In our last post, following a distraught patient’s request and one of our own reader’s queries, and at the risk of putting myself out of a job, we began to address the issue of safety of obstetrical ultrasound procedures for the developing baby. I will admit at the outset, that I do not know. We have used ultrasound in obstetrics for 40+ years throughout its remarkable evolution of technological advances. There has been, and currently is moreso, widespread use of ultrasound beginning in early first trimester (sometimes with multiple early scans during the critical periods of embryogenesis) and throughout pregnancy and yet there are no convincing reports (to date) of specific, reproducible, visible, physical birth defects associated with its use.
The overall impression has been and continues to be, in the eyes of patients and providers, that ultrasound is, generally, safe and the potential benefits of ultrasound-based procedures far outweigh the small risks and the downsides of other options currently available for fetal evaluation. But, it is also true that over the years, despite the fact that most babies don’t lose their fingers, toes, or genitalia following ultrasound procedures, questions have been raised about the possibility of more subtle ‘bioeffects’ of even the low-level ultrasound intensities used in B-mode scanning. Hence the question of our astute (and most loyal!) reader regarding lefthandedness and ultrasound exposures.
The first study of which I am aware that raised the issue of an association between ‘lefthandedness’ and in utero ultrasound exposure was published in 1993 by Salvesen and colleagues (BMJ 1993;28:159-64). This study reported the follow-up of 2161 eight and nine year old children whose mother’s participated in two randomized, controlled trials of routine ultrasonography during pregnancy in Norway from 1979-1981. Using questionnaires and an abbreviated version of the Denver developmental screening test, handedness and developmental milestones in this cohort of children were retrospectively assessed. Although they found “no clear differences…between the groups with regard to deficits in attention, motor control, and perception or neurological development during the first year of life,” they did find that the odds of lefthandedness “were higher among children who had been screened in utero than among the control children (odds ratio 1.32; 95% confidence interval 1.02 to 1.71). The authors were fast to point out, however, that even though their data suggested a “possible association…the observed results may be due to chance.” A subsequent report by this same group (Salvesen, et al., Ultrasound Obstet Gynecol 1999;13:241-6) presented a metaanalysis of two follow-up studies from their three randomized (ultrasound vs no ultrasound) trials of 4715 children and concluded that “there was no statistically significant difference in the presence of non-right handedness between the ultrasound-screened children and the controls, but there was a statistically significant difference…among the boys.” Again, they stated “the results …must be interpreted with caution.”
A subsequent study out of Sweden (Kieler, et al., Epidemiology 2001;12:618-23) looked at a cohort of men born between 1973 and 1978 and estimated relative risks for being born left-handed according to in utero ultrasound exposure. Risk estimates were based on 6,858 men born in hospitals “that included ultrasound scanning in standard antenatal care (exposed) and 172,537 men born in hospitals without ultrasound scanning programs (unexposed).” They found no differences between the groups during the early years (1973 to 1975) of ultrasound use, but “when ultrasound was offered more widely (1976 to 1978; multiple scans during pregnancy), the risk of left-handedness was higher among those exposed to ultrasound…(odds ratio = 1.32, 95% CI = 1.16-1.51).” Such a study cannot be completely ignored but, even if compelling by virtue of the number of individuals evaluated, it suffers from the faults of any retrospective analysis. For example, not analyzed is the ‘handedness’ and socioeconomic status of the parents of the children born at the different centers; nor are the indications for birth at an ‘ultrasound center’ adequately differentiated from those indications for birth at centers that did not have ultrasound.
Yet to be adequately explained is why this difference only was found in the subgroup of randomized males and not in the entire population of ‘exposed vs nonexposed’ children (Kieler, et al., Early Hum Dev 1998;50:233-45). It should also be noted that subsequent studies by this group involving 3,265 eight to nine year old children whose mothers participated in a randomized (ultrasound vs. no ultrasound) study between 1985-1987 showed no significant differences in behavioral disorders or impaired neurologic development (Kieler, et al., Obstet Gynecol 1998;91:750-6), nor in abnormalities of childhood growth, hearing, or vision (Kieler, et al., Br J Obstet Gynaecol 1997;104:1276-72). Furthermore, when Kieler and colleagues (Epidemiology 2005;16:304-10) focused on men born in Sweden between 1973 and 1978 and broke this down into ultrasound exposed (N = 7998) vs unexposed (N = 197,829) individuals enrolling in military service between 1991 and 1996, they “failed to demonstrate a clear association between ultrasound scanning and intellectual performance.”
To my knowledge, the only other group that has systematically evaluated ultrasound exposure and childhood outcome is that of Newnham’s in Western Australia. In their first study (Newnham, et al., Lancet 1993;342:887-91), they randomized women with singleton pregnancies at 16-20 weeks to receive either a single scan (18 weeks) or 5 scans (18, 24, 28, 34, and 34 weeks) with the latter including continuous-wave Doppler flow studies (higher intensity and longer ultrasound exposure). Interestingly, they found a higher rate of intrauterine growth restriction at both < 10th (RR = 1.35; 95% CI 1.09-1.67; p = 0.006) and < 3rd (RR = 1.65; 95% CI 1.09-2.49; p = 0.020) centiles. However, a follow-up study of these infants (Newnham, et al., Lancet 2004;364:2038-44) demonstrated no differences in physical size by age 1 and thereafter, and more importantly, no differences in “standard tests of childhood speech, language, behaviour, and neurological development.”
In summary, based on the relative risks estimated in the Swedish and Norwegian studies, males exposed to ultrasound in utero in the 1970’s and early 1980’s have a 32% greater chance of being lefthanded than those who were not. To translate this into real numbers and put the results in perspective, these results suggest that about one in 50 ultrasound-exposed males who were expected to be born righthanded came out lefthanded. Is this simply the result of chance? Is it the result of differences in the demographic characteristics of the children who were more likely to be exposed to ultrasound? Is it the result of subtle brain damage? Is it the result of some form of ‘reprogramming’ of the genetic predisposition to lefthandedness? The bottomline is, we do not know! If brain damage is a concern, the follow-up studies have not clearly shown a more global effect on behaviour and neurodevelopment.
Should we brush these results off and resume a cavalier stance toward ultrasound imaging during pregnancy? Absolutely not! As the number and intensity of ultrasound-based procedures has increased across all socioeconomic groups, we should be especially vigilant for adverse outcomes. But, I must admit, I am grateful for our reader and patient who piqued my interest in this subject, because I have come away from this review feeling much more comfortable about what I do every day and about the admonition of primum non nocere!