sign in sign up
<<
Home , Insulin pump

Editorials POINT-COUNTERPOINT (SEE P. 1453)

Are Pumps Underutilized in Type 1 ? Yes

ontinuous subcutaneous insulin in- also a key determinant of cost effective- glycemia without stipulating what “re- fusion (CSII), popularly called insu- ness, the estimates of which are limited peated” means. A recent cross-sectional C lin pump therapy, has evolved from (8). study of 1,076 consecutive adult type 1 its invention in the 1970s as an experi- diabetic patients treated according to mental treatment designed to test the re- The problem of severe modern guidelines at four centers in the lationship between glycemic control and Reduction of severe hypoglycemia (where U.K. and Denmark is informative in this diabetic tissue complications (1) to its third-party assistance is required for re- respect (16). As many as 21% had two or present status as a routine therapy for se- suscitative measures) in more severe hypoglycemic episodes in the lected type 1 diabetic patients (2). How- was first identified in the mid-1980s, ei- previous year, as compared with a mean ever, the use of insulin pump therapy ther with matched groups of injection- or of 13% suffering severe hypoglycemia in varies markedly throughout the world; pump-treated type 1 diabetic subjects (9) the previous year over a 12-year study of there are some notable high-use coun- or in a randomized controlled trial of mul- an intensified insulin program in Ger- tries, e.g., the U.S. and Israel, where it is tiple insulin injection (MDI) therapy (and many (17). However, the definition of hy- estimated that ϳ20% of type 1 diabetic nonoptimized injection therapy) versus poglycemia here was the requirment of patients use CSII (manufacturers’ esti- CSII (the Oslo Study [10]). Many subse- intravenous glucose or injec- mates), whereas in other countries, such quent studies have confirmed the hypo- tion, so the frequency of hypoglycemia as the U.K. and Denmark, ϳ1% use glycemia-reducing effect of insulin requiring any assistance would be higher. pump therapy (3). pumps (2,11–13), with typical reduc- The reasons for this variation include tions in frequency of severe hypoglycemia Yet the distribution of severe hypoglyce- the availability of financial resources and ϳ mia in the type 1 diabetic population is of 70% compared with MDI. ϳ health care professionals to supervise CSII The clinical impact of this beneficial extremely skewed, with 5% of patients and a lack of knowledge on the effective- effect was undervalued until recently be- having 70% of all episodes (16). This ness of CSII (3), but there is also disagree- cause of the untypically high frequency of therefore represents a reasonable mini- ment on which diabetic subjects should severe hypoglycemia reported in the mum target group, although many more be treated with CSII, as evidenced by both pump-treated subjects in the Diabetes might suffer hypoglycemia, which is dis- the different intercountry usage and the Control and Complications Trial (DCCT) abling for them. large number of reasons for starting insu- (0.5 vs. ϳ0.1–0.25 episodes/patient-year lin pump therapy (4–6). in other pump studies) (14). The expla- As noted by Schade and Valentine (7), nation for this discrepancy is unclear but The impact of long-acting insulin “the challenge for the health care provider may relate to a large number of centers in analogs on severe hypoglycemia is to select the diabetic patients who will the DCCT that were using pump therapy There is no strong evidence that using really benefit from pump usage.” What for the first time. glargine or detemir (with their proportion and what types of type 1 dia- flatter profile and improved predictabil- betic patients should then be offered a Frequency estimation for severe ity), instead of isophane-based regimens, trial of CSII on clinical grounds alone, hypoglycemia will reduce the frequency of severe hypo- leaving aside the legitimate issues of sup- Estimating the proportion of type 1 dia- glycemia (18). Although minor hypogly- ply on the basis of patient preference and betic subjects with severe hypoglycemia is cemia during the night is less with long- restrictions due to availability of funding difficult because it critically depends on acting analogs, there is no difference in and staffing? I shall argue that the target patient selection. Many factors influence the rate of severe hypoglycemia when an proportion best treated by CSII, or offered hypoglycemia frequency (15,16), includ- MDI regimen using isophane as the basal a trial of CSII, can be derived from an ing the definition of severe hypoglycemia insulin is compared with either glargine- estimate of the effectiveness of this ther- (e.g., requiring any assistance or, specifi- based (18,19) or detemir-based (20–22) apy compared with the best insulin injec- cally intravenous glucose/glucagon injec- injection regimens. Severe hypoglycemia tion treatment for particular clinical tion), the type of treatment (intensive over extended periods has not been well problems in type 1 diabetes. versus conventional regimens), prevailing studied in randomized trials directly com- Most current guidelines (4,6) or re- glycemic level, diabetes duration, con- paring glargine-based regimens and CSII views of the evidence base on CSII (2) do comitant drug usage, alcohol intake, pres- not take into account recent studies on ence of autonomic neuropathy and renal because of the relatively short-term na- the effectiveness of pump therapy in the disease, smoking, educational level, and ture of the studies (23–25). However, putative target groups of hypoglycemia- history of previous hypoglycemia and hy- since severe hypoglycemia does not ap- prone diabetes and the worst controlled poglycemia awareness. pear, based on current evidence, to be re- subjects or the possible impact of recently A further issue is the judgement about duced with MDI, based on long-acting introduced long-acting insulin analogs on what frequency of hypoglycemia is dis- insulin analogs compared with isophane the quality of control achievable with in- abling. Some guidelines for insulin pump regimens, the use of CSII to improve hy- jection therapy. The efficacy of CSII in the therapy (6) define this as the “repeated poglycemia frequency during MDI is still most appropriate groups of patients is and unpredictable occurrence” of hypo- justified.

DIABETES CARE, VOLUME 29, NUMBER 6, JUNE 2006 1449 Point-Counterpoint

The problem of and The impact of glargine and detemir The problem of the dawn elevated glycated HbA1c on MDI on hyperglycemia phenomenon Until recently, the belief was that the dif- Though there may be lowered fasting The dawn phenomenon refers to the rise ference in average glycemia achievable on blood glucose concentrations, overall gly- in blood glucose concentration in some pump therapy was relatively small com- cemia as measured by A1C is usually not diabetic patients occurring in the few pared with MDI (2,6,26). For example, a improved by glargine or detemir com- hours before breakfast, without preced- meta-analysis of 12 randomized con- pared with NPH-based injection regi- ing hypoglycemia; it is thought to be due trolled trials indicated that glycemic con- mens in type 1 diabetes (18,19,21,22,32). to a combination of trol on pump therapy was slightly but In studies comparing glargine with pump caused by surges in growth hormone dur- significantly better than on MDI, with a therapy, A1C or values were ing the night and insulin deficiency difference in HbA1c (A1C) of 0.5% and improved on CSII versus glargine caused by waning of the effects of the pre- mean blood glucose concentration of 1 (23,24,32), but in relatively well-controlled ceding evening’s insulin injection (33). mmol/l (26). However, recent work from subjects, A1C percentages were similar Increasing the evening long-acting insulin several groups, including a pooled analy- (25). Thus, the evidence to date indicates dose or delaying its injection to bedtime sis of randomized controlled trials, has that long-acting –based injec- to extend action are useful strategies for shown that the fall in A1C on switching tion regimens are not as effective as pump treating the dawn phenomenon, but both type 1 diabetic subjects who have failed to therapy in lowering glycemia in most can lead to nocturnal hypoglycemia. The achieve good control on MDI to CSII is poorly controlled type 1 diabetic patients, phenomenon can be successfully man- directly proportional to the initial A1C on and the target group with elevated A1C on aged by CSII because the basal insulin in- MDI (27–29). Thus, the best improve- injections suitable for a trial of CSII will fusion rate can be preset to increase ment is seen in the worst-controlled sub- remain at ϳ15%. during the dawn hours (33). jects (who are the likely candidates for The frequency of the dawn phenom- pump therapy), a fact that was obscured enon during MDI is difficult to judge. Rel- in previous trials of unselected, general The syndrome of hyperglycemia, atively few patients are referred to our type 1 diabetic patients without clinical blood glucose variability, and specialist pump clinic because of a problems (26). When, for example, the unpredictable hypoglycemia marked dawn phenomenon and probably starting A1C is 10% on MDI, the fall in Who are the patients who remain hyper- many can now be managed by glargine or A1C on switching to CSII is likely to be glycemic on MDI? It might be thought detemir regimens, which are often very ϳ2% but, in a relatively well-controlled effective at lowering fasting blood glucose that elevated blood glucose concentra- subject with an A1C of 7%, the difference without increasing hypoglycemia (18– tions can be obviated by increasing the in A1C could be Ͻ0.5% (28). 22). Moreover, the long-term clinical insulin dosage. However, the patients consequences of a marked dawn phe- with the highest A1C during MDI also nomenon are unclear. A mean fasting have the widest swings in blood glucose Frequency estimation for markedly plasma glucose concentration as high as levels, and it is probable that they (or their elevated A1C 10 mmol/l was associated with a mean A good estimate of the quality of control health care professionals) resist attempts A1C of only 7.5% in one study (19). Prob- that is achievable on MDI comes from to lower the mean glucose level for fear of ably, then, there will be relatively few pa- studies where therapy combines basal/ inducing hypoglycemia (28). This ex- tients with the dawn phenomenon who bolus insulin injection, frequent blood plains why many patients who are consid- will need to be treated by pump therapy. glucose self-monitoring, dietary advice, ered hypoglycemia prone have actually insulin dosage adjustment according to had few recent serious hypoglycemic ep- Patient choice and patient suitability meal composition and size, structured pa- isodes (30) but are characterized by fre- If patients were allowed to choose insulin tient education, and adequate contact and quent, unpredictable glycemic oscillations pump therapy as their routine treatment, advice from health care professionals. The and episodes of biochemical or moderate without respect to cost or whether there level at which glycemia is so elevated that hypoglycemia. They usually maintain a was a clinical problem with their diabetes CSII should be considered is debatable, high A1C. Both within- and between-day control on injection therapy, there would but, as an example, I have calculated the blood glucose variability is significantly im- undoubtedly be a large enrollment in- mean percentage of subjects with an A1C proved by switching from insulin injections crease. Quality of life is reported to be Ͼ9.5% from the distribution of reported to pump therapy (28,30). better during pump therapy than MDI values in a number of trials and surveys We do not know if such patients with (13), and in a survey in the U.S. in 2000, describing the injection regimen as MDI variable control always belong to the same more than half of the health care diabetes (12,24,28,30) or as “intensified,” with a population as those who are hyperglyce- specialists who themselves have type 1 di- description of dosage-adjusted basal- mic on best attempts with injection ther- abetes were being treated by pump ther- bolus therapy in the methods (13,17,31) apy, and glycemic predictability (if not apy (34). There are probably more now. as 15%. This estimate needs to be con- A1C) is often improved by glargine and There is need for much more research on firmed by a more extensive survey of the detemir (20–22). However, it seems the degree of improvement in lifestyle af- literature, but I suggest that this would likely that, after a period of attempting to forded by insulin pumps, but here I re- represent a reasonable first estimate of the improve control with MDI, at least 15% of strict estimates of individuals who might target population for a trial of pump ther- type 1 diabetic patients are markedly un- be treated by CSII to the clinical problems apy, i.e., the ϳ15% who remain very controlled, with either an elevated A1C or outlined above, which also might form poorly controlled (elevated A1C) after glycemic variability or both, and are at an improved basis for future cost- best attempts with MDI. least candidates for a trial of CSII. effectiveness analyses.

1450 DIABETES CARE, VOLUME 29, NUMBER 6, JUNE 2006 Point-Counterpoint

Conclusions might be best treated by CSII on clinical 8. Roze S, Valentine WJ, Zakrzewska KE, There are some 5% of type 1 diabetic sub- grounds. Neither do we know how Palmer AJ: Health-economic comparison jects treated by MDI with severe, recur- changes in pump technology in the com- of continuous subcutaneous insulin infu- rent hypoglycemia. At least another 5% ing decades will influence these difficul- sion with multiple daily injection for the suffer severe hypoglycemia at such a fre- ties: smaller and cheaper pumps may treatment of type 1 diabetes in the UK. Diabet Med 22:1239–1245, 2005 quency that it is markedly disabling to make pump usage more widespread but ϳ 9. Bending JJ, Pickup JC, Keen H: Frequency them. I estimate that 15% of type 1 di- more sophisticated and expensive devices of and hypoglyce- abetic subjects on MDI have the syn- may not. mic coma during treatment with continu- drome of markedly elevated A1C and I do not take “many” to mean “most” ous subcutaneous insulin infusion. Am J wide swings in blood glucose concentra- and do not believe, based on current evi- Med 79:685–691, 1985 tion, often with unpredictable, moderate dence, that the majority of type 1 diabetic 10. Dahl-Jørgensen K, Brinchman-Hansen O, (nonsevere) hypoglycemia. A small per- patients should be treated by CSII. How- Hanssen KF, Ganes T, Kierulf P, Smeland centage will have the dawn phenomenon. ever, there is a good evidence base for the E, Sandvik L, Aagenaes Ø: Effect of near- These clinical problems are at least as im- substantial minority (the many) who can- normoglycaemia for two years on the pro- portant in children as adults (35). not be well treated by MDI to be so man- gression of early : the Oslo Study. BMJ 293:1195–1199, 1986 According to present evidence, the pro- aged. 11. Bode BW, Steed RD, Davidson PC: Re- portion of subjects with severe hypogly- duction in severe hypoglycemia with cemia or unacceptable hyperglycemia JOHN C. PICKUP, DPHIL, FRCPATH long-term continuous subcutaneous in- who are improved by regimens using new sulin infusion in type 1 diabetes. Diabetes long-acting insulin analogs is likely to be From the Metabolic Unit, King’s College London Care 19:324–327, 1996 School of Medicine, Guy’s Hospital, London, U.K. 12. Boland EA, Grey M, Oesterle A, Fredrick- small. Address correspondence to Prof. John Pickup, Some patients are known to be un- Metabolic Unit, King’s College London School of son L, Tamborlane WV: Continuous sub- suitable for insulin pump treatment (2) Medicine, Guy’s Hospital, London SE1 9RT, U.K. cutaneous insulin infusion: a new way to because they are unable to perform pump E-mail: [email protected]. lower risk of severe hypoglycemia, im- procedures or are psychologically unsuit- J.C.P. has received honoraria for speaking en- prove metabolic control, and enhance gagements from Medtronic and Roche and receives able or simply decline this treatment op- coping in adolescents with type 1 diabe- funds for unrestricted research from Medtronic. tes. Diabetes Care 22:1799–1784, 1999 tion and prefer MDI. Even using the most DOI: 10.2337/dc06-0011 13. Linkeschova R, Raoul M, Bott U, Berger conservative estimate that this number is © 2006 by the American Diabetes Association. M, Spraul M: Less severe hypoglycaemia, as much as one-quarter of those with the better metabolic control, and improved above clinical problems, a reasonable quality of life in type 1 diabetes mellitus Acknowledgments—The author thanks the minimum target for those type 1 diabetic with continuous subcutaneous insulin in- Diabetes Foundation and the Engineering and fusion (CSII) therapy: an observational patients who should be offered a trial of Physical Sciences Research Council for addi- ϳ study of 100 consecutive patients fol- insulin pump therapy is therefore 15– tional grant support. 20% of type 1 diabetic subjects. This a lowed for a mean of 2 years. Diabet Med 19:746–751, 2002 percentage of patients similar to that al- ●●●●●●●●●●●●●●●●●●●●●●● ready treated as such in the U.S. and some 14. Diabetes Control and Complications Trial Group: Implementation of treatment pro- other countries. References 1. Pickup JC, Keen H, Parsons JA, Alberti tocols in the Diabetes Control and It must be emphasized that most KGMM: Continuous subcutaneous insu- Complications Trial. Diabetes Care 18: health care professionals recommend a lin infusion: an approach to achieving 361–376, 1995 sequential approach to selecting patients normoglycaemia. BMJ i:204–207, 1978 15. Cryer PE, Davis SN, Shamoon H: Hypo- for CSII on clinical grounds, with best ef- 2. Pickup J, Keen H: Continuous subcutane- glycemia in diabetes. Diabetes Care 26: forts first being applied to MDI (with ous insulin infusion at 25 years: evidence 1902–1912, 2003 long-acting insulin analogs if necessary) base for the expanding use of insulin 16. Pedersen-Bjergaard U, Pramming S, and with appropriate education and diet pump therapy in type 1 diabetes. Diabetes Heller SR, Wallace TM, Rasmusen AK, before offering a trial of CSII to those who Care 25:593–598, 2002 Jorgensen HV, Matthews DR, Hougaard fail to achieve satisfactory glycemic con- 3. Nøgaard K: A nationwide study of contin- P, Thorsteinsson B: Severe hypoglycaemia in 1076 adult patients with type 1 diabe- trol on such a regimen. I strongly recom- uous subcutaneous insulin infusion (CSII) in Denmark. Diabet Med 20:307– tes: influence of risk markers and selec- mend that this practice continue and that 311, 2003 tion. Diabetes Metab Res Rev 20:479–486, only after MDI has been tried and hypo- 4. Bode BW, Tamborlane WV, Davidson PC: 2004 glycemia and/or an elevated A1C persist Insulin pump therapy in the 21st century. 17. Plank J, Kohler G, Rakovak I, Semlitsch should a trial of CSII be considered. Postgrad Med 111:69–77, 2002 BM, Horvath K, Bock G, Kraly B, Pieber I do not underestimate the organiza- 5. American Diabetes Association: Continu- TR: Long-term evaluation of a structured tional, financial, staffing, and political ous subcutaneous insulin infusion. Diabe- outpatient education programme for in- challenges that that must be faced in tes Care 27 (Suppl. 1):S110, 2004 tensified insulin therapy in patients with meeting this target in many countries or 6. National Institute for Clinical Excellence: type 1 diabetes: a 12-year follow-up. Dia- the need for continuing research into Guidance on the use of continuous sub- betologia 47:1370–1375, 2004 quality of life, cost effectiveness, and cutaneous insulin infusion for diabetes. In 18. Warren E, Weatherley-Jones E, Chilcott J, Technology Appraisal Guidance. No. 57. Beverley C: Systematic review and eco- other possible benefits of CSII. Some London, National Institute for Clinical nomic evaluation of a long-acting insulin countries may not be able to achieve this Excellence, 2003, p. 1–23 analogue, insulin glargine. Health Technol suggested level of pump usage in the near 7. Schade DS, Valentine V: To pump or not Assess 8:1–57, 2004 future for several reasons, but that should to pump (Editorial). Diabetes Care 25: 19. Raskin P, Klaff L, Bergenstal R, Halle J-P, not influence our estimates of those who 2100–2102, 2002 Donley D, Mecca T: A 16-week compari-

DIABETES CARE, VOLUME 29, NUMBER 6, JUNE 2006 1451 Point-Counterpoint

son of the novel insulin analog insulin 24. Hirsch IB, Bode BW, Garg S, Lane WS, 29. Hammond P: NICE guidance on insulin glargine (HOE 901) and NPH human in- Sussman A, Hu P, Santiago OM, Kola- pump therapy: time for a re-appraisal? sulin used with in patients czynski JW: Continuous subcutaneous Pract Diabet Int 22:115–116, 2005 with type 1 diabetes. Diabetes Care 23: insulin infusion (CSII) of 30. Pickup JC, Kidd J, Burmiston S, Yemane 1666–1671, 2000 versus multiple daily injections of insulin N: Effectiveness of continuous subcutane- 20. Hermansen K, Fontaine P, Kukolja KK, aspart/insulin glargine in type 1 diabetic ous insulin infusion in hypoglycaemia- Peterkova V, Leth G, Gall MA: Insulin an- patients previously untreated with CSII. prone type 1 diabetes: implications for alogues (insulin detemir and insulin as- Diabetes Care 28:533–538, 2005 NICE guidelines. Pract Diabet Int 22:10– part) versus traditional human insulins 25. Bolli GB, Capani F, Home PD, Kerr D, 14, 2005 (NPH and regular human insulin) in bas- Thomas R, Torlone E, Selam J-L, Sola- 31. DAFNE Study Group: Training and flexi- al-bolus therapy for patients with type 1 Gazagnes A, Vitacollona E: Comparison ble, intensive insulin management to en- diabetes. Diabetologia 47:622–629, 2004 of a multiple daily injection regimen with able dietary freedom in people with type 1 21. Home P, Bartley P, Russell-Jones D, once daily insulin glargine basal insulin diabetes: dose adjustment for normal eat- Hanaire-Broutin H, Heeg JE, Abrams P, and mealtime lispro, to continuous sub- ing (DAFNE) randomised controlled trial. Landin-Olsson M, Hylleberg B, Lang H, cutaneous insulin infusion: a randomised, BMJ 325:746–756, 2002 Draeger E, the Study to Evaluate the Ad- open, parallel study. Diabetes 53 (Suppl. 32. Lepore G, Dodesini AR, Nosari I, Trevisan ministration of Detemir Insulin Efficacy, 2):A107, 2004 R: Both continuous subcutaneous insulin Safety and Suitability (STEADINESS) 26. Pickup JC, Mattock MB, Kerry S: Gly- infusion and a multiple daily injection Study Group: Insulin detemir offers im- caemic control with continuous subcu- regimen with glargine as basal insulin are proved glycemic control compared with taneous insulin infusion compared to equally better than traditional multiple NPH insulin in people with type 1 diabe- intensive insulin injection therapy in type tes: a randomized clinical trial. Diabetes 1 diabetes: meta-analysis of randomised daily injection treatment (Letter). Diabetes Care 27:1081–1087, 2004 controlled trials. BMJ 324:705–708, 2002 Care 26:1321–1322, 2003 22. Russel-Jones D, Simpson R, Hylleberg B, 27. Retnakaran R, Hochman J, DeVries JH, 33. Koivisto VA, Yki-Harvinen H, Helve E, Draeger E, Bolinder J: Effects of QD insu- Hanaire-Broutin H, Heine RJ, Melki V, Karonen S-L, Pelkonen R: Pathogenesis lin detemir or neutral proamine Hage- Zinman B: Continuous subcutaneous in- and prevention of the dawn phenomenon dorn on blood glucose control in patients sulin infusion versus multiple daily injec- in diabetic patients treated with CSII. Di- with type 1 diabetes mellitus using a bas- tions: the impact of baseline A1c. Diabetes abetes 35:78–82, 1986 al-bolus regimen. Clin Ther 26:724–736, Care 27:2590–2596, 2004 34. Graff MR, Rubin RR, Walker EA: How di- 2004 28. Pickup JC, Kidd J, Burmiston S, Yemane abetes specialists treat their own diabetes: 23. Doyle EA, Weinzimer SA, Steffen AT, N: Determinants of glycaemic control in findings from a study of the AADE and Ahern JA, Vincent M, Tamborlane WV: A type 1 diabetes during intensified therapy ADA membership. Diabetes Educ 26:460– randomized, prospective trial comparing with multiple daily insulin injections or 467, 2000 the efficacy of continuous subcutaneous continuous subcutaneous insulin infu- 35. Jones TW, Davis EA: Hypoglycemia in insulin infusion with multiple daily injec- sion: importance of blood glucose vari- children with type 1 diabetes: current is- tions using insulin glargine. Diabetes Care ability. Diabetes Metab Res Rev [Epub sues and controversies. Pediatr Diabetes 27:1554–1558, 2004 ahead of print] 4:143–150, 2003

1452 DIABETES CARE, VOLUME 29, NUMBER 6, JUNE 2006