The Effects of Subcutaneous Insulin Infusion Versus Multiple Insulin Injections on Glucose Variability in Young Adults with Type 1 Diabetes: The 2-Year Follow-Up of the Observational METRO Study

Abstract

Background:

Type 1 diabetic patients have high instability of daily glucose levels. The aim of this study was to evaluate the long-term effects of continuous subcutaneous insulin infusion (CSII) therapy, compared with multiple daily injections of insulin (MDI), on glucose variability, in young type 1 diabetic patients transitioned to the adult diabetes care.

Methods:

Patients aged 18–30 years and considered eligible for insulin pump therapy were included in the study. Ninety-eight patients who started CSII therapy and 125 who remained in MDI completed a 2-year follow-up. Glucose variability was assessed with continuous glucose monitoring using blood glucose standard deviation (BGSD), mean amplitude of glycemic excursion (MAGE), continuous overall net glycemic action (CONGA-2 h), low blood glucose index, high blood glucose index, and average daily risk range.

Results:

MAGE and BGSD decreased in both groups, with adjusted differences at 2 years of −0.74 mM (95% confidence interval [CI] −1.22 to −0.26, P = 0.003) and −0.3 (CI −0.52 to −0.1, P = 0.005) favoring the pump-therapy group. No significant differences between groups in the other variability indexes were observed. HbA1c decreased in both groups without significant difference (0.05%, −0.26, 0.35, P = 0.77); fasting glucose, insulin dose, and overall hypoglycemia (daily, nocturnal, and severe) decreased more in patients with CSII, compared with those with MDI.

Conclusions:

Among young adults with type 1 diabetes transitioning from the pediatric care, the use of CSII is associated with lower glucose variability, fasting glycemia, and overall hypoglycemic events than MDI during a 2-year period of follow-up.

Introduction

Emerging adulthood is a particularly challenging period for young patients with type 1 diabetes, who are at risk of poor glycemic control, appearance of microvascular complications, and early mortality.^1

–4 As incidence rates of type 1 are growing,⁵ the quality of care for young adults is paramount to optimize diabetes control and prevent adverse diabetes-related outcomes. Current guidelines for the management of type 1 diabetes from the American Diabetes Association (ADA) recommend a HbA1c target <7.0% (53 mmol/mol) for many adults⁶ to prevent and delay the late microvascular complications, whereas a less stringent target (<7.5%, 58 mmol/mol) is suggested for both children and adolescents by both ADA and ISPAD (International Society for Pediatric and Adolescent Diabetes).^7,8 Achieving the glycemic target without hypoglycemia and with limited glucose variability is a desirable objective for type 1 diabetic patients, in which the absolute insulin deficiency leads to high instability of daily glucose levels.⁹

Intensive insulin therapy, including multiple daily injections of insulin (MDI, up to four or more injections per day) or continuous subcutaneous insulin infusion (CSII or insulin pump), is the recommended treatment regimen to lower HbA1c levels and prevent microvascular complications in type 1 diabetes.^10,11 There is some evidence from short-term randomized trials (RCTs)^12
–14 and case–control^15,16 or observational¹⁷ studies that CSII therapy is associated with reduced glucose variability in type 1 diabetes, although other studies^18,19 did not find any effect. However, no prospective study specifically investigated the long-term effect of intensive glucose control (IGC) on glucose variability in young type 1 diabetic patients transitioning from the pediatric clinic to the adult healthcare.

The aim of the present study was to evaluate the long-term effects of CSII, compared with MDI, on glucose variability, assessed with continuous glucose monitoring (CGM), in a population of young adults with type 1 diabetes with suboptimal glycemic control. To this purpose, we used data of the Management and Technology for Transition (METRO) study, a longitudinal observational study of type 1 diabetic patients in transition from the pediatric clinic to the adult diabetes care center. In accordance with the ADA recommendations for transition from pediatric to adult diabetes care systems,²⁰ we focused on the age range of 18–30 years.

Subjects and Methods

Study design and participants

The METRO study is a single-center, observational prospective study designed to evaluate the effects of CSII versus MDI on glycemic and metabolic outcomes in young adults with type 1 diabetes in transition from the Pediatric Clinic to the Diabetes Unit at the Teaching Hospital of University of Campania “Luigi Vanvitelli” (Naples, Italy). The study was approved by the local ethics committee, and all participants signed an informed consent before enrollment. The protocol of the study has been described elsewhere.^21,22 Briefly, patients aged 18–30 years, with type 1 diabetes for at least 12 months and previously treated with MDI, were recruited from March 2012 to March 2015 and followed over time. Participants were excluded if they were pregnant or planning to become pregnant in the next 2 years, were not able to use the study devices, used therapeutic CGM or needed retrospective CGM outside the planned study visits, had history of severe chronic diseases, or drug or alcohol abuse. Patients were considered eligible for CSII if they showed persistent HbA1c levels ≥7.5% (58 mmol/mol) despite optimized education therapy, recurrent severe hypoglycemic episodes or high glucose variability, and the willingness to wear the insulin pump. Patients eligible for CSII therapy who preferred to remain on MDI therapy were selected as control group.

All participants in the study used blood glucose meters and test strips and were asked to perform self-monitored plasma glucose measurements four times daily (bedtime and pre-prandial). Moreover, they were instructed to the correct use of the insulin devices (insulin pump or prefilled pen) and trained to carbohydrate counting, insulin dose adjustment, and management of hyper- and hypoglycemic episodes. All patients starting CSII therapy received a rapid acting insulin analog (lispro, aspart, or glulisine) through a multiprogrammable insulin infusion pump, whereas patients of MDI group were treated with three injections of a rapid acting insulin analog at meals and one injection of insulin glargine or degludec, usually at bedtime.

Participants of both groups (CSII and MDI) had regular follow-up visits at the Diabetes Center, each 2 weeks for the first month and each 12–16 weeks for the following period. All participants in the study were followed for 2 years; data on fasting glucose, HbA1c, insulin dose, weight, and mean amplitude of glycemic excursion (MAGE) were collected at 6, 12, and 24 months, and the other outcome measures at baseline and after 2 years.

Assessment of glucose variability

Participants of both groups underwent blinded CGM for 14 days at basal evaluation, and before the study visits at 6, 12, and 24 months. Patients used the Dexcom G4 CGM system (Dexcom, Inc., San Diego, CA) composed of a 7-day transcutaneous sensor, a transmitter, and a receiver. The sensor was implanted in the anterior abdominal wall. All patients were instructed to change the sensor every 7 days; moreover they were asked to perform the required sensor calibration procedure according to the manufacturer's instruction within 2 h from placing the sensor and then each 12 h. Glucose data have been downloaded with Dexcom STUDIO. Glycemic readings from CGM were entered in the EasyGV^© software (www.phc.ox.ac.uk/research/diabetes/software/easygv/) to calculate different glucose variability indexes, as there is no clear evidence of the superiority of one method over the others: the MAGE, the blood glucose standard deviation (BGSD), the continuous overall net glycemic action (CONGA-2 h), the low blood glucose index (LBGI), the high blood glucose index (HBGI), and the average daily risk range (ADRR). MAGE represents the most popular measure of glucose variability and has been proposed as the “gold standard” metric system for its evaluation.²³ Both MAGE and BGSD are good indicators of the within-day glucose excursions and mainly reflect wide variations of glucose levels, which often occur in insulin-treated patients.^9,23 In contrast, CONGA represents a more sensible index to detect glucose fluctuations in condition of good glycemic control (HbA1c <7.5% or <58 mmol/mol).²³ Furthermore, the other indexes best describe the measure and extend of hypoglycemia (LBGI), hyperglycemia (HBGI), and the average of the maximal daily amplitudes of glucose excursions across several days (ADRR).^9,23

Study measurement

A complete medical history, including smoking status and concomitant pathologies, was obtained by patients' interview or clinical chart review. Hypoglycemia was defined as symptoms or signs associated with hypoglycemia experienced by the patient and self-treated, or the glucose level reduction below 70 mg/dL (3.9 mM), whereas severe hypoglycemic episodes referred to an episode requiring assistance of another person for its resolution.²⁴ Daily and nocturnal hypoglycemia were defined as hypoglycemic events occurring between 7.01–23.00 and 23.01–7.00, respectively. Both daily and nocturnal hypoglycemic events were extracted from the glucose readings collected with CGM. Severe hypoglycemic episodes were collected from participants if reported at study visits. Height and weight were recorded with participants wearing lightweight clothing and no shoes using a Seca 200 scale (Seca, Hamburg, Germany) with attached stadiometer. Body mass index (BMI) was calculated as weight in kg divided by the square of height in meters (kg/m²). Total insulin dose was recorded at each study visit and expressed as units per kilograms of body weight. Arterial blood pressure was measured thrice, at the end of the physical examination with the subjects in sitting position, after a 15 min rest. Patients whose average blood pressure levels were greater or equal to 140/90 mmHg or who were under medication were classified as hypertensive. Eye complications were defined as the presence of any grade of diabetic retinopathy or maculopathy on dilated eye examination. Renal complications included micro- or macroalbuminuria. Diabetic neuropathy was assessed according to the recent American Diabetes Association guidelines for somatic and autonomic neuropathy.²⁵

Fasting plasma glucose, total and high density lipoprotein cholesterol, triglycerides, and HbA1c were measured at the hospital's chemistry laboratory.

Diabetes treatment satisfaction

The validated Italian version of the Diabetes Treatment Satisfaction Questionnaire (DTSQ) was administered at the beginning of the study and at the end of the 2-year follow-up.²⁶ This instrument evaluates changes in patients' satisfaction related to therapy modifications, but is also useful for comparing levels of satisfaction in subjects using different treatment strategies. The questionnaire consisted of eight questions: six questions were grouped in a total score addressing general satisfaction with a score from 0 to 6 for each question (0 = worst); one concerned the perception of hyperglycemia and another the perception of hypoglycemia, both with a score from 0 (none of the time) to 6 (most of the time).

Statistical analysis

The primary end point was blood glucose variability, measured as change in MAGE. Secondary end points were changes in HbA1c levels, fasting glucose concentrations, number and severity of hypoglycemia, total insulin dose, body weight, lipid profile, blood pressure, and diabetes treatment satisfaction. Continuous variables were reported as mean and standard deviation (SD) or median and interquartile range (IQR) according to distribution and compared with unpaired Student's t-test or Wilcoxon rank-sum test as appropriate. Categorical variables were reported as absolute number and percentages and compared with chi-square test or Fisher exact test when appropriate. Difference between the two treatment regimens (MDI vs. CSII) was estimated using propensity scores to adjust for the bias inherent to the different characteristics of patients. The propensity scores were estimated by fitting a logistic regression model with intensive insulin regimen (CSII or MDI) as dependent variable. The covariates included in the propensity score models were age, sex, diabetes duration, BMI, fasting glucose, glycated hemoglobin, insulin units, MAGE, total cholesterol, and systolic blood pressure at basal level. Repeated measure ANOVA analyses were performed for each variable, including as covariates propensity score as continuous variable, type of treatment regimen, the time points of observations (Time: 0, 6, 12, and 24 months), and their interaction. Adjusted differences (and 95% confidence interval) between the two treatment regimens at 2 years were estimated from these analyses. The same model was used for the other measures of variability. A sensitivity analysis was performed, including quintile categories of propensity score as covariates, instead of the linear term. A two-tailed P-value <0.05 was considered significant. Data were analyzed using STATA software version 14.0.

Results

Four hundred and fifty-two type 1 diabetic patients transitioning from the pediatric clinic to our Diabetes Unit were initially examined for eligibility for CSII therapy. Among the 252 eligible patients, 123 started CSII therapy, and the remaining 164 chose to continue MDI. A total of 223 patients completed the 2-year follow-up (125 patients in the MDI group and 98 patients in the CSII group) (Fig. 1). Mean age was 24.8 years, and mean diabetes duration was 13.9 years. Mean baseline HbA1c level was 8.5% (69 mmol/mol). Table 1 shows the baseline demographics and characteristics of the study population according to the treatment regimen. The two groups were well matched for demographic and clinical characteristics. A similar percentage of patients in both groups had microvascular complications or concomitant autoimmune diseases. The frequency of the overall hypoglycemic events (daily, nocturnal, and severe) was also similar in both groups.

FIG. 1.

Process of patients selection. CSII, continuous subcutaneous insulin infusion; MDI, multiple daily injections of insulin.

Table 1.

Baseline Characteristics of Participants in the Study

	All patients (n = 223)	CSII (n = 98)	MDI (n = 125)	P
Age, years	24.8 ± 3.1	25.3 ± 3.3	24.5 ± 2.9	0.06
Percentage of males (%)	57.3	52.1	61.6	0.19
Diabetes duration, year	13.9 ± 4.5	14.2 ± 4.9	13.7 ± 4.1	0.63
Weight, Kg	70.3 ± 10.1	70.7 ± 10.7	69.9 ± 9.5	0.53
BMI, Kg/m²	24.5 ± 3.1	24.8 ± 3.3	24.2 ± 2.9	0.15
Fasting glucose, mg/dL	220.5 ± 64.1	224 ± 67.1	217.8 ± 61.7	0.47
Fasting glucose, mM	12.3 ± 3.6	12.4 ± 3.7	12.1 ± 3.4	0.47
HbA1c,%	8.5 ± 1.1	8.6 ± 1.1	8.5 ± 1.2	0.94
HbA1c, mmol/mol	69.4	70.5	69.4	0.94
MAGE, mM	6.5 ± 1.8	6.6 ± 2.1	6.4 ± 1.6	0.54
CONGA-2 h, mM	6.6 ± 1.7	6.6 ± 1.6	6.6 ± 1.8	0.76
BGSD, mM	3.5 ± 0.8	3.3 ± 0.8	3.6 ± 0.8	0.05
HBGI	8.7 ± 3.7	8.5 ± 3.4	8.9 ± 3.9	0.41
LBGI	5.5 ± 1.4	5.5 ± 1.7	5.5 ± 1.2	0.88
ADRR	32 ± 7.9	31.9 ± 8.6	32.1 ± 7.5	0.87
Daily hypoglycemia, n (%)
0	1 (0.45)	0 (0.0)	1 (0.8)
≥1	222 (99.55)	98 (100.00)	124 (99.20)	0.37
Nocturnal hypoglycemia, n (%)
0	122 (54.71)	53 (54.08)	69 (55.20)
≥1	101 (45.29)	45 (45.92)	56 (44.80)	0.87
Severe hypoglycemia, n (%)
0	153 (68.61)	69 (70.41)	84 (67.20)
≥1	70 (31.39)	29 (29.59)	41 (32.80)	0.61
Insulin dose, UI/kg	0.73 ± 0.2	0.73 ± 0.2	0.72 ± 0.2	0.58
Prandial	0.36 ± 0.2	0.36 ± 0.1	0.35 ± 0.2	0.65
Basal	0.37 ± 0.1	0.37 ± 0.2	0.37 ± 0.1	0.96
Lipids, mg/dL
Total cholesterol	165.16 ± 28.1	167.3 ± 27.8	163.5 ± 28.3	0.33
LDL cholesterol	91 (76, 112)	96.5 (75, 115)	88 (76, 104)	0.50
HDL cholesterol	57.3 ± 12.8	57.6 ± 12.4	57.05 ± 13.1	0.77
Triglycerides	72 (54, 89)	74.5 (53, 97)	71 (54, 85)	0.24
Blood Pressure, mmHg
SBP	120 (90, 140)	120 (90, 140)	120 (90, 140)	0.05
DBP	75 (60, 90)	77.5 (60, 85)	70 (60, 90)	0.66
DTSQ
DTSQ total score	27.5 ± 3.5	27.7 ± 3.3	27.4 ± 3.7	0.67
perceived hypoglycemia	3 (2, 4)	3 (2, 3)	3 (2, 4)	0.12
perceived hyperglycemia	2 (2, 3)	2 (2, 3)	2 (2, 3)	0.28
Microvascular complications, n (%)	19 (8.52)	9 (9.18)	10 (8.0)	0.75
Comorbidities, n (%)
Thyroiditis	45 (20.2)	24 (24.5)	21 (16.8)	0.61
Celiac disease	16 (7.17)	6 (6.12)	10 (8.0)	0.81
Smoke, n (%)	51 (22.80)	17 (17.30)	34 (27.20)	0.11

Values are expressed as mean ± standard deviation (SD) or median (interquartile range).

MDI, multiple daily injections of insulin; CSII, continuous subcutaneous insulin infusion; BMI, body mass index; MAGE, mean amplitude of glycemic excursion; CONGA-2 h, continuous overall net glycemic action; BGSD, blood glucose standard deviation; HBGI, high blood glucose index; LBGI, low blood glucose index; ADRR, average daily risk range; SBP, systolic blood pressure; DBP, diastolic blood pressure; DTSQ, diabetes treatment satisfaction questionnaire.

Primary end point

Among patients treated with CSII, MAGE fell from baseline to 6 months and remained steady for the remainder of the study with significant differences compared with patients on MDI (Fig. 2). This change was significant between the two groups even after adjustment with the propensity score for the interaction treatment/time (Table 2). Table 3 summarizes the results at the end of the follow-up. After 2 years, both MAGE and BGSD decreased to 5.7 and 3.2 mM, respectively, in the CSII group, compared with 6.3 and 3.5 mM in the MDI group; the between-group adjusted differences favored the pump-therapy group by −0.74 mM (95% CI, −1.22 to −0.26, P < 0.01) and −0.3 (−0.52 to −0.1, P < 0.01), respectively. CONGA, HBGI, LBGI, and ADRR decreased in both CSII and MDI groups, without any significant difference between them.

FIG. 2.

HbA1c levels, MAGE, fasting glucose, and insulin dose at 6, 12, and 24 months in all the study patients according to insulin regimen. Values are mean ± SE. Asterisks denote significant differences for all comparisons between pump therapy and injection therapy at each time point. * P < 0.05; **P < 0.01; ***P < 0.001. MAGE, mean amplitude of glycemic excursion.

Table 2.

Adjusted Models with Propensity Scores for Treatment, Time, and Interaction Treatment/Time

Variable	P	Treatment	Time	Treatment/time
MAGE	Unadjusted	0.01	<0.001	<0.01
MAGE	Adjusted	<0.01	<0.001	<0.01
Fasting glycemia	Unadjusted	0.01	<0.001	<0.001
Fasting glycemia	Adjusted	0.02	<0.001	<0.001
HbA1c	Unadjusted	0.69	<0.001	0.86
HbA1c	Adjusted	0.55	<0.001	0.86
Insulin dose	Unadjusted	<0.001	<0.001	<0.001
Insulin dose	Adjusted	0.001	<0.001	<0.001

Table 3.

Main Outcomes at 2 Years

	CSII (n = 98)			MDI (n = 125)
Parameter	Baseline	End point	Difference	Baseline	End point	Difference	Adjusted difference ^a (95% CI)	P value
Weight, Kg	70.7 ± 10.7	71.6 ± 10.7	0.85 ± 3.73	69.9 ± 9.5	70.3 ± 10.0	0.45 ± 4.48	0.63 (−2.11, 3.42)	0.65
BMI, kg/m²	24.8 ± 3.3	25.1 ± 3.2	0.32 ± 1.49	24.2 ± 2.9	24.4 ± 2.7	0.23 ± 1.75	0.12 (−0.65, 0.93)	0.75
Fasting glucose, mg/dL	224.1 ± 67.1	162.3 ± 54.5	−61.73 ± 72.4	217.8 ± 61.7	193.3 ± 67.5	−24.5 ± 74.6	−31.2 (−47.91, −14.42)	<0.001
Fasting glucose, mM	12.4 ± 3.7	9.0 ± 3.0	−3.4 ± 4.0	12.1 ± 3.4	10.7 ± 3.8	−1.4 ± 4.1	−1.7 (−2.7, −0.8)	<0.001
HbA1c,%	8.6 ± 1.1	8.1 ± 1.0	−0.41 ± 1.07	8.5 ± 1.2	8.1 ± 1.3	−0.42 ± 1.0	0.05 (−0.26, 0.35)	0.77
HbA1c, mmol/mol	70.5	65.0	—	69.4	65.0	—	—	0.77
Insulin dose, UI/kg	0.73 ± 0.2	0.63 ± 0.1	−0.10 ± 0.14	0.72 ± 0.2	0.73 ± 0.2	0.01 ± 0.14	−0.1 (−0.15, −0.06)	<0.001
Prandial	0.36 ± 0.1	0.31 ± 0.2	−0.05 ± 0.15	0.35 ± 0.2	0.35 ± 0.1	0.01 ± 0.12	−0.06 (−0.14, 0.02)	0.13
Basal	0.37 ± 0.2	0.32 ± 0.1	−0.05 ± 0.13	0.37 ± 0.1	0.38 ± 0.1	0.01 ± 0.14	−0.06 (−0.13, 0.01)	0.09
MAGE, mM	6.6 ± 2.1	5.7 ± 1.9	−0.93 ± 2.55	6.4 ± 1.6	6.3 ± 1.6	−0.12 ± 1.03	−0.74 (−1.22, −0.26)	<0.01
CONGA-2 h, mM	6.6 ± 1.6	6.3 ± 1.7	−0.31 ± 1.47	6.6 ± 1.8	6.4 ± 1.8	−0.19 ± 0.92	−0.01 (−0.48, 0.46)	0.97
SD, mM	3.3 ± 0.8	3.2 ± 0.8	−0.17 ± 0.79	3.6 ± 0.8	3.5 ± 0.8	−0.06 ± 0.54	−0.3 (−0.52, −0.10)	<0.01
HBGI	8.5 ± 3.4	7.6 ± 3.0	−0.87 ± 3.28	8.9 ± 3.9	8.4 ± 3.3	−0.53 ± 2.16	−0.9 (−1.8, 0.05)	0.06
LBGI	5.5 ± 1.7	5.0 ± 1.9	−0.56 ± 1.3	5.5 ± 1.2	5.2 ± 1.4	−0.29 ± 1.06	−0.15 (−0.56, 0.26)	0.47
ADRR	31.9 ± 8.6	30.3 ± 8.0	−1.61 ± 5.22	32.1 ± 7.5	30.5 ± 6.8	−1.56 ± 2.61	−0.05 (−2.10, 2.05)	0.97
Total cholesterol	167.3 ± 27.8	167.8 ± 30.2	0.58 ± 23.54	163.5 ± 28.3	159.3 ± 28.3	−4.19 ± 23.2	4.5 (−3.01, 12.01)	0.24
LDL cholesterol	96.5 (75, 115)	86.0 (71.0, 114.0)	−2 (−12, 9)	88.4 (76, 104)	88.6 (72, 102)	−5 (−15, 5)	1.7 (−5.63, 8.91)	0.65
HDL cholesterol	57.6 ± 12.4	58.6 ± 14.4	1.1 ± 10.4	57.1 ± 13.1	58.7 ± 14.7	1.6 ± 9.6	−0.7 (−4.42, 3.04)	0.70
Triglycerides	74.5 (53, 97)	75.5 (56, 90)	2 (−15, 9)	71.4 (54, 85)	69.7 (58, 88)	0 (−10, 9)	0.14 (−7.21, 7.53)	0.97
SBP	120 (90, 140)	115 (110, 120)	0 (−10, 0)	120 (90, 140)	115 (110, 120)	0 (0, 0)	−1.5 (−4.03, 1.10)	0.26
DBP	75 (60, 90)	70 (70, 80)	0 (−10, 0)	70 (60, 90)	70 (70, 80)	0 (−5, 0)	−1.15 (−3.13, 0.81)	0.25
DTSQ total score	27.5 ± 3.5	29.1 ± 3.7	1 (0, 3)	27.4 ± 3.7	28.2 ± 3.4	1 (−1, 2)	0.5 (−0.42, 1.52)	0.27
Perceived hypoglycemia	3 (2, 4)	2 (1, 3)	−1 (−2,0)	3 (2, 4)	3 (2, 3)	−1 (−1, 0)	−0.5 (−0.83, −0.21)	<0.01
Perceived hyperglycemia	2 (2, 3)	2 (1, 3)	0 (−1, 1)	2 (2, 3)	2 (1, 3)	−0.5 (−1, 0)	−0.02 (−0.35, 0.30)	0.93

Difference between groups adjusted with the propensity score.

For definition of abbreviations, refer to Table 1 footnotes.

Secondary end points

HbA1c decreased by ∼0.4% (P < 0.01) at 2 years in both groups, without significant difference between them (Table 3). The lowest value was reached at 12 months and then slightly increased over the remaining follow-up period (Fig. 2). Fasting blood glucose levels decreased in both groups, with significant differences at 1 and 2 years favoring patients in CSII group [adjusted mean difference at the end point −31 mg/dL (−47.9 to −14.4, P < 0.001; 1.7 mM, −2.6 to −0.8)] (Fig. 2). Insulin dose was reduced in CSII group, and remained unchanged in MDI group, with significant differences between groups at 6 months, which remained significant over the entire follow-up [adjusted mean difference at the end point −0.1 UI/kg (−0.15 to −0.06, P < 0.001)] (Fig. 2). All these findings were consistent with the results of the repeated measures ANOVA analyses (Table 2). At 2 years, changes in lipid levels and blood pressure were not significantly different between the two groups. There were no significant changes in both the DTSQ overall scores and the rate of perceived hyperglycemia between CSII and MDI groups, whereas the score relative to the perceived hypoglycemia decreased more in patients with CSII than in those with MDI, with an adjusted difference of −0.5 (−0.8 to −0.2, P < 0.01). Sensitivity analysis, including quintile categories of propensity score as covariates, instead of the linear term showed in the repeated measures analyses confirmed these results (data not shown).

Table 4 describes the frequency of hypoglycemic events in the two groups at baseline and after 2 years. The proportion of patients with one or more daily hypoglycemic events decreased more in the CSII group and remained unchanged in the MDI group (P < 0.01). Moreover, the percentage of patients with one or more episode of nocturnal hypoglycemia decreased in the CSII group and increased in the MDI group (P < 0.01). Similar results were obtained when comparing the proportion of patients with one up to six nocturnal hypoglycemic episodes between groups (Table 5). The proportion of patients experiencing one or more episodes of severe hypoglycemia decreased in both groups; in both cases the difference at 2 years favored patients in the CSII group (P < 0.05).

Table 4.

Frequency of Hypoglycemic Events According to the Treatment Regimen

	Baseline			2 years
Events, n	CSII (n = 98)	MDI (n = 125)	P	CSII (n = 98)	MDI (n = 125)	P
Daily hypoglycemia, n (%)
0	0 (0.00)	1 (0.80)		10 (10.20)	1 (0.80)
≥1	98 (100.0)	124 (99.20)	0.37	88 (89.80)	124 (99.20)	<0.01
Nocturnal hypoglycemia, n (%)
0	53 (54.08)	69 (55.20)		71 (72.45)	64 (51.20)
≥1	45 (45.92)	56 (44.80)	0.87	27 (27.55)	61 (48.80)	<0.01
Severe hypoglycemia, n (%)
0	69 (70.41)	84 (67.20)		88 (89.80)	99 (79.20)
≥1	29 (29.59)	41 (32.80)	0.61	10 (10.20)	26 (20.80)	<0.05

Table 5.

Frequency of Nocturnal Hypoglycemia According to the Treatment Regimen at Baseline and After 2 Years

	MDI	CSII
	n (%)	n (%)	P
Nocturnal hypoglycemia at baseline (events)
0	69 (55.20)	53 (54.08)	0.363
1	33 (26.40)	22 (22.45)
2	18 (14.40)	12 (12.24)
3	4 (3.20)	6 (6.12)
4	1 (0.80)	4 (4.08)
5	0 (0.00)	1 (1.02)
Total	125 (100.00)	98 (100.00)
Nocturnal hypoglycemia at 2 years (events)
0	64 (51.20)	71 (72.45)	0.010
1	34 (27.20)	20 (20.41)
2	11 (8.80)	5 (5.10)
3	7 (5.60)	1 (1.02)
4	6 (4.80)	0 (0.00)
5	2 (1.60)	1 (1.02)
6	1 (0.80)	0 (0.00)
Total	125 (100.00)	98 (100.00)

In the first column, 0–5 (at baseline) and 0–6 (at 2 years) indicate the number of nocturnal hypoglycemic events. The number and the percentage of patients in both MDI and CSII group experiencing a certain number of nocturnal hypoglycemia are reported in the second and in the third column, respectively.

Discussion

Among young adults with type 1 diabetes with suboptimal glycemic control in transition from the pediatric clinic, the use of insulin pump, compared with multiple insulin injections, resulted in a greater decrease in glucose variability, fasting glucose, and insulin dose, but in similar reduction of HbA1c levels, during a 2-year follow-up period; moreover, CSII was better in decreasing overall hypoglycemia (daily, nocturnal and severe) and improving the rate of perceived hypoglycemia. Although previous studies have demonstrated benefits of insulin pump therapy on glucose variability in patients with type 1 diabetes,^12

–17 no study evaluated young adults over a long follow-up period using CGM to determine glucose variability indexes. CGM may be considered the best instrument for the evaluation of glycemic fluctuation in diabetes, given the very high number of determinations necessary to evaluate the parameters such as BGSD, MAGE, and CONGA. A sustained reduction in MAGE and BGSD over the other indexes of glucose variability occurred in patients treated with CSII but not in those with MDI and translated in less glucose fluctuations and more homogeneous glycemic levels. Interestingly, compared with patients on MDI, patients in the CSII group showed a greater decrease in fasting glucose and overall hypoglycemia, which could have contributed to the reduction of glycemic fluctuations.

The role of glucose variability in the development of diabetes complications is still debated. A positive association between glucose variability and diabetes-related complications has been reported in type 2 diabetes, although this evidence mostly comes from retrospective studies.²⁷ In contrast, there is still a contradictory evidence about the role of glucose variability in vascular complications of type 1 diabetes. Data from the Diabetes Control and Complication Trial failed to demonstrate a role of the glycemic variability in the development of diabetic microvascular complications in the long term.^28,29 The use of the 7-point glucose profile to determine the measures of glucose variability remains the main limitation of these studies, as it fails to capture the full degree of glycemic fluctuations as the CGM can capture. Despite this, in vitro studies^30,31 showed a more detrimental effect of glucose fluctuations than persistent hyperglycemia on endothelial function, mainly due to the generation of oxidative stress, which has been suggested as the key link between hyperglycemia and diabetic complications.

We found a similar reduction in HbA1c levels in both CSII and MDI groups. Our findings are novel, given the lack of long-term studies comparing the effects of insulin pump or multiple injections on glycemic outcomes in transition-age youth with type 1 diabetes. A 2-year RCT of 267 adults with type 1 diabetes (mean age 41.5 years) assigned to CSII or MDI therapy reported a not significant mean difference of −0.24% (−0.53 to 0.05) in HbA1c levels favoring participants with insulin pump.³² In contrast, two recent meta-analyses of RCTs concluded that use of insulin pump is more effective than MDI at decreasing HbA1c in adults with type 1 diabetes.^33,34 Moreover, previous long-term (up to 5 years) observational studies of both children³⁵ and adults^36,37 with type 1 diabetes showed that CSII therapy, compared with MDI, may produce a significant decrease in HbA1c levels which declines over time. Obtaining a good level of glycemic control for the transition-age patients is difficult, as this represents a particularly vulnerable population of patients, subjected to the influence of the “metabolic memory”.¹¹ According to the American T1D Exchange registry, only the 14% of young adults aged 18–25 are currently achieving the ADA recommended target of HbA1c.³⁸

The prevention of severe hypoglycemia remains a challenge in the management of type 1 diabetic patients. The DCCT/EDIC study³⁹ showed that, after about 30-year follow-up, the rate of severe hypoglycemia was similar in both the intensive and conventional treatment groups, in association with advancing duration of diabetes and similar HbA1c levels. In our study, severe hypoglycemia decreased with both intensive insulin regimens, although the number of patients with more than one episode of hypoglycemia was smaller in the CSII group. Compared with patients of the DCCT/EDIC cohort, our population of young type 1 diabetic patients may reflect an earlier diabetes stage which can primarily take advantage of insulin pump to reduce the risk of severe hypoglycemia. A longer follow-up could clarify whether the rate of severe hypoglycemia will equilibrate over time.

In a shorter follow-up (3 months), young type 1 diabetic patients treated with CSII had a higher degree of satisfaction for treatment, compared with those treated with MDI.²² With the longer follow-up (2 years), the overall DTSQ score was similar between patients treated with both IGC regimens, suggesting that the initial advantage obtained with CSII is lost over time. The situation may be different in adults with type 1 diabetes, as the largest case–control, cross-sectional study of 1341 adults with type 1 diabetes treatment with CSII was associated with a markedly higher DTSQ score compared with MDI.⁴⁰

Compared with young patients with MDI, those with CSII showed lower levels of fasting glycemia, associated with a significant reduction of insulin dose. This may have occurred because patients using MDI have less flexibility in adjusting their insulin basal delivery than pump users, who can modify the basal insulin rate in the different times of the day and night according to increasing or decreasing glucose concentrations or planned activities, including physical exercise.⁴¹ Moreover, the higher rate of nocturnal hypoglycemic events occurring in patients with MDI, compared with those treated with CSII, may have contributed to increasing fasting glucose levels of patients using insulin injections, as a result of both a counter-regulatory response and corrections made by patients themselves.

Strengths of this study are the narrow age range of type 1 diabetic patients investigated (18–30 years), the largest sample so far investigated in this age range, the long-term follow-up, the contemporary evaluation of many indexes of glucose variability, and the use of a valid tool (CGM) to measure glucose fluctuations. This study also has limitations. First, the lack of randomization due to its observational nature may have introduced a selection bias. However, this study represents a real life experience, reflecting the indication by ADA and EASD Diabetes Technology Working group to an appropriate selection of candidates for pump therapy, which must have poor glycemic control and be willing and able to use the device.⁴² Moreover, the use of the propensity score allowed to eliminate, at least in part, the allocation bias. The single-center nature of the study may be seen as another limitation due to the lack of generalizability of our findings. In contrast, homogeneity is much higher in single-center studies, compared with multicenter studies, as indicated by the lack of any differences in clinical and metabolic characteristics between the two groups at baseline.

In conclusion, in young adults with type 1 diabetes transitioning from the pediatric care, the use of CSII is associated with lower glucose variability, fasting glycemia, and overall hypoglycemic events than MDI during a 2-year period of follow-up. Our data support the use of CSII therapy as an appropriate treatment regimen to improve glycemic control in transition-age adults with type 1 diabetes, with the advantage of less glucose variability and hypoglycemia.

Footnotes

Acknowledgments

The authors thank all the members of the METRO study group, and specifically Dr. Filomena Castaldo, Dr. Maria Rosaria Improta, and Dr. Annalisa Sarnataro for the medical expertise; Dr. Mariangela Caputo for the dietetic educational support; and Sig. Concetta Verazzo for nursing assistance.

Author Disclosure Statement

No competing financial interests exist.

References

Bryden

, Dunger

, Mayou

, et al.: Poor prognosis of young adults with type 1 diabetes: a longitudinal study. Diabetes Care, 2003; 26:1052–1057.

Dabelea

, Stafford

, Mayer-Davis

, et al.; SEARCH for Diabetes in Youth Research Group: Association of Type 1 Diabetes vs Type 2 diabetes diagnosed during childhood and adolescence with complications during teenage years and young adulthood. JAMA, 2017; 317:825–835.

Lotstein

, Seid

, Klingensmith

, et al.; SEARCH for Diabetes in Youth Study Group. Transition from pediatric to adult care for youth diagnosed with type 1 diabetes in adolescence. Pediatrics, 2013; 131:e1062–e1070.

Laing

, Jones

, Swerdlow

, et al.: Psychosocial and socioeconomic risk factors for premature death in young people with type 1 diabetes. Diabetes Care, 2005; 28:1618–1623.

Mayer-Davis

, Lawrence

, Dabelea

, et al.; SEARCH for Diabetes in Youth Study: Incidence trends of type 1 and type 2 diabetes among youths, 2002–2012. N Engl J Med, 2017; 376:1419–1429.

American Diabetes Association: 6. Glycemic Targets. In Standards of Medical Care in diabetes-2017. Diabetes Care, 2017; 40 (Suppl 1):S48–S56.

American Diabetes Association: 12. Children and Adolescents. In Standards of Medical Care in Diabetes-2017. Diabetes Care, 2017; 40 (Suppl 1):S105–S113.

Rewers

, Pillay

, de Beaufort

, et al.: International Society for Pediatric and Adolescent Diabetes.: ISPAD Clinical Practice Consensus Guidelines 2014. Assessment and monitoring of glycemic control in children and adolescents with diabetes. Pediatr Diabetes, 2014; 15 (Suppl 20):102–114.

Candido

: Which patients should be evaluated for blood glucose variability?. Diabetes Obes Metab, 2013; 15( Suppl 2):9–12.

10.

American Diabetes Association: 8. Pharmacologic Approaches to Glycemic Treatment. In Standards of Medical Care in Diabetes-2017. Diabetes Care, 2017; 40 (Suppl 1):S64–S74.

11.

Nathan

; DCCT/EDIC Research Group: The diabetes control and complications trial/epidemiology of diabetes interventions and complications study at 30 years: overview. Diabetes Care, 2014; 37:9–16.

12.

Bruttomesso

, Crazzolara

, Maran

, et al.: In type 1 diabetic patients with good glycaemic control, blood glucose variability is lower during continuous subcutaneous insulin infusion than during multiple daily injections with insulin glargine. Diabet Med, 2008; 25:326–332.

13.

Hoogma

, Hammond

, Gomis

, et al.: 5-Nations Study Group.: Comparison of the effects of continuous subcutaneous insulin infusion (CSII) and NPH-based multiple daily insulin injections (MDI) on glycaemic control and quality of life: results of the 5-nations trial. Diabet Med, 2006; 23:141–147.

14.

Hirsch

, Bode

, Garg

, et al.; Insulin Aspart CSII/MDI Comparison Study Group; Insulin Aspart CSII/MDI Comparison Study Group: Continuous subcutaneous insulin infusion (CSII) of insulin aspart versus multiple daily injection of insulin aspart/insulin glargine in type 1 diabetic patients previously treated with CSII. Diabetes Care, 2005; 28:533–538.

15.

Lepore

, Corsi

, Dodesini

, et al.: Continuous subcutaneous insulin infusion is better than multiple daily insulin injections in reducing glucose variability only in type 1 diabetes with good metabolic control. Diabetes Care, 2010; 33:e81.

16.

Schreiver

, Jacoby

, Watzer

, et al.: Glycaemic variability in paediatric patients with type 1 diabetes on continuous subcutaneous insulin infusion (CSII) or multiple daily injections (MDI): a cross-sectional cohort study. Clin Endocrinol (Oxf), 2013; 79:641–647.

17.

Lepore

, Dodesini

, Nosari

, et al.: Effect of continuous subcutaneous insulin infusion vs multiple daily insulin injection with glargine as basal insulin: an open parallel long-term study. Diabetes Nutr Metab, 2004; 17:84–89.

18.

Bolli

, Kerr

, Thomas

, et al.: Comparison of a multiple daily insulin injection regimen (basal once-daily glargine plus mealtime lispro) and continuous subcutaneous insulin infusion (lispro) in type 1 diabetes: a randomized open parallel multicenter study. Diabetes Care, 2009; 32:1170–1176.

19.

Alemzadeh

, Palma-Sisto

, Parton

, et al.: Continuous subcutaneous insulin infusion and multiple dose of insulin regimen display similar patterns of blood glucose excursions in pediatric type 1 diabetes. Diabetes Technol Ther, 2005; 7:587–596.

20.

Peters

, Laffel

; American Diabetes Association Transitions Working Group: Diabetes care for emerging adults: recommendations for transition from pediatric to adult diabetes care systems: a position statement of the American Diabetes Association, with representation by the American College of Osteopathic Family Physicians, the American Academy of Pediatrics, the American Association of Clinical Endocrinologists, the American Osteopathic Association, the Centers for Disease Control and Prevention, Children with Diabetes, The Endocrine Society, the International Society for Pediatric and Adolescent Diabetes, Juvenile Diabetes Research Foundation International, the National Diabetes Education Program, and the Pediatric Endocrine Society (formerly Lawson Wilkins Pediatric Endocrine Society). Diabetes Care, 2011; 34:2477–2485.

21.

Maiorino

, Bellastella

, Petrizzo

, et al.: Treatment satisfaction and glycemic control in young Type 1 diabetic patients in transition from pediatric health care: CSII versus MDI. Endocrine, 2014; 46:256–262.

22.

Maiorino

, Casciano

, Della Volpe

, et al.: Reducing glucose variability with continuous subcutaneous insulin infusion increases endothelial progenitor cells in type 1 diabetes: an observational study. Endocrine, 2016; 52:244–252.

23.

Frontoni

, Di Bartolo

, Avogaro

, et al.: Glucose variability: an emerging target for the treatment of diabetes mellitus. Diabetes Res Clin Pract, 2013; 102:86–95.

24.

Seaquist

, Anderson

, Childs

, et al.; American Diabetes Association; Endocrine Society: Hypoglycemia and diabetes: a report of a workgroup of the American Diabetes Association and the Endocrine Society. J Clin Endocrinol Metab, 2013; 98:1845–1859.

25.

Pop-Busui

, Boulton

, Feldman

, et al.: Diabetic neuropathy: a position statement by the American Diabetes Association. Diabetes Care, 2017; 40:136–154.

26.

Nicolucci

, Giorgino

, Cucinotta

, et al.: Validation of the Italian version of the WHO-well-being questionnaire (WHO-WBQ) and the WHO-diabetes treatment satisfaction questionnaire (WHO-DTSQ). Diabetes Nutr Metab, 2004; 17:235–243.

27.

Smith-Palmer

, Brändle

, Trevisan

, et al.: Assessment of the association between glycemic variability and diabetes-related complications in type 1 and type 2 diabetes. Diabetes Res Clin Pract, 2014; 105:273–284.

28.

Kilpatrick

, Rigby

, Atkin

: Effect of glucose variability on the long-term risk of microvascular complications in type 1 diabetes. Diabetes Care, 2009; 32:1901–1903.

29.

Lachin

, Bebu

, Bergenstal

, et al.; DCCT/EDIC Research Group: Association of glycemic variability in type 1 diabetes with progression of microvascular outcomes in the diabetes control and complications trial. Diabetes Care, 2017; 40:777–783.

30.

Quagliaro

, Piconi

, Assaloni

, et al.: Intermittent high glucose enhances apoptosis related to oxidative stress in human umbilical vein endothelial cells: the role of protein kinase C and NAD(P)H-oxidase activation. Diabetes, 2003; 52:2795–2804.

31.

Piconi

, Quagliaro

, Da Ros

, et al.: Intermittent high glucose enhances ICAM-1, VCAM-1, E-selectin and interleukin-6 expression in human umbilical endothelial cells in culture: the role of poly(ADP-ribose) polymerase. J Thromb Haemost, 2004; 2:1453–1459.

32.

Heller

, White

, Lee

, et al.: A cluster randomised trial, cost-effectiveness analysis and psychosocial evaluation of insulin pump therapy compared with multiple injections during flexible intensive insulin therapy for type 1 diabetes: the REPOSE Trial. Health Technol Assess, 2017; 21:1–278.

33.

Yeh

, Brown

, Maruthur

, et al.: Comparative effectiveness and safety of methods of insulin delivery and glucose monitoring for diabetes mellitus: a systematic review and meta-analysis. Ann Intern Med, 2012; 157:336–347.

34.

Benkhadra

, Alahdab

, Tamhane

, et al.: Continuous subcutaneous insulin infusion versus multiple daily injections in individuals with type 1 diabetes: a systematic review and meta-analysis. Endocrine, 2017; 55:77–84.

35.

Fendler

, Baranowska

, Mianowska

, et al.: Three-year comparison of subcutaneous insulin pump treatment with multi-daily injections on HbA1c, its variability and hospital burden of children with type 1 diabetes. Acta Diabetol, 2012; 49:363–370.

36.

Carlsson

, Attvall

, Clements

, et al.: Insulin pump-long-term effects on glycemic control: an observational study at 10 diabetes clinics in Sweden. Diabetes Technol Ther, 2013; 15:302–307.

37.

Rosenlund

, Hansen

, Andersen

, et al.: Effect of 4 years subcutaneous insulin infusion treatment on albuminuria, kidney function and HbA1c compared with multiple daily injections: a longitudinal follow-up study. Diabet Med, 2015; 32:1445–1452.

38.

Miller

, Foster

, Beck

, et al.: Current state of type 1 diabetes treatment in the U.S.: updated data from the T1D exchange clinic registry. Diabetes Care, 2015; 38:971–978.

39.

Gubitosi-Klug

, Braffett

, White

, et al.; Diabetes Control and Complications Trial (DCCT)/Epidemiology of Diabetes Interventions and Complications (EDIC) Research Group: Risk of Severe Hypoglycemia in Type 1 Diabetes Over 30 Years of Follow-up in the DCCT/EDIC Study. Diabetes Care, 2017; 40:1010–1016.

40.

Nicolucci

, Maione

, Franciosi

, et al.: EQuality Study Group—Evaluation of QUALITY of Life and Costs in Diabetes Type 1. Quality of life and treatment satisfaction in adults with Type 1 diabetes: a comparison between continuous subcutaneous insulin infusion and multiple daily injections. Diabet Med, 2008; 25:213–220.

41.

Pickup

: Insulin-pump therapy for type 1 diabetes mellitus. N Engl J Med, 2012; 366:1616–1624.

42.

Heinemann

, Fleming

, Petrie

, et al.: Insulin pump risks and benefits: a clinical appraisal of pump safety standards, adverse event reporting, and research needs: a joint statement of the European Association for the Study of Diabetes and the American Diabetes Association Diabetes Technology Working Group. Diabetes Care, 2015; 38:716–722.