Gestational Diabetes: beyond the label

Updated: November 2023

I have written this blog post in response to readers’ requests. Trying to make sense of the research and guidelines hurt my brain, and I almost gave up a few times. So, for those who asked – I hope this post meets your expectations! The post explores blood glucose levels (BGLs) in pregnancy, and attempts to make some sense of the fairly nonsense diagnosis and management of ‘gestational diabetes’ (GD). This post is not about Type 1 or Type 2 diabetes, and I am assuming you already know about the relationship between blood glucose (sugar) and insulin – if not do some googling.

Blood glucose and insulin in a healthy pregnancy

Babies needs glucose to grow, and the demand for glucose increases as pregnancy progresses and the baby develops. From around 20 weeks, placental hormones cause insulin resistance in the mother’s cells. Insulin resistant cells are less able to convert glucose into energy, resulting in a peak of blood glucose after eating a meal which goes through the placenta to ‘feed’ the baby. In response to this peak, the woman’s pancreas increases the production of insulin to bring BGLs back down to a healthy pre-meal range. So, during pregnancy the woman’s body needs to bump up insulin production to counteract the effect of insulin resistant cells. Once the baby is born, the placental hormones stop entering the woman’s circulation and her insulin metabolism returns to her pre-pregnant state.

High blood glucose in pregnancy

[NOTE: the clear as mud definition of ‘high’ is discussed below in ‘parameters of normal’]

Some women’s bodies are unable to produce the additional insulin required during pregnancy. This results in high levels of glucose remaining in the blood instead of being converted into energy by insulin. The exact cause of this situation is not clear. However, pregnancy places additional demands on the body’s metabolism, and pre-existing health issues influence the ability of the body to meet these demands. High BGLs in pregnancy are associated with an increased chance of health problems during pregnancy (eg. pre-eclampsia) and later in life (eg. cardiovascular disease and Type 2 diabetes). Therefore, pregnancy may offer a glimpse into the general health of a woman, and her ability to meet physical challenges. Rather than causing ill health, abnormal BGLs may reflect underlying ill health.

What is known is that high maternal BGLs influence the development of the baby. In early pregnancy (before 14 weeks) high BGLs are associated with an increased chance of miscarriage, congenital abnormality and subsequent stillbirth (Murphy et al. 2017). This is because the structural development of the major organs is taking place at this time, and any toxin, including excessive glucose can cause damage. However, BGLs are only high in early pregnancy in poorly controlled, pre-existing diabetes.

In contrast, ‘pregnancy induced’ high BGLs do not occur until after 20 weeks when insulin resistance kicks in. By 20 weeks all of the baby’s major organs have formed, and the baby grows mostly in size rather than in complexity. Therefore, pregnancy induced high BGLs primarily effect the weight/shape of the baby. In response to maternal high BGLs passing through the placenta, the baby increases their own insulin production. This insulin converts the excess blood glucose into additional fat stores resulting in a heavier baby. This extra fat is concentrated around the baby’s upper body, in particular around the shoulders. Chunky shoulders increase the chance of shoulder dystocia and perineal tearing during birth. Insulin can also delay the production of surfactant, which prepares the lungs for breathing. This can cause breathing issues at birth, particularly if the baby is born early (eg. by early induction or c-section – which are more common when GD is diagnosed).

Once the baby is born, they no longer need to produce high insulin. However, adjusting and re-balancing insulin and BGLs can be a bit of a bumpy ride for the baby. The withdrawal of high BGLs is sudden (as soon as the placenta stops functioning); but it can take some hours before the baby’s insulin levels drop. During this time the high insulin can covert too much of the baby’s blood glucose into energy resulting in low BGLs (hypoglycaemia).

The baby’s high insulin levels during pregnancy also increase their red blood cells. After birth the baby needs to break down and excrete these additional red blood cells. A by-product of breaking down red blood cells is bilirubin. If there is a lag between breaking down the red blood cells and excreting them out of the body, bilirubin builds up causing jaundice. Jaundice is common in babies who produced high insulin in the uterus.

The effects of high BGLs and high insulin in utero may also cause long term epigenetic changes to the baby’s metabolism. These babies have an increased chance of developing obesity and Type 2 diabetes later in life.

If you want a more in-depth explanation watch this movie:

In summary high BGLs in pregnancy are not ideal, and can alter the growth and development of the baby.

‘Gestational Diabetes’

In an attempt to identify and manage women with high BGLs, the maternity system has defined a disease and created a label that clinical guidelines can be based around. When high BGLs are identified for the first time during a pregnancy it is referred to as ‘gestational diabetes’ (GD) or ‘gestational diabetes mellitus’ (GDM). Most cases of GD are pregnancy induced ie. caused by an inability to meet the additional insulin needs of pregnancy – as described above. Occasionally, Type 2 diabetes was already present but only identified in pregnancy. Either way – high BGLs will be termed GD until proven otherwise ie. after pregnancy when BGLs fail to return to normal in the case of undiagnosed Type 2.

However, due to inconsistencies in who is tested, and how and what parameters are applied, there is a huge variation in whether an individual woman gets diagnosed and labelled with GD or not. For example, the incidence of GD varies globally from 2% to 26% depending on the definition used, the approach to screening, and the population of women tested.

Applying the label

There are two main approaches to screening for GD – universal screening (every woman is offered a test) and risk factor-based (only women with an increased chance of developing GD are offered a test). There is no evidence to demonstrate that either approach improves outcomes for mothers and babies. A Cochrane Review concluded: “There is not enough evidence to guide us on effects of screening for GDM based on different risk profiles or settings on outcomes for women and their babies… Low-quality evidence suggests universal screening compared with risk factor-based screening leads to more women being diagnosed with GDM.”

TYPES OF TESTING

The Oral Glucose Tolerance Test (OGTT) is offered between 24 and 28 weeks gestation, or earlier for women considered ‘at risk’ of GD. It is the standard recommended test for GD diagnosis in most clinical guidelines worldwide. It involves fasting overnight, then drinking a glucose solution, followed by a blood test to assess BGLs. The dose of glucose can vary from 50g, 75g to 100g; and the timing of the blood test varies from 1 hour, 2 hours or 3 hours afterwards. There is no evidence to support any of these variations, however most guidelines recommend 75g of glucose and a 2 hour blood test. The OGTT assesses how well a woman’s body responds to a huge bolus of glucose (and chemicals – read the label).

The Glucose Challenge Test (GCT) was previously recommended as a screening assessment (24-28 weeks) to determine which women went on the have the OGTT. The test involves drinking a 50g glucose solution and having a blood test 1 hour later. However, the test lacks both sensitivity and specificity and is no longer recommended (except in the US).

The Glycated haemoglobin (HbA1c) is only recommended for identifying pre-existing diabetes during the first trimester of pregnancy. The results of the blood test provide an indication of what the average BGLs have been over a 2-3 month period. This test cannot effectively identify pregnancy induced diabetes – only previously undiagnosed Type 2 diabetes.

Self testing is not recommended in any guidelines – however some women choose to do this rather than an OGTT. The woman tests her own BGLs over a few days to get an idea about what her BGLs are doing when she is following her usual diet and lifestyle.

PARAMETERS OF NORMAL

It is generally agreed that the normal range of blood glucose for non-pregnant people is 4.0 to 6.0 mmol/L (millimoles per litre) when fasting, and up to 7.8 mmol/L two hours after eating. Diagnosis of non-pregnant diabetes occurs when an OGTT identifies fasting BGLs  ≥ 7 mmol/L or BGLs ≥ 11.1 two hours after 75g glucose.

However, when it comes to pregnancy, definitions and parameters of normal are not so clear. Various organisations advocate differing diagnostic parameters, and Bonventura, Ernest & Dee (2015) describe a number of them. However, I’ll stick to the most commonly used criteria initiated by the International Association of Diabetes and Pregnancy Study Group (IADPSG). In 2010 the IADPSG Consensus Panel lowered the threshold for GD diagnosis. This move was based on the findings of one study – the HAPO study. This was an observational study looking at the risk of ‘adverse outcomes’ (see above) associated with 7 different categories of fasting BGLs; and with 1 hour and 2 hour BGLs after 75g glucose. The findings identified an association between fasting BGLs and the frequency of particular ‘adverse outcomes’ (see association vs causation in this post). The study reported: “frequencies in the lowest and highest [of the 7 fasting BGL] categories, respectively, were 5.3% and 26.3% for birth weight above the 90th percentile, 13.3% and 27.9% for primary [first ie. not VBAC] cesarean section, 2.1% and 4.6% for clinical neonatal hypoglycemia [low BGL], and 3.7% and 32.4% for C-peptide level [which reflects baby’s insulin levels] above the 90th percentile .” The amount that the 1 hour and 2 hour BGLs went up also influenced the frequency of ‘adverse outcomes’ – although the associations for primary c-section and neonatal hypoglycaemia (low BGLs) were weak.

The IADPSG Consensus Panel concluded that: “because associations were continuous with no obvious thresholds at which risks increased… a consensus was required to translate these results into clinical practice.” And so the new GD diagnostic threshold was created: OGTT results of BGL ≥ 5.1 mmol/l fasting or ≥ 8.5 mmol/l two hours after 75g glucose load. These thresholds are based on the average BGL values that increased the odds of a big baby by 1.75 times. Whilst this threshold may reduce the rates of babies over 4kg, there is no evidence that it will reduce the rate of birth/newborn complications Bonventura, Ernest & Dee (2015).

WHO changed their recommendations to align with IADPSG’s. WHO even state in the recommendation that the quality of evidence to support this new threshold is ‘very low’, and the strength of the recommendation is ‘weak’. This threshold results in up to 16% of pregnant women meeting the criteria for GD (previously 5%). Kevat et al. (2014) raised a number of concerns about the impact of the lower threshold for Australian women – many of which can be applied to other populations. However, despite an initial wave of concern from care providers, consumers, maternity organisations and researchers – these new thresholds made it into clinical guidelines and practice worldwide.

Research carried out after the implementation of a lower BGL threshold demonstrates no improvements, quite the opposite. A Cochrane review looked at ‘different intensities of glycaemic control’ for women with GDM. They found that with lower BGL targets there was no difference in the rates of c-section, induction of labour or body fat percentage of the baby. However, there was an increased risk of the women developing high blood pressure and protein in the urine. The review concludes that: There remains limited evidence regarding the benefit of different glycaemic targets for women with GDM to minimise adverse effects on maternal and infant health. Glycaemic target recommendations from international professional organisations vary widely and are currently reliant on consensus given the lack of high-certainty evidence.

Two Australian studies investigated the benefits and harms of lower BGL thresholds for GDM. The first study (Cade et al. 2018) examined the impact of introducing the new thresholds and found that: “There was an increase in annual incidence of GDM of 74% without overall improvements in primary health outcomes. This incurred a net cost increase of AUD$560 093. Babies of women with GDM had lower rates of neonatal hypoglycaemia and special care nursery admissions after the change, suggesting a milder spectrum of disease.” The latest Australian study (Hegerty & Ostini 2023) compared two large retrospective cohorts, one before the BGL threshold changed (2011-2013) and one after the change (2016-2018). The rate of GDM diagnosis increased from 7.8% to 14.3% between the cohorts. For the cohort with lower BGL thresholds, there was no improvement in the rate of shoulder dystocia, c-sections or hypertensive disorders. Instead, there was an increased rate of planned birth (induction or c-section with no labour) from 62.9% to 71.8% and planned early birth (before 39 weeks) from 35% to 45%. These rates were even higher for women prescribed insulin. Early planned birth is important to consider in relation to short and long-term newborn health and breastfeeding. For women diagnosed with GDM the chance of going into spontaneous labour and giving birth vaginally (SLVB) went from 30.01% to 23.6% between the two cohorts. Note that I can’t work out from the article if ‘vaginal birth’ includes an instrumental birth or if this ‘spontaneous labour’ includes augmentation. However, I think it is safe to say that physiological birth is likely to be very uncommon for women diagnosed with GDM. The researchers conclude that: Outcomes were not apparently improved with increased GDM diagnosis. The merits of increased IOL or decreased IOL or decreased SLVB depend on the views of individual women, but categorising more pregnancies as abnormal, and exposing more babies to the potential effects of early birth, medication effects and growth limitation may be harmful.

Increasing calls for a review of the diagnostic criteria (from medicine) are being ignored as the new threshold norm continues (Bilous et al. 2021; Doust et al. 2022).

Treating the label

Once a woman has been labelled with GD she is usually diverted into ‘GD-centred’ antenatal care. There is often stigma attached to having GD, and additional medical surveillance and restricted choices regarding birth setting. Management of GD centres on keeping the BGLs within a certain range via diet and exercise, and/or insulin medication. The issues around dietary recommendations are a whole other issue that I can’t fit into this blog post. Long story short – the usual GD recommendations involve a high carb (ie. sugar) diet. Alternatively, Lily Nichols has written a couple of great books about diet in pregnancy and for GD (see below in further resources).

Although there are varying opinions about what BGLs should be maintained by women diagnosed with GD (Bonventura, Ernest & Dee (2015). In general the fasting BGL target is around 5.0-5.5 mmol/l fasting and the 2 hour post meal BGL is 6.7-7.1 mmol/l (by capillary blood, ie. finger prick test). Not surprisingly, hypoglycaemia (low BGLs) is a common problem for women trying to keep their BGLs within this range.

WHO summarised the evidence into the effectiveness of GD treatment. The only outcome categorised as ‘high quality’ is that treatment for GD reduces the chance of having a  baby 4kg+ (number needed to treat NNT = 11.4 to prevent 1 large baby). However, the evidence indicating a reduction in shoulder dystocia is of ‘low quality’ (NNT = 48.8 to prevent one shoulder dystocia). There is ‘moderate quality’ evidence that treatment reduces the chance of hypertension (NNT 18.1) and pre-eclampsia (NNT 21). For all other outcomes evidence was ‘moderate’ to ‘low’ quality. Bear in mind the research in the WHO summary was carried out before the new lower GD thresholds were introduced. A more recent Cochrane Review compared lifestyle interventions (diet and exercise) with ‘usual’ care or another intervention and found no difference in any outcomes except the size of the baby.

It is also important to note that only 14-22% of women diagnosed with GDM will have a baby over 4kg and ultrasound assessment of size is ineffective.

Labour and birth care for women labelled GD

The IADPSG Consensus Panel acknowledged that the “bias of caregivers toward expectation of adverse outcomes may increase morbidity due to increased intervention” for women diagnosed with GD. Women are often coerced into early induction of labour or even c-section because they have been diagnosed with GD. By coerced, I mean they are advised to have an intervention, rather than discussing the risks and benefits of various options, and their individual situation, then making their own decision.

Large-scale research exploring birth outcomes for GD tends to focus on the label rather than on BGLs. This results in 3 groups of women being mixed into the research sample:

  1. women with pre-existing diabetes only diagnosed in pregnancy
  2. women diagnosed with GD who had high BGLs during pregnancy
  3. women diagnosed with GD who maintained normal BGLs during pregnancy

For this mixed up group of GD women a Cochrane Review concluded: “There is insufficient evidence to clearly identify if there are differences in health outcomes for women with gestational diabetes and their babies when elective birth is undertaken compared to waiting for labour to start spontaneously or until 41 weeks’ gestation if all is well.” (the ’41 weeks’ is because induction at this gestation tends to be standard for all women).

However, things look different when we consider women based on what their BGLs have been in pregnancy rather than their GD label. In this case there are 2 distinct groups:

1. Women with normal BGLs (and a GD label)

These women do not have babies effected by high BGLs – because they didn’t have consistently high BGLs during pregnancy. Their babies are as likely to be over 4kg as women without GD. They should be cared for in the same way as women without a GD label because their ‘risk profile’ is the same. For this group of women induction is not supported by evidence or clinical guidelines. ACOG (US) state that “women with GDM with good glycemic control and no other complications are commonly managed expectantly until term.” Queensland Health (Australia) recommend that if blood glucose is well managed, there is no indication for induction for gestational diabetes. Despite this clear guidance women, are often booked in for an early induction by their care provider based simply on their GD label.

2. Women with abnormal BGLs (and a GD label)

This group of women are at increased chance of experiencing complications associated with high BGLs during pregnancy (see above). However, even for this group of women there is a lack of evidence to support induction. A paper by Berger and Melamed (2014) discusses the research relating to the timing of birth for women with GD, including the risks of induction for women and babies with GD. Like the Cochrane review above, they found inadequate evidence to support induction of labour for women with GD and concluded that “until such data are available, the clinician should consider the maternal, fetal and neonatal implications of induction of labour versus expectant management, involve the patient in the decision process and as usual follow the maxim of ‘‘first do no harm.’’

The main concern regarding high BGLs in pregnancy is the size of the baby (see above). This is often used as the reason for recommending induction. Babies with big shoulders are more likely to experience shoulder dystocia. For example, in non-GD pregnancies, shoulder dystocia occurs with around 1% of babies weighing less than 4kg compared to 5-9% of babies weighing over 4kg (Politi et al. 2010). These figures may be higher for babies subjected to high BGL in pregnancy because of the distribution of their additional weight (ie. upper body and shoulders). However, increased shoulder dystocia rates may also be partially due to the interventions women with suspected big babies experience. For example, if a care provider suspects a ‘big baby’ the woman is more likely to experience interventions (syntocinon, c-section, instrumental birth, etc) and complications regardless of whether her baby is actually big (Sadeh-Mestechkin et al. 2008; Blackwell et al. 2009).

Not surprisingly, induction before 40 weeks does reduce the chance of shoulder dystocia. A baby will be smaller before 40 weeks than after 40 week, and therefore statistically less likely to get stuck. A Cochrane Review comparing induction of labour before 40 weeks for a suspected big baby with waiting for spontaneous labour, found that induction decreased the incidence of shoulder dystocia from 6.8% to 4.1%. However, they also found an increased rate of severe perineal tearing in the induction group of 2.6% vs 0.7% in the spontaneous labour group; and an increase in the treatment of jaundice for the baby (11% vs 7%). Both NICE guidelines and WHO guidelines state that induction of labour should not be carried out simply because the baby is suspected of being big. Which is interesting because both guidelines support induction in the case of GD with high BGLs where the only significant risk factor for birth is a suspected big baby.

Women with high BGLs in pregnancy need to consider the risks of possible shoulder dystocia with the risks of induction (see Berger and Melamed 2014 ) and their own individual situation and preferences. Many women with abnormal BGLs can and do have physiological births, however most follow care provider recommendations and have their labour induced. The following are some suggestions for reducing/managing complications associated with birth for women who had high BGLs in pregnancy. Most of these suggestions can be applied to physiological labour or induced labour – although an induced labour is likely to result in a smaller baby.

  • Maximise the size of the pelvis – avoid positions that restrict the movement of pelvic bones (eg. don’t sit on the back of the pelvis)
  • Maximise the ability of the baby to rotate – A mobile mother offers lots of opportunities for baby to move – water immersion is good for this. Resting space between contractions also allows the baby to move when the uterus is relaxed. If syntocinon is regulating contractions, make sure there is a good ‘resting space’ between the contractions (no more than 4 contractions in 10 mins). If the woman has an epidural her care providers / support people will need to assist her to move her pelvis (eg. pelvic rocking using the drawsheet or towel).
  • BG management – if the woman is insulin dependent it may be necessary to check BGLs during labour.
  • Avoid interventions that cause wounds eg. c-section or episiotomy. High BGLs can interfere with healing and increase the chance of infections.
  • Avoid any interventions that interfere with instinctive behaviour as the woman pushes her baby out. If she has an epidural then avoid directed pushing until the baby’s head is on the perineum – and then keep it gentle with spaces in between for re-oxygenation of mother/baby and a chance for baby to rotate and move. Do not pull on the baby’s head immediately after it has birthed – this can wedge the shoulders into the pelvis before they have had a chance to rotate. If there is no change with the next contraction (no rotation or descent) – then suspect shoulder dystocia. and manage accordingly.
  • After birth do not remove the baby from their mother – this will result in a stress response that will burn up the baby’s glycogen (glucose stores). These stores will be needed as the baby re-balances their metabolism. Any resuscitation should be done with mother and baby together.
  • Prolonged skin-to-skin with mother will stabilise the baby’s heart rate and temperature; reduce stress; and encourage early breastfeeding – all great for maintaining BGLs.
  • Ensure the baby feeds early and often. Colostrum provides a nutrient dense package of glucose to help the baby keep their BGLs within a normal range. Even a few drops can increase the baby’s BGLs significantly. Woman can express and store colostrum at the end of pregnancy to provide additional colostrum for the first hours after birth.
  • The baby may have their BGLs monitored as they adjust to the withdrawal of high maternal BGLs. Any monitoring and/or management can be done with mother and baby together. Separating mother and baby is detrimental for all kinds of reasons, including BGL stabilisation.
  • Observe the baby during the first week for jaundice. As discussed above, significant jaundice is fairly common for babies who produced high insulin during pregnancy. The baby may need light therapy to resolve their jaundice.
How to reduce the chance of hypoglycaemia (low BGLs)

Summary

High BGLs in pregnancy alter the growth and development of the baby, increasing the chance of particular complications occurring. However, the label ‘gestational diabetes’ is problematic because it is poorly defined and there is a lack of evidence to demonstrate that labelling and treatment improves outcomes. Guidelines do not support induction of labour for GD unless BGLs are high. Inducing women before 40 weeks with high BGLs reduces the chance of a large baby and shoulder dystocia, but increases the chance of other complications. Labour and birth care for women with high BGLs should centre on minimising the chance of shoulder dystocia, and supporting the baby to regulate their own BGLs after birth.

Additional Resources