Hyperechogenic fetal bowel: a prospective analysis of sixty consecutive cases
Correspondence: Dr M. D. Stringer, Department of Paediatric Surgery, Clarendon Wing, Leeds General Infirmary, Leeds LS2 9NS, UK.
A two year prospective analysis of all second trimester fetuses (16–22 weeks of gestation) with hyperechogenic bowel was undertaken. Hyperechogenic fetal bowel (sonographic echogenicity similar to or greater than that of surrounding fetal bone) was diagnosed using strict criteria. Outcome of affected fetuses was ascertained from hospital records, health care workers and autopsy reports, up to six months of age. Sixty consecutive fetuses were identified, of which 48 (80%) were liveborn. Six pregnancies were terminated, four ended with an intrauterine fetal death and two died at birth. The incidence of cystic fibrosis and aneuploidy were each 5%, and there were no cases of congenital infection. Intrauterine growth retardation was recorded in six fetuses (10%), four of whom died perinatally. Eighty-two percent of fetuses (28/34) with isolated hyperechogenic bowel had a normal outcome.
During the last 15 years hyperechogenic fetal bowel has become an established sonographic marker of fetal pathology. However, despite the publication of several large studies, there is no consensus about its prognostic significance. Most reports have been retrospective and have been flawed by inclusion of cases with different grades of echogenic bowel1,2, a failure to control for inter-observer variation3, and the study of a priori, high risk groups3. Only one large, prospective study in a low risk population has been published2. These issues cause considerable difficulties in providing accurate antenatal counselling. We report the results of a two year prospective analysis of second trimester fetuses with hyperechogenic fetal bowel managed in accordance with a defined clinical protocol.
Hyperechogenic fetal bowel was defined as bowel with ultrasonographic echogenicity similar to or greater than that of surrounding fetal bone. Detailed second trimester fetal anomaly scans were undertaken according to routine clinical practice. During each fetal scan where hyperechogenic bowel was suspected, the diagnosis was confirmed or rejected by the same two independent observers (G.C.M. and D.M.). Using an ATL Ultramark 9 machine and a 3.5 MHz transducer with a curvilinear array, images were stored on an Elonex Pentium computer with a DT 55LC frame grabber and Global Lab image analysis software. A previous study demonstrated that this approach resulted in an inter-observer agreement of 95% (Kappa statistic = 0.71) and intra-observer agreement of 98%–100% (Kappa = 0.88–1.0)4. From the outset of the study, fetuses with an abdominal wall defect, extraintestinal calcification, ascites or bowel which was less echogenic than that of surrounding fetal bone were not included in the analysis.
Between February 1996 and December 1997, clinical details of all second trimester fetuses (16–22 weeks of gestation) with hyperechogenic fetal bowel were recorded. These cases were detected either during routine antenatal scans at our institution or referred from regional obstetricians. All women were counselled and offered investigations in accordance with our routine clinical guidelines for investigating a fetus with hyperechogenic fetal bowel: serological testing for toxoplasmosis, cytomegalovirus and rubella, fetal karyotyping (amniocentesis or chorionic villus sampling) and parental blood DNA testing to determine the carrier status for cystic fibrosis. The latter was originally confined to delta F508, G551 D and R553X mutations, but after January 1997 laboratory screening was extended to include 12 of the more common cystic fibrosis mutations.
The outcome of all affected fetuses was ascertained from hospital records and by contact with referring consultants. In cases of termination of pregnancy or intrauterine fetal death, available autopsy reports were requested. For all liveborn infants, details of congenital anomalies, disturbances of gastrointestinal function, and information about growth and development were obtained from available records by one of the authors (I. G.) during follow up periods ranging from one to six months of age.
Sixty consecutive fetuses with hyperechogenic fetal bowel were identified during the two year period. Of these, 54 were singletons and six were one of twins or triplets. Maternal age range was 16 to 41 years, with a median of 28.5 years. Forty-eight pregnancies (80%) resulted in liveborn infants. Of the remainder, six were terminated, four resulted in intrauterine fetal deaths and two died within hours of delivery.
Termination of pregnancy was performed in six (10%) pregnancies: two for trisomy 21, one for trisomy 18, one for cystic fibrosis (delta F508 homozygous), one for severe oligohydramnios with intrauterine growth retardation and one for multiple congenital anomalies including a major cardiac malformation. Autopsy findings confirmed the diagnosis in all cases except for the fetus with cystic fibrosis who was not examined. There were four intrauterine fetal deaths (6.7%) at 23, 25 and 38 weeks, respectively, in singleton pregnancies and at 24 weeks in a twin pregnancy. Two of these deaths (24 and 38 weeks) were unexplained. The other two fetuses, both with normal karyotypes, were severely growth retarded; maternal systemic lupus erythematosus complicated one of these pregnancies. Two fetuses (3.3%) died following premature birth, one at 26 and the other at 32 weeks of gestation. In the first case autopsy was refused, but antenatal screening investigations including fetal karyotyping were normal. In the second case a complete autopsy showed marked intrauterine growth retardation and a phenotype consistent with Edward's syndrome, but the karyotype was normal (46XY).
Forty-three (90%) of 48 liveborn infants (24 males and 24 females) were normal at birth, as well as on follow up which range from one to six months. Five infants had the following abnormalities: a high anorectal malformation (n= 1), cystic fibrosis (n= 2), cerebral palsy (n= 1), and Nezelov's syndrome (n= 1). In the case of both infants with cystic fibrosis, their mothers had declined antenatal screening after hyperechogenic fetal bowel had been detected as an isolated finding. The infant with Nezelov's syndrome (autosomal recessive) came from a family where two siblings had died previously from die condition; this infant died at four months of age. Small-for-gestational-age infants (< 10th centile) were not observed more commonly than expected.
Hyperechogenic fetal bowel was seen as an isolated anomaly in 34 (57%) cases. Of these 28 (47%) were completely normal at birth and have remained well on follow up ranging from one to six months. Six fetuses had a variety of problems which included one unexplained intrauterine fetal death, one with trisomy 21, three with cystic fibrosis, and one with a high anorectal malformation (Table 1).
Table 1. Outcome of all fetuses with hyperechogenic bowel (n= 60) compared with those in whom this was an isolated finding (n= 34). Values are given as n (%).
|Normal||43 (72)||28 (82)||15 (58)|
|Perinatal and intrauterine fetal death||6(10)||1(3)||5(19)|
Serological testing was performed in 50 women and all investigations proved negative. Of the remaining 10 women, there was one intrauterine fetal death in a twin pregnancy and the surviving twin was normal at birth and on follow up. None of the liveborn infants have been reported to have signs or symptoms of congenital infection.
Forty eight women and their partners consented to cystic fibrosis screening and specific gene mutations were identified in two couples. In one, the fetus was found to be homozygous for delta F508 and the pregnancy was terminated. In the second, the fetus was found to be unaffected and remains well on follow up. From the 12 pregnancies in which cystic fibrosis screening was refused, there were two infants who were subsequently found to have cystic fibrosis. One required surgery for meconium ileus in the neonatal period, and the other was detected on routine postnatal screening. Both had isolated hyperechogenic fetal bowel and both are homozygous for delta F508.
Fetal karyotyping was performed in 31 cases, yielding three instances of aneuploidy (two trisomy 21 and one trisomy 18). Maternal age in these three cases was 19 years, 41 years and 41 years, respectively. Of the 29 fetuses not tested, one subsequently died from Nezelov's syndrome which is not associated with aneuploidy. Other than two cases of intrauterine fetal death and one termination of pregnancy where karyotyping was not possible (either antenatally or after attempted culture of fetal tissue), there was no suggestion of any other instance of aneuploidy.
Since the early descriptions of hyperechogenic fetal bowel the definition, assessment and clinical significance of this finding have remained controversial. Although several authors have identified a spectrum of increased fetal bowel echogenicity1,5,6, most have defined hyperechogenic fetal bowel as having an echogenicity similar to or greater than surrounding bone7. This is the definition which we adopted.
The ability to detect hyperechogenic fetal bowel is influenced by the type of sonographic equipment used and the number of observers. In the review by Bahado-Singh et al.3, all but one of the reported series had relied on more than one model of ultrasound machine combined with different transducers3. In none of the publications was the number of sonographers specifically stated. All cases of hyperechogenic fetal bowel in our analysis were regarded as being as echogenic as surrounding fetal bone by two experienced, independent observers whose inter- and intra-observer variation were validated by prior study. The same ultrasound machine and transducer were used in all cases. Our definition remains subjective but is more robust than that used by others.
The clinical significance of hyperechogenic fetal bowel is controversial. There has been only one truly prospective study2. Many reports have been biased by their focus on fetal populations at increased risk of aneuploidy and/or cystic fibrosis1,5,7. For example, Scioscia et al.5 reported a 27% incidence of aneuploidy in their study of hyperechogenic fetal bowel which was skewed towards pregnancies associated with advanced maternal age and/or abnormal maternal serum alpha-fetoprotein levels. This is much higher than the incidence of aneuploidy in other series8,9. All cases of hyperechogenic fetal bowel in our study were identified after routine second trimester ultrasound scans in a low-risk population.
The decision to offer routine serological testing for congenital infection, cystic fibrosis mutation analysis and fetal karyotyping to all women with affected fetuses was based on known associations1,5–9. We did not identify a single fetus with hyperechogenic fetal bowel and evidence of congenital infection (cytomegalovirus, toxoplasmosis or rubella), but this may reflect our small sample size. Muller et al.10 reported seven congenital infections (cytomegalovirus, toxoplasmosis and parvovirus) from a retrospective review of 182 fetuses with HEFB and Mac-Gregor et al.8 two instances in their review of 45 cases.
Three cases of cystic fibrosis were identified from among 60 fetuses with hyperechogenic fetal bowel, an incidence of 5%. This may be an under-estimate since not all couples consented to screening. However, this is unlikely since our detailed postnatal follow up identified two of these cases, which were from the small group who had rejected antenatal testing. In previous reports the incidence of cystic fibrosis in fetuses with hyperechogenic fetal bowel has varied between 0% and 60%5,10,11. Muller et al.10 identified only one case of cystic fibrosis among 182 fetuses with hyperechogenic fetal bowel; cystic fibrosis was excluded reliably by postnatal assessment in liveborn infants, but there was uncertainty about pregnancies ending in fetal death or termination. More recently, Sepulveda et al.12 suggested that hyperechogenic fetal bowel is not associated with an increased incidence of cystic fibrosis mutations in low risk pregnancies in London. However, Slotnick and Abuhamad2 in the United States found an incidence of 6.7% in their subgroup of fetuses with bowel as echogenic as bone. Such variations in incidence may reflect differences in a priori risks in study populations, screening techniques, the number of mutations studied and the length and accuracy of postnatal follow up. In addition, more densely hyperechogenic bowel may be associated with greater risk. The 5% incidence of cystic fibrosis in our cohort (compared with a UK population incidence of 1 in 2500 live births) justifies our decision to continue offering cystic fibrosis mutation analysis and to recommend postnatal screening of liveborn infants with hyperechogenic fetal bowel.
The overall incidence of aneuploidy in our series of fetuses with hyperechogenic bowel was 5%. Hyperechogenic fetal bowel was an isolated anomaly in one fetus with trisomy 21 (Table 1). Not all cases were karyotyped, but postnatal follow up did not suggest any additional chromosomal anomalies. Aneuploidy could not be excluded in three fetuses (two intrauterine fetal deaths and one termination of pregnancy). Our incidence figure for aneuploidy compares well with that of 4% reported by Muller et al.10 and 7.6% from Slotnick and Abuhamad2. Other authors have reported a higher incidence of 13%–27%, but these studies were biased toward pregnancies at higher risk of aneuploidy1,5.
Intrauterine growth retardation was recorded in six fetuses, one of which was a triplet and the other a twin. Intrauterine fetal death occurred in three of these growth retarded fetuses and a fourth died shortly after birth. The incidence of intrauterine growth retardation in our series (10%) is less than the 14% to 23% incidence reported by other authors3,7. The combination of hyperechogenic fetal bowel and intrauterine growth retardation appears to be associated with an adverse outcome since four of the six affected fetuses died. This is a complex area with probable links between hyperechogenic fetal bowel, uteroplacental insufficiency and functional intestinal obstruction.
This prospective analysis of 60 second trimester fetuses with hyperechogenic bowel confirms the associations with cystic fibrosis and aneuploidy previously suggested in retrospective studies. An objective measure of this sonographic fetal marker is needed to assist comparison between centres. Further studies are required to determine when hyperechogenic fetal bowel is a normal finding, the importance of persistent versus transient hyperechogenicity, the significance of hyperechogenic fetal bowel that is less echodense than surrounding bone, and the significance of luminal vs bowel wall hyperechogenicity.
The authors would like to thank Ms L. Ellis and Ms G. Linton at the Cytogenetics Department, St. James's University Hospital, and our regional obstetric and pae-diatric colleagues for their invaluable assistance with this project.