History and use of CSII
Insulin pump therapy or CSII has been in use for nearly 40 years. Based on achieving strict long-term glycaemic control by variable-rate subcutaneous insulin infusion from a portable pump, it was originally a research procedure, but it quickly developed into a routine treatment for selected people with type 1 diabetes. It is now used by more than 0.75 million people throughout the world, though uptake varies widely from one country to another. First pumps were relatively simple without alarms and with limited basal rate and bolus dose adjustment, but modern pumps are sophisticated with flexible dosing, safety alarms, data download to computers and sometimes connectivity to continuous glucose monitoring.
CSII was initially developed in the 1970s by a team at Guy’s Hospital in London as a research procedure aiming to test the effect of strict metabolic control on the development of diabetic microvascular disease . At that time, there was some doubt as to whether diabetic tissue complications are caused by the disease process itself (probably through hyperglycaemia) or whether retinopathy, nephropathy and other manifestations only run in parallel with the metabolic disorder. Insulin injection regimens at the time were not able to achieve and maintain near-normoglycaemia for long enough to show a possible effect on the complications, and alternative means of achieving strict control were being sought.
Gérard Slama and colleagues in Paris had shown in 1974  that a few days of open-loop intravenous (IV) insulin infusion in type 1 diabetes using a portable pump held in a shoulder bag produced good glycaemic control – indeed similar to that of the first closed-loop, artificial pancreas devices that were then being tested. Professor Harry Keen in London then reasoned that, since multiple small subcutaneous injections of short-acting insulin generally produce good control in type 1 diabetes, short-acting insulin infused subcutaneously would both improve metabolic control and avoid the risks of thrombosis, phlebitis and septicaemia associated with long-term IV infusion.
The first insulin pump (the ‘Mill Hill Infuser’). Top: pump showing end button (red) that engaged the bolus dose, i.e. the higher infusion rate. Lower: pump with casing removed to show the syringe drive. Keen and Dr John Pickup at Guy’s Hospital then went on to develop CSII as both an experimental tool and as a routine treatment for selected people with type 1 diabetes. A battery-operated, 159-g miniature syringe pump that was being used by the pharmacologist, Dr John Parsons, at the National Institute for Medical Research in Mill Hill, London, to administer parathyroid hormone in animals was adapted for insulin delivery in diabetes using a dual-rate infusion strategy - a single slow basal rate throughout the 24 hours and a higher preprandial rate, thus mimicking non-diabetic insulin delivery. This pump became known as the ‘Mill Hill Infuser’, and was the first experimental and (shortly afterwards) commercially available, dedicated insulin pump. In initial studies, insulin was infused via a nylon cannula that was implanted by the physician into the subcutaneous tissue of the abdomen using an introducing needle , but later this was changed to a winged infusion set with metal needle (‘Butterfly’) that patients could implant themselves.
Early confirmation of the effectiveness of CSII in comparison with conventional injection therapy came from groups at Yale University, USA (William Tamborlane and colleagues ), the University of Western Ontario, Canada (John Dupré and colleagues ), and Düsseldorf, Germany (Michael Berger et al. ).
With the wider use of CSII in routine practice in the 1980s, anecdotal or uncontrolled reports began to appear detailing episodes of severe hypoglycaemia or diabetic ketoacidosis (DKA) in some patients on insulin pumps, and because of this, enthusiasm for the technology dimmed for a number of years, particularly in the UK. However, a series of controlled studies  and, later, meta-analyses  was to show that the frequency of severe hypoglycaemia is in fact markedly less on CSII than injection therapy, and though patients on CSII might be at increased risk of DKA in the event of pump malfunction, with proper education of patients and healthcare professionals and frequent blood glucose monitoring, DKA occurs no more frequently than on injection therapy .
By the 1990s, CSII was increasing in usage, especially in the USA and especially in response to the positive results of the Diabetes Control and Complications Trial (DCCT, reported 1993 ) which finally showed the value of intensive insulin therapy and strict glycaemic control in slowing the development of diabetic microangiopathy. Criteria for starting patients on insulin pump therapy in routine practice and the effectiveness vs. multiple daily insulin injections (MDI) were, however, still ill-defined at this time.
A number of pre-DCCT, relatively short-term randomised controlled clinical trials (RCTs) in the 1980s tested the effect of CSII-induced strict glycaemic control vs. conventional or MDI therapy on the progression of microvascular disease, including the multinational Kroc Study (UK, USA and Canada) , the Steno Study in Denmark  and the Oslo Study in Norway . In the DCCT , 59% of the patients in the intensively treated group were managed by CSII at some stage, though they were not randomised between pump therapy and MDI.
Uptake of CSII in clinical practice has continued to increase over the last three decades or so, driven particularly by three factors: the development of more reliable and sophisticated pump technology, an accumulating evidence base for the efficacy of pumps vs. best injection therapy and the issuing of national guidelines on the indications for using CSII, based on clinical benefit and cost-effectiveness .
First insulin pumps lacked many of the features that are now standard in modern pumps such as alarms for malfunction, low-battery state and cannula occlusion. Apart from alarms, notable developments in pump technology over the years include the introduction of flexible basal rates that allow patients to alter infusion rates both on demand and at a preset time of the night or day. This now allows an increase to be pre-programmed, for example for the early hours of morning so as to counter the dawn phenomenon, or a reduction in infusion rate can be employed before and after exercise to reduce the risk of hypoglycaemia.
Integrated bolus calculators were included in pumps from about the year 2002 and give advice on meal insulin based on intended carbohydrate intake, insulin: carbohydrate ratio and insulin sensitivity, pre-meal glucose and target glucose concentrations, and an estimation of the insulin remaining from the previous bolus (‘active insulin’ or ‘insulin onboard’). In most first pumps, bolus doses were delivered in as short a time as possible but in subsequent years models with additional bolus profile options became available – square wave for high fat and protein meals or the slow digestion of diabetic gastroparesis, and dual or combination waves for high-carbohydrate/high-fat meals. Modern pumps also have the facility for download of data to a computer which helps to detect problems and optimise control, and some pumps have connectivity with continuous glucose monitoring .
A newer generation of smaller ‘patch pumps’ are now appearing which often have a short, pump-integrated cannula (‘tubeless’) and are attached to the body via an adhesive patch . Some of these pumps might be more suitable for type 2 diabetes, where pump therapy has not been traditionally recommended and RCT evidence of effectiveness is weak but where there is increasing evidence for a role in selected patients.
An example of the guidelines that have been issued in recent years for the most appropriate usage of CSII is the UK National Institute for Health and Clinical Excellence (NICE) Technology Appraisal of 2008  which summarised the clinical evidence base and health economic value of CSII. It concluded that insulin pump therapy is a cost effective use of resources for the UK National Health Service and recommended a trial of CSII should be considered in adults when HbA1c has remained elevated (≥8.5%, 69 mmol/mol) or when there is persistent disabling hypoglycaemia, in spite of best attempts with MDI.
NICE and other national guidelines have been widely influential in promoting insulin pump usage but the uptake of CSII varies markedly between countries and, in many cases, within a country. In December 2012, manufacturers’ estimates indicated that high uptake countries included USA, Norway and Israel (>30% of those with type 1 diabetes) and low uptake countries included UK, Belgium, Spain and Portugal (<10%). It is unclear why there is this variation but determining factors might include differing availability of financial and manpower resources, particularly diabetes nurse educators, the lack of knowledge of healthcare professionals (HCPs) about the indications for CSII and about its procedures, and poor HCP engagement and persisting negative beliefs about the value and safety of insulin pumps (and possibly medical technology in general).
^ Pickup JC et al. Continuous subcutaneous insulin infusion: an approach to achieving normoglycaemia. BMJ 1978; i: 204-207.
^ Slama, G et al. One to five days of continuous insulin infusion on seven diabetic patients. Diabetes 1974; 23: 732-738.
^ Tamborlane WV et al. Reduction to normal of plasma glucose in juvenile diabetes by subcutaneous administration of insulin with a portable infusion pump. N Engl J Med 1979; 300: 573-578.
^ Champion MC et al. Continuous subcutaneous insulin infusion in the management of diabetes mellitus. Diabetes 1980; 29: 206-212.
^ Berger M et al. Insulin pump treatment for diabetes: some questions can be answered already. Clin Physiol 1982; 2: 351-355.
^ Bending JJ et al. Frequency of diabetic ketoacidosis and hypoglycemic coma during treatment with continuous subcutaneous insulin infusion. Am J Med 1985; 79: 685-691.
^ Dahl-Jørgensen K et al. Effect of near-normoglycaemia for two years on the progression of early diabetic retinopathy: the Oslo Study. BMJ 1986; 293: 1195-1199.
^ Pickup JC et al. Severe hypoglycaemia and glycaemic control in type 1 diabetes: meta-analysis of multiple daily insulin injections versus continuous subcutaneous insulin infusion. Diabetic Med 2008; 25: 765-774.
^ Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 1993; 329: 977-986.
^ Kroc Collaborative Study Group. Blood glucose control and the evolution of diabetic retinopathy and albuminuria. NEJM 1984; 311: 365-372.
^ Pickup JC. Insulin-pump therapy for type 1 diabetes mellitus. NEJM 2012; 366: 1616-1624.
^ Pickup JC. Management of diabetes mellitus: is the pump mightier than the pen? Nature Rev Endocrinol 2012; 8: 425-433.
^ National Institute for Health and Clinical Excellence. Continuous Subcutaneous Insulin Infusion for the Treatment of Diabetes Mellitus. Technology Appraisal Guidance 151 (Review of Technology Appraisal Guidance 57). London, England: NICE, 2008.