycaemia in a district general hospital were therefore reviewed to see whether investigations performed were both appropriate and complete, as well as whether abnormal results were actioned.
Patients aged ≤16 years who were recorded as having an episode of proven or suspected hypoglycaemia over a 1-year period in a district general hospital were selected. Hypoglycaemia was defined as a serum glucose ≤2.6 mmol/l3 ,7 confirmed on laboratory testing, or a mention of the word ‘hypoglycaemia’ or similar on the laboratory test request form. A total of 51 patients were identified, and 2 patients were excluded as clinical notes could not be obtained from medical records.
Clinical notes for the remaining 49 patients were reviewed for details of the events surrounding the episode of reported hypoglycaemia. Details collected included; age, weight, background to the admission, admission details, treatments given, investigations done, diagnosis reached and follow-up arrangements. Follow-up included further testing, diagnosis with appropriate management plan or outpatient review. Investigations were audited against those recommended by guidelines on MetBioNet.3 As a urine toxicology screen would only be done in exceptional circumstances this was not included as part of the audit criteria.
The results of all investigations were reviewed in order to identify the abnormal results, as well as those not available due to laboratory processing errors. Further investigations or follow-up of these results were noted. Results considered abnormal included a cortisol level <200 nmol/L,8 and any detectable levels of insulin in samples taken during hypoglycaemia. Intermediary metabolites were interpreted depending on the lipolytic and ketogenic response to hypoglycaemia.
All results, investigations and outcomes were reviewed by a paediatric registrar, paediatric endocrinologist and clinical scientist. Irrespective of investigations, patients were assigned into the categories of ‘Hyposcreen’ required or not required depending on whether the hypoglycaemia could be explained, in order to assess whether investigations were targeted appropriately. Reasons for hypoglycaemic children to be assigned to the ‘hypoglycaemia screen not required’ category included known diabetes on insulin, known fructose 1,6 bisphosphatase deficiency and non-compliance with growth hormone treatment in a patient with septo-optic dysplasia. In neonates reasons included non-refractory hypoglycaemia in neonates less than 48 hours of age, maternal insulin dependent diabetes and non-refractory hypoglycaemia, known poor feeding with improvement in blood glucose once adequate feeds given, and possible sepsis with improvement in blood glucose once treatment commenced.
Two patients were excluded as their clinical notes were unable to be reviewed. Of the 49 patients, 35 were neonates, 14 were children, 5 were planned admissions and 44 were emergency cases. Forty-two cases met the biochemical criteria for hypoglycaemia, of which 34 were neonates and 8 were children.
Thirty-four of 35 (97.1%) neonates had proven hypoglycaemia. The remaining neonate presented on day 1 of life with concerns regarding hypoglycaemia on routine postnatal ward monitoring, however when a venous sample was taken, it was not proven. This neonate appropriately did not have further investigations performed.
Of the neonates with proven hypoglycaemia, it was refractory or recurrent in 15 of 34 (44.1%) and therefore they should have been investigated. Of these only 12 of 15 (80%) had a Hyposcreen performed. Ten of these babies were less than 48 hours old at presentation.
Nineteen of 34 (55.9%) neonates were not considered to require further investigations into hypoglycaemia, as they were all <48 hours of age and had hypoglycaemia that was not recurrent or resistant to treatment. Three of 19 (15.8%) of these had a Hyposcreen performed unnecessarily.
In total 12 (80%) neonates had tests from a ‘Hyposcreen’ sent when required but 3 (20%) did not. Conversely 20 (57%) neonates did not require further testing as they had an identifiable cause for their hypoglycaemia. Despite this 3 (15%) patients had investigations sent as part of a ‘Hyposcreen’. Therefore a total of 83% (29/35) of neonatal investigations were appropriate.
Eight of 14 (57.1%) children presented with proven hypoglycaemia, and 3 of 8 (37.5%) required a Hyposcreen. All 3 (100%) of these children had a Hyposcreen performed. The remaining 5 of 8 children had known reasons for their hypoglycaemia and did not require investigation. Three of five had known insulin dependent diabetes, 1 of 5 had fructose 1,6 bisphosphatase deficiency and 1 of 5 had septo-optic dysplasia and had been admitted to the ward for blood sugar monitoring as they were known to be non-compliant with growth hormone replacement. None of these children were investigated unnecessarily.
Six of 14 (42.9%) children presented without proven hypoglycaemia. Two of 6 (33.3%) of these children presented with a gastroenteritis illness, and were hypoglycaemic on point-of-care testing. They had investigations sent when intravenous access was obtained, although the lab glucose result did not prove hypoglycaemia. It was felt in these cases that investigation had been appropriate as the judgement had been made with the best information available at the time. One of 6 children had insulin-dependent diabetes and had presented with vomiting but was not hypoglycaemic on testing. The remaining 3 of 6 children were admitted electively for a supervised fast and did not become hypoglycaemic, but despite this all 3 patients had hypoglycaemia investigations sent.
In total 5 of 14 (35.7%) children required investigation with a ‘Hyposcreen’ and 5 of 5 (100%) of these children had tests sent. Conversely 9 of 14 (64%) children did not require further testing as they had an identifiable cause for their hypoglycaemia. Despite this 3 of 9 (33%) patients had investigations sent as part of a ‘Hyposcreen’. Therefore a total of 79% (11/14) of investigations in children were appropriate.
Completeness of investigations
Twenty-three of 49 patients had investigations taken but only 2 of 23 (9%) had a full ‘Hyposcreen’ completed. The actual number of investigations carried out in each patient ranged from 6 to 13 with 9 being the median number of investigations performed.
The investigations performed in the 23 patients included intermediary metabolites in 22 (96%), lactate in 23 (100%), cortisol in 23 (100%), insulin in 19 (83%), growth hormone in 20 (87%), acylcarnitine in 12 (52%), plasma amino acids in 19 (83%), ammonia in 7 (30%) and urine samples in 12 (52%).
Although samples were taken and sent to the lab the results were not available in all cases due to laboratory processing errors (7), insufficient sample volumes (29) and transport errors (1).
As a result of investigations the diagnoses of growth hormone deficiency, adrenocorticotrophic hormone (ACTH) deficiency and hypothyroidism secondary to pituitary aplasia was made in one patient, ketotic hypoglycaemia in three patients and transient hyperinsulinism in one patient. Despite ketotic hypoglycaemia being a diagnosis of exclusion, these three patients did not have comprehensive investigation. Hypoglycaemia was most commonly attributed to known disorders, ketotic hypoglycaemia and gastroenteritis in children, and low birth weight, maternal diabetes and poor feeding in neonates (table 2).
Hypoglycaemia in the paediatric population has a number of aetiologies.3 ,4 It is more frequent and less concerning in neonates under 48 hours of age than in children presenting to the emergency department with an unexplained hypoglycaemia. Junior doctors are generally aware of the need for urgent investigation in these cases, but the requirement to perform urgent investigations in a time pressured situation can lead to difficulty. This difficulty can be technical and exaggerated by situational factors such as stress, availability of equipment as well as a lack of knowledge regarding what investigations need to be performed.
A premade pack containing the required specimen tubes and biochemistry request forms, is one method to potentially improve the adequacy of investigation of hypoglycaemia.9 These ‘Hypoglycaemia packs’ contain all the relevant equipment and therefore aim to facilitate complete and timely investigations. It is important however to be clear about who and when to investigate in order to use the packs appropriately. The threshold at which investigation should take place had previously been taken to be blood glucose ≤2.6 mmol/L,3 ,7 and this threshold was used during this audit. More recently, new guidance has been published by the Pediatric Endocrine Society following review of the available evidence which recommends that investigation is only necessary in younger children who cannot report symptoms with glucose <60 mg/dL (3.3 mmol/L) and that laboratory evaluation is performed only when the glucose is <50 mg/dL (<2.8 mmol/L), as this is the level below which autonomic symptoms occur.1
Different patient groups are likely to have different underlying causes and hypoglycaemic neonates in particular may have an explainable cause such as poor feeding. The recent guidance states that investigations are only necessary if hypoglycaemia is refractory or persistent after 48 hours of life.1 In children there is often a known cause but in those without, investigation is crucial in order to identify any metabolic or endocrine disorders that have not been diagnosed previously. In these cases complete investigation is important for diagnosis and to prevent future investigations including supervised fasting which would require hospital admission. In those admitted for supervised fasting, further investigation is only required should there be recurrence of hypoglycaemia.1 The introduction of a clear local guideline for medical staff incorporating this new guidance will help to target testing more effectively and reduce the cost of superfluous investigations.
A ‘Hyposcreen’ consists of several tests which are otherwise done infrequently and there is likely to be less familiarity with the result and its interpretation, particularly among junior doctors. It is therefore important to have clear guidance about result interpretation. The MetBioNet guideline does give some advice regarding result interpretation but there is no overall consensus about thresholds for abnormality for certain tests.
The cost of healthcare can be a difficult but important topic. In particular as NHS services are stretched with more stringent management of hospital budgets, attention needs to be paid by clinicians to the value and cost of investigations being requested. The current laboratory cost of investigating children with hypoglycaemia as per recommendations is approximately £4505 (correct as of July 2015). This is financially important, and the potential safety cost to the patient of inadequate investigation, along with the inconvenience of repeat investigations and price of recurrent hospital stays also requires consideration. The cost of laboratory investigation of these 49 patients was £6977, of which £1630 has subsequently been deemed to be unnecessary. If those patients who required further investigation had a ‘Hyposcreen’ completed, expenditure would have been £9000, which is only £2023 more than the total that was actually spent on investigations done in these 49 patients. Clearly this is a more worthwhile cost than the cost of partial or inappropriate investigations. Conversely, if we consider a potential expenditure of £13 000 to investigate the 29 patients in whom a ‘Hyposcreen’ was deemed unnecessary, this highlights the importance of clear guidance for who and when to screen. It is therefore important to target investigations appropriately as well as ensuring they are completed. In addition to this, consideration should be given to the five patients requiring elective admission for further investigations at a cost of approximately £400 per day10 not including the social costs associated with this.
Three quarters of all paediatric hypoglycaemia occurred in neonates, who are better adapted to cope with this metabolic stress. It is not necessary to immediately investigate hypoglycaemia in this age group, provided it is not persistent and there is a reasonable explanation. Within our sample group only 40% of all patients with hypoglycaemia required further investigation in order to establish a cause. Our results show that presently, resources are not being targeted appropriately with 21% of patients overinvestigated; but of greater concern, 15% are being underinvestigated. In those requiring investigation it is important that these are performed quickly and completely. Unfortunately present figures demonstrate that 91% of ‘Hyposcreens’ are incomplete, and half of abnormal results are not followed up. This has major cost implications for both the NHS and the individual who is investigated inadequately or incorrectly. We need to consider stricter guidance to improve completion and follow-up of tests, avoid inappropriate investigations and prevent delay in diagnosis. This hospital has subsequently introduced ‘Hypoglycaemia packs’ containing the necessary equipment for testing and a precompleted request form as a logical step for aiming to improve test completion.
The majority of investigations into hypoglycaemia are incomplete.
Hypoglycaemia investigations are not always targeted appropriately, but new guidance has been recently released.
Incomplete and inappropriate investigations are costly to both the individual patient and NHS.
Current research questions
Is paediatric hypoglycaemia investigated appropriately and adequately nationally or is this a wider problem?
Does the use of ‘Hypoglycaemia packs’ improve adequacy of investigation of paediatric hypoglycaemia?
What is the threshold of an abnormal cortisol response to hypoglycaemia in infants less than 3 months of age?
The authors thank the Doncaster and Bassetlaw Hospitals NHS Foundation Trust Clinical Audit & Effectiveness Department for their contribution to this project.
Contributors LR completed data collection, review of clinical notes and was the main contributing author. AN was the main project supervisor and contributed to data review and editing of the main article. KW identified investigation of paediatric hypoglycaemia as an area of concern and initiated data collection and review.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Anonymised data regarding all subjects included in this audit are available. Please contact the corresponding author for further information.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.