Do you think a pump is for you? Insulin infusion pumps are modern, computerised devices which infuse insulin at a low dose continuously under the skin via a catheter. You should consider a pump if you are-

• having hypos at night, and adjusating your basal insulin isn't helping.
• You are not getting your HbA1c down, even though you are doing lots of tests and making lots of changes to you doses.
• doing lots of different activities or sport which are difficult to predict.
• you are having to take a lot of extra glucose or food to stop you going low during or after sport.
• you are fed up of doing 4+ doses per day!

Continuous subcutaneous insulin infusion and hypoglycaemia during exercise, by Alistair Lumb.

CSII has become an increasingly important option in the therapy of type 1 diabetes over recent years. Treatment involves the insertion of a temporary cannula into the subcutaneous tissue through which rapid-acting insulin is infused continuously using an electronically controlled pump. CSII therapy permits numerous bolus insulin doses to be given without the need for a physical injection each time, and also means that background insulin levels can easily be adjusted as they are provided by a variable infusion of rapid acting analogue insulin rather than a subcutaneous bolus of longer-acting insulin. This gives much greater flexibility in insulin delivery than MDI therapy.

It is recognised in people without diabetes that aerobic exercise is associated with a fall in insulin concentrations to below fasting levels. In view of this, the strategies for adjustment of CSII therapy during exercise have focussed on the reduction or cessation of insulin infusion during exercise, based on the reasonable theory that this should approximate more closely to the physiological situation and therefore reduce rates of hypoglycaemia. Early studies, carried out in adults using CSII therapy in the 1980s (and therefore in the era pre-dating analogue insulins) showed mixed results(8;9). As expected, continuing basal insulin infusion at the usual rate was demonstrated to result in hypoglycaemia. A 50% reduction in the prandial isulin bolus with a meal taken 90 minutes prior to exercise was shown to reduce rates of hypoglycaemia, and a reduction in insulin basal rate of between 50% and 100% was also shown to be beneficial in reducing rates of hypoglycaemia(9).

In slight contrast, cessation of basal infusion 30 minutes prior to 45 minutes of exercise at 60% VO2 max, with basal infusion stopped for a total of 3 hours (and hence restarted at the usual rate 105 minutes following exercise), resulted in the avoidance of hypoglycaemia during post-prandial exercise in the morning, but not in the afternoon. 3 out of 7 participants performing post-prandial exercise in the afternoon suffered hypoglycaemia, and comparison with previous results suggested little benefit of this strategy when the infusion had been left running at the usual rate. Interestingly, in these 3 participants, insulin levels did not decline during exercise, and this failure to achieve the hoped for reduction in circulating insulin by reducing infused insulin likely explains why the strategy failed to produce the expected results(8).

Avoiding hypoglycaemia during exercise is a particular focus for children, adolescents and their parents as children’s play tends to be very active. Strategies to allow children to play safely while avoiding significant hypoglycaemia are clearly important for children to live as normal a life as possible with diabetes, and therefore the majority of the more recent evidence regarding the management of CSII therapy for sport and exercise therefore comes from work with younger people with type 1 diabetes.

Studies in children using CSII with analogue insulins have yielded clearer evidence of the benefits of reducing insulin basal infusion rate during exercise. One such study, designed to simulate unplanned post-prandial exercise in children and adolescents aged between 10 and 19 years, compared the effect of reducing basal rate to 50% of normal during exercise with that of removing the insulin pump altogether for the exercise period(10). Exercise occurred on average just under 2 hours following the most recent meal, which was accompanied by the usual insulin bolus in order to simulate the unplanned exercise, and consisted of 40 to 45 minutes exercise on a cycle ergometer at 60% VO2max. Little difference was found in the physiological response to exercise between the 2 groups, and 2 of 10 participants suffered hypoglycaemia (defined as blood glucose <70 mg/dl, equivalent to < 3.9 mmol/l) in each group. Interestingly, the exercise sessions with a hypoglycaemic event began at lower glucose levels and higher insulin levels (for 3 subjects, 10–50% higher) than the other sessions performed by the same participants, again suggesting that the problem for these individuals was a failure to achieve the intended reduction in circulating insulin by the reduction in insulin actually being infused. This suggests that the strategy of reducing insulin delivery is the correct one, but that if this does not result in a reduction in circulating insulin levels then problems with hypoglycaemia may persist. It also demonstrates that insulin levels at the start of exercise are important, and therefore reducing infused insulin some time before exercise may be more useful than making the reduction at the start of exercise. We shall return to this later.

A study in children and adolescents aged between 8 and 17 investigated whether stopping basal insulin at the start of exercise could reduce the frequency of hypoglycaemia compared to when basal insulin infusion is continued at its usual rate(11). Exercise occurred approximately 4 hours following the most recent meal, and consisted of four sessions of 15 minutes walking on a treadmill to a target heart rate of 140bpm with a 5 minute rest between each session. The two conditions were presented in a crossover design. In the condition where the basal insulin infusion was suspended, this was done at the start of exercise with the basal infusion re-started after 2 hours (and thus 45 minutes after the exercise period). Exercise started with blood glucose between 120 and 200 mg/dl (6.7 and 11.1 mmol/l). There was a significant reduction in the fall in blood glucose during exercise when basal insulin was suspended, leading to a reduction in rates of hypoglycaemia from 43% to 16%. Only 9% of the children who suspended basal insulin and started exercise with a blood glucose > 130mmol/l (7.2 mmol/l) suffered hypoglycaemia. The beneficial effect of suspending basal insulin was consistent in subgroups based on HbA1c, age, gender and usual frequency of exercise. However, a consequence of stopping basal insulin was a significant increase in the rates of hyperglycaemia. No abnormal ketone levels were recorded during the exercise period, although it should be noted that ketone readings were not recorded in the 45 minutes after exercise finished.

Taken together, the results of these studies suggest how CSII therapy might be adjusted for exercise, although the optimal strategy has not been identified and will likely vary from person to person. With CSII therapy using rapid-acting insulin analogues, it has been demonstrated that a reduction in basal insulin infusion at the start of a relatively short period of moderate exercise will help to avoid hypoglycaemia both during the period of exercise and also immediately afterwards. The optimal reduction is not clear, but for most people it probably lies somewhere between 50% and 100% (the suspension of basal insulin). Whether there is a benefit of an earlier reduction in basal rate has not been investigated. The available evidence suggests that circulating insulin levels at the start of exercise are important, and the time-action profile of the available rapid-acting analogue insulins suggests that the earliest effect of a reduction in basal rate will occur at around 10-15 minutes(12;13). In order to reduce circulating insulin levels at the start of exercise, therefore, it would sensible to suggest that a reduction in basal infusion rate should be made some time prior to starting exercise. Unfortunately there is as yet no evidence to suggest when this should take place. An extension of this would be the consideration of whether basal rate should be increased again before exercise finishes, as insulin levels in those without diabetes would increase at this time. Again, while this strategy makes physiological sense and has been used with some success in individual athletes, there is as yet no data to support whether this might be beneficial in general.

As one might suspect, suspending basal insulin at the start of exercise does increase the risk of hyperglycaemia but was not observed to increase the risk of ketosis during a short period of exercise and may be the ideal option in groups (e.g. very young children) where significant hypoglycaemia may have long term consequences. However, it should be noted that the avoidance of hypoglycaemia may not be the only focus of attention for athletes of any age with type 1 diabetes. Many also report problems with hyperglycaemia affecting performance, and hyperglycaemia also has implications for overall glycaemic control. Furthermore, it is important to note that while significant ketosis was not observed during 45 minutes of exercise following the suspension of basal insulin, the theoretical risk of ketosis remains for longer suspension of basal infusion, and this has also not been investigated in detail.

Continuous glucose monitoring and exercise

Continuous glucose monitoring (CGM) has been increasingly available since the late 1990s. This technology employs a sensor to measure glucose concentrations in the subcutaneous tissue using a glucose oxidase reaction or microdialysis method, and circulating glucose is then estimated from this concentration using an algorithm and some assumptions about the equilibrium in glucose levels between the two regions(5). Earlier systems recorded glucose data which could only be accessed once the sensor had been removed, but “real-time” CGM systems have become available which allow users to access glucose readings and information about their rate and direction of change while the sensor is being worn. Many experts feel that CGM will prove to be a useful technology to help people with Type 1 diabetes improve metabolic control around exercise - both through the ability to react to changes detected at the time exercise, and also through the provision of information which will allow individuals to plan more accurately how to balance carbohydrate intake and the adjustment of insulin doses for exercise in the future(5).

The use of real-time continuous glucose monitoring systems in general has been shown to reduce hypoglycaemia in well-controlled adults and children with type 1 diabetes while also allowing for an improvement in overall glycaemic control(14;15). Data looking at the accuracy of CGM during exercise have been encouraging. CGM accurately determines interstitial glucose levels during 1 hour of intensive cycling exercise (spinning), and also accurately reflects the direction of change of blood glucose levels(16). In a separate study, the Freestyle Navigator CGM system was found accurately to reflect the magnitude of the fall in blood glucose levels seen with moderate treadmill exercise in a group of children , albeit with a 10 minute delay (17).

This delay, however, represents one of the major challenges with using CGM to prevent hypoglycaemia during exercise. There is a recognised time lag between changes in blood glucose and changes in glucose levels in the interstitial compartment at rest(18), which is similar to the 10 minute time lag reported above. As a result, in both of the above studies the CGM was unable to keep up with the rapidly falling glucose levels seen during intense aerobic exercise, and therefore tended to overestimate the actual blood glucose reading when compared with capillary blood samples(16;17). In one study designed to examine the factors affecting CGM system calibration, CGM accurately detected only 65% of hypoglycaemic events during exercise when 3 calibrations per day were used, and only 69% when 4 calibrations per day were used(19).

Strategies are now being developed to overcome this limitation. Using real-time CGM with an alarm, people with uncomplicated Type 1 Diabetes and no evidence of hypoglycaemia unawareness suffered significantly fewer episodes of hypoglycaemia during exercise (30 minutes at 40% VO2 MAX) when a low glucose warning alarm was set to 5.5 mmol/l compared to when the alarm was set to 4 mmol/l or no alarm was used(20). The alarm was used to trigger carbohydrate intake to avoid incipient hypoglycaemia. Interestingly, the CGM still overestimated capillary glucose by an average of 1.6 mmol/l, meaning that even using the higher alarm threshold did not completely eliminate hypoglycaemia. Based on their results, the authors therefore recommend this strategy in situations where glucose levels can be expected to fall rapidly, such as during moderate exercise similar to that used in their experimental protocol.

An extension of this strategy has been piloted during a sports camp for adolescents with diabetes(7). Participants aged between 9 and 17 wore real-time CGM during a variety of different exercise situations. An algorithm was used in which carbohydrate intake was advised based both on real-time CGM readings and also the sensor’s indication of their rate of change (see figure 1). Blood glucose levels were maintained within target levels to a great extent, with a reduction in hypoglycaemia compared to expected levels and no hyperglycaemia. The authors recognise that this is a pilot study in which there was no control group, and variables such as exercise intensity, age and body weight were not taken into account. Also of concern was that no results were obtained from 6 of 25 participants recruited because of sensor data loss or the sensor falling out. However, the algorithm was surprisingly successful in spite of these limitations, and this certainly suggests a possible important future role for CGM in the management of diabetes in the context of sport and exercise.

CSII, CGM and Nocturnal Hypoglycaemia

Nocturnal hypoglycaemia following exercise is a well recognised problem in Type 1 Diabetes, and has been observed following both aerobic(21;22) and mixed forms(23) of exercise. The exact mechanism for this is not clear, although it is certainly possible that the exercise-induced recruitment of GLUT-4 receptors to the surface of the muscle cell may be involved. Another contributing factor is likely to be that exercise blunts the counter-regulatory response to subsequent hypoglycaemia(24). One strategy which has been used to combat the risk of nocturnal hypoglycaemia following exercise is to reduce basal insulin doses on the night after an exercise bout. This can be problematic when done in the context of a MDI regimen as it increases the risk of subsequent hyperglycaemia.

In those using CSII it would be reasonable to suspect that reducing basal insulin for some of the night following exercise might be able to reduce the risk of nocturnal hypoglycaemia without significant hyperglycaemia subsequently. As with many other aspects of the management of CSII for exercise, this hypothesis has been tested in children(25). Children and adolescents with Type 1 Diabetes treated with CSII underwent 4x15minutes bursts of treadmill exercise at around 4pm with 5 minute rest periods in between. For the night after exercise they either reduced their basal insulin by 20% from 9pm to 3am, took 2.5mg orally of terbutaline (a β-adrenoceptor agonist) or received no intervention. Both ingestion of terbutaline and reduction in basal insulin infusion resulted in a reduction of nocturnal hypoglycaemia, but ingestion of terbutaline did result in an increase in morning hyperglycaemia as had previously been found with adults(26). While further studies may permit a more appropriate dose of terbutaline to be selected in future, being able to reduce nocturnal basal insulin infusion for a limited period of time offers a means to reduce nocturnal hypoglycaemia without the need for additional pharmacological therapy. In those treated with CSII this represents a successful solution to the problem of nocturnal hypoglycaemia following exercise. As the authors themselves say "The flexibility to adjust basal rates by the hour remains one of the most attractive features of an insulin pump and is ... particularly useful for the active person with T1DM”.

An attractive strategy, combining CSII and CGM, may be of particular benefit in those experiencing nocturnal hypoglycaemia following exercise. There is an insulin pump commercially available in the UK (Paradigm® Veo™, Medtronic Inc., Northridge, CA) which can be set to cease insulin infusion for a period of up to 2 hours in response to CGM glucose readings below a certain threshold, referred to as the low-glucose suspend (LGS) function. In a six-centre trial, 31 adults with type 1 diabetes were studied during a period of standard CSII therapy compared with a 3-week period using CSII with LGS(27). In those with the highest frequency of nocturnal hypoglycaemia there was a significant reduction in the duration of nocturnal hypoglycaemia (defined as CGM glucose <2.2mmol/l or 40mg/dl) from 46.2 min/day for conventional CSII to 1.8 min/day for CSII with LGS. Clearly this strategy could be beneficial in avoiding hypoglycaemia following exercise. Furthermore, knowing that hypoglycaemia can affect the counter-regulatory hormone response to subsequent exercise and that this effect increases with increasing severity of hypoglycaemia(28), it is possible that preventing nocturnal hypoglycaemia in this way could also help to prevent hypoglycaemia and maintain performance during exercise the following day as well.

Novel strategies for preventing hypoglycaemia during exercise

The majority of strategies designed to reduce dysglycaemia (primarily hypoglycaemia) during exercise involve adjustments to carbohydrate intake or insulin dosing. More recently consideration has been given to trying to prevent exercise-induced hypoglycaemia through augmentation of the counter-regulatory response to exercise. One very inventive way of has been to use a 10 second maximal effort sprint either before(29) or after(30) exercise. When such a sprint was performed at the beginning of a recovery period after 20 minutes of moderate exercise at 40% VO2 MAX there was no further fall in blood glucose levels, whereas a further significant fall in glucose levels was seen in the control condition when no sprint was performed(30). Performing the sprint was associated with an increase in catecholamine, cortisol, growth hormone and lactate levels, although it is not clear which of these were important for attenuating the fall in blood glucose. Interestingly, performing a 10 second sprint prior to similar exercise did not attenuate the drop in blood glucose levels seen during the exercise, but it did again attenuate the drop in blood glucose levels seen after exercise in the control group where no sprint was performed(29). This was in spite of a significant rise in circulating catecholamine and lactate levels immediately following the sprint.

Both of these studies therefore demonstrate the benefit of the 10 second maximal sprint performed either before or after exercise in preventing a post-exercise fall in blood glucose. This is particularly useful in that it provides a strategy which does not require pre-planning and which does not require the ingestion of significant amounts of carbohydrate. Interestingly, it is also possible that the benefit of performing a 10 second sprint prior to exercise was underestimated. The augmented catecholamine response was not shown to affect the fall in blood glucose levels during exercise but this was with blood glucose levels well above hypoglycaemic levels. Glucose levels fell on average by 3mmol/l during the exercise period, and given that exercise commenced at around 11mmol/l this suggests a fall from 11mmol/l to 8mmol/l. It has been demonstrated that carbohydrate oxidation is favoured over fat oxidation as the source of energy for exercising muscle at a blood glucose level of 11mmol/l when compared with 7 mmol/l(31), so participants in the trial would have been predisposed to preferential use of carbohydrate and hence a fall in blood glucose. This may have masked any benefit of the higher levels of catecholamines during the exercise period, and may explain why the benefit was only seen when glucose levels fell into the normal range during the period of recovery. It would be useful to see whether the 10s sprint might attenuate the fall in blood glucose levels during moderate exercise if blood glucose levels at the start of exercise were closer to the normal range.

Caffeine has been demonstrated to have beneficial effects on hypoglycaemia in type 1 diabetes, particularly nocturnal hypoglycaemia, when the hypoglycaemia is not specifically related to exercise(32;33). In particular, caffeine enhances the counter-regulatory hormone response to hypoglycaemia, as well as increasing symptoms accompanying hypoglycaemia which allow earlier treatment of hypoglycaemia and therefore reduce the chance of neurglycopaenia developing(34). With exercise, we have found that caffeine in doses of 5mg/kg taken 30 minutes prior to exercise reduces the need for carbohydrate treatment to prevent hypoglycaemia during exercise in people with type 1 diabetes(35). This is a preliminary study, but offers another strategy for the prevention of hypoglycaemia during exercise in type 1 diabetes which does not require much planning and does not involve the ingestion of extra carbohydrate.

Practical aspects of CSII and exercise

While CSII appears to provide an excellent solution to many of the problems posed by managing type 1 diabetes for sport and exercise, there are important practical considerations which need to be taken into account. Insulin pump therapy is expensive to provide relative to MDI, with a significant initial outlay for the pump and then ongoing costs for consumables. Improving athletic performance may not be seen as an appropriate justification for the extra cost, although a reduction in hypoglycaemia and improvement in metabolic control could be. Whilst the pumps are robust, it may not be practical to wear them for some contact sports because of the risk of damage even if they are carried in a protective case. There are some reports that newer patch pumps can be placed in locations on the body where they are protected from damage (e.g. the inner thigh), but these may not always be ideal locations for insulin delivery. Participation in such sports will also increase the risk of cannula displacement, which carries the risk of ketoacidosis if not detected early enough. Not all pumps are adequately waterproof for swimming, and increased exposure to treated swimming pool water or sea water may reduce the useful life of any waterproof seals. The risk to the pump can be reduced by keeping in a waterproof container while in the water, but this reduces access to the pump and may result in excessive bulkiness.

Summary and practical advice

The importance of insulin in the regulation of fuel production during exercise means that there is a significant risk of dysglycaemia during exercise in type 1 diabetes where insulin is exogenously administered. The flexibility in basal insulin infusion afforded by CSII is an attractive solution to the problem of hypoglycaemia with aerobic exercise, in theory at least allowing the person with diabetes to approximate more closely the metabolic state during exercise which is seen in those without diabetes. In doing this it is hoped that performance will be optimised.

It is likely that insulin infusion rates should be reduced prior to exercise, as circulating insulin levels at the start of exercise are a predictor of hypoglycaemia during exercise. Exactly when this reduction should be made is not clear. For practical purposes, based on the pharmacokinetics of the rapid-acting analogue insulins used in CSII therapy, the reduction should probably be made 30-45 minutes prior to the start of exercise. The exact amount by which basal infusion should be reduced is also not clear, but based on available evidence it is likely to be somewhere between 50 and 100% (i.e. cessation of basal insulin infusion). It is not clear when a normal basal rate should be restarted, but doing so at the end of exercise may reduce the risk of post-exercise hyperglycaemia (although, as detailed below, later reductions may be required for the avoidance of nocturnal hypoglycaemia). Complete removal of the insulin pump at the start of aerobic exercise is a strategy that will help to avoid hypoglycaemia, but at the expense of hyperglycaemia. It is possible that this will result in an impairment of athletic performance, and therefore may not be the best solution for those in whom performance is of paramount importance.

CGM can provide useful information regarding blood glucose levels during exercise, although using standard real-time CGM there is still a significant risk of hypoglycaemia in aerobic exercise when blood glucose levels fall rapidly. However, strategies are being developed to help to combat this. Low-glucose thresholds need to be set significantly higher than the minimum glucose level hoped for, with benefit seen for alarms used to trigger carbohydrate replacement and checking of capillary glucose when glucose falls 5.5mmol/l. An algorithm used to guide carbohydrate replacement taking into account both the level of interstitial glucose and its rate of change may point the way to how this technology could best be employed in the future to help to avoid hypoglycaemia during exercise.

Nocturnal hypoglycaemia following exercise is a well-recognised phenomenon in type 1 diabetes. Reducing the CSII basal rate by 20% between 9pm and 3am has been shown to reduce the risk of this in children without increasing the risk of moring hyperglycaemia. This result may well transfer to adults, although the timing of the reduction may need to be altered due to adults going to sleep later in the day. Low glucose suspend insulin pumps have been shown to be useful in reducing nocturnal hypoglycaemia in those most at risk of this, and this may be a useful technique to avoid nocturnal hypoglycaemia following exercise.

Augmenting the counter-regulatory hormone response to exercise offers an alternative means of avoiding hypoglycaemia which can help to avoid the significant planning often required for the adjustment of insulin doses. It may also help to avoid, or at least reduce, the need for extra carbohydrate supplementation which can be problematic if one of the aims of exercise is weight control. A 10 second maximal sprint either immediately before or after exercise attenuates a post-exercise fall in blood glucose. While it has not been shown to attenuate the fall in glucose during exercise, this was tested during exercise in conditions of hyperglycaemia when carbohydrate oxidation is favoured. High doses of caffeine have also been shown to reduce the need for carbohydrate supplementation to avoid hypoglycaemia during exercise.

Areas for future research

Further research is needed to identify the optimal way to manage CSII for exercise. Firstly, a clearer understanding of the pharmacokinetics and pharacodynamics of rapid-acting insulin analogues when used for CSII will be useful to understand this. Further research is also required to identify both the optimal time to alter basal insulin rate prior to exercise, and also to identify exactly what this alteration should be to permit athletes with diabetes to optimise their performance. Furthermore, it would be helpful to test whether restarting the normal basal rate prior to finishing an exercise session would avoid post-exercise hyperglycaemia, and may permit greater reductions to (or even cessation of) basal insulin infusion to be made prior to exercise without subsequent hyperglycaemia. It may well be that the alterations to basal rate which provide the best approximation to normal physiology will provide the optimum means of managing CSII for exercise, but this also requires further testing.

More detailed analysis of strategies using CGM to guide carbohydrate replacement during exercise is required, to see whether the early promise shown by these strategies can be fulfilled. The use of low-glucose suspend technology to avoid nocturnal hypoglycaemia following exercise should also be assessed, as well as the effect this has on the counter-regulatory response to any subsequent exercise. Assessment of the effect of the 10s during euglycaemic exercise would be interesting, to see whether such a strategy might be able to protect against the fall in blood glucose in these conditions. Similarly, using other means to augment the counter-regulatory hormone response to exercise might also offer further strategies to help to avoid hypoglycaemia during exercise without the requirement for significant carbohydrate ingestion.

Conclusions
The ability to adjust basal insulin infusion rates in CSII therapy means that, in our clinic, we now class CSII therapy as the gold standard in athletes with diabetes where it is practical. The difficulties in accurate assessment of rapidly-changing blood glucose levels, as during aerobic exercise, using CGM, mean that closed-loop insulin delivery is likely to remain difficult in this context for some time. However, individuals can use current knowledge to develop extremely successful strategies for the management of diabetes for sport and exercise using CSII. Real-time CGM is likely to play an increasingly important role in this, permitting accurate carbohydrate replacement based on individual requirements. Managing diabetes for sport and exercise is not easy, but improvements in the available technology and our understanding of how best to use it can only help to increase the numbers of successful athletes with diabetes.

Reference List

(1) Heise T, Nosek L, Ronn BB, Endahl L, Heinemann L, Kapitza C, et al. Lower within-subject variability of insulin detemir in comparison to NPH insulin and insulin glargine in people with type 1 diabetes. Diabetes 2004 Jun;53(6):1614-20.
(2) Lepore M, Pampanelli S, Fanelli C, Porcellati F, Bartocci L, Di VA, et al. Pharmacokinetics and pharmacodynamics of subcutaneous injection of long-acting human insulin analog glargine, NPH insulin, and ultralente human insulin and continuous subcutaneous infusion of insulin lispro. Diabetes 2000 Dec;49(12):2142-8.
(3) Ratner RE, Hirsch IB, Neifing JL, Garg SK, Mecca TE, Wilson CA. Less hypoglycemia with insulin glargine in intensive insulin therapy for type 1 diabetes. U.S. Study Group of Insulin Glargine in Type 1 Diabetes. Diabetes Care 2000 May;23(5):639-43.
(4) Hermansen K, Fontaine P, Kukolja KK, Peterkova V, Leth G, Gall MA. Insulin analogues (insulin detemir and insulin aspart) versus traditional human insulins (NPH insulin and regular human insulin) in basal-bolus therapy for patients with type 1 diabetes. Diabetologia 2004 Apr;47(4):622-9.
(5) Riddell M, Perkins BA. Exercise and glucose metabolism in persons with diabetes mellitus: perspectives on the role for continuous glucose monitoring. J Diabetes Sci Technol 2009 Jul;3(4):914-23.
(6) Brazeau AS, Rabasa-Lhoret R, Strychar I, Mircescu H. Barriers to physical activity among patients with type 1 diabetes. Diabetes Care 2008 Nov;31(11):2108-9.
(7) Riddell MC, Milliken J. Preventing Exercise-Induced Hypoglycemia in Type 1 Diabetes Using Real-Time Continuous Glucose Monitoring and a New Carbohydrate Intake Algorithm: An Observational Field Study. Diabetes Technol Ther 2011 May 20.
(8) Edelmann E, Staudner V, Bachmann W, Walter H, Haas W, Mehnert H. Exercise-induced hypoglycaemia and subcutaneous insulin infusion. Diabet Med 1986 Nov;3(6):526-31.
(9) Sonnenberg GE, Kemmer FW, Berger M. Exercise in type 1 (insulin-dependent) diabetic patients treated with continuous subcutaneous insulin infusion. Prevention of exercise induced hypoglycaemia. Diabetologia 1990 Nov;33(11):696-703.
(10) Admon G, Weinstein Y, Falk B, Weintrob N, Benzaquen H, Ofan R, et al. Exercise with and without an insulin pump among children and adolescents with type 1 diabetes mellitus. Pediatrics 2005 Sep;116(3):e348-e355.
(11) Tsalikian E, Kollman C, Tamborlane WB, Beck RW, Fiallo-Scharer R, Fox L, et al. Prevention of hypoglycemia during exercise in children with type 1 diabetes by suspending basal insulin. Diabetes Care 2006 Oct;29(10):2200-4.
(12) Heise T, Nosek L, Spitzer H, Heinemann L, Niemoller E, Frick AD, et al. Insulin glulisine: a faster onset of action compared with insulin lispro. Diabetes Obes Metab 2007 Sep;9(5):746-53.
(13) Arnolds S, Rave K, Hovelmann U, Fischer A, Sert-Langeron C, Heise T. Insulin glulisine has a faster onset of action compared with insulin aspart in healthy volunteers. Exp Clin Endocrinol Diabetes 2010 Oct;118(9):662-4.
(14) Battelino T, Phillip M, Bratina N, Nimri R, Oskarsson P, Bolinder J. Effect of continuous glucose monitoring on hypoglycemia in type 1 diabetes. Diabetes Care 2011 Apr;34(4):795-800.
(15) Effectiveness of continuous glucose monitoring in a clinical care environment: evidence from the Juvenile Diabetes Research Foundation continuous glucose monitoring (JDRF-CGM) trial. Diabetes Care 2010 Jan;33(1):17-22.
(16) Iscoe KE, Campbell JE, Jamnik V, Perkins BA, Riddell MC. Efficacy of continuous real-time blood glucose monitoring during and after prolonged high-intensity cycling exercise: spinning with a continuous glucose monitoring system. Diabetes Technol Ther 2006 Dec;8(6):627-35.
(17) Wilson DM, Beck RW, Tamborlane WV, Dontchev MJ, Kollman C, Chase P, et al. The accuracy of the FreeStyle Navigator continuous glucose monitoring system in children with type 1 diabetes. Diabetes Care 2007 Jan;30(1):59-64.
(18) Boyne MS, Silver DM, Kaplan J, Saudek CD. Timing of changes in interstitial and venous blood glucose measured with a continuous subcutaneous glucose sensor. Diabetes 2003 Nov;52(11):2790-4.
(19) Buckingham BA, Kollman C, Beck R, Kalajian A, Fiallo-Scharer R, Tansey MJ, et al. Evaluation of factors affecting CGMS calibration. Diabetes Technol Ther 2006 Jun;8(3):318-25.
(20) Iscoe KE, Davey RJ, Fournier PA. Increasing the low-glucose alarm of a continuous glucose monitoring system prevents exercise-induced hypoglycemia without triggering any false alarms. Diabetes Care 2011 Jun;34(6):e109.
(21) McMahon SK, Ferreira LD, Ratnam N, Davey RJ, Youngs LM, Davis EA, et al. Glucose requirements to maintain euglycemia after moderate-intensity afternoon exercise in adolescents with type 1 diabetes are increased in a biphasic manner. J Clin Endocrinol Metab 2007 Mar;92(3):963-8.
(22) Tsalikian E, Mauras N, Beck RW, Tamborlane WV, Janz KF, Chase HP, et al. Impact of exercise on overnight glycemic control in children with type 1 diabetes mellitus. J Pediatr 2005 Oct;147(4):528-34.
(23) Maran A, Pavan P, Bonsembiante B, Brugin E, Ermolao A, Avogaro A, et al. Continuous glucose monitoring reveals delayed nocturnal hypoglycemia after intermittent high-intensity exercise in nontrained patients with type 1 diabetes. Diabetes Technol Ther 2010 Oct;12(10):763-8.
(24) Sandoval DA, Guy DL, Richardson MA, Ertl AC, Davis SN. Effects of low and moderate antecedent exercise on counterregulatory responses to subsequent hypoglycemia in type 1 diabetes. Diabetes 2004 Jul;53(7):1798-806.
(25) Taplin CE, Cobry E, Messer L, McFann K, Chase HP, Fiallo-Scharer R. Preventing post-exercise nocturnal hypoglycemia in children with type 1 diabetes. J Pediatr 2010 Nov;157(5):784-8.
(26) Raju B, Arbelaez AM, Breckenridge SM, Cryer PE. Nocturnal hypoglycemia in type 1 diabetes: an assessment of preventive bedtime treatments. J Clin Endocrinol Metab 2006 Jun;91(6):2087-92.
(27) Choudhary P, Shin J, Wang Y, Evans ML, Hammond PJ, Kerr D, et al. Insulin pump therapy with automated insulin suspension in response to hypoglycemia: reduction in nocturnal hypoglycemia in those at greatest risk. Diabetes Care 2011 Jul 13.
(28) Galassetti P, Tate D, Neill RA, Richardson A, Leu SY, Davis SN. Effect of differing antecedent hypoglycemia on counterregulatory responses to exercise in type 1 diabetes. Am J Physiol Endocrinol Metab 2006 Jun;290(6):E1109-E1117.
(29) Bussau VA, Ferreira LD, Jones TW, Fournier PA. A 10-s sprint performed prior to moderate-intensity exercise prevents early post-exercise fall in glycaemia in individuals with type 1 diabetes. Diabetologia 2007 Sep;50(9):1815-8.
(30) Bussau VA, Ferreira LD, Jones TW, Fournier PA. The 10-s maximal sprint: a novel approach to counter an exercise-mediated fall in glycemia in individuals with type 1 diabetes. Diabetes Care 2006 Mar;29(3):601-6.
(31) Jenni S, Oetliker C, Allemann S, Ith M, Tappy L, Wuerth S, et al. Fuel metabolism during exercise in euglycaemia and hyperglycaemia in patients with type 1 diabetes mellitus--a prospective single-blinded randomised crossover trial. Diabetologia 2008 Aug;51(8):1457-65.
(32) Watson J, Kerr D. The best defense against hypoglycemia is to recognize it: is caffeine useful? Diabetes Technol Ther 1999;1(2):193-200.
(33) Richardson T, Thomas P, Ryder J, Kerr D. Influence of caffeine on frequency of hypoglycemia detected by continuous interstitial glucose monitoring system in patients with long-standing type 1 diabetes. Diabetes Care 2005 Jun;28(6):1316-20.
(34) Debrah K, Sherwin RS, Murphy J, Kerr D. Effect of caffeine on recognition of and physiological responses to hypoglycaemia in insulin-dependent diabetes. Lancet 1996 Jan 6;347(8993):19-24.
(35) Gallen IW, Ballav C, Lumb AN, Carr J. Caffeine supplementation reduces exercise induced decline in blood glucose and subsequent hypoglycaemia in adults with type 1 diabetes (T1DM) treated with multiple daily insulin injection (MDI). Poster 1184-P presented at the 70th Scientific Sessions of the American Diabetes Association, Orlando, Florida: 25-29 June 2010.