DIABETICMedicine DOI: 10.1111/dme.12615

Research: Pregnancy Glutamic acid decarboxylase autoantibody-positivity post-partum is associated with impaired b–cell function in women with gestational diabetes mellitus T. P. Lundberg1,2, K. Højlund1,2, L. S. Snogdal1,2 and D. M. Jensen1,2 1

Department of Endocrinology, Odense University Hospital, Odense and 2The Clinical Research Institute, University of Southern Denmark, Odense, Denmark

Accepted 15 October 2014

Abstract Aims To investigate whether the presence of glutamic acid decarboxylase (GAD) autoantibodies post-partum in women with prior gestational diabetes mellitus was associated with changes in metabolic characteristics, including b–cell function and insulin sensitivity.

During 1997–2010, 407 women with gestational diabetes mellitus were offered a 3–month post-partum follow-up including anthropometrics, serum lipid profile, HbA1c and GAD autoantibodies, as well as a 2–h oral glucose tolerance test (OGTT) with blood glucose, serum insulin and C–peptide at 0, 30 and 120 min. Indices of insulin sensitivity and insulin secretion were estimated to assess insulin secretion adjusted for insulin sensitivity, disposition index (DI).

Methods

Results Twenty-two (5.4%) women were positive for GAD autoantibodies (GAD+ve) and the remainder (94.6%) were negative for GAD autoantibodies (GAD–ve). The two groups had similar age and prevalence of diabetes mellitus. Women who were GAD+ve had significantly higher 2–h OGTT glucose concentrations during their index-pregnancy (10.5 vs. 9.8 mmol/l, P = 0.001), higher fasting glucose (5.2 vs. 5.0 mmol/l, P = 0.02) and higher 2–h glucose (7.8 vs. 7.1 mmol/l, P = 0.05) post-partum. Fasting levels of C–peptide and insulin were lower in GAD+ve women compared with GAD–ve women (520 vs. 761 pmol/l, P = 0.02 and 33 vs. 53 pmol/l, P = 0.05) Indices of insulin sensitivity were similar in GAD+ve and GAD–ve women, whereas all estimates of DI were significantly reduced in GAD+ve women.

GAD+ve women had higher glucose levels and impaired insulin secretion adjusted for insulin sensitivity (DI) compared with GAD–ve women. The combination of OGTT and GAD autoantibodies post-partum identify women with impaired b–cell function. These women should be followed with special focus on development of Type 1 diabetes. Conclusion

Diabet. Med. 32, 198–205 (2015)

Introduction Gestational diabetes mellitus (GDM), defined as glucose intolerance with onset or first diagnosis during pregnancy, complicates an increasing number of pregnancies worldwide [1]. It is well established that women with previous GDM are characterized by several metabolic abnormalities, such as insulin resistance and b–cell dysfunction, leading to an increased risk of later diabetes mellitus [2]. Impaired insulin sensitivity is believed to play a role in the development of Type 2 diabetes following GDM [3]. In agreement, women with GDM have a seven-fold risk of developing Type 2 diabetes compared with women without GDM [4]. MoreCorrespondence to: Trine Præstekjær Lundberg. E-mail: [email protected]

198

over, it has been estimated that more than one-third of women with Type 2 diabetes have had GDM previously [5]. However, the risks of developing autoimmune diabetes [Type 1 diabetes and/or latent autoimmune diabetes of adulthood (LADA)] after GDM has not been studied to the same extend. A recent literature review by Lapolla et al. [6] recommends that autoimmune GDM be considered as a distinct clinical entity in the traditional classification of GDM. The presence of glutamic acid decarboxylase (GAD) autoantibody is an important marker of Type 1 diabetes and LADA. This autoantibody is a detectable marker of an ongoing destructive process in the pancreatic b–cells and therefore provides a tool to identify women at risk of developing autoimmune diabetes after GDM. The consequences of the presence of GAD autoantibodies in healthy

ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK

Research article

What’s new? • Women positive for glutamic acid decarboxylase (GAD) autoantibodies with recent gestational diabetes mellitus were characterized by impairment in insulin secretion when adjusted for insulin sensitivity. • Metabolic characterization and GAD autoantibodies measurement were done concomitantly. • GAD autoantibody-positive and GAD autoantibody-negative women had different phenotypical characteristics 3 months post-partum.

DIABETICMedicine

for the prevention of later diabetes and planning of future pregnancies. During pregnancy, women with GDM were managed at the Department of Obstetrics and Endocrinology, Odense University Hospital. Medical charts and laboratory data with information of pregnancy outcome were available, including the time and result of the diagnostic OGTT during pregnancy. Maternal and infant characteristics recorded were: pre-pregnancy BMI, age at delivery, parity, duration of gestation, birth weight of the infant, insulin treatment, ethnicity and information on family history of diabetes. The study was in accordance with the Declaration of Helsinki.

Screening procedure and diagnosis of GDM

individuals with a normal glucose tolerance test are unknown. Several studies [7–9] have shown that women with GDM and autoantibodies have significantly increased risk of developing Type 1 diabetes or LADA post-partum. These studies underline the importance of detecting women at risk of autoimmune GDM in order to diagnose insulin-dependent diabetes at an early stage during the years after pregnancy. Studies [10–16] have investigated insulin secretion and insulin sensitivity post-partum in women with GDM. In summary, all studies reported that women with a history of GDM showed impaired insulin secretion and sensitivity. However, at present, no studies have examined how the presence of GAD autoantibodies affects these measurements. We hypothesized that women positive for GAD autoantibodies (GAD+ve) compared with women negative GAD autoantibodies (GAD–ve) with recent GDM have clinical features and metabolic characteristics that resemble patients with Type 1 diabetes and LADA, and are characterized by an impairment in b–cell function. Thus, our aim in this study was to characterize the metabolic profiles including indices of insulin sensitivity, insulin secretion and disposition index (DI) in GAD+ve women versus GAD–ve women with recent GDM.

On the first visit to the obstetric department, pregnant women with risk factors for GDM (family history of diabetes mellitus, 20% overweight or more in the non-pregnant state or BMI ≥ 27 kg/m2, history of unexplained stillbirth, previous macrosomic baby, i.e. birth weight ≥ 4500 g, age ≥ 35 years, previous GDM and glucosuria) were selected for screening. During 1997–1998, GDM was diagnosed based on a 75–g, 3–h OGTT (n = 48). The OGTT was defined abnormal if two or more glucose values exceeded the mean + 3 SD on a curve based on a group of 40 Danish healthy, non-obese, non-pregnant women without a family history of diabetes [17]. Diagnostic cut-off levels for plasma glucose were: fasting, 5.7 mmol/l; 30 min, 11.9 mmol/l; 60 min, 12.0 mmol/l; 90 min, 9.7 mmol/l; 120 min, 8.9 mmol/l; 150 min, 8.5 mmol/l; and 180 min, 7.4 mmol/ l. In 1999, the 75–g, 3–h OGTT was replaced by a 75–g, 2–h OGTT with measurements of capillary whole blood, fasting capillary blood glucose and 2–h whole blood glucose (n = 359). The OGTT was defined as abnormal when fasting capillary blood glucose was ≥ 6.1 mmol/l and/or 2–h whole blood glucose was ≥ 9.0 mmol/l [18]. Thus, the two screening tests—the 75–g, 3–h OGTT and 75–g, 2–h OGTT— comprise similar diagnostic fasting and 2–h values.

Participants and methods

Biochemical methods at follow-up

Participants

During 1997–2010 women with GDM delivering at Odense University Hospital in Southern Denmark were routinely offered post-partum follow-up with an oral glucose tolerance test (OGTT) and clinical examination at the Department of Endocrinology, Odense University Hospital. The women were examined using a 75–g, 2–h OGTT three months after delivery. Capillary blood glucose, serum insulin and C–peptide were measured at 0, 30 and 120 min. In addition, fasting lipids, HbA1c and GAD autoantibodies were analysed and blood pressure, height and weight were measured. Only GAD autoantibodies and not any other antibodies known to predict the development of diabetes were measured routinely. This information was used to provide individual counselling

ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK

Blood glucose was measured on capillary whole blood using € HemoCue (HemoCue Glucose 201 DM RT Analyzer, Angelholm, Sweden). HbA1c was measured on venous blood using Tosoh G8 chromatograms, which provide direct determination of stable HbA1c [19]. Plasma total cholesterol, HDL cholesterol and triglycerides were analysed using enzymatic colorimetric reactions (Modular P, Roche, Switzerland), and LDL cholesterol was calculated using the Friedewald equation [20]. Serum insulin and C–peptide were analysed using a commercial time-resolved fluoroimmunoassay (Perkin-Elmer Life Sciences, Turku, Finland). GAD autoantibodies were measured using the GAD Autoantibody RIA kit (RSR, Cardiff, UK) in units per ml (U/ml) with 95% specificity and 84% sensitivity [21]. Units are RSR arbitrary units and analysis of NIBSC 97/550 in RSR’s GAD antibody assay indicates that

199

DIABETICMedicine

GAD autoantibodies in women with gestational diabetes mellitus  T. P. Lundberg et al.

one RSR kit unit per ml is equivalent to 25 U/ml of 97/550 (WHO units). A value > 1 U/ml (25 WHO units) is considered positive. Titres > 1.0 U/ml are defined as positive and all GAD+ve women had titres > 1.0 U/ml. Derived estimates of insulin release and insulin sensitivity from OGTT results (Table S1)

Indices of oral glucose-stimulated insulin secretion are reported as: 1. Insulinogenic index [22], calculated as [serum insulin30 min serum insulin0 min (pmol/l)]/[plasma glucose30 min plasma glucose0 min (mmol/l)]; 2. Corrected insulin response [23], calculated as: [serum insulin30 min 9 100/(plasma glucose30 min 9 (plasma glucose30 min 3.89)]; and 3. BIGTT acute insulin response (AIR) index [24]. The surrogate measures of insulin sensitivity are reported as: 1. Homeostatic model assessment (HOMA)–IS, calculated as the 1/HOMA–IR, where HOMA–IR is calculated as [plasma glucose0 min (mmol/l) 9 serum insulin0 min (mU/ L)]/22.5; 2. Stumvoll index includes glucose and insulin during a 2–h OGTT, and the index is validated by the euglycaemic– hyperinsulinaemic and the hyperglycaemic clamp technique, respectively and was calculated as reported previously [25]; and 3. BIGTT sensitivity index (SI). The BIGTT indices, which apply information on sex and BMI, combined with plasma glucose and serum insulin during an OGTT, were calculated as reported by Hansen et al. [24]. b–Cell function adjusted for insulin sensitivity was assessed as the disposition index (DI), using three different estimates: 1. multiplying corrected insulin response and HOMA–IS, because these measures are not intrinsically interdependent [26]; 2. multiplying BIGTT–SI and BIGTT–AIR; and 3. multiplying corrected insulin response and Stumvoll index. Here, the calculated disposition indices (DI1–DI3) were used as indices of b–cell function adjusted for insulin sensitivity.

regulation is defined as diabetes mellitus, impaired fasting glycaemia, impaired glucose tolerance or more than one of these. Weight and height were measured with the participants wearing indoor clothes and no shoes. Blood pressure was measured using a standard mercury sphygmomanometer with an appropriate cuff size after 5 min rest. Systolic blood pressure ≥ 140 mmHg and/or diastolic blood pressure ≥ 90 mmHg were defined as hypertension. Statistical analysis

Statistical analysis was carried out by using STATA version 11.1 (Stata Corporation, College Station, TX, USA). Because some of the parameters were not normally distributed, we decided to apply non-parametric tests in all analysis. Data are given as median and interquartile (IQ) range or number and per cent. The Mann–Whitney test was used to compare differences between two groups. The v2-test was used for comparison of proportions in two groups. Correlation analysis was performed using Spearman’s correlation analysis. For all analyses, a two-tailed P-value < 0.05 was considered significant.

Results Twenty-two of 407 women (5.4%) were GAD+ve at follow-up 3 months post-partum. The median time for performing the OGTT was 98 (84–133) days post-partum. GAD+ve and GAD–ve women were of a similar age: 32.0 (29.0–34.3) vs. 32.2 (28.8–36.2) years (P = 0.64). Stratifying for ethnicity did not significantly change results.

Baseline characteristics

Baseline characteristics from the index-pregnancy of GAD+ve and GAD–ve women are presented in Table 1. Diagnostic 2–h glucose levels for GDM during index-pregnancy were significantly higher in GAD+ve women (P = 0.001), and all women in this group were Caucasian compared with 81.8% among GAD–ve women (P = 0.03). There were no significant differences between GAD+ve and GAD–ve women at index-pregnancy regarding: pre-pregnancy BMI, family history of diabetes, need for insulin treatment in pregnancy (%), infant birth weight, gestational age at GDM diagnosis or parity (Table 1). Fasting glucose concentrations were available in a subcohort (n = 251), but showed no difference between GAD+ve and GAD–ve women (5.7 (4.5–6.8) vs. 5.0 (4.3– 5.7) mmol/l; P = 0.16). Mean (SD) GAD titres were 21.2 (22.2) U/ml in the GAD+ve group.

Definitions

Normal glucose tolerance, impaired fasting glycaemia, impaired glucose tolerance and diabetes mellitus were defined by diagnostic levels for capillary blood glucose according to the 1999 WHO criteria [27]. Impaired glucose

200

Clinical characteristics and glucose tolerance status at follow-up OGTT

The clinical characteristics and glucose tolerance status of GAD+ve vs. GAD–ve women at OGTT post-partum are

ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK

Research article

DIABETICMedicine

Table 1 Baseline characteristics from index-pregnancy of GAD+ve vs. GAD–ve women Variable

GAD+ve

GAD–ve

P

Age at delivery (years) Pre-pregnancy BMI (kg/m2) Etnicity Caucasian Family history of diabetes 2-h glucose (mmol/l) Insulin treatment in pregnancy Birth weight of infant (g) Gestational age at GDM diagnosis (days) Nulliparous

32.0 (29.0–34.3) 26.8 (23.0–31.6)

32.2 (28.8–36.2) 29.0 (24.4–33.4)

0.64 0.16

22 12 10.5 8 3686 229 9

315 189 9.8 107 3565 209 132

0.03 0.62 0.001 0.39 0.71 0.41 0.53

(100) (54.5) (9.6–13.0) (36.4) (3356–4000) (171–253) (41)

(81.8) (49.1) (9.2–10.8) (27.8) (3152–3945) (191–234) (34)

Data are given as median (interquartile range) or number (%). The values were available for n = 19–22 in the GAD+ve group and for n = 292–385 in the GAD–ve group. P-values were calculated using the Mann–Whitney test and v2-test.

presented in Table 2. In total, 180 (44%) women with recent GDM had impaired glucose regulation. The prevalences of diabetes mellitus, impaired fasting glycaemia, impaired glucose tolerance and impaired glucose regulation did not differ between GAD+ve and GAD–ve women. GAD+ve women had significantly lower fasting serum insulin and C– peptide and higher fasting and 2–h capillary blood glucose compared with GAD–ve women. LDL cholesterol and post-partum BMI were similar in the two groups. Diastolic and systolic blood pressures measured in a subcohort (n = 210) showed no differences between GAD+ve and

Table 2 Phenotypes of GAD+ve vs. GAD-ve women at OGTT post-partum. Variable Impaired glucose regulation (%) Diabetes mellitus (%) Impaired fasting glucose (%) Impaired glucose tolerance (%) Normal glucose tolerance (%) LDL cholesterol (mmol/l) Fasting C–peptide (pmol/l) Fasting insulin (pmol/l) Fasting glucose (mmol/l) 2-h glucose (mmol/l) HbA1c (mmol/mol) BMI post-partum (kg/m2)

GAD+ve

GAD–ve women (diastolic blood pressure: 79 (74–80) vs. 80 (70–85) mmHg, P = 0.62; and systolic blood pressure: 120 (113–122) vs. 120 (112–130) mmHg, P = 0.53). Glucose, C–peptide and insulin at follow-up OGTT (Fig. 1)

Fasting and 2–h glucose concentrations were significantly higher in GAD+ve women compared with GAD–ve women (P = 0.02 and 0.050, respectively). Fasting C–peptide and fasting insulin levels were significantly lower in GAD+ve women than in GAD–ve women (both P < 0.05) (Table 2). Estimates of insulin secretion and insulin sensitivity

The surrogate measures of insulin secretion—BIGTT–AIR and insulinogenic index—tended to be lower in GAD+ve women than in GAD–ve women (P = 0.07 and 0.05, respectively) (Table 3). The corrected insulin response was significantly lower in GAD+ve women compared with GAD– ve women (P = 0.002). Measures of insulin sensitivity— BIGTT-SI, HOMA–IS and Stumvoll index—were similar in the two groups (Table 3).

GAD–ve

P

13 (59)

167 (43)

0.15

5 (23)

57 (15)

0.32

0 (0)

15 (4)

0.35

8 (36)

95 (25)

0.22

9 (41)

214 (56)

0.18

Disposition index (estimates of b–cell function adjusted for

3.0 (2.5–3.4)

3.0 (2.5–3.6)

0.91

insulin sensitivity)

520 (374–715)

761 (563–1035) 0.02

33 (28–56)

53 (33–82)

0.05

5.2 (5.1–5.8)

5.0 (4.6–5.5)

0.02

7.8 (6.9–9.8)

7.1 (6.1–8.9)

0.05

36 (33–40) 36 (33–39) 0.85 26.9 (23.2–33.3) 29.0 (25.2–33.0) 0.18

Data are given as median (interquartile range). The values were available for n = 20–22 in the GAD+ve group and n = 359–380 in the GAD–ve group. All P-values were calculated using the Mann–Whitney test.

ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK

GAD+ve women had significantly lower DI based on corrected insulin response and HOMA–IS compared with GAD–ve women (DI1: P = 0.0455) (Table 3). The two other estimates of DI confirmed this finding—DI2, based on BIGTT–AIR and BIGTT–SI (P = 0.0311) and DI3 based on corrected insulin response and Stumvoll index (P = 0.0055) —were both significant lower in GAD+ve women compared with GAD–ve women.

Correlation analysis

To explore the effect of GAD antibody titres on insulin sensitivity and b–cell function, we examined the relationship

201

GAD autoantibodies in women with gestational diabetes mellitus  T. P. Lundberg et al.

DIABETICMedicine

(a)

Plasma Glucose (mmol/l)

10 *

9 8

Discussion

7 6

GAD-ve GAD+ve

*

5 4

0

20

40

60

80

100

120

min (b)

400

Serum Insulin (pmol/l)

*** 300

200 GAD-ve GAD+ve 100

0

*

0

20

40

60

80

100

120

min **

Serum C-peptide (pmol/l)

(c) 3000

*** 2000

1000

0

GAD-ve GAD+ve

*

0

20

40

60

80

100

120

min

FIGURE 1 Concentrations of (a) blood glucose, (b) serum insulin and (c) C-peptide during an OGTT performed 3-month post-partum in women with gestational diabetes mellitus with (GAD+ve, white circles) or without (GAD ve, black circles) the presence of GAD autoantibody. *P < 0.05, **P < 0.01 and ***P < 0.001 for GAD+ve vs. GAD ve. Data represent mean +/ SEM.

between GAD antibody titres and measures of glucose, insulin and C–peptide, as well as all estimates of insulin sensitivity, insulin secretion and disposition indices. In

202

GAD+ve women, GAD antibody titres correlated positively with 2–h plasma glucose concentrations (r = 0.615; P = 0.004), and negatively with insulinogenic index (r = 0.498; P = 0.035) and D3 (r = 0.565; P = 0.023). No other significant correlations were observed.

Pregnancy is characterized by alterations in the immune system with fluctuations in symptoms and markers of autoimmune diseases [28]. Thus markers of b–cell destruction like GAD autoantibodies might be higher outside pregnancy. Most studies on GDM and the role of autoimmune markers as predictors of later diabetes mellitus have used GAD autoantibodies measured in pregnancy [7,29–31]. Thus, the role of GAD autoantibodies might be underestimated. In this follow-up study, we aimed to study the metabolic and clinical characteristics in GAD+ve and GAD– ve women with recent GDM, including estimates of insulin sensitivity, insulin secretion and insulin secretion adjusted for insulin sensitivity (DI). To the best of our knowledge, this is the first study with concomitant measures of GAD autoantibodies and post-partum anthropometrics and detailed metabolic characterization. We found a significant impairment in insulin secretion adjusted for insulin sensitivity (DI) in GAD+ve women compared with GAD–ve women 3 months post-partum. GAD+ve women had significantly higher fasting and 2–h glucose compared with GAD–ve women. This is in keeping with the results from a Swedish study in which women with GDM and controls had a follow-up OGTT 1–2 years post-partum [32]. The authors reported a higher risk of diabetes among women with previous GDM and the presence of GAD autoantibodies. They also noted that a number of women with GDM had developed diabetes 1–2 years after pregnancy and therefore were not included in the follow-up study. A number of studies [10–16] have investigated insulin secretion and sensitivity in women with a history of GDM. These studies have shown the presence of insulin resistance and impaired b–cell function compared with women without a history of GDM. Previously, Ryan et al. [13] studied 14 women with GDM 8 years post-partum and found impaired insulin secretion and insulin action before the development of hyperglycaemia during OGTT post-partum. Retnakaren et al. [10] investigated women with various degrees of glucose intolerance in pregnancy and found that both insulin sensitivity and pancreatic b–cell function based on OGTT post-partum decreased progressively with severity of glucose intolerance in pregnancy. Thus, GDM is a strong predictor of abnormal glucose homeostasis in later life. In a London multi-ethnic population, Kousta et al. [11] studied women with previous GDM and normal glucose tolerance, thus focusing on the group without severe metabolic derangement. By dividing the women according to

ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK

Research article

DIABETICMedicine

Table 3 Indices of insulin sensitivity and insulin secretion from an OGTT Variable Insulin sensitivity BIGTT-SI/0-30-120 HOMA-IS Stumvoll index HOMA-IR Insulin secretion BIGTT-AIR0-30-120 Insulinogenic index Corrected insulin response (CIR) Disposition index (DI) DI1 (CIR 9 HOMA-IS) DI2 (BIGTT-SI/0-30-120 9 BIGTT-AIR0-30-120) DI3 (CIR 9 Stumvoll index)

GAD+ve

GAD–ve

13.9 0.73 0.0898 1.38

11.6 0.49 0.0857 2.03

(7.1–15.5) (0.36–1.1) (0.0595–0.1004) (0.95–2.93)

1468 (1349–2130) 28.5 (24.6–62.9) 322 (194–477) 197 (130–313) 20 194 (14 581–22 044) 25.4 (13.9–35.6)

(8.4–15.0) (0.31–0.87) (0.0635–0.0998) (1.15–3.27)

2050 (1460–2858) 50.8 (31.5–87.1) 588 (356–942) 320 (174–521) 24 078 (16 689–33 275) 45.3 (25.2–73.2)

P

0.88 0.21 0.85 0.21 0.07 0.05 < 0.01 0.05 0.03 < 0.01

Data are given as median (interquartile range). The values were available for n = 16–20 in the GAD+ve group and n = 331–357 in the GAD– ve group. All P-values were calculated using the Mann–Whitney test.

ethnicity, the authors demonstrated different metabolic patterns for different ethnic groups. As 100% of GAD+ve and 80% of GAD–ve women in our study were Caucasian we do not anticipate ethnicity to have major influence on our results. Tura et al. [12] compared 24 lean, normotolerant women with previous GDM with 23 controls and found that impairment of b–cell glucose sensitivity may be an intrinsic risk factor for Type 2 diabetes independent of obesity and hyperglycaemia. Their study is not directly comparable with our work. Thus, in our study, we compared GAD+ve with GAD–ve women rather than controls, and our population consists of a heterogeneous group of women, the majority of whom were obese and glucose intolerant. However, we also observed that the higher fasting and 2–h glucose levels and impaired pancreatic b–cell function in GAD+ve women with former GDM were independent of BMI and glycaemic control as well as insulin sensitivity. Thus, we found no significant difference between GAD+ve and GAD–ve women regarding BMI and weight, and no significant difference in surrogate markers of insulin resistance between the groups. However, we found that GAD+ve women had significant impairment in OGTT-derived indices of glucose-stimulated insulin secretion adjusted for insulin sensitivity (DI) compared with GAD–ve women. This emphasizes the potential importance of identifying this subgroup of women, because a low DI has been shown to be an early marker of b–cell depletion and has been found to predict the development of diabetes [26]. Although autoimmune GDM only accounts for a relatively small proportion—about 10%—of all GDM cases [6], several studies have shown that women with GDM and autoantibodies against b–cell have a significantly increased risk of developing Type 1 diabetes or LADA after pregnancy [7–9]. Most studies refer to years of follow-up compared with our study, when evaluating the development of diabetes post-partum [7,8]. L€ obner et al. [15] found an 8-year ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK

diabetes risk of 52.7% in women with GDM and even higher (97%) in those women who were GAD+ve. In a recent study by Nilsson et al. [29], 50% of autoantibody-positive women had developed diabetes 8 years post-partum and among these women, 41.7% developed diabetes within 6 months and 50% within 1 year post-partum. In a short-term follow-up study from Korea [30], GAD+ve women (measured during pregnancy) with previous GDM were evaluated according to glucose homeostasis 6 weeks post-partum. This study found 33% of GAD+ve women with diabetes compared with 7% of GAD–ve women at OGTT 6 weeks post-partum. However, the development of impaired glucose tolerance was not significantly different between GAD+ve and GAD–ve women (33% vs. 25%). Consistent with these studies, we found that already 3 months after pregnancy, five GAD+ve women (23%) were diagnosed with diabetes mellitus and eight women (38%) had impaired glucose tolerance according to OGTT 3 months post-partum. The most important finding in our study compared with other studies of GAD autoantibodies in women with former GDM, is the significant impairment in insulin secretion adjusted for markers of both hepatic (HOMA–IS) and peripheral insulin sensitivity (DI) in GAD+ve women compared with GAD–ve women. In our study, 36% of the GAD+ve women were insulin treated during pregnancy compared with 28% of GAD–ve women (P = 0.39). One might have expected a significantly higher rate of insulin treatment in GAD+ve women due to relative insulin deficiency, as indicated by higher 2–h glucose levels during index pregnancy as well as higher fasting and 2– h glucose levels post-partum, than in GAD–ve women. The lack of significance might be due to low numbers. The association between positive GAD autoantibodies and a higher frequency of insulin treatment during pregnancy was found in the 6-week follow-up study in Korea [30]. Thus, Yu et al. found that GAD+ve women required more frequent insulin therapy than GAD–ve women (47% vs. 23%;

203

DIABETICMedicine

GAD autoantibodies in women with gestational diabetes mellitus  T. P. Lundberg et al.

P = 0.04). Yu et al. reported higher area under the curve (AUC) of glucose during a 3–h OGTT in GAD+ve women, whereas fasting plasma insulin and AUC of insulin were not different between the groups [30]. In comparison, we found both significantly higher fasting and 2–h glucose levels and lower insulin and C–peptide levels during OGTT in GAD+ve women compared with GAD–ve women. This discrepancy might be explained by a tendency for insulin resistance in GAD+ve women in the Korean study [30]. This will lead to increased insulin secretion to compensate for the degree of insulin resistance, and therefore it is important to adjust for this when evaluating b–cell function. In our study, two estimates of glucose-stimulated insulin secretion (BIGTT– AIR and insulinogenic index) tended to be reduced only in GAD+ve women with previous GDM. However, all three estimates of DI demonstrated that GAD+ve women have impaired insulin secretion when adjusted for insulin sensitivity. The strength of this follow-up study is our large study population of women with previous GDM. Insulin sensitivity, insulin secretion and insulin secretion adjusted for insulin sensitivity (DI) were estimated and autoimmune markers were determined prospectively a few months after pregnancy. We have used the surrogate measures of insulin secretion and sensitivity to characterize the glucose homeostasis in our study population. Ideally, the clamp technique would have been the best method to evaluate the glucose homeostasis. Only capillary blood glucose samples were available at OGTT after pregnancy. Venous plasma glucose would have been preferable as a more standardized method to evaluate glucose tolerance during OGTT. In daily clinical practice, we used measurements of capillary blood glucose in order to give the women immediate information of their status. We used formulas based on venous plasma blood to calculate insulin sensitivity, HOMA–IR and the b–cell function. A potential minor limitation of this study is the change of diagnostic OGTT in pregnancy during the study period. However, the diagnostic thresholds in the two diagnostic tests are similar and we consider the women diagnosed before and after the change of test to have glucose intolerance in pregnancy of similar severity. Only women at risk for GDM were included in the GDM screening programme, due to the national standard screening criteria during the study period. Furthermore, the diagnostic criteria were higher than the criteria recently proposed by the WHO. Using the latter, we estimate a two- to threefold higher GDM frequency in the Danish population. Thus, the women in the present study represent a high-risk group in which the prevalence of GAD autoantibodies is likely to be markedly higher than in GDM populations detected by universal screening. A study with screening in a non-selected population of pregnant women and lower diagnostic criteria might detect another metabolic characterization of GAD+ve women.

204

In conclusion, GAD+ve women with recent GDM are characterized by impairment in insulin secretion when adjusted for insulin sensitivity. The combination of OGTT and measurement of GAD autoantibodies post-partum highlights women with GDM having a greater risk of developing insulin-dependent diabetes. Future follow-up studies in this cohort will investigate the significance of these findings in relation to the development of different types of diabetes.

Funding sources

Funding was received from the Odense University Hospital Free Research Fund.

Competing interests

The authors declare no conflict of interest.

References 1 Hunt KJ, Schuller KL. The increasing prevalence of diabetes in pregnancy. Obstet Gynecol Clin North Am 2007 Jun; 34(2): 173–199, vii. 2 Vambergue A, Dognin C, Boulogne A, Rejou MC, Biausque S, Fontaine P. Increasing incidence of abnormal glucose tolerance in women with prior abnormal glucose tolerance during pregnancy: DIAGEST 2 study. Diabet Med 2008; 25: 58–64. 3 Retnakaran R, Qi Y, Connelly PW, Sermer M, Hanley AJ, Zinman B. Low adiponectin concentration during pregnancy predicts postpartum insulin resistance, beta cell dysfunction and fasting glycaemia. Diabetologia 2010; 53: 268–276. 4 Bellamy L, Casas JP, Hingorani AD, Williams D. Type 2 diabetes mellitus after gstational diabetes: a systematic review and meta-analysis. Lancet 2009 May 23; 373: 1773–1779. 5 Cheung NW, Byth K. Population health significance of gestational diabetes. Diabetes Care 2003; 26: 2005–2009. 6 Lapolla A, Dalfra MG, Fedele D. Diabetes related autoimmunity in gestational diabetes mellitus: is it important? Nutr Metab Cardiovasc Dis 2009; 19: 674–682. 7 Jarvela IY, Juutinen J, Koskela P, Hartikainen AL, Kulmala P, Knip M et al. Gestational diabetes identifies women at risk for permanent type 1 and type 2 diabetes in fertile age: predictive role of autoantibodies. Diabetes Care 2006; 29: 607–612. 8 Ivarsson SA, Ackefors M, Carlsson A, Ekberg G, Falorni A, Kockum I et al. Glutamate decarboxylase antibodies in non-diabetic pregnancy precedes insulin-dependent diabetes in the mother but not necessarily in the offspring. Autoimmunity 1997; 26: 261–269. 9 Lauenborg J, Hansen T, Jensen DM, Vestergaard H, Molsted-Pedersen L, Hornnes P et al. Increasing incidence of diabetes after gestational diabetes: a long-term follow-up in a Danish population. Diabetes Care 2004; 27: 1194–1199. 10 Retnakaran R, Qi Y, Sermer M, Connelly PW, Hanley AJ, Zinman B. Glucose intolerance in pregnancy and future risk of pre-diabetes or diabetes. Diabetes Care 2008; 31: 2026–2031. 11 Kousta E, Lawrence NJ, Godsland IF, Penny A, Anyaoku V, Millauer BA et al. Early metabolic defects following gestational diabetes in three ethnic groups of anti-GAD antibodies negative women with normal fasting glucose. Hormones (Athens) 2007; 6: 138–147.

ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK

Research article

12 Tura A, Mari A, Winzer C, Kautzky-Willer A, Pacini G. Impaired beta-cell function in lean normotolerant former gestational diabetic women. Eur J Clin Invest 2006; 36: 22–28. 13 Ryan EA, Imes S, Liu D, McManus R, Finegood DT, Polonsky KS et al. Defects in insulin secretion and action in women with a history of gestational diabetes. Diabetes 1995; 44: 506–512. 14 Kautzky-Willer A, Prager R, Waldhausl W, Pacini G, Thomaseth K, Wagner OF et al. Pronounced insulin resistance and inadequate beta-cell secretion characterize lean gestational diabetes during and after pregnancy. Diabetes Care 1997; 20: 1717–1723. 15 L€ obner K, Knopff A, Baumgarten A, Mollenhauer U, Marienfeld S, Garrido-Franco M et al. Predictors of postpartum diabetes in women with gestational diabetes mellitus. Diabetes 2006; 55: 792–797. 16 Tura A, Grassi A, Winhofer Y, Guolo A, Pacini G, Mari A et al. Progression to type 2 diabetes in women with former gestational diabetes: time trajectories of metabolic parameters. PLoS One 2012; 7: e50419. 17 Damm P, Handberg A, Kuhl C, Beck-Nielsen H, Molsted-Pedersen L. Insulin receptor binding and tyrosine kinase activity in skeletal muscle from normal pregnant women and women with gestational diabetes. Obstet Gynecol 1993; 82: 251–259. 18 Jensen DM, Molsted-Pedersen L, Beck-Nielsen H, Westergaard JG, Ovesen P, Damm P. Screening for gestational diabetes mellitus by a model based on risk indicators: a prospective study. Am J Obstet Gynecol 2003; 189: 1383–1388. 19 Tosoh. Available at: www.diagnostics.us.tosohbioscience.com/ Products/G8þHPLCþAnalyzer/2012. 20 Vossen M, Vettorazzi E, Wegscheider K, Laatsch A. Estimation of LDL cholesterol by the Friedewald formula during an oral glucose tolerance test. Diabet Med 2011; 28: 1433–1434. 21 Powell M, Prentice L, Asawa T, Kato R, Sawicka J, Tanaka H et al. Glutamic acid decarboxylase autoantibody assay using 125I-labelled recombinant GAD65 produced in yeast. Clin Chim Acta 1996; 256: 175–188. 22 Phillips DI, Clark PM, Hales CN, Osmond C. Understanding oral glucose tolerance: comparison of glucose or insulin measurements during the oral glucose tolerance test with specific measurements of insulin resistance and insulin secretion. Diabet Med 1994; 11: 286–292. 23 Sluiter WJ, Erkelens DW, Reitsma WD, Doorenbos H. Glucose tolerance and insulin release, a mathematical approach I. Assay of the beta-cell response after oral glucose loading. Diabetes 1976; 25: 241–244. 24 Hansen T, Drivsholm T, Urhammer SA, Palacios RT, Volund A, Borch-Johnsen K et al. The BIGTT test: a novel test for

ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK

DIABETICMedicine

25

26

27

28

29

30

31

32

simultaneous measurement of pancreatic beta-cell function, insulin sensitivity, and glucose tolerance. Diabetes Care 2007; 30: 257–262. Stumvoll M, Van Haeften T, Fritsche A, Gerich J. Oral glucose tolerance test indexes for insulin sensitivity and secretion based on various availabilities of sampling times. Diabetes Care 2001; 24: 796–797. Utzschneider KM, Prigeon RL, Faulenbach MV, Tong J, Carr DB, Boyko EJ et al. Oral disposition index predicts the development of future diabetes above and beyond fasting and 2–h glucose levels. Diabetes Care 2009; 32: 335–341. Alberti KG, Zimmet PZ. Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: diagnosis and classification of diabetes mellitus provisional report of a WHO consultation. Diabet Med 1998; 15: 539–553. Adams Waldorf KM, Nelson JL. Autoimmune disease during pregnancy and the microchimerism legacy of pregnancy. Immunol Invest 2008; 37: 631–644. Nilsson C, Ursing D, Torn C, Aberg A, Landin-Olsson M. Presence of GAD antibodies during gestational diabetes mellitus predicts type 1 diabetes. Diabetes Care 2007; 30: 1968–1971. Yu SH, Park S, Kim HS, Park SY, Yim CH, Han KO et al. The prevalence of GAD antibodies in Korean women with gestational diabetes mellitus and their clinical characteristics during and after pregnancy. Diabetes Metab Res Rev 2009; 25: 329–334. Bo S, Menato G, Pinach S, Signorile A, Bardelli C, Lezo A et al. Clinical characteristics and outcome of pregnancy in women with gestational hyperglycaemia with and without antibodies to beta-cell antigens. Diabet Med 2003; 20: 64–68. Papadopoulou A, Lynch KF, Anderberg E, Landin-Olsson M, Hansson I, Agardh CD et al. HLA–DQB1 genotypes and islet cell autoantibodies against GAD65 and IA–2 in relation to development of diabetes post partum in women with gestational diabetes mellitus. Diabetes Res Clin Pract 2012; 95: 260–264.

Supporting Information Additional Supporting Information may be found in the online version of this article: Table S1. Calculations of estimates.

205

Copyright of Diabetic Medicine is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.

Glutamic acid decarboxylase autoantibody-positivity post-partum is associated with impaired β-cell function in women with gestational diabetes mellitus.

To investigate whether the presence of glutamic acid decarboxylase (GAD) autoantibodies post-partum in women with prior gestational diabetes mellitus ...
154KB Sizes 0 Downloads 4 Views