- Email: [email protected]
Hemoglobin A1c as a tool for the diagnosis of type 2 diabetes in 208 premenopausal women with polycystic ovary syndrome
Hemoglobin A1c as a tool for the diagnosis of type 2 diabetes in 208 premenopausal women with polycystic ovary syndrome Line Velling Magnussen, M.D., Hanne Mumm, M.D., Marianne Andersen, M.D., Ph.D., and Dorte Glintborg, M.D., Ph.D. Department of Endocrinology and Metabolism, Odense University Hospital, Odense, Denmark
Objective: To study hemoglobin A1c (HbA1c) as a tool for diagnosing diabetes and to study HbA1c as a cardiovascular risk marker in patients with polycystic ovary syndrome (PCOS). Design: Retrospective observational study. Setting: Academic tertiary-care medical center. Patient(s): Two hundred eight premenopausal women with PCOS. Intervention(s): Patients underwent clinical evaluation (Ferriman-Gallwey score, body mass index, waist, blood pressure), hormone analyses (T, sex hormone–binding globulin, fasting lipids, insulin, glucose, HbA1c), transvaginal ultrasound, and 2-hour oral glucose tolerance tests (OGTT) measuring capillary blood glucose (BG) at 0 (BG 0) and 120 (BG 120) minutes, insulin, and C-peptide. Main Outcome Measure(s): Results of OGTT, HbA1c values. Result(s): Twenty patients were diagnosed with type 2 diabetes during OGTT. The sensitivity and specificity of HbA1c R6.5% for the diagnosis of diabetes were 35% and 99%, respectively, compared with the diagnosis established by OGTT. Hemoglobin A1c showed closer correlation with waist, body mass index, and lipid profile than BG 120, suggesting that HbA1c could be a cardiovascular risk marker. Conclusion(s): The clinical utility of HbA1c for diagnosing impaired glucose tolerance and type 2 diabetes in PCOS in daily practice is low. Long-term prospective studies are needed to determine whether HbA1c is superior to glucose levels as a cardiovascular risk marker in patients with PCOS. (Fertil Steril 2011;96: 1275–80. 2011 by American Society for Reproductive Medicine.) Key Words: PCOS, HbA1c, diabetes, OGTT, cardiovascular risk, sensitivity, specificity
Polycystic ovary syndrome (PCOS) is a common endocrine condition that is most often diagnosed according to the Rotterdam criteria (1). More than 50% of women with PCOS are insulin resistant, and patients with PCOS have an estimated five- to eightfold increased risk of type 2 diabetes mellitus (T2D) compared with age- and weight-matched controls (2, 3). Polycystic ovary syndrome is characterized by an increased activation of the inflammatory system, with decreased adiponectin levels and increased levels of adipokines, chemokines, and interleukins (3, 4). Abdominal obesity and increased activation of the inflammatory system in PCOS are associated with an increased risk of dyslipidemia and a five- to 10-fold increased risk of diabetes (3). The cardiovascular risk in PCOS is debated, but a recent study confirmed that postmenopausal women with PCOS had more cardiovascular events than controls, and PCOS was associated with angiographic coronary artery disease (5, 6). It is recommended that patients with PCOS are screened for diabetes using an oral glucose tolerance test (OGTT) (1, 7). Fasting plasma glucose (FPG) may be considered in low-risk patients or alternatively where OGTTs are not available (7). In daily practice Received July 4, 2011; revised August 24, 2011; accepted August 25, 2011; published online October 6, 2011. L.V.M. has nothing to disclose. H.M. has nothing to disclose. M.A. has nothing to disclose. D.G. has nothing to disclose. Reprint requests: Dorte Glintborg, M.D., Odense University Hospital, Department of Endocrinology and Metabolism, Kløvervænget 6, 3rd floor, 5000 Odense C, Denmark (E-mail: [email protected]).
0015-0282/$36.00 doi:10.1016/j.fertnstert.2011.08.035
the performance of OGTT may be inconvenient and timeconsuming because the patient has to be seen in a fasting state and therefore often must attend the clinic on two different days. Hemoglobin A1c (HbA1c) is a widely used marker of chronic glycemia and reflects the average blood glucose levels over a 2- to 3-month period (8). Hemoglobin A1c has higher repeatability than fasting glucose and can be assessed in the nonfasting state (8). The use of HbA1c instead of OGTT can avoid the problem of day-to-day variability of glucose values and the need for fasting and preceding dietary preparations. Hemoglobin A1c may, however, be affected by genetic, hematologic, and illness-related factors (8, 9). Recently it was reported from an International Expert Committee composed of members of the European Association for the Study of Diabetes, the International Diabetes Federation, and the American Diabetes Association that values of HbA1c R6.5% can be applied as a cutoff point for diagnosing T2D in asymptomatic patients (8). The cutoff point for HbA1c was largely based on the established association between HbA1c and microvascular disease (8). The guideline also stated that HbA1c levels between 6% and 6.5% should lead to an evaluating OGTT, whereas HbA1c levels of <6% should lead to no further tests (8). Increased HbA1c levels could be used as a marker of cardiovascular risk in nondiabetic adults (10). Hemoglobin A1c may therefore be superior to glucose levels for characterizing long-term risk in nondiabetic adults (10). In the present study we wanted to establish the value of HbA1c for the diagnosis of T2D and as a cardiovascular risk marker in women with PCOS.
Fertility and Sterility Vol. 96, No. 5, November 2011 Copyright ª2011 American Society for Reproductive Medicine, Published by Elsevier Inc.
1275
MATERIALS AND METHODS Inclusion Criteria
Methods
All premenopausal women referred to the outpatient clinic at the Department of Endocrinology during 1997–2010 with a diagnosis of PCOS were included. All patients met at least two of the following Rotterdam criteria: hirsutism, biochemical hyperandrogenism, menstrual disturbances, and/or polycystic ovaries. Patients suspended oral contraceptives and metformin for at least 3 months before evaluation. All included patients had normal thyroid function. During 1997–2005, an OGTT was part of the routine evaluation program at the department, whereas HbA1c measurements were not routinely performed. During 2005–2010, diabetes was excluded by fasting blood glucose and HbA1c measurement, whereas OGTT was performed in high-risk patients for diabetes or the metabolic syndrome. In the present study we included patients for whom data on both HbA1c and OGTT were available. No weight criteria were applied in the study.
Exclusion Criteria Women younger than 15 years were referred to the department of Pediatrics at Odense University Hospital, and patients with a primary complaint of infertility were referred to the local fertility clinic. To exclude patients with serious endocrine disease, s-prolactin and s-17-hydroxyprogesterone were measured. Two measurements of 24-hour urinary cortisol or a short dexamethasone suppression test were performed in the patients with a clinical suspicion of Cushing’s syndrome (2, 11). Patients with anemia and adrenal diseases were excluded to avoid misinterpretation of HbA1c levels. The study was designed as a retrospective observational study. According to our local ethics committee consult and Danish regulations, this study was exempt from institutional review board review. Clinical and paraclinical evaluation of the referred patients were part of the routine evaluation program, and therefore no informed consent from the patients was needed.
Routine evaluation included medical history, clinical examination, transvaginal ultrasound, and fasting blood samples. Hirsutism was assessed by the Ferriman-Gallwey score (12). Waist circumference was measured to the nearest centimeter in a standing position midway between the lower costal margin and the iliac crest. Body mass index (BMI) was calculated as the weight in kilograms divided by the height in meters squared. Transvaginal ultrasound was performed at the Department of Gynecology, Odense University Hospital. Fasting blood samples were drawn during the morning in follicular phase (cycle days 2–8) in patients with a cycle length shorter than 3 months. Patients with cycle length >3 months had the blood samples drawn on a random cycle day. Blood tests included androgens (total T, free T, sex hormone– binding globulin [SHBG], DHEAS, 17-hydroxyprogesterone), LH, FSH, PRL, and lipid profile. An OGTT was performed at 8:00 AM on a random day of the cycle in 208 fasting individuals. P-insulin and capillary blood glucose were measured at baseline and at 30 minutes, 60 minutes, and 120 minutes after oral ingestion of a glucose load containing the equivalent of 75 g anhydrous glucose dissolved in water. According to the World Health Organization diagnosis criteria, T2D was defined by fasting capillary glucose (BG 0) R6.1 mmol/ L or 2-hour glucose (BG 120) R11.1 mmol/L during OGTT. Impaired glucose tolerance (IGT) was defined by BG 0 <6.1 mmol/L and BG 120 7.8– 11.0 mmol/L (13). Because of the above-mentioned recommendation by the International Expert Committee, we divided our patients into three groups according to their HbA1c level: [1] <6%, [2] 6–6.5%, and [3] R6.5%.
Assays Insulin was analyzed by time-resolved flouroimmunoassay using commercial kits (AutoDELFIA; Wallac Oy). Intra-assay coefficient of variation (CV) was 2.1%–3.7%, and interassay CV was 3.4%–4.0%. Serum total T and SHBG were analyzed using a specific RIA after extraction, as previously described (14). In this method, T, dihydrotestosterone, and androstenedione are extracted before applying RIA, and overestimation of T
TABLE 1 Clinical and paraclinical data in patients divided according to HbA1c level (n [ 208). Parameter a
Age (y) Waist (cm)a BMI (kg/m2)a Ferriman-Gallwey total Total T (nmol/L)d SHBG (nmol/L) Free testosterone index Cholesterol (mmol/L)d TG (mmol/L)a HDL (mmol/L) LDL (mmol/L)d Fasting insulin (pmol/L)a AUC insulin (104 pmol/L) HOMA (pmol mmol L2)a BG 0 (mmol/L)d BG 120 (mmol/L)a AUC glucose (102 mmol/L)a
HbA1c <6% (n [ 190) b,c
29 (23–34) 91 (81–102)b 27.9 (24.4–32.8)b,c 9 (5–14) 1.73 (1.26–2.61)c 42 (31–62) 0.032 (0.019–0.051) 4.6 (4.0–5.1)c 1.0 (0.7–1.4)b,c 1.4 (1.2–1.6) 2.7 (2.1–3.3)c 65 (47–111)c 4.7 (3.1–7.9) 14.0 (10.5–24.4)b,c 4.8 (4.6–5.3)b,c 6.5 (5.7–7.6)b,c 9.4 (8.2–10.7)b,c
HbA1c 6–6.5% (n [ 10)
HbA1c R6.5% (n [ 8)
39 (32–42) 113 (108–124) 37.7 (32.1–38.9) 13 (8–19) 1.24 (0.86–2.06) 25 (19–48) 0.029 (0.018–0.054) 4.6 (4.3–7.2) 1.8 (1.5–2.3) 1.2 (1.1–1.7) 2.6 (2.5–4.4) 94 (67–176) 7.2 (5.0–8.4) 25.1 (16.2–54.8) 5.6 (4.6–6.2) 8.6 (7.6–12.5) 12.0 (10.7–15.2)c
38 (36–41) 107 (99–115) 34.8 (33.3–40.9) 14 (9–26) 1.05 (0.89–1.27) 40 (35–50) 0.024 (0.010–0.033) 6.5 (5.7–7.5) 2.7 (2.0–3.5) 1.2 (1.0–1.5) 4.5 (3.5–5.0) 162 (109–285) 6.6 (3.4–7.5) 62.0 (28.9–102.2) 6.3 (5.1–8.8) 12.2 (9.5–15.6) 15.0 (12.3–17.5)
Note: Data presented as median (quartiles). a P< .001, Kruskall-Wallis test between all groups. b P< .05, Mann-Whitney vs. patients with HbA1c 6–6.5%. c P< .05, Mann-Whitney vs. patients with HbA1c R6.5%. d P< .05, Kruskall-Wallis test between all groups. Velling Magnussen. HbA1c and OGTT in PCOS. Fertil Steril 2011.
1276
Velling Magnussen et al.
HbA1c and OGTT in PCOS
Vol. 96, No. 5, November 2011
TABLE 2 Outcome of HbA1c measurement and OGTT in individual patients. HbA1c OGTT
<6%
6–6.5%
R6.5%
Total
Normal IGT T2D Total
148 32 10 190
3 4 3 10
1 0 7 8
152 36 20 208
Note: Patients were divided according to HbA1c and OGTT outcome. Velling Magnussen. HbA1c and OGTT in PCOS. Fertil Steril 2011.
levels is avoided. The method shows a close correlation with the determination of T levels by using mass spectrometry. The intra-assay CV for total T was 8.2%, and for SHBG it was 5.2%. The interassay CV for total T was 13.8%, and for SHBG it was 7.5%. Free T levels were calculated from measurements of total T and SHBG. High-density lipoprotein (HDL) cholesterol and triglyceride (TG) were analyzed by enzymatic colorimetric reactions (Modular P; Roche), and low-density lipoprotein (LDL) cholesterol was calculated using the Friedewald equation. Blood glucose was measured on capillary ear blood using HemoCue. We calculated Homeostasis Model Assessment (HOMA)-r as fasting insulin BG 0/22.5 (15), reference %15 pmol mmol L2. Area under the curve (AUC) for insulin and glucose during OGTT was calculated using the trapezium rule.
Hemoglobin A1c was measured on venous blood with Tosoh G8 chromatograms, which provide direct determination of stable HbA1c. The system is used for in vitro diagnostic measurement of HbA1c in blood specimens. The G8 uses a nonporous column and microcomputer technology to quickly and accurately measure the stable portion of HbA1c as a percentage of the total amount of hemoglobin present in the sample. The G8 provides accurate and precise separation of HbA1c from other hemoglobin fractions. Offline pretreatment is not required, and there is no interference from the labile portion of HbA1c (http:// www.diagnostics.us.tosohbioscience.com/Products/G8þHPLCþAnalyzer/).
Statistical Methods Parameters were not normally distributed, and therefore nonparametric tests were used. Data were described as medians and quartiles. The MannWhitney test was used to compare differences between two groups. The Kruskal-Wallis test was used to compare and test for differences in frequencies between the three groups of sample data. Sensitivity and specificity measures were presented with calculated 95% confidence intervals, and the c2 test was used to test for differences in frequencies between groups. Spearman’s rank correlation analyses were performed to test associations among BG 120, HbA1c, and metabolic and hormonal variables. The variables with the closest association to BG 120 and HbA1c were entered in multiple regression analyses. Multiple regression analyses were performed with BG 120 and HbA1c as the dependent variables and BMI, HOMA, TG, and AUC glucose as explanatory variables. We used commercial software (SPSS 17) for our calculations. A P value of < .05 was considered significant.
RESULTS The clinical and paraclinical characteristics of the 208 included patients are listed in Table 1 according to HbA1c level. Increasing
TABLE 3 Clinical and paraclinical data in individual patients diagnosed with T2D.
No.
HbA1c (%)
Age (y)
BMI (kg/m2)
Cholesterol (mmol/L)
LDL (mmol/L)
Data for the 20 patients diagnosed with T2D according to the OGTT 1 5.0 25 26.2 3.9 2 5.1 31 25.1 4.9 0.6 3 5.1 30 23.8 3.3 1.0 4 5.3 37 37.5 7.6 0.9 5 5.3 18 36.0 4.6 1.2 6 5.4 41 35.2 5.6 1.0 7 5.4 21 33.7 5.6 1.4 8 5.5 26 29.4 4.7 1.8 9 5.8 40 23.5 5.3 0.6 10 5.8 31 35.4 3.8 0.9 11 6.0 40 37.7 4.5 2.5 12 6.1 32 31.2 7.2 1.7 13 6.3 39 38.7 7.2 2.3 14 6.5 38 35.1 7.5 2.6 15 6.6 49 32.8 6.5 3.5 16 6.6 41 33.2 4.2 0.8 17 6.6 37 41.9 6.8 2.7 18 6.9 38 49.0 6.5 3.0 19 7.7 40 33.8 20 9.0 36 37.7 8.0 3.9 Median 5.9 37 34.4 5.6 1.6 Data for the one patient with HbA1c R6.5% and normal glucose tolerance 2 6.6 33 34.4 5.7
TG (mmol/L)
HDL (mmol/L)
BG 0 (mmol/L)
BG 120 (mmol/L)
1.5 1.7 1.5 1.0 1.4 1.7 1.4 1.5 1.0 1.0 1.3 1.7 1.5 1.1 1.5 0.9 1.2
3.2 1 5.7 3.0 3.8 3.4 2.6 3.5 2.4 2.6 5.2 4.4 4.9 4.2 2.4 4.7 3.9 5.3 3.6
4.9 5.5 6.2 4.9 4.6 6.3 6.2 6.4 6.4 6.5 6.5 6.4 5.9 6.1 6.1 6.5 4.7 8.9 8.3 10.7 6.2
11.9 12.8 6.0 11.1 12.0 10.1 8.2 6.9 7.9 7.7 8.8 13.8 12.5 12.1 9.4 9.9 12.3 13.6 20.2 16.3 11.5
1.0 1.4 2.0
1.2
3.8
6.1
Note: Patients are listed according to increasing HbA1c. Horizontal lines represent cut off values for HbA1c. Velling Magnussen. HbA1c and OGTT in PCOS. Fertil Steril 2011.
Fertility and Sterility
1277
HbA1c level was associated with higher age, BMI, HOMA, BG 0, BG 120, and a more adverse lipid profile. Twenty patients were diagnosed with T2D during OGTT: nine by BG 0, six by BG 120, and five by both tests (Table 2). Of 20 patients, 7 (35%) had HbA1c levels R6.5%. Seven of eight patients with HbA1c R6.5% had diabetes during OGTT. Data on individual patients diagnosed with diabetes can be seen in Tables 2 and 3. The sensitivity and the specificity of HbA1c for the diagnosis of diabetes were 35% (range, 14%–55%) and 99% (range, 98%– 100%), respectively, compared with the diagnosis established by OGTT. The sensitivity of HbA1c increased to 50% (range, 28%– 72%) when HbA1c R6% was applied as a cutoff, and the specificity was 96% (range, 93%–99%) (c2 test, P<.001). Fifteen of 20 patients diagnosed with T2D according to the OGTT had a BMI >30 kg/m2. All patients with HbA1c R6.5% had a BMI >30 kg/m2 (Table 3). Impaired glucose tolerance was diagnosed in 32 patients during OGTT. All patients with IGT had HbA1c <6.5% (Table 2). If we defined abnormal OGTT as an outcome with IGT or T2D, the sensitivity of HbA1c R6.5% as a diagnostic marker decreased to 13% (range, 4%–21%), and the specificity was 99% (range, 98%– 100%) (c2 test, P<.001). Hemoglobin A1c showed closer correlation with age, waist, BMI, and lipid profile than BG 120. The BG 120 showed higher correlation with HOMA than HbA1c (Table 4). Hemoglobin A1c and BG 120 showed no significant associations with T (total or free) (data not shown). Multiple regression analyses were performed to test the independent effect of fat mass (BMI), inflammation (TG), insulin sensitivity (HOMA), and glucose levels (AUC glucose) on HbA1c and BG 120 (Table 4). Triglycerides, HOMA, and AUC glucose showed significant associations with HbA1c and BG 120, whereas no significant association was found with BMI.
TABLE 4 A. Association analyses for HbA1c and BG 120: bivariate correlations with clinical/biochemical parameters. Parameter
HbA1c
BG 120
Age (y) Waist (cm) BMI (kg/m2) SHBG (nmol/L) Cholesterol (mmol/L) TG (mmol/L) Fasting insulin (pmol/L) AUC insulin (104 pmol/L) HOMA (pmol mmol l2) BG 0 (mmol/L) BG 120 (mmol/L) AUC glucose (102 mmol/L) HbA1c (%)
0.25a 0.31a 0.40a 0.25 0.25a 0.43a 0.35a 0.28a 0.40a 0.19a 0.37a 0.52a 1a
0.21a 0.17a 0.24a 0.23a 0.21a 0.31a 0.39a 0.29a 0.44a 0.31a 1a 0.77a 0.37a
B. Multiple regression analyses on the independent effects of metabolic risk factors on BG 120 and HbA1c. Dependent variable BMI HbA1c BG 120
NS NS
TG
HOMA
AUC glucose
R2 model
0.29a 0.70b
0.007b 0.03b
0.001a 0.002a
0.46a 0.40a
Note: Spearman correlations between HbA1c and BG 120 and diverse hormonal and metabolic outcomes. Multiple regression analyses were performed with BG 120 and HbA1c as the dependent variables and BMI, TG, HOMA, and AUC glucose as explanatory variables. Data are presented as B value. NS ¼ nonsignificant. a P< .001. b P< .05. Velling Magnussen. HbA1c and OGTT in PCOS. Fertil Steril 2011.
DISCUSSION In the present study we tested HbA1c for the diagnosis of T2D in 208 patients with PCOS. Hemoglobin A1c was a relatively poor diagnostic marker for diabetes, but increasing HbA1c was associated with higher waist, BMI and a more adverse lipid profile, suggesting that HbA1c levels could be applied as a cardiovascular risk marker in PCOS. We are not aware of previous studies that in a similar manner compared the outcome of OGTT to HbA1c levels in study populations with PCOS. Low diagnostic performance of HbA1c for the diagnosis of T2D was, however, reported in non-PCOS populations (16–18). We found that HbA1c had a low sensitivity of 35% for the diagnosis of T2D. However, the specificity of HbA1c for the diagnosis of diabetes was high, and the more severe cases of T2D established by the OGTT were also identified by the HbA1c method. Previous studies evaluated the effects of reducing the HbA1c cutoff to 6.0% (18). In agreement with these studies the sensitivity of HbA1c increased from 35% to 50% when applying HbA1c R6.0% as a cutoff, and the specificity decreased to from 99% to 96%. Impaired glucose tolerance is an important predictor of future T2D in patients with PCOS (19, 20). Diagnosing IGT may therefore be used in the daily clinic to identify patients who may benefit from lifestyle and medical intervention. In the present study all patients with IGT had HbA1c levels <6.5%, and the sensitivity for HbA1c R6.5% as a diagnostic test for abnormal glucose tolerance was only 13%. Hemoglobin A1c therefore seems to be a poor marker of patients at risk for diabetes.
1278
Velling Magnussen et al.
HbA1c and OGTT in PCOS
In the present study all patients had one OGTT performed, and the majority of the 20 patients diagnosed with T2D during OGTT had glucose levels close to the diagnostic cutoff point. Approximately 50% of the patients were diagnosed with diabetes because of elevated fasting glucose levels, whereas BG 120 and HbA1c levels were normal. There are considerable intraindividual variations in fasting and BG 120 levels, which may lead to misclassification of abnormal glucose tolerance (21, 22). Therefore, two diagnostic tests are recommended in asymptomatic patients before the diabetes diagnosis can be established. A second fasting blood glucose may have resulted in a normal value concordant with the measured levels of HbA1c. Selvin et al. (23) found that the accuracy of HbA1c as a diagnostic test was significantly increased when glucose measurements were repeated. Low sensitivity of HbA1c as a diagnostic marker for T2D may be in part explained by the inaccuracy of glucose measurements. Selvin et al. (9) suggested that the usefulness of HbA1c as a diagnostic test should be evaluated against gold standards such as vascular events or mortality. Approximately 75% of women with PCOS are overweight, but a high waist/hip ratio, indicating increased abdominal fat mass, is seen in both normal and overweight patients with PCOS (3). Abdominal obesity is associated with decreased adiponectin and increased chemokines and interleukins, reflecting a high inflammatory state (24, 25). In population-based studies, increased BMI and increased Vol. 96, No. 5, November 2011
waist circumference were associated with a higher mortality due to cardiovascular causes and cancer (26). The close associations observed between HbA1c and BMI, waist, and lipid profile in the present study suggests that HbA1c may be used as an inflammatory marker in PCOS. Hemoglobin A1c and BG 120 levels showed similar associations with measures of insulin resistance. Our results are in agreement with the results from a community-based population of nondiabetic adults (10). In this study, HbA1c and fasting glucose levels were similarly associated with risk of diabetes, but HbA1c was more strongly associated with risk of cardiovascular disease and death from any cause than fasting glucose (10). Different study populations may therefore be identified when HbA1c and glucose levels are applied for the diagnosis of diabetes. Population-based studies found that increasing HbA1c levels within the reference range were associated with cardiovascular disease (10, 27, 28). Subjects with HbA1c levels <5% had the lowest risk for cardiovascular disease (27, 28). As discussed by Khaw et al. (27), future intervention studies are needed to evaluate the effect of decreased HbA1c levels on cardiovascular disease also in nondiabetic populations. Randomized studies in PCOS patients should look at the effect of lifestyle intervention and/or metformin treatment on HbA1c levels and cardiovascular risk markers. Sex hormone–binding globulin is a glycoprotein that binds to T. Low SHBG levels are associated with insulin resistance in PCOS (29), and recent studies suggested that SHBG is an independent risk marker for diabetes (30). We found no association between T levels and HbA1c, but HbA1c levels were inversely correlated
with SHBG levels. Our findings are in agreement with a recent population-based cohort study of 1,629 young adult women followed for 18 years (30). The SHBG levels were negatively associated with subclinical cardiovascular disease, independent of BMI and insulin resistance (30). No significant associations were found between T and cardiovascular outcomes (30). Recent studies found that genetic alterations in the SHBG gene were associated with risk of diabetes and coronary artery disease (31, 32). These findings suggest that high HbA1c and low SHBG levels may be risk markers for cardiovascular disease. The patients included in the present study only represent a subgroup of PCOS patients evaluated during 1997–2010 at our department. Ideally patients should be included consecutively. Hemoglobin A1c was not applied routinely during 1997–2005, and therefore the patients included from this period were not well defined regarding metabolic risk. The current data were based on results of diagnostic screening according to current recommendations. During the study period the recommendations for metabolic screening of PCOS patients were changed, and this may have affected results. The data from the present study are in agreement with studies in non-PCOS populations, but clearly more studies are needed to reproduce our results. In conclusion, HbA1c showed a low sensitivity but high specificity in diagnosing T2D using a cutoff HbA1c value R6.5% compared with the OGTT. Long-term prospective studies are needed to determine whether HbA1c is superior to glucose levels as a cardiovascular risk marker in patients with PCOS.
REFERENCES 1. Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome. Fertil Steril 2004;81:19–25. 2. Glintborg D, Henriksen JE, Andersen M, Hagen C, Hangaard J, Rasmussen PE, et al. Prevalence of endocrine diseases and abnormal glucose tolerance tests in 340 Caucasian premenopausal women with hirsutism as the referral diagnosis. Fertil Steril 2004;82: 1570–9. 3. Glintborg D, Andersen M. An update on the pathogenesis, inflammation, and metabolism in hirsutism and polycystic ovary syndrome. Gynecol Endocrinol 2010;26:281–96. 4. Glintborg D, Hojlund K, Andersen M, Henriksen JE, Beck-Nielsen H, Handberg A. Soluble CD36 and risk markers of insulin resistance and atherosclerosis are elevated in polycystic ovary syndrome and significantly reduced during pioglitazone treatment. Diabetes Care 2008;31:328–34. 5. Wild RA, Carmina E, Diamanti-Kandarakis E, Dokras A, Escobar-Morreale HF, Futterweit W, et al. Assessment of cardiovascular risk and prevention of cardiovascular disease in women with the polycystic ovary syndrome: a consensus statement by the Androgen Excess and Polycystic Ovary Syndrome (AE-PCOS) Society. J Clin Endocrinol Metab 2010;95:2038–49. 6. Shaw LJ, Bairey Merz CN, Azziz R, Stanczyk FZ, Sopko G, Braunstein GD, et al. Postmenopausal women with a history of irregular menses and elevated androgen measurements at high risk for worsening cardiovascular event-free survival: results from the National Institutes of Health–National Heart, Lung, and Blood Institute sponsored Women’s Ischemia Syndrome Evaluation. J Clin Endocrinol Metab 2008;93:1276–84. 7. Salley KE, Wickham EP, Cheang KI, Essah PA, Karjane NW, Nestler JE. Glucose intolerance in poly-
Fertility and Sterility
8.
9. 10.
11.
12.
13. 14.
15.
16.
17.
cystic ovary syndrome—a position statement of the Androgen Excess Society. J Clin Endocrinol Metab 2007;92:4546–56. American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care 2011;34(Suppl 1):S62–9. Selvin E, Brancati FL. A conundrum addressed: the prognostic value of HbA1c. Nat Rev Endocrinol 2011;7:c1. Selvin E, Steffes MW, Zhu H, Matsushita K, Wagenknecht L, Pankow J, et al. Glycated hemoglobin, diabetes, and cardiovascular risk in nondiabetic adults. N Engl J Med 2010;362:800–11. Glintborg D, Mumm H, Hougaard D, Ravn P, Andersen M. Ethnic differences in Rotterdam criteria and metabolic risk factors in a multiethnic group of women with PCOS studied in Denmark. Clin Endocrinol (Oxf) 2010;73:732–8. Ferriman D, Gallwey JD. Clinical assessment of body hair growth in women. J Clin Endocrinol 1961;21: 1440–7. World Health Organization. Definition and diagnosis of diabetes. Geneva: WHO; 2011. Lykkesfeldt G, Bennett P, Lykkesfeldt AE, Micic S, Moller S, Svenstrup B. Abnormal androgen and oestrogen metabolism in men with steroid sulphatase deficiency and recessive X-linked ichthyosis. Clin Endocrinol (Oxf) 1985;23:385–93. Radziuk J. Insulin sensitivity and its measurement: structural commonalities among the methods. J Clin Endocrinol Metab 2000;85:4426–33. Peter A, Fritsche A, Stefan N, Heni M, Haring HU, Schleicher E. Diagnostic value of hemoglobin A1c for type 2 diabetes mellitus in a population at risk. Exp Clin Endocrinol Diabetes 2011;119:234–7. Unwin N, Shaw J, Zimmet P, Alberti KG. Impaired glucose tolerance and impaired fasting glycaemia: the current status on definition and intervention. Diabet Med 2002;19:708–23.
18. Cavagnolli G, Comerlato J, Comerlato C, Renz PB, Gross JL, Camargo JL. HbA(1c) measurement for the diagnosis of diabetes: is it enough? Diabet Med 2011;28:31–5. 19. Ehrmann DA, Barnes RB, Rosenfield RL, Cavaghan MK, Imperial J. Prevalence of impaired glucose tolerance and diabetes in women with polycystic ovary syndrome. Diabetes Care 1999;22:141–6. 20. Legro RS, Gnatuk CL, Kunselman AR, Dunaif A. Changes in glucose tolerance over time in women with polycystic ovary syndrome: a controlled study. J Clin Endocrinol Metab 2005;90:3236–42. 21. Rasmussen SS, Glumer C, Sandbaek A, Lauritzen T, Carstensen B, Borch-Johnsen K. Short-term reproducibility of impaired fasting glycaemia, impaired glucose tolerance and diabetes The ADDITION study, DK. Diabetes Res Clin Pract 2008;80:146–52. 22. Schousboe K, Henriksen JE, Kyvik KO, Sorensen TI, Hyltoft PP. Reproducibility of S-insulin and B-glucose responses in two identical oral glucose tolerance tests. Scand J Clin Lab Invest 2002;62:623–30. 23. Selvin E, Steffes MW, Gregg E, Brancati FL, Coresh J. Performance of A1C for the classification and prediction of diabetes. Diabetes Care 2011; 34:84–9. 24. Glintborg D, Andersen M, Richelsen B, Bruun JM. Plasma monocyte chemoattractant protein-1 (MCP1) and macrophage inflammatory protein-1alpha are increased in patients with polycystic ovary syndrome (PCOS) and associated with adiposity, but unaffected by pioglitazone treatment. Clin Endocrinol (Oxf) 2009;71:652–8. 25. Glintborg D, Andersen M, Hagen C, Frystyk J, Hulstrom V, Flyvbjerg A, et al. Evaluation of metabolic risk markers in polycystic ovary syndrome (PCOS). Adiponectin, ghrelin, leptin and body composition in hirsute PCOS patients and controls. Eur J Endocrinol 2006;155:337–45.
1279
26. Pischon T, Boeing H, Hoffmann K, Bergmann M, Schulze MB, Overvad K, et al. General and abdominal adiposity and risk of death in Europe. N Engl J Med 2008;359:2105–20. 27. Khaw KT, Wareham N, Bingham S, Luben R, Welch A, Day N. Association of hemoglobin A1c with cardiovascular disease and mortality in adults: the European prospective investigation into cancer in Norfolk. Ann Intern Med 2004;141:413–20. 28. Levitan EB, Liu S, Stampfer MJ, Cook NR, Rexrode KM, Ridker PM, et al. Hemoglobin A1c
1280
Velling Magnussen et al.
measured in stored erythrocytes and mortality rate among middle-aged and older women. Diabetologia 2008;51:267–75. 29. Nestler JE. Sex hormone-binding globulin: a marker for hyperinsulinemia and/or insulin resistance? J Clin Endocrinol Metab 1993;76:273–4. 30. Calderon-Margalit R, Schwartz SM, Wellons MF, Lewis CE, Daviglus ML, Schreiner PJ, et al. Prospective association of serum androgens and sex hormone-binding globulin with subclinical cardiovascular disease in young adult women: the
HbA1c and OGTT in PCOS
‘‘Coronary Artery Risk Development in Young Adults’’ women’s study. J Clin Endocrinol Metab 2010;95:4424–31. 31. Alevizaki M, Saltiki K, Xita N, Cimponeriu A, Stamatelopoulos K, Mantzou E, et al. The importance of the (TAAAA)n alleles at the SHBG gene promoter for the severity of coronary artery disease in postmenopausal women. Menopause 2008;15:461–8. 32. Ding EL, Song Y, Manson JE, Hunter DJ, Lee CC, Rifai N, et al. Sex hormone-binding globulin and risk of type 2 diabetes in women and men. N Engl J Med 2009;361:1152–63.
Vol. 96, No. 5, November 2011
Objective: To study hemoglobin A1c (HbA1c) as a tool for diagnosing diabetes and to study HbA1c as a cardiovascular risk marker in patients with polycystic ovary syndrome (PCOS). Design: Retrospective observational study. Setting: Academic tertiary-care medical center. Patient(s): Two hundred eight premenopausal women with PCOS. Intervention(s): Patients underwent clinical evaluation (Ferriman-Gallwey score, body mass index, waist, blood pressure), hormone analyses (T, sex hormone–binding globulin, fasting lipids, insulin, glucose, HbA1c), transvaginal ultrasound, and 2-hour oral glucose tolerance tests (OGTT) measuring capillary blood glucose (BG) at 0 (BG 0) and 120 (BG 120) minutes, insulin, and C-peptide. Main Outcome Measure(s): Results of OGTT, HbA1c values. Result(s): Twenty patients were diagnosed with type 2 diabetes during OGTT. The sensitivity and specificity of HbA1c R6.5% for the diagnosis of diabetes were 35% and 99%, respectively, compared with the diagnosis established by OGTT. Hemoglobin A1c showed closer correlation with waist, body mass index, and lipid profile than BG 120, suggesting that HbA1c could be a cardiovascular risk marker. Conclusion(s): The clinical utility of HbA1c for diagnosing impaired glucose tolerance and type 2 diabetes in PCOS in daily practice is low. Long-term prospective studies are needed to determine whether HbA1c is superior to glucose levels as a cardiovascular risk marker in patients with PCOS. (Fertil Steril 2011;96: 1275–80. 2011 by American Society for Reproductive Medicine.) Key Words: PCOS, HbA1c, diabetes, OGTT, cardiovascular risk, sensitivity, specificity
Polycystic ovary syndrome (PCOS) is a common endocrine condition that is most often diagnosed according to the Rotterdam criteria (1). More than 50% of women with PCOS are insulin resistant, and patients with PCOS have an estimated five- to eightfold increased risk of type 2 diabetes mellitus (T2D) compared with age- and weight-matched controls (2, 3). Polycystic ovary syndrome is characterized by an increased activation of the inflammatory system, with decreased adiponectin levels and increased levels of adipokines, chemokines, and interleukins (3, 4). Abdominal obesity and increased activation of the inflammatory system in PCOS are associated with an increased risk of dyslipidemia and a five- to 10-fold increased risk of diabetes (3). The cardiovascular risk in PCOS is debated, but a recent study confirmed that postmenopausal women with PCOS had more cardiovascular events than controls, and PCOS was associated with angiographic coronary artery disease (5, 6). It is recommended that patients with PCOS are screened for diabetes using an oral glucose tolerance test (OGTT) (1, 7). Fasting plasma glucose (FPG) may be considered in low-risk patients or alternatively where OGTTs are not available (7). In daily practice Received July 4, 2011; revised August 24, 2011; accepted August 25, 2011; published online October 6, 2011. L.V.M. has nothing to disclose. H.M. has nothing to disclose. M.A. has nothing to disclose. D.G. has nothing to disclose. Reprint requests: Dorte Glintborg, M.D., Odense University Hospital, Department of Endocrinology and Metabolism, Kløvervænget 6, 3rd floor, 5000 Odense C, Denmark (E-mail: [email protected]).
0015-0282/$36.00 doi:10.1016/j.fertnstert.2011.08.035
the performance of OGTT may be inconvenient and timeconsuming because the patient has to be seen in a fasting state and therefore often must attend the clinic on two different days. Hemoglobin A1c (HbA1c) is a widely used marker of chronic glycemia and reflects the average blood glucose levels over a 2- to 3-month period (8). Hemoglobin A1c has higher repeatability than fasting glucose and can be assessed in the nonfasting state (8). The use of HbA1c instead of OGTT can avoid the problem of day-to-day variability of glucose values and the need for fasting and preceding dietary preparations. Hemoglobin A1c may, however, be affected by genetic, hematologic, and illness-related factors (8, 9). Recently it was reported from an International Expert Committee composed of members of the European Association for the Study of Diabetes, the International Diabetes Federation, and the American Diabetes Association that values of HbA1c R6.5% can be applied as a cutoff point for diagnosing T2D in asymptomatic patients (8). The cutoff point for HbA1c was largely based on the established association between HbA1c and microvascular disease (8). The guideline also stated that HbA1c levels between 6% and 6.5% should lead to an evaluating OGTT, whereas HbA1c levels of <6% should lead to no further tests (8). Increased HbA1c levels could be used as a marker of cardiovascular risk in nondiabetic adults (10). Hemoglobin A1c may therefore be superior to glucose levels for characterizing long-term risk in nondiabetic adults (10). In the present study we wanted to establish the value of HbA1c for the diagnosis of T2D and as a cardiovascular risk marker in women with PCOS.
Fertility and Sterility Vol. 96, No. 5, November 2011 Copyright ª2011 American Society for Reproductive Medicine, Published by Elsevier Inc.
1275
MATERIALS AND METHODS Inclusion Criteria
Methods
All premenopausal women referred to the outpatient clinic at the Department of Endocrinology during 1997–2010 with a diagnosis of PCOS were included. All patients met at least two of the following Rotterdam criteria: hirsutism, biochemical hyperandrogenism, menstrual disturbances, and/or polycystic ovaries. Patients suspended oral contraceptives and metformin for at least 3 months before evaluation. All included patients had normal thyroid function. During 1997–2005, an OGTT was part of the routine evaluation program at the department, whereas HbA1c measurements were not routinely performed. During 2005–2010, diabetes was excluded by fasting blood glucose and HbA1c measurement, whereas OGTT was performed in high-risk patients for diabetes or the metabolic syndrome. In the present study we included patients for whom data on both HbA1c and OGTT were available. No weight criteria were applied in the study.
Exclusion Criteria Women younger than 15 years were referred to the department of Pediatrics at Odense University Hospital, and patients with a primary complaint of infertility were referred to the local fertility clinic. To exclude patients with serious endocrine disease, s-prolactin and s-17-hydroxyprogesterone were measured. Two measurements of 24-hour urinary cortisol or a short dexamethasone suppression test were performed in the patients with a clinical suspicion of Cushing’s syndrome (2, 11). Patients with anemia and adrenal diseases were excluded to avoid misinterpretation of HbA1c levels. The study was designed as a retrospective observational study. According to our local ethics committee consult and Danish regulations, this study was exempt from institutional review board review. Clinical and paraclinical evaluation of the referred patients were part of the routine evaluation program, and therefore no informed consent from the patients was needed.
Routine evaluation included medical history, clinical examination, transvaginal ultrasound, and fasting blood samples. Hirsutism was assessed by the Ferriman-Gallwey score (12). Waist circumference was measured to the nearest centimeter in a standing position midway between the lower costal margin and the iliac crest. Body mass index (BMI) was calculated as the weight in kilograms divided by the height in meters squared. Transvaginal ultrasound was performed at the Department of Gynecology, Odense University Hospital. Fasting blood samples were drawn during the morning in follicular phase (cycle days 2–8) in patients with a cycle length shorter than 3 months. Patients with cycle length >3 months had the blood samples drawn on a random cycle day. Blood tests included androgens (total T, free T, sex hormone– binding globulin [SHBG], DHEAS, 17-hydroxyprogesterone), LH, FSH, PRL, and lipid profile. An OGTT was performed at 8:00 AM on a random day of the cycle in 208 fasting individuals. P-insulin and capillary blood glucose were measured at baseline and at 30 minutes, 60 minutes, and 120 minutes after oral ingestion of a glucose load containing the equivalent of 75 g anhydrous glucose dissolved in water. According to the World Health Organization diagnosis criteria, T2D was defined by fasting capillary glucose (BG 0) R6.1 mmol/ L or 2-hour glucose (BG 120) R11.1 mmol/L during OGTT. Impaired glucose tolerance (IGT) was defined by BG 0 <6.1 mmol/L and BG 120 7.8– 11.0 mmol/L (13). Because of the above-mentioned recommendation by the International Expert Committee, we divided our patients into three groups according to their HbA1c level: [1] <6%, [2] 6–6.5%, and [3] R6.5%.
Assays Insulin was analyzed by time-resolved flouroimmunoassay using commercial kits (AutoDELFIA; Wallac Oy). Intra-assay coefficient of variation (CV) was 2.1%–3.7%, and interassay CV was 3.4%–4.0%. Serum total T and SHBG were analyzed using a specific RIA after extraction, as previously described (14). In this method, T, dihydrotestosterone, and androstenedione are extracted before applying RIA, and overestimation of T
TABLE 1 Clinical and paraclinical data in patients divided according to HbA1c level (n [ 208). Parameter a
Age (y) Waist (cm)a BMI (kg/m2)a Ferriman-Gallwey total Total T (nmol/L)d SHBG (nmol/L) Free testosterone index Cholesterol (mmol/L)d TG (mmol/L)a HDL (mmol/L) LDL (mmol/L)d Fasting insulin (pmol/L)a AUC insulin (104 pmol/L) HOMA (pmol mmol L2)a BG 0 (mmol/L)d BG 120 (mmol/L)a AUC glucose (102 mmol/L)a
HbA1c <6% (n [ 190) b,c
29 (23–34) 91 (81–102)b 27.9 (24.4–32.8)b,c 9 (5–14) 1.73 (1.26–2.61)c 42 (31–62) 0.032 (0.019–0.051) 4.6 (4.0–5.1)c 1.0 (0.7–1.4)b,c 1.4 (1.2–1.6) 2.7 (2.1–3.3)c 65 (47–111)c 4.7 (3.1–7.9) 14.0 (10.5–24.4)b,c 4.8 (4.6–5.3)b,c 6.5 (5.7–7.6)b,c 9.4 (8.2–10.7)b,c
HbA1c 6–6.5% (n [ 10)
HbA1c R6.5% (n [ 8)
39 (32–42) 113 (108–124) 37.7 (32.1–38.9) 13 (8–19) 1.24 (0.86–2.06) 25 (19–48) 0.029 (0.018–0.054) 4.6 (4.3–7.2) 1.8 (1.5–2.3) 1.2 (1.1–1.7) 2.6 (2.5–4.4) 94 (67–176) 7.2 (5.0–8.4) 25.1 (16.2–54.8) 5.6 (4.6–6.2) 8.6 (7.6–12.5) 12.0 (10.7–15.2)c
38 (36–41) 107 (99–115) 34.8 (33.3–40.9) 14 (9–26) 1.05 (0.89–1.27) 40 (35–50) 0.024 (0.010–0.033) 6.5 (5.7–7.5) 2.7 (2.0–3.5) 1.2 (1.0–1.5) 4.5 (3.5–5.0) 162 (109–285) 6.6 (3.4–7.5) 62.0 (28.9–102.2) 6.3 (5.1–8.8) 12.2 (9.5–15.6) 15.0 (12.3–17.5)
Note: Data presented as median (quartiles). a P< .001, Kruskall-Wallis test between all groups. b P< .05, Mann-Whitney vs. patients with HbA1c 6–6.5%. c P< .05, Mann-Whitney vs. patients with HbA1c R6.5%. d P< .05, Kruskall-Wallis test between all groups. Velling Magnussen. HbA1c and OGTT in PCOS. Fertil Steril 2011.
1276
Velling Magnussen et al.
HbA1c and OGTT in PCOS
Vol. 96, No. 5, November 2011
TABLE 2 Outcome of HbA1c measurement and OGTT in individual patients. HbA1c OGTT
<6%
6–6.5%
R6.5%
Total
Normal IGT T2D Total
148 32 10 190
3 4 3 10
1 0 7 8
152 36 20 208
Note: Patients were divided according to HbA1c and OGTT outcome. Velling Magnussen. HbA1c and OGTT in PCOS. Fertil Steril 2011.
levels is avoided. The method shows a close correlation with the determination of T levels by using mass spectrometry. The intra-assay CV for total T was 8.2%, and for SHBG it was 5.2%. The interassay CV for total T was 13.8%, and for SHBG it was 7.5%. Free T levels were calculated from measurements of total T and SHBG. High-density lipoprotein (HDL) cholesterol and triglyceride (TG) were analyzed by enzymatic colorimetric reactions (Modular P; Roche), and low-density lipoprotein (LDL) cholesterol was calculated using the Friedewald equation. Blood glucose was measured on capillary ear blood using HemoCue. We calculated Homeostasis Model Assessment (HOMA)-r as fasting insulin BG 0/22.5 (15), reference %15 pmol mmol L2. Area under the curve (AUC) for insulin and glucose during OGTT was calculated using the trapezium rule.
Hemoglobin A1c was measured on venous blood with Tosoh G8 chromatograms, which provide direct determination of stable HbA1c. The system is used for in vitro diagnostic measurement of HbA1c in blood specimens. The G8 uses a nonporous column and microcomputer technology to quickly and accurately measure the stable portion of HbA1c as a percentage of the total amount of hemoglobin present in the sample. The G8 provides accurate and precise separation of HbA1c from other hemoglobin fractions. Offline pretreatment is not required, and there is no interference from the labile portion of HbA1c (http:// www.diagnostics.us.tosohbioscience.com/Products/G8þHPLCþAnalyzer/).
Statistical Methods Parameters were not normally distributed, and therefore nonparametric tests were used. Data were described as medians and quartiles. The MannWhitney test was used to compare differences between two groups. The Kruskal-Wallis test was used to compare and test for differences in frequencies between the three groups of sample data. Sensitivity and specificity measures were presented with calculated 95% confidence intervals, and the c2 test was used to test for differences in frequencies between groups. Spearman’s rank correlation analyses were performed to test associations among BG 120, HbA1c, and metabolic and hormonal variables. The variables with the closest association to BG 120 and HbA1c were entered in multiple regression analyses. Multiple regression analyses were performed with BG 120 and HbA1c as the dependent variables and BMI, HOMA, TG, and AUC glucose as explanatory variables. We used commercial software (SPSS 17) for our calculations. A P value of < .05 was considered significant.
RESULTS The clinical and paraclinical characteristics of the 208 included patients are listed in Table 1 according to HbA1c level. Increasing
TABLE 3 Clinical and paraclinical data in individual patients diagnosed with T2D.
No.
HbA1c (%)
Age (y)
BMI (kg/m2)
Cholesterol (mmol/L)
LDL (mmol/L)
Data for the 20 patients diagnosed with T2D according to the OGTT 1 5.0 25 26.2 3.9 2 5.1 31 25.1 4.9 0.6 3 5.1 30 23.8 3.3 1.0 4 5.3 37 37.5 7.6 0.9 5 5.3 18 36.0 4.6 1.2 6 5.4 41 35.2 5.6 1.0 7 5.4 21 33.7 5.6 1.4 8 5.5 26 29.4 4.7 1.8 9 5.8 40 23.5 5.3 0.6 10 5.8 31 35.4 3.8 0.9 11 6.0 40 37.7 4.5 2.5 12 6.1 32 31.2 7.2 1.7 13 6.3 39 38.7 7.2 2.3 14 6.5 38 35.1 7.5 2.6 15 6.6 49 32.8 6.5 3.5 16 6.6 41 33.2 4.2 0.8 17 6.6 37 41.9 6.8 2.7 18 6.9 38 49.0 6.5 3.0 19 7.7 40 33.8 20 9.0 36 37.7 8.0 3.9 Median 5.9 37 34.4 5.6 1.6 Data for the one patient with HbA1c R6.5% and normal glucose tolerance 2 6.6 33 34.4 5.7
TG (mmol/L)
HDL (mmol/L)
BG 0 (mmol/L)
BG 120 (mmol/L)
1.5 1.7 1.5 1.0 1.4 1.7 1.4 1.5 1.0 1.0 1.3 1.7 1.5 1.1 1.5 0.9 1.2
3.2 1 5.7 3.0 3.8 3.4 2.6 3.5 2.4 2.6 5.2 4.4 4.9 4.2 2.4 4.7 3.9 5.3 3.6
4.9 5.5 6.2 4.9 4.6 6.3 6.2 6.4 6.4 6.5 6.5 6.4 5.9 6.1 6.1 6.5 4.7 8.9 8.3 10.7 6.2
11.9 12.8 6.0 11.1 12.0 10.1 8.2 6.9 7.9 7.7 8.8 13.8 12.5 12.1 9.4 9.9 12.3 13.6 20.2 16.3 11.5
1.0 1.4 2.0
1.2
3.8
6.1
Note: Patients are listed according to increasing HbA1c. Horizontal lines represent cut off values for HbA1c. Velling Magnussen. HbA1c and OGTT in PCOS. Fertil Steril 2011.
Fertility and Sterility
1277
HbA1c level was associated with higher age, BMI, HOMA, BG 0, BG 120, and a more adverse lipid profile. Twenty patients were diagnosed with T2D during OGTT: nine by BG 0, six by BG 120, and five by both tests (Table 2). Of 20 patients, 7 (35%) had HbA1c levels R6.5%. Seven of eight patients with HbA1c R6.5% had diabetes during OGTT. Data on individual patients diagnosed with diabetes can be seen in Tables 2 and 3. The sensitivity and the specificity of HbA1c for the diagnosis of diabetes were 35% (range, 14%–55%) and 99% (range, 98%– 100%), respectively, compared with the diagnosis established by OGTT. The sensitivity of HbA1c increased to 50% (range, 28%– 72%) when HbA1c R6% was applied as a cutoff, and the specificity was 96% (range, 93%–99%) (c2 test, P<.001). Fifteen of 20 patients diagnosed with T2D according to the OGTT had a BMI >30 kg/m2. All patients with HbA1c R6.5% had a BMI >30 kg/m2 (Table 3). Impaired glucose tolerance was diagnosed in 32 patients during OGTT. All patients with IGT had HbA1c <6.5% (Table 2). If we defined abnormal OGTT as an outcome with IGT or T2D, the sensitivity of HbA1c R6.5% as a diagnostic marker decreased to 13% (range, 4%–21%), and the specificity was 99% (range, 98%– 100%) (c2 test, P<.001). Hemoglobin A1c showed closer correlation with age, waist, BMI, and lipid profile than BG 120. The BG 120 showed higher correlation with HOMA than HbA1c (Table 4). Hemoglobin A1c and BG 120 showed no significant associations with T (total or free) (data not shown). Multiple regression analyses were performed to test the independent effect of fat mass (BMI), inflammation (TG), insulin sensitivity (HOMA), and glucose levels (AUC glucose) on HbA1c and BG 120 (Table 4). Triglycerides, HOMA, and AUC glucose showed significant associations with HbA1c and BG 120, whereas no significant association was found with BMI.
TABLE 4 A. Association analyses for HbA1c and BG 120: bivariate correlations with clinical/biochemical parameters. Parameter
HbA1c
BG 120
Age (y) Waist (cm) BMI (kg/m2) SHBG (nmol/L) Cholesterol (mmol/L) TG (mmol/L) Fasting insulin (pmol/L) AUC insulin (104 pmol/L) HOMA (pmol mmol l2) BG 0 (mmol/L) BG 120 (mmol/L) AUC glucose (102 mmol/L) HbA1c (%)
0.25a 0.31a 0.40a 0.25 0.25a 0.43a 0.35a 0.28a 0.40a 0.19a 0.37a 0.52a 1a
0.21a 0.17a 0.24a 0.23a 0.21a 0.31a 0.39a 0.29a 0.44a 0.31a 1a 0.77a 0.37a
B. Multiple regression analyses on the independent effects of metabolic risk factors on BG 120 and HbA1c. Dependent variable BMI HbA1c BG 120
NS NS
TG
HOMA
AUC glucose
R2 model
0.29a 0.70b
0.007b 0.03b
0.001a 0.002a
0.46a 0.40a
Note: Spearman correlations between HbA1c and BG 120 and diverse hormonal and metabolic outcomes. Multiple regression analyses were performed with BG 120 and HbA1c as the dependent variables and BMI, TG, HOMA, and AUC glucose as explanatory variables. Data are presented as B value. NS ¼ nonsignificant. a P< .001. b P< .05. Velling Magnussen. HbA1c and OGTT in PCOS. Fertil Steril 2011.
DISCUSSION In the present study we tested HbA1c for the diagnosis of T2D in 208 patients with PCOS. Hemoglobin A1c was a relatively poor diagnostic marker for diabetes, but increasing HbA1c was associated with higher waist, BMI and a more adverse lipid profile, suggesting that HbA1c levels could be applied as a cardiovascular risk marker in PCOS. We are not aware of previous studies that in a similar manner compared the outcome of OGTT to HbA1c levels in study populations with PCOS. Low diagnostic performance of HbA1c for the diagnosis of T2D was, however, reported in non-PCOS populations (16–18). We found that HbA1c had a low sensitivity of 35% for the diagnosis of T2D. However, the specificity of HbA1c for the diagnosis of diabetes was high, and the more severe cases of T2D established by the OGTT were also identified by the HbA1c method. Previous studies evaluated the effects of reducing the HbA1c cutoff to 6.0% (18). In agreement with these studies the sensitivity of HbA1c increased from 35% to 50% when applying HbA1c R6.0% as a cutoff, and the specificity decreased to from 99% to 96%. Impaired glucose tolerance is an important predictor of future T2D in patients with PCOS (19, 20). Diagnosing IGT may therefore be used in the daily clinic to identify patients who may benefit from lifestyle and medical intervention. In the present study all patients with IGT had HbA1c levels <6.5%, and the sensitivity for HbA1c R6.5% as a diagnostic test for abnormal glucose tolerance was only 13%. Hemoglobin A1c therefore seems to be a poor marker of patients at risk for diabetes.
1278
Velling Magnussen et al.
HbA1c and OGTT in PCOS
In the present study all patients had one OGTT performed, and the majority of the 20 patients diagnosed with T2D during OGTT had glucose levels close to the diagnostic cutoff point. Approximately 50% of the patients were diagnosed with diabetes because of elevated fasting glucose levels, whereas BG 120 and HbA1c levels were normal. There are considerable intraindividual variations in fasting and BG 120 levels, which may lead to misclassification of abnormal glucose tolerance (21, 22). Therefore, two diagnostic tests are recommended in asymptomatic patients before the diabetes diagnosis can be established. A second fasting blood glucose may have resulted in a normal value concordant with the measured levels of HbA1c. Selvin et al. (23) found that the accuracy of HbA1c as a diagnostic test was significantly increased when glucose measurements were repeated. Low sensitivity of HbA1c as a diagnostic marker for T2D may be in part explained by the inaccuracy of glucose measurements. Selvin et al. (9) suggested that the usefulness of HbA1c as a diagnostic test should be evaluated against gold standards such as vascular events or mortality. Approximately 75% of women with PCOS are overweight, but a high waist/hip ratio, indicating increased abdominal fat mass, is seen in both normal and overweight patients with PCOS (3). Abdominal obesity is associated with decreased adiponectin and increased chemokines and interleukins, reflecting a high inflammatory state (24, 25). In population-based studies, increased BMI and increased Vol. 96, No. 5, November 2011
waist circumference were associated with a higher mortality due to cardiovascular causes and cancer (26). The close associations observed between HbA1c and BMI, waist, and lipid profile in the present study suggests that HbA1c may be used as an inflammatory marker in PCOS. Hemoglobin A1c and BG 120 levels showed similar associations with measures of insulin resistance. Our results are in agreement with the results from a community-based population of nondiabetic adults (10). In this study, HbA1c and fasting glucose levels were similarly associated with risk of diabetes, but HbA1c was more strongly associated with risk of cardiovascular disease and death from any cause than fasting glucose (10). Different study populations may therefore be identified when HbA1c and glucose levels are applied for the diagnosis of diabetes. Population-based studies found that increasing HbA1c levels within the reference range were associated with cardiovascular disease (10, 27, 28). Subjects with HbA1c levels <5% had the lowest risk for cardiovascular disease (27, 28). As discussed by Khaw et al. (27), future intervention studies are needed to evaluate the effect of decreased HbA1c levels on cardiovascular disease also in nondiabetic populations. Randomized studies in PCOS patients should look at the effect of lifestyle intervention and/or metformin treatment on HbA1c levels and cardiovascular risk markers. Sex hormone–binding globulin is a glycoprotein that binds to T. Low SHBG levels are associated with insulin resistance in PCOS (29), and recent studies suggested that SHBG is an independent risk marker for diabetes (30). We found no association between T levels and HbA1c, but HbA1c levels were inversely correlated
with SHBG levels. Our findings are in agreement with a recent population-based cohort study of 1,629 young adult women followed for 18 years (30). The SHBG levels were negatively associated with subclinical cardiovascular disease, independent of BMI and insulin resistance (30). No significant associations were found between T and cardiovascular outcomes (30). Recent studies found that genetic alterations in the SHBG gene were associated with risk of diabetes and coronary artery disease (31, 32). These findings suggest that high HbA1c and low SHBG levels may be risk markers for cardiovascular disease. The patients included in the present study only represent a subgroup of PCOS patients evaluated during 1997–2010 at our department. Ideally patients should be included consecutively. Hemoglobin A1c was not applied routinely during 1997–2005, and therefore the patients included from this period were not well defined regarding metabolic risk. The current data were based on results of diagnostic screening according to current recommendations. During the study period the recommendations for metabolic screening of PCOS patients were changed, and this may have affected results. The data from the present study are in agreement with studies in non-PCOS populations, but clearly more studies are needed to reproduce our results. In conclusion, HbA1c showed a low sensitivity but high specificity in diagnosing T2D using a cutoff HbA1c value R6.5% compared with the OGTT. Long-term prospective studies are needed to determine whether HbA1c is superior to glucose levels as a cardiovascular risk marker in patients with PCOS.
REFERENCES 1. Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome. Fertil Steril 2004;81:19–25. 2. Glintborg D, Henriksen JE, Andersen M, Hagen C, Hangaard J, Rasmussen PE, et al. Prevalence of endocrine diseases and abnormal glucose tolerance tests in 340 Caucasian premenopausal women with hirsutism as the referral diagnosis. Fertil Steril 2004;82: 1570–9. 3. Glintborg D, Andersen M. An update on the pathogenesis, inflammation, and metabolism in hirsutism and polycystic ovary syndrome. Gynecol Endocrinol 2010;26:281–96. 4. Glintborg D, Hojlund K, Andersen M, Henriksen JE, Beck-Nielsen H, Handberg A. Soluble CD36 and risk markers of insulin resistance and atherosclerosis are elevated in polycystic ovary syndrome and significantly reduced during pioglitazone treatment. Diabetes Care 2008;31:328–34. 5. Wild RA, Carmina E, Diamanti-Kandarakis E, Dokras A, Escobar-Morreale HF, Futterweit W, et al. Assessment of cardiovascular risk and prevention of cardiovascular disease in women with the polycystic ovary syndrome: a consensus statement by the Androgen Excess and Polycystic Ovary Syndrome (AE-PCOS) Society. J Clin Endocrinol Metab 2010;95:2038–49. 6. Shaw LJ, Bairey Merz CN, Azziz R, Stanczyk FZ, Sopko G, Braunstein GD, et al. Postmenopausal women with a history of irregular menses and elevated androgen measurements at high risk for worsening cardiovascular event-free survival: results from the National Institutes of Health–National Heart, Lung, and Blood Institute sponsored Women’s Ischemia Syndrome Evaluation. J Clin Endocrinol Metab 2008;93:1276–84. 7. Salley KE, Wickham EP, Cheang KI, Essah PA, Karjane NW, Nestler JE. Glucose intolerance in poly-
Fertility and Sterility
8.
9. 10.
11.
12.
13. 14.
15.
16.
17.
cystic ovary syndrome—a position statement of the Androgen Excess Society. J Clin Endocrinol Metab 2007;92:4546–56. American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care 2011;34(Suppl 1):S62–9. Selvin E, Brancati FL. A conundrum addressed: the prognostic value of HbA1c. Nat Rev Endocrinol 2011;7:c1. Selvin E, Steffes MW, Zhu H, Matsushita K, Wagenknecht L, Pankow J, et al. Glycated hemoglobin, diabetes, and cardiovascular risk in nondiabetic adults. N Engl J Med 2010;362:800–11. Glintborg D, Mumm H, Hougaard D, Ravn P, Andersen M. Ethnic differences in Rotterdam criteria and metabolic risk factors in a multiethnic group of women with PCOS studied in Denmark. Clin Endocrinol (Oxf) 2010;73:732–8. Ferriman D, Gallwey JD. Clinical assessment of body hair growth in women. J Clin Endocrinol 1961;21: 1440–7. World Health Organization. Definition and diagnosis of diabetes. Geneva: WHO; 2011. Lykkesfeldt G, Bennett P, Lykkesfeldt AE, Micic S, Moller S, Svenstrup B. Abnormal androgen and oestrogen metabolism in men with steroid sulphatase deficiency and recessive X-linked ichthyosis. Clin Endocrinol (Oxf) 1985;23:385–93. Radziuk J. Insulin sensitivity and its measurement: structural commonalities among the methods. J Clin Endocrinol Metab 2000;85:4426–33. Peter A, Fritsche A, Stefan N, Heni M, Haring HU, Schleicher E. Diagnostic value of hemoglobin A1c for type 2 diabetes mellitus in a population at risk. Exp Clin Endocrinol Diabetes 2011;119:234–7. Unwin N, Shaw J, Zimmet P, Alberti KG. Impaired glucose tolerance and impaired fasting glycaemia: the current status on definition and intervention. Diabet Med 2002;19:708–23.
18. Cavagnolli G, Comerlato J, Comerlato C, Renz PB, Gross JL, Camargo JL. HbA(1c) measurement for the diagnosis of diabetes: is it enough? Diabet Med 2011;28:31–5. 19. Ehrmann DA, Barnes RB, Rosenfield RL, Cavaghan MK, Imperial J. Prevalence of impaired glucose tolerance and diabetes in women with polycystic ovary syndrome. Diabetes Care 1999;22:141–6. 20. Legro RS, Gnatuk CL, Kunselman AR, Dunaif A. Changes in glucose tolerance over time in women with polycystic ovary syndrome: a controlled study. J Clin Endocrinol Metab 2005;90:3236–42. 21. Rasmussen SS, Glumer C, Sandbaek A, Lauritzen T, Carstensen B, Borch-Johnsen K. Short-term reproducibility of impaired fasting glycaemia, impaired glucose tolerance and diabetes The ADDITION study, DK. Diabetes Res Clin Pract 2008;80:146–52. 22. Schousboe K, Henriksen JE, Kyvik KO, Sorensen TI, Hyltoft PP. Reproducibility of S-insulin and B-glucose responses in two identical oral glucose tolerance tests. Scand J Clin Lab Invest 2002;62:623–30. 23. Selvin E, Steffes MW, Gregg E, Brancati FL, Coresh J. Performance of A1C for the classification and prediction of diabetes. Diabetes Care 2011; 34:84–9. 24. Glintborg D, Andersen M, Richelsen B, Bruun JM. Plasma monocyte chemoattractant protein-1 (MCP1) and macrophage inflammatory protein-1alpha are increased in patients with polycystic ovary syndrome (PCOS) and associated with adiposity, but unaffected by pioglitazone treatment. Clin Endocrinol (Oxf) 2009;71:652–8. 25. Glintborg D, Andersen M, Hagen C, Frystyk J, Hulstrom V, Flyvbjerg A, et al. Evaluation of metabolic risk markers in polycystic ovary syndrome (PCOS). Adiponectin, ghrelin, leptin and body composition in hirsute PCOS patients and controls. Eur J Endocrinol 2006;155:337–45.
1279
26. Pischon T, Boeing H, Hoffmann K, Bergmann M, Schulze MB, Overvad K, et al. General and abdominal adiposity and risk of death in Europe. N Engl J Med 2008;359:2105–20. 27. Khaw KT, Wareham N, Bingham S, Luben R, Welch A, Day N. Association of hemoglobin A1c with cardiovascular disease and mortality in adults: the European prospective investigation into cancer in Norfolk. Ann Intern Med 2004;141:413–20. 28. Levitan EB, Liu S, Stampfer MJ, Cook NR, Rexrode KM, Ridker PM, et al. Hemoglobin A1c
1280
Velling Magnussen et al.
measured in stored erythrocytes and mortality rate among middle-aged and older women. Diabetologia 2008;51:267–75. 29. Nestler JE. Sex hormone-binding globulin: a marker for hyperinsulinemia and/or insulin resistance? J Clin Endocrinol Metab 1993;76:273–4. 30. Calderon-Margalit R, Schwartz SM, Wellons MF, Lewis CE, Daviglus ML, Schreiner PJ, et al. Prospective association of serum androgens and sex hormone-binding globulin with subclinical cardiovascular disease in young adult women: the
HbA1c and OGTT in PCOS
‘‘Coronary Artery Risk Development in Young Adults’’ women’s study. J Clin Endocrinol Metab 2010;95:4424–31. 31. Alevizaki M, Saltiki K, Xita N, Cimponeriu A, Stamatelopoulos K, Mantzou E, et al. The importance of the (TAAAA)n alleles at the SHBG gene promoter for the severity of coronary artery disease in postmenopausal women. Menopause 2008;15:461–8. 32. Ding EL, Song Y, Manson JE, Hunter DJ, Lee CC, Rifai N, et al. Sex hormone-binding globulin and risk of type 2 diabetes in women and men. N Engl J Med 2009;361:1152–63.
Vol. 96, No. 5, November 2011