SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. References

Objective To test a hypothesis of no association between ultrasound exposure in early fetal life and growth or impaired vision or hearing during childhood.

Design Follow up of eight to nine year old children born to women who participated in a randomised controlled trial on ultrasound screening during pregnancy.

Setting Nineteen antenatal care clinics run by three central hospitals in Sweden from 1985 to 1987.

Population and methods Of 4637 eligible singleton pregnancies, 3265 (71%) were followed up through a questionnaire sent to their mothers. Analyses were performed both according to randomised groups and to ultrasound exposure.

Main outcome measures Parents’ report of vision and hearing tests as recorded on child's record card. Parents’ report of their child's weight and height at 1, 4 and 7 years of age.

Results Reduced hearing was reported by 3.4% in the screening group compared with 3.5% in the nonscreening group (odds ratio [OR] 1.0; 95% confidence interval [CI] 0.67–1.41). The same prevalences were found when analysed according to ultrasound exposure (OR 1.0; 95% CI 0.67–1.42). Reduced vision was reported by 6.3% in the screening group compared with 7.8% in the nonscreening group (OR 0.8; 95% CI 0.60–1.03). Corresponding figures for ultrasound exposed and unexposed were 6.2% and 8.0%, respectively (OR 0.8; 95% CI 0.58–1.00). No statistically significant differences in body weight or height at 1, 4 or 7 years of age between screened and not screened children or between exposed and unexposed were found.

Conclusion This study found no association between ultrasound exposure in early fetal life and growth or impaired vision or hearing during childhood.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. References

The clinical value of routinely performed obstetric ultrasound is still being debated1,2. The safety of ultrasound should be an important part of this discussion especially as some studies3,4 have presented data that implicate long term harmful effects on children exposed in fetal life.

Ultrasound may damage human tissue by a rise in temperature or cavitation5. Local cell death or cell membrane damage, which in turn affects cell differentiation6, might be the result of ultrasound exposure. Mole5 implied that the relay system between the retina and visual cortex and the cochlea of the inner ear are obvious sites for irreparable damage because of one-to-one cell connections. Even the eye lens is a sensitive organ because of the lack of continuous blood circulation, so a rise in temperature caused by ultrasound can not be counteracted5. Consequently, minor visual or hearing losses could be possible results of ultrasound exposure in fetal life.

There are still controversies on how to interpret the association between ultrasound exposure in pregnancy and altered fetal weight found in both animal and human studies, as both decrease as well as increase in weight has been reported3,4,7. Altered growth during childhood might be expected if birthweight is affected by ultrasound.

Only two studies8,9 have focused on antenatal ultrasound exposure and subsequent vision or hearing and two10,11 on growth during childhood according to two reviews of epidemiological studies on ultrasound exposure3,4. None of these reports8–11 have found any associations between ultrasound and the studied outcomes. However, more studies in this field are needed as the two cohort studies9,10 had limitations because of small sample size10, imperfect matching9 and low response rate9. In the randomised follow up study by Salvesen et al.8,11 the scans were performed in gestational week 19 and 32. Nowadays there is a tendency to perform routine ultrasound earlier in pregnancy and to omit scans in the third trimester. No study has so far investigated the association between growth, vision and hearing during childhood and ultrasound scanning at about 15 weeks.

The aim of this study was to test a hypothesis of no difference in growth, vision and hearing during childhood among children in a follow up of a single stage randomised trial on ultrasound screening during gestational week 1512. Boys were studied separately as male fetuses are more vulnerable than female fetuses13.

METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. References

Participants in the original trial

Between October 1985 and March 1987 all women who booked for antenatal care in three Swedish hospitals were asked to participate in a randomised trial on ultrasound screening during pregnancy12.

Of a total of 8768 women, 1414 were excluded either because they booked too late to have an ultrasound scan before 19 gestational weeks, had already had an ultrasound scan, or intended to move to another region or to another antenatal clinic. Another 2357 women were excluded because they were assigned to ultrasound scanning due to a medical indication12.

The remaining 4997 women, who all had regular menstrual cycles, were randomised into two groups by the opaque, sealed envelope technique. The screening group of 2482 women was offered an ultrasound scan at about 15 weeks of gestation. The nonscreening group of 2511 women was not offered a routine scan but the standard antenatal care provided an opportunity for a later scan on clinical indication. There were four women lost to follow up. In the screening group, 32 women (1.3%) did not attend the ultrasound screening. In the non-screening group, 103 women (4.1%) were subjected to ultrasound scanning before 19 weeks of gestation and 786 women (31%) had a scan later in pregnancy. There were 28% smokers in the screening and 29% in the non-screening group at the first antenatal visit.

Apart from the ultrasound screening procedure, the groups received the same standard antenatal care. Linear real-time machines (Hitachi EUB 22, EUB 25, EUB 26, EUB 400 and Siemens Imager 2380) calibrated to sound velocity 1540 m/s were used. The intensity output levels were 11–12 mW/cm2 for the EUB machines and 35 mW/cm2 for the Siemens Imager 2380 given as spatial peak temporal average. The screening examination included a check for multiple pregnancies and fetal viability, measurement of biparietal diameter and a general examination of the fetus. The exposure times were not recorded individually, but the booking interval was 15 minutes.

Subjects in the follow up trial

Among the screened women, 87 had a miscarriage and six a legal abortion. There were 24 pairs of twins and one triplet pregnancy. As 52 children could not be followed up either because they or their mothers had died, emigrated, had an unlisted address or could not be traced, 2312 singletons were included in the follow up study. In the nonscreening group there were 96 women who had a miscarriage, three who had a legal abortion, and 20 pairs of twins. As 67 children could not be followed up for the above reasons, 2325 singletons remained to be included.

In January 1995, the mothers of all 4637 children were sent a questionnaire together with an information letter and a postage-paid return envelope. Return of the questionnaire was taken as informed consent for the child's participation. Two reminders were sent to non-responders.

The questionnaire consisted of 52 questions, of which two concerned the child's hearing and three his or her vision (Table 1). The parents were asked to report their child's weight and height at 1, 4 and 7 years of age together with the child's age (in years and months) at the time of measurement. They were requested to use the data from the child's health record, which is kept by the parents.

Table 1.  Response rates on vision, hearing and growth among 4634 eight and nine year old children whose mothers participated in a trial on ultrasound screening during pregnancy. Values are given as n (%).
Items in questionnaireResponse
Reduced hearing3238 (71.0)
Referred to otorhinolaryngolist3224 (70.7)
Reduced vision3240 (71.1)
Use of spectacles3251 (71.3)
Referred to ophtalmologist3241 (71.1)
Body weight 
 At 1 year2806 (61.5)
 At 4 years2616 (57.4)
 At 7 years2062 (45.2)
Body height 
 At 1 year2764 (60.6)
 At 4 years2608 (57.2)
 At 7 years2019 (44.3)

Swedish children are regularly examined from birth up to six years of age at the child health care centres by physicians and specially trained nurses. From seven years of age the examinations are performed at school by school physicians and nurses. Vision and hearing are first tested at the age of four years, and at the age of eight years all children have been tested at least three times. The results of the vision and hearing tests are recorded in the health record. Children with a risk factor for reduced hearing or vision are usually referred directly and examined by a specialist. Weight and height are registered in the child's health record at all visits at the child health care centres.

Distant visual acuity is generally tested with the HVOT matching test (a chart containing the letters H, V, O and T). Each eye is tested separately at a distance of 3 m. The child has a copy of the chart in his or her hand and is asked to indicate the letter pointed at by the examiner. Children with a visual acuity of less than 0.8 of one eye, a difference between the eyes of two rows or more or with signs of squinting are referred to an ophthalmologist. Visual acuity of less than 0.8 is regarded as impaired vision14. Hearing tests are carried out with the use of pure tone audiometry15. Both ears are tested with frequencies in the range 0.250–8 kHz at 20 dB. A hearing reduction of more than 20 dB at three or more frequencies or 30 dB at one frequency is regarded as impaired hearing15.

All comparisons were made between children in the screening and nonscreening group and between those exposed to ultrasound before 19 weeks of gestation and those unexposed. Children who were exposed to ultrasound later in fetal life were not considered as ultrasound-exposed. Separate analyses were made for boys.

Power calculations and statistical analyses

Power calculations were performed before the study to estimate the magnitude of shift in prevalence of impaired vision and hearing that we would be able to detect with the given sample size of 2000 children in each group. The prevalence of impaired vision and hearing in a normal population of children was presumed to be 4%. With a two-sided α of 0.05 and β of 0.10 a 57% increase or decrease in the prevalence of impaired vision and hearing could be detected. Statistical analyses were performed with the SAS program16.

Possible confounders such as maternal age, parity, education, smoking, gender, gestational age and perinatal asphyxia (Apgar score < 7 at 5 min) were controlled for in logistic regression analyses. Gestational age was calculated from the last menstrual period in order to use similar estimates for screened and not screened women.

Comparisons of impaired vision and hearing between the groups were analysed and expressed as odds ratios with 95% confidence intervals. Mean values of body weight and height at 1, 4 and 7 years of age were calculated, but as weight and height were not recorded exactly at these ages for all children the estimated values were calculated by use of a multiple linear regression model. The possible confounders were included as variables.

The Ethical Committee of the Medical Faculty at Uppsala University gave consent to both the original and the follow up study.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. References

In the screening group 661 mothers did not respond or declined participation, leaving 1651 (71.5%) of 2312 screened children to be included in the follow up study. In the nonscreening group 711 did not respond or declined participation, leaving 1614 (69.4%) of 2325 children from the nonscreening group to be included. In early fetal life (before 19 weeks of gestation), 1704 children were exposed to ultrasound whereas 1561 were not.

The responders were older (27.7 years) and less often smokers (26%) compared with the nonresponders (27.0 years and 33% smokers; P < 0.01). There were no differences between responders and nonresponders in terms of parity, gender, gestational age and perinatal asphyxia. Family and social variables of the children followed up are presented in Table 2.

Table 2.  Family and social variables for 8 and 9 year old children whose mothers participated in a randomised trial on ultrasound screening during pregnancy. Variables are shown both according to randomised groups and to ultrasound exposure. Values are given as n or n (%).
 Screening group (n = 1651)Nonscreening group (n = 1614)Ultrasound-exposed (n = 1704)Unexposed (n = 1561)
Years of education (mother)    
 No. who responded1617158616701533
 <10 years178 (11.0)194 (12.2)182 (10.9)190 (12.4)
 10–12 years849 (52.5)764 (48.2)867 (51.9)746 (48.7)
 College or university590 (36.5)628 (39.6)621 (37.2)597 (38.9)
Years of education (father)    
 No. who responded1583154316381488
 <10 years286 (18.1)270 (17.5)293 (17.9)263 (17.7)
 10–12 years785 (49.6)747 (48.4)810 (49.5)722 (48.5)
 College or university512 (32.3)526 (34.1)535 (32.7)503 (33.8)
Family economy    
 No. who responded1609157816631524
 Very good184 (11.4)174 (11.0)199 (12.0)159 (10.4)
 Good707 (43.9)689 (43.7)729 (43.8)667 (43.8)
 Medium657 (40.8)644 (40.8)671 (40.4)630 (41.3)
 Poor61 (3.8)71 (4.5)64 (3.9)68 (4.5)
Lived with both parents during childhood    
 No. who responded1610157616631523
 Lived with both parents during childhood1362 (84.6)1317 (83.6)1413 (85.0)1266 (83.1)

The overall response rates for questions on vision, hearing, weight and height at ages 1, 4 and 7 years are shown in Table 1. There were no significant differences in response rates for the different items between th screening and nonscreening group nor between ultrasound-exposed and unexposed children.

The number of children with impaired vision and hearing together with odds ratios and 95% confidence intervals for comparisons between groups is presented in Table 3. The number of referrals for specialist examination is also shown in Table 3. When controlling for possible confounding factors such as maternal age, parity, education, smoking, gender, gestational age and perinatal asphyxia (Apgar score < 7 at 5 mins) the odds ratios remained practically unchanged.

Table 3.  Prevalence of items regarding hearing and vision for 8 and 9 year old children whose mothers participated in a randomised trial on ultrasound screening during pregnancy. Results are shown both according to randomised groups and to ultrasound exposure. Values are given as n (%); comparisons are expressed as odds ratios (OR) with 95% confidence intervals (95% CI).
 Screening group (n = 1651)Nonscreening group (n = 1614)OR[95%CI]Ultrasound exposed (n = 1704)Unexposed (n = 1561)OR[95%CI]
Reduced hearing56 (3.4)56 (3.5)0.97 [0.67–1.41]58 (3.4)54 (3.5)0.98 [0.67–1.42]
Referred to otorhinolaryngolist152 (9.3)172 (10.8)0.84 [0.67–1.06]159 (9.4)165 (10.8)0.86 [0.68–1.08]
Reduced vision103 (6.3)126 (7.8)0.79 [0.60–1.03]105 (6.2)124 (8.0)0.76 [0.58–1.00]
Use of spectacles141 (8.6)149 (9.3)0.92 [0.72–1.17]149 (8.8)141 (9.1)0.97 [0.76–1.23]
Referred to ophthalmologist189 (11.5)219 (13.7)0.82 [0.67–1.01]200 (11.8)208 (13.4)0.86 [0.70–1.06]

Mean body weight and height from the health controls at 1, 4 and 7 years of age for all included children are shown in Table 4. The estimated differences in body weight and height at the exact ages of 1, 4 and 7 years, and shown in Table 4, were calculated using multiple regression analyses. No significant differences were found between screened and not screened or between exposed and unexposed children, respectively.

Table 4.  Mean body weight and height from the health controls at 1, 4 and 7 years of age together with estimated differences in weight and height between screened and not screened children and between ultrasound exposed and unexposed, respectively. Negative values indicate higher values among not screened and unexposed compared with screened and exposed, respectively. No differences were significant. Values are given as mean and mean difference* [95% CI]. Scr = screened; Exp = exposed.
 Body weight (kg)Body height (cm)
Age (years)Body weightScr – not ScrExp – unExpBody weightScr – not ScrExp – unExp
  1. *Controlling for exact age and sex

1 year9.90.03 [−0.06 to 0.12]0.06 [−0.03 to 0.15]76.050 [−0.24 to 0.23]-0.01 [−0.25 to 0.23]
4 years17.55-0.03 [−0.21 to 0.16]0.01 [−0.18 to 0.19]104.95-0.01 [−0.36 to 0.35]-0.03 [−0.39 to 0.32]
7 years25.06-0.04 [−0.41 to 0.32]-0.01 [−0.38 to 0.35]124.65-0.05 [−0.57 to 0.47]-0.04 [−0.55 to 0.48]

The estimated difference in body weight at one year of age between exposed and unexposed children to smoking mothers was 129 g in favour of the exposed children. At four and seven years of age the unexposed children had the highest mean body weight. The mean difference at four years was 156 g and at seven years 547 g. None of these differences was statistically significant. Separate comparisons for boys made on vision, hearing, mean body weight and height showed only minor and nonsignificant differences (results not shown).

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. References

No association between exposure to ultrasound in early fetal life and impaired vision or hearing among eight or nine year old children could be found in this study. Neither were there any statistically significant differences in body weight or height at 1, 4 or 7 years of age between children from the screening and nonscreening group nor between exposed or unexposed. The results are in agreement with previous studies8–11.

A major limitation of validity for a study like this in which mailed questionnaires were used is nonresponse bias. The overall response rate here was 71%, which is in accordance with studies using this approach17,18. No difference in response rate between the screening and nonscreening group nor between exposed and unexposed children was observed. The responders were older and less often smokers compared with the nonresponders (P < 0.01). There were, however no differences between responders and nonresponders in terms of parity, gender, gestational age and perinatal asphyxia. The response rate for questions on hearing and vision corresponded well to the overall response rate, but the response rate for questions on weight and height was somewhat lower. The two lowest rates were recorded for questions on body weight and height at seven years of age.

Although the original study was a randomised trial, we chose to analyse also according to ultrasound exposure as it was possible adverse effects of ultrasound that were of interest19.

Of the women in the nonscreening group, 4.1% had an ultrasound scan before 19 weeks of gestation and 1.3% of the women in the screening group had no scan. Children who were exposed to ultrasound later in fetal life were not considered as ultrasound-exposed as the aim was to investigate possible adverse effects of routine ultrasound scanning in the second trimester. Analyses performed according to exposure did not change the results and the odds ratios remained practically unchanged after controlling for possible confounders.

The 75 g higher birthweight (P= 0.01) of screened compared with not screened children to mothers who smoked found in the original study12 could be observed as a nonsignificant difference only up to one year of age, which is in agreement with the findings by Salvesen et al.11. If ultrasound exposure caused reduction in smoking during pregnancy12, the effect on growth only lasted for a short period in childhood. Nonetheless, growth in childhood is affected by many factors, of which smoke exposure in fetal life plays only a minor part20. This study could not support the idea of ultrasound initiating improved growth in childhood as suggested by Salvesen et al.11, but it is important to note that nonresponse bias might have been introduced because smoking women had a lower response rate.

In this study, no direct testing of the children was performed as vision and hearing among the participating children were evaluated by parent assessments. The parents’ report should be based on the results from tests used in the Swedish child and school health care systems14,15. The Swedish screening program for vision and hearing impairment at child and school health care centres has a very high standard, with an attendance and detection rate above 95%14. Although the possibilities for detecting disabling vision or hearing by these screening procedures are very good14, minor visual or hearing losses such as those hypothetically caused by ultrasound exposure in fetal life might not be revealed. Another important factor to consider is that power calculations showed that an increase in the prevalence of impaired vision or hearing of less than 57% could not be detected even with a sample size of 2000 in each group, and our final sample sizes contained only around 1600 children in each group.

Impaired vision was reported by 6% to 8%, which is more than expected but in agreement with the figures reported by the parents in the study by Salvesen et al.8. As about 9% reported use of spectacles which are used not only to treat impaired vision but also refractive errors and squinting, it is possible that many parents included these conditions when answering the question about impaired vision. We do not believe that this misclassification of impaired vision by the parents affects the results. A lower prevalence of squinting or refractive errors among the ultrasound exposed children compared with the unexposed might, in theory, conceal a potentially higher prevalence of impaired vision, but such an effect of ultrasound is unlikely.

The prevalence of impaired hearing found in this study for ultrasound exposed (3.4%) and unexposed (3.5%) is in agreement with the expected prevalence of 4%15 and with the findings by Salvesen8. Persistent middle ear effusion is the most common cause of reduced hearing in children15,21 but there is no reason to suspect an uneven distribution of middle ear effusion between the groups as there is no theoretical correlation between this condition and ultrasound exposure in fetal life.

As male fetuses probably are more vulnerable than female fetuses13, and both impaired vision and hearing are more common in males14,22, boys were analysed separately, but no differences were found.

CONCLUSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. References

No associations between antenatal ultrasound exposure before 19 gestational weeks and disabling hearing or vision impairments or growth in children up to seven years of age were found in this study.

Acknowledgements

This study was supported by grants from the Research Council of Dalarna; the Foundation of Astrid Karlsson, Uppsala University; and the Foundation of Medical Research and Evaluation in Dalarna in co-operation with Uppsala University.

References

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. References
  • 1
    Reed KL. Editorial. Why (not) do obstetric ultrasound? An observation on uncertainty. Ultrasound Obstet Gynecol 1996; 8: 12.
  • 2
    Keirse MJNC. Frequent prenatal ultrasound: time to think again [editorial]. Lancet 1993; 342: 878879.
  • 3
    Salvesen KÅ, Eik-Nes SH. Is ultrasound unsound? A review of epidemiological studies of human exposure to ultrasound. Ultrasound Obstet Gynecol 1995; 6: 293298.
  • 4
    European Committee for Ultrasound Radiation Safety—Tutorial paper. Epidemiology of diagnostic ultrasound exposure during human pregnancy. EurJ Ultrasound. 1996; 4: 6973.
  • 5
    Mole R. Possible hazards of imaging and Doppler ultrasound in obstetrics. Birth 1986; 13: 2937.
  • 6
    Dinno MA, Dyson M, Young SR, Mortimer AJ, Hart J, Crum LA. The significance of membrane changes in the safe and effective use of therapeutic and diagnostic ultrasound. Phys Med Biol 1989; 34: 15431552.
  • 7
    O'Brien WD, Safety of ultrasound with selective emphasis for obstetrics. Semin Ultrasound CTMR 1984; 5: 105120.
  • 8
    Salvesen KÅ, Vatten LJ, Jacobsen G, Eik-Nes SH, Økland O, Molne K, Bakketeig LS. Routine ultrasonography in utero and subsequent vision and hearing at primary school age. Ultrasound Obstet Gynecol 1992; 2: 243247.
  • 9
    Stark CR, Orleans M, Haverkamp AD, Murphy J. Short and long term risks after exposure to diagnostic ultrasound in utero. Obstet Gynecol 1984; 63:194200.
  • 10
    Lyons EA, Dyke C, Toms M, Cheang M. In utero exposure to diagnostic ultrasound: a 6-year follow-up. Radiology 1988; 166: 687690.
  • 11
    Salvesen KÅ, Jacobsen G, Vatten LJ, Eik-Nes SH, Bakketeig LS. Routine ultrasonography in utero and subsequent growth during childhood. Ultrasound Obstet Gynecol 1993; 3: 610.
  • 12
    Waldenström U, Axelsson O, Nilsson S, Eklund G, Fall O, Lindeberg S, Sjödin Y. Effects of routine one-stage ultrasound screening in pregnancy: a randomised controlled trial. Lancet 1988; 2: 585588.
  • 13
    Geschwind N, Galaburda AM. Cerebral lateralization. Biological mechanisms, associations, and pathology: I. A hypothesis and a program for research. Arch Neural 1985; 42: 42859.
  • 14
    Lennerstrand G, Jakobsson P, Kvarnström G. Screening for ocular dysfunction in children: Approaching a common program. Acta Ophthalmol Scand 1995; 73 Suppl 214: 2638.
  • 15
    Lidén G. Audiologi [in Swedish]. Stockholm : Almquist & Wiksell, 1985.
  • 16
    SAS Institute Inc. SAS User's Guide, Version 6. Cary, North Carolina, USA : SAS Institute Inc, 1989.
  • 17
    Siemiatycki J and Campbell S. Nonresponse bias and early versus all responders in mail and telephone surveys. Am J Epidemiol 1984; 120: 291301.
  • 18
    Barton J, Bain C, Hennekens CH, Rosner B Belanger C, Roth A, Speizer FE. Characteristics of respondents and non-respondents to a mailed questionnaire. Am J Public Health 1980; 70: 823825.
  • 19
    Pocock SJ. Clinical trials: A Practical Approach. Chichester : John Wiley and Sons Ltd, 1983.
  • 20
    Johnston C. Cigarette smoking and the outcome of human pregnancies: a status report on the consequences. Clin Toxicol 1981; 18: 189209.
  • 21
    Augustsson I, Nilsson C, Engstrand I. The preventive value of audiometric screening of preschool and young school-children. Int J Pediatr Otorhinolaryngol 1990; 20: 5162.
  • 22
    Axelsson A, Aniansson G, Costa O. Hearing loss in school children. Scand Audiol 1987; 16: 137143.