In the past years, treatment with continuous subcutaneous insulin infusion (CSII) has been increasingly used in younger age-groups of patients with type 1 diabetes, and several studies have reported advantages of CSII, but little information is based on case-controlled analyses (1,2). Although short-term effects of CSII were described (3), there is no clear evidence whether its use leads to a continuous improvement of glycemic control. In order to evaluate long-term effects, we performed a matched-pair analysis comparing the outcome in young patients with CSII versus multiple daily injection (MDI) over 1 year of treatment under routine conditions representative of a pediatric diabetes clinic.
RESEARCH DESIGN AND METHODS
Fifty-nine of 85 patients starting on CSII from 2001 to 2004 with treatment duration of at least 1 year were eligible. The control cohort was comprised of 580 patients treated with MDI within the same time period. Fifty-two matched pairs could be identified fulfilling the matching criteria: sex, age (± 1 year), diabetes duration (± 1 year), and HbA1c (A1C) level (± 0.3%) at study entry (start of CSII). In the case of multiple matched MDI subjects, only the subject closest to A1C and diabetes duration was selected.
Patients with MDI received treatment based on the basal-bolus principle and had 3.5 ± 0.9 injections per day. Forty-one patients (79%) used NPH, eight (15%) NPH and Semilente MC, and three (6%) Glargine as basal insulin.
Patients on CSII were treated with a short-acting insulin analog. Indications of CSII were patient’s preference or erratic blood glucose fluctuations (hypoglycemia, dawn phenomenon). CSII was started according to a local standardized approach (4). Treatment goals based on International Society for Pediatric and Adolescent Diabetes guidelines were the same in CSII and MDI patients (5).
Variables of interest were total daily insulin dose per kilogram body weight, BMI-SDS (standard deviation score), and A1C (DCA2000 Analyzer; normal range 4.3–5.7%) at study entry and 6 and 12 months. Incidence of severe hypoglycemia (as defined by the Diabetes Control and Complications Trial [6]) and diabetic ketoacidosis (DKA) (requiring intravenous infusion of insulin and fluids) was evaluated in the year before and after study entry. BMI-SDS was calculated based on German reference data (7). The study was approved by the local ethics committee. Informed consent was obtained from parents and/or patients.
Data were analyzed by SPSS 11.5. Differences for continuous variables between matched groups and within each treatment group at subsequent time points were assessed by two-sided paired t test with P ≤ 0.05 considered as statistically significant. The power to detect a difference in A1C at 12 months of at least 0.5% between CSII and MDI groups was 93%.
RESULTS
Sex (40.4% male), age (6.72 ± 3.42 years), diabetes duration (4.83 ± 2.96 years), and BMI-SDS (0.35 ± 0.89) were virtually identical in both groups. At study entry, mean A1C (±SD) was 8.17 ± 1.03% in the CSII and 8.17 ± 1.02% in the MDI group (P = 0.498). In patients with CSII, A1C decreased slightly at 6 months (8.02 ± 1.12%) and increased again after 12 months of treatment (8.27 ± 1.01%), although these changes were not statistically significant. In the MDI group, A1C significantly increased after 6 (8.41 ± 1.09%, P = 0.032) and 12 months (8.51 ± 1.19%, P = 0.028) as compared with baseline values.
In patients with CSII aged <12 years, A1C remained stable over 12 months and was significantly lower than in matched control subjects with MDI at 6 and 12 months, respectively (Table 1).
During 12 months, total daily insulin requirements remained stable in patients with CSII (0.96 ± 0.29 vs. 0.93 ± 0.24 units/kg) but increased in patients with MDI (0.88 ± 0.24 vs. 0.95 ± 0.25 units/kg, P = 0.012).
Body weight (BMI-SDS) did not change significantly in patients with CSII throughout the study period but increased in patients with MDI aged >12 years (Table 1).
At baseline, the incidence of severe hypoglycemia per 100 patient-years (±SE) was comparable between the groups (CSII 19.2 ± 7.3, MDI 28.8 ± 8.8; P = 0.389). During the study, the incidence was lower with CSII (5.8 ± 3.3) than with MDI (23.1 ± 10.5), although not statistically significant (P = 0.107).
Episodes of DKA occurred five times in the CSII and two times in the MDI group during the study period. DKA incidence per 100 patient-years (±SE) increased in both groups (CSII 0.0–9.6 ± 4.1, P = 0.024; MDI 0.0–3.8 ± 2.7, P = 0.159).
CONCLUSIONS
In the total cohort, introduction of CSII led to a temporarily small improvement of glycemic control within the first 6 months, which could not be sustained during the further course of treatment. However, in patients <12 years, A1C was consistently lower when treated with CSII compared with MDI, while virtually no difference was found in patients starting pump therapy during puberty.
There was a tendency toward weight gain in patients with MDI, while BMI-SDS remained unchanged in CSII patients throughout the study.
The risk of DKA has been reported to be enhanced in patients with CSII (2,8). In the present study, the incidence was 2.5-fold higher with CSII than with MDI; however, statistical significance was not achieved in the face of rare events. The reason for DKA in patients with CSII was catheter displacement followed by incorrect therapy management and lack of measurement of ketone bodies. Thus, intensified and repeated education is necessary to prevent DKA in patients with CSII.
The incidence of severe hypoglycemia decreased in patients with CSII, although no significant difference was detected between the groups during therapy due to insufficient power.
Patients accepting and taking advantage of CSII do represent a cluster with specific clinical, behavioral, and psychological characteristics that make them different from MDI-treated patients in many respects. As these factors have not been taken into account for matching, results should be regarded cautiously. Additionally, CSII patients received more education and attention for several weeks, while MDI patients continued to receive routine care. However, despite methodological limitations, our data indicate that CSII is effective in young children.
Clinical data of patients with CSII compared with matched controls with MDI according to age-groups
. | Study group CSII . | Control group MDI . | Paired t test (P) . | |||
---|---|---|---|---|---|---|
Age <12 years | ||||||
n | 23 | 23 | ||||
BMI (SDS) | 0.30 ± 0.78 | 0.29 ± 0.92 | 0.977 | |||
6 months | 0.28 ± 0.81 | 0.36 ± 0.99 | 0.796 | |||
12 months | 0.28 ± 0.77 | 0.29 ± 1.03 | 0.983 | |||
Total insulin (units/kg) | 0.902 ± 0.317 | 0.785 ± 0.195 | 0.084 | |||
6 months | 0.860 ± 0.260 | 0.780 ± 0.205 | 0.153 | |||
12 months | 0.931 ± 0.291 | 0.850 ± 0.246 | 0.113 | |||
A1C (%) | 7.97 ± 0.93 | 7.98 ± 0.90 | 0.257 | |||
6 months | 7.72 ± 0.79 | 8.19 ± 0.95 | 0.031 | |||
12 months | 7.97 ± 0.82 | 8.31 ± 0.85* | 0.033 | |||
Age ≥12 years | ||||||
n | 29 | 29 | ||||
BMI (SDS) | 0.43 ± 0.92 | 0.38 ± 0.97 | 0.806 | |||
6 months | 0.39 ± 0.90 | 0.52 ± 0.98* | 0.543 | |||
12 months | 0.40 ± 0.90 | 0.56 ± 0.89* | 0.367 | |||
Total insulin (units/kg) | 1.003 ± 0.258 | 0.950 ± 0.251 | 0.360 | |||
6 months | 0.872 ± 0.237* | 0.975 ± 0.258 | 0.066 | |||
12 months | 0.928 ± 0.196 | 1.021 ± 0.225 | 0.100 | |||
A1C (%) | 8.32 ± 1.09 | 8.32 ± 1.10 | 1.000 | |||
6 months | 8.26 ± 1.29 | 8.58 ± 1.18 | 0.234 | |||
12 months | 8.51 ± 1.10 | 8.67 ± 1.40 | 0.572 |
. | Study group CSII . | Control group MDI . | Paired t test (P) . | |||
---|---|---|---|---|---|---|
Age <12 years | ||||||
n | 23 | 23 | ||||
BMI (SDS) | 0.30 ± 0.78 | 0.29 ± 0.92 | 0.977 | |||
6 months | 0.28 ± 0.81 | 0.36 ± 0.99 | 0.796 | |||
12 months | 0.28 ± 0.77 | 0.29 ± 1.03 | 0.983 | |||
Total insulin (units/kg) | 0.902 ± 0.317 | 0.785 ± 0.195 | 0.084 | |||
6 months | 0.860 ± 0.260 | 0.780 ± 0.205 | 0.153 | |||
12 months | 0.931 ± 0.291 | 0.850 ± 0.246 | 0.113 | |||
A1C (%) | 7.97 ± 0.93 | 7.98 ± 0.90 | 0.257 | |||
6 months | 7.72 ± 0.79 | 8.19 ± 0.95 | 0.031 | |||
12 months | 7.97 ± 0.82 | 8.31 ± 0.85* | 0.033 | |||
Age ≥12 years | ||||||
n | 29 | 29 | ||||
BMI (SDS) | 0.43 ± 0.92 | 0.38 ± 0.97 | 0.806 | |||
6 months | 0.39 ± 0.90 | 0.52 ± 0.98* | 0.543 | |||
12 months | 0.40 ± 0.90 | 0.56 ± 0.89* | 0.367 | |||
Total insulin (units/kg) | 1.003 ± 0.258 | 0.950 ± 0.251 | 0.360 | |||
6 months | 0.872 ± 0.237* | 0.975 ± 0.258 | 0.066 | |||
12 months | 0.928 ± 0.196 | 1.021 ± 0.225 | 0.100 | |||
A1C (%) | 8.32 ± 1.09 | 8.32 ± 1.10 | 1.000 | |||
6 months | 8.26 ± 1.29 | 8.58 ± 1.18 | 0.234 | |||
12 months | 8.51 ± 1.10 | 8.67 ± 1.40 | 0.572 |
Data are given as means ± SD and were recorded at study entry.
Significant change from baseline values at study entry (P < 0.05, paired t test within groups).
Article Information
This work was supported by a research grant from Medtronic MiniMed Deutschland.
References
A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C Section 1734 solely to indicate this fact.