A baby's abdominal circumference can climb several percentile points between two growth scans for two quite separate reasons: because he grew, and because the measurement is not exact. Wright and colleagues, writing in Ultrasound in Obstetrics & Gynecology, concluded that measurement error in fetal biometry is large enough to misclassify babies as large or small for gestational age. That is the piece almost nobody hands you, and it belongs at the front.

Not medical advice. This article cannot see your scan, your sugars or your baby, and it will not tell you whether anything about your plan is right. Only your obstetrician can weigh your case.

What those two numbers actually measure

Abdominal circumference is a line traced around your baby's middle on a frozen ultrasound image. It is one of three measurements — abdomen, head, thigh bone — fed into a formula that estimates weight, and in gestational diabetes it is the one watched hardest.

It is also less precise than a percentile makes it sound. A multicentre analysis of 11,049 term deliveries published in BMC Pregnancy and Childbirth found the sonographic estimate was within ten per cent of actual birth weight in 71.6 per cent of cases, with a mean percent error of 2.39 per cent plus or minus 9.13 per cent. Roughly three estimates in ten land more than a tenth away from the truth. In Wright's modelling the error standard deviation for the abdominal measurement alone was 6.93 mm — a few millimetres of caliper placement, on a curve where a few millimetres is several percentile points.

So part of a jump between two scans is real growth and part of it is noise. Nobody can tell you the split for your baby — and treating the number as meaningless would be as wrong as treating it as a confession.

Amniotic fluid index is the sum of the deepest pocket of fluid in each of four quadrants of your uterus. A 2024 review in the Journal of Medical Ultrasound defines polyhydramnios as an index above 24 cm and notes that the index correlates poorly with the actual amount of fluid present. A review in Geburtshilfe und Frauenheilkunde draws the line lower, calling 8.1 to 18 cm normal, and concludes that because normal volumes vary so widely, fluid volume does not play an important role in monitoring gestational diabetes. Authorities genuinely disagree about where "too much" starts. A reading of twenty-one sits under one threshold and over the other.

Why "the chocolate did this" is not how it works

Fetal growth in gestational diabetes tracks the glucose your baby is bathed in over weeks, not the area under any one meal. The test clinicians use to judge that exposure, HbA1c, reflects average blood glucose over roughly the past three months. The US National Institute of Diabetes and Digestive and Kidney Diseases describes high blood glucose levels during pregnancy as what can make a baby weigh too much and make delivery difficult. The exposure that shapes a baby's abdomen is cumulative and slow.

There is also something probably missing from the story you have been telling yourself. The NHS notes that blood sugar levels can rise as pregnancy progresses, so even when they improve at first, medicine may become necessary later. A woman whose numbers held on diet at twenty-eight weeks and needed tablets by thirty-two is not a woman who failed; she is the ordinary course of this condition.

None of which makes what you eat irrelevant, and you would not believe us if we said so. A Cochrane overview of systematic reviews found that lifestyle treatment probably reduces large-for-gestational-age birth compared with usual care, with a risk ratio of 0.60 across six randomised trials of 2,994 women. Diet moves the odds. It moves them over a pregnancy, at the level of populations, and never with the precision of "this packet, that percentile."

The honest answer sits in between, and it is uncomfortable: your eating was one input among several into a number that is itself fuzzy, in a condition that was going to get harder anyway.

Why these measurements move the delivery conversation

They move it because size carries risk. ACOG's practice bulletin on macrosomia states that as birth weight increases, the likelihood of labor abnormalities, shoulder dystocia, birth trauma, and permanent injury to the newborn increases. Rising fluid and a rising abdominal measurement together are read as a signal about glucose control, not just about size.

Guidance differs on what to do about it, and it is worth knowing whose figure is whose. NICE advises that women with gestational diabetes give birth no later than forty weeks plus six days, and that an earlier elective birth be considered where there are maternal or fetal complications. India's Ministry of Health and Family Welfare guidelines say routine delivery before thirty-nine weeks is not recommended, that women with poor control or other complications should be delivered earlier, and that the timing should be individualised by the obstetrician — which is precisely the judgement your doctor is making and we are not. We have written separately about the reasoning behind early induction when growth is heading the other way, how timing is weighed in gestational diabetes and how long an induction actually takes.

Questions that tend to be worth asking at the next appointment:

  • Was the same sonographer, on the same machine, responsible for both scans?
  • What is the estimated weight with its error range, rather than the percentile alone?
  • What specifically is the earlier delivery intended to avoid, and what does waiting cost?
  • What would change the plan between now and then?
  • How do the cord loops factor in, if at all?

That last one is your obstetrician's call entirely, and there is no useful general answer to give you about it here.

The fear you are actually carrying

ACOG's patient guidance states that children of women who had gestational diabetes may be at risk of becoming overweight or obese during childhood, and have a higher risk of developing diabetes. That association is real, and it would be dishonest to wave it off.

Its shape, though, is not a verdict on your son. An individual participant data meta-analysis of 160,757 mother-child pairs from 34 cohorts, published in Lancet Child & Adolescent Health, found gestational diabetes associated with higher odds of overweight throughout childhood — but those associations attenuated towards the null once maternal body mass index was accounted for, and the authors concluded prevention should focus on maternal BMI rather than on pregnancy complications. In the HAPO Follow-Up Study, published in JAMA, the association with childhood overweight or obesity defined by body mass index was no longer statistically significant after full adjustment, while higher body fat percentage and sum of skinfolds remained associated, at odds ratios of 1.35 and 1.57.

Read those together and the picture is a modest, partly confounded, population-level signal — detectable across tens of thousands of children, invisible in any one of them. It is not a sentence passed on a particular boy because of four weeks of rice and chapati.

And the levers that matter most arrive after he does. The NIDDK advises that showing a child how to make healthy lifestyle choices helps the whole family and may protect him from developing obesity or diabetes later in life. Childhood is long, and most of it is still ahead of you both.

You are thirty-six weeks pregnant with a condition that was always going to tighten in the third trimester, reading a fuzzy number as a confession. Take the questions to your appointment. The guilt is not evidence.