ASSESSMENT of /3-cell function in insulin-treated

Transcription

1 X/87/ $02.00/0 Journal of Clinical Endocrinology and Metabolism Copyright 1987 by The Endocrine Society Vol. 65, No. 1 Printed in U.S.A. Effects of Age, Duration and Treatment of Insulin- Dependent Diabetes Mellitus on Residual /?-Cell Function: Observations During Eligibility Testing for the Diabetes Control and Complications Trial (DCCT) THE DCCT RESEARCH GROUP* ABSTRACT. To examine the effects of age and duration and treatment of insulin-dependent diabetes (IDDM) on residual 0- cell function, we measured the fasting and Sustacal-stimulated serum C-peptide levels in 610 conventionally treated IDDM patients (age, yr; duration of diabetes, 1-15 yr) during eligibility screening for the Diabetes Control and Complications Trial (DCCT). Fasting and stimulated C-peptide values were closely correlated (r = 0.83; P < 0.001), and both declined with increasing duration of disease. However, among patients who had been diabetic for more than 5 yr, 11% (33 of 296) of adults compared with 0 of 75 adolescents (P < 0.001) retained substantial insulin secretory capacity. Patients with stimulated C-peptide levels greater than 0.2 pmol/ml had a significantly lower mean fasting plasma glucose level [177 ± 6 (±SEM) US. 222 ± 6 mg/dl; P < 0.001), a smaller rise in glucose after Sustacal administration (151 ± 5 us. 184 ± 3 mg/dl; P < 0.001), and lower hemoglobin A, c (8.4 ± 0.2% vs. 9.3 ± 0.1%; P < 0.001) than the patients with a stimulated C-peptide level of 0.05 pmol/ ASSESSMENT of /3-cell function in insulin-treated diabetic patients has been facilitated by the development of RIAs for the measurement of circulating C- peptide levels. While undetectable serum C-peptide values support the diagnosis of insulin-dependent diabetes mellitus (IDDM), it has become apparent that many such patients retain some insulin secretory capacity (1-10). Indeed, a surprising degree of variability in C-peptide concentrations has been reported in patients thought to have IDDM on clinical grounds. Interpretation of these reports is hindered by the inclusion of heterogeneous patient populations. There is, however, general agreement that the remission or "honeymoon" period is followed by a gradual loss of endogenous insulin secretion and some suggestion that the rate of decline of /3-cell function is related to the age of onset of IDDM (5). The impact of detectable but subnormal C-peptide values on Received September 12,1986. Address requests for reprints to: Box NDIC/DCCT, Bethesda, MD * Prepared for the DCCT by: Leslie J. Klaff, William V. Taniborlane, Patricia A. Cleary, Michael W. Steffes, and Dorothy J. Becker. A full list of investigators and members of the DCCT Research Group appeared in Diabetes 35:530,1986. ml or less, even though the C-peptide secretors were receiving less insulin (0.52 ± 0.02 us ± 0.02 U/kgday; P < 0.001). To determine the effects of treatment of /3-cell function, 33 patients with stimulated C-peptide values between 0.2 and 0.5 pmol/ml at entry in the DCCT were restudied 1 yr after randomization to standard treatment (n = 15) or an experimental (n = 18) treatment designed to achieve and maintain nearnormal glucose levels. Although C-peptide levels declined in both groups, experimental treatment was associated with slightly less of a decline in stimulated C-peptide values compared to Standard treatment. The results of C-peptide measurements in this large and well defined population of IDDM patients demonstrate that residual j8-cell function continues for a longer period of time in adults compared to adolescents with IDDM. This endogenous insulin secretion contributes significantly to metabolic control and may be prolonged by intensive insulin treatment regimens. (J Clin Endocrinol Metab 65: 30,1987) glycemic control (6-9) and microvascular complications of diabetes (3,10) is controversial. The Diabetes Control and Complications Trial (DCCT) is a multicenter randomized clinical trial designed to determine whether an intensive treatment regimen directed at maintaining blood glucose concentrations as close to normal as possible will affect the appearance or progression of early vascular complications in patients with IDDM (11). Each patient had to fulfill eligibility criteria that were thought to be important in establishing the clinical diagnosis of IDDM. Patients who fulfilled these criteria and were not excluded for other reasons also had measurements of their serum C- peptide levels before and after a standard mixed meal feeding (Sustacal, Mead-Johnson, Evansville, IN). The results of C-peptide testing in this large and well defined population of IDDM patients provided the opportunity to examine residual /3-cell function in IDDM, its relationship to factors such as age, disease duration and treatment, and its effects on metabolic control. Materials and Methods The initial eligibility screening interviews for the DCCT were designed to enroll a group of patients between yr old 30

2 0-CELL FUNCTION IN IDDM 31 with a clinical history of IDDM of 1- to 15-yr duration who were by history free of other medical problems and advanced diabetic complications. Other pertinent exclusion criteria included body weight more than 130% of ideal, previous use of 3 or more daily injections of insulin or an insulin pump, 3 or more episodes of diabetic ketoacidosis during the previous year, or a history of hypertension. Details concerning eligibility procedures have been reported previously (11). Six hundred and fifty-six patients passed this screening process and gave written informed consent for further eligibility testing, which included a comprehensive evaluation of ophthalmological, renal, and neurological status as well as the C-peptide response to Sustacal. Although the order of performance of eligibility studies varied from clinic to clinic, only 46 of the 656 patients were excluded from the DCCT before measurement of serum C-peptide levels; 20 patients voluntarily withdrew consent, and 26 patients were excluded on the basis of the medical assessments. The clinical characteristics of the remaining 610 patients who did undergo C-peptide testing and whose results are reported herein are summarized in Table 1. Basal and Sustacal-stimulated serum C-peptide levels below 0.2 pmol/l were two of the eligibility criteria for enrollment in the DCCT. In addition, patients with IDDM of 1- to 5-yr duration were included if their stimulated C-peptide values were between pmol/ml. Thirty-four of the 278 patients ultimately enrolled in the feasibility phase of the DCCT had had IDDM between 1-5 yr and had stimulated C-peptide values between pmol/ml; 19 were randomized to experimental treatment (insulin pump or multiple injections) and 15 to standard treatment (1 or 2 daily insulin injections). To assess the persistence of the C-peptide response and to examine the effect of metabolic control of diabetes on residual insulin secretory capacity, repeat Sustacal tests were performed in 33 of the 34 patients (18 experimental and 15 standard) 12 months after randomization. C-Peptide testing All Sustacal tests were performed in the postabsorptive state after an 8- to 12-h overnight fast. The usual morning insulin TABLE 1. Clinical characteristics of the 610 IDDM patients who had serum C-peptide concentrations measured 1. Age (yr) 2. % <18 yr old 3. Gender (% female) 4. Duration of IDDM (months) 5. Race (% white) 6. Daily insulin dose (U/kg) 7. HbA.C {%) 8. Hospitalizations for DKA in previous year (%; n = 555) Retinopathy (n = 466) Absent Nonproliferative Preproliferative or greater 24.9 ± ± (n = 557) 0.71 ± 0.28 (n = 553) 9.2 ± 1.8 (n = 548) n, The number of patients for whom data are available if less than Mean ± SD. injection was not given. Upon arrival at the center, a baseline blood sample was obtained for determination of serum glucose and C-peptide levels. The patient then drank 6 ml/kg (to a maximum of 360 ml) Sustacal (1 cal/ml; 55% carbohydrate, 24% protein, and 21% fat) in a period not exceeding 10 min. Ninety minutes after ingestion of Sustacal another blood sample was obtained for determination of glucose and C-peptide levels. Laboratory analyses Blood samples were kept at 4 C, and serum was separated within 30 min, frozen at -20 C, and shipped on dry ice within 7 days of collection to the Central Biochemistry Laboratory at the University of Minnesota. Serum glucose was measured within 1 week of sample collection by a glucose oxidase method using a glucose analyzer (Beckman Instruments, Inc., Fullerton, CA). C-Peptide was measured within 4 weeks of sample collection using the M-1230 antiserum and other reagents obtained from Novo Industri (Bagsvaerd, Denmark). The procedure followed that recommended by Novo (12), with the additional step of first precipitating serum samples, blanks, and standards with polyethylene glycol (25%, wt/vol) in distilled water. The interassay coefficient of variation was 3.3% at 0.28 pmol/ml, 3.9% at 0.15 pmol/ml, and 12% at 0.07 pmol/ml. The lower limit of detection of the assay was 0.01 pmol/ml. To determine the effect of short term storage at -20 C on C-peptide levels, blood samples were drawn from 3 nondiabetic subjects and 2 IDDM patients, and the serum was split into aliquots for assay at later times. There was no significant difference in C-peptide levels when these specimens were assayed on the day of blood collection [0.21 ± 0.08 (±SEM) pmol/ ml] or after 14 days (0.20 ± 0.08 pmol/ml) or 28 days (0.20 ± 0.08 pmol/ml) of storage at -20 C. As a further check of the storage and shipping procedures used in the DCCT, 43 blood specimens collected at various centers were divided into 2 aliquots. One aliquot was sent promptly to the Central Biochemistry Laboratory and assayed immediately, while the second was sent and assayed after local storage for 7-30 days at -20 C. The C-peptide levels in the first and second aliquots were similar (coefficient of reliability, 0.922). Hemoglobin A ic (HbA lc ) levels were measured by the Central Hemoglobin Laboratory at the Joslin Diabetes Center (Boston, MA) by a high performance liquid chromatography method (13). Data collection and statistical analysis The statistical analyses were conducted with the Statistical Analysis System (SAS). Coefficients of reliability were used to estimate the proportion of total variability between values that was not due to measurement error. The distributions of basal and stimulated C-peptide levels were tested for normality. Although the distributions clearly cannot be considered to originate from a normally distributed population, analyses of covariance and t tests perform well even if the data do deviate somewhat from the requirements of normality, homoscedasticity, and additivity (14). Linear regression was used to examine the relationship between basal and stimulated C-peptide, and basal and stimulated C-peptide and duration of IDDM, age, HbA lc and serum glucose

3 32 DCCT RESEARCH GROUP JCE & M 1987 Vol 65 No 1 levels. Tests of mean difference were performed using analysis of variance, analysis of covariance or two-sample t tests. All data are presented as the mean ± SEM, except as specified. C- Peptide levels below the detection limit of the assay were assigned a value of 0.01 pmol/ml for statistical analysis. Results We first examined the relationship between basal and Sustacal-stimulated serum C-peptide levels in the 610 individual patients. There was a highly significant direct correlation between basal and stimulated serum C-peptide levels (r = 0.83; P < 0.001). In 81 (19%) of the 423 of patients who had basal levels of 0.05 pmol/ml or less (mean, 0.03 pmol/ml), C-peptide increased above this value (to 0.13 ± 0.01 pmol/ml; range, pmol/ ml) when they were challenged with Sustacal. This subgroup of patients (defined as responders) had significantly lower fasting serum glucose values than those patients in whom both basal and stimulated C-peptide concentrations were 0.05 pmol/ml or less (179 ± 11 vs. 222 ± 5 mg/dl; P < ). On the other hand, the HbAic levels of the responders and nonresponders were similar (9.5 ± 0.2% vs. 9.3 ± 0.1%, respectively; P = NS). In the entire study population, stimulated C-peptide levels correlated more closely than basal C-peptide values with duration of IDDM (r = -0.4 and P < vs. r = -0.3 and P < 0.001), HbA lc (r = -0.3 and P < vs. r = -0.2; P < 0.01), and fasting glucose (r = -0.2 and P < 0.01 vs. r = -0.1 and P = NS). The effect of duration of diabetes on basal and stimulated C-peptide levels in the 610 patients is shown in Fig. 1. As expected, there was a gradual reduction in mean basal and stimulated C-peptide levels during the first 5 yr of diabetes, after which both mean levels were less than 0.10 pmol/l. The data were then examined to determine whether the decline in /3-cell function with time was similar in adults and adolescents with IDDM (Table 2 and Fig. 2). There was no significant difference in basal levels between adults and adolescents with 5-yr or less duration of IDDM, but mean stimulated values were slightly higher in adults (0.26 ± 0.02 vs ± 0.03 pmol/ml; P < 0.05). Although a greater proportion of adults (45%) than adolescents (37%) with 1- to 5-yr duration of diabetes had stimulated values greater than 0.20 pmol/ml, the difference was not significant. In patients with diabetes of more than 5-yr duration, mean stimulated C-peptide levels were also significantly greater in adults than in adolescents (P < 0.03; Table 2). Even more important, none of the adolescents but 11% of the adults with diabetes of more than 5-yr duration had stimulated values over 0.20 pmol/ml (P < 0.001; Fig. 2). To examine further the impact of residual insulin secretion on control of diabetes, the patient population was divided arbitrarily into four groups according to the C-peptide level achieved after Sustacal administration (Table 3). There were no significant differences in fasting glucose, the rise in glucose after Sustacal administration, or HbAic levels among the three groups of patients with stimulated C-peptide values of 0.20 pmol/ml or less. In contrast, each of these parameters of metabolic control was significantly reduced in the group of patients with stimulated C-peptide concentrations above 0.20 pmol/ ml (P < 0.03). This group had lower glycemic indices even though they were taking significantly less insulin than any of the other patient groups (P < 0.005; Table 3). Although glycemic control was not different in the group with stimulated C-peptide levels greater than pmol/ml, patients in this group were taking less insulin than those with C-peptide values of 0.05 pmol/ ml or less (P = 0.005). Thirty-three patients enrolled in the DCCT, who had had diabetes between 1-5 yr [mean, 28 ± 11 (±SD) months] and stimulated C-peptide values between pmol/ml, had repeat C-peptide measurements 1 yr after randomization. As shown in Table 4, mean basal and stimulated C-peptide levels at entry were no different in patients randomized to experimental or standard treatment. After 1 yr, basal and stimulated values declined significantly in both groups. Experimental treatment patients had a slightly but not significantly smaller decline in stimulated C-peptide values (0.21 ± 0.03 vs ± 0.02 pmol/ml). \ II Duration of IDDM (years) FIG. 1. Mean (±SEM) basal (D) and stimulated serum C-peptide (ffl) levels in relation to duration of IDDM. Discussion The patient eligibility criteria for entry into the DCCT were designed to define a reasonably homogenous group

4 0-CELL FUNCTION IN IDDM 33 TABLE 2. Basal and stimulated serum C-peptide levels by age group and duration of disease Duration <5 yr Duration >5 yr n Basal C-peptide (pmol/ml) Stimulated C-peptide (pmol/ml) Adolescent ± ± 0.03 Adult ± ± 0.02* Total ± ± 0.02 Adolescent ± ± 0.00 Adult ± ± 0.01 c Total ± ± 0.01 Mean ± SEM. * P < 0.05 vs. adolescents with disease duration of 5 yr or less. c P < 0.03 us. adolescents with disease duration of more than 5 yr. of patients with IDDM ranging in age from yr and in disease duration from 1-15 yr. In particular, patients with clinical characteristics or coexisting conditions that might confound analysis of the effects of standard and experimental treatment (11) on the complications of IDDM were excluded. Retention of substantial insulin secretory capacity was considered a potential confounding variable, since such patients might not be truly classified as insulin dependent (15) and might respond differently to the two treatment regimens (6, 16). A standardized protocol was developed for determining basal and stimulated C-peptide levels as part of eligibility testing. Because the testing was to be performed at 21 centers, a Central Biochemistry Laboratory was established to perform all assays. Quality control procedures carried out by the laboratory ensured that no deterioration of samples occurred during shipping and storage. Since almost all patients who were eligible for the DCCT by history also underwent C-peptide testing, the large scale screening of adults and adolescents for the DCCT provided a new opportunity to examine some unresolved issues concerning C-peptide secretion in IDDM. As previously reported, many of our patients were able to secrete C-peptide and presumably insulin for several years after the onset of IDDM (1-10). The expected gradual decline in basal and stimulated C-peptide levels during the first 5 yr of diabetes was also found (5, 6). Because 25% of our patients were adolescents between the ages of yr, we were able to examine whether the rate of decline of /3-cell function is influenced by age, as was suggested in a previous study (5). Mean stimulated C-peptide values were indeed modestly reduced in adolescents within the first 5 yr of IDDM compared to values in adults. However, a similar proportion of adults and adolescents with this duration retained substantial insulin secretory capacity. On the other hand, after 5 yr there was a sharp difference in the proportion of adolescents and adults with C-peptide secretory capacity. Only three of the adolescents with a disease duration greater than 5 yr achieved a stimulated C-peptide level greater than 0.10 pmol/ml, and none had values above 0.20 pmol/ml. In contrast, 11% of adults had stimulated C- peptide values of 0.20 pmol/ml or more -after 5 yr. The adolescents thus had a much sharper decline in C-peptide than did adults, even though both groups had to meet identical clinical criteria and were drawn from the same clinic populations. It has been suggested previously that very young children lose insulin secretory capacity even more quickly (17). Our data support the importance of performing stimulation tests when assessing residual /3-cell function in IDDM. Like previous studies (18), there was a close correlation between fasting and stimulated C-peptide levels in our patients. However, 19% of our patients with very low basal C-peptide values (<0.05 pmol/ml) were able to increase their serum levels substantially when challenged with Sustacal. Even though HbAic levels in these responder patients were no different from values in nonresponders, fasting glucose concentrations on the morning of the study were lower. Thus, preceding hypoglycemia and/or higher serum insulin concentrations (from exogenous sources) may have contributed to suppression of basal C-peptide concentrations in these individuals. Alternatively, the use of a standardized mixed meal feeding may have helped identify the responders. In IDDM patients residual /?-cells that are relatively insensitive to changes in serum glucose may remain responsive to other insulin secretagogues, such as amino acids and glucagon (19). In the total study population, disease duration, HbAic, and fasting glucose correlated more closely with stimulated than basal C-peptide, suggesting that the response to a mixed meal feeding is also more clinically important and that stimulated C-peptide is a more accurate index of residual 0-cell function. The insulin secretion that usually characterizes the honeymoon or remission period of IDDM has been found to reduce glycemic lability (20, 21). The influence of lower levels of residual insulin secretion on metabolic control in well established diabetes has not been demonstrated, even though most investigations have used stimulated C-peptide values below 0.1 pmol/ml to indicate severe insulin deficiency (16, 22, 23). Several crosssectional studies examining this issue yielded conflicting results (6-9), but their value was limited by relatively small patient populations. The large sample size in this study allowed us to subdivide our patients into four

5 34 DCCT RESEARCH GROUP JCE&M«1987 Vol 65 No 1 ADOLESCENTS < N - 154) Q o Q 0.4- UJ I < I 0. J - DURATION OF IDDM (years) ADULTS (N DURATION OF IDDM (yeors) FIG. 2. Stimulated serum C-peptide levels in adolescents (A) and adults (B) in relation to duration of IDDM. Eleven percent of adults and none of the adolescents with disease duration of 5 yr or more had stimulated values greater than 0.02 pmol/ml (P < 0.001). groups on the basis of their stimulated C-peptide values. There was no difference in plasma glucose or HbAic in the three groups with stimulated values of 0.20 pmol/ml or less. On the other hand, patients who could raise their C-peptide levels to over 0.20 pmol/ml had better metabolic control; their HbAic and fasting and postprandial glucose values were all lower even though they were receiving less insulin. C-Peptide values as low as pmol/ml (approximately equal to serum insulin concentrations of /xu/ml) may also have an im- >i

6 0-CELL FUNCTION IN IDDM 35 TABLE 3. Serum glucose, HbAic, and insulin dosage in patients grouped by stimulated serum C-peptide levels n Age (yr) Duration of IDDM (months) Fasting glucose {tag/ dl) Glucose post-sustacal (mg/dl) HbA lc (%) Daily insulin dosage (U/kg) Group 1: < ± 0.4" 105 ±2 222 ±6 183 ±3 9.3 ± ± 0.02 Stimulated C-peptide level (pmol/ml) Group 2: > ± ±6 206 ± ±7 9.8 ± ± 0.04 Group 3: > ± ±6 217 ± ± ± 0.02* Group 4: > ± ±3 177±6».cd 151±5"' d ' e 8.4±0.2 6 ' d ' / 0.52 ± *' P values were determined by analysis of covariance, adjusting for differences in age and duration of disease. " Mean ± SEM. * P < us. group 1. c P < 0.05 us. group 2. " P < us. group 3. ' P < vs. group 2. f P< us. group 2. * P < us. group 1. TABLE 4. Basal and stimulated serum C-peptide and HbA lc levels before and after 1 yr of experimental (n = 18) or standard (n = 15) treatment in C-peptide secretors Basal C-peptide (pmol/ml) Experimental Standard Stimulated C-peptide (pmol/ml) Experimental Standard HbA lc (%) Experimental Standard Entry 0.14 ± ±0.01 c 0.15 ± ± ± ± ± ± ± ± ± ± 0.5 Entry vs P values Exp vs. standard (iyr) Stimulated C-peptide level at entry into study, pmol/ml. 0 Values are the mean ± SEM. pact on diabetic management, since patients in this group required less exogenous insulin to achieve the same level of control as those whose stimulated levels were 0.05 pmol/ml or less. As might be expected, the four groups also differed with respect to duration of diabetes. However, the differences in serum glucose, HbAic, and insulin doses between the groups were statistically significant even when results were adjusted using duration of diabetes as a covariate. A number of investigators have examined whether intensive insulin treatment might help preserve residual /?-cell function. Most studies of this type were done in patients with recent onset of IDDM (24, 25). In patients with long-standing disease, no improvement in glucagonstimulated C-peptide levels was found after 8 months of either continuous sc insulin infusion or standard treatment in the Kroc Collaborative Study (23). However, in that study the mean duration of IDDM was 17 yr, and all of the patients were severely insulin deficient (postglucagon C-peptide, <0.10 pmol/ml). In our study, C- peptide stimulation tests were repeated 1 yr after randomization to either experimental or standard treatment in patients with diabetes of intermediate duration (mean ± SD, 28 ± 11 months) who had appreciable insulin secretory capacity at entry (stimulated C-peptide, pmol/ml). An overall decline in basal and stimulated C-peptide levels was found after 1 yr of standard or experimental treatment. These data suggest that intensive diabetes management does not reverse the underlying destructive process affecting the j8-cell. Alternatively, residual 0-cell function may be more efficient at lower serum glucose levels (26). In conclusion, our study has reemphasized the heterogeneity of IDDM, as we found a distinct difference in the pattern of residual /3-cell secretion in adolescents and adults. We also more clearly defined the degree of endogenous insulin secretion that may help facilitate metabolic control of IDDM. Acknowledgments We would like to thank the members of the Publications and Presentations Committee of the DCCT for their careful review of our manuscript, especially its chairman, Dr. David M. Nathan. We also appreciate Sonja Kornienko's expert help in preparation of the manuscript. References 1. Ludvigsson J, Heding LG 1976 C-Peptide in children with juvenile diabetes. Diabetologia 12: Henriksen C, Faber OK, Drejer J, Binder C 1977 Prevalence of residual B-cell function in insulin-treated diabetics evaluated by the plasma C-peptide response to intravenous glucagon. Diabetologia 13: Madsbad S 1983 Prevalence of residual B-cell function and its metabolic consequences in type 1 (insulin-dependent diabetes). Diabetologia 24: Hoekstra JBL, van Rijn HJM, Erkelens DW, Thijssen JHH 1982 C-Peptide. Diabetes Care 5: Madsbad S, Faber OK, Binder C, McNair P, Christiansen C, Transbol I 1978 Prevalence of residual B-cell function in insulindependent diabetes in relation to age at onset and duration of diabetes. Diabetes 27: Madsbad S, McNair P, Faber OK, Binder C, Christiansen C, Transbol I 1980 Beta-cell function and metabolic control in insulin-treated diabetics. Acta Endocrinol (Copenh) 93: Gonen B, Goldman J, Baldwin D, Goldberg RB, Ryan WG, Blix PM, Schanzlin D, Fritz KJ, Rubenstein AH 1979 Metabolic control in diabetic patients effect of insulin-secretory reserve (measured by plasma C-peptide levels) and circulating insulin antibodies. Diabetes 28: Ludvigsson J, Heding LG, Larsen Y 1977 C-Peptide in juvenile diabetics beyond the post initial remission period: relation to clinical manifestations at onset of diabetes remission and diabetes control. Acta Paediatr Scand 66: Dahlquist G, Blom L, Bolme P, Hagenfeldt L, Lindgren F, Persson B, Thalme B, Theorell M, Westin S 1982 Metabolic control in 131

7 36 DCCT RESEARCH GROUP JCE & M 1987 Vol 65 No 1 juvenile-onset diabetic patients as measured by HbAic: relation to age, duration, C-peptide, insulin dose and one or two insulin injections. Diabetes Care 5: Eff C, Faber 0, Deckert T 1978 Persistant insulin secretion, assessed by plasma C-Peptide estimation in long-term diabetics with low insulin requirements. Diabetologia 15: The DCCT Research Group 1986 The Diabetes Control and Complications Trial (DCCT). Design and methodological considerations for the feasibility phase. Diabetes 35: Heding LG 1975 Radioimmunological determination of human C- peptide in serum. Diabetologia 11: Dunn PJ, Cole RA, Soeldner JS 1979 Further development and automation of a high-pressure liquid chromatography method for the determination of glycosylated hemoglobins. Metabolism 28: Zar J 1974 Biostatistical Analyses. Prentice-Hall, Englewood Cliffs 15. Madsbad S, Krarup T, McNair P, Christiansen C, Faber OK, Transbol I, Binder C 1981 Practical clinical value of the C-peptide response to glucagon stimulation in the choice of treatment in diabetes mellitus. Acta Med Scand 210: Rodger NW, Dupre J, Canny CLB, Brown WF 1985 Continuous subcutaneous insulin infusion in adults: glycemic advantage is predicted by venous plasma C-peptjde concentrations. Diabetes Care 8: Becker DJ, Arslanian SA, Waters T 1984 Determinants of C- peptide levels at onset and follow-up in childhood IDD. Diabetes [Suppl 1] 33:103A (Abstract) 18. Garcia-WebbP, Bonser A, Welborn TA 1982 Correlation between fasting serum C-peptide and B-Cell secretory capacity in diabetes mellitus. Diabetologia 22: Ganda OP, Srikanta S, Brink SJ, Morris MA, Gleason RE, Soeldner JS, Eisenbarth GS 1984 Differential sensitivity to B-cell secretagogues in "early" type 1 diabetes mellitus. Diabetes 33: Reynolds C, Molnar GD, Horwitz DL, Rubenstein AH, Taylor WF, Jiang NS 1977 Abnormalities of endogenous glucagon and insulin in unstable diabetes. Diabetes 26: Shima K, Tanaka R, Morishita S, Tarui S, Kumahara Y, Nishikawa M 1977 Studies on the etiology of brittle diabetes: relationship between diabetic instability and insulinogenic reserve. Diabetes 26: Faber OK Binder C 1977 C-peptide response to glucagon: a test for residual B-cell function in diabetes mellitus. Diabetes 26: Bergenstal RM, Dupre J, Lawson PM, Rubenstein AH, for the Kroc Collaborative Study Group 1985 Observations on C-peptide and free insulin in the blood during continuous subcutaneous insulin infusion and conventional insulin therapy. Diabetes [Suppl 3] 34: Mirouze J, Selam JL, Phan TC, Mendoza E, Orsett A 1984 Sustained normoglycemia in newly diagnosed type I diabetic subjects: short term effects and one year follow-up. Diabetes 33: Perlman K, Erlich RM, Filler RM, Albisser AM 1984 Sustained normoglycemia in newly diagnosed type I diabetic subjects: short term effects and one year follow-up. Diabetes 33: Unger RH, Grundy S 1985 Hyperglycemia as an inducer as well as a consequence of impaired islet cell function and insulin resistance: implications for the management of diabetes. Diabetologia 28:119