Stålberg, Karin*; Haglund, Bengt*†; Axelsson, Ove*; Cnattingius, Sven‡; Hultman, Christina M.‡§; Kieler, Helle*‡
Prenatal ultrasound seems to affect the risk of atypical lateralization, with an increase in left-hand or mixed-hand preference in men.1–3 Almost all fetuses in developed countries are exposed to ultrasound, often on a routine basis, and therefore it is important to investigate whether prenatal ultrasound might have potential effects on the developing brain.
The reported association between ultrasound and handedness has been hypothesized to be caused by an influence of ultrasound on neuronal migration.4 From the 6th to the 32nd week of gestation, neurons migrate from the neural tube toward their final destinations in the cerebral cortex.5 It has been shown previously that neuronal migration can be disrupted by environmental insults such as heat and radiation, and ultrasound has the potential to damage tissue by heating, cavitation, and streaming.6–8 In mice exposed to prenatal ultrasound in clinically relevant doses, a substantial number of neurons failed to acquire their proper positions in the cerebral cortex.9
Abnormal development of the brain during fetal life is thought to contribute to the etiology of schizophrenia.10 Abnormalities of neuronal proliferation and migration have been hypothesized to be related to critical specific insults during the fetal period11,12 or to genes that regulate neuronal migration.13,14 Moreover, it has been suggested that patients with schizophrenia have less cerebral asymmetry than healthy individuals,7,15,16 and patients with schizophrenia have also been found to have a higher prevalence of left or mixed hand preference.15 If prenatal ultrasound disrupts neuronal migration, this effect may result in an increased incidence of schizophrenia among exposed individuals.
Prenatal factors previously shown to be associated with the risk of schizophrenia are pregnancy complications (diabetes, bleedings, multiparity), disturbed fetal growth (low birth weight and small head circumference for gestational age), and delivery complications (cesarean section, asphyxia).17,18 The aim of the present study was to investigate a possible association between prenatal ultrasound exposure and risk of schizophrenia. Although the previous associations with prenatal risk factors appear to be stronger for schizophrenia than for other psychoses, an influence on all related diagnoses can not be excluded.18 Accordingly, we assessed possible effects of prenatal ultrasound exposure on both schizophrenia and other psychoses.
The Swedish National Board of Health and Welfare gave access to medical information from the Medical Birth Register, the Hospital Discharge Register and the Cause of Death Register. Information about place of residence during follow-up time was gathered from the Register of Total Population, held by Statistics Sweden. Individual record linkage between the registers was possible through the unique personal identification number assigned to each Swedish resident.
The Swedish Medical Birth Register
The Swedish Medical Birth Register contains prospectively collected information on more than 99% of all births in Sweden since 1973.19 The register includes data on maternal demographics, reproductive history and complications during pregnancy, delivery and the neonatal period. The information is provided through antenatal, obstetrical and neonatal records, which are filled in by midwives and physicians.
The Hospital Discharge Register
The Swedish Hospital Discharge Register includes data on dates of each hospital admission, discharge, main discharge diagnosis and secondary diagnoses if any. The diagnoses are classified and recorded by the treating physician according to the International Classification of Diseases (ICD) at the time of discharge from hospital. The agreement of Swedish register diagnoses versus review of medical records is very high for psychotic disorders.20–22 Recorded diagnoses are forwarded by computer medium to the Hospital Discharge Register held by the National Board of Health and Welfare. The routines of recording and forwarding diagnoses are standardized across Sweden. The register provides nationwide coverage since 1987 for psychiatric diagnoses of inpatient care, and includes care in psychiatric as well as medical clinics.
The Cause of Death Register
The Cause of Death Register contains data on dates and causes of death for Swedish citizens since 1961. Coverage is more than 99.5% and data are updated yearly.
Population Databases, Statistics Sweden
Information about emigration dates and place of residence during follow-up were retrieved from population databases of the total population in Sweden. Information about all births from 1969 through 1972 was collected from Statistics Sweden's Birth Register.
We included singletons born at hospitals with reliable information on ultrasound scanning. Of 97 hospitals, 49 had reliable information on their ultrasound program regarding the current time period (1973 to 1978).1,23 To increase the homogeneity of the study population, we included only children of mothers who themselves had been born in one of the Nordic countries (Sweden, Denmark, Norway, Finland, and Iceland). All included children were alive and living in Sweden at the age of twelve (data from the Cause of Death Register). Of 593,917 singleton live births in Sweden between 1973 and 1978, we could include 370,945 individuals (62.5%) with reliable exposure information, of whom 190,405 were men and 180,540 were women.
Diagnoses of schizophrenia and other psychoses were obtained from the Hospital Discharge Register, and all individuals eligible for inclusion were followed up after reaching 12 years of age. The follow-up period was 1987 to 2004. A schizophrenia diagnosis included the following ICD codes: ICD-9 codes 295A-295E, 295G, 295W, 295X and ICD-10 code F20. Other psychotic disorders included ICD-9 codes 295F, 296–298 and ICD-10 codes F21-31.
The University Hospital in Malmö was the first hospital in Sweden to introduce ultrasound scanning as part of standard antenatal care. Since October 1972, 90% of all pregnant women living in the city of Malmö have been scanned. The use of ultrasound from this period is well documented. Consequently, children born in Malmö to mothers registered as citizens of Malmö were considered exposed to ultrasound. During the introduction period, 1973–1975, the examinations were performed at approximately 28 weeks of gestation (83% had their routine scan before gestational week 30). Around 50% had additional scans, mostly to confirm measurements (because of lack of experience during this introduction period). The risk pregnancies having additional scans were mainly performed to assess intrauterine growth on request from clinicians. These repeated scans were generally performed with very short intervals.
From October 1976 through December 1978, 2 examinations were performed (full-scale period). For most mothers (95%) the first scan was in gestational week 18 to 20 and the second in week 32. There were no additional scans performed between these examinations, except for patients with bleeding from a suspected placenta previa. Of the more than 5000 women scanned during this period, there were 17 with a clinical diagnosis of placenta previa. The ultrasound machines used were the Kretz-Technic 4100 MGS and Combison II echoscope, Philips Diagnost B, and ADR-Kranzbüler. Each examination was scheduled for 15 minutes.24
Before 1980, 48 hospitals did not practice ultrasound scanning. We considered children born at any of these hospitals to be unexposed to ultrasound unless their mothers were registered as residents of Malmö, in which case they were excluded from the study.
We used Poisson regression analysis to estimate the effect of ultrasound exposure on the incidence of schizophrenia and other psychotic disorders. The results are presented as incidence rate ratios (IRRs) with 95% confidence intervals (CIs). The time at risk was calculated from the 12th birthday, but no earlier than 1 January 1987, to the first date of event, death, emigration or end of follow-up (31 December 2004). The incidence was calculated as cases per 100,000 person years. The following variables from the Medical Birth Register were included as potential confounders: maternal age, parity, gestational age, intrauterine growth (calculated as the ratio between birth weight and expected birth weight for gestational age) and Apgar score. In addition, by retrieving information from the Hospital Discharge Register, we were able to adjust for the mother's number of hospitalizations at a psychiatric clinic. We also controlled for subjects’ attained age during follow-up. Preliminary analyses indicated that gestational age, intrauterine growth, and Apgar score had practically no impact on the estimates of interest. Hence, to avoid problems with missing values, we excluded these covariates from analyses.
To evaluate whether unmeasured factors related to hospital level might affect the results, we performed analyses in which hospitals of tertiary level were compared with all other hospitals. Apart from Malmö there were 7 hospitals of tertiary level in Sweden during the study period. We adjusted for county of residence during the follow-up period to control for geographical variation in diagnostic and admittance routines.
Because other factors related to place of birth might affect the risk of being diagnosed with schizophrenia, we compared the incidence of schizophrenia according to place of birth before and after the introduction of ultrasound scanning in Malmö. It was not possible to obtain information on hospital of birth before establishment of the Swedish Medical Birth Register in 1973. However, for individuals born in the cities of Malmö, Umeå, Uppsala, and Örebro between 1967 and 1978, hospital of birth could be inferred from the fact that each of these 4 cities only had one hospital. Consequently, we compared incidence of schizophrenia for individuals born in each of these 4 cities with those born in other parts of Sweden. For these analyses, we included 362,315 singletons born 1967 to 1972 (before the start of ultrasound scanning) and 315,989 singletons born 1973 to 1978 (after introduction of ultrasound scanning). Only mothers living in communities that fulfilled the criteria for the main study were included.
To assess the impact of number and frequency of ultrasound examinations on the risk for schizophrenia, we divided the exposed cohort into 2 subcohorts born 1973 to 1975 and 1976 to 1978. These 2 time periods coincide approximately with the introduction phase and full-scale phase of the ultrasound scanning program in Malmö.
The male fetal brain is considered more vulnerable than the female brain16 and higher risks of mixed-handedness/left-handedness in connection with prenatal ultrasound have been found only in men.1,3,25 Hence, we included sex as an effect modification term in all analyses and we present the results according to sex. Statistical analyses were conducted with the SAS 9.1 software package (SAS Institute, Cary, NC). Ethics Committees at Uppsala University and at Karolinska Institutet approved the study.
In Table 1, maternal and infant characteristics and associated incidence rates for schizophrenia and other psychotic disorders are presented stratified by ultrasound exposure (ie, exposed vs. nonexposed). We found a higher incidence of schizophrenia among individuals born at Malmö University Hospital (exposed to ultrasound) when compared with individuals born at all other hospitals (unexposed). The difference appears to be more pronounced among men (17.2 versus 10.9/100,000 person-years) than women (7.5 versus 6.0/100,000 person-years). In almost all categories of variables included in Table 1, the incidence of schizophrenia was higher in the exposed compared with the unexposed cohort. The same pattern could not be seen for other psychoses. Among exposed as well as nonexposed children, male sex, high maternal age, preterm births, and mother's psychiatric care were associated with high incidence rates of schizophrenia and other psychotic disorders. The estimated crude IRR for schizophrenia when exposed to ultrasound was 1.58 (95% CI = 0.99–2.51) for men and 1.26 (0.62–2.55) for women. The same estimates for other psychotic disorders were 1.12 (0.80–1.58) for men and 0.92 (0.62–1.37) for women.
We continued with detailed analyses of ultrasound exposure and risks of schizophrenia and other psychoses. However, because there was no association between ultrasound exposure and other psychoses, we limit the presentation to ultrasound exposure and risk of schizophrenia. Because we found practically no differences in risks for schizophrenia between the subcohorts born during the introduction and the full-scale period, we present data for the whole cohort.
To control for possible confounders we performed multivariate analyses adjusting for maternal age, parity and maternal psychiatric care and subjects attained age during follow-up (Table 2). In Model I, individuals born at Malmö University Hospital were compared with those born at all other included hospitals. Among men, we found higher risks of schizophrenia if born in Malmö. For women, the point estimate was slightly increased but the confidence interval did not support an association. In Model II, individuals born at Malmö University Hospital and the 7 other hospitals of tertiary level were compared with those born at primary and secondary hospitals. Risks of schizophrenia remained essentially unchanged for both men and women born in Malmö. The highest risks of schizophrenia were found among men born at the university hospitals in Malmö and Uppsala. For women, the highest risk of schizophrenia was found for those born in Uppsala. Similarly, when adjusted for county of residence during follow-up (Model III), the highest risk of schizophrenia was found among men born in Malmö (IRR = 1.60; CI = 0.90–2.83) and women born in Uppsala (1.66; 0.89–3.10).
Finally, we compared risk of schizophrenia for singletons born in 4 cities between 1967 and 1972, before the introduction of ultrasound scanning, with risk of schizophrenia for those born between 1973 and 1978. Men born in Malmö had the highest ratio of IRRs between cohorts born before and after the introduction of ultrasound scanning (Table 3).
Individuals born at the university hospital in Malmö and assumed to have been exposed to prenatal ultrasound had higher incidence rates of schizophrenia than those born in other hospitals without ultrasound. However, when we extended the analyses and compared individuals born at other hospitals of tertiary level with those born at primary and secondary hospitals, some of the other tertiary level hospitals also had high incidence rates of schizophrenia. Accordingly, factors other than ultrasound (factors that are unknown to us and seem to be related to place of birth) probably affect the risk of being diagnosed with schizophrenia. For other psychoses, the differences between ultrasound exposure and nonexposure were smaller and more ambiguous, which is in agreement with the opinion that nonschizophrenic psychoses are less associated with preand perinatal characteristics than schizophrenia.26
The women in our study had lower incidence rates of psychotic diseases than the men, which was most obvious for schizophrenia. The lifetime risk of schizophrenia is thought to be equal in men and women, although onset generally occurs about 4 years later in women (mean 26.5 years in men and 30.6 years in women).27 In the present study we were able to follow all individuals until 2004, which limited the maximum age to 31 years. This truncated follow-up time may have influenced the risk estimates, especially among women, and may also explain our lower incidence rates of schizophrenia among women. Thus, our conclusions should be restricted to subjects with a relatively early age of onset of schizophrenia.
Because the male fetal brain is considered more vulnerable than the female brain,16 and since higher risks of left-hand or mixed-hand preference in connection with prenatal ultrasound have been found only in men, we analyzed data according to sex. For most of the hospitals there was a reasonable agreement in risks of schizophrenia between men and women, ie, both men and women at a specific tertiary level hospital had either increased or decreased risks. Although women born in Malmö had increased risks of schizophrenia, women born at 3 other tertiary level hospitals had even higher risks. Accordingly, prenatal ultrasound does not seem to be associated with increased risks of schizophrenia in women.
The lowest risks of schizophrenia were found in individuals born in Umeå. As we have adjusted for county of residence during the follow-up period, the differences in risks between the tertiary hospitals should not reflect geographical variation in diagnostic and admittance routines. We believe that a certain degree of variation in incidence of schizophrenia between the hospitals is to be expected and that the variations found in this study may reflect differences in genetic, socio-demographic and other factors.
For the same cohort, we have previously reported an increased risk of left-handedness among men born in Malmö and exposed to ultrasound, when compared with unexposed. In that study all the subanalyses on hospital level supported the association.2 However, in a previous study on intellectual performance, the subanalyses on hospital level did not support an association with ultrasound.30 Similarly, the association with schizophrenia reported here, was not supported in the subanalyses on hospital level. Nevertheless, men born in Malmö had the highest risks of schizophrenia and poor intellectual performance.30 Though these may be unrelated to ultrasound exposure or result of chance, it is somewhat alarming that men born in Malmö and assumed to be exposed to ultrasound deviate in all 3 outcomes that we have assessed.
We adjusted for potential confounding factors, such as maternal age, parity and mother's number of hospitalizations for psychiatric care. Immigration, especially by persons from economically and genetically distinct populations, has been shown to increase risk of developing schizophrenia.29 In the present study we included individuals with Swedish mothers or with mothers that had immigrated to Sweden from the other Nordic countries, as the Nordic population is genetically and socioeconomically similar. Socioeconomic factors might affect risks of developing psychoses,28 yet in our previous studies we saw only minor differences in socioeconomic status between families living in Malmö compared with the rest of Sweden.30 We were unable to control for maternal smoking, as information on maternal smoking was not included in the Medical Birth Register for the birth cohorts included in the present investigation. Maternal smoking is causally related to fetal growth,31 and pregnant smokers may therefore be more likely to be exposed to repeated ultrasound scans during pregnancy. We suggest that in future studies maternal smoking should be included as a confounding factor in the analyses.
We were limited to include only patients with inpatient psychiatric treatment, as the information on outpatient treatment in the Hospital Discharge Register is incomplete. Routines for admitting psychiatric patients might vary among regions of Sweden. To minimize the effects of restricting inclusions to inpatients, we adjusted for county of residence during the follow-up period. The adjustments for living place had minor effects on the results. The psychiatric diagnoses in the Hospital Discharge Register are reliable as Swedish diagnostic practice is generally considered to be of high quality.20–22 Risk of misclassification in diagnosed cases is probably small, as the diagnosis of schizophrenia is most certainly used with caution.
Strengths of the present study include the population-based design, the large size, and the prospective collection of data on birth characteristics, which precludes recall bias. We considered children of mothers from Malmö and born in Malmö as exposed to ultrasound. To avoid risk of misclassification of exposure we included only hospitals with reliable documentation on introduction of ultrasound programs. More than 90% of women attending antenatal care in Malmö had prenatal ultrasound scans from 1973 and onwards. However, some of the women in Malmö might not have had ultrasound, and some of those considered as unexposed may have had an ultrasound examination done in another part of Sweden. This possible error might lead to an underestimation of the association and reduce the possibility of finding an existing relation between exposure and outcome. The ultrasound examinations were performed as part of a routine screening program and were not performed on clinical indication, which minimizes the risk of exposure bias.
Although we had sufficient power to detect even small increases in risk in the main analyses, the strata included in subanalyses were too small to demonstrate small changes in schizophrenia incidence when comparing each tertiary hospitals with other hospitals.
The present findings do not provide clear evidence of an association between prenatal ultrasound exposure and schizophrenia. Men born in Malmö had the highest risks compared with other men, whereas women born in Malmö were similar to women born at several other tertiary level hospitals. It seems that factors related to place of birth, rather than ultrasound exposure, may have influenced the results for both men and women. To overcome obstacles regarding uncertainty of exposure, adjustments for confounders and case ascertainment, we suggest that a case-control study should be performed using all available information from medical records.
1. Kieler H, Cnattingius S, Haglund B, et al. Sinistrality—a side-effect of prenatal sonography: a comparative study of young men. Epidemiology. 2001;12:618–623.
2. Kieler H, Cnattingius S, Palmgren J, et al. First trimester ultrasound scans and left-handedness. Epidemiology. 2002;13:370.
3. Salvesen KA, Vatten LJ, Eik-Nes SH, et al. Routine ultrasonography in utero and subsequent handedness and neurological development. BMJ. 1993;307:159–164.
4. Kieler H, Ahlsten G, Haglund B, et al. Routine ultrasound screening in pregnancy and the children's subsequent neurologic development. Obstet Gynecol. 1998;91:750–756.
5. Geschwind N, Galaburda AM. Cerebral lateralization. Biological mechanisms, associations, and pathology: I. A hypothesis and a program for research. Arch Neurol. 1985;42:428–459.
6. Merritt CR, Kremkau FW, Hobbins JC. Diagnostic ultrasound: bioeffects and safety. Ultrasound Obstet Gynecol. 1992;2:366–374.
7. Nowakowski RS. Basic concepts of CNS development. Child Dev. 1987;58:568–595.
8. Barnett S. Can diagnostic ultrasound heat tissue and cause biological effects? Safety of Diagnostic Ultrasound. Lancashire: The Partenon Publishing Group, 1998.
9. Ang ESBC Jr., Gluncic V, Duque A, et al. From the Cover: Prenatal exposure to ultrasound waves impacts neuronal migration in mice. Proc Natl Acad Sci U S A. 2006;103:12903–12910.
10. Cannon M, Clarke MC. Risk for schizophrenia—broadening the concepts, pushing back the boundaries. Schizophr Res. 2005;6:6.
11. Rice D, Barone S Jr. Critical periods of vulnerability for the developing nervous system: evidence from humans and animal models. Environ Health Perspect. 2000;108(Suppl 3):511–533.
12. Rehn AE, Rees SM. Investigating the neurodevelopmental hypothesis of schizophrenia. Clin Exp Pharmacol Physiol. 2005;32:687–696.
13. Harrison PJ, Law AJ. Neuregulin 1 and schizophrenia: genetics, gene expression, and neurobiology. Biol Psychiatry. 2006;24:24.
14. Tabares-Seisdedos R, Escamez T, Martinez-Gimenez JA, et al. Variations in genes regulating neuronal migration predict reduced prefrontal cognition in schizophrenia and bipolar subjects from mediterranean Spain: a preliminary study. Neuroscience. 2006;139:1289–1300.
15. Sommer I, Ramsey N, Kahn R, et al. Handedness, language lateralisation and anatomical asymmetry in schizophrenia: meta-analysis. Br J Psychiatry. 2001;178:344–351.
16. Witelson SF, Nowakowski RS. Left out axons make men right: a hypothesis for the origin of handedness and functional asymmetry. Neuropsychologia. 1991;29:327–333.
17. Cannon M, Jones PB, Murray RM. Obstetric complications and schizophrenia: historical and meta-analytic review. Am J Psychiatry. 2002;159:1080–1092.
18. Hultman CM, Sparen P, Takei N, et al. Prenatal and perinatal risk factors for schizophrenia, affective psychosis, and reactive psychosis of early onset: case-control study. BMJ. 1999;318:421–426.
19. Cnattingius S, Ericson A, Gunnarskog J, et al. A quality study of a medical birth registry. Scand J Soc Med. 1990;18:143–148.
20. Ekholm B, Ekholm A, Adolfsson R, et al. Evaluation of diagnostic procedures in Swedish patients with schizophrenia and related psychoses. Nord J Psychiatry. 2005;59:457–464.
21. Dalman C, Broms J, Cullberg J, et al. Young cases of schizophrenia identified in a national inpatient register–are the diagnoses valid? Soc Psychiatry Psychiatr Epidemiol. 2002;37:527–531.
22. Kristjansson E AP, Wistedt B. Validity of the diagnosis of schizophrenia in a psychiatric inpatient register. Nord Psykiatr Tidsskr. 1987;41:229–234.
23. Hogberg U, Larsson N. Early dating by ultrasound and perinatal outcome. A cohort study. Acta Obstet Gynecol Scand. 1997;76:907–912.
24. Grennert L, Persson PH, Gennser G. Benefits of ultrasonic screening of a pregnant population. Acta Obstet Gynecol Scand Suppl. 1978;78:5–14.
25. Kieler H, Axelsson O, Haglund B, et al. Routine ultrasound screening in pregnancy and the children's subsequent handedness. Early Hum Dev. 1998;50:233–245.
26. Hultman CM, Ohman A, Cnattingius S, et al. Prenatal and neonatal risk factors for schizophrenia. Br J Psychiatry. 1997;170:128–133.
27. Hafner H. Gender differences in schizophrenia. Psychoneuroendocrinology. 2003;28(Suppl 2):17–54.
28. Wicks S, Hjern A, Gunnell D, Lewis G, et al. Social Adversity in Childhood and the Risk of Developing Psychosis: a national cohort study. Am J Psychiatry. 2005;162:1652–1657.
29. Cantor-Graae E, Selten JP. Schizophrenia and migration: a meta-analysis and review. Am J Psychiatry. 2005;162:12–24.
30. Kieler H, Haglund B, Cnattingius S, et al. Does prenatal sonography affect intellectual performance? Epidemiology. 2005;16:304–310.
31. Windham GC, Hopkins B, Fenster L, et al. Prenatal active or passive tobacco smoke exposure and the risk of preterm delivery or low birth weight. Epidemiology. 2000;11:427–433.
© 2007 Lippincott Williams & Wilkins, Inc.