Subjects with type 1 diabetes were studied at the five DirecNet clinical centers. A data and safety monitoring board and institutional review boards at each center approved the study protocol and consent and assent forms. A parent or guardian and each subject aged >7 years gave written consent and assent, respectively.
Eligibility criteria for type 1 diabetic subjects were 1) age 3 to <8 years or 12 to <18 years, 2) duration of type 1 diabetes of ≥1 year, 3) use of an insulin pump, and 4) A1C ≤10.0% (DCA2000+ analyzer; Siemens Healthcare Diagnostics, Indianapolis, IN). Exclusion criteria included a severe hypoglycemic event resulting in seizure or loss of consciousness in the last month, use of systemic or inhaled corticosteroids in the last month, or cystic fibrosis.
Subjects with type 1 diabetes were admitted to the research center on the evening before the study. An intravenous catheter was inserted in an arm vein for blood sampling. Glucose measurements were made using the One Touch Ultra meter at bedtime, 12:00 a.m., 3:00 a.m., 6:00 a.m., and 7:00 a.m. Oral carbohydrates were given to prevent low glucose levels with the goal of having the glucose values ≥110 mg/dl at 8:00 a.m. If glucose levels were high, and it was predicted that it would take a long time for the glucose level to drop to 110 mg/dl, small insulin correction doses were given.
At the start of the test, a bolus dose of insulin equal to ~1 hour of the subject's basal rate at that time was given, and the basal insulin infusion rate was increased by 25–50%. The basal insulin rate was increased further, and additional insulin doses were given, if needed, to achieve a gradual decline in plasma glucose.
Blood samples for meter glucose measurements were obtained every 15 min until the glucose value was ≤100 mg/dl and at 5- to 10-min intervals thereafter, until the end of the study. Blood samples were obtained for determination of glucose, epinephrine, norepinephrine, cortisol, glucagon, and growth hormone at baseline, when the meter glucose was between 95 and 110 mg/dl (in duplicate) and then <90, <80, <70, and <60 mg/dl. Once the glucose level fell to <60 mg/dl, intravenous glucose was given, the basal rate was returned to normal, and breakfast was eaten. An additional blood sample was collected for glucose and hormone concentrations 15 min after treatment with glucose.
At each blood sampling time, parents were asked whether their child “looked low,” and adolescents were asked whether they “felt low.” Parents and adolescents were masked to the meter glucose, and they ranked their response on a 4-point scale where 0/1 denoted “Not at all”/“Very little,” 2 denoted “Some,” and 3 denoted “Very much.”
Blood samples were frozen before shipping. Glucagon, cortisol, growth hormone, and glucose concentrations were measured at the DirecNet Central Laboratory (University of Minnesota). Glucagon was measured by a radioimmunoassay (Linco Research, St. Charles, MO) with the primary antibody from guinea pig and the secondary antibody from goat. The lower limit of detection was 20 pg/ml (6 pmol/l). Coefficients of variation (CVs) were 6.5–8.8% on three control samples. Cortisol was assayed with a competitive chemiluminescence assay (Bayer Advia Centaur; Bayer HealthCare, Diagnostics Division, Tarrytown, NY), using a polyclonal rabbit antibody and a mouse monoclonal antibody coupled with paramagnetic particles. The lower limit of detection was 0.5 μg/dl (14 nmol/l). CVs were 11–12% on two control samples. Growth hormone was measured by a sandwich chemiluminescence assay (DPC Immulite; Diagnostic Products Corporation, Los Angeles, CA). Monoclonal mouse antibody was coated on the bead with a rabbit polyclonal antibody in the reagent. The lower limit of detection was 0.1 ng/ml (4 pmol/l). CVs were 5.9–9.1%. Glucose determinations were made using a hexokinase enzymatic method in the laboratory (9
) and measured by the meter at the bedside (5
Epinephrine and norepinephrine concentrations were measured at the Mayo Clinic Laboratory (Rochester, MN) using a reverse-phase (C18) high-pressure liquid chromatography column to separate norepinephrine and epinephrine, which were detected coulometrically, using an ESA Coulochem II instrument. The lower limit of detection was 10 pg/ml (0.06 nmol/l and 55 pmol/l, respectively). CVs were 7–11% and 6–7% on two control samples. Catecholamine samples were collected in EDTA tubes and frozen immediately. This method was determined to not cause any difference in values versus when samples were collected in EDTA-sodium bisulfite tubes.
Studies in 14 nondiabetic adolescents between 12 and 17 years of age (14.8 ± 2.1 years, mean ± SD) were performed at the Children's Hospital of Pittsburgh (Pittsburgh, PA) under a protocol approved by its institutional review board between 1999 and 2002. Subjects were studied in the morning after an overnight fast. Two intravenous catheters, one for blood sampling and one for glucose and insulin infusion, were inserted in the nondominant arm. After baseline blood samples for glucose and catecholamines were obtained, an intravenous insulin infusion was begun at a rate of 0.1 unit · kg−1 · h−1 throughout the study. Glucose levels were maintained at 103 ± 6 mg/dl for 60 min by a variable rate infusion of 20% dextrose. Subsequently, glucose levels were allowed to fall to 64 ± 4 mg/dl by decreasing the intravenous glucose infusion over a 20- to 25-min period. The hypoglycemic nadir was maintained for 60 min. Blood was withdrawn every 15 min for measurement of plasma epinephrine concentrations.
Epinephrine concentrations in these samples were originally measured using a high-pressure liquid chromatographic method on either the LCEC capsule N-46 Bioanalytical System (Lafayette, IN) or the ESA system (ESA, Chelmsford, MA). Extra plasma was frozen and stored at −70°C. A subset (n = 18) of these samples was reassayed in the DirecNet Central Laboratory. The 25th, 50th, and 75th quartile levels for the original analysis were 96, 174, and 312 pg/ml and the corresponding values from the DirecNet laboratory were 10, 73, and 186 pg/ml. In the nine samples obtained during the 1-h hypoglycemic phase of the Pittsburgh study, the original median peak plasma epinephrine concentration was 312 pg/ml (range 126–848) vs. 186 pg/ml (range 10–758) on reanalysis in the DirecNet Central Laboratory.
Based on the study by Jones et al. (11
), 50 subjects (25 in each of the two age-groups with type 1 diabetes) were estimated to be needed to detect, with 90% power and a type 1 error rate of 5%, a difference between the two groups in the peak epinephrine response, assuming a true mean difference of 300 pg/ml with an SD of 306 pg/ml. The study was discontinued after a preplanned interim analysis, when it was determined that a significant mean difference would be unachievable, even with a much larger sample size.
The comparison of the diabetes duration in the two age-groups was performed using a two-sample t test. Analyses involving plasma glucose concentrations used laboratory rather than meter glucose values. Time intervals between the <90 mg/dl and postglucose treatment blood samplings were compared using a paired t test. The plasma glucose threshold that stimulated a counterregulatory hormone response was defined as the point at which the hormone concentration was ≥3 SD above baseline (with the SD based on the duplicate blood samples at baseline). The 3-SD limits were 26 pg/ml for epinephrine, 44 pg/ml for norepinephrine, 0.83 μg/dl for cortisol, 1.2 ng/ml for growth hormone, and 16 pg/ml for glucagon. The proportions of subjects who had hormone responses in each age-group were compared using Fisher's exact test. Group comparisons of the change in hormone concentration from baseline to peak were performed using an ANCOVA model based on van der Waerden normal scores and adjusted for the corresponding baseline value.
The study was completed by 30 subjects with type 1 diabetes. One subject in each age-group was not included in the analysis because of blood sampling problems. Because the Pittsburgh study used a one-step hypoglycemic clamp, lowering glucose in one step from ~100 to 60 mg/dl, plasma epinephrine responses in this study were used only to describe the magnitude of the plasma epinephrine responses in nondiabetic subjects.