In one line
Every normal pregnancy deliberately makes the mother insulin resistant so that glucose flows to the fetus, and a healthy pancreas answers by tripling its insulin output. Gestational diabetes is what happens when a pancreas cannot answer. That single idea tells you why the disease appears when it does, why the injury depends on when the sugar was high, and why the most important appointment in the whole story is the one six weeks after delivery.
Where this sits. The endocrine groundwork this chapter builds on, the hormone axes and the normal control of fuel metabolism, is in the Primary chapter on endocrine physiology. The consultant layer above it, where the subtype mechanisms, the named insulin regimens and the guideline disagreements are argued out, is the Finals chapter on diabetes in pregnancy.
Pregnancy is a metabolic stress test the woman never asked for
By the third trimester, maternal insulin sensitivity has fallen by roughly 50%. In any non-pregnant adult that number would describe advanced metabolic disease. In pregnancy it is the intended state.
The healthy pancreas compensates by raising insulin secretion two to three-fold. Nothing shows on a glucose reading, because the extra insulin exactly cancels the extra resistance. The woman walks around with the metabolism of a person with insulin resistance and the glucose readings of a person without it.
Now take a woman whose beta cells were already working near their ceiling. She has never had an abnormal glucose in her life, because nothing has ever asked her pancreas for three times its baseline output. Pregnancy asks. The reserve is not there, and her glucose rises.
This reframes the whole topic. Gestational diabetes is not a disease that pregnancy causes. It is a limitation that pregnancy reveals. The hyperglycaemia is new; the defect is not. That is why roughly 50% of women with gestational diabetes develop type 2 diabetes within 10 to 20 years, why it recurs in later pregnancies, and why finding it matters at least as much for the woman's next thirty years as for this baby.
It also explains something you will see constantly in a South African antenatal clinic. Young women are presenting with obesity and type 2 diabetes at ages where a generation ago you would not have looked. Pregnancy is often the first time any of them has had a glucose measured. Whether you catch that is decided by whether you screen.
How glucose is normally held steady
Insulin does three things, and holding the three separately is what makes the rest of this chapter easy.
It tells the liver to stop making glucose. Between meals the liver produces glucose by glycogenolysis and gluconeogenesis. Insulin switches that off. This is what sets the fasting glucose.
It tells skeletal muscle and adipose tissue to take glucose up. Insulin binds its receptor, the receptor phosphorylates insulin receptor substrate 1 (IRS-1) on tyrosine residues, IRS-1 activates phosphoinositide 3-kinase (PI3K), and the pathway ends with GLUT4 transporters moving from intracellular vesicles to the cell surface. Glucose then flows in down its gradient. This is what clears the postprandial rise.
It tells adipose tissue to stop lipolysis. Insulin suppresses hormone-sensitive lipase. Without insulin, free fatty acids pour into the circulation and the liver converts them to ketones.
Hold those three. Fasting glucose is a liver readout. Postprandial glucose is a muscle readout. Ketones are a lipolysis readout. Every abnormality in this chapter is one of those three failing.
What the placenta does to that system, and why
The placenta is not a passive organ. It is a large and growing endocrine gland whose commercial interest is the fetus, and it manipulates maternal metabolism accordingly.
Human placental lactogen is produced by syncytiotrophoblast, and its concentration tracks placental mass. It rises through the second trimester and peaks in the third. It drives lipolysis, flooding the mother with free fatty acids. Those fatty acids activate protein kinase C, which inhibits insulin signalling at IRS-1. So hPL does not block insulin directly. It causes a fuel switch that blocks insulin downstream.
Placental growth hormone progressively replaces pituitary growth hormone from mid-gestation and is a potent driver of insulin resistance and of maternal lipolysis.
Progesterone interferes with insulin receptor signalling and impairs GLUT4 translocation. Oestrogen runs the other way and improves peripheral glucose uptake, which is one reason the net effect is not simply proportional to placental hormone load.
Cortisol is raised, and placental 11beta-hydroxysteroid dehydrogenase type 1 overactivity contributes. Cortisol acts through glucocorticoid receptors to reduce insulin sensitivity and to promote hepatic gluconeogenesis.
Prolactin rises steeply and contributes to the resistance, while also driving the beta-cell mass expansion that is supposed to compensate for it.
Then two signals from adipose tissue, which matter enormously in an obese population. Leptin rises, and in an already leptin-resistant woman it adds nothing useful. Adiponectin falls, and this one is mechanistically important. Adiponectin normally activates AMP-activated protein kinase (AMPK) in skeletal muscle, and AMPK promotes GLUT4 translocation. Hypoadiponectinaemia therefore removes an insulin-independent route for muscle to take up glucose at exactly the moment the insulin-dependent route is being suppressed.
Finally the inflammatory signals. TNF-alpha and interleukin-6 rise, from adipose tissue and from the placenta itself. They activate IkappaB kinase, which activates protein kinase C, which phosphorylates IRS-1 on serine rather than tyrosine residues. A serine-phosphorylated IRS-1 cannot propagate the insulin signal. This is the same molecular lesion that underlies obesity-associated insulin resistance outside pregnancy, which is why obesity and pregnancy are not two separate risks but the same one, arriving twice.
The purpose of all of this is a maternal-fetal glucose gradient. Glucose crosses the placenta by facilitated diffusion through GLUT1, down its concentration gradient, with no active transport and no regulation. The fetus cannot pull glucose across. It can only receive what the gradient gives it. So the only way to guarantee fetal supply is to keep maternal glucose higher than fetal glucose at all times, and the placenta achieves that by making the mother's own tissues stop competing for it.
Where the system breaks
Two compartments fail, and the distinction matters clinically.
The peripheral tissue. In gestational diabetes the post-receptor defects described above are more severe than physiology requires. IRS-1 serine phosphorylation is increased, PI3K activation is reduced, GLUT4 translocation is impaired, and the AMPK route is blunted by low adiponectin. Muscle stops clearing the postprandial load. This produces the postprandial peaks.
The beta cell. The compensatory expansion fails. Several mechanisms are described: reduced expression of PDX1, the transcription factor that maintains beta-cell identity and insulin gene expression; endoplasmic reticulum stress, with rising CHOP and GRP78, because a cell forced to triple its output of a secreted protein is being asked to triple its protein-folding capacity; mitochondrial dysfunction with impaired ATP production, which matters because glucose-stimulated insulin secretion is an ATP-dependent process; oxidative damage from accumulating reactive oxygen species; and reduced incretin signalling, with lower GLP-1 and GIP, removing the amplification that normally makes an oral glucose load release far more insulin than the same load given intravenously.
Put those together and you have the phenotype. A woman with gestational diabetes typically has a normal or near-normal fasting glucose and abnormal postprandial peaks, because the liver arm is relatively preserved and the muscle arm has failed. A woman with unrecognised type 2 diabetes has both abnormal, because her liver has stopped listening too. That is not a rule you memorise. It is a consequence of which compartment failed.
Why the fasting glucose is the confusing one
Here is the finding that catches people out, and once it makes sense you will never forget the diagnostic thresholds.
In normal pregnancy the fasting glucose falls. It falls for two reasons: the fetus siphons glucose continuously across the placenta, twenty-four hours a day, including overnight when the mother is not eating; and maternal plasma volume expands, diluting everything. So a healthy pregnant woman has a lower fasting glucose than she did before conception.
Postprandial glucose, meanwhile, rises, because that is the arm the placental hormones have deliberately impaired.
Now look at what that does to the numbers. A fasting glucose of 5.4 mmol/L in a non-pregnant adult is unremarkable, comfortably below the 5.6 mmol/L that defines impaired fasting glucose. The same 5.4 mmol/L at 26 weeks is diagnostic of gestational diabetes, because pregnancy should have pushed it down and instead it has stayed flat. The threshold is not lower because pregnancy is treated more anxiously. It is lower because the baseline moved.
Three different diseases share one name
The World Health Organization classification, adopted by SEMDSA and used across South African practice, separates hyperglycaemia in pregnancy into categories that behave differently.
Pre-existing diabetes. Type 1, type 2 or another form, diagnosed before this pregnancy, with or without established complications. Type 2 is by far the commonest form encountered in pregnancy.
Diabetes mellitus in pregnancy, also called overt diabetes. Hyperglycaemia detected for the first time in pregnancy that meets the non-pregnant diagnostic criteria for diabetes. Almost always this is type 2 diabetes that was there before conception and had never been looked for.
Gestational diabetes. A lesser degree of glucose intolerance, first recognised in pregnancy, typically appearing in the second half, and expected to resolve after delivery.
The reason the classification earns its place is that these are not three severities of one condition. Overt diabetes and pre-existing diabetes were present at conception. Gestational diabetes was not. That single fact decides the injury.
Timing decides the injury
Organogenesis is complete by about eight weeks after conception. Hyperglycaemia during that window is a teratogenic exposure. Hyperglycaemia after it is a growth exposure. They produce entirely different babies.
Hyperglycaemia at conception and in the first trimester raises the risk of miscarriage and of congenital malformation. Cardiac defects and neural tube defects are the commonest. Caudal regression, or sacral agenesis, is rare but so strongly associated with maternal diabetes that finding it should send you looking for undiagnosed disease in the mother. The risk is graded by periconceptional glycaemia, which is the entire argument for preconception care: by the time a woman with type 2 diabetes books at 14 weeks, the window in which her HbA1c mattered most has already closed. SASOG regards an HbA1c above 10% as a relative contraindication to pregnancy, on grounds of teratogenicity.
Hyperglycaemia in the second half of pregnancy does something quite different, and the mechanism is worth building properly because it explains a whole cluster of neonatal problems that otherwise look unrelated.
Maternal glucose crosses the placenta freely. Maternal insulin does not. The fetus therefore sees the mother's glucose and must make its own insulin to deal with it. In sustained maternal hyperglycaemia the fetal beta cells undergo hyperplasia and the fetus becomes hyperinsulinaemic. This is the Pedersen hypothesis, and Freinkel extended it to amino acids and lipids, so the fetus is over-supplied with fuel generally, not glucose alone.
Fetal insulin is a growth factor. Follow it into each tissue and the neonatal picture assembles itself:
- Insulin-sensitive tissues grow. Liver, cardiac muscle, subcutaneous fat and the shoulder girdle enlarge. The brain is relatively insulin-independent and does not. The result is asymmetric macrosomia: abdominal circumference above the 90th centile with a head circumference near the 50th. This is the diabetic morphometry SASOG asks you to look for, and it is why a 4.2 kg baby of a diabetic mother carries a higher shoulder dystocia risk than a constitutionally large 4.2 kg baby, whose growth is proportionate. See Macrosomia.
- Insulin antagonises cortisol at the type II pneumocyte, delaying surfactant maturation. Hence respiratory distress syndrome at gestations where you would not expect it. See Initiation of respiration.
- Hyperinsulinaemia raises fetal oxygen consumption, producing relative tissue hypoxia, erythropoietin release and polycythaemia, which later becomes hyperbilirubinaemia as that red cell mass is broken down.
- The hyperinsulinaemia does not switch off at delivery, but the glucose supply does. The cord is clamped, the maternal glucose infusion stops, and a baby still secreting insulin for a hyperglycaemic environment becomes profoundly hypoglycaemic. This is why the paediatrician must be told before, not after. See Neonatal transition.
- Hypocalcaemia and hypomagnesaemia follow a delayed parathyroid response.
And the same fuel over-supply leaves a durable mark. Intrauterine hyperglycaemia is associated with obesity and type 2 diabetes in the offspring, so the disease propagates into the next generation through the uterus rather than through the genome alone.
What it does to the mother
Diabetes in pregnancy is a pre-eclampsia disease. Pregestational diabetes is one of the established strong clinical risk factors in the ISSHP classification, alongside prior pre-eclampsia, chronic hypertension, BMI above 30, antiphospholipid syndrome and assisted reproduction. Add polyhydramnios (fetal polyuria from fetal hyperglycaemia), preterm delivery, operative delivery, accelerated vascular disease, and diabetic ketoacidosis. See Pre eclampsia and HELLP and Hypertension in pregnancy.
That ISSHP listing has a practical consequence that is easy to miss. Aspirin prophylaxis has to start before 16 weeks, and certainly before 20. A woman with pregestational diabetes qualifies at booking. A woman with gestational diabetes diagnosed on a 26-week glucose tolerance test does not qualify in time, because the window has closed. Aspirin in diabetes is a preconception and first-trimester decision, not a response to a glucose result.
Screening: who, when, and with what
Two questions have to be answered separately: whom do you test, and against what numbers.
On whom, the international position and the South African position differ for a reason that is about money, not about medicine.
FIGO and the IADPSG recommend universal one-step testing: every pregnant woman gets a 75 g oral glucose tolerance test at 24 to 28 weeks. SEMDSA notes that universal screening may be adopted in well-resourced settings, and SASOG says that universal screening is advisable in ideal circumstances but that the recommendation is dependent on local resources.