Prenatal ultrasound examinations and risk of childhood leukaemia: case-control study
BMJ 2000; 320 doi: https://doi.org/10.1136/bmj.320.7230.282 (Published 29 January 2000) Cite this as: BMJ 2000;320:282All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
Editor,
Naumburg et al. (1) test the hypothesis of an association between
prenatal ultrasound examination and risk of childhood leukaemia in an
elegant case-control design. By using prospectively assembled data from
Swedish population-based registries rather that interviews with parents of
affected children they manage to reduce bias and increase statistical
power. Their main finding is that children prenatally exposed to
ultrasound have an odds ratio of 0.85 (95% confidence interval 0.62 to
1.17) of developing lymphatic leukaemia relative to unexposed children.
The authors conclude that exposure to ultrasound does not influence the
risk of subsequent development of leukaemia.
We find this interpretation plausible, considering that the evidence
of relevant biological effects of ultrasound is limited (albeit not
absent). Yet we argue against the sole use of point estimates of risk and
related confidence intervals when studying the association between an
exposure that by now has become virtually universal and an outcome that is
very serious. In such situations it will be appropriate to also assess the
"worst case" scenario at population level, i.e. to estimate the highest
population attributable risk (PAR%) that is consistent with the study
results. The PAR% gives the proportion of cases of a disease in a
population that is attributable to exposure to a particular risk factor
(2). In a case-control study it can be estimated as Pe(RR-1) / (1+ Pe(RR-
1)) where Pe is the proportion of controls exposed and RR the relative
risk estimate. Assuming that today at least one ultrasound examination
will be carried out in 95% of all pregnancies in Sweden and that there is
a 2.5% probability of an OR of 1.17 or higher the PAR% could be as high as
14% for myeloid leukaemia. If, in view of the apparent absence of a dose-
effect pattern, the odds ratios based on a regression modelling of linear
trend are used for this calculation, the PAR% could still be up to 5% for
myeloid and 26% for lymphatic leukaemia. This means that the data
presented by Naumburg et al. are consistent with a 2.5% probability that a
small but not negligible proportion of childhood leukaemia cases in Sweden
today might be attributable to prenatal ultrasound exposure.
A slight increase in risk would be tolerable if ultrasound
examinations were only performed when medically indicated. Hardly any
medical test is completely without risk; evidently, any potential risk
incurred has to be balanced against the potential benefit of the test (3).
Yet ultrasound is used liberally and often without a rational indication,
with increasing intensity levels (4) and despite evidence that its routine
use in antenatal screening does not necessarily improve outcomes (5).
From a public health perspective, the discussion should not be
confined to how safe ultrasound is. It should extend to whether an
ultrasound examination is indicated, and if indicated, how the benefits
balance against the risk. For specific obstetric conditions, the benefits
clearly outweigh the risk (3). However, this may not be the case when
ultrasound is used as "baby TV" as is common practice in Germany and other
countries.
Dr Oliver Razum, MSc
Department of Tropical Hygiene and Public Health
Im Neuenheimer Feld 324, 69120 Heidelberg, Germany
e-mail: oliver.razum@urz.uni-heidelberg.de
Dr Albrecht Jahn, MSc
Department of Tropical Hygiene and Public Health
Im Neuenheimer Feld 324, 69120 Heidelberg, Germany
e-mail: oliver.razum@urz.uni-heidelberg.de
1. Naumburg E, Bellocco R, Cnattingius S, Hall P, Ekbom A. Prenatal
ultrasound examinations and risk of childhood leukaemia: a case-control
study. Br Med J 2000;320:282-3.
2. A Dictionary of Epidemiology. New York, Oxford, Toronto: Oxford
University Press, 1995.
3. Enkin M, Keirse MJCN, Renfrew MJ, Neilson JP. A Guide to Effective
Care in Pregnancy & Childbirth. Oxford: Oxford University Press, 1995.
4. Grab D, Sterzik K, Terinde R. [Ultrasound studies in pregnancy:
technical principles and clinical safety]. Geburtshilfe Frauenheilkd.
1992;52:721-9.
5. Jahn A, Razum O, Berle P. Routine screening for intrauterine
growth retardation in Germany: low sensitivity and questionable benefit
for diagnosed cases. Acta Obstet Gynecol Scand 1998;77:643-8.
Competing interests: No competing interests
It is strange how little 'asides' occasionally turn the lights on.
One such aside is noted in Eleni Petridou's letter of Feb 4 2000 '(e.g.
childhood leukaemia is known to be more common among the first born and
the higher income groups ) "
I had occasion to study clustering patterns of this disease in the
North Shore area of Vancouver (population 100,000 ) back in 1974 (Lancet
1974 i : 210-211 ). 20 cases were admitted to the local hospital between
1958 and 1973. 2 of these were diagnosed elsewhere. Two features of
interest were noted. A high degree of mobility and interpersonal contact
with some adults who later developed leukaemia.
8 of 18 families had been at their presenting domicile for less than
a year. In a control group where juvenile diabetes was diagnosed over
this same period 7 of 59 families had moved house within a year of
diagnosis. i.e. The mobility of leukaemia families was four times that
of the diabetes families. While documentation of interpersonal contact
was confirmed by examination of hospital reccords in 3 situations, in 2
further cases this was anecdotal. The possibility of analogy between man
and cats regarding retroviral leukaemic aetiology was entertained.
In the summer of 1974 I spent a morning in Oxford discussing the
above with P.G Smith and Professor Doll. The mobility patterns were of
particular interest. My concerns had to do with the definition of
'closeness'. Subsequently Dr L J Kinlen has very elegantly demonstrated
the role of mobility in large populations as this may relate to the
aetiology of leukaemia.
This brings me to the question posed 'Why first born ?.' The
question of birth order didn't arise when I was studying the North
Vancouver group. In reviewing these cases again at least 11 of 18 fall
in the first born group to my knowledge. There may be more.
The significance of this may also have to do with 'mobility'. First
borns are picked up more often, cuddled more often and generally receive
more handling. They are after all 'experimental models' !. Kinlen's
group have suggested that mobility patterns are compatible with a
leukaemia aetiology caused by a horizontally transmitted agent of low
grade infectivity. Can this not also be reconciled with incidence in
first borns ?. Are first borns peculiarly immuno incompetent in this
regard ? Is there scope here for further epidemiologic modelling ?
James E Parker.
Competing interests: No competing interests
EDITOR - The recently published study by Naumburg and her colleagues1
concerning prenatal ultrasound examinations in relation to childhood
leukemia is excellent and the results reassuring. However, the design of
this study did not allow control for confounders other than maternal age,
birth weight and twin membership. Because during the earlier period of the
Swedish study, that has covered children born between 1973-89, ultrasound
was not generally applied, selection bias or subtle confounding by
exposures that have not been ascertained in the study or even recorded
(e.g. childhood leukemia is known to be more common among the first born
and the higher income groups), cannot be excluded.
The authors are
justified to be unconcerned by these possible sources of error, because
controlling for them would have further reduced the point estimates of the
odds ratio linking ultrasound examination to the risk of childhood
leukemia. Nevertheless, it is preferable to rely on data than even on the
most reasonable assumptions and a published study provides such data2.
Among 153 cases of childhood leukemia diagnosed throughout Greece during
1993-94, 100 have been exposed to ultrasound during their pregnancy,
whereas among 300 age, gender, and residence- matched controls 195 have
been exposed, for an odds ratio of 1.02 (95% confidence intervals 0.68 to
1.53). Controlling for twenty-four different socio-demographic, lifestyle,
environmental and biomedical variables reduced this odds ratio to 0.74
(95% confidence intervals 0.41 to 1.32)2.
Our results are, thus,
compatible to those by Naumburg in indicating the safety of pregnancy
ultrasound but they are also compatible with the possibility that there
may be some moderate bias and confounding in the estimation of the odds
ratio of childhood leukemia in relation to pregnancy ultrasound when the
procedure is not universally applied. Indeed, selection and confounding
could be responsible for the non-significant excess risk noted by Naumburg
et al., when ultrasound during the second semester was evaluated1.
Eleni Petridou,
Associate Professor of Epidemiology
Dept of Hygiene and Epidemiology, Athens University Medical School,
75 M. Asias Str., Goudi, Post Code 115 27, Athens, Greece and
Dept. of Epidemiology, Harvard School of Public Health
677 Huntington Avenue, Boston MA 02115, USA
Dimitrios Trichopoulos,
Professor of Epidemiology
Dept of Hygiene and Epidemiology, Athens University Medical School
75 M. Asias Str., Goudi, Post Code 115 27, Athens, Greece and
Dept. of Epidemiology, Harvard School of Public Health
677 Huntington Avenue, Boston MA 02115, USA
1. Naumburg E, Belocco R, Cnattingius S, Hall P, Ekbom A.Renatal
ultrasound examinations and risk of childhood leukemia: a case-control
study. Br Med J 2000;320:282-283.
2. Petridou E, Trichopoulos D, Kalapothaki V, Pourtsidis A, Kogevinas
M, Kalmanti M, et al. The risk profile of childhood leukemia in Greece: a
nationwide case-control study. Br J Cancer 1997;76:1241-1247
Competing interests: No competing interests
Re: Risk of prenatal ultrasound examination: a public health perspective
Dr. Razum makes some excellent points about ultrasound being used for
non-medically required cercumstances. It is also used widely as "Baby TV"
in the United States.
However, by far the greatest exposure to ultrasound comes during
labor, when external fetal monitors are used almost universally, at least
in the US. These monitors are often in continuous operation for many
hours. This amount of exposure dwarfs anything that occurs during most
pregnancies.
This is a practice that may not have been in use during much of the
period of the sudy by Naumberg et al, although it is not mentioned in
their study. However, if this level of ultrasound exposure did not occur
in their subjects, I think that the results of their study may not be
relevant to current practices.
Competing interests: No competing interests