Association of isolated short femur in the mid-trimester fetus with perinatal outcome
To evaluate the prevalence of fetal isolated short femur in a cohort of women screened for Down syndrome by the integrated test, and to compare the outcome of fetuses with isolated short femur in the mid-trimester with that of fetuses with normal femur length (controls).
This was a retrospective cohort study of 1262 women booked for antenatal care and delivery at University College London Hospital. All women had integrated testing in the late first and early second trimesters and a detailed anomaly scan in the mid-trimester. All scan reports, screening results and neonatal data were analyzed statistically.
The fetal femur was short (< 5th percentile) in 5.1% of patients and 4.7% had isolated short femur. In pregnancies with isolated short femur, the birth weight was significantly lower and there were higher rates of small-for-gestational age (SGA) and low birth weight (LBW) infants, compared with controls (P < 0.01). The odds ratios for SGA and LBW in pregnancies with isolated short femur were 3.0 (95% CI, 1.5–5.9) and 2.60 (95% CI, 1.1–6.2), respectively. Isolated short femur was associated significantly with low levels of pregnancy-associated plasma protein-A (P = 0.001).
Isolated short femur in the mid-trimester fetus is associated with fetal growth restriction and SGA. In the context of normal Down syndrome screening and a normal anomaly scan, this marker should be regarded as a predictor for SGA, and fetal growth should be monitored during these pregnancies. Copyright © 2008 ISUOG. Published by John Wiley & Sons, Ltd.
The detection of a fetal femur length below the expected value (< 5th percentile) at mid-trimester ultrasound examination is a diagnostic challenge: it may be a normal finding or it may be a marker of aneuploidy (trisomy 21) or associated with a skeletal dysplasia. Some studies have shown that a short femur is a feature of intrauterine growth restriction (IUGR)1 and small-for-gestational age (SGA) neonates2. It has been suggested that short femur might be associated with placental insufficiency and altered biochemical markers. However, these studies were descriptive case series3 or were based on referrals to fetal medicine units1, which may introduce bias.
The aims of this study were: to assess the prevalence of short femur in the mid-trimester fetus in a cohort of women screened for Down syndrome by the integrated test, and to compare the outcome of fetuses with isolated short femur in the mid-trimester with that of fetuses with normal femur length.
This was a retrospective cohort study of women booked for antenatal care at University College London Hospital (UCLH) between January 2003 and April 2004. Women whose fetus had a structural or chromosomal abnormality and those with multiple pregnancies were excluded from the study.
All women underwent screening for Down syndrome by the integrated test4 (measurement of first-trimester maternal pregnancy-associated plasma protein-A (PAPP-A) and fetal nuchal translucency thickness (NT) and second-trimester biochemistry (maternal serum alpha-fetoprotein (α-FP), beta-human chorionic gonadotropin (β-hCG), unconjugated estriol (uE3) and inhibin-A)). Fetal NT was measured routinely in the ultrasound department at 11 to 13 + 6 weeks (CRL, 44–84 mm) as previously described5. All were performed by a qualified ultrasound technician (sonographer or midwife) who had been trained by one of the authors (P.P.P.). Following the NT measurement, blood was taken for routine obstetric care (booking bloods) and for the PAPP-A measurement. All biochemical analysis was performed at the Wolfson Institute of Preventive Medicine, Bart's and the London School of Medicine and Dentistry. Gestational age was determined by the last menstrual period. In cases in which the gestational age was discordant with the first-trimester crown–rump length (CRL) by more than 7 days, the gestational age was based on CRL. Patients returned at or after 15 weeks of gestation for the second component of the integrated test (biochemical markers). Results of the biochemical markers were absolute concentrations, and calculated multiples of the median (MoM), NT-MoM and risk of Down syndrome were obtained using Alpha software (Logical Medical Systems, London, UK).
All patients had a routine mid-trimester (gestational age, 18–22 weeks) detailed anomaly scan performed in the ultrasound department by a qualified sonographer. In cases in which the patient had a high risk for chromosomal abnormality (based on the integrated screening test) or when a structural anomaly was suspected, the ultrasound examination was repeated by a fetal medicine specialist. Ultrasound examination details were entered into a computer database (Viewpoint software, General Electric, Wauwatosa, WI, USA) at the time of the examination, from where we retrieved them for this study. A short femur was defined as a femur length below the 5th percentile for gestational age6. Isolated short femur was diagnosed when the fetus had a normal (> 5th percentile) abdominal circumference (AC)7 and estimated fetal weight (EFW)8. Early-onset IUGR was defined as EFW or AC < 5th percentile at the mid-trimester anomaly scan.
Perinatal outcome characteristics including gestational age at delivery, birth weight, Apgar score and occurrence of pregnancy-induced hypertension (PIH) or pre-eclampsia was retrieved from the electronic institutional obstetric database (Patient Administration System). PIH was defined as two blood-pressure readings above 140/90 mmHg, taken at least 4 hours apart and newly diagnosed after 20 weeks' gestation. Pre-eclampsia was defined as PIH and proteinuria, with either > 300 mg protein in a 24-h urine collection or ‘+ + ’ at urinalysis with a laboratory-confirmed sterile urine culture. Gestational hypertension was defined as PIH or pre-eclampsia. The case notes of all women defined as having PIH, pre-eclampsia, an SGA fetus (birth weight < 10th percentile8, 9) or a short fetal femur length were examined to confirm the diagnosis. Low birth weight (LBW) was defined as birth weight < 2500 g.
Analysis was performed using SPSS version 11.5 statistical software (SPSS Inc., Chicago, IL, USA). The significance level was set to P ≤ 0.05. Continuous variables (birth-weight and serum markers) were analyzed by one-way ANOVA. Categorical variables (SGA, LBW, gestational hypertension, preterm delivery, low Apgar score) were analyzed by Fisher's exact test.
The study, being part of a clinical audit, was exempt from UCL institutional review board approval.
During the study period, 1651 women underwent integrated screening for Down syndrome, of whom 1262 (76%) had an anomaly scan and delivered at UCLH. There were 22 women excluded from our analysis, 18 due to aneuploidy or major structural anomalies and four due to incorrect dating. There were no fetuses subsequently identified with a skeletal dysplasia. Of the 1240 women included in this study, 1177 women (94.9%) had a femur length above the 5th centile for gestational age (normal femur). In 63 women (5.1%) the fetus was found to have a short femur at the anomaly scan: 58 (4.7%) were classified as isolated short femur and five (0.4%) fetuses had short femur and early-onset IUGR, with AC < 5th centile (n = 1), or EFW < 5th centile (n = 3) or both (n = 1). There was no difference in the mean maternal age of women carrying a fetus with isolated short femur compared with normal femur length (33.1 vs. 32.5 years, P = 0.38). There was no difference in the ethnic origin of women carrying a fetus with isolated short femur compared with normal femur length (Caucasian, 71.9% vs. 67.7%; Afro-Caribbean, 4.5% vs. 5.4%; South Asian, 3.8% vs. 5.1%; Oriental, 2.2% vs. 2.9%; all P > 0.5). The ethnic origin was not detailed or stated as ‘Other’ in 17.9% of cases with a fetus with isolated short femur and 18.7% of cases with normal femur length (P = 1.0).
Women carrying a fetus with isolated short femur were significantly more likely to deliver an infant with a lower mean birth weight by an average of 194 g (P = 0.008), SGA (P = 0.004) and LBW (P = 0.04), than were women carrying a fetus with normal femur length (Table 1). The odds ratio for SGA and LBW in the isolated short femur group were 3.0 (95% CI, 1.5–5.9) and 2.60 (95% CI, 1.1–6.2), respectively. After excluding 63 patients who delivered preterm (< 37 completed weeks of gestation) these differences still remained significant (P < 0.001 for SGA and P = 0.001 for LBW). When short femur was not present, the negative predictive value (NPV) was 93%.
Table 1. Perinatal outcome in fetuses with short and normal-length femur
|Birth weight (g)||3380 ± 541||3057 ± 703||3186 ± 522||0.008‡|
|Small-for-gestational age||85 (7.2)||16 (25.4)||11 (19.0)||0.004§|
|Low birth weight||50 (4.2)||11 (17.5)||6 (10.3)||0.043§|
|Gestational hypertension*||61 (5.2)||7 (11.1)||3 (5.2)||1.0§|
|Spontaneous delivery < 37 weeks||32 (2.7)||2 (3.2)||2 (3.4)||0.670§|
|5-min Apgar score < 7||21 (1.8)||6 (9.5)||2 (3.4)||0.294§|
Sixty-eight patients (5.5%) developed either PIH (n = 32) or pre-eclampsia (n = 36). These women delivered significantly earlier (mean gestational age at delivery, 37.1 ± 2.7 weeks vs. 39.5 ± 21.0 weeks; P = 0.001) and had a larger proportion of deliveries before 34 weeks of gestation (10.2% vs. 1.2%, P = 0.001). Isolated short femur was not associated with gestational hypertension. Interestingly, women carrying a fetus with short femur and early onset IUGR had a highly significant risk of having gestational hypertension (P < 0.001, Fisher's exact test), with four of these five patients developing pre-eclampsia or PIH (odds ratio, 73.2; 95% CI, 10.7–492.6).
There was one case of neonatal death, due to asphyxia, in the isolated short-femur group (2650-g fetus delivered at 38 weeks of gestation with Apgar scores of 3 and 4 at 5 and 10 min, respectively, and severe acidemia). The rate of spontaneous preterm delivery before 37 weeks of gestation was similar between the isolated short-femur and normal groups (2.7% vs. 3.4%, P = 0.67). Women carrying fetuses with isolated short femur were more likely to deliver between 34 and 37 weeks (4/58 (6.9%) cases with isolated short femur were induced before 37 weeks of gestation compared with 23/1154 (2.0%) cases with normal femur length; Fisher's exact test, P = 0.034). Isolated short femur was not associated with a 5-min Apgar score below 7.
When compared with controls, women carrying a fetus with short femur had significantly lower levels of PAPP-A (0.90 ± 0.46 vs. 1.19 ± 0.89 MoM, P < 0.001) and uE3 (0.95 ± 0.28 vs. 1.02 ± 0.26 MoM, P = 0.023), and higher levels of inhibin-A (1.35 ± 1.11 vs. 1.07 ± 0.59 MoM, P < 0.001). However, this was mostly due to the group of short-femur fetuses with early onset IUGR. Comparing the cases of isolated short femur with the group with normal femur length, PAPP-A was the only serum marker to be significantly lower (Table 2). The NT-MoM did not differ between patients with isolated short femur and those with normal femur length (1.07 vs. 1.03, P = 0.34).
Table 2. Maternal serum markers of the integrated test in fetuses with short and normal-length femur
|PAPP-A||1.19 ± 0.89||0.90 ± 0.46||0.87 ± 0.45||1.01 ± 0.71||0.001|
|α-FP||1.09 ± 1.01||1.31 ± 1.66||1.10 ± 0.52||4.19 ± 5.5||0.973|
|uE3||1.02 ± 0.26||0.95 ± 0.28||0.96 ± 0.25||0.71 ± 0.31||0.099|
|β-hCG||1.27 ± 0.93||1.50 ± 1.07||1.36 ± 0.94||2.88 ± 1.30||0.510|
|Inhibin-A||1.07 ± 0.59||1.35 ± 1.11||1.20 ± 0.70||3.40 ± 2.74||0.123|
This study describes the association of isolated short femur in the mid-trimester fetus with perinatal outcome in an unselected population that has been screened for Down syndrome. In our study, isolated short femur at the mid-trimester was associated with a smaller birth weight and an increased risk of LBW and SGA infants. Early-onset IUGR in the mid-trimester fetus with short femur and small AC was significantly associated with gestational hypertension and poor fetal outcome.
Some authors have advocated the use of ethnic-specific femur-length growth charts in a multiethnic population10–12. This is mainly due to the shorter femur length attributed to Asian and oriental populations. Due to the retrospective nature of this study, we did not use an ethnic-specific chart for each case; this may have introduced bias and is therefore a limitation of the study. However, the majority of patients in this study, both in the short-femur and normal femur-length groups, were Caucasian in origin. Moreover, the distribution of the other ethnic groups (including the Asian-Oriental groups) did not differ with respect to femur-length group.
The strength of our study lies in the fact that is was a large cohort rather than a case–control study. Several case series have shown an association between short femur and IUGR. Bromley et al.2 reported four fetuses with a short femur measuring more than 2 SD below the mean compared with the biparietal diameter. These fetuses were subsequently found to have severe IUGR with no evidence of skeletal dysplasia. Two other studies have described a short femur in cases with IUGR. O'Brien and Queenan13 reported delayed growth in femur length measurements in seven growth-restricted fetuses and Woo et al.14 observed that 12 of 30 IUGR fetuses had a femur length below 2 SD for gestation. Two studies have described referrals to a fetal medicine unit for short femur length. Todros et al.1 reviewed 86 consecutive referrals for femur length below the 10th percentile, finding that 21% of the women delivered structurally normal, euploid, SGA neonates and the diagnosis of SGA fetuses was made about 9 weeks after the finding of short femur. Interestingly, a fifth of cases with short femur resulted in SGA fetuses, despite normal uterine and umbilical Doppler studies at 22–24 weeks. Our study supports this finding of about 20% SGA in cases with isolated short femur, but it did not incorporate Doppler analysis.
We found that cases of isolated short femur were associated with lower mean birth weight, LBW and SGA but not with gestational hypertension. Compared with fetuses with normal femur length, these cases of isolated short femur were associated with significantly lower levels of PAPP-A (but there was no difference in levels of β-hCG, inhibin-A or α-FP). PAPP-A is a protease secreted by human fibroblasts which acts on insulin-like growth factor-binding protein-4 and increases its bioactivity15. Recent studies have shown that PAPP-A levels (at 11–14 weeks) are associated with femur length measured at the end of the first16 and second17 trimesters.
In our study, early-onset IUGR, with short femur as one of its features, was associated with poor outcome; two of the five cases resulted in extremely LBW neonates (< 1000 g). However, this group was too small to allow definitive conclusions to be drawn. Since these fetuses are likely to be diagnosed in the mid-trimester, they should be offered close follow-up with frequent assessment of Doppler measurement and growth. We found that isolated short femur was not associated with an increased risk of pre-eclampsia or PIH. In fact, the only perinatal outcomes associated with isolated short femur were mean birth weight, LBW and SGA. The positive predictive value (PPV) and NPV of short femur for SGA (PPV, 25%; NPV, 92%) or isolated short femur for SGA (PPV, 19%; NPV, 92%) are similar to those described for abnormal uterine artery Doppler in mid-gestation (PPV, 5–36%; NPV, 92–97%)18–21. Therefore, fetuses with short femur at mid-trimester might be considered as having an increased risk of IUGR, even in an unselected population.
The etiology of femur shortening in SGA fetuses is unclear. A study of body proportionality of SGA fetuses found that in most cases of SGA due to abnormal placentation, the reduced femur length was concordant with the small AC22. This might be explained by redistribution of blood flow, with increased flow to the heart and brain and decreased flow to the lower body23, 24. Another study described nine cases, referred due to short long bones, which resulted in IUGR fetuses; three of them had chronic hypertension and one had severe PIH leading to termination of pregnancy3. Most (6/9) of these cases were associated with elevated β-hCG levels at mid-trimester. The authors speculated that abnormal placentation altered the level of fibroblast growth factor receptor-2, resulting in growth restriction of the fetal long bones. It is likely that IUGR and short femur are linked to pre-eclampsia by the common mechanism of placental insufficiency.
This cohort study demonstrates that short femur, even as an isolated finding at mid-trimester, is associated with smaller and SGA fetuses. We believe that a finding of short femur at mid-trimester should be regarded as a risk factor for SGA and possibly pre-eclampsia, and that these patients should have fetal growth scans throughout pregnancy.