SEARCH

SEARCH BY CITATION

Keywords:

  • age;
  • availability;
  • continuous subcutaneous insulin infusion;
  • sex

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and methods
  5. Results
  6. Discussion
  7. Competing interests
  8. Acknowledgements
  9. References

Diabet. Med. 29, 1055–1059 (2012)

Aim  To examine the availability of insulin pump therapy in patients with Type 1 diabetes.

Methods  Patients using insulin pumps among a cohort of 7224 patients with Type 1 diabetes were studied.

Results  In logistic regression, used to evaluate variables not changing over time among the total cohort, use of insulin pumps varied by outpatient clinic (P < 0.001) and sex (P < 0.001). Cox regression analysis in 5854 patients with detailed patient data prior to use of an insulin pump showed higher HbA1c (P < 0.0001), lower creatinine (P = 0.002), high and low insulin doses (P < 0.0001), younger age (P < 0.0001) and female sex (P < 0.0001) to be associated with use of an insulin pump. Women were 1.5-fold more likely to start using an insulin pump (hazard ratio 1.52, 95% confidence interval 1.29–1.79) and patients in the 20- to 30-years age range were more than twice as likely to begin use of an insulin pump than patients aged 40–50 years (hazard ratio 8.63, 95% confidence interval 5.91–12.59 and hazard ratio 3.98, 95% confidence interval 2.80–5.64, respectively). A 10-μmol/l higher level of creatinine was associated with a hazard ratio of 0.56 (95% confidence interval 0.39–0.81) of starting use of an insulin pump.

Conclusions  At 10 hospital outpatient clinics in Sweden, use of insulin pumps therapy varied by clinic. A higher proportion of women began using insulin pumps. Younger patients and patients with fewer complications were also more likely to start using an insulin pump. Further research is needed to confirm these findings in other geographical regions and to understand whether the availability of insulin pumps today is optimized.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and methods
  5. Results
  6. Discussion
  7. Competing interests
  8. Acknowledgements
  9. References

Good glycaemic control is associated with fewer diabetic complications [1,2]. Patients with Type 1 diabetes are currently treated with either multiple daily insulin injections or insulin pumps, also referred to as continuous subcutaneous insulin infusion [3,4]. Meta-analyses have shown insulin pump use to be associated with HbA1c approximately 3 mmol/mol (0.3%) lower than multiple daily insulin injections, and data in clinical practice have shown an improvement of 4 mmol/mol (0.4%) after switching from multiple daily insulin injections to an insulin pump [5–7].

In light of these data showing reductions in HbA1c, there may be incentives for increasing the use of insulin pumps. In Sweden, insulin pump therapy is a high priority in recently presented national guidelines, and one goal is to increase insulin pump use in patients who have not achieved target HbA1c with multiple daily insulin injections [8]. Although extensive research has been performed on the effects of insulin pumps on HbA1c, there is little information on the availability of the therapy in clinical practice. Previous work has shown that patients with higher HbA1c experience greater benefits from insulin pump therapy, while, to our knowledge, superior effects have not been shown for other patient groups [7]. While the effects of insulin pumps on HbA1c are favourable, markedly lower use in certain patient groups may indicate that clinicians are not optimizing the availability of insulin pumps. The aim of this study was to examine the probability of treatment with an insulin pump by age, sex, diabetes duration, grade of diabetic complications, hospital outpatient clinic, insulin dose and level of glycaemic control. This information may lead to better understanding of the accessibility of insulin pumps for different patient groups and, by extension, possibly to increased use for patients, with favourable effects on HbA1c and consequently on risks of diabetic complications.

Subjects and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and methods
  5. Results
  6. Discussion
  7. Competing interests
  8. Acknowledgements
  9. References

Data were obtained from a medical record system (Diab-Base; Journalia AB, Kungälv, Sweden) with the possibility of electronic tracking of prescriptions of medications, diabetic therapies, risk factors and diabetic complications. The system has previously been described in detail [9–11]. Currently, a total of 10 specialized outpatient clinics in Sweden use Diab-Base.

The probability of using an insulin pump was compared in terms of age, sex, diabetes duration, grade of complications, hospital outpatient clinic, insulin dose and HbA1c. Creatinine level, examined as a continuous variable, was used to indicate whether the use of insulin pumps differs for patients with severe complications and those with moderate, if any, complications.

Differences in patient characteristics (age, sex, diabetes duration and hospital outpatient clinic) at latest appointment were compared cross-sectionally in terms of patients who had used vs. never used an insulin pump. Patients who began treatment with an insulin pump prior to the start of their registration in the system were also included in this analysis.

To understand potential characteristics associated with use of an insulin pump, a survival analysis was performed to account for changes in HbA1c, creatinine, age, diabetes duration and insulin dose over time. Patients were only included if they began using an insulin pump after registration in the database and had at least one measure of HbA1c, creatinine level and insulin dose before starting use of an insulin pump.

As diabetes clinics in Sweden used HbA1c methods calibrated to the HPLC Mono-S method until September 2010, all HbA1c values have been converted to International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) and National Glycohemoglobin Standardization Program (NGSP) levels [12].

Statistical analysis

Multiple logistic regression was performed to compare differences in patients with or without an insulin pump at any time by the last appointment. Age, sex, diabetes duration and hospital outpatient clinic were used as independent variables, while use of an insulin pump at any time was the dependent variable. Cox regression was performed to compare differences in HbA1c, age, sex, diabetes duration, insulin dose and creatinine level at the time for starting use of an insulin pump. All variables were updated monthly until the patient began using an insulin pump, was lost to follow up or until the end of the study period. The proportional hazards assumption was not fulfilled for the outpatient clinic covariate and was thus used to stratify the data. All analyses were performed with SAS version 9.1.3 (SAS Institute, Cary, NC, USA).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and methods
  5. Results
  6. Discussion
  7. Competing interests
  8. Acknowledgements
  9. References

A total of 7224 patients with Type 1 diabetes were studied. Of these, 1348 used an insulin pump. The number of patients at each clinic ranged from 222 to 1693. The overall proportion of patients using an insulin pump increased over time, with a decline in 2003–2005 (Fig. 1). Patient characteristics are described in Table 1. There were 177 patients who stopped using insulin pumps. Patient characteristics of those discontinuing insulin pump use were similar to the entire group.

image

Figure 1.  Proportion of patients with multiple daily insulin injections resulting in starting use of an insulin pump from 1999 to 2010.

Download figure to PowerPoint

Table 1.   Patient demographics at baseline for the total cohort of 7224 patients with Type 1 diabetes analysed with logistic regression
Total cohort
 Multiple daily insulin injectionsContinuous subcutaneous insulin infusion Starting
n = 5876n = 1348
Sex—female, n (%)2432 (41%)762 (57%)
Age (years)40 (16.1)32 (12.0)
Duration of diabetes (years)
 Median (interquartile range)12 (20)13 (15)
Part of total cohort used in time-dependent analysis
 Multiple daily insulin injectionsContinuous subcutaneous insulin infusion Starting
n = 5156n = 698
  1. Cox regression was used for subgroup analysis in patients who began to use an insulin pump after the establishment of the database and in patients with multiple daily insulin injections who had valid data for all patient characteristics used in the Cox regression.

  2. Data are mean (sd) or number (%) if not otherwise specified.

Sex—female, n (%)2122 (41%)397 (57%)
Age (years)42 (15.9)35 (11.5)
Duration of diabetes (years)
 Median (interquartile range)15 (22)15 (16)
Creatinine (μmol/l)
 Median (interquartile range)82 (24)78 (21)
HbA1c (mmol/mol)66.1(19.9)68.2 (16.1)
HbA1c (%)8,2 (1.8)8,4 (1.5)
Insulin dose (U/kg) 0.7 (0.6)0.7 (0.5)

In logistic regression analysis, the probability of using an insulin pump at any time differed substantially by outpatient clinic (P <  0.0001), age (P <  0.0001), diabetes duration (P <  0.0001) and sex (P <  0.0001) (Table 2). There was a statistically significantly difference by sex, with more women using an insulin pump at any time. Younger age and longer diabetes duration also showed statistically significant associations with use of an insulin pump, as compared with middle-aged patients and patients with shorter diabetes duration (Table 2).

Table 2.   Odds ratios for insulin pump use at any time determined by multiple logistic regression
 Odds ratio95% confidence intervalContinuous subcutaneous insulin infusion, nMultiple daily insulin injections, n
Women vs. men
 Women1.831.62–2.087392247
 Men (reference)1 5703178
Outpatient clinic
 10.950.71–1.2775343
 20.580.47–0.731371049
 31.551.22–1.95152419
 40.810.66–1.00173830
 51.020.80–1.30118482
 61.941.51–2.50126298
 70.980.71–1.3755232
 81.360.94–1.9744169
 91.230.95–1.60107364
10 (reference)1 3221239
Age, years
 < 204.682.54–8.601670
 20 to < 305.224.06–6.71283861
 30 to < 403.963.13–5.01297904
 40 to < 503.422.75–4.26323963
 50 to < 602.201.76–2.752251014
 ≥ 60 (reference)1 1651613
Diabetes duration, years
 < 100.260.20–0.41241166
 10 to < 200.760.60–1.003791307
 20 to < 300.940.74–1.203551153
 30 to < 401.271.01–1.60286853
 ≥ 40 (reference)1 165946
Total number of patients  13095425

There were 5854 patients who did not have an insulin pump when registered in the database and who had valid information on HbA1c, insulin dose and creatinine before using an insulin pump. In this group, 698 patients started to use a pump during the study period. According to Cox regression analysis, patient characteristics associated with starting use of an insulin pump were HbA1c (P < 0.0001), creatinine (P = 0.002), insulin dose (P < 0.0001), age (P < 0.0001) and sex (P < 0.0001) (Table 3). Diabetes duration was of borderline significance (P = 0.051). Women were 1.5-fold more likely to begin using an insulin pump (hazard ratio 1.52, 95% CI 1.29–1.79) and patients in the 20- to 30-years age group were twice as likely to start use of an insulin pump than patients aged 40–50 years (hazard ratio 8.63, 95% CI 5.91–12.59 and hazard ratio 3.98, 95% CI 2.80–5.64, respectively). By insulin dose, the hazard ratios were highest for patients with very low and very high insulin doses (Table 3). A higher HbA1c level was also associated with a greater hazard ratio for starting use of an insulin pump (Table 3). Among the 698 patients who began using an insulin pump, 86 discontinued use. Characteristics of patients discontinuing insulin pump use were similar to those of the entire group.

Table 3.   Cox regression showing hazard ratios and confidence intervals for changing therapy to insulin pump for different patient characteristics
 Hazard ratio95% confidence interval
  1. Model is stratified by outpatient clinic.

Creatinine, 10 μmol/l higher0.560.39–0.81
Female1.521.29–1.79
HbA1c, mmol/mol (%)
 < 42 (< 6.0%) (reference)1 
 42 to < 52 (6.0 to < 6.9)1.981.05–3.75
 52 to < 63 (6.9 to < 7.9)2.701.45–5.02
 63 to < 73 (7.9 to < 8.8)3.752.02–6.96
 73 to < 83 (8.8 to < 9.8)4.822.58–9.03
 ≥ 83 (≥ 9.8)5.652.98–10.72
Insulin dose, U/kg
 < 0.5 U/kg (reference)1 
 0.5 to < 1.00.290.25–0.35
 1.0 to < 1.50.170.12–0.25
 ≥ 1.50.400.18–0.86
Diabetes duration, years
 < 100.820.56–1.20
 10 to < 201.050.74–1.50
 20 to < 301.120.79–1.59
 30 to < 401.240.88–1.77
 ≥ 40 (reference)1 
Age, years
 < 209.373.58–24.50
 20 to < 308.635.91–12.59
 30 to < 406.184.35–8.77
 40 to < 503.982.80–5.64
 50 to < 602.631.86–3.73
 ≥ 60 (reference)1 

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and methods
  5. Results
  6. Discussion
  7. Competing interests
  8. Acknowledgements
  9. References

In this study cohort of 7224 patients with Type 1 diabetes, the use of insulin pump treatment differed by age, sex, glycaemic control, grade of diabetic complications and outpatient clinic. Younger patients aged 20–30 years were twice as likely to begin using an insulin pump as patients aged 40–50 years, and approximately 50% more women than men started to use a pump. High HbA1c levels were associated with a greater chance of starting use of an insulin pump, possibly indicating rational use, as better effects in lowering HbA1c have previously been reported for these patients [7]. The absolute risk reduction for microvascular complications is also higher in this patient group [13]. Outpatient clinics also showed a strong association with use of insulin pumps, as was the case for extent of complications described in terms of creatinine levels, with less use in patients with higher creatinine levels.

To our knowledge, the use of insulin pumps among patients with Type 1 diabetes has not previously been studied. At this stage, our findings are only speculative. It can be noted that health-related behaviours have been proposed to differ by sex [14], and men and women may, in some instances, have different attitudes about using a pump. Differences in the use of insulin pumps by sex are probably not influenced by pregnancy as insulin pumps in Sweden are generally not prescribed to women with current or planned pregnancy. Regarding age, caregivers may view younger patients as more technologically knowledgeable or ‘savvy’, and hence more appropriate users of pumps. However, as far as we know, different effects of insulin pumps by age have not been reported. One could speculate that differences by outpatient clinic could be attributable to variations in level of staff training and preferences, experience in recommended treatments and differences in local incentives. It is possible that local finance policies were previously a factor at some clinics, but today national guidelines generally encourage caregivers to increase prescriptions for insulin pumps. The finding of fewer insulin pumps among patients with high creatinine levels is difficult to explain.

Limitations of the present study include insufficient data on the frequency of hypoglycaemia, diabetic ketoacidosis and number of blood glucose measurements, attributable to non-registration at appointments. In Sweden, both the physician and the diabetic nurse can propose insulin pump therapy, although the prescription is issued by the physician. At times, patients also propose the therapy themselves. No information was available on the characteristics of the caregiver prescribing the pump, the individual who proposed the therapy or whether patients may have been offered but rejected using a pump. Further research is needed to understand why use differs among patient groups and outpatient clinics, and this could be addressed in qualitative studies consisting of patient and caregiver interviews. It would be of interest, for example, to determine whether male and female patients, younger and older, with different degrees of complications and patients at different clinics have been recommended the use of an insulin pump to varying extents. Furthermore, it would be of interest to ask whether more patients in these groups have refused the recommended insulin pump and, if so, why.

In conclusion, at 10 hospital outpatient clinics in Sweden, use of insulin pump therapy varied by clinic. Fewer men than women began using insulin pumps, fewer older than younger patients and fewer patients with a high grade of complications. This implies that if we understand how availability could be increased in these patient groups, especially among those patients with high HbA1c, diabetic complications could probably be reduced.

Further research is needed to confirm these findings in other geographical regions and to understand factors associated with low use and how availability can be optimized.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and methods
  5. Results
  6. Discussion
  7. Competing interests
  8. Acknowledgements
  9. References

We would like to thank all of the outpatient clinics working with Diab-Base that contributed to the data used in this study. We would also like to thank Joseph Murphy and Linda Schenk for editorial assistance in the preparation of this manuscript.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and methods
  5. Results
  6. Discussion
  7. Competing interests
  8. Acknowledgements
  9. References
  • 1
    DCCT Study 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: 977986.
  • 2
    Nathan DM, Cleary PA, Backlund JY, Genuth SM, Lachin JM, Orchard TJ et al. Intensive diabetes treatment and cardiovascular disease in patients with type 1 diabetes. N Engl J Med 2005; 353: 26432653.
  • 3
    Pickup JC, Renard E. Long-acting insulin analogs versus insulin pump therapy for the treatment of type 1 and type 2 diabetes. Diabetes Care 2008; 31: S140S145.
  • 4
    Hanaire H, Lassmann-Vague V, Jeandidier N, Renard E, Tubiana-Rufi N, Vambergue A et al. Treatment of diabetes mellitus using an external insulin pump: the state of the art. Diabetes Metab 2008; 34: 401423.
  • 5
    Misso ML, Egberts KJ, Page M, O’Connor D, Shaw J. Continuous subcutaneous insulin infusion (CSII) versus multiple insulin injections for type 1 diabetes mellitus. Cochrane Database Syst Rev 2010; 1: CD005103.
  • 6
    Monami M, Lamanna C, Marchionni N, Mannucci E. Continuous subcutaneous insulin infusion versus multiple daily insulin injections in type 1 diabetes: a meta-analysis. Acta Diabetol 2010; 47: 7781.
  • 7
    Fahlén M, Eliasson B, Odén A. Optimization of basal insulin delivery in Type 1 diabetes: a retrospective study on the use of continuous subcutaneous insulin infusion and insulin glargine. Diabet Med 2005; 22: 382386.
  • 8
    National Board of Health and Welfare. Nationella Riktlinjer för diabetesvården (National Guidelines for Diabetes Care). 2010. Available at http://www.socialstyrelsen.se/nationellariktlinjerfordiabetesvarden Last accessed 1 July 2010.
  • 9
    Journalia. Available at http://www.journalia.se Last accessed 1 July 2010.
  • 10
    Lind M, Fahlén M, Happich M, Odén A, Eliasson B. The effect of insulin lispro on glycemic control in a large patient cohort. Diabetes Technol Ther 2009; 11: 5156.
  • 11
    Lind M, Odén A, Fahlén M, Eliasson B. The true value of HbA1c as a predictor of diabetic complications: simulations of HbA1c variables. PLoS One 2009; 4: e4412.
  • 12
    Hoelzel W, Weykamp C, Jeppsson JO, Miedema K, Barr JR, Goodall I et al.; IFCC Working Group on HbA1c Standardization IFCC reference system for measurement of hemoglobin A1c in human blood and the national standardization schemes in the United States, Japan, and Sweden: a method-comparison study. Clin Chem 2004; 50: 166174.
  • 13
    DCCT Study Group. The relationship of glycaemic exposure (HbA1c) to the risk of development and progression of retinopathy in the Diabetes Control and Complications Trial. Diabetes 1995; 44: 968983.
  • 14
    Courtenay WH. Constructions of masculinity and their influence on men’s well-being: a theory of gender and health. Soc Sci Med 2000; 50: 13851401.