Ultrasound estimates of fetal weight are typically accurate to within about 10 percent of the baby’s actual birth weight, but that margin is an average that hides a lot of variability. One large study found the mean error was only around 2 percent, yet the spread around that average was wide enough that individual estimates could be off by considerably more. The accuracy depends on the baby’s size, when the scan happens relative to delivery, the mother’s body type, and who is holding the probe. For a measurement that drives major decisions about labor and delivery, understanding what ultrasound weight estimates can and cannot tell you matters more than most parents realize.
The Typical Margin of Error
When researchers compare ultrasound-estimated fetal weight to what the baby actually weighs at birth, the average error looks reassuringly small. A study of over 4,600 deliveries found the mean percent error was about 2.4 percent, which sounds excellent until you look at the standard deviation: roughly 9 percent.1PubMed Central. Percent error of ultrasound examination to estimate fetal weight at term in different categories of birth weight with focus on maternal diabetes and obesity That means while the average estimate lands close, plenty of individual estimates miss by 10, 15, or even 20 percent. In practical terms, for a baby who ends up weighing about 3,500 grams (roughly 7 pounds 12 ounces), the ultrasound guess could easily be off by 350 grams or more in either direction.
A separate systematic review confirmed that the main sources of inaccuracy include the sheer difficulty of getting precise measurements late in pregnancy, when there is less room for the baby to move and less fluid around it, along with differences in operator skill.2PubMed Central. The accuracy of ultrasound estimation of fetal weight in comparison to birth weight: A systematic review The upshot is that ultrasound weight estimates are a useful ballpark figure, not a precise measurement. Treating them as gospel can lead to unnecessary anxiety or, worse, clinical decisions based on a number that might be wrong.
Accuracy Breaks Down at the Extremes
The biggest limitation of ultrasound weight estimation is that it performs worst exactly where accuracy matters most: at the high and low ends of the birth-weight spectrum. That same large study found that for babies born in the 3,000 to 3,250 gram range (the sweet spot near average), the error was smallest. But for the smallest babies, ultrasound overestimated weight by about 5 percent on average, and for the largest babies, it underestimated by roughly 12 percent.1PubMed Central. Percent error of ultrasound examination to estimate fetal weight at term in different categories of birth weight with focus on maternal diabetes and obesity Both directions of error create real problems: overestimating a small baby’s weight can mask growth restriction, while underestimating a large baby’s weight can lead to unexpected complications during delivery.
For suspected large babies, the false-positive rate is strikingly high. A population-based cohort study found that about 36 percent of fetuses flagged as large-for-gestational-age on ultrasound turned out to be normal-sized at birth.3American Journal of Obstetrics and Gynecology. Accuracy of antenatal ultrasound in predicting large-for-gestational-age babies: population-based cohort study In one New Zealand cohort cited in the same study, the number was even worse: 58 percent of fetuses estimated above the 90th percentile were not actually large at birth. That means more than half of the “big baby” calls were wrong. At the small end, around half of truly small-for-gestational-age babies were missed on scan and categorized as normal.
Detecting Small Babies Is Harder Than You Would Think
Growth restriction is one of the leading risk factors for stillbirth, so catching it matters enormously.4PubMed. Antenatal detection of fetal growth restriction and risk of stillbirth: population-based case-control study Yet ultrasound’s ability to identify small babies is surprisingly modest. A meta-analysis pooling data from multiple studies found that a 36-week ultrasound picks up only about 48 percent of small-for-gestational-age babies when using an estimated weight below the 10th percentile as the cutoff. At 32 weeks, that drops to around 36 percent.5PubMed. Diagnostic performance of 32 vs 36 weeks ultrasound in predicting late-onset fetal growth restriction and small-for-gestational-age neonates: a systematic review and meta-analysis Specificity is high, over 93 percent at both time points, meaning false alarms for smallness are rare. But missing roughly half of truly small babies is a significant limitation.
When ultrasound is done routinely on everyone rather than only on pregnancies already flagged for concern, the detection rate improves. One Scandinavian study found that routine screening caught about 67 percent of small fetuses, compared with only 34 percent when ultrasound was ordered based on clinical suspicion alone.6PubMed Central. Detection and clinical outcome of small‐for‐gestational‐age fetuses in the third trimester—A comparison between routine ultrasound examination and examination on indication The reason is straightforward: clinical palpation (feeling the belly) misses many small babies, so if the only pregnancies scanned are those where something seems off by hand, a lot of growth-restricted fetuses fly under the radar. Routine third-trimester scanning catches more of them, though it is not standard practice in every country.
Why the Estimate Gets Less Accurate Over Time
Ultrasound captures a snapshot of the baby’s size at the moment of the scan. Between that moment and birth, the baby keeps growing, and how much it grows is anyone’s guess. A study of small-for-gestational-age fetuses found that even after statistically adjusting for known growth rates, an estimate taken 7 days before delivery differed from birth weight by an average of about 111 grams. At 14 days before delivery, the gap widened to roughly 200 grams.7PubMed. Comparing the relation between ultrasound-estimated fetal weight and birthweight in cohort of small-for-gestational-age fetuses That second number amounts to nearly half a pound of discrepancy introduced purely by the passage of time.
Another study looking at scans done very close to delivery found that accuracy held reasonably steady for about the first three days, but beyond that, the estimate became less reliable.8PubMed. Influence of ultrasound-to-delivery interval and maternal-fetal characteristics on validity of estimated fetal weight The practical takeaway: a weight estimate from a scan done two weeks before your due date is already somewhat stale. If your provider is making delivery decisions based on fetal weight, the timing of the scan relative to the decision matters. A scan from four weeks earlier should be interpreted with more caution than one from two days ago.
Maternal Body Type Affects Image Quality
Ultrasound works by bouncing sound waves through tissue. The more tissue those waves have to travel through, the more the signal degrades. Higher maternal body mass index makes it harder to get clean images of fetal anatomy, and this directly affects the accuracy of weight estimates. A study of extremely preterm infants found that the mean percentage error between estimated and actual weight was about 6.7 percent in women with normal pre-pregnancy weight, but rose to 9.5 percent in women with obesity.9American Journal of Obstetrics & Gynecology MFM. Accuracy of estimated fetal weight in extremely preterm infants and the impact of prepregnancy body mass index The difference was statistically meaningful, and it compounds the challenges that already exist at extreme fetal weights. If you are in a higher BMI category, it is worth knowing that your weight estimates carry a wider band of uncertainty.
The Person Holding the Probe Matters
Ultrasound is an operator-dependent technology. The accuracy of the weight estimate depends partly on who takes the measurements and how precisely they place the calipers on the screen. Research has shown that fetal presentation (whether the baby is head-down, breech, or transverse) and the sonographer’s experience both influence the estimate, and their effects vary depending on which mathematical formula is used to convert the measurements into a weight.10PubMed. Sonographic fetal weight estimation – is there more to it than just fetal measurements?
Counterintuitively, dedicated sonographers who perform scans all day long tend to outperform physicians. One study compared registered sonographers with maternal-fetal medicine specialists and found that 70 percent of the sonographers’ predictions fell within 10 percent of actual birth weight, versus only 54 percent of the specialists’ estimates. The sonographers were also significantly better at detecting growth restriction.11PubMed. Sonographic estimate of birth weight: relative accuracy of sonographers versus maternal-fetal medicine specialists The likely explanation is repetition: someone who takes fetal measurements dozens of times a day develops a consistency that is hard to match in a physician who divides time between scans and other clinical work. If you are given a choice between a dedicated ultrasound technologist and a quick scan by a busy clinician, the technologist’s estimate may be more reliable.
Fetal Anatomy and Fluid Levels Add Noise
Ultrasound weight formulas rely on a handful of fetal body measurements, typically head circumference, abdominal circumference, and thigh-bone length. The formulas assume that babies have roughly standard proportions. When they do not, accuracy suffers. Research has shown that babies with unusually narrow or wide heads (measured by the cephalic index) throw off estimates: those with narrow heads tend to have their weight underestimated, while those with rounder heads get slightly overestimated.12PubMed. Accurate ultrasonic estimation of fetal weight. Effect of head shape, growth patterns, and amniotic fluid volume Variations in skull shape can also affect estimates of gestational age, which in turn affects whether a baby is categorized as small or large for its age.13PubMed. Estimating fetal age: effect of head shape on BPD
Amniotic fluid volume plays a role too. When fluid is abnormally low (oligohydramnios), the baby’s limbs and body are compressed into positions that make measurements harder, and weight tends to be underestimated. Large-for-gestational-age babies were also slightly underestimated in the same analysis, which aligns with the broader pattern of accuracy degrading at the extremes. Fetal sex, interestingly, did not affect the accuracy of the estimate.12PubMed. Accurate ultrasonic estimation of fetal weight. Effect of head shape, growth patterns, and amniotic fluid volume
Population-Specific Growth Standards Change the Picture
Most of the weight-estimation formulas used worldwide were developed from data sets that skewed heavily toward White European or North American populations. When applied to other groups, they can systematically mislabel healthy babies as abnormally small or large. The NICHD Fetal Growth Studies found that using a standard derived solely from White pregnancies would erroneously classify as much as 15 percent of non-White fetuses as growth-restricted.14PubMed Central. Racial/Ethnic Standards for Fetal Growth, the NICHD Fetal Growth Studies That is a staggering rate of misclassification, potentially triggering unnecessary interventions or surveillance for babies who are growing normally for their background.
Part of the issue is that fetal body proportions differ across populations. One study found that Black fetuses had smaller abdominal circumferences than White fetuses between 17 and 23 weeks, the window when anatomic surveys are typically performed. Because abdominal circumference contributes heavily to the weight calculation, this difference could lead to overestimating growth restriction in Black patients.15Journal of Maternal-Fetal and Neonatal Medicine. Fetal biometry: Does patient ethnicity matter? Maternal height, weight, parity, ethnic group, and smoking status all have significant, independent effects on birth weight.16PubMed. An adjustable fetal weight standard Ethnicity-specific or customized growth charts are increasingly recommended to reduce the false-alarm rate, but they are still not used universally.
Ultrasound Versus Feeling the Belly
Before ultrasound, the main way to estimate fetal weight was clinical palpation, where a trained provider feels the uterus through the abdomen and makes a judgment call. A meta-analysis comparing the two found that ultrasound was significantly more likely than clinical palpation to land within 10 percent of the actual birth weight.17Acta Obstetricia et Gynecologica Scandinavica. Comparing the accuracy of maternal, clinical, and ultrasound estimations to predict birthweight: a meta-analysis Perhaps more surprising, the same meta-analysis found no significant difference between clinical palpation and mothers’ own guesses. Both were in the same accuracy range for normal-weight pregnancies. The advantage of ultrasound over palpation is real but modest, and it grows at the extremes of fetal weight where manual estimation becomes even less reliable.
MRI Is More Accurate but Rarely Used
Magnetic resonance imaging can estimate fetal weight more precisely than standard two-dimensional ultrasound. A review of roughly ten comparative studies found that MRI carries a mean relative error of about 2.6 to 3.7 percent when performed within a week of delivery, compared with 6.3 to 11.4 percent for ultrasound on the same babies.18PubMed. Magnetic resonance imaging for prenatal estimation of birthweight in pregnancy: review of available data, techniques, and future perspectives The difference is even more dramatic when you look at extreme weights. In a series of 270 fetuses scanned within 48 hours of birth, MRI detected 98 percent of large-for-gestational-age neonates at a fixed 10 percent false-positive rate, compared with 67 percent for ultrasound. For small-for-gestational-age babies, MRI hit 100 percent detection versus 78 percent for ultrasound.
Multiple individual studies have confirmed this advantage.19PubMed. Comparison of magnetic resonance imaging to ultrasound in the estimation of birth weight at term20PubMed. Optimization of fetal weight estimates using MRI: comparison of acquisitions So why isn’t MRI the standard? Cost, access, and time. An MRI scanner is far more expensive to operate, the exam takes longer, and availability is limited in many hospitals. MRI also cannot be performed in real time at the bedside the way ultrasound can. For most pregnancies, the added precision does not change management enough to justify the expense. But for cases where the weight estimate is genuinely driving a major decision, such as whether to attempt vaginal delivery of a suspected macrosomic baby or to deliver a very preterm infant early because of growth restriction, MRI can provide a meaningfully better number.
What Happens When the Estimate Is Wrong
Inaccurate weight estimates are not just a trivia problem. They drive clinical decisions with real consequences. A retrospective study of over 38,000 estimated fetal weights found that about 17 percent were overestimates and 13 percent were underestimates (defining accuracy as within 10 percent of birth weight). Overestimation was independently associated with higher rates of both labor induction and cesarean delivery. Underestimation was linked to increased odds of shoulder dystocia and severe perineal tearing.21PubMed Central. Factors Affecting Clinical over and Underestimation of Fetal Weight—A Retrospective Cohort
Being told your baby is very large can set off a cascade of interventions. Women with a macrosomia diagnosis (estimated weight above 4,000 or 4,500 grams, depending on the guideline) were more likely to have elective cesarean sections or premature induction of labor.22PubMed. Fetal macrosomia in non-diabetic mothers: antenatal diagnosis and delivery outcome Given the high false-positive rate for large-baby predictions discussed earlier, a substantial number of those interventions happened for babies who were ultimately normal-sized. On the other end, missing a growth-restricted baby deprives the care team of the chance to monitor more closely, adjust timing of delivery, and reduce the risk of stillbirth.
The American College of Obstetricians and Gynecologists acknowledges these limitations, noting that while the risks of labor complications rise with increasing birth weight, the methods available to estimate that weight have meaningful limitations.23PubMed Central. Macrosomia: ACOG Practice Bulletin, Number 216 In practice, this means that a single weight estimate should rarely be the sole basis for a major intervention. It is one data point among many, including maternal history, glucose control, prior delivery history, and the baby’s growth trend across multiple scans.
How to Interpret Your Own Scan Results
If you have a late-pregnancy ultrasound coming up or just got one, a few things are worth keeping in mind. First, the weight number on your report is an estimate, not a measurement. Unlike weighing something on a scale, ultrasound calculates weight from a formula that converts a handful of linear measurements into a single number. Small errors in those measurements get amplified by the formula. Second, estimates at the extremes carry more uncertainty than estimates in the middle. If you are told the baby is measuring very large or very small, the probability that the actual weight is significantly different from the estimate is higher than if you are told the baby is average.
Third, ask when the scan was done relative to the decision being discussed. A growth scan from three or four weeks ago is considerably less reliable than one from yesterday. If an intervention is being recommended based on an older estimate, it is reasonable to ask whether a repeat scan would help. Fourth, trends matter more than single numbers. A baby that has been tracking along the 30th percentile for months and is still at the 30th percentile is less concerning than one that drops from the 50th to the 10th in two weeks, even if the absolute estimated weight looks similar. Your provider should be looking at the trajectory, not just the latest number.
Finally, remember that about 70 percent of ultrasound weight estimates land within 10 percent of the actual birth weight. That leaves roughly 30 percent that do not, and there is no reliable way to know in advance whether yours is in the accurate group or the outlier group. Treat the number as helpful guidance, not as a certainty, and do not be afraid to ask your provider how confident they are in the estimate and what the plan would be if the actual weight turns out to be different.