Article Text

Download PDFPDF

What are the main research findings during the last 5 years that have changed my clinical practice in diabetes medicine?
  1. John W Gregory
  1. Department of Child Health, Wales School of Medicine, Cardiff University, Cardiff, UK
  1. Correspondence to John W Gregory, Department of Child Health, Wales School of Medicine, Cardiff University, Cardiff CF14 4XN, UK; gregoryjw{at}

Statistics from


The world of childhood diabetes has seen more changes in clinical care being introduced in recent years than many subspecialty areas of clinical practice. This review identifies some of the key underlying evidence and considers the implications for clinical teams that personal experience of these developments has highlighted.


The last 5 years have been a particularly exciting time in diabetes medicine with dramatic implications for a few patients. There have been advances in our understanding of the genetic basis of disease, and furthermore, technological advances in monitoring glycaemic control and delivering insulin leading to significant changes in the provision of care to many children with diabetes. This article focuses on a few major discoveries in recent years that have impacted on the diagnosis and treatment of young people with diabetes attending my clinical service.

Diagnosis of diabetes

The diagnosis of diabetes is now classified into four main groups:

  • type 1 diabetes (T1D) associated with autoimmune insulin deficiency

  • type 2 diabetes (T2D) characterised by insulin resistance

  • other types of diabetes such as genetic forms, which include various causes of neonatal diabetes and maturity onset of diabetes in youth

  • gestational diabetes.

Type 2 diabetes

The distinction between T1D and T2D is critically important because the therapeutic approaches are completely distinct. T1D requires immediate treatment with insulin to avoid death whereas lifestyle changes combined with insulin-sensitising agents, and sometimes in the longer term, insulin therapy form the mainstay of established treatment for T2D. T2D should be suspected in the presence of other features of insulin resistance, including obesity, hyperlipidaemia, increased blood pressure, acanthosis nigricans, ovarian hyperandrogenism and non-alcoholic fatty liver disease.1 However, insulin secretion in T2D has been shown to vary markedly from elevated levels to absolute deficiency. In children with features suggestive of T2D, assessment of insulin reserve through glucose tolerance testing has become an important tool in identifying those who will benefit from early introduction of insulin therapy.2

Neonatal diabetes

Neonatal diabetes is a relatively rare but often clinically severe form of diabetes presenting in the first 3 months of life and can be subdivided into transient and permanent forms. Transient forms resolve by a median age of 12 weeks and do not require therapy thereafter, although 50% of patients ultimately relapse. Neonatal diabetes has been shown to be caused by single gene mutations, and in the case of permanent neonatal diabetes, mutations of the KCNJ11 gene which encodes the Kir6.2 subunit of the pancreatic ß-cell KATP channel.3 This observation led Hattersley and Ashcroft to demonstrate the significant potential benefits to patients of understanding the molecular basis of their disease. Closure of this channel was known to be the first step in insulin release, precipitating ß-cell wall depolarisation, calcium influx and exocytosis of insulin granules. Subsequent clinical studies demonstrated that patients with this molecular defect respond positively to sulphonylurea therapy and can be successfully weaned off insulin treatment. This finding now makes it mandatory to undertake genetic testing of all patients with apparent insulin-dependent diabetes (30% of those with KCNJ11 mutations present with ketoacidosis) presenting under the age of 6 months. Sulphonylurea treatment has been shown to produce better glycaemic control and less hypoglycaemia than insulin in children and adults of all ages whose diabetes is due to this underlying genetic cause.4 For this subgroup at least, there is the exciting possibility that their apparent insulin-dependent form of diabetes may not require injections of insulin for life.

Management of diabetes

In 1993 the Diabetes Control and Complications Trial (DCCT) reported on a 10-year study that confirmed beyond a shadow of doubt, the close relationship between glycaemic control and the complications of diabetes. A median of 7.4 years of intensified insulin therapy produced a haemoglobin A1c (HbA1c) concentration of 8.1% compared with 9.8% in controls. Even in a subgroup of 195 adolescents there were significant reductions in background retinopathy by 53%, neuropathy by 60% and microalbuminuria by 54%.5 A follow-up study (the Epidemiology of Diabetes Interventions and Complications Study) showed a ‘memory effect’ of this period of tight glycaemic control. After participation in the trial, blood glucose control equilibrated among patients in both arms of the trial but the benefits of intensive therapy persisted, with 74% less retinopathy, 48% less microalbuminuria and 85% less albuminuria.6 This implies that periods of tight glycaemic control may have surprisingly long benefits in terms of reduced complications and has stimulated extensive effort in subsequent years to find ways of tightening blood glucose management. The relevance of tight glycaemic control to those of us currently working in the UK has been further emphasised by the findings of a recent long-term prospective observational follow-up study in the Oxford region,7 which included a cohort of over 500 people who developed T1D in childhood. The study showed a cumulative presence of microalbuminuria of 25.7% after 10 years of diabetes and 50.7% after 19 years of diabetes, with 13.9% progressing to macroalbuminuria at a mean age of 18.5 years. The only modifiable predictor was that of poor glycaemic control.

Insulin pumps

It has long been recognised that subcutaneous injections of insulin two to four times daily fails to achieve adequate glycaemic control in the majority of children and young people with T1D. The management of children is particularly complicated by their unpredictable and increased frequency of food intake, variable patterns of physical activity and increased incidence of intercurrent illnesses. Measured against these challenges, continuous subcutaneous pump infusions of insulin offered potential advantages, theoretically allowing a better matching of insulin supply to requirements.

Although insulin pump therapy has been available for over 20 years and is considered the ‘gold-standard’ of care for adults with T1D,8 it is only in the last decade that evidence has emerged on the potential benefits of pump therapy in children with T1D. Observational studies in school-aged children and adolescents9 appear to confirm previous evidence that this form of insulin delivery is associated with benefits, including lowering of HbA1c and a decreased risk of ketoacidosis and hypoglycaemia. Similar benefits have been reported in a systematic review of pump therapy in younger, largely preschool-aged children.8 However, interpretation of the published data on pumps is complicated by the limited number of randomised trials which avoid the significant bias that may be inherent in cohorts of patients recruited for observational studies, and the small size and relatively short duration of many published studies. Recent systematic reviews and meta-analyses of paediatric trials have concluded that pump therapy10,,12 produces improved quality of life and modest reductions of about 0.2% in HbA1c compared with multiple injection regimens, which is less dramatic than the results seen in observational studies. What is less clear is the extent to which any of these benefits results from the intrinsic advantages of delivering insulin by pump instead of the intensive education package that is required before and during pump therapy. Nonetheless, these findings have encouraged many clinics to treat more young patients with pump therapy, particularly when they have struggled to achieve optimal glycaemic control (HbA1c<8.5%) using multiple injection therapy in the absence of disabling hypoglycaemia.13

Blood glucose monitoring

Optimal manipulation of insulin dosages requires a clear understanding of prevailing blood glucose concentrations. Conventionally, most children monitor their diabetes by intermittent finger-prick blood tests for glucose measurement, ideally four or more times daily. Such testing has a number of disadvantages, particularly the limited information on overall blood glycaemic control that is derived from such occasional sampling within each 24 h period. Minimally invasive sensors using a catheter or small plastic chip containing a sensor inserted every few days into the subcutaneous space have been developed to monitor interstitial glucose concentrations at frequencies of 1–20 min. This technology allows patients to ‘continuously’ monitor the effects of insulin, food and exercise on glucose concentrations in real time and to make appropriate therapeutic decisions. This may be especially useful in patients attempting to intensify their insulin treatment (particularly when combined with pump therapy) and in those who may have been prone to recurrent, possibly asymptomatic hypoglycaemia. Alternatively, when reviewed retrospectively, these data may provide a useful educational tool for healthcare professionals to discuss blood glucose management with their patients, particularly young patients who have experienced great difficulties interpreting their overall blood glucose control from intermittent blood glucose measurements.

Initial studies14 suggested continuous glucose monitoring had the potential to help young people make changes to their blood glucose management to reduce HbA1c. However, a larger trial produced disappointing results, suggesting that teenagers may struggle to engage in the longer term with this approach to monitoring their glycaemic control.15 16 More recent evidence from randomised controlled trials seems more optimistic, showing that young people maintaining optimal glycaemic control can reduce their time spent in hypoglycaemia. This technology may in fact be acceptable to many young people with T1D17 and continuous monitoring of glucose may be particularly effective at improving glycaemic control when combined with an additional intervention such as insulin pump therapy.18

Insulin resistance and T2D

The start of the 21st century has coincided with dramatic increases in the prevalence of obesity in many developed countries, including the UK. Although absolute numbers of children with T2D remain small in the UK compared with the USA, where T2D is anticipated to be the commonest form of childhood diabetes within 10 years, in a British paediatric surveillance unit survey in 2004–2005 the incidence of T2D was shown to have increased significantly in the early years of this century, with particular risk factors including ethnic minority origin, a family history of T2D and obesity.19 There is limited experience of the use of insulin-sensitising agents to reduce the risks of developing T2D in childhood. Therapeutic approaches have been largely influenced by experience in adult practice. However, a recent systematic review20 identified four randomised controlled studies which have shown clear evidence that treatment with metformin promotes increased insulin sensitivity and reduces body mass indices in children in the prediabetes phase, potentially reducing the risks of progression to T2D.

As for the management of established T2D, a systematic approach has been advised, again based largely on the experience of treatment of adults with T2D.1 This focuses initially on changing lifestyle to promote physical activity and to alter dietary intake to reduce weight, along with the use of pharmacological agents if necessary to normalise blood glucose concentrations. In addition, interventions designed to impact on the co-morbidities of hypertension, dyslipidaemia, nephropathy and non-alcoholic fatty liver disease should be considered. Early introduction of metformin has been recommended for some time to improve glycaemic control in children with T2D,21 with the use of sulphonylureas and insulin as adjunctive therapy when there has been a poor response to oral hypoglycaemic agents as potential additional therapeutic options.1 Trials are ongoing evaluating the effectiveness and safety in children of newer agents which influence glycaemic control through the effects on incretin physiology (glucagon-like peptide-1 receptor agonists and dipeptidyl peptidase IV inhibitors) and the results are eagerly awaited.

Cystic fibrosis

Improvements in the care of children with cystic fibrosis have resulted in dramatic improvements in survival in recent decades. With prolonged life expectancy, a greater incidence of cystic-fibrosis-related diabetes (CFRD) has occurred so that in many paediatric diabetes services, these patients now account for the second largest group of patients. CFRD may develop insidiously over many years and seems to be associated with a six-fold increase in morbidity and mortality, particularly in girls.22 Measurement of HbA1c to diagnose CFRD is unreliable and the standard diagnostic test is oral glucose tolerance testing. In recent years there has been increased interest in the potential of continuous glucose monitoring to recognise abnormalities of glucose homeostasis earlier than would be the case with glucose tolerance testing. This approach has now been validated for use in young people with cystic fibrosis, though its clinical prognostic significance remains to be determined.23

The observation of increased mortality associated with CFRD has stimulated a more aggressive approach to its management, including the early use of insulin therapy. There are important differences in the treatment regimes recommended for CFRD compared with those required for T1D. Affected children have energy requirements that often exceed 150% of normal and insulin is indicated, not just to normalise blood glucose values but for its anabolic effects too. There is no clear evidence to recommend one insulin regime over another, but the use of either multiple injection regimes or insulin pumps when insulin delivery is matched to carbohydrate intake is recommended24 and results in improved body composition and lung function.25 Anecdotal evidence suggests that an aggressive approach to insulin treatment during intercurrent illnesses is particularly important in cystic fibrosis, during which time dramatic increases in insulin requirements may occur.

Reflections on how research findings in the last five years have impacted on clinical practice

The invitation to write this article provides an opportunity to reflect on the relatively slow pace of change in clinical practice and the delay between research findings being made public and their practical integration into clinical services. Why should this be? Maybe it reflects a relatively conservative approach to innovation in my own clinical service or perhaps the need to see new developments applied successfully in a broader clinical context before incorporating them into personal clinical practice. An audit of outcomes (HbA1c) across the UK and around the developed world has shown a wide variation in outcomes but to date there is no clear evidence that clinics which embrace change early on are more likely to demonstrate better outcomes.26 The lack of evidence from high-quality, long-term (greater than 1 year) trials that have used study designs which limit the risk of bias in patient recruitment (and therefore make it difficult to understand the relevance of study findings to routine clinical practice) has certainly contributed.

This review highlights how developments in our understanding of the genetic basis of disease are particularly beneficial for subpopulations of children with diabetes. Investigation of the genetic basis of neonatal diabetes has allowed two children and their father to be weaned off insulin therapy.4 Furthermore, the database of childhood onset diabetes in Wales has been searched for patients (many of whom have progressed to adult services) known to have developed diabetes in very early life to identify other patients who might benefit. The requirements and skills necessary to track down affected patients require investment in specialist nurses with the relevant expertise in diabetes and genetic counselling to deliver these services.

Despite my personal clinical and scientific interest in psychoeducational interventions for the management of childhood diabetes,27 the prominence placed on this aspect of care by organisations such as Diabetes UK representing the interests of children with diabetes, and evidence that these interventions are capable of producing a beneficial effect on glycaemic control equivalent to that achieved through many technological advances,28 it is surprising what little impact they have had on our clinical practice over the last 5 years. In a thoughtful review of this topic by Skinner and Cameron,29 the point is made that ‘intensification’ of insulin therapy is only likely to optimise outcomes when appropriate psychoeducational support is provided, such as that during participation in the intensification arm of the DCCT study.30 A major contribution to this problem is the lack of psychological support available in UK services, with only 21% reporting integrated psychological expertise in 2008.31

This review highlights the potential benefits that might arise from the introduction of technological advances such as insulin pumps and continuous glucose monitoring. A practical problem for clinics such as mine arises from the complexity of many of these ‘high-tech’ devices and the tendency for just one or two members of the nursing team to develop the clinical skills necessary to support patients who choose to use these approaches to self-management. To avoid patients becoming confused by seeing team members with different skill levels and knowledge about their diabetes management, it is important to ensure that all team members are adequately trained. It is likely that a key factor in optimising outcomes relates to the clinician's belief that an intervention will work.29 Furthermore, the results of a study investigating outcomes in clinics from around the world have suggested that a major contribution to the variance in HbA1c levels seen among clinics is the extent to which the healthcare team are agreed on the treatment goals within the service.32 It is therefore imperative that sufficient consideration is given within teams to ensure that there are agreed strategies for the management of children with T1D and that staff are adequately skilled for all the tasks they undertake in clinic. When training needs are identified, adequate resources need to be made available to ensure such training can be provided. These practical concerns support a view that pump services should only be provided by centres of excellence with adequate numbers of staff who have the appropriate specialist expertise and a specified minimum numbers of patients receive pump therapy to ensure maintenance of skills and standards. With respect to young people with T2D and CFRD, close working relationships with colleagues in adult services may be especially helpful for paediatricians to gain advice and expertise in the use of treatments in which they may have relatively limited experience. In addition, optimal transition of teenagers with diabetes to adult services needs to be ensured.


The author thanks Justin Warner, Lesley Lowes and Corinna Bretland for their helpful comments on an earlier draft of this review.



  • Provenance and peer review Commissioned; externally peer reviewed.

Request Permissions

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.