Glycaemic control with continuous subcutaneous insulin infusion compared with intensive insulin injections in patients with type 1 diabetes: meta-analysis of randomised controlled trials
BMJ 2002; 324 doi: https://doi.org/10.1136/bmj.324.7339.705 (Published 23 March 2002) Cite this as: BMJ 2002;324:705All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
I have been a Type I diabetic for 38 years and have been an insulin
pump user only for the last two. My A1Cs have dropped from 6.5 to 5.1
since using the pump. One of the biggest advantages that my pump gives me
over a daily regime of pre-meal injections of Humalog and a bedtime
injection of Lantus is: I can change the basal rate to match my
requirements during the day. Once I give myself a nightly shot of Lantus,
I can't reduce the amount if I get more exercise.
With the pump, I can increase or decrease my basal rate depending on the
circumstances during the day. I cannot do the same with long-acting
insulin. I can now sleep in and program my pump to cover the dawn effect.
I can also program an extended bolus to cover high carb or high fat meals,
instead of juggling 3 different types of insulin (Lantus, Regular, and
Humalog). If I don't eat as much as I set the extended bolus for, I can
stop the bolus and reprogram for less. I have had far less hypoglycemic
episodes in the last 2 years on the pump.
Type I diabetes is a HEALTH MANAGEMENT NIGHTMARE for the unfortunate
person that has to deal with this disease or the parent with a diabetic
child. I will never willingly go back to multiple injections. The medical
profession has made the empty promises of a cure for Type I for 40 years,
and the best its done is the insulin pump (which for the record is a long
way from a cure). For the team of researchers that say a pump isn't
necessary, let's surgically remove their pancreas and put them on multiple
injection therapy and see how fast they change their recommendations. It
looks like that study was funded by medical insurance providers.
Competing interests:
None declared
Competing interests: No competing interests
Dear sir,
Although I have not read your paper in detail I just wonder whether you
have considered the importance of use of the more modern insulins
(Humalog or Novo Rapid) for differences in effects when comparing effects
of pumps vs insulin injections.
With best regards
M.Kollind, MD, PhD
Competing interests: No competing interests
Dr. Pickup's excellent meta analysis of the effect of insulin pumps
on metabolic control is very timely. More and more individuals with type 1
diabetes are using insulin pumps. Dr. Pickups article shows that blood
glucose and HbA1C levels are improved and if sustained will reduce the
risk of long term complications. I take issue with his final conclusion
that pumps should be limited to individuals having difficulty in
controlling their diabetes by injections. His data only addressed
metabolic control not who should or should not use an insulin pump.
Insulin pumps are clearly the closest we can come at the present time to
mimicking insulin secretion by the normal pancreas. I believe more
individuals should be (and will be) using pumps because if Dr Pickup's
data is correct pump use should be encouraged and used more widely. The
major deterrent is the cost.
Competing interests: No competing interests
You explain to your young child that he/she must stop her current
activity or wake from sound sleep and eat, hungry or not. You tell your
child, sure, you can have more to eat, but you must have an extra shot.
You call your child awake for school every morning and when he/she
doesn't answer immediately, feel your heart race while "dead in bed"
syndrome runs through your mind along with all the blood sugar readings,
food and exercise the child had for the previous 24 hours.
Don't take it personally when your child yells, cries, and screams
that he/she hates you, hates diabetes, and wishes he/she was dead; tell
yourself it's just low blood sugar talking.
I hope that after you have walked a day in the shoes of a parent of a
child with diabetes, you will not have insurance turn down your request
for a pump because the circumstances are not special enough.
Competing interests: No competing interests
Pickup et al are to be commended with their thorough meta-analysis of
continuous subcutaneous insulin (CSII) as compared to intensified insulin
injection therapy.1 Their main conclusion is that glycemic control,
expressed as HbA1c, is better with CSII, with a difference of 0.51%
favouring CSII. We have three comments, which may cast a different light
on this conclusion.
First, the results of the largest study in the meta-analysis, the so
called Dusseldorf study, needs careful interpretation.2 The authors have
estimated this study to show a positive result for continuous subcutaneous
insulin infusion, with an advantage in percentage glycated haemoglobin of
0.68%. This number will be an estimation made from a graph, as exact data
are not given in the original paper. This 0.68% difference seems to be a
correct interpretation of the six month data of this study. However, as
the authors acknowledge in their paper, the total study duration was two
years. At 12, 18 and 24 months the ‘advantage’ of continuous subcutaneous
insulin infusion can be estimated to be 0.35%, -0.1% and -0.2%,
respectively. Thus, another interpretation of this one study, representing
918 months of the meta-analysis’ total of 2522 patient months of pump
treatment, will have a substantial impact on the overall outcome of the
meta-analysis.
Second, the authors did not include the study of Reeves et al in
their analysis. This study, albeit small, did not show a difference in
HbA1c between intensified injection therapy and insulin pump therapy.3
Third, modified rapid-acting insulins have have recently been shown
to be advantageous with respect to HbA1c.4;5 Therefore, the most relevant
comparison is the one between continuous subcutaneous insulin infusion and
multiple injection therapy, both using rapid-acting insulin analogues.
Only two studies making this comparison have been published. The first
found a 0.35% lower HbA1c with insulin pump treatment as compared to
injection therapy in 41 patients using a cross-over design.6 However,
patients had been on insulin pump therapy using unmodified human insulin
for a mean of 5.5 years before entering the trial, limiting the external
validity of this study. The second has been published too recently to be
included in the meta-analysis, but was not able to show a difference in
HbA1c over a nine month period in 27 patients.7
Therefore, we feel the jury is still out on the question whehter
insulin pump treatment in type 1 diabetes lowers HbA1c. Two large multi-
centre trials on insulin pump therapy as compared to optimized injection
schemes with rapid-acting analogues have been recently completed and will
hopefully provide clinically useful information.
1. Pickup J, Mattock M, Kerry S. Glycaemic control with continuous
subcutaneous insulin infusion compared with intensive insulin injections
in patients with type 1 diabetes: meta-analysis of randomised controlled
trials. BMJ 2002;324:705.
2. The Dusseldorf Study Group. Comparison of continuous subcutaneous
insulin infusion
and intensified conventional therapy in the treatment of type I diabetes:
a two-year randomized study. Diab.Nutr.Metab. 1990;3:203-13.
3. Reeves ML, Seigler DE, Ryan EA, Skyler JS. Glycemic control in
insulin-dependent diabetes mellitus. Comparison of outpatient intensified
conventional therapy with continuous subcutaneous insulin infusion.
Am.J.Med. 1982;72:673-80.
4. Raskin P, Guthrie RA, Leiter L, Riis A, Jovanovic L. Use of
insulin aspart, a fast-acting insulin analog, as the mealtime insulin in
the management of patients with type 1 diabetes. Diabetes Care 2000;23:583
-8.
5. Home PD, Lindholm A, Riis A. Insulin aspart vs. human insulin in
the management of long-term blood glucose control in Type 1 diabetes
mellitus: a randomized controlled trial. European Insulin Aspart Study
Group. Diabet.Med. 2000;17:762-70.
6. Hanaire-Broutin H, Melki V, Bessieres-Lacombe S, Tauber JP.
Comparison of continuous subcutaneous insulin infusion and multiple daily
injection regimens using insulin lispro in type 1 diabetic patients on
intensified treatment: a randomized study. The Study Group for the
Development of Pump Therapy in Diabetes. Diabetes Care 2000;23:1232-5.
7. Tsui E, Barnie A, Ross S, Parkes R, Zinman B. Intensive Insulin
Therapy With Insulin Lispro: A randomized trial of continuous subcutaneous
insulin infusion versus multiple daily insulin injection. Diabetes Care
2001;24:1722-7.
Declaration of competing interests: JHD has received reimbursement
for attending a scientific meeting, and RJH has received funding for
research from Disetronic Medical Systems BV, an insulin pump manufacturer.
Competing interests: No competing interests
While I appreciate the viewpoints of those responding to the
scientific study, my experience, both as a scientist and as a diabetic
would seem to favor the conclusions of Dr. Pickup. From my investigations
and through talking with my own endocrinologist, I have come to the
conclusion that for most brittle diabetics and those with wildly
flucuating blood glucose values, an insulin pump would appear to be the
best choice of treatment. As one who maintained a hemoglobin A1c level of
4-4.5 during pregnancy with twins while talking 2 injections per day, it
is possible to obtain and maintain an almost "non-diabetic" level of
glycemic control. I tend to think of myself in stable terms and not one
who would benefit from pump therapy. I assumed from Dr. Pickup's study
that his conclusions were in line with equating "special needs" diabetics
with brittle, unstable diabetics.
Competing interests: No competing interests
Dr. Pickup is a well respected expert and advocate of pumps, but I
must disagree that pumps should be reserved for those with unusual
problems with diabetes. Once again, the 'experts' are leaving an element
out of the control equation that is as important as preventing
complications. People with diabetes should be able to expect and attain as
high a quality of life as people without diabetes.
As an individual with diabetes, I am constantly angered that medical
professionals expect that I will have to eat at certain times, expect that
I will pass out once or twice a year, expect that I will, sooner or later,
have numb toes and eye damage. This is not acceptable; nor is it
acceptable for someone to have to leave a business meeting or social event
in order to take insulin.
Quality of life is, indeed, the issue we need to target in diabetes
care, because preventing complications is a subset of that issue. I
experienced a highly tailored, quite effective MDI insulin regime that
gave me much more flexibility compared to many diabetics. But I still had
to leave meetings, set my alarm at weekends, plan exercise around peaks of
long acting insulin. My injected regime was remarkable, but it gave
nowhere near the quality of life I receive with my pump. And in
conjunction with that improved quality of life I have also benefitted from
smaller insulin doses, a higher carb-to-insulin ratio, and an improved
glycosylated haemoglobin. Every pumper I know has similar experiences.
Certainly there are diabetics who can not or should not use pumps;
screening programs can easily identify people who would not benefit from a
pump. The rest of us, whether we have 'special' or 'unusual' problems with
diabetes, or fairly good control, should be given the option of using a
pump because of the manifold benefits it brings that are far beyond
prevention of microvascular complications, benefits which have a much more
immediate and beneficial impact.
Competing interests: No competing interests
Having recently changed to pump therapy for my 7 year old child, and
programmed his basal rates, it has become clear to me why his prior
injection regimens had failed. The basal insulin amounts required to keep
his blood glucose constant, in the absence of food, range from .2 units
per hour during the period of highest insulin sensitivity (3 a.m. to 7
a.m.), to 1.6 units per hour during the highest period of insulin
resistance in pre-pubertal children, 9 p.m. to 12 p.m.). There has been
inordinate focus on what a patient eats, as compared to controlling the
hepatic output in periods between meals, and during the night, which are
by far more significant.
Accordingly, I take exception to your statement that pump therapy is
indicated only for people with special problems. There is no injection
regime that I know of that can provide eight-fold differences in basal
insulin. I could not achieve any reasonable A1C for my child, even giving
several injections to correct blood sugars during the evening while he was
sleeping. A long-acting insulin adequate for controlling the insulin
resistant period 9 p.m. to 12 p.m. (or from 12 a.m. to 3 a.m. in post-
pubertal children) will more than likely cause hypoglycemia in the early
morning hours of insulin sensitivity, giving rise to the statistic that an
individual with type 1 DM has a low once every four nights. This is
unacceptable given risks of brain damage and coma. Since most
practitioners titrate only based on morning BGL numbers, and make changes
when the BGL is high (Somagi effect) or otherwise out of range, the
glucose excursions are not as evident, but still manifest themselves in
high A1Cs.
Competing interests: No competing interests
"However we consider that in general insulin pump should be reserved
for those with special problems such as unpredictable hypoglycaemia or a
marked increase in blood glucose concentration at dawn, despite best
attempts to improve control with optimised injection regimens."
I find it appalling that 'professionals' would make such a statement.
When I consider my child, any difference in control and thus reduction in
risk is appreciable. Furthermore, this study does not take into account
the emotional effect of diabetes and the flexibility that the pump can
provide. Yes this can be achieved with MDI, but at a price of a lower
quality of life.
To study the pump, establish an increased ability to control glycemic
levels and still reserve it for "special problems" is absurd! It's clear
to me that most likely none of you lives with this disease. In the
unlikely event that one or more of you do, I am certain that you have not
tried CSII.
You cannot study the treatment of disease without taking into account
the patients whom you purport to treat. I implore you to seriously
rethink your conclusion. As members in the medical field, if you are not
looking out for the best interest of patients, you should choose another
field.
With due respect,
Marisa Schmidt
Competing interests: No competing interests
Insulin Pumps: Cost-Effective in the UK
Almost three and a half years have passed since Professor John Pickup
and colleagues reported a thorough meta-analysis of continuous
subcutaneous insulin infusion (CSII) versus multiple daily injections
(MDI) of insulin in patients with type 1 diabetes. During this period, a
number of publications have supported many of the conclusions made by
Pickup et al, including the assertion that CSII is associated with
improvements in glycemic control versus MDI (1;2). Perhaps the most
important of these, at least in terms of influencing decision making in
the UK, was the Health Technology Assessment on the clinical and cost-
effectiveness of CSII for diabetes for the National Institute of Clinical
Excellence (NICE) by Colquitt et al (3). Having read this article with
considerable interest, we felt that there were some important issues to
communicate with the readers. However, on contacting the NHS Health
Technology Assessment Programme, we were informed that their journal does
not offer the opportunity to respond to previous appraisals in forthcoming
publications (although the programme’s website does offer a forum for
discussion: http://www.ncchta.org/correspond/). Given the potential impact
of such publications in the UK and beyond, we found this “closed book”
approach concerning, as we believe the website discussion would only reach
a very limited audience. We have therefore written to the British Medical
Journal to raise this issue and outline what we believe are noteworthy
shortcomings of the Colquitt et al. assessment of the cost-effectiveness
of implementing CSII versus MDI treatment.
The article by Colquitt et al. was based on a literature review made
prior to August 2002. By the time of its publication in October 2004, much
of its content had become obsolete because material over two years old in
this field of diabetes research is of limited value, and probably not
applicable to modern pumps that are notably less expensive to maintain
than earlier versions. At the time of the review, there were no published
cost-effectiveness analyses on CSII versus MDI treatment, and Colquitt et
al. went on to state that they were “unable to identify health outcomes
that can be quantified for the purposes of cost-effectiveness analysis,”
particularly in terms of translating observed glycemic control benefits
into a cost per quality-adjusted life year. However, we have identified at
least four health economic models that had been published at the time of
the review, that were able to project the long-term change in risk of
complications dependent on glycosylated haemoglobin levels, and there
calculate clinical and cost outcomes (4-8). At least one of these models
was capable of assessing the cost-effectiveness of CSII versus MDI in a UK
setting (4).
Colquitt et al. failed to recognize that, whilst very long-term
clinical trial data on the impact of CSII are currently unavailable,
combining cost and clinical data (often from different studies) to project
long-term cost and clinical outcomes is a well-recognized tool for
assisting health policy decision-makers. Computerized mathematical
modelling allows the best empirical data currently available (usually
short- to medium-term) to be projected over considerably longer time
horizons and provide valuable information that would otherwise not be
available. If empirical data is lacking or absent and assumptions need to
be made, sensitivity analyses are used, with model value inputs varied
through plausible ranges and re-run through the model. If the outcomes
remain unchanged or within a certain interval, they are considered to be
robust. Further, statistical analysis allows confidence interval levels to
be calculated.
Since the development of the earlier models (and the completion of
the Colquitt et al. review), a number of other peer-reviewed diabetes
models have been published which can project cost-effectiveness, based on
glycosylated haemoglobin changes (9-12). Two further studies have actually
examined the cost-effectiveness of implementing CSII versus MDI treatment
in the UK setting (13;14).
Scuffham and Carr used a Markov modelling technique and estimated
that CSII compared to MDI treatment in type patients with type 1 diabetes,
over an 8-year period, would lead to an average gain of 0.48 quality-
adjusted life years per patient with an incremental cost of £5,462 (an
incremental cost-effectiveness ratio of £11,461/quality-adjusted life
year). This was a very attractive level of cost-effectiveness. However,
the short-time horizon limited the usefulness of this analysis (15).
Roze et al. used an validated and peer-reviewed diabetes model, the
CORE Diabetes Model, to perform a similar analysis of CSII and MDI
treatment in type 1 diabetes (9;13). The ability of this model to project
robust and realistic clinical and cost outcomes has been thoroughly
examined, with the results of 66 separate analyses with the model closely
matching outcomes from published clinical epidemiological studies. In the
analysis of CSII versus MDI treatment, this model projected an improvement
of 0.76±0.19 quality-adjusted life years with CSII and incremental costs
of £19,407±1,727. This led to an incremental cost-effectiveness ratio of
£25,648 per quality-adjusted life year gained for CSII versus MDI
treatment, over a lifetime horizon, which represents good value for money
by generally accepted UK standards.
Another shortcoming of the Colquitt et al review was a failure to
acknowledge the considerable impact of improved glycemic control on the
development of diabetes-related complications (although severe
hypoglycaemia and diabetic ketoacidosis are discussed at length). It is
well establish, based on data from a number of landmark diabetes trials
(DCCT, UKPDS, WESDR etc.), that improvements in glycemic control leads to
a reduced risk of micro- and macrovascular complications. Failure to take
long-term complications, associated costs and their impact on quality of
life into account would undoubtedly lead to a misrepresentation of the
cost-effectiveness of CSII versus MDI. Indeed, the importance of taking
such long-term complications into account has been outlined in guidelines
for the computer modelling of diabetes (16).
It is disappointing that Colquitt et al. did not seek further
consultation from competent health economic modelling groups with
experience in diabetes before publishing their report. Having done so,
they would have been unlikely to conclude that they “have not found a
satisfactory method, in the time available, of converting observed
benefits into a cost per QALY.” A method has been available for some time
and, in the years between completion of the review and its publication,
several transparent, validated and generally well-accepted computer
simulation models of diabetes have been described in the medical press.
REFERENCES
(1) Weissberg-Benchell J, Antisdel-Lomaglio J, Seshadri R. Insulin pump
therapy: a meta-analysis. Diabetes Care 2003; 26(4):1079-1087.
(2) DeWitt DE, Hirsch IB. Outpatient insulin therapy in type 1 and type 2
diabetes mellitus: scientific review. JAMA 2003; 289(17):2254-2264.
(3) Colquitt JL, Green C, Sidhu MK, Hartwell D, Waugh N. Clinical and
cost-effectiveness of continuous subcutaneous insulin infusion for
diabetes. Health Technol Assess 2004; 8(43):iii, 1-iii171.
(4) Palmer AJ, Weiss C, Sendi PP, Neeser K, Brandt A, Singh G et al. The
cost-effectiveness of different management strategies for type I diabetes:
a Swiss perspective. Diabetologia 2000; 43:13-26.
(5) Gozzoli V, Palmer AJ, Brandt A, Spinas GA. Economic and clinical
impact of alternative disease management strategies for secondary
prevention in type 2 diabetes in the Swiss setting. Swiss Med Wkly 2001;
131(21-22):303-310.
(6) Eastman RC, Javitt JC, Herman WH, Dasbach EJ, Copley-Merriman C,
Maier W et al. Model of complications of NIDDM. II. Analysis of the health
benefits and cost-effectiveness of treating NIDDM with the goal of
normoglycemia [see comments]. Diabetes Care 1997; 20(5):735-744.
(7) Brown JB, Russell A, Chan W, Pedula K, Aickin M. The global diabetes
model: user friendly version 3.0. Diabetes Res Clin Pract 2000; 50 Suppl
3:S15-46.:S15-S46.
(8) Bagust A, Hopkinson PK, Maier W, Currie CJ. An economic model of the
long-term health care burden of Type II diabetes. Diabetologia 2001;
44(12):2140-2155.
(9) Palmer AJ, Roze S, Valentine W, Minshall M, Foos V, Lurati F et al.
Validation of the CORE Diabetes Model against Epidemiological and Clinical
Studies. Current Medical Research and Opinion 2004; 20(Supplement 1):S27-
S40.
(10) Eddy DM, Schlessinger L. Validation of the archimedes diabetes
model. Diabetes Care 2003; 26(11):3102-3110.
(11) Hauner H, Maxion-Bergemann S, Muller E, Schulz M, Huppertz E,
Bergemann R. [Disease management program (DMP) diabetes mellitus:
simulation of therapeutic results of different guidelines. A new diabetes
mellitus model (DMM)]. Dtsch Med Wochenschr 2003; 128(21):1167-1172.
(12) Clarke PM, Gray AM, Briggs A, Farmer AJ, Fenn P, Stevens RJ et al. A
model to estimate the lifetime health outcomes of patients with Type 2
diabetes: the United Kingdom Prospective Diabetes Study (UKPDS) Outcomes
Model (UKPDS no. 68). Diabetologia 2004; 47:1747-1759.
(13) Roze S, Valentine WJ, Zakrzewska KE, Palmer AJ. Health-economic
comparison of continuous subcutaneous insulin infusion with multiple daily
injection for the treatment of Type 1 diabetes in the UK. Diabet Med 2005;
22(9):1239-1245.
(14) Scuffham P, Carr L. The cost-effectiveness of continuous
subcutaneous insulin infusion compared with multiple daily injections for
the management of diabetes. Diabet Med 2003; 20(7):586-593.
(15) Palmer AJ, Roze S, Valentine WJ, Spinas G. The cost-effectiveness of
continuous subcutaneous insulin infusion compared with multiple daily
injections for the management of diabetes: response to Scuffham
[corrected] and Carr. Diabet Med 2004; 21(12):1372-1373.
(16) American Diabetes Association. Guidelines for computer modeling of
diabetes and its complications. Diabetes Care 2004; 27(9):2262-2265.
Competing interests:
Andrew Palmer, Daniel Tucker, Joshua Ray and William Valentine are all employees of CORE. CORE has received an unrestricted grant from Medtronic to perform an analysis evaluating the cost-effectiveness of subcutaneous pumps.
Competing interests: No competing interests