Abstract
Aims
To examine the association between islet autoantibody positivity and clinical characteristics, residual β-cell function (C-peptide) and prevalence of complications in a childhood-onset (age <17 years), long-duration (≥32 years) type 1 diabetes cohort.
Methods
Islet autoantibodies (glutamic acid decarboxylase, insulinoma-associated protein 2 and zinc transporter-8 antibodies) were measured in the serum of participants who attended the 2011–2013 Pittsburgh Epidemiology of Diabetes Complications study follow-up examination (n=177, mean age 51 years, diabetes duration 43 years).
Results
Prevalences of islet autoantibodies were: glutamic acid decarboxylase, 32%; insulinoma-associated protein 2, 22%; and zinc transporter-8, 4%. Positivity for each islet autoantibody was associated with older age at diabetes onset (glutamic acid decarboxylase antibodies, P=0.03; insulinoma-associated protein 2 antibodies, P=0.001; zinc transporter-8 antibodies, P<0.0001). Older age at onset was also associated with an increasing number of autoantibodies (P = 0.001). Glutamic acid decarboxylase antibody positivity was also associated with lower HbA1c (P = 0.02), insulinoma-associated protein 2 antibody positivity was associated with lower prevalence of severe hypoglycaemic episodes (P=0.02) and both distal and autonomic neuropathy (P=0.04 for both), and zinc transporter-8 antibody positivity was associated with higher total and LDL cholesterol (P=0.01). No association between autoantibody positivity and C-peptide was observed.
Conclusions
The strong association between islet autoantibody positivity and older age at type 1 diabetes onset supports the hypothesis of a less aggressive, and thus more persistent, immune process in those with older age at onset. This observation suggests that there may be long-term persistence of heterogeneity in the underlying autoimmune process.
Introduction
In the classic paradigm, soon after the onset of overt type 1 diabetes, complete destruction of β cells occurs, with a corresponding cessation of islet autoimmunity [1]. However, a growing body of evidence has demonstrated that many individuals retain residual β-cell function, even decades after diabetes onset [2–6]. Generally, studies have shown that ~30–50% of individuals with long-duration type 1 diabetes have detectable C-peptide levels [3–6]. Concomitant with this apparent β-cell preservation is the hypothesis that these individuals may also have continuing islet autoimmunity, including autoantibodies to glutamic acid decarboxylase (GAD)-65, insulinoma-associated antigen-2 (IA-2) and zinc transporter-8 (ZnT8). Data regarding persistence of islet autoimmunity after long diabetes durations are sparse [6,8–11], but such data have consistently shown persistence of autoimmunity in a significant proportion of individuals with long-standing type 1 diabetes.
Despite an expected relationship between autoantibodies and residual β-cell function (i.e. C-peptide), reported associations have been inconsistent [4,6,11]. Additionally, in most reports, a large proportion of each cohort had <20 years’ diabetes duration, so it is unclear whether the findings are relevant to longer-duration diabetes. There is even less information regarding the association of autoantibodies with other clinical characteristics and complications in long-duration type 1 diabetes. Higher autoantibody levels have been associated with worse glycaemic control at shorter (i.e.< 5 years) diabetes durations [12], but it is unknown whether this association continues as duration increases. A few studies have examined the association between GAD antibodies and neuropathy, with inconsistent results [7,9,12,13], but studies looking at the relationship between autoantibodies and other diabetes complications are lacking. The aim of the present exploratory study, therefore, was to examine whether autoantibody positivity, comprising GAD, IA-2 and ZnT8 antibodies, was associated with residual β-cell function, glycaemic control, other clinical characteristics and prevalence of complications in individuals who attended the 2011–2013 follow-up examination of the Pittsburgh Epidemiology of Diabetes Complications (EDC) study, a cohort which now has a minimum diabetes duration of 32 years.
Participants and methods
Study population
The EDC Study is a prospective cohort study of childhood-onset (<17 years) type 1 diabetes. All participants (n=658) were diagnosed, or seen within 1 year of diagnosis, at the Children’s Hospital of Pittsburgh between 1950 and 1980. The cohort has been described in detail elsewhere [14,15]. Briefly, participants have been followed since 1986–1988, initially with biennial examinations for the first 10 years of follow-up (until 1996–1998), and thereafter with biennial questionnaires, with further examinations at 18 years (2004–2006) and 25 years (2011–2013). In the present study we used stored serum specimens from participants who attended the 25-year EDC examination (Fig. S1). Of the 223 participants examined, 12 had undergone a pancreatic transplant and 34 did not have sufficient volume of stored serum. Thus, islet autoantibodies were measured in a total of 177 participants.
Islet autoantibodies
The islet autoantibody panel was measured at the Autoantibody/HLA Core Laboratory at the Barbara Davis Center for Childhood Diabetes (Aurora, CO), an National Institute of Health (NIH)/ National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) North America Autoantibody/HLA Core Laboratory. GAD and IA-2 antibodies were measured using a radiobinding assay [16], with the result expressed as DK units/ml with an NIDDK standard curve. ZnT8 antibodies were measured using a radiobinding assay [17], with the result expressed as an index against standard internal positive and negative controls. All individuals using exogenous insulin would be expected to be positive for insulin autoantibodies, thus we did not examine insulin autoantibodies in this cohort. Autoantibody positivity was defined using the core laboratory’s established thresholds: GAD antibodies >20 DK units/ml, IA-2 antibodies >5 DK units/ml and ZnT8 antibodies >0.02.
Demographic and clinical characteristics
Dates of birth and diabetes diagnosis were extracted from the Children’s Hospital of Pittsburgh diabetes registry. Age and diabetes duration were calculated as decimal years at the date of the 25-year examination. Age at diabetes onset was calculated as decimal years from the date of birth until diagnosis.
HbA1c was measured using the DCA 2000 analyser (Bayer Healthcare LLC. Elkhart, IN, USA) and converted to Diabetes Control and Compliations Trial (DCCT)-aligned HbA1c: DCCT HbA1c = (EDC HbA1c –1.13)/0.81 [18]. Serum C-peptide was measured using ultra-sensitive ELISA, with a detection limit ≤25 pmol/L (Mercodia AB, Uppsala, Sweden). Human leukocyte antigen-DR isotype (HLA-DR) typing was performed serologically. Total, HDL and LDL cholesterol and triglycerides were measured using the Cholestech LDX (Cholestech Corp., Hayward, CA, USA). Non-HDL cholesterol was calculated by subtracting HDL cholesterol from total cholesterol. White blood cell count and differential were determined using electronic cell sizing and counting (Quest Diagnostics, Secaucus, NJ, USA). Blood pressure was measured according to the Hypertension Detection and Follow-Up protocol [19], with an aneroid sphygmomanometer. Hypertension was defined as blood pressure ≥140/90 mmHg or use of blood pressure-lowering medication. Serum creatinine was measured enzymatically (Quest Diagnostics, Secaucus, NJ, USA). GFR was estimated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) creatinine equation [20]. Urinary albumin and creatinine were measured using the DCA Vantage Analyser (Siemens Healthcare GmbH, Erlangen, Germany) in each of three timed urine collections (24-h, overnight and 4-h), collected within a 2-week period. Height and weight were measured using standard methods to calculate BMI.
Insulin dose was the number of self-reported units of insulin administered per day/body weight (kg). Smoking status was assessed by self-report. Recent severe hypoglycaemia was defined as reporting at least one hypoglycaemic episode which resulted in unconsciousness and/or hospitalization within the preceding 2 years or a hypoglycaemic episode requiring assistance from another person in the preceding 12 months. ‘Ever’ severe hypoglycaemia was defined as at least one such event reported at any previous EDC assessment.
Complication ascertainment
Coronary artery disease comprised a history of myocardial infarction (including clinical events and subclinical myocardial infarction on ECG, i.e. Minnesota code 1.1 or 1.2), coronary revascularization procedure, blockage ≥50%, ischaemic EGC at an EDC study visit (Minnesota codes 1.3, 4.1–4.3, 5.1–5.3, 7.1), or EDC physician-diagnosed angina. Cardiovascular disease comprised coronary artery disease or stroke. Cardiovascular disease events were confirmed with medical records.
To assess proliferative diabetic retinopathy, stereoscopic colour fundus photographs of three standard fields (1, 2 and 4) were taken with a Zeiss camera (Carl Zeiss, Oberkochen, Germany) after pupil dilation. Grading of fundus photographs was performed at the Fundus Photograph Reading Centre at the University of Wisconsin-Madison, using a modification of the Early Treatment Diabetic Retinopathy Study (ETDRS) adaptation of the modified Airlie House classification of proliferative retinopathy [21]. Proliferative retinopathy was defined as an ETDRS score ≥60 or laser photocoagulation for retinopathy.
At baseline, 2-, 4-, 6-, 8-, 10- and 18-year EDC examinations, urinary albumin was measured by immunonephelometry [4]. Albumin excretion rate was calculated for three timed urine samples (24-h, overnight and 4-h collections obtained over a 2-week period. Prevalence of microalbuminuria was defined as ever having an albumin excretion rate ≥20 μg/min in two of three timed urine samples prior to the 25-year examination, urinary albumin/creatinine ratio ≥30 mg/g at the 25-year examination, or history of dialysis or renal transplantation. Overt nephropathy was defined as ever having an albumin excretion rate ≥200 μg/min in two of three timed urine samples prior to the 25-year examination, albumin/creatinine ≥300 mg/g at the 25-year examination, or history of dialysis or renal transplantation.
Confirmed distal symmetric polyneuropathy was defined as any history of experiencing an abnormal age-specific vibratory threshold using the Vibratron II Tester (Physitemp Instruments, Clifton, NJ, USA) and at least two of the following: symptoms consistent with distal symmetric polyneuropathy; sensory and/or motor signs; or absent/reduced tendon reflexes based on the DCCT clinical examination protocol [22]. Cardiac autonomic neuropathy was defined as any history of an abnormal heart rate response to deep breathing (expiration–inspiration ratio of <1.1).
Statistical methods
Characteristics at the time of autoantibody measurement were compared between those positive and negative for each autoantibody using t-tests for continuous variables (or a Wilcoxon rank-sum test if non-normally distributed). For dichotomous variables (sex, detectable C-peptide, HLA-DR3 or -DR4, severe hypoglycaemia, smoking, and complication prevalence) chi-squared tests (or Fisher’s exact test) were used to assess the association with autoantibody positivity. Characteristics were also examined according to total number of positive autoantibodies in general linear models using contrasts to test for linear trend for continuous variables or using the Cochran–Armitage test for trend for dichotomous variables. Complete case analysis was performed for variables with missing data (insulin dose missing, n=15; LDL cholesterol missing, n=21; albumin/creatinine ratio missing, n=16). To account for multiple comparisons, given the three autoantibodies being examined, we adjusted the significance level to P=0.017 (i.e. 0.05 divided by three). Pubertal diabetes onset was defined as age at diabetes onset ≥11 years in girls or ≥12 years in boys. Three age-at-onset categories were defined: <6 years old (early diabetes onset), 6–11/12 years old (later, but pre-pubertal diabetes onset), and ≥11/12 (pubertal diabetes onset).
Ethics
Research protocols were approved by the University of Pittsburgh institutional review board (approval #19040065) and all participants provided written informed consent.
Results
Overall characteristics of the EDC cohort at the 2011–2013 follow-up visit are in Table 1. The median (range) age was 50 (33–70) years, diabetes duration was 42 (32–60) years, and age at diabetes onset was 8 (<1–16) years. The mean HbA1c level was 65 mmol/mol (8.1%); 10% had detectable C-peptide. The proportion positive for each autoantibody and for each combination of autoantibodies is shown in Fig. 1. The proportion of EDC participants positive for at least one autoantibody was 41.2% (n=73). Nearly 32% were positive for GAD, 22% were positive for IA-2 and 4% were positive for ZnT8 antibodies. GAD antibodies comprised the largest proportion of those positive for a single autoantibody (19%), compared to 9% for IA-2 antibodies. None were positive for ZnT8 antibodies alone. The proportion positive for multiple autoantibodies was 14%.
Table 1.
Overall characteristics of the Epidemiology of Diabetes Complications (EDC) study cohort at the 25-year follow-up examination
| Number of participants, n | 177 |
| Age, years | |
| Mean (SD) | 51.0 (7.5) |
| Median (25th percentile, 75th percentile) | 50 (46, 56) |
| Range | 33–70 |
| Type 1 diabetes duration, years | |
| Mean (SD) | 42.7 (6.8) |
| Median (25th percentile, 75th percentile) | 42 (37, 46) |
| Range | 32–60 |
| Age at type 1 diabetes onset, years | |
| Mean (SD) | 8.3 (4.1) |
| Median (25th percentile, 75th percentile) | 8 (4, 11) |
| Range | 0.3–16 |
| Women, % (n) | 50 (89) |
| HbA1c mmol/mol | |
| Mean (SD) | 65 (16.6) |
| Range | 31–147 |
| HbA1c, % | |
| Mean (SD) | 8.1 (1.5) |
| Range | 5.0–15.6 |
| C-peptide detectable, % (n) | 10 (18) |
FIGURE 1.
Venn diagram of percent positive for combinations of islet autoantibodies at the 25-year Epidemiology of Diabetes Complications (EDC) study follow-up examination (n=177). GAD, glutamic acid decarboxylase; IA-2, insulinoma-associated protein 2; ZnT8, zinc transporter-8.
Individual islet autoantibodies and clinical characteristics
The participant characteristics by individual autoantibody positivity status are in Table 2. IA-2 antibody-positive participants had a significantly older age at diabetes onset (P=0.001) compared to IA-2 antibody-negative participants. Those positive for ZnT8 antibodies also had a significantly older age at diabetes onset (P<0.0001) and had significantly higher total (P=0.008), LDL (P=0.01) and non-HDL cholesterol (P=0.007) compared to those negative for ZnT8 antibodies.
Table 2.
Participant characteristics by islet autoantibody positivity status at the 25-year Epidemiology of Diabetes Complications (EDC) study follow-up examination
| GAD antibody-positive n=56 | GAD antibody-negative n=121 | P | IA-2 antibody-positive n=39 | IA-2 antibody-negative n=138 | P | ZnT8 antibody-positive n=7 | ZnT8 antibody-negative n=170 | P | |
|---|---|---|---|---|---|---|---|---|---|
| Age, years | 52.0 (7.3) | 50.5 (7.6) | 0.21 | 51.5 (7.7) | 50.8 (7.5) | 0.62 | 52.2 (5.9) | 50.9 (7.6) | 0.67 |
| Type 1 diabetes duration, years | 42.7 (6.6) | 42.7 (6.9) | 0.96 | 41.3 (6.8) | 43.1 (6.7) | 0.16 | 40.6 (4.8) | 42.8 (6.8) | 0.41 |
| Age at type 1 diabetes onset, years | 9.3 (3.7) | 7.8 (4.2) | 0.03 | 10.2 (3.6) | 7.8 (4.1) | 0.001 | 11.5 (1.2) | 8.2 (4.1) | <0.0001 |
| Women, % (n) | 59 (33) | 46 (56) | 0.12 | 59 (23) | 48 (66) | 0.22 | 57 (4) | 50 (85) | 0.99 |
| C-peptide detectable, % (n) | 8.9 (5) | 11 (13) | 0.71 | 10 (4) | 10 (14) | 0.99 | 0 (0) | 11 (18) | 0.99 |
| HLA-DR3 or -DR4, % (n) | 89.3 (50) | 91.7 (111) | 0.60 | 94.9 (37) | 89.9 (124) | 0.53 | 100 (7) | 90.6 (154) | 0.99 |
| HbA1c, mmol/mol | 61 (11.3) | 66 (18.3) | 0.02 | 64 (11.2) | 65 (17.8) | 0.62 | 65 (6.5) | 65 (16.9) | 0.93 |
| HbA1c, % | 7.7 (1.0) | 8.2 (1.7) | 8.0 (1.0) | 8.1 (1.6) | 8.1 (0.6) | 8.1 (1.5) | |||
| Insulin dose, units/kg body weight* | 0.56 (0.22) | 0.60 (0.32) | 0.39 | 0.60 (0.34) | 0.58 (0.28) | 0.70 | 0.49 (0.17) | 0.59 (0.29) | 0.38 |
| Severe hypoglycaemia: ever, % (n) | 91 (51) | 92 (111) | 0.99 | 82 (32) | 94 (130) | 0.02 | 71 (5) | 92 (157) | 0.11 |
| Severe hypoglycaemia: past 2 years, % (n) | 69 (34) | 67 (72) | 0.85 | 61 (20) | 69 (86) | 0.34 | 60 (3) | 68 (103) | 0.66 |
| BMI, kg/m2 | 27.4 (4.4) | 28.9 (5.5) | 0.06 | 28.8 (5.3) | 28.3 (5.2) | 0.62 | 30.2 (6.1) | 28.3 (5.2) | 0.35 |
| Current smoking, % (n) | 13 (7) | 7.6 (9) | 0.26 | 18 (7) | 6.7 (9) | 0.05 | 14 (1) | 9 (15) | 0.50 |
| Smoking history, % (n) | |||||||||
| Never | 63 (34) | 64 (76) | 0.53 | 61 (23) | 65 (87) | 0.18 | 86 (6) | 63 (104) | 0.49 |
| Past | 24 (13) | 28 (33) | (trend) | 21 (8) | 28 (38) | (trend) | 0 (0) | 28 (46) | (trend) |
| Current | 13 (7) | 7.6 (9) | 18 (7) | 6.7 (9) | 14 (1) | 9 (15) | |||
| White blood cell count, ×109/l | 6.3 (2.2) | 6.6 (1.8) | 0.48 | 6.4 (2.5) | 6.5 (1.8) | 0.64 | 5.9 (1.4) | 6.5 (2.0) | 0.39 |
| Median (25th percentile, 75th percentile) albumin/creatinine ratio† | 9.7 (6.3, 45.5) | 11.6 (7.3, 41.0) | 0.31 | 11.5 (6.0, 41.0) | 10.7 (6.9, 43.9) | 0.80 | 28.7 (11.2, 150.3) | 10.6 (6.6, 40.6) | 0.24 |
| eGFR, mL/min/1.73m2 | 78.3 (23.3) | 81.0 (20.3) | 0.44 | 84.4 (21.4) | 78.9 (21.1) | 0.15 | 75.8 (27.7) | 80.3 (21.1) | 0.58 |
| Systolic blood pressure, mmHg | 113.0 (14.9) | 117.7 (17.0) | 0.08 | 115.2 (15.5) | 116.5 (16.8) | 0.64 | 112.9 (21.2) | 116.4 (16.3) | 0.58 |
| Diastolic blood pressure, mmHg | 64.7 (9.6) | 66.1 (9.0) | 0.37 | 66.6 (10.8) | 65.4 (8.8) | 0.54 | 67.9 (11.3) | 65.5 (9.2) | 0.52 |
| Total cholesterol, mmol/l | 4.58 (0.86) | 4.72 (1.02) | 0.37 | 4.85 (1.01) | 4.63 (0.96) | 0.21 | 5.63 (1.19) | 4.64 (0.95) | 0.008 |
| LDL cholesterol‡, mmol/l | 2.50 (0.70) | 2.72 (0.82) | 0.11 | 2.77 (0.79) | 2.62 (0.79) | 0.30 | 3.51 (0.91) | 2.63 (0.77) | 0.01 |
| Non-HDL cholesterol, mmol/l | 2.98 (0.80) | 3.14 (1.01) | 0.30 | 3.34 (0.98) | 3.01 (0.93) | 0.06 | 4.04 (1.31) | 3.05 (0.92) | 0.007 |
| HDL cholesterol, mmol/l | 1.60 (0.52) | 1.59 (0.50) | 0.82 | 1.51 (0.48) | 1.61 (0.52) | 0.28 | 1.59 (0.54) | 1.59 (0.51) | 0.99 |
| Median (25th percentile, 75th percentile) triglycerides, mmol/l | 0.78 (0.6, 0.9) | 0.81 (0.5, 1.3) | 0.30 | 0.92 (0.6, 1.5) | 0.79 (0.5, 1.2) | 0.11 | 1.22 (0.5, 2.8) | 0.80 (0.5, 1.2) | 0.21 |
GAD, glutamic acid decarboxylase; HLA-DR, human leukocyte antigen-DR isotype; IA-2, insulinoma-associated protein 2; ZnT8, zinc transporter-8.
Values are mean (SD), unless otherwise stated.
Missing n=15
missing n=16
missing n=21.
Characteristics where P<0.05 are presented in bold.
Number of positive islet autoantibodies and clinical characteristics
Associations between clinical characteristics and number of autoantibodies are shown in Table 3. There was a significant trend towards an association between greater number of positive autoantibodies and older age at onset (P=0.0004). The association between greater number of autoantibodies and age at onset was reinforced when examining categories of age at diabetes onset (Fig. 2). The group with an early age at onset (<6 years old) included the greatest proportion with no positive autoantibodies and none was positive for all three autoantibodies. Conversely, those with a pubertal age at diabetes onset (11–16 years for girls, 12–16 years for boys) had the smallest proportion with no positive autoantibodies and the greatest proportion positive for two or three autoantibodies.
Table 3.
Participant characteristics by number of positive islet autoantibodies at the 25-year Epidemiology of Diabetes Complications (EDC) study follow-up examination
| 0 autoantibody (n=104) | 1 autoantibody (n=48) | ≥2 autoantibody (n=25) | P for trend | |
|---|---|---|---|---|
| Age, years | 50.6 (7.6) | 51.0 (7.1) | 52.6 (7.7) | 0.24 |
| Type 1 diabetes duration, years | 43.1 (6.8) | 42.3 (6.9) | 41.8 (6.6) | 0.42 |
| Age at type 1 diabetes onset, years | 7.6 (4.2) | 8.7 (3.8) | 10.7 (3.2) | 0.0004 |
| Women, % (n) | 46 (48) | 52 (25) | 64.0 (16) | 0.11 |
| C-peptide detectable, % (n) | 11 (11) | 10 (5) | 8.0 (2) | 0.74 |
| HLA-DR3 or -DR4, % (n) | 91 (95) | 88 (42) | 96.0 (24) | 0.76 |
| HbA1c, mmol/mol | 66 (19.2) | 64 (12.9) | 61 (9.3) | 0.18 |
| HbA1c, % | 8.2 (1.8) | 8.0 (1.2) | 7.7 (0.85) | |
| Insulin dose*, units/kg | 0.59 (0.29) | 0.57 (0.32) | 0.58 (0.23) | 0.83 |
| Severe hypoglycaemia: ever, % (n) | 95 (99) | 85 (41) | 88 (22) | 0.08 |
| Severe hypoglycaemia, past 2 years, % (n) | 69 (64) | 66 (29) | 65 (68) | 0.68 |
| BMI, kg/m2 | 28.6 (5.5) | 28.1 (4.6) | 28.2 (5.2) | 0.71 |
| Smoking: current, % (n) | 6.9 (7) | 8.5 (4) | 21 (5) | 0.06 |
| White blood cell count ×109/l | 6.67 (1.83) | 6.15 (1.70) | 6.47 (2.76) | 0.65 |
| Median (25th percentile, 75th percentile) albumin/creatinine ratio † | 12.6 (7.3, 43.9) | 9.1 (6.4, 19.7) | 13.2 (5.9, 45.5) | 0.85 |
| eGFR, mL/min/1.73m2 | 79.8 (20.8) | 80.2 (21.3) | 81.5 (24.0) | 0.72 |
| Systolic blood pressure, mmHg | 117.4 (17.0) | 115.7 (16.6) | 112.2 (13.4) | 0.15 |
| Diastolic blood pressure, mmHg | 65.6 (8.7) | 65.6 (9.7) | 65.8 (10.8) | 0.92 |
| Total cholesterol, mmol/l | 4.64 (0.98) | 4.74 (0.93) | 4.72 (1.0) | 0.67 |
| LDL cholesterol‡, mmol/l | 2.67 (0.82) | 2.60 (0.69) | 2.68 (0.88) | 0.88 |
| Non-HDL cholesterol, mmol/l | 3.06 (0.99) | 3.06 (0.80) | 3.24 (1.05) | 0.40 |
| HDL cholesterol, mmol/l | 1.58 (0.49) | 1.67 (0.57) | 1.48 (0.43) | 0.44 |
| Median (25th percentile, 75th percentile) triglycerides | 0.80 (0.6, 1.3) | 0.78 (0.6, 0.9) | 0.85 (0.6, 1.1) | 0.48 |
HLA-DR, human leukocyte antigen-DR isotype.
Values are mean (SD), unless otherwise stated.
Missing n=15
missing n=16
missing n=21.
Characteristics where P<0.05 are presented in bold.
FIGURE 2.
Proportion of age at type 1 diabetes onset category positive for zero, one, two and three islet autoantibodies at the 25-year Epidemiology of Diabetes Complications (EDC) study follow-up examination (P=0.001 for difference in trend between <6 years and 11/12–16 years age categories). M, male; F, female.
Islet autoantibodies and complication prevalence
Figure 3 shows the prevalence of each complication by individual autoantibody status. There were no significant differences in prevalence of any complication by GAD or ZnT8 antibody status. Positivity for IA-2 antibodies was marginally associated with lower prevalence of both polyneuropathy (P=0.04) and cardiac autonomic neuropathy (P=0.04). Prevalence of complications by number of positive autoantibodies is shown in Table S1.
FIGURE 3.
Prevalence of complications by islet autoantibody positivity status at the 25-year Epidemiology of Diabetes Complications (EDC) study follow-up examination. Ab+, antibody-positive; Ab-, antibody-negative.
Discussion
In this exploratory study examining islet autoantibodies in a cohort with 32–60 years’ type 1 diabetes duration, we examined the clinical characteristics and complication prevalence associated with autoantibody positivity. Positivity for each autoantibody and total number of positive autoantibodies were associated with older age at diabetes onset. Additionally, ZnT8 antibody positivity was associated with higher LDL cholesterol. There were no significant associations between islet autoantibody positivity and diabetes duration, detectable C-peptide status, HLA-DR, or sex. The comparability of the autoantibody frequencies observed in the present study with those in prior reports varied. A summary of the results of published studies examining persistence of islet autoantibodies in type 1 diabetes of >20 years’ duration is shown in Table S2.
The most striking finding of the present study is the consistent association between positivity for any of the three individual autoantibodies and older age at type 1 diabetes onset. This association between older age at onset and presence of islet autoantibodies at long diabetes durations has also been observed in other studies [6,10]; however, in the present study, we demonstrate that, even within the relatively narrow age-at-onset range (0–16 years) in this cohort that exclusively had childhood-onset type 1 diabetes, this association between autoantibody positivity and older age at onset was still observed. We additionally observed that older age at onset was associated with a greater number of positive autoantibodies. Notably, 25% of those with a pubertal age at onset were positive for ≥2 autoantibodies, compared with only 5% of those with an age at onset of <6 years. While in the prediabetes period, multiple autoantibodies are associated with a more aggressive immune response, in the present study we were examining long-term persistence of autoantibodies at least three decades after diabetes onset, which probably indicates a more moderate, but persistent, immune process over time. It has previously been hypothesized that older age at type 1 diabetes onset may be characterized by a less aggressive autoimmune process [23–25]. Indeed, in a recent report examining insulitis, or infiltration of leukocytes into pancreatic islets, a dramatically higher average CD20+ : CD4+ ratio, representing a more aggressive autoimmune phenotype, was observed in the islets of people with type 1 diabetes onset at age <7 years compared to those with onset at age >13 years [25]. Our observation of lower islet autoantibody prevalence in a group with similar age at diabetes onset (<6 years), compared to older age at onset, is consistent with those findings, and provides evidence that there may be long-term persistence of heterogeneity in the autoimmune process underlying type 1 diabetes.
The lack of associations between islet autoantibodies and both C-peptide [3,9,26] and HLA-DR status [9,10] has also been reported in other studies and contradicts the concept of a milder autoimmune process being associated with greater β-cell function. The persistence of islet cell autoimmunity suggests that at least some β cells continue to function, so this lack of association between autoantibody positivity and C-peptide is somewhat surprising. One obstacle to elucidating this relationship is that it is unclear whether functioning β cells in long-duration diabetes were preserved from before diabetes onset and were somehow protected from autoimmune destruction, or whether β cells are regenerated or newly formed through other mechanisms (e.g. proliferation or transdifferentiation) [27]. If regeneration or proliferation of β cells is a dynamic process that changes over time, then the association between C-peptide and islet autoantibodies may be difficult to assess cross-sectionally. Additionally, there is recent evidence that examining C-peptide alone may underestimate β-cell function [28]. In a recent report, nearly 90% of individuals with undetectable C-peptide had detectable levels of serum pro-insulin [28]. The authors of that report suggest that this differential reflects a ‘potential hierarchy of β-cell dysfunction’, where C-peptide secretion declines before pro-insulin production, as β-cell destruction progresses. Thus, future research on the association between autoantibody persistence and β-cell function should incorporate pro-insulin as another measure of β-cell function in long-duration type 1 diabetes.
The small number of participants positive for ZnT8 antibodies limited the study’s power for detecting associations. Despite this limitation, we observed large differences in age at onset and LDL and non-HDL cholesterol by ZnT8 antibody positivity status. The association between ZnT8 antibodies and higher LDL cholesterol has previously been reported in latent autoimmune diabetes in adults [29], but we are unaware of any studies reporting this association in type 1 diabetes. These intriguing findings warrant further study.
Other findings of note include the association between IA-2 antibody positivity and lower prevalence of both polyneuropathy and cardiac autonomic neuropathy that did not reach statistical significance (P=0.04 for both). The relationship between islet autoantibody and polyneuropathy has been examined in several studies, with varying results [8,12,13,30]. A recent study by Louraki et al. [30] found both GAD antibodies and IA-2 antibodies to be positively associated with the development of axonal degeneration in children and adolescents with type 1 diabetes. The mean diabetes duration in that study, however, was only 5.5 years, so this association may not be apparent at longer durations. In another study of recent-onset type 1 diabetes, GAD antibodies, but not IA-2 antibodies, were positively associated with peripheral neuropathy [13]. Other studies have found no association between GAD antibodies and neuropathy, but did not examine IA-2 antibodies [8,9,12].
The major strength of the EDC study is that it is a well-characterized cohort that facilitates examination of many clinical characteristics and complications. We have also measured islet autoantibodies at an NIH/NIDDK North America Autoantibody/HLA Core Laboratory using standardized assays. One limitation is the possibility of a healthy survivor effect. As we were assessing autoantibody prevalence in long-duration type 1 diabetes in the present study, we examined a survivor cohort by definition. The subset examined was slightly younger at study baseline (1986–1988; median age 25 years vs 27 years in the entire cohort), with shorter diabetes duration (median 16 years vs 18 years in the entire cohort). These small differences are not likely to have influenced our findings. Importantly, age at diabetes onset did not differ. If persistence of autoantibodies is strongly associated with mortality or morbidity, then the highest-risk individuals may not be represented and our ability to detect associations between autoantibodies and complications could be limited. However, it is important to note that there was no association between age at onset and mortality in the EDC cohort, so there is no evidence of selection bias by age at onset affecting the associations described in the present study. A further limitation is the cross-sectional design, a limitation shared with the other reports on islet autoantibodies in long-duration type 1 diabetes. The aforementioned limitations underscore the need for longitudinal studies incorporating repeated measures of autoantibodies over time. Additionally, associations that have not been examined in other reports should be considered as hypothesis-generating and require confirmation in other cohorts. Another limitation is that the EDC cohort is 98% white, so we were unable to examine differences by race. Finally, pubertal status was defined using age (no Tanner staging or hormone data were available).
In conclusion, the strong association between islet autoantibody positivity and age at type 1 diabetes onset observed in the present study provides additional evidence that heterogeneity in the autoimmune process underlying type 1 diabetes may persist for decades after onset. Given the potential that islet autoimmunity may be a dynamic process, longitudinal analyses are needed to examine associations between islet autoantibodies and concomitant clinical characteristics, as well as future risk of complications.
Supplementary Material
Figure S1. EDC islet antibody study participant flow chart.
Table S1. Complication prevalence by number of positive islet autoantibodies at the 25-year EDC study follow-up examination.
Table S2. Results of published studies examining persistence of islet autoantibodies in long-duration (>20 years) type 1 diabetes, sorted by descending mean/median diabetes duration.
What’s new?
In the people with childhood-onset (<17 years old) type 1 diabetes for >30 years in this study cohort, there was an association between older age at diabetes onset and current positivity for islet autoantibodies, as well as increasing total number of islet autoantibodies.
These findings support the hypothesis of a less aggressive, but more persistent immune process in those with older age at type 1 diabetes onset.
If there is long-term persistence of heterogeneity in the autoimmune process underlying type 1 diabetes, immune intervention therapies may have potential to benefit individuals with long duration of diabetes.
Acknowledgements
We thank the EDC study participants and staff for their invaluable contributions.
Funding sources
The EDC study is supported by NIH/NIDDK (Grant No. R01-DK034818) and the Rossi Memorial Fund. Funding for the islet autoantibody study was provided by the University of Pittsburgh Central Research Development Fund. R.G.M. is supported by American Diabetes Association Grant number 1-19-JDF-109.
Footnotes
Competing interests
T.J.O. is a consultant for Boehringer Ingelheim International GmbH. R.G.M., L.Y., D.J.B. and T.C. have no conflicts of interest.
Supporting Information
Additional Supporting Information may be found in the online version of this article:
References
- 1.Hagopian W, Lernmark Å. Autoimmune Diabetes Mellitus In: Rose N, Mackay I, ed. The Autoimmune Diseases II. New York: Academic Press, 1992. pp. 235–278. [Google Scholar]
- 2.Meier JJ, Bhushan A, Butler AE, Rizza RA, Butler PC. Sustained beta cell apoptosis in patients with long-standing type 1 diabetes: Indirect evidence for islet regeneration? Diabetologia 2005;48:2221–2228. [DOI] [PubMed] [Google Scholar]
- 3.Keenan HA, Sun JK, Levine J, Doria A, Aiello LP, Eisenbarth G et al. Residual insulin production and pancreatic β-cell turnover after 50 years of diabetes: Joslin medalist study. Diabetes 2010; 59: 2846–2853. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Wang L, Lovejoy NF, Faustman DL. Persistence of prolonged C-peptide production in type 1 diabetes as measured with an ultrasensitive C-peptide assay. Diabetes Care 2012;35:465–470. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Oram RA, Jones AG, Besser REJ, Knight BA, Shields BM, Brown RJ et al. The majority of patients with long-duration type 1 diabetes are insulin microsecretors and have functioning beta cells. Diabetologia 2014;57:187–191. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Williams GM, Long AE, Wilson IV, Aitken RJ, Wyatt RC, McDonald TJ et al. Beta cell function and ongoing autoimmunity in long-standing, childhood onset type 1 diabetes. Diabetologia 2016; 59: 2722–2726. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Zanone MM, Petersen JS, Peakman M, Mathias CJ, Watkins PJ, Dyrberg T et al. High prevalence of autoantibodies to glutamic acid decarboxylase in long- standing IDDM is not a marker of symptomatic autonomic neuropathy. Diabetes 1994;43:1146–1151. [DOI] [PubMed] [Google Scholar]
- 8.Roll U, Nuber A, Schröder A, Gerlach E, Janka HU, Ziegler AG. No association of antibodies to glutamic acid decarboxylase and diabetic complications in patients with IDDM. Diabetes Care 1995;18:210–215. [DOI] [PubMed] [Google Scholar]
- 9.Jaeger C, Allendörfer J, Hatziagelaki E, Dyrberg T, Bergis K, Federlin K et al. Persistent GAD 65 Antibodies in Longstanding IDDM are not Associated with Residual Beta-Cell Function, Neuropathy or HLA-DR Status. Horm Metab Res 1997; 29:510–515. [DOI] [PubMed] [Google Scholar]
- 10.Richardson CC, Dromey JA, McLaughlin KA, Morgan D, Jonathan Bodansky H, Feltbower RG et al. High frequency of autoantibodies in patients with long duration type 1 diabetes. Diabetologia 2013;56:2538–2540. [DOI] [PubMed] [Google Scholar]
- 11.Panero F, Novelli G, Zucco C, Fornengo P, Perotto M, Segre O et al. Fasting plasma C-peptide and micro- and macrovascular complications in a large clinic-based cohort of type 1 diabetic patients. Diabetes Care 2009; 32:301–305. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hoeldtke RD, Hampe CS, Bekris LM, Hobbs G, Bryner KD, Lernmark A. Antibodies to GAD65 and peripheral nerve function in the DCCT. J Neuroimmunol 2007;185(1–2):182–189. [DOI] [PubMed] [Google Scholar]
- 13.Hoeldtke RD, Bryner KD, Hobbs GR, Horvath GG, Riggs JE, Christie I et al. Antibodies to glutamic acid decarboxylase and peripheral nerve function in type 1 diabetes. J Clin Endocrinol Metab 2000; 85: 3297–308. [DOI] [PubMed] [Google Scholar]
- 14.Orchard TJ, Dorman JS, Maser RE, Becker DJ, Ellis D, LaPorte RE et al. Factors associated with avoidance of severe complications after 25 yr of IDDM. Pittsburgh Epidemiology of Diabetes Complications Study I. Diabetes Care 1990;13:741–747. [DOI] [PubMed] [Google Scholar]
- 15.Orchard TJ, Dorman JS, Maser RE, Becker DJ, Drash AL, Ellis D et al. Prevalence of complications in IDDM by sex and duration. Pittsburgh Epidemiology of Diabetes Complications Study II. Diabetes 1990; 39:1116–1124. [DOI] [PubMed] [Google Scholar]
- 16.Bonifacio E, Yu L, Williams AK, Eisenbarth GS, Bingley PJ, Marcovina SM et al. Harmonization of Glutamic Acid Decarboxylase and Islet Antigen-2 Autoantibody Assays for National Institute of Diabetes and Digestive and Kidney Diseases Consortia. J Clin Endocrinol Metab 2010; 95:3360–3367. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Wenzlau JM, Juhl K, Yu L, Moua O, Sarkar SA, Gottlieb P et al. The cation efflux transporter ZnT8 (Slc30A8) is a major autoantigen in human type 1 diabetes. Proc Natl Acad Sci U S A 2007;104:17040–17045. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Prince CT, Becker DJ, Costacou T, Miller RG, Orchard TJ. Changes in glycaemic control and risk of coronary artery disease in type 1 diabetes mellitus: findings from the Pittsburgh Epidemiology of Diabetes Complications Study (EDC). Diabetologia 2007; 50: 2280–2288. [DOI] [PubMed] [Google Scholar]
- 19.The hypertension detection and follow-up program. Hypertension detetection and follow-up program cooperative group. Prev Med (Baltim) 1976; 5:207–215. [DOI] [PubMed] [Google Scholar]
- 20.Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF, Feldman HI et al. A new equation to estimate glomerular filtration rate. Ann Intern Med 2009;150:604–612. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Grading diabetic retinopathy from stereoscopic color fundus photographs–an extension of the modified Airlie House classification. ETDRS report number 10. Early Treatment Diabetic Retinopathy Study Research Group. Ophthalmology 1991; 98(5 Suppl):786–806. [PubMed] [Google Scholar]
- 22.Feldman EL, Stevens MJ, Thomas PK, Brown MB, Canal N, Greene DA. A practical two-step quantitative clinical and electrophysiological assessment for the diagnosis and staging of diabetic neuropathy. Diabetes Care 1994;17:1281–1289. [DOI] [PubMed] [Google Scholar]
- 23.Pipeleers D, Ling Z. Pancreatic beta cells in insulin-dependent diabetes. Diabetes Metab Rev 1992; 8:209–227. [DOI] [PubMed] [Google Scholar]
- 24.Lernmark A, Klöppel G, Stenger D, Vathanaprida C, Fält K, Landin-Olsson M et al. Heterogeneity of islet pathology in two infants with recent onset diabetes mellitus. Virchows Arch 1995;425:631–640. [DOI] [PubMed] [Google Scholar]
- 25.Leete P, Willcox A, Krogvold L, Dahl-Jørgensen K, Foulis AK, Richardson SJ et al. Differential Insulitic Profiles Determine the Extent of β-Cell Destruction and the Age at Onset of Type 1 Diabetes. Diabetes 2016; 65:1362–1369. [DOI] [PubMed] [Google Scholar]
- 26.Wenzlau JM, Walter M, Gardner TJ, Frisch LM, Yu L, Eisenbarth GS et al. Kinetics of the Post-Onset Decline in Zinc Transporter 8 Autoantibodies in Type 1 Diabetic Human Subjects. J Clin Endocrinol Metab 2010; 95:4712–4719. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Oram RA, Sims EK, Evans-Molina C. Beta cells in type 1 diabetes: mass and function; sleeping or dead? Diabetologia 2019;62:567–577. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Sims EK, Bahnson HT, Nyalwidhe J, Haataja L, Davis AK, Speake C et al. Proinsulin Secretion Is a Persistent Feature of Type 1 Diabetes. Diabetes Care 2019;42:258–264. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Huang G, Xiang Y, Pan L, Li X, Luo S, Zhou Z. Zinc transporter 8 autoantibody (ZnT8A) could help differentiate latent autoimmune diabetes in adults (LADA) from phenotypic type 2 diabetes mellitus. Diabetes Metab Res Rev 2013; 29:363–368. [DOI] [PubMed] [Google Scholar]
- 30.Louraki M, Katsalouli M, Kanaka-Gantenbein C, Kafassi N, Critselis E, Kallinikou D et al. The prevalence of early subclinical somatic neuropathy in children and adolescents with Type 1 diabetes mellitus and its association with the persistence of autoantibodies to glutamic acid decarboxylase (GAD) and islet antigen-2 (IA-2). Diabetes Res Clin Pract 2016;117:82–90. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1. EDC islet antibody study participant flow chart.
Table S1. Complication prevalence by number of positive islet autoantibodies at the 25-year EDC study follow-up examination.
Table S2. Results of published studies examining persistence of islet autoantibodies in long-duration (>20 years) type 1 diabetes, sorted by descending mean/median diabetes duration.