- Email: [email protected]
Predictors of glycemic control on insulin pump therapy in children and adolescents with type I diabetes
Diabetes Research and Clinical Practice 74 (2006) 217–221 www.elsevier.com/locate/diabres
Predictors of glycemic control on insulin pump therapy in children and adolescents with type I diabetes Zeina M. Nabhan *, Laura Rardin, Jill Meier, Erica A. Eugster, Linda A. DiMeglio Section of Pediatric Endocrinology and Diabetology, Department of Pediatrics, James Whitcomb Riley Hospital for Children, Indiana University School of Medicine, Indianapolis, IN, United States Received 11 January 2006; accepted 20 March 2006 Available online 15 May 2006
Abstract Objective: To determine variables predictive of glycemic control in a large population of pediatric patients with type 1 diabetes on continuous subcutaneous insulin infusion (CSII). Methods: Charts of patients on CSII for 1 year were reviewed. ‘‘Good’’ control was a priori defined as HbA1c 9% in patients under 12 years of age, and 8% in patients over 12 years. Results: Ninety-three patients were identified (57 girls and 36 boys). Their mean age at pump start was 11.6 3.1 years with duration of diabetes of 4.7 3.1 years. Average time on pump therapy was 2.4 0.8 years. HbA1C decreased from 8.7 0.9% prior to pump therapy to 8.3 0.6% while on CSII ( p < 0.01). Despite analysis of a large number of possible predictors, only number of basal rates (4.4 versus 3.4) and younger age (10.0 years versus 13.1 years) correlated with good control. Conclusion: Only younger age and use of more basal rates were predictive of good diabetes control in children using CSII. Decisions regarding which pediatric patients are most appropriate for CSII must continue to be individualized. # 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Type 1 diabetes; Insulin pump therapy; Glycemic control; Children and adolescents
1. Introduction Insulin pump therapy for the treatment of diabetes began to be utilized in the late 1970s [1] and reemerged in the 1990s due to technological advances in insulin delivery systems. Interest in continuous subcutaneous insulin infusion (CSII) via pumps for children has also
Abbreviations: CSII, continuous subcutaneous insulin infusion; HbA1c, hemoglobin A1c * Corresponding author at: Pediatric Endocrinology/Diabetology, Riley Hospital for Children, Room 5960, 702 Barnhill Drive, Indianapolis, IN 46202-5225, United States. Tel.: +1 317 274 3889; fax: +1 317 274 3882. E-mail address: [email protected] (Z.M. Nabhan).
been fueled by an increased emphasis on achieving strict glycemic control to minimize the risk of complications [2]. CSII therapy has multiple potential advantages over injection therapy, including an enhanced ability to mimic physiologic insulin release as well as permitting greater flexibility in food intake and physical activity. Several studies in adult populations have demonstrated a high degree of patient satisfaction on CSII therapy as well as improvements in quality of life [3,4]. Studies regarding safety and efficacy of CSII therapy in children and adolescents have provided mixed results. Some studies suggest an increased risk of diabetic ketoacidosis (DKA) on CSII compared to injections [5] whereas others show a reduction in DKA rates [6,7].
0168-8227/$ – see front matter # 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.diabres.2006.03.020
218
Z.M. Nabhan et al. / Diabetes Research and Clinical Practice 74 (2006) 217–221
Similarly, glycemic control has been reported to improve [8–10] or remain unchanged [11–13]. Most studies, however, concur that episodes of severe hypoglycemia appear to decrease while on CSII [7–9]. Few studies have tried to elucidate individual factors that predict or affect glycemic control of patients on CSII [14,15]. In adults, the frequency of self-monitoring of blood glucose has been found to be the most influential factor. Given the inconsistencies in the literature and the cost and complexity of CSII, identification of the best candidates for this therapy is inherently challenging. No analysis of multiple factors predictive of control in a large cohort of diabetic children and adolescents on insulin pumps has been previously performed. In this study, we investigated whether variables could be identified that correlated with good or poor control in a large cohort of pediatric patients on insulin pump therapy. We hypothesized that chart reviews would reveal clinically relevant factors that would enable better prospective selection of patients in whom pump therapy would be beneficial and well tolerated. 2. Methods Following institutional review board approval, medical records of patients with type 1 diabetes on insulin pump therapy followed at James Whitcomb Riley Hospital for Children were identified. All patients were started on insulin pump therapy by the clinical pediatric diabetology team at Riley Hospital and had been utilizing CSII for at least 1 year. Patient records were excluded if children had begun on pumps as part of research protocols. Charts for eligible patients were all reviewed by one author (L.R.). The following variables were extracted from the medical records: sex, race, age at diabetes onset, diabetes duration, age at CSII initiation, duration of pump therapy, hemoglobin A1c (HbA1c) values for 1 year prior to pump therapy through 1 year post-pump initiation, HbA1c at the last available follow-up, fast-acting insulin regimen prior to CSII (sliding scale or carbohydrate to insulin ratios), average number of injections per day prior to pump start, number of basal rates on pumps, pump brand, diabetes camp attendance, participation in a series of advanced diabetes education classes (ADAPT), family structure, and clinic visit frequency. ‘‘Good’’ control was defined by HbA1c stratified by age, with a target HbA1c (DCA 2000, Roche Diagnostics) of 9% for patients under 12 years of age, and 8% in patients over 12. The cut-off point for the HbA1c was made a priori based upon our diabetes program’s target blood sugar ranges for age. HbA1c at last available follow-up after pump start was used for analysis of predictors of glycemic control and ‘‘poor’’ and ‘‘good’’ control for age.
3. Pump program at Riley Hospital Patients were recommended for insulin pump therapy by their primary diabetologist if they had been diagnosed with type 1 diabetes for at least 12 months and had a history of compliance with physicians’ visits and home blood glucose monitoring. Before initiation of pump therapy, all families and patients attended a half-day education session, during which insulin types, insulin adjustment, and carbohydrate counting were reviewed. Families that had not previously used carbohydrate/insulin ratios or corrective doses began using these techniques after the education sessions. At the time of pump therapy initiation, all patients held their morning dose of long acting insulin, began using their pump by late morning, and remained through the afternoon for a second education session with a nurse educator during which patients and families were taught the mechanics of pump use. For the first 10 days, families were asked to check their child’s blood sugar at least eight times per day, with tests before meals, before snacks, at midnight, and at 3 a.m. Families were also asked to change infusion sets every 3 days. For the first 10 days, families were contacted daily by a nurse educator to make insulin adjustments for basal rates, carbohydrate/ insulin ratio, and correction boluses. Subsequently, phone follow-up occurred twice weekly until the insulin dose stabilized. The first follow-up appointment was within 2–3 months of the pump start. 3.1. Statistical analysis Statistics were done using Microsoft Excel 2000 (Microsoft Corp., Redmond, Wash) for descriptive statistics and SPSS Version 11.5 (SPSS Inc., Chicago Ill) for other analyses. Data are expressed as means standard deviations. T-tests were used for comparisons between groups when the data were normally distributed. Bonferroni correction was done for multiple comparisons. Correlations were done using Pearson’s test. 4. Results Ninety-three patients (57 girls and 36 boys) fit inclusion criteria and had charts available for review. Patients started on pumps between November 1998 and June 2002. Subject characteristics at the time of pump start are summarized in Table 1. Fifty-three children were over the age of 12 years. Forty-four (16 boys and 28 girls) of the 93 children were in ‘‘good’’ control for age at an average age of 10.4 3.2 years (range, 3.4–18.2 years)
Z.M. Nabhan et al. / Diabetes Research and Clinical Practice 74 (2006) 217–221
219
Table 1 Baseline patient characteristics Age at pump start (y)
11.6 3.1
Gender
F: 57 (60%) M: 36 (40%)
Diabetes duration at pump start (y) Baseline HbA1c
4.7 3.1 8.7 0.9%
Parental marital status
Married: 68 (73%) Divorced: 26 (27%)
Insulin regimen prior to CSII
Sliding scale: 30 (32%) Carb/Insulin ratio: 63 (68 %)
Number of injections/day Diabetes camp attendance Advanced diabetes educational session attendance (ADAPT) a
2.7 0.9 37 (40%) 37 (40%)
Data expressed as means standard deviations (S.D.); y: years; mo: months; F: female; M: male. a ADAPT is an outpatient education program for children and young adults with diabetes and their families. Through ADAPT patients learn how to develop good habits in managing diabetes.
and mean HbA1c of 8.0 0.6% (range, 6.6–9.0%). Those in poor control for age had an average age of 12.6 2.5 years (range, 5.7–17.9 years) and HbA1c of 9.3 0.8% (range, 8.1–11.9%). Most (77%) of the patients in poor control prior to pump start were over the age of 12 years. Following pump start, patients were seen in diabetes clinic every 3.3 0.4 months. At the time of last available follow-up, mean age of the patients was 13.9 3.1 years (range, 6.5–20.2 years). Average HbA1c prior to pump start was 8.7 0.9%. By 6 months post-CSII start, average HbA1c dropped to 8.1 0.9% ( p < 0.001) and then rose to 8.2 0.8% at 12 months ( p < 0.001 versus baseline, p = 0.34 versus 6 months). At the time of the last available follow-up after a mean duration of pump therapy of 2.4 0.8 years (range, 1.1–4.6 years), HbA1c was 8.3 0.6% (range, 6.9–9.7; p < 0.01 versus baseline) as shown in Fig. 1. At this time, 46 patients (18 boys and 28 girls) were in good control with an average HbA1c of 7.8 0.4% (range, 7.0–9.0%) and 47 (17 boys and 30 girls) were in poor control with an average HbA1c of 8.7 0.6% (range, 8.0–11.0%). 4.1. Predictors of glycemic control There was no correlation between the children’s age and HbA1c prior to pump start (r = 0.049; p = 0.64). However, there were positive correlations of both age at pump start and age at last visit with HbA1c at the time
Fig. 1. Mean HbA1c in 93 children with type 1 diabetes at pump start, 6 months post-CSII, 12 months post-CSII, and 28 months (mean time at the last available follow-up). p < 0.001 at 6 and 12 months vs. baseline and p < 0.01 at 28 months vs. baseline. Error bars show mean 1.0S.D.
of the last available follow-up visit (r = 0.227 for age at pump start, r = 0.216 for age at last visit, both p < 0.05). The mean number of basal rates was negatively correlated with mean HbA1c after pump start (r = 0.269; p < 0.01). Mean HbA1c after pump start did not correlate with diabetes duration, duration of CSII therapy, pre-pump insulin regimen, frequency of clinic visits, family structure, gender, diabetes camp attendance, or participation in advanced diabetes education sessions. The same variables were analyzed to identify predictors of change in HbA1c (difference between average pre-pump HbA1c and HbA1c at last available follow-up). Only average pre-pump HbA1c positively correlated with change in HbA1c (r = 0.751; p < 0.001) as shown in Fig. 2. 4.2. Predictors of ‘‘good’’ and ‘‘poor’’ control for age The data were also analyzed to examine determinants of good and poor control of diabetes (stratified by HgbA1c for age). Comparisons between the two groups are summarized in Table 2. Younger age at pump initiation and more basal rates predicted better control. Diabetes duration at pump start, duration of CSII therapy, pre-pump insulin regimen, pump brand,
220
Z.M. Nabhan et al. / Diabetes Research and Clinical Practice 74 (2006) 217–221
Table 2 Comparison between good control group and poor control group Comparison
Good control
Poor control
T-test
On pump A1C Pre-pump A1C Patient age at pump start a No of basal ratesa Diabetes duration at pump start Duration of CSII Gender (% female) Frequency of clinic visits (months) Pre-pump insulin regimen (number of shots/day) Family structure (% married) Pump brand (% medtronic)
7.8 0.4 8.6 0.9 10.0 3.1 4.4 1.3 4.3 2.6 2.3 0.7 58.7 3.3 0.4 2.7 0.9 76.1 66
8.7 0.6 8.8 1.0 13.1 2.2 3.4 1.1 5.24 3.3 2.4 0.8 63.8 3.3 0.4 2.7 0.1 68.1 67
p < 0.01 p = 0.26 p < 0.01 p < 0.01 p = 0.22 p = 0.26 p = 0.61 p = 0.69 p = 0.96 p = 0.39 p = 0.88
a
Good control correlated with age at pump start (r = 0.495; p < 0.01) and number of basal rates (r =
Fig. 2. Average pre-pump HbA1c positively correlated with change in HbA1c on CSII (r = 0.751; p < 0.001).
frequency of clinic visits, family structure, gender, diabetes camp attendance, and attendance of advanced diabetes education sessions did not differ between the groups. 5. Discussion Identifying variables that predict glycemic control in children and adolescents on insulin pump therapy is paramount in optimizing the use of this expensive, intensive therapy. However, little information is available to inform clinical diabetes management decisions. In our study, younger age at pump initiation
0.381; p < 0.01).
and number of basal rates on pump therapy correlated with better glycemic control. Although we did not have a direct measure of parental involvement, the younger age of the patients in good control likely reflects increased parental involvement with CSII therapy. Similarly, the number of basal rates may be a surrogate for the intensity and frequency of insulin adjustment. Our results concur with those of other investigators [14– 16] who have reported that greater parental involvement and younger age are associated with improved control on CSII. Consistent with prior studies [8–10], we also found a significant decrease in HbA1c following the initiation of CSII, suggesting that CSII improves metabolic control in children and adolescents with type 1 diabetes. Unlike what has been reported by other studies [5,17], this decrease was sustained for an average of 2.4 years in our patients on pumps. Although HbA1c increases slightly after the initial 6 months post-CSII level, it plateaus at 1 year without reverting to pre-pump values. The decrease at 6 months likely represents intensive involvement in diabetes management provided by the pump nurses that follows transition to pump therapy or it may also be due to more patient or parent attention with use of new technology. Another interesting finding in our study was the greater improvement in HbA1c in patients with worse control at baseline prior to initiation of pump therapy indicating that CSII does have the capacity to greatly improve outcomes in some previously poorly controlled patients. This finding has been reported by other investigators [18]. However, neither our study nor the other studies was controlled. It is possible that the greater decrease in HbA1c in patients with worse glycemic control prior to pump therapy represents simply improvement from increased contact for insulin
Z.M. Nabhan et al. / Diabetes Research and Clinical Practice 74 (2006) 217–221
adjustments with members of the diabetes team and increased frequency of blood glucose monitoring, rather than due to factors exclusive to the use of CSII. The strengths of this study are the large number of patients with a broad age range (5–20 years) and followup data for as long as 4.5 years after pump start. Also, our data collection incorporated multiple HbA1C assessments after pump start, as patients were followed every 3 months on average. Our study, however, has several limitations including its retrospective design. Data such as frequency of blood glucose monitoring, total daily dose of insulin, timing of boluses, bolus omission, family involvement, and rate of emergency room and hospital utilization were not available in the patient records. In addition, HbA1c was used to identify predictors of good control on CSII; however, good HbA1c levels are not the sole basis for choosing who would benefit from pump therapy. Pump therapy has been demonstrated to have several other advantages such as improving lifestyle flexibility and decreasing hypoglycemia and DKA episodes. These factors, however, were not evaluated in this study because of its retrospective design. Finally, the patients who received pump therapy generally came from families with a history of motivation and compliance with therapy and might not represent results that are applicable to a general population of youth with diabetes. In summary, our study corroborates the mounting evidence that CSII is effective in children and adolescents with type 1 diabetes. However, only a few specific factors help predict good glycemic control on insulin pump therapy. Further research into this area is needed. At present, decisions regarding which pediatric patients are most appropriate for CSII must continue to be highly individualized. References [1] M.J. Leonard, G.D. Reeves, Continuous subcutaneous insulin infusion: a comprehensive review of insulin pump therapy, Arch. Intern. Med. 19 (2001) 2293–2300. [2] The Diabetes Control, Complications Trial, Research Group, The effect of intensive diabetes treatment on the development and progression of long-term complications in insulin-dependant diabetes mellitus, N. Engl. J. Med. 329 (1993) 977–986. [3] R. Linkeschova, M. Raoul, U. Bott, M. Berger, M. Spraul, Less severe hypoglycemia, better metabolic control, and improved quality of life in type 1 diabetes mellitus with continuous subcutaneous insulin infusion (CSII) therapy; an observational
[4]
[5]
[6]
[7] [8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
221
study of 100 consecutive patients followed for a mean of 2 years, Diabet. Med. 19 (2002) 746–751. M. Graff, R.R. Rubin, E.A. walker, How diabetes specialists treat their own diabetes: findings from a study of the AADE and ADA membership, Diabetes Educ. 26 (2000) 460–467. N.H. White, Hollander As, M. Sadler, L. Daniels, Risks and benefits of continuous subcutaneous insulin infusion (CSII) in children (Abstract), Diabetes 50 (Suppl. 2) (2001) A66. B. Steindel, T.R. Roe, G. Costin, M. Carlson, F. Kaufman, Continuous subcutaneous insulin infusion (CSII) in children and adolescents with chronic poorly controlled type 1 diabetes mellitus, Diabetes Res. Clin. Pract. 27 (1995) 199–204. P. Blackett, Insulin pump treatment for recurrent ketoacidosis in adolescence, Diabetes Care 18 (1995) 881–882. E.A. Boland, M. Grey, A. Oesterle, L. Fredrickson, W.V. Tamborlane, Continuous subcutaneous insulin infusion: a new way to lower risk of severe hypoglycemia, improve metabolic control, and enhance coping in adolescents with type 1 diabetes, Diabetes Care 22 (1999) 1779–1784. F.R. Kaufman, M. Halvorson, D. Miller, M. Mackenzie, L.K. Fisher, P. Pitukcheewanont, Insulin pump therapy in type 1 pediatric patients: now and into the year, Diabetes Metab. Res. Rev. 15 (1999) 338–352. A.K. Maniatis, G.J. Klingensmith, R.H. Slover, C.J. Mowry, H.P. Chase, Continuous subcutaneous insulin infusion therapy for children and adolescents; an option for routine diabetes care, Pediatrics 107 (2001) 351–356. A. Schiffrin, M.M. Belmonte, Comparison between continuous subcutaneous insulin infusion and multiple injections of insulin—a one-year prospective-study, Diabetes 31 (1982) 255–264. N. Weintrob, H. Benzaquen, A. Galatzer, S. Shalitian, L. Lazar, G. Fayman, et al., Comparison of continuous subcutaneous insulin infusion and multiple daily injection regimens in children with type 1 diabetes: a randomized open crossover trial, Pediatrics 112 (2003) 559–564. L.A. DiMeglio, T.M. Pottorff, S.R. Boyd, L. France, N. Fineberg, E.A. Eugster, A randomized, controlled study of insulin pump therapy in diabetic preschoolers, J. Pediatr. 145 (2004) 380–384. B. Bode, D. Steed, P. Davidson, Determinants of glycemic control in insulin pump therapy, Diabetes 46 (Suppl. 1) (1997) 143. L.P. Plotnick, L.M. Clark, F.L. Brancati, T. Erlinger, Safety and effectiveness of insulin pump therapy in children and adolescents with type 1 diabetes, Diabetes Care 26 (4) (2003) 1142–1146. B. Anderson, J. Ho, J. Brackett, D. Finkelstein, L. Laffel, Parental involvement in diabetes management tasks: relationships to blood glucose monitoring adherence and metabolic control in young adolescents with insulin dependant diabetes mellitus, J. Pediatr. 130 (1997) 257–265. L. Laffel, C. Loughlin, N. Ramchan-Dani, D. Butler, N. Laffel, B. Levine, et al., Glycemic challenges of pump therapy in youth with type-diabetes (Abstract), Diabetes 50 (Suppl. 2) (2001) A66–A67. D. Iafusco, L. Fortunato, F. De Luca, F. Prisco, The egg or the chicken? Should good compliance to multiinjection insulin therapy criterion for insulin pump therapy, or does insulin pump therapy improve compliance?, J Pediatr. 148 (2006) 421.
Predictors of glycemic control on insulin pump therapy in children and adolescents with type I diabetes Zeina M. Nabhan *, Laura Rardin, Jill Meier, Erica A. Eugster, Linda A. DiMeglio Section of Pediatric Endocrinology and Diabetology, Department of Pediatrics, James Whitcomb Riley Hospital for Children, Indiana University School of Medicine, Indianapolis, IN, United States Received 11 January 2006; accepted 20 March 2006 Available online 15 May 2006
Abstract Objective: To determine variables predictive of glycemic control in a large population of pediatric patients with type 1 diabetes on continuous subcutaneous insulin infusion (CSII). Methods: Charts of patients on CSII for 1 year were reviewed. ‘‘Good’’ control was a priori defined as HbA1c 9% in patients under 12 years of age, and 8% in patients over 12 years. Results: Ninety-three patients were identified (57 girls and 36 boys). Their mean age at pump start was 11.6 3.1 years with duration of diabetes of 4.7 3.1 years. Average time on pump therapy was 2.4 0.8 years. HbA1C decreased from 8.7 0.9% prior to pump therapy to 8.3 0.6% while on CSII ( p < 0.01). Despite analysis of a large number of possible predictors, only number of basal rates (4.4 versus 3.4) and younger age (10.0 years versus 13.1 years) correlated with good control. Conclusion: Only younger age and use of more basal rates were predictive of good diabetes control in children using CSII. Decisions regarding which pediatric patients are most appropriate for CSII must continue to be individualized. # 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Type 1 diabetes; Insulin pump therapy; Glycemic control; Children and adolescents
1. Introduction Insulin pump therapy for the treatment of diabetes began to be utilized in the late 1970s [1] and reemerged in the 1990s due to technological advances in insulin delivery systems. Interest in continuous subcutaneous insulin infusion (CSII) via pumps for children has also
Abbreviations: CSII, continuous subcutaneous insulin infusion; HbA1c, hemoglobin A1c * Corresponding author at: Pediatric Endocrinology/Diabetology, Riley Hospital for Children, Room 5960, 702 Barnhill Drive, Indianapolis, IN 46202-5225, United States. Tel.: +1 317 274 3889; fax: +1 317 274 3882. E-mail address: [email protected] (Z.M. Nabhan).
been fueled by an increased emphasis on achieving strict glycemic control to minimize the risk of complications [2]. CSII therapy has multiple potential advantages over injection therapy, including an enhanced ability to mimic physiologic insulin release as well as permitting greater flexibility in food intake and physical activity. Several studies in adult populations have demonstrated a high degree of patient satisfaction on CSII therapy as well as improvements in quality of life [3,4]. Studies regarding safety and efficacy of CSII therapy in children and adolescents have provided mixed results. Some studies suggest an increased risk of diabetic ketoacidosis (DKA) on CSII compared to injections [5] whereas others show a reduction in DKA rates [6,7].
0168-8227/$ – see front matter # 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.diabres.2006.03.020
218
Z.M. Nabhan et al. / Diabetes Research and Clinical Practice 74 (2006) 217–221
Similarly, glycemic control has been reported to improve [8–10] or remain unchanged [11–13]. Most studies, however, concur that episodes of severe hypoglycemia appear to decrease while on CSII [7–9]. Few studies have tried to elucidate individual factors that predict or affect glycemic control of patients on CSII [14,15]. In adults, the frequency of self-monitoring of blood glucose has been found to be the most influential factor. Given the inconsistencies in the literature and the cost and complexity of CSII, identification of the best candidates for this therapy is inherently challenging. No analysis of multiple factors predictive of control in a large cohort of diabetic children and adolescents on insulin pumps has been previously performed. In this study, we investigated whether variables could be identified that correlated with good or poor control in a large cohort of pediatric patients on insulin pump therapy. We hypothesized that chart reviews would reveal clinically relevant factors that would enable better prospective selection of patients in whom pump therapy would be beneficial and well tolerated. 2. Methods Following institutional review board approval, medical records of patients with type 1 diabetes on insulin pump therapy followed at James Whitcomb Riley Hospital for Children were identified. All patients were started on insulin pump therapy by the clinical pediatric diabetology team at Riley Hospital and had been utilizing CSII for at least 1 year. Patient records were excluded if children had begun on pumps as part of research protocols. Charts for eligible patients were all reviewed by one author (L.R.). The following variables were extracted from the medical records: sex, race, age at diabetes onset, diabetes duration, age at CSII initiation, duration of pump therapy, hemoglobin A1c (HbA1c) values for 1 year prior to pump therapy through 1 year post-pump initiation, HbA1c at the last available follow-up, fast-acting insulin regimen prior to CSII (sliding scale or carbohydrate to insulin ratios), average number of injections per day prior to pump start, number of basal rates on pumps, pump brand, diabetes camp attendance, participation in a series of advanced diabetes education classes (ADAPT), family structure, and clinic visit frequency. ‘‘Good’’ control was defined by HbA1c stratified by age, with a target HbA1c (DCA 2000, Roche Diagnostics) of 9% for patients under 12 years of age, and 8% in patients over 12. The cut-off point for the HbA1c was made a priori based upon our diabetes program’s target blood sugar ranges for age. HbA1c at last available follow-up after pump start was used for analysis of predictors of glycemic control and ‘‘poor’’ and ‘‘good’’ control for age.
3. Pump program at Riley Hospital Patients were recommended for insulin pump therapy by their primary diabetologist if they had been diagnosed with type 1 diabetes for at least 12 months and had a history of compliance with physicians’ visits and home blood glucose monitoring. Before initiation of pump therapy, all families and patients attended a half-day education session, during which insulin types, insulin adjustment, and carbohydrate counting were reviewed. Families that had not previously used carbohydrate/insulin ratios or corrective doses began using these techniques after the education sessions. At the time of pump therapy initiation, all patients held their morning dose of long acting insulin, began using their pump by late morning, and remained through the afternoon for a second education session with a nurse educator during which patients and families were taught the mechanics of pump use. For the first 10 days, families were asked to check their child’s blood sugar at least eight times per day, with tests before meals, before snacks, at midnight, and at 3 a.m. Families were also asked to change infusion sets every 3 days. For the first 10 days, families were contacted daily by a nurse educator to make insulin adjustments for basal rates, carbohydrate/ insulin ratio, and correction boluses. Subsequently, phone follow-up occurred twice weekly until the insulin dose stabilized. The first follow-up appointment was within 2–3 months of the pump start. 3.1. Statistical analysis Statistics were done using Microsoft Excel 2000 (Microsoft Corp., Redmond, Wash) for descriptive statistics and SPSS Version 11.5 (SPSS Inc., Chicago Ill) for other analyses. Data are expressed as means standard deviations. T-tests were used for comparisons between groups when the data were normally distributed. Bonferroni correction was done for multiple comparisons. Correlations were done using Pearson’s test. 4. Results Ninety-three patients (57 girls and 36 boys) fit inclusion criteria and had charts available for review. Patients started on pumps between November 1998 and June 2002. Subject characteristics at the time of pump start are summarized in Table 1. Fifty-three children were over the age of 12 years. Forty-four (16 boys and 28 girls) of the 93 children were in ‘‘good’’ control for age at an average age of 10.4 3.2 years (range, 3.4–18.2 years)
Z.M. Nabhan et al. / Diabetes Research and Clinical Practice 74 (2006) 217–221
219
Table 1 Baseline patient characteristics Age at pump start (y)
11.6 3.1
Gender
F: 57 (60%) M: 36 (40%)
Diabetes duration at pump start (y) Baseline HbA1c
4.7 3.1 8.7 0.9%
Parental marital status
Married: 68 (73%) Divorced: 26 (27%)
Insulin regimen prior to CSII
Sliding scale: 30 (32%) Carb/Insulin ratio: 63 (68 %)
Number of injections/day Diabetes camp attendance Advanced diabetes educational session attendance (ADAPT) a
2.7 0.9 37 (40%) 37 (40%)
Data expressed as means standard deviations (S.D.); y: years; mo: months; F: female; M: male. a ADAPT is an outpatient education program for children and young adults with diabetes and their families. Through ADAPT patients learn how to develop good habits in managing diabetes.
and mean HbA1c of 8.0 0.6% (range, 6.6–9.0%). Those in poor control for age had an average age of 12.6 2.5 years (range, 5.7–17.9 years) and HbA1c of 9.3 0.8% (range, 8.1–11.9%). Most (77%) of the patients in poor control prior to pump start were over the age of 12 years. Following pump start, patients were seen in diabetes clinic every 3.3 0.4 months. At the time of last available follow-up, mean age of the patients was 13.9 3.1 years (range, 6.5–20.2 years). Average HbA1c prior to pump start was 8.7 0.9%. By 6 months post-CSII start, average HbA1c dropped to 8.1 0.9% ( p < 0.001) and then rose to 8.2 0.8% at 12 months ( p < 0.001 versus baseline, p = 0.34 versus 6 months). At the time of the last available follow-up after a mean duration of pump therapy of 2.4 0.8 years (range, 1.1–4.6 years), HbA1c was 8.3 0.6% (range, 6.9–9.7; p < 0.01 versus baseline) as shown in Fig. 1. At this time, 46 patients (18 boys and 28 girls) were in good control with an average HbA1c of 7.8 0.4% (range, 7.0–9.0%) and 47 (17 boys and 30 girls) were in poor control with an average HbA1c of 8.7 0.6% (range, 8.0–11.0%). 4.1. Predictors of glycemic control There was no correlation between the children’s age and HbA1c prior to pump start (r = 0.049; p = 0.64). However, there were positive correlations of both age at pump start and age at last visit with HbA1c at the time
Fig. 1. Mean HbA1c in 93 children with type 1 diabetes at pump start, 6 months post-CSII, 12 months post-CSII, and 28 months (mean time at the last available follow-up). p < 0.001 at 6 and 12 months vs. baseline and p < 0.01 at 28 months vs. baseline. Error bars show mean 1.0S.D.
of the last available follow-up visit (r = 0.227 for age at pump start, r = 0.216 for age at last visit, both p < 0.05). The mean number of basal rates was negatively correlated with mean HbA1c after pump start (r = 0.269; p < 0.01). Mean HbA1c after pump start did not correlate with diabetes duration, duration of CSII therapy, pre-pump insulin regimen, frequency of clinic visits, family structure, gender, diabetes camp attendance, or participation in advanced diabetes education sessions. The same variables were analyzed to identify predictors of change in HbA1c (difference between average pre-pump HbA1c and HbA1c at last available follow-up). Only average pre-pump HbA1c positively correlated with change in HbA1c (r = 0.751; p < 0.001) as shown in Fig. 2. 4.2. Predictors of ‘‘good’’ and ‘‘poor’’ control for age The data were also analyzed to examine determinants of good and poor control of diabetes (stratified by HgbA1c for age). Comparisons between the two groups are summarized in Table 2. Younger age at pump initiation and more basal rates predicted better control. Diabetes duration at pump start, duration of CSII therapy, pre-pump insulin regimen, pump brand,
220
Z.M. Nabhan et al. / Diabetes Research and Clinical Practice 74 (2006) 217–221
Table 2 Comparison between good control group and poor control group Comparison
Good control
Poor control
T-test
On pump A1C Pre-pump A1C Patient age at pump start a No of basal ratesa Diabetes duration at pump start Duration of CSII Gender (% female) Frequency of clinic visits (months) Pre-pump insulin regimen (number of shots/day) Family structure (% married) Pump brand (% medtronic)
7.8 0.4 8.6 0.9 10.0 3.1 4.4 1.3 4.3 2.6 2.3 0.7 58.7 3.3 0.4 2.7 0.9 76.1 66
8.7 0.6 8.8 1.0 13.1 2.2 3.4 1.1 5.24 3.3 2.4 0.8 63.8 3.3 0.4 2.7 0.1 68.1 67
p < 0.01 p = 0.26 p < 0.01 p < 0.01 p = 0.22 p = 0.26 p = 0.61 p = 0.69 p = 0.96 p = 0.39 p = 0.88
a
Good control correlated with age at pump start (r = 0.495; p < 0.01) and number of basal rates (r =
Fig. 2. Average pre-pump HbA1c positively correlated with change in HbA1c on CSII (r = 0.751; p < 0.001).
frequency of clinic visits, family structure, gender, diabetes camp attendance, and attendance of advanced diabetes education sessions did not differ between the groups. 5. Discussion Identifying variables that predict glycemic control in children and adolescents on insulin pump therapy is paramount in optimizing the use of this expensive, intensive therapy. However, little information is available to inform clinical diabetes management decisions. In our study, younger age at pump initiation
0.381; p < 0.01).
and number of basal rates on pump therapy correlated with better glycemic control. Although we did not have a direct measure of parental involvement, the younger age of the patients in good control likely reflects increased parental involvement with CSII therapy. Similarly, the number of basal rates may be a surrogate for the intensity and frequency of insulin adjustment. Our results concur with those of other investigators [14– 16] who have reported that greater parental involvement and younger age are associated with improved control on CSII. Consistent with prior studies [8–10], we also found a significant decrease in HbA1c following the initiation of CSII, suggesting that CSII improves metabolic control in children and adolescents with type 1 diabetes. Unlike what has been reported by other studies [5,17], this decrease was sustained for an average of 2.4 years in our patients on pumps. Although HbA1c increases slightly after the initial 6 months post-CSII level, it plateaus at 1 year without reverting to pre-pump values. The decrease at 6 months likely represents intensive involvement in diabetes management provided by the pump nurses that follows transition to pump therapy or it may also be due to more patient or parent attention with use of new technology. Another interesting finding in our study was the greater improvement in HbA1c in patients with worse control at baseline prior to initiation of pump therapy indicating that CSII does have the capacity to greatly improve outcomes in some previously poorly controlled patients. This finding has been reported by other investigators [18]. However, neither our study nor the other studies was controlled. It is possible that the greater decrease in HbA1c in patients with worse glycemic control prior to pump therapy represents simply improvement from increased contact for insulin
Z.M. Nabhan et al. / Diabetes Research and Clinical Practice 74 (2006) 217–221
adjustments with members of the diabetes team and increased frequency of blood glucose monitoring, rather than due to factors exclusive to the use of CSII. The strengths of this study are the large number of patients with a broad age range (5–20 years) and followup data for as long as 4.5 years after pump start. Also, our data collection incorporated multiple HbA1C assessments after pump start, as patients were followed every 3 months on average. Our study, however, has several limitations including its retrospective design. Data such as frequency of blood glucose monitoring, total daily dose of insulin, timing of boluses, bolus omission, family involvement, and rate of emergency room and hospital utilization were not available in the patient records. In addition, HbA1c was used to identify predictors of good control on CSII; however, good HbA1c levels are not the sole basis for choosing who would benefit from pump therapy. Pump therapy has been demonstrated to have several other advantages such as improving lifestyle flexibility and decreasing hypoglycemia and DKA episodes. These factors, however, were not evaluated in this study because of its retrospective design. Finally, the patients who received pump therapy generally came from families with a history of motivation and compliance with therapy and might not represent results that are applicable to a general population of youth with diabetes. In summary, our study corroborates the mounting evidence that CSII is effective in children and adolescents with type 1 diabetes. However, only a few specific factors help predict good glycemic control on insulin pump therapy. Further research into this area is needed. At present, decisions regarding which pediatric patients are most appropriate for CSII must continue to be highly individualized. References [1] M.J. Leonard, G.D. Reeves, Continuous subcutaneous insulin infusion: a comprehensive review of insulin pump therapy, Arch. Intern. Med. 19 (2001) 2293–2300. [2] The Diabetes Control, Complications Trial, Research Group, The effect of intensive diabetes treatment on the development and progression of long-term complications in insulin-dependant diabetes mellitus, N. Engl. J. Med. 329 (1993) 977–986. [3] R. Linkeschova, M. Raoul, U. Bott, M. Berger, M. Spraul, Less severe hypoglycemia, better metabolic control, and improved quality of life in type 1 diabetes mellitus with continuous subcutaneous insulin infusion (CSII) therapy; an observational
[4]
[5]
[6]
[7] [8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
221
study of 100 consecutive patients followed for a mean of 2 years, Diabet. Med. 19 (2002) 746–751. M. Graff, R.R. Rubin, E.A. walker, How diabetes specialists treat their own diabetes: findings from a study of the AADE and ADA membership, Diabetes Educ. 26 (2000) 460–467. N.H. White, Hollander As, M. Sadler, L. Daniels, Risks and benefits of continuous subcutaneous insulin infusion (CSII) in children (Abstract), Diabetes 50 (Suppl. 2) (2001) A66. B. Steindel, T.R. Roe, G. Costin, M. Carlson, F. Kaufman, Continuous subcutaneous insulin infusion (CSII) in children and adolescents with chronic poorly controlled type 1 diabetes mellitus, Diabetes Res. Clin. Pract. 27 (1995) 199–204. P. Blackett, Insulin pump treatment for recurrent ketoacidosis in adolescence, Diabetes Care 18 (1995) 881–882. E.A. Boland, M. Grey, A. Oesterle, L. Fredrickson, W.V. Tamborlane, Continuous subcutaneous insulin infusion: a new way to lower risk of severe hypoglycemia, improve metabolic control, and enhance coping in adolescents with type 1 diabetes, Diabetes Care 22 (1999) 1779–1784. F.R. Kaufman, M. Halvorson, D. Miller, M. Mackenzie, L.K. Fisher, P. Pitukcheewanont, Insulin pump therapy in type 1 pediatric patients: now and into the year, Diabetes Metab. Res. Rev. 15 (1999) 338–352. A.K. Maniatis, G.J. Klingensmith, R.H. Slover, C.J. Mowry, H.P. Chase, Continuous subcutaneous insulin infusion therapy for children and adolescents; an option for routine diabetes care, Pediatrics 107 (2001) 351–356. A. Schiffrin, M.M. Belmonte, Comparison between continuous subcutaneous insulin infusion and multiple injections of insulin—a one-year prospective-study, Diabetes 31 (1982) 255–264. N. Weintrob, H. Benzaquen, A. Galatzer, S. Shalitian, L. Lazar, G. Fayman, et al., Comparison of continuous subcutaneous insulin infusion and multiple daily injection regimens in children with type 1 diabetes: a randomized open crossover trial, Pediatrics 112 (2003) 559–564. L.A. DiMeglio, T.M. Pottorff, S.R. Boyd, L. France, N. Fineberg, E.A. Eugster, A randomized, controlled study of insulin pump therapy in diabetic preschoolers, J. Pediatr. 145 (2004) 380–384. B. Bode, D. Steed, P. Davidson, Determinants of glycemic control in insulin pump therapy, Diabetes 46 (Suppl. 1) (1997) 143. L.P. Plotnick, L.M. Clark, F.L. Brancati, T. Erlinger, Safety and effectiveness of insulin pump therapy in children and adolescents with type 1 diabetes, Diabetes Care 26 (4) (2003) 1142–1146. B. Anderson, J. Ho, J. Brackett, D. Finkelstein, L. Laffel, Parental involvement in diabetes management tasks: relationships to blood glucose monitoring adherence and metabolic control in young adolescents with insulin dependant diabetes mellitus, J. Pediatr. 130 (1997) 257–265. L. Laffel, C. Loughlin, N. Ramchan-Dani, D. Butler, N. Laffel, B. Levine, et al., Glycemic challenges of pump therapy in youth with type-diabetes (Abstract), Diabetes 50 (Suppl. 2) (2001) A66–A67. D. Iafusco, L. Fortunato, F. De Luca, F. Prisco, The egg or the chicken? Should good compliance to multiinjection insulin therapy criterion for insulin pump therapy, or does insulin pump therapy improve compliance?, J Pediatr. 148 (2006) 421.