Diabetologia (1985) 28:264-268 Diabetologia Springer-Verlag 1985 C-peptide measurement in the differentiation of Type 1 (insulin-dependent) and Type 2 (non-insulin-dependent) diabetes mellitus H. L. Katzeff 1, P.J. Savage 2, B. Barclay-White, M. Nagulesparan, and R H. Bennett Epidemiology and Field Studies Branch and Phoenix Clinical Research Section, National Institute of Arthritis, Diabetes and Digestive and Kidney Disease, Phoenix, Arizona, USA Summary. To determine whether individual subjects with and there was no overlap in the individual values between the Type 1 (insulin-dependent) diabetes or Type 2 (non-insulin- groups. The Pima Indians with early onset diabetes have been dependent) diabetes, who are treated with insulin, could be re- previously shown to have Type 2 diabetes, and the Caucasoids liably distinguished, C-peptide concentrations and urinary C- with an early onset are most likely to have Type I diabetes. peptide excretion were measured in 10 Caucasoids and 10 Pi- These results suggest that distinction between these two major ma Indians. All the subjects had developed diabetes before types of diabetes can be made effectively by using C-peptide 21 years of age and were receiving insulin treatment. Fasting measurements provided that overt renal disease is absent. This C-peptide concentrations were significantly higher in the Pi- differentiation between insulin-treated patients will be useful ma Indians (0.73 +0.17 versus 0.02+0.01nmol/1 in Cauca- for a variety of research applications and possibly in making soids; p < 0.001), but there were slight overlaps in individual clinical management decisions. values. Urinary C-peptide excretion, an index of 24-h-insulin excretion, was also higher in the Pima Indian group (27.6_ Keywords: Type 1 diabetes, Type2 diabetes, plasma C-pep- 1.85 versus 0.72+ 0.18 pmol/min in Caucasoids; p <0.001) tide, urinary C-peptide Pima Indian. Heterogeneity in diabetes mellitus is well recognized [1, Subjects and methods 2] as is reflected in the recent classification of the Na- tional Diabetes Data Group and World Health Organi- Subjects zation [3, 4]. Many patients with Type 2 diabetes are treated with insulin to control hyperglycaemia, and al- Ten Caucasoids and ten Pima Indians with the onset of diabetes be- though they are not prone to spontaneous ketosis, ke- fore age 21 years were admitted to the Clinical Research Section for investigation. The study was approved by the National Institutes of toacidosis can occur in Type 2 diabetes with severe in- Health Clinical Research Committee, and each subject gave written tercurrent illness or other stress. Conversely, many pat- consent to participate in the study. ients with well controlled Type 1 diabetes, who receive The subjects were teenagers or young adults, who had been treat- insulin therapy regularly, may never experience an ep- ed with insulin. They included a group of Caucasoids, who were as- sumed almost certainly to have Type 1 diabetes, and Pima Indians isode of ketoacidosis. Thus, in some insulin-treated pat- who, on the basis of previous evidence, were believed to have Type 2 ients, unless insulin is withdrawn, it may not be possible diabetes [5]. The clinical characteristics of the individual subjects are to establish the type of diabetes by clinical history given in Table 1. The Pima participants were either full-blooded Pima alone. This is particularly true in adolescents and young Indians or a combination of Pima/Papago. The Papago tribe is close- adults who receive insulin treatment from the time of ly related to the Pima Indians and also has a high prevalence of Type 2 diabetes. diagnosis of diabetes, as well as in older patients who may have received insulin treatment for many years. The present study was undertaken to determine Study design whether it was possible to discriminate between Types 1 and 2 diabetes among subjects who had insulin treat- After a 12-h overnight fast, subjects were given their usual insulin ment from the time of diagnosis of diabetes by examin- dose and 30 min later received a 100 g carbohydrate load (Koladex- Custom Laboratories, Baltimore, Maryland). Venous blood, for deter- ing plasma C-peptide concentrations, fasting and in re- mination of plasma glucose and C-peptide concentrations, was drawn sponse to glucose and arginine, and 24 h-urinary C-pep- at -30, 0, 30, 60, 90, 120, and 180 min. On day 2 insulin was withheld tide excretion. and after an overnight fast, arginine monohydrochloride (5.0 rag. kg-1. min-a) was infused intravenously for 40 rain. Venous blood for Present addresses: measurement of glucose and C-peptide concentrations was with- 1 Department of Medicine, Cornell University Medical College, drawn from the contralateral forearm at- 20, - 10, 0, 5,10, 20, 30, and S-807 515 East 71 Street, New York, New York 10021, USA 40 rain. A 24-h urine collection for measurement of C-peptide, creati- 2 Department of Internal Medicine, Wayne State University, VA nine and protein was obtained. Blood for HLA typing and islet cell Medical Center. Allen Park. Michigan. USA antibodies was also collected. H. A. Katzeff et al.: C-Peptide in Type 1 and Type 2 Diabetes 265 Table 1. Individual characteristics and C-peptide results of subjects studied Sex Age at Duration Body mass History a HLA-B8 Islet Basal C-peptide examination of diabetes Index of keto- antigen cell blood Basal Urinary (years) (years) (kg/m 2) acidosis antibodies glucose plasma excretion (mmol/l) (nmol/1) (pmol/min) Caucasoids M 19 5 20.8 + + + 3.1 0.13 4.5 M 20 6 21.1 - + - 11.6 0.03 1.6 M 23 8 22.9 - + + 4.7 0.02 0.0 F 24 10 20.9 + - NT 5.8 0.00 0.3 M 24 11 23.0 + + + 3.6 0.00 0.0 F 30 14 18.0 - - + 10.4 0.00 0.0 F 29 16 21.5 - - NT 3.1 Renal failure F 22 18 21.5 - + NT 13.9 0.00 0.0 F 27 18 21.3 - + + 5.0 0.00 0.1 F 29 25 20.3 + + NT 9.2 0.00 0.0 Mean • SEM 24.2+0.4 13.1+0.7 20.9• 7.0• 0.02• 0.7• Pima Indians F 17 0.5 25.2 - - NT 10.5 0.54 30.2 M 14 2 28.3 + - NT 14.5 1.74 56.2 M 23 4 26.2 - - - 14.1 0.16 28.8 F 25 5 27.0 + - - 5.8 1.29 39.0 M 27 6 23.5 + - - 12.9 0.77 20.2 F 19 7 25.1 - - - 13.3 0.31 11.3 F 28 7 30.9 - - - 13.3 0.88 40.1 M 37 20 21.1 + - - 9.3 0.44 8.2 F 42 23 29.6 - - - 14.3 0.46 14.5 F 37 27 30.8 - - - 6.0 Renal failure Mean + SEM 27.9+1.0 10.2+1.0 26.3-4-0.9 ~ c b 11.4+0.8 ~ 0.73 • 0.17 d 27.6 +_ 1.85 d + = present; NT = Not tested; a with blood pH ~< 7.2 or serum bicarbonate ~< 12 mmol/1; b p < 0.05 ; c p < 0.01 ; d p < 0.001 1,8 ~ 54 Methods x_ c 1,5 ~ 45 Samples for plasma glucose determinations were collected in tubes o containing sodium fluoride, stored at- 20 ~ and were analyzed in ~ 1.2 c 36 duplicate using the glucose oxidase method on a glucose analyzer ~~ (Beckman Instruments, Anaheim, California). Plasma for C-peptide ~o 0.9 ~ 2z determinations was stored at- 20 ~ and measured using a non-equi- o-E librium ethanol precipitation radioimmunoassay [6]. The sensitivity of 0.6 this assay is 0.02 pmol/ml with an intra- and interassay variation of o | 5.8% and 9.6%, respectively. Urine for C-peptide measurements was 0,3 c1~ 9 centrifuged at 2500 g for 30 min to remove particulate Matter and the | c c.o "7-- pH adjusted to 7.0 with NaOH before analysis [7]. HLA typing was D Z o,o -3 0,0 performed using standard microlymphocytotoxicity techniques for A Caucasoid Pima Caucasoid Pima Subjects Indians Subjects Indians and B loci antigens only [8]. Islet cell cytoplasmic antibodies were a b measured by the method of Bottazzo et al. [9]. Fig. 1l. A and B Individual C-peptide levels in the nine Caucasoid and nine Pima Indian subjects. The horizontal bars represent the mean Statistical methods value of each group. A Fasting plasma concentrations; p<0.01. B Urinary excretion rates; p < 0.005 The fasting C-peptide concentrations were taken as the mean of 6G the - 30 and 0 rain samples for each subject. The mean fasting plasma and urinary concentrations of C-pepfide were compared using Stu- c dent's t-test A Fisher exact test for 2 x 2 tables was utilized to analyze -~ 50 the HLA and islet cell antibody data and Pearson correlation coeffi- cients were used to examine relationships between C-peptide concen- ~ 40 trations, excretion rates, obesity and duration of disease within each group of subjects [10]. Analysis of variance with repeated measures ~ 30 was used to test for differences from baseline in responses to oral glu- cose tolerance and arginine infusion tests [10]. ,3 2o o c Results o The age of onset of diabetes was somewhat lower and o o o9o----o, •, 9-- 5 10 15 20 25 the duration of diabetes slightly longer in the Cauca- Duration of diobetes (years) soids compared with the Pima Indians.