SEARCH

SEARCH BY CITATION

Keywords:

  • Gestational diabetes;
  • postpartum reclassification

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. References
  12. Supporting Information

Please cite this paper as: Megia A, Näf S, Herranz L, Serrat N, Yañez R, Simón I, Vendrell J. The usefulness of HbA1c in postpartum reclassification of gestational diabetes. BJOG 2012;119:891–894.

To investigate the role of HbA1c in postpartum reclassification of gestational diabetes (GDM) we studied 364 women with GDM attending the postpartum reclassification assessment of their glucose tolerance status. A 75-g oral glucose tolerance test (OGTT) was performed and HbA1c was determined. Diabetes was diagnosed in 12 (3.3%), 7 (1.9%) and 2 (0.6%) women according to the fasting plasma glucose (FPG) and/or the 2-hour OGTT, the FPG alone and HbA1c levels, respectively. The sensitivity and specificity for HbA1c to diagnose diabetes was 16.7% and 100%, respectively, for FPG and OGTT criteria. The combination of a cutoff value of 5.5% for HbA1c and FPG allowed us to identify 95.1% of women with any kind of glucose intolerance. We conclude that in the early postpartum period, the cutoff of 6.5% for HbA1c alone has low sensitivity for the diagnosis of diabetes compared with OGTT, but the combination of FPG and HbA1c at a lower cutoff value is very useful to identify women with any kind of glucose intolerance.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. References
  12. Supporting Information

Gestational diabetes mellitus (GDM) affects up to 14% of all pregnancies and identifies a group of women at risk for the development of type II diabetes, impaired glucose tolerance and metabolic syndrome in the future. Recommendations of the Fifth Workshop Conference on GDM1 are that women with GDM should undergo postpartum glucose tolerance testing with an oral glucose tolerance test (OGTT). The rationale for this recommendation is not only to identify women with apparent diabetes but also to identify women with impaired glucose tolerance in whom diabetes can be delayed or prevented. Unfortunately, the rates of postpartum evaluation are low.2 The discontinuation of care after delivery as a result of mothers underestimating their diabetes risk, the stress of adapting to the challenge of caring for a young baby, and the time needed in the re-evaluation testing are some of the reasons implicated. Hence, the adoption of easier strategies that do not require fasting or more than a single blood draw could help to increase the rate of postpartum testing.

Recently, HbA1c has been proposed as a diagnostic tool to identify people with undiagnosed diabetes, or who are at risk of diabetes.3 The HbA1c assay is in most cases universally standardised and traceable to the Diabetes Control and Complications Trial assay. HbA1c does not require fasting, it has better pre-analytical stability and reflects long-term glycaemic exposure better than current diagnostic tests based on fasting or post-load glucose measures. Nevertheless, the diagnostic rate in the general and at-risk populations has been shown to be lower than fasting plasma glucose (FPG) or OGTT. In previous reports, HbA1c has been shown to be less cost effective in the long-term follow up than OGTT for the diagnosis of diabetes mellitus in women with GDM.4 In a recent report, the same group assessed the usefulness of HbA1c in the postpartum evaluation of a small group of women over a longer period of time.5 But as far as we are aware, there are no studies assessing the usefulness of HbA1c in the early postpartum reclassification of GDM. The aim of this study was to analyse whether the use of HbA1c may be useful in the postpartum reclassification of women with GDM in a large cohort of women.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. References
  12. Supporting Information

This analysis was conducted in the setting of an ongoing observational study at the University Hospital Joan XXIII in Tarragona, Spain. Women with GDM are recruited during pregnancy and are encouraged to return for postpartum re-evaluation. The study protocol was approved by our centre’s Research Ethics Board.

Women were selected from an ongoing database in which data for all women are entered prospectively. Before inclusion, all participants provide their written informed consent. At our centre, GDM is diagnosed following the National Diabetes Data Group criteria. During pregnancy, all women are advised to follow a diet with 40% carbohydrates, and if targets of fasting glucose <5.3 mmol/l (<95 mg/dl) and 1 hour postprandial glucose <7.3 mmol/l (<140 mg/dl) are not attained, insulin therapy is added. Neither insulin therapy nor dietary counselling is continued postpartum. In this analysis, we included women that returned for the post-delivery study visit within the first year postpartum, between January 2006 and March 2011, and had HbA1c measured at the time of the postpartum 2-hour 75-g OGTT. Results of the OGTT were evaluated according to the 1999 World Health Organization criteria.

Blood for HbA1c was collected in ethylenediaminetetraacetic acid vials and estimated by high-performance liquid chromatography-based ion exchange chromatography (ADAMS-A1c HA-8160; Menarini Diagnostics, Firenze, Italy). This method corrects the most common haemoglobin (Hb) traits. Labile HbA1c, carbamylated Hb and acetylated Hb are eluted separately from the stable HbA1c peak. The HbA1c assay is standardised and traceable to the Diabetes Control and Complications Trial assay. The intra-assay coefficients for people without and with diabetes are 1.5 and 0.5%, respectively.

Data are presented as numbers and proportions for categorical variables or as median and interquartile range for continuous variables. Accordingly, a Pearson chi-square test or nonparametric tests were used to test for differences between groups. Spearman’s correlation was used to analyse the relationship between glucose and HbA1c values. The receiver operating characteristic (ROC) curve was used to analyse the performance of the HbA1c test to diagnose diabetes or any kind of glucose intolerance by OGTT. A significance level of P < 0.05 was adopted.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. References
  12. Supporting Information

Three hundred and sixty-four women had postpartum OGTT and HbA1c performed simultaneously. They were white European (91.5%), Arabic (5.5%), Hispanic (1.6%) and others (1.4%). Among them, according to the OGTT results, 263 women (72.3%) had normal glucose tolerance, 89 (24.5%) had pre-diabetes (impaired fasting glucose, impaired glucose tolerance or both) and 12 (3.3%) had diabetes. Seven of the women with diabetes fulfilled the FPG criteria (2.7%) and two had an HbA1c of 6.5% or more (0.6%). No difference in the time of postpartum evaluation [3 (2–5) versus 3 (3–5) months], age [33 (30–36) versus 36 (30.5–39.75) years] or breastfeeding (65.8 versus 77.8%) was observed between women without and with diabetes, respectively. Women with diabetes compared with those without had higher HbA1c levels [5.6 (5.4–6.3) versus 5.1 (4.8–5.3) %; P < 0.001], pre-gravid and postpartum body mass index values [30.1 (26.8–32.7) versus 24.8 (22.2–25.6) kg/m2; P < 0.01 and 29.2 (26.4–33.5) versus 25.7 (22.7–30.2) kg/m2; P < 0.05, respectively]. All women diagnosed with diabetes had been treated with insulin during pregnancy compared with 47% in the non-diabetic group (P < 0.001).

Correlations were r = 0.296 (P < 0.001) for HbA1c versus FPG and r = 0.196 (P < 0.001) for HbA1c versus 2-hour glucose. We also performed the same analysis separating the women into three groups according to the time of postpartum evaluation: group 1 (from 6 weeks to 3 months after delivery, n = 260), group 2 (from 4 to 6 months postpartum, n = 69), and group 3 (from 7 months to 1 year, n = 35). The correlation coefficients for HbA1c versus FPG and HbA1c versus 2-hour OGTT were r = 0.281 (P < 0.001) and r = 0.162 (< 0.05), respectively, for group 1; r = 0.399 (P = 0.004) and r = 0.160 (P = 0.263), respectively, for group 2; and r = 0.243 (P = 0.232) and r = 0.054 (P = 0.795), respectively, for group 3.

To determine differences in HbA1c levels across the first year postpartum, we compared mean HbA1c levels observed in each of the three groups considered, and a significantly lower HbA1c value was observed in group 2 compared with group 1 [HbA1c in group 1: 5.19 (4.94–5.39) %, in group 2: 4.89 (4.79–5.30) % and in group 3: 5.18 (4.79–5.39) %; P = 0.025]. We also evaluated whether there were any differences between HbA1c in the third trimester of pregnancy and in the postpartum period, and we found that both measures were correlated (r = 0.604; P < 0.001) and a significant increase in HbA1c was observed in the postpartum period [4.89 (4.60–5.19) versus 5.09 (4.80–5.3); P < 0.001].

Sensitivity and specificity of HbA1c for diabetes diagnosis for a cutoff of 6.5% or more were 16.7% and 100%, respectively. When we used the FPG criteria alone to diagnose diabetes, sensitivity increased to 28.6% and specificity remained unchanged. To analyse whether the information of a single blood draw could improve sensitivity against FPG alone, we tested the combination of FPG and HbA1c. No increase in sensitivity was observed. All the women with HbA1c over 5.9% had diabetes. Table 1 shows sensitivity and specificity of HbA1c at various cutoff levels for the OGTT or FPG criteria. The area under the ROC curve for diagnosis of diabetes was 0.870 (see Figure S1A).

Table 1.   Sensitivity and specificity of HbA1C at various cutoff levels in the postartum reclassification assessment attending to the OGTT or FPG criteria
HbA1C cutoff level (%)OGTT criteriaFPG criteria
Sensitivity (%)Specificity (%)Sensitivity (%)Specificity (%)
5.391.6772.4410077.03
5.475.0082.6785.7182.07
5.566.6788.0785.7182.07
5.641.6792.0557.1491.88
5.741.6796.3157.1496.08
5.841.6798.8657.1497.76
5.933.3310057.1498.60
6.025.0010042.8699.72
6.516.6710028.57100

The sensitivity and specificity of HbA1c ≥5.7% to diagnose any kind of glucose intolerance was 13.5% and 97.3%, respectively. The area under the ROC curve for HbA1c for any kind of glucose intolerance was 0.674 (see Figure S1B) with a sensitivity of 44.6% and a specificity of 78.8% for a cutoff value of 5.3%. The use of this threshold to select women for the OGTT might reduce the number of women for testing to about 30%, but 15.4% of women with some kind of glucose intolerance will be overlooked. To reduce the number of women with some kind of glucose intolerance lost without increasing the number of women for testing, we combined FPG and HbA1c with a cutoff value of 5.5%. To select women for the OGTT the method had a sensitivity of 82.2% and a specificity of 92.0% to identify women with pre-diabetes reducing the number of women for testing to 28.5%.

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. References
  12. Supporting Information

In this prospective study, we showed that in early postpartum evaluation the cutoff values of 6.5% and 5.7% for HbA1c have low sensitivity to diagnose diabetes and pre-diabetes, respectively, despite their high specificity, but FPG and HbA1c with a lower cutoff value can be useful to select women for OGTT. The low agreement between HbA1c and 2-hour OGTT has also been observed in other high-risk populations.6 In a recent report, Kim et al.5 compared HbA1c with the results of FPG and/or OGTT in 54 women from 6 weeks after delivery to more than 2 years after delivery. In contrast to our results, they reported a higher sensitivity and specificity for HbA1c ≥5.7% to detect glucose intolerance in the postpartum evaluation. There are some differences between both studies that must be pointed out. In the study by Kim et al., the postpartum evaluation was performed over a longer period of time, the sample studied is smaller, women were heavier and sedentary, and a higher percentage of women belonging to high-risk ethnic groups was included. These differences could be responsible for the higher rate of diabetes, impaired glucose tolerance and impaired fasting glucose found in their study compared with ours.

In our population we observed significant correlation coefficients between FPG and HbA1c among the women evaluated up to 6 months postpartum, and this association was lost among women evaluated from 7 to 12 months after delivery. But in accordance with our data, Kim et al.5 also found that the relationship between fasting plasma glucose and HbA1c period differ in the different periods studied.

More interesting is the finding that HbA1c can be useful to detect women with any degree of glucose intolerance alone or in combination with FPG, and it might be used to select women candidates for the OGTT. When HbA1c ≥5.5% with plasma glucose ≥5.5 mmol/l, information that could be obtained in a single blood draw are used to select women for OGTT, the number of women for testing would decrease by more than 70%, overlooking <4.9% of women with any kind of glucose intolerance.

Little is known about the change of HbA1c levels during the first year postpartum. We observed higher HbA1c in the postpartum period than in the third trimester of pregnancy, which could be because of the lower glucose levels and the increased turnover of red blood cells at the end of pregnancy. The blood loss at delivery, a factor that could contribute to falsely decreased HbA1c values in the early postpartum period, seems not to be a relevant factor,3 at least in our series. In fact, we found lower mean HbA1c concentrations in the group of women evaluated between the fourth and sixth months postpartum compared with those evaluated earlier.

Taking these results into account and considering that early detection and treatment of diabetes are of utmost importance in preventing diabetic complications and improving short-term and long-term outcomes in this subset of women, we believe that the recently introduced HbA1c criterion cannot be considered a good reclassification test for GDM in the early postpartum evaluation, but lowering the cutoff for HbA1c to 5.5% with the combination of FPG to select women for the OGTT is a useful tool to select women candidates for OGTT that allows us to decrease the number of candidates for OGTT.

However, longitudinal studies are needed to determine whether postpartum HbA1c could be considered a good predictor of diabetes mellitus, cardiovascular disease or metabolic syndrome in women with previous GDM.

Contribution to authorship

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. References
  12. Supporting Information

AM conducted the study, data collection, analysis and wrote the manuscript. SN contributed to the data collection and reviewed the manuscript. LH designed the study. REY contributed with data collection. NS contributed with data analysis. IS and JV contributed to writing the manuscript.

Details of ethics approval

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. References
  12. Supporting Information

The study protocol was approved by the Joan XXIII Hospital Research Ethics Board (p01/06/2009; 22p/2009). Informed consent was obtained from all the participants at inclusion.

Funding

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. References
  12. Supporting Information

This work was supported by grants FIS PS09/02152, Instituto de Salud Carlos III, cofunded by Fondos FEDER and CIBER de Diabetes y Enfermedades Metabólicas asociadas CIBERDEM (CB07/08/0012). AM is grateful to the IISPV for the 2010 grant: ‘Programa d’ajuts per el desenvolupament de la recerca’. CIBERDEM de Diabetes y Enfermedades Metabólicas Asociadas (CB07708/0012) is an initiative of the Instituto de Salud Carlos III.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. References
  12. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. References
  12. Supporting Information

Figure S1. ROC curves for HbA1c used for detection of diabetes (A) and any kind of glucose intolerance from an OGTT (B).

FilenameFormatSizeDescription
BJO_3325_sm_FigS1.pdf99KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.