Introduction
Gestational diabetes mellitus (GDM) is associated with an increased risk of pregnancy-related complications for both mother and child [
1,
2]. International guidelines recommend active screening for GDM since many of these risks can be reduced by its detection and management [
3,
4]. However, these guidelines lack uniformity in terms of their diagnostic thresholds.
In 2010, the International Association of the Diabetes and Pregnancy Study Groups (IADPSG) proposed more stringent thresholds for diagnosing GDM that were based on the results of the international prospective Hyperglycaemia and Adverse Pregnancy Outcomes (HAPO) study [
5,
6]. This study demonstrated a linear association between maternal glucose levels at fasting and during an OGTT and the risk of adverse pregnancy outcomes such as increased birthweight, primary Caesarean section and neonatal hypoglycaemia [
6]. The IADPSG diagnostic criteria (fasting plasma glucose ≥5.1 mmol/l; and/or 1 h plasma glucose ≥10.0 mmol/l; and/or 2 h plasma glucose (2HG) ≥8.5 mmol/l) have now been adopted by many guideline committees and expert groups, including the WHO in 2013 [
5,
7].
However, evidence in support of applying the new criteria to diagnose GDM to improve pregnancy outcomes is limited. The optimal glucose thresholds to define GDM remain uncertain and international consensus has not yet been reached [
8,
9]. Applying the new criteria gives rise to more women being diagnosed with GDM and the resulting cost increases and medicalisation of pregnancy are causes for concern for healthcare managers and caregivers [
10,
11]. Studies into clinical outcomes and cost-effectiveness analyses are required to better appraise the value of these new glucose thresholds. In the Netherlands, the new WHO 2013 criteria have not yet been endorsed. In their 2010 guideline ‘Diabetes and Pregnancy’, the Dutch Society of Obstetrics and Gynaecology instead recommends using the WHO 1999 criteria to diagnose GDM (fasting glucose ≥7.0 mmol/l and/or 2HG ≥7.8 mmol/l) [
12,
13]. When compared with the new WHO 2013 criteria, these use a much higher threshold for fasting glucose and a lower threshold for 2HG.
The consequences of adopting the WHO 2013 thresholds need to be evaluated in order to answer the following crucial questions: Do the additional women diagnosed with GDM using the new WHO 2013 fasting glucose criteria (fasting glucose ≥5.1 but ≤6.9 mmol/l) indeed have unfavourable pregnancy outcomes? What are the pregnancy outcomes of the women who would be missed owing to the higher 2HG threshold using the WHO 2013 criteria (i.e. women with 2HG ≥7.8 but ≤8.4 mmol/l)?
The aim of this study was therefore to evaluate the possible impact on the number of GDM diagnoses and pregnancy outcomes when applying the new WHO 2013 criteria instead of the older WHO 1999 criteria.
Results
Pregnancy outcomes
Maternal and neonatal outcomes according to the different GDM classification groups are given in Table
3. Compared with women in the NGT group, women classified as having GDM (using the WHO 2013 criteria or WHO 1999 criteria) were more likely to develop gestational hypertension or preeclampsia and to have had a planned Caesarean section delivery or induced labour.
Table 3
Pregnancy outcomes according to the GDM classification groups
N
| 29,562 | 4431 | 2851 | 913 | 1346 | 667 | 234 |
Treated for GDM, n
| | | 0 | 913 | 679 | 0 | 234 |
Maternal | | | | | | | |
Gestational hypertension, n (%) | | 4427 | 139 (4.9) | 62 (6.8)* | 98 (7.3)** | 52 (7.8)** | 16 (6.9) |
Preeclampsia, n (%) | | 4427 | 41 (1.4) | 28 (3.1)** | 35 (2.6)** | 12 (1.8) | 5 (2.1) |
Induction of labour, n (%) | | 4405 | 793 (28.0) | 587 (64.3)*** | 670 (50.0)*** | 230 (34.8)** | 147 (62.8)*** |
Mode of delivery, n (%) | | 4410 | | | | | |
Vaginal | | | 2051 (72.3) | 618 (67.7)** | 904 (67.4)** | 451 (68.1)* | 165 (70.5) |
Emergency CS | | | 327 (11.5) | 116 (12.7) | 177 (13.2) | 89 (13.4) | 28 (12.0) |
Planned CS | | | 185 (6.5) | 103 (11.3)*** | 150 (11.2)*** | 68 (10.3)** | 21 (9.0) |
Instrumental | | | 272 (9.6) | 76 (8.3) | 110 (8.2) | 54 (8.2) | 20 (8.5) |
Gestational age at delivery (weeks) | | 4431 | 39.7 (38.7–40.6) | 38.3 (38.0–39.0)*** | 38.7 (38.0–39.9)*** | 39.6 (38.3–40.4)*** | 38.6 (38.1–39.4)*** |
Neonatal | | | | | | | |
LGA, n (%) | 3246 (11.0) | 4430 | 514 (18.0) | 167 (18.3) | 271 (20.1) | 140 (21.0) | 36 (15.4) |
Macrosomia, n (%) | 4275 (14.5) | 4431 | 595 (20.9) | 108 (11.8)*** | 226 (16.8)** | 148 (22.2) | 30 (12.8)** |
Small for gestational age, n (%) | 2364 (8.0) | 4430 | 195 (6.8) | 36 (3.9)** | 69 (5.1)* | 38 (5.7) | 5 (2.1)** |
Birthweight (g) | | 4431 | 3544 ± 579 | 3391 ± 550*** | 3477 ± 590** | 3580 ± 596 | 3437 ± 498** |
Birth trauma, n (%) | | 4420 | 64 (2.3) | 27 (3.0) | 43 (3.2) | 20 (3.0) | 4 (1.7) |
Hypoglycaemia, n (%)a
| | 4418 | NA | 38 (4.2)*** | NA | NA | 4 (1.7) |
Hyperbilirubinaemia, n (%)a
| | 4418 | NA | 24 (2.6)** | NA | NA | 5 (2.1) |
Stillbirth, n (%) | | 4431 | 10 (0.4) | 2 (0.2) | 6 (0.4) | 4 (0.6) | 0 |
Preterm delivery, n (%) | | 4431 | 146 (5.1) | 57 (6.2) | 92 (6.8)* | 46 (6.9) | 11 (4.7) |
Respiratory support, n (%) | | 4418 | 116 (4.1) | 34 (3.7) | 51 (3.8) | 27 (4.1) | 10 (4.3) |
Apgar score <7 at 5 min, n (%) | | 4414 | 74 (2.6) | 30 (3.3) | 57 (4.3)** | 29 (4.4)* | 2 (0.9) |
Admission to neonatology, n (%) | | 4423 | 315 (11.1) | 130 (14.2)* | 206 (15.3)*** | 100 (15.0)** | 24 (10.3) |
Compared with women in the NGT group, women classified as having GDM based only on the WHO 2013 criteria for fasting glucose were more likely to have gestational hypertension (7.8% vs 4.9%, p = 0.003), to have a planned Caesarean section (10.3% vs 6.5%, p = 0.001) and induced labour (34.8% vs 28.0%, p = 0.001).
Women classified as having GDM based only on the WHO 1999 criteria for 2HG were more likely to have induced labour (62.8% vs 28.0%, p < 0.001) compared with women in the NGT group. There were no significant differences in gestational hypertension, preeclampsia or mode of delivery between this group and women with NGT.
Neonates from mothers classified as having GDM (using the WHO 2013 criteria or WHO 1999 criteria) had a lower birthweight, a lower gestational age at delivery and were less likely to have macrosomia compared with those from mothers with NGT. However, the likelihood of these neonates being born large for gestational age (LGA) [
16] did not differ significantly from that of neonates born to women with NGT. The likelihood of these neonates being born small for gestational age (SGA) [
16] was lower than that of neonates born to women with NGT.
Compared with neonates from mothers with NGT, neonates from mothers classified as having GDM based only on the WHO 2013 criteria for fasting glucose had similar birthweight (3580 ± 596 g vs 3544 ± 579 g, p = 0.145), likelihood of having macrosomia (22.2% vs 20.9%, p = 0.452) or being born LGA (21.0% vs 18.0%, p = 0.077). However, these neonates were more likely to have had an Apgar score <7 after 5 min (4.4% vs 2.6%, p = 0.015) and to have been admitted to the neonatology department (15.0% vs 11.0%, p = 0.004). None of the other neonatal outcomes showed significant differences between these two groups.
Compared with neonates from mothers with NGT, neonates from mothers classified as having GDM based only on the WHO 1999 criteria for 2HG had a lower birthweight (3437 ± 498 g vs 3544 ± 579 g, p = 0.01) and were less likely to have macrosomia (12.8% vs 20.9%, p = 0.003). The likelihood of these neonates being born LGA was similar to those from mothers with NGT (15.4% vs 18.0%, p = 0.309). However, 20.5% of the women in this group were treated with insulin. None of the other neonatal outcomes showed significant differences between these two groups.
When we compared the percentage of LGA neonates in our data with those found in the general obstetric population in the north of the Netherlands (11%), we found that all GDM classification groups as well as women with NGT had a higher percentage of LGA neonates.
Discussion
This large retrospective cohort study to evaluate the possible impact of applying the new WHO 2013 criteria demonstrates that the number of GDM diagnoses would increase by 45%, relative to the WHO 1999 criteria. We also show that applying these new criteria indeed identifies a new group of women (with fasting glucose ≥5.1 but ≤6.9 mmol/l) who have unfavourable characteristics and more adverse pregnancy outcomes when compared either with women found to have NGT upon testing or with the general obstetric population. Our results show that women potentially missed owing to the higher 2HG threshold (2HG ≥7.8 but ≤8.4 mmol/l) of the WHO 2013 criteria have similar pregnancy outcomes to women with NGT. Our results also indicate that neonates from mothers who are tested for GDM but are found to have NGT are more likely to be born LGA or with macrosomia than those born to mothers in the general obstetric population in our region.
Pregnancy outcomes
Although uncertainty remains regarding the optimal glucose threshold to define GDM, hyperglycaemia during pregnancy is clearly associated with an increased risk of adverse pregnancy outcomes [
6]. Indeed, women in this study classified as having GDM using any criteria had higher rates of adverse maternal outcomes, including hypertensive disorders during pregnancy, planned Caesarean section and induced labour when compared with women with NGT. Moreover, the neonates of mothers classified as having GDM by any criteria were likely to have been admitted to the neonatology department.
Concerns have been raised about the ‘medicalisation’ of pregnancy should the new WHO 2013 criteria be implemented [
10]. Based on the WHO 1999 criteria currently applied in the Netherlands, these women are not diagnosed with GDM and are therefore not treated with diet and/or insulin. However, our findings suggest that these women already have higher intervention rates. We demonstrated that, in contrast to women with NGT, women classified as having GDM based only on the WHO 2013 criteria for fasting glucose had higher rates of gestational hypertension, planned Caesarean section and induced labour and their neonates were more likely to have an Apgar score <7 at 5 min and to be admitted to the neonatology department. A number of other studies have also shown that women reclassified as having gestational diabetes using the new WHO 2013 criteria are at increased risk of adverse pregnancy outcomes including gestational hypertension, Caesarean section, neonatal intensive care admission and LGA neonates [
19‐
22].
In terms of the likelihood of having an LGA neonate, we found no significant differences between the women classified as having GDM based on the WHO 2013 threshold for fasting glucose and women with NGT. However, the percentage of women in this group having an LGA neonate was much higher than for the general obstetric population (21% vs 11%). On the basis of these findings, it seems that women classified with GDM based only on the WHO 2013 criteria for fasting glucose should not be left untreated. This is supported by the results of a study by Landon et al, 2009, suggesting that early treatment in women with mild GDM reduces the percentage of women giving birth to LGA neonates by 7% [
4].
Our study has also shown that implementing the new WHO 2013 criteria with a higher 2HG threshold may exclude a group of women who now benefit from treatment. The women classified as having GDM based only on the WHO 1999 criteria for 2HG had pregnancy outcomes similar to those of women with NGT. A notable finding was that they had the lowest rate of LGA neonates of all other diagnostic groups. The only obstetric variable that differed in comparison with the NGT group was the rate of induced labour, but it has to be borne in mind that induction of labour at 38/39 weeks of gestation is more likely to be recommended in women being treated for GDM, especially those receiving insulin therapy.
All women classified as having GDM based on the WHO 1999 criteria were actively treated with diet and/or insulin. These interventions normalised their glycaemic profile and outcomes for this group [
14,
15]. Therefore, it is unclear whether these women can be safely left untreated after implementing the new WHO 2013 criteria. Indeed, Farrar et al, 2015, showed that even women with a 2HG glucose level ≥7.5 mmol/l are at increased risk of adverse outcomes (i.e. birthweight >90th percentile, high infant adiposity, and Caesarean section) [
23]. These authors therefore recommend using a 2HG glucose threshold even lower than those recommended by both the WHO 1999 and WHO 2013 criteria.
A notable finding of our study is that the women who had undergone an OGTT and were subsequently found to have NGT also had a rate of LGA neonates higher than that of the women receiving treatment after being diagnosed with GDM based on the WHO 1999 criteria for 2HG (18.0% vs 15.4%). Although this finding was not statistically significant, it was a large difference compared with the incidence of LGA neonates in the general obstetric population (18% vs 11%). This coincidental finding shows that even NGT women without a positive diagnosis of GDM are at increased risk of giving birth to an LGA neonate. This finding is in agreement with a study by Meek et al, 2015, who demonstrated that women diagnosed and treated for GDM according to the National Institute for Health and Care Excellence (NICE) criteria in the UK had lower rates of LGA neonates than women negative for GDM according to both the NICE and IADPSG/WHO 2013 criteria [
21]. A possible explanation for this finding is that these women were tested too early in pregnancy and were therefore not diagnosed with GDM at this time. Some studies have shown that the OGTT has a poor reproducibility, suggesting that some women who first test negative for GDM can test positive on a second test [
24]. We therefore agree with the suggestion made by Meek et al, 2015, that standard lifestyle interventions (including dietary advice) given to women with GDM might also benefit NGT women.
Strengths and limitations
A major strength of our study is the relatively large cohort of laboratory results from 75 g OGTTs and the extensive and detailed information regarding pregnancy outcomes in a subset of 4431 women with singleton pregnancies. All women with GDM were treated according to a detailed protocol in two large hospitals [
14,
15]. Maternal and pregnancy outcome data were collected manually from individuals’ charts at their midwives’ offices. This study also has limitations that should be noted. First, since universal testing for GDM is not currently recommended in the Netherlands, only women with one or more risk factors for GDM or signs suggestive of GDM, such as macrosomia, were tested. The number of pregnancies affected by GDM found in our study is therefore not a reflection of the general obstetric population, and represents a selected group of women at higher risk of GDM. Universal testing is now recommended in several countries around the world. However, literature regarding the best method of screening (universal or risk-based) remains controversial [
8]. Second, the WHO 2013 criteria also recommend that the diagnosis of GDM should include a 1 h plasma glucose of ≥10.0 mmol/l following a 75 g OGTT. Since we did not have data for 1 h glucose levels, the number of GDM diagnoses reported here might be an underestimation. Third, all women diagnosed according to the WHO 1999 criteria were offered treatment for GDM. Finally, we have compared outcomes with those in the general population in the north of the Netherlands between 2011 and 2013. Unfortunately data after 2013 have not yet been made available from public datasets. Furthermore, this dataset does not indicate which women were tested with an OGTT.
Acknowledgements
The authors wish to thank the endocrinologists, gynaecologists, diabetes specialist nurses, and dietitians of the University Medical Center and Martini Hospital Groningen. Special thanks are expressed to the participating midwife practices: De Verloskundigenpraktijk van Groningen, Verloskundigenpraktijk Hoogezand, Verloskundigenpraktijk La Vie, Verloskundigenpraktijk New Life, Verloskundige Stadspraktijk, Verloskundigenpraktijk ‘t Stroomdal, Verloskundigenpraktijk Veendam. We would also like to thank H. Hepkema-Geerligs (customer relations manager Laboratory of Clinical Chemistry, Certe, the Netherlands) and the students S. Klöppner (University Medical Center Groningen) and J. van Amstel (University Medical Center Groningen) for their contribution to the data collection. Finally, we thank epidemiologist H. Groen (Department of Epidemiology, University Medical Center Groningen), the Dutch Perinatal Registry and the Municipal Health Service Groningen for providing the data on the reference population in the northern region of the Netherlands.
Some of the data were presented as an abstract at the 53rd EASD Annual Meeting in 2017.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.