Please note: In 2003, the CTF updated its Grades of Recommendations to include an "I Recommendation" for situations where insufficient evidence exists to allow a recommendation to be made.  (Formerly, these situations were captured under a "C Recommendation".)  This change is not retroactive, and all "C Recommendations" made prior to 2003 have not been reevaluated in light of the new "I" recommendation grade.  For a discussion of these recommendation grades, please link to the 2003 article in the Canadian Medical Association Journal here.

Routine Prenatal Ultrasound Screening

Chapter Overview
Ultrasound examination has been suggested as a prenatal screening tool for various purposes, including the estimation of gestational age, the detection of multiple pregnancies and fetal anomalies, and the identification of intrauterine growth retardation (IUGR). The goal of prenatal ultrasound screening is to reduce the rates of perinatal illness and death from several causes, some of which (e.g., IUGR) are etiologically non-specific. Therefore, the Task Force has reviewed the evidence on the impact of prenatal ultrasound screening on measures of perinatal illness and death rather than on its ability to detect specific abnormalities. Ultrasound is also considered as part of the protocol in screening for neural tube defects (Chapter 7).
 

Use of Prenatal Ultrasound in Canada
The use of perinatal ultrasound has become increasingly common in Canada. Between 1981/82 and 1989/90 the number of prenatal ultrasound (PNU) services in Ontario and British Columbia more than doubled and the rate of PNU per delivery increased from 1.06 to 2.18 in Ontario and from 0.88 to 1.75 in British Columbia.<1> In 1989/90 an average of $160 was spent on PNU per delivery in Ontario and $130 per delivery in British Columbia. This rapid increase in the use and cost of ultrasonography underlines the need for careful assessment of the evidence on the benefits of this procedure.

Maneuver
A single ultrasound examination in the second trimester is used to estimate gestational age and to detect multiple pregnancies and malformations. Two serial examinations (one in the second trimester and one in the third) is used to screen for IUGR and to detect multiple pregnancies and malformations.

Effectiveness of Preventive Intervention
The published trials were separated into two groups. One included four trials that examined the impact of a single ultrasound examination performed in the second trimester.<2-6> The other group included four studies of the impact of serial examinations, one in the second trimester and one in the third.<7-10>
 

Single Ultrasound Scan
In the first trial<2,3> of this group, a total of 1,621 women underwent an ultrasound examination, including measurement of the biparietal diameter (BPD), at about 16 weeks’ gestation. Of these women, 836 were randomly allocated to an experimental group whose ultrasound results were released to the attending physician. The remaining 785 patients were allocated to a control group whose ultrasound results were not released to the attending physician. However, during the trial 30% of the control subjects had their ultrasound results released in response to specific requests from their physician because of clinical concerns. There were eight perinatal deaths in the experimental group and seven in the control group; this difference was not statistically significant. The Apgar score at 1 minute was 7 or less among 172 singleton infants in the experimental group and among 142 singleton infants in the control group; again, this difference was not statistically significant.

The second trial involved 4,997 women with no clinical indication for an ultrasound examination.<4> A total of 2,482 women were randomly assigned to undergo one ultrasound examination and BPD measurement at 15 weeks’ gestation (experimental group). The remaining 2,515 women did not have an examination before 19 weeks’ gestation (control group). Both groups then received the same antenatal care, including ultrasound examinations later in pregnancy. In the experimental group 32 (1.3%) of the women did not undergo the examination; in the control group 103 (4.1%) had the examination before the 19th week.

The experimental group had 3,068 ultrasound scans, as compared with 1,279 in the control group. Of the women in the control group 68% never underwent ultrasonography. The numbers of prenatal hospital days and hospital admissions were the same in each group. Labour was induced in 41 women in the experimental group and 88 in the control group (p=0.0001).

There were 12 perinatal deaths in each group. In the experimental group eight were in singleton infants and four in twins. In the control group singleton infants accounted for all of the deaths. After the perinatal period but before discharge from hospital there were two additional deaths in each group.

Analysis of the 4,776 singleton births showed no significant difference between the two groups in the proportion of infants with Apgar scores of 7 or less at 1 or at 5 minutes. In each group mechanical ventilation was required at delivery in seven cases and neonatal seizures occurred in four. There was a tendency for fewer babies in the experimental group than in the control group (231 vs. 275) to be admitted to the neonatal ward; however, this difference was not significant.

Singleton infants in the experimental group were 42 g heavier on average than those in the control group (p=0.008). The average birth weight of infants of nonsmoking women did not differ significantly between the two groups. Infants of smoking women in the experimental group were 75 g heavier on average than those of smoking women in the control group (p=0.013).

The third trial involved 9,310 women from Finland,<5> or about 95% of the pregnant women in the catchment area. A total of 648 women were not followed through to delivery: 569 had a miscarriage, 58 underwent an induced abortion, 17 were found not to be pregnant, and 4 were lost to follow-up.

There were 4,353 deliveries (4,317 of singletons and 72 of twins) in the experimental group. In this group ultrasound screening was done between 16 and 20 weeks’ gestation. The BPD was measured, the placenta located and the number of fetuses registered. Most of the 318 women who did not undergo the examination at the study hospital did so elsewhere. In the experimental group 1.6% of the women did not undergo the examination.

In the control group there were 4,309 deliveries (4,271 of singletons and 76 of twins). Although the women were not assigned to undergo an ultrasound examination 77% did so.

The overall perinatal death rate among the infants who were delivered was 4.6 per 1,000 in the experimental group and 9.0 per 1,000 in the control group (p=0.013). In the experimental group 18 singleton babies died (11 were stillborn and 7 died within 1 week after delivery). In the control group 34 singleton babies died (22 were stillborn and 12 died within 1 week after birth). Of the babies that died, only 2 (11%) in the experimental group had major anomalies, as compared with 10 (29%) in the control group. Eleven induced abortions were performed because of the ultrasound findings in the experimental group; there were no such abortions in the control group. There were two deaths in twins in the experimental group and five in the control group.

Among the singleton infants there was no significant difference between the two groups in 1) mean birth weight; 2) proportion of infants with a birth weight of less than 2500 g; 3) mean Apgar score at 1 minute; 4) proportion with an Apgar score of less than 7 at 1 minute; 5) rate of admission to special care unit; and 6) proportion with hospital stay of more than 5 days. The mean birth weight of the twins and the proportion of twins with a birth weight of less than 2500 g did not differ significantly between the two groups; however, all of the twins in the experimental group were detected by 21 weeks’ gestation, as compared with 76% of those in the control group (p=0.005).

The difference in the perinatal death rates between the two groups lost its significance when the induced abortions resulting from the ultrasound findings were included as deaths in the analysis. Also, 10 of the "malformations" detected at the ultrasound examination had disappeared by the time of the follow-up examination.

The final study took place in the U.S.<6> The trial involved 915 women at low risk who were randomly allocated to routine prenatal ultrasound screening at 10 to 12 weeks’ gestation or routine care. The two groups did not differ significantly in total adverse perinatal outcomes, as measured by the number of perinatal deaths, intensive care admissions and babies with an Apgar score of less than 6 at 5 minutes.

Serial Ultrasonography
The main features of the four trials in this group are summarized in Table 1. The first two trials involved random allocation of patients selected from the general population.<7,8> The third involved random allocation of women who had no clinical indications of IUGR.<9> The most recent trial excluded 60% of the general population.<10>

In the first study<7> three of the infants in the experimental group and eight in the control group died. Of the deaths in the experimental group one was intrauterine and of unexplained cause, and two were due to severe preeclampsia. In the control group four deaths were intrauterine and associated with IUGR, one was due to a severe malformation, and one was of unknown cause; the two postnatal deaths involved a premature twin and an infant with hydrops fetalis. There were no late neonatal deaths in the experimental group and three in the control group. The number of days of pediatric care for malformations due to "overdue pregnancy" (p<0.01) and for hyperbilirubinemia (p<0.05) was significantly lower in the experimental group than in the control group. The significantly fewer hospital admissions in the experimental group (p<0.01) did not result in a difference in the number of days of prenatal care (828 days in the experimental group and 829 in the control group).

There was little description of the causes of perinatal death in the second study.<8> Two of the deaths in the control group involved twins identified through ultrasonography at 24 weeks’ gestation. They were delivered 2 weeks later; one was stillborn, and the other died 2 hours postnatally. The distribution of birth weights did not differ significantly between the two groups. Forty-nine of the women in the control group were referred for ultrasonography as part of their routine care. The estimated cost of the screening program was U.S. $250 per pregnancy. Two-thirds of the cost was from the increased use of in-patient services by the experimental group.

The only perinatal death in the third study<9> involved a child born with open spina bifida and microcephaly. The mean birth weight was the same in the two groups. The number of inductions and the delivery methods did not differ significantly between the two groups.

The results of these three smaller trials of serial prenatal ultrasound screening are summarized in Table 2.

The fourth trial was based in 99 obstetrical and family practices in six different American states. A total of 55,744 women registered for the trial, 2,377 were lost before screening and 32,317 were excluded after screening. The major reason for exclusion was because the attending physician indicated that they planned to perform a prenatal ultrasound. The attending physician was not asked to indicate the clinical reason for the proposed ultrasound. The vast majority of the ultrasound exams were performed in one of the 28 laboratories participating in the trial. These laboratories used standardized reporting techniques and scanning equipment. In the screened group, 94% of the women had ultrasound examinations at both 15-22 weeks and 31-35 weeks. Only 2% of the women in the control group had ultrasounds at both of these times. The mean number of ultrasounds in the screened group was 2.2 and the mean number of ultrasounds in the control group was 0.6 with 55% of the group having no prenatal ultrasounds.

The trial used somewhat different outcome measures than the other 3 trials of serial PNU. Three main outcome measures were assessed; 1) perinatal mortality (fetal or neonatal death up to 28 days of age); 2) severe morbidity (including severe intraventricular hemorrhage, seizures, documented sepsis and prolonged use of special care nursery); and 3) moderate morbidity (including use of oxygen for more than 48 hours, nerve injury, short stay in special care nursery). There was no statistically significant difference between the control and screened groups for any of these outcomes (Table 3). There was a total of 52 perinatal deaths in the screened group with 48 of these deaths occurring in singeltons. There was a total of 41 perinatal deaths in the control group with 37 of these deaths occurring in singeltons. There were no statistically significant differences in gestational age at delivery or birth weight between the two groups.

A total of 350 fetuses had at least one major anomaly. Of the 187 fetuses with major anomalies in the ultrasound group 65 (35%) were detected before delivery and 31 (17%) were detected before 24 weeks. Of the 163 fetuses with major anomalies in the control group 18 (11%) were detected before delivery and 8 (5%) were detected before 24 weeks. There were 9 induced abortions for anomalies in the screened group and 5 in the control group.

Systematic Reviews of Trials
The Cochrane Database of Systematic Reviews has published reviews of routine prenatal ultrasound examination in early<11> and late<12> pregnancy. The review of early prenatal ultrasound trials concluded that screening results in early detection of twin pregnancies, a reduced rate of induction of labour, higher birth weights in singeltons and increased rates of abortion for fetal abnormalities, but no statistically significant effects on perinatal mortality or Apgar scores was found. The review of late prenatal ultrasound trials, which did not include the most recent American trial<10>, concluded that there was no statistically significant effects of late screening on mortality, morbidity or induction of labour. Another recent meta-analysis, which included the two large studies of single stage screening, concluded that there was no statistically significant effect of screening on live births, although there was a significantly lower perinatal mortality in the screened population.<13>

Recommendations of Others
In Canada, the Federal Task Force on High Risk Pregnancies and Prenatal Record Systems<14> in 1982 stated that "there seems to be very good evidence that ultrasound is a useful adjunct to clinical identification and assessment of intrauterine growth retardation" but that "the use of routine ultrasound without specific indications in pregnancy should be discouraged." The 1984 U.S. National Institutes of Health Consensus Conference on the use of diagnostic ultrasonography during pregnancy<15> concluded that "the data on the clinical efficacy and safety do not allow a recommendation for routine screening at this time." Recently the Society of Obstetrics and Gynecology of Canada recommended the routine use of single PNU in the second trimester.<16> The use of PNU is currently under review by the U.S. Preventive Services Task Force.

Conclusions and Recommendations
Although trials of a single PNU in the second trimester have not shown a statistically significant effect on the rate of live births or Apgar scores, these trials indicate that a single PNU early in pregnancy results in lower rates of induction (presumably through better estimates of gestational age), earlier detection of twin pregnancies, increased birth weight in singeltons and higher rates of therapeutic abortion for fetal abnormalities. Based on these clinical effects there is fair evidence to include a single PNU examination in routine prenatal care (B Recommendation). Trials of serial PNU have not shown any statistically significant effect on perinatal mortality, morbidity or birth weights. Based on this there is poor evidence for the inclusion or exclusion of serial PNU in routine prenatal care (C Recommendation).

Prenatal ultrasound examination may not only provide the clinician with information on perinatal anomalies and intrauterine problems but may also reassure the expectant mother and provide her with useful information.<17> On the other hand, false positive results can have an adverse psychologic effect on the expectant mother.

Unanswered Questions (Research Agenda)
The benefits and disadvantages of ultrasound examination require further analysis. Such analysis should focus on the impact of the maneuver on fetal survival and perinatal illness rates and should include bi-directional measurement of the psychologic effects of screening on the parents.

Evidence
To identify studies of the effectiveness of routine prenatal ultrasound screening we searched MEDLINE up to October 1993 (with the keywords ultrasonography and randomized controlled trial) for pertinent articles as well as using the references of those articles. Two criteria were used: (a) the study had to have allocated subjects randomly to undergo or not to undergo routine screening and (b) the outcomes had to have included measures of perinatal illness and death. This review was initiated in October 1993, and based on a previous report.<18> Recommendations were finalized by the Task Force in March 1994.

Full Citation
Anderson G. Routine prenatal ultrasound screening. In: Canadian Task Force on the Periodic Health Examination. Canadian Guide to Clinical Preventive Health Care. Ottawa: Health Canada, 1994; 4-14.