Re-examining the link between prenatal maternal anxiety and child emotional difficulties, using a sibling design

Re-examining the link between prenatal maternal anxiety and child emotional difficulties, using a... Abstract Background Prenatal exposure to maternal anxiety has been associated with child emotional difficulties in a number of epidemiological studies. One key concern, however, is that this link is vulnerable to confounding by pleiotropic genes or environmental family factors. Methods Data on 82 383 mothers and children from the population-based Mother and Child Cohort Study and data on 21 980 siblings were used in this study. Mothers filled out questionnaires for each unique pregnancy, for infant difficulties at 6 months and for emotional difficulties at 36 months. The link between prenatal maternal anxiety and child difficulties were examined using logistic regression analyses and multiple linear regression analyses for the full study sample and the sibling sample. Results In the conventional full-cohort analyses, prenatal exposure to maternal anxiety was associated with child difficulties at both 6 months [odds ratio (OR) = 2.1 (1.94–2.27)] and 36 months [OR = 2.72 (2.47–2.99)]. The findings were essentially the same whether we examined difficulties at 6 months or at 36 months. However, these associations were no longer present once we controlled for potential social and genetic confounders in the sibling comparison analyses, either at 6 months [OR = 1.32 (0.91–1.90)] or at 36 months [OR = 1.28 (0.63–2.60)]. Findings from multiple regression analyses with continuous measures were essentially the same. Conclusions Our finding lends little support for there being an independent prenatal effect on child emotional difficulties; rather, our findings suggest that the link between prenatal maternal anxiety and child difficulties could be confounded by pleiotropic genes or environmental family factors. Prenatal, anxiety, sibling design, emotional difficulties Key Messages In the full cohort (n = 82 383), we found that children exposed to maternal anxiety in pregnancy had about twice the risk for short- and long-term difficulties in the child. When the same association was examined within a sibling design (n = 21 980) of this cohort, this association disappeared. The findings from our sibling analyses suggest that the association between prenatal maternal anxiety and emotional difficulties is confounded by genetic or other familial factors. Introduction During the past half-century, animal research and parallel human findings suggest that prenatal exposure to maternal stress and anxiety influence offspring psychopathology.1 Prenatal maternal anxiety has been associated with both short- and long-term mental health problems in the child (i.e. difficulties such as emotional reactivity, symptoms of anxiety, somatic complaints and sleep problems).1–7 This link may be mediated through a programming effect, perhaps through increased production of cortisol, which can influence fetal brain development.8–10 Although findings in animal models have been robust in showing the link between elevated glucocorticoid levels following stress and adverse outcome in the offspring,1,10,11 scepticism has been raised as to the validity of the concepts and methods of several of the findings in human studies. Some interpretations are based on inconsistent results, low sample sizes, poor research designs and lack of appropriate genetic and postnatal controls.12,13 One key concern is that observational studies are vulnerable to confounding by pleiotropic genes.14,15 Rice and colleagues16 used a ‘prenatal cross-fostering’ design to examine the intra-uterine environment where pregnant mothers were either related or unrelated to their child, as a result of in vitro fertilization (IVF). The results suggest that the associations with prenatal stress were caused by either postnatal risks or common genes. However, this finding was limited by the use of retrospective recall of prenatal stress, and small sample size. Another powerful way of examining prenatal risk effects within epidemiological studies is through the use of sibling designs. For example, D’Onofrio and colleagues17 examined the well-established link between smoking during pregnancy and antisocial behaviour in a sibling design where they compared siblings that differed in their exposure to maternal smoking during pregnancy. Findings showed that siblings discordant in exposure to prenatal smoking did not differ in their risk for antisocial behaviour. This finding has been confirmed by other quasi-experimental studies.18–20 The implication from these studies is that relying on measured covariates to account for confounding could result in false-positive conclusions, because of insufficient control for postnatal exposure or pleiotropic genes. A sibling comparison design, where one sibling has been exposed to prenatal maternal anxiety and the other has not, would control for both genetic and environmental factors that are shared by the siblings. Because each child receives a random set of its mother´s genes through the process of meiosis,21 the sibling design provides a good control for pleiotropic genes that influence maternal anxiety and behaviour as well as emotional difficulties in the child.15 Using data from pregnancy questionnaires as well as responses to questions on child behaviour at 6 and 36 months, we aim to examine the risk associated with maternal anxiety in pregnancy on short- (6 months) and long-term (36 months) outcomes in the child, and to control potential social and genetic confounding using a sibling comparison design. Methods Study population and data collection This study is a subproject of the Norwegian Mother and Child Cohort Study (MoBa) conducted by the Norwegian Institute of Public Health.22 MoBa is a cohort comprising 113 000 pregnancies recruited from 1999 to 2008 with a participation rate of 40.6%.22 All but two of a total of 52 hospitals around Norway agreed to participate in the recruitment to the study. Women were invited to participate when they attended the routine ultrasound examination offered to all pregnant women at 17–18 weeks of gestation [www.fhi.no/moba-en]. The mothers filled out questionnaires about characteristics including their age, education, smoking and drinking habits, partner harmony and anxiety at the 17th and 30th gestational week. In addition, the current study used data from the questionnaires about child somatic health, infant difficulties and emotional difficulties at 6 months and 36 months after birth. The cohort was also linked to the Medical Birth Registry of Norway (MBRN).23 The MBRN contains detailed medical information about the infant, including birthweight and birth complications. The maternal questionnaire response rates at the 17th and 30th week of gestation, and at 6 and 36 months after birth, were approximately 95.1%, 91.4%, 87.0% and 61.4%, respectively.24 Written informed consent was obtained from all participating women. The Regional Committees for Medical and Health Research Ethics (REK) and the Norwegian Data Inspectorate have approved the study. Within MoBa, a total of 15 256 mothers participated with more than one pregnancy. Mothers filled out questionnaires for each unique pregnancy and for infant difficulties at 6 months, as well as for emotional difficulties at 36 months. For women participating with three or more pregnancies, two siblings were randomly selected. We used version 7 of the quality-assured data files for participants recruited in the period 1999 to 2009, see flow chart (Figure 1). Figure 1 View largeDownload slide Flow chart of sibling sample. Figure 1 View largeDownload slide Flow chart of sibling sample. Measure of maternal general anxiety Mothers reported on symptoms of anxiety using a validated short version of the Hopkins Symptom Checklist, the SCL-525 and SCL-8.26 The Symptom Checklist scale (SCL) is scored on a Likert scale ranging from 1 to 4. The short scale has been found to be valid with a correlation of 0.92 with the SCL-25.26 Assessments of anxiety were made twice during pregnancy (17th and 30th gestational weeks), and again when the child was 6 months old. Two out of five questions from the SCL-5 and four out of eight questions from the SCL-8 measured anxiety at 17th and 30th gestational weeks and at 6 months, respectively. The mean score for the 17th and 30th gestational weeks and 6 months measure ranged 1–2 and 1–4, respectively. Both continuous scores and a quasi-clinical cut-off (in the absence of an established cut-off, the top 15% of the sample was identified, which is equivalent to other studies, e.g. O’Donnell and colleagues,3 and Brandlistuen and colleagues27), are used in the analyses. Exploring alternative cut-points suggested essentially the same results. Based on these dichotomized variables, a four-level factor variable was constructed to represent ‘no exposure to maternal anxiety’, ‘high anxiety in week 17 only’, ‘high anxiety in week 30 only’ or ‘both’. Main outcome variables: Infant difficulties at 6 months were measured at 6 months postpartum by nine items from the Infant Characteristic Questionnaire (ICQ).28 Mothers responded to seven ordered response categories. The responses were scored 1–7, and the mean score calculated (range 1–7). The categorical variable was based on one standard deviation (SD) above the mean, and coded as low infant difficulties (‘0’), and high infant difficulties (‘1’). That is, the top 15%, which is equivalent to one standard deviation away from the centre, is regarded as high infant difficulties. Emotional difficulties were measured by 10 items from the Child Behaviour Checklist (CBCL/TRF),29 which represented four subscales: emotionally reactive, anxious/depressed, somatic complaints and sleep problems. Mothers reported the extent to which they agreed with the difficulties statements on a three-point Likert scale, from not true (‘0’), sometimes true (‘1’) and often true (‘2’). Mean scores were calculated and ranged 0–2, and the Cronbach`s alpha was 0.56. The categorical variable was based on one SD above the mean, and coded as low emotional difficulties (‘0’), and high emotional difficulties (‘1’). That is, approximately the top 15% is regarded as high emotional difficulties. A team of clinical and developmental psychologists selected the items from the CBCL used in MoBa. The selected items were based on theoretical and empirical representativeness and have been found to be representative with a correlation of 0.92 to the full scale.30 Assessment of potential confounders Potential confounding factors were considered if the variable could be associated with both prenatal maternal anxiety and child outcomes (i.e. causing the two to correlate), and were included in the adjusted model if associated with the exposure (prenatal maternal anxiety) or one of the two outcome measures. The following variables were considered as potential controls: alcohol consumption during pregnancy (coded as never ‘0’ and more than once a month ‘1’); and smoking in pregnancy (coded as never ‘0’, sometimes ‘1’ or daily ‘2’). In addition we controlled for: maternal anxiety reported at 6 months, low = ‘0’ and high = ‘1’; partner (dis)harmony (good partner relationship = ‘0’, and poor partner relationship = ‘1’); somatic disease (not present at 6 months ‘0’ and present ‘1’); marital status (married/living together ‘0’ and single ‘1’); maternal education (coded as higher university degree of +4 years ‘0’, college/university of 3 years ‘1’, college 1–2 years ‘2’, secondary school ‘3’); and maternal age (coded as <25 ‘0’, 25–29 ‘1’, 30–34 ‘2’ and >= 35 ‘3’). We also controlled for the following variables extracted from the MBRN: parity was coded 0 ‘0’, and >=1 ‘1’; gestational age (coded as <37 ‘0’, and >=37 ‘1’; birth complications, coded as yes ‘1’ or no ‘0’; child sex (girl ‘0’, boy ‘1’); and birthweight (coded as >=2500 g = ‘0’, <2500 g = ‘1’). Statistical analyses We present descriptive data on the full sample and the sibling sample, including the distribution on confounders and covariates for infant difficulties at 6 months and emotional difficulties at 36 months. After reporting descriptive data, we present the analyses for the full sample (n = 82 383) and the sibling sample (n = 21 980). We present the results from the logistic regression analyses and multiple linear regression analyses for the full study sample and the sibling sample. We also adjusted for the following control variables: maternal age, educational level, marital status, parity, partner relationship, prenatal alcohol consumption, prenatal smoking, birth complications, child’s gender, birthweight, child’s somatic diseases and child’s gestational age. We also examined the potential mediating effect of the control variables: birth complication, birthweight, gestational age and child somatic disease, in a proportion; however, the mediation effect, was not clearly present. The mediating effect of the four variables was 0.11%, 0.09%, 6.27% and 1.42% respectively. For the sibling data with the binary outcomes, we used a conditional logistic regression model containing a family-specific term in the linear predictor that captures environmental and genetic factors common to the family. When conditioning on the number of siblings in each family with the reported outcome, these nuisance terms vanish, but the remaining terms are unchanged. This gives a conditional logistic regression model and we used the ‘clogit’ function in R to fit the model. Also for the continuous scores, we can include family-specific terms in the linear predictors that will cancel out when conditioning on the total reported scores in each family. Equivalently, as we considered only two siblings in each family, we analysed the differences in the siblings’ scores using differences in sibling covariates as explanatory variables. To handle missing values, we used pairwise deletion of missing. For the full cohort (n = 82 383), the crude model deleted 2145 observations due to missing. In the sibling design (n = 21 980), only 413 were taken out when we have the crude model fitted. The total number of discordant siblings was 1476 at 6 months and 786 at 36 months. Results The characteristics of the full cohort sample are shown in Table 1 and the sibling sub-sample in Table 2. In the sibling sub-sample, the proportion of children with difficulties at 6 months after birth was 20.4% among mothers reporting anxiety at week 17 and 20% among mothers with anxiety at week 30. This is comparable to 22%, which was found in the full cohort. In the sibling sub-sample, approximately 26% of the children exposed to high prenatal maternal anxiety at both 17th and 30th weeks of gestation had emotional difficulties at 36 months. This is somewhat lower than for the full cohort, which was approximately 30% at both 17th and 30th weeks of gestation. Of 10 990 pairs of siblings (n = 21 980), 86.3% and 86.9% turned out to be concordant for not being exposed to maternal anxiety at the 17th and 30th weeks of gestation, respectively. Also, 1.8% (17th week) and 2.5% (30th week) of the sibling pairs were both exposed to anxiety at these two points in time. However, 9.9% and 9.7% differed in exposure for maternal anxiety at the 17th and 30th week, respectively. The proportion of children with high score on child difficulties at both 6 and 36 months was 3.4%. Table 1 Characteristics of the Full Cohort   Infant Difficulties at 6 months   Emotional Difficulties 36 months   N, total  n  %  N, total  n  %  Total  82,383  12,613  15.3  53,917  9,750  18.1  General anxiety week 17   Low  73,028  10,488  14.36  48,431  8,151  16.83   High  8,074  1,852  22.94  4,802  1,446  30.11   Missing  1,281  273  21.31  684  153  22.37  General anxiety week 30   Low  73,230  10,569  14.43  48,536  8,145  16.78   High  8,547  1,893  22.15  5,087  1,528  30.04   Missing  606  151  24.92  294  77  26.19  General anxiety 6 months   Low  75,207  10,801  14.36  49,732  8,404  16.90   High  6,828  1,721  25.21  4,041  1,306  32.32   Missing  348  91  26.15  144  40  27.78  Mother's age   <25  8,405  1,566  18.63  4,833  1,077  22.28   25–29  27,478  4,409  16.05  17,957  3,266  18.19   30–34  32,081  4,643  14.47  21,412  3,673  17.15   >35  14,314  1,980  13.83  9,646  1,721  17.84   Missing  105  15  14.29  69  13  18.84  Education   University 4+  18,828  2,803  14.89  13,191  2,232  16.92   College/University 3y  32,948  4,756  14.43  22,525  3,701  16.43   College 1–2y  11,316  1,738  15.36  6,938  1,386  19.98   Secondary school  15,253  2,647  17.35  8,790  1,900  21.62   Missing  4,038  669  16.57  2,473  531  21.47  Marital status   Married/partners  79,325  12,050  15.19  52,069  9,311  17.88   Single  2,953  548  18.56  1,779  426  23.95   Missing  105  15  14.29  69  13  18.84  Partner harmony   High  69,814  10,115  14.49  46,386  7,964  17.17   Low  7,990  1,647  20.61  4,910  1,239  25.23   Missing  4,579  851  18.58  2,621  547  20.87  Alcohol in pregnancy   Never  67,959  10,167  14.96  45,772  8,140  17.78   >1 a month  480  82  17.08  341  93  27.27   Missing  13,944  2,364  16.95  7,804  1,517  19.44  Smoking status   Never  74,456  11,341  15.23  49,406  8,744  17.70   Sometimes  3,502  557  15.91  2,022  439  21.71   Daily  3,146  483  15.35  1,716  408  23.78   Missing  1,279  232  18.14  773  159  20.57  Birth complication   No  59,757  9,327  15.61  39,589  7,376  18.63   Yes  22,521  3,271  14.52  14,259  2,361  16.56   Missing  105  15  14.29  69  13  18.84  Parity   0(=0)  37,734  6,399  16.96  25,788  5,347  20.73   >=1(=1)  44,544  6,199  13.92  28,060  4,390  15.65   Missing  105  15  14.29  69  13  18.84  Child sex   Girls  40,170  5,720  14.24  26,317  4,902  18.63   Boys  42,108  6878  16.33  27,531  4835  17.56   Missing  105  15  14.29  69  13  18.84  Birth weight   >2500  80,245  12,168  15.16  52,468  9,412  17.94   <2500  1,989  422  21.22  1,358  321  23.64   Missing  149  23  15.44  91  17  18.68  Somatic disease 6m   No  56,759  7,100  12.51  39,115  6,610  16.90  vYes  17,590  4,420  25.13  11,271  2,451  21.75   Missing  8,034  1,093  13.60  3,531  689  19.51  Gestational age  v<37  3,533  681  19.28  2,412  515  21.35  v>=37  78,420  11,858  15.12  51,219  9,184  17.93  vMissing  430  74  17.21  286  51  17.83    Infant Difficulties at 6 months   Emotional Difficulties 36 months   N, total  n  %  N, total  n  %  Total  82,383  12,613  15.3  53,917  9,750  18.1  General anxiety week 17   Low  73,028  10,488  14.36  48,431  8,151  16.83   High  8,074  1,852  22.94  4,802  1,446  30.11   Missing  1,281  273  21.31  684  153  22.37  General anxiety week 30   Low  73,230  10,569  14.43  48,536  8,145  16.78   High  8,547  1,893  22.15  5,087  1,528  30.04   Missing  606  151  24.92  294  77  26.19  General anxiety 6 months   Low  75,207  10,801  14.36  49,732  8,404  16.90   High  6,828  1,721  25.21  4,041  1,306  32.32   Missing  348  91  26.15  144  40  27.78  Mother's age   <25  8,405  1,566  18.63  4,833  1,077  22.28   25–29  27,478  4,409  16.05  17,957  3,266  18.19   30–34  32,081  4,643  14.47  21,412  3,673  17.15   >35  14,314  1,980  13.83  9,646  1,721  17.84   Missing  105  15  14.29  69  13  18.84  Education   University 4+  18,828  2,803  14.89  13,191  2,232  16.92   College/University 3y  32,948  4,756  14.43  22,525  3,701  16.43   College 1–2y  11,316  1,738  15.36  6,938  1,386  19.98   Secondary school  15,253  2,647  17.35  8,790  1,900  21.62   Missing  4,038  669  16.57  2,473  531  21.47  Marital status   Married/partners  79,325  12,050  15.19  52,069  9,311  17.88   Single  2,953  548  18.56  1,779  426  23.95   Missing  105  15  14.29  69  13  18.84  Partner harmony   High  69,814  10,115  14.49  46,386  7,964  17.17   Low  7,990  1,647  20.61  4,910  1,239  25.23   Missing  4,579  851  18.58  2,621  547  20.87  Alcohol in pregnancy   Never  67,959  10,167  14.96  45,772  8,140  17.78   >1 a month  480  82  17.08  341  93  27.27   Missing  13,944  2,364  16.95  7,804  1,517  19.44  Smoking status   Never  74,456  11,341  15.23  49,406  8,744  17.70   Sometimes  3,502  557  15.91  2,022  439  21.71   Daily  3,146  483  15.35  1,716  408  23.78   Missing  1,279  232  18.14  773  159  20.57  Birth complication   No  59,757  9,327  15.61  39,589  7,376  18.63   Yes  22,521  3,271  14.52  14,259  2,361  16.56   Missing  105  15  14.29  69  13  18.84  Parity   0(=0)  37,734  6,399  16.96  25,788  5,347  20.73   >=1(=1)  44,544  6,199  13.92  28,060  4,390  15.65   Missing  105  15  14.29  69  13  18.84  Child sex   Girls  40,170  5,720  14.24  26,317  4,902  18.63   Boys  42,108  6878  16.33  27,531  4835  17.56   Missing  105  15  14.29  69  13  18.84  Birth weight   >2500  80,245  12,168  15.16  52,468  9,412  17.94   <2500  1,989  422  21.22  1,358  321  23.64   Missing  149  23  15.44  91  17  18.68  Somatic disease 6m   No  56,759  7,100  12.51  39,115  6,610  16.90  vYes  17,590  4,420  25.13  11,271  2,451  21.75   Missing  8,034  1,093  13.60  3,531  689  19.51  Gestational age  v<37  3,533  681  19.28  2,412  515  21.35  v>=37  78,420  11,858  15.12  51,219  9,184  17.93  vMissing  430  74  17.21  286  51  17.83  Note: n and % corresponds to infant difficulties and emotional difficulties. Table 2 Characteristics of the Siblings Sub-sample   Infant Difficulties at 6 months   Emotional Difficulties 36 months   N, total  n  %  N, total  n  %  Total  21,980  3,067  14.0  13,252  1,926  14.5  General anxiety week 17   Low  20,241  2,721  13.44  12,385  1,705  13.77   High  1,505  307  20.40  756  198  26.19   Missing  234  39  16.67  111  23  20.72  General anxiety week 30   Low  20,248  2,719  13.43  12,372  1,698  13.72   High  1,628  326  20.02  830  218  26.27   Missing  104  22  21.15  50  10  20.00  General anxiety 6 months   Low  20,546  2,743  13.35  12,497  1,706  13.65   High  1,374  309  22.49  734  210  28.61   Missing  60  15  25.00  21  10  47.62  Mother's age   <25  1,889  310  16.41  944  168  17.80   25–29  7,828  1,120  14.31  4,683  680  14.52   30–34  9,122  1,221  13.39  5,673  811  14.30   >35  3,111  409  13.15  1,933  262  13.55   Missing  30  7  23.33  19  5  26.32  Education   University 4+  5,781  790  13.67  3,711  530  14.28   College/University 3y  9,837  1,297  13.18  6,087  827  13.59   College 1–2y  2,504  370  14.78  1,411  237  16.80   Secondary school  2,967  476  16.04  1,551  242  15.60   Missing  891  134  15.04  492  90  18.29  Marital status   Married/partners  21,543  2,997  13.91  13,031  1,896  14.55   Single  407  63  15.48  202  25  12.38   Missing  30  7  23.33  19  5  26.32  Partner harmony   High  19,213  2,559  13.32  11,713  1,637  13.98   Low  1,905  377  19.79  1,083  223  20.59   Missing  862  131  15.20  456  66  14.47  Alcohol in pregnancy   Never  18,580  2,553  13.74  11,447  1,650  14.41   >1 a month  89  9  10.11  55  11  20.00   Missing  3,311  505  15.25  1,750  265  15.14  Smoking status   Never  20,571  2,855  13.88  12,535  1,815  14.48   Sometimes  601  84  13.98  306  50  16.34   Daily  483  82  16.98  225  38  16.89   Missing  325  46  14.15  186  23  12.37  Birth complication   No  15,369  2,209  14.37  9,350  1,410  15.08   Yes  6,581  851  12.93  3,883  511  13.16   Missing  30  7  23.33  19  5  26.32  Parity   0(=0)  8,295  1,289  15.54  5,049  889  17.61   >=1(=1)  13,655  1,771  12.97  8,184  1,032  12.61   Missing  30  7  23.33  19  5  26.32  Child sex   Girls  10,707  1,384  12.93  6,407  989  15.44   Boys  11,243  1676  14.91  6,826  932  13.65   Missing  30  7  23.33  19  5  26.32  Birth weight   >2500  21,505  2,968  13.80  12,964  1,873  14.45   <2500  433  90  20.79  263  48  18.25   Missing  42  9  21.43  25  5  20.00  Somatic disease 6M   No  15,393  1,758  11.42  9,732  1,323  13.59   Yes  4,751  1,090  22.94  2,813  489  17.38   Missing  1,836  219  11.93  707  114  16.12  Gestational age   <37  816  148  18.14  499  81  16.23   >=37  21,065  2,900  13.77  12,698  1,837  14.47   Missing  99  19  19.19  55  8  14.55    Infant Difficulties at 6 months   Emotional Difficulties 36 months   N, total  n  %  N, total  n  %  Total  21,980  3,067  14.0  13,252  1,926  14.5  General anxiety week 17   Low  20,241  2,721  13.44  12,385  1,705  13.77   High  1,505  307  20.40  756  198  26.19   Missing  234  39  16.67  111  23  20.72  General anxiety week 30   Low  20,248  2,719  13.43  12,372  1,698  13.72   High  1,628  326  20.02  830  218  26.27   Missing  104  22  21.15  50  10  20.00  General anxiety 6 months   Low  20,546  2,743  13.35  12,497  1,706  13.65   High  1,374  309  22.49  734  210  28.61   Missing  60  15  25.00  21  10  47.62  Mother's age   <25  1,889  310  16.41  944  168  17.80   25–29  7,828  1,120  14.31  4,683  680  14.52   30–34  9,122  1,221  13.39  5,673  811  14.30   >35  3,111  409  13.15  1,933  262  13.55   Missing  30  7  23.33  19  5  26.32  Education   University 4+  5,781  790  13.67  3,711  530  14.28   College/University 3y  9,837  1,297  13.18  6,087  827  13.59   College 1–2y  2,504  370  14.78  1,411  237  16.80   Secondary school  2,967  476  16.04  1,551  242  15.60   Missing  891  134  15.04  492  90  18.29  Marital status   Married/partners  21,543  2,997  13.91  13,031  1,896  14.55   Single  407  63  15.48  202  25  12.38   Missing  30  7  23.33  19  5  26.32  Partner harmony   High  19,213  2,559  13.32  11,713  1,637  13.98   Low  1,905  377  19.79  1,083  223  20.59   Missing  862  131  15.20  456  66  14.47  Alcohol in pregnancy   Never  18,580  2,553  13.74  11,447  1,650  14.41   >1 a month  89  9  10.11  55  11  20.00   Missing  3,311  505  15.25  1,750  265  15.14  Smoking status   Never  20,571  2,855  13.88  12,535  1,815  14.48   Sometimes  601  84  13.98  306  50  16.34   Daily  483  82  16.98  225  38  16.89   Missing  325  46  14.15  186  23  12.37  Birth complication   No  15,369  2,209  14.37  9,350  1,410  15.08   Yes  6,581  851  12.93  3,883  511  13.16   Missing  30  7  23.33  19  5  26.32  Parity   0(=0)  8,295  1,289  15.54  5,049  889  17.61   >=1(=1)  13,655  1,771  12.97  8,184  1,032  12.61   Missing  30  7  23.33  19  5  26.32  Child sex   Girls  10,707  1,384  12.93  6,407  989  15.44   Boys  11,243  1676  14.91  6,826  932  13.65   Missing  30  7  23.33  19  5  26.32  Birth weight   >2500  21,505  2,968  13.80  12,964  1,873  14.45   <2500  433  90  20.79  263  48  18.25   Missing  42  9  21.43  25  5  20.00  Somatic disease 6M   No  15,393  1,758  11.42  9,732  1,323  13.59   Yes  4,751  1,090  22.94  2,813  489  17.38   Missing  1,836  219  11.93  707  114  16.12  Gestational age   <37  816  148  18.14  499  81  16.23   >=37  21,065  2,900  13.77  12,698  1,837  14.47   Missing  99  19  19.19  55  8  14.55  Note: n and % corresponds to infant difficulties and emotional difficulties. Full cohort analyses In the logistic regression analyses on the full cohort, we found that children exposed to maternal anxiety at both 17th and 30th week of gestation had a higher risk of infant difficulties at 6 months and emotional difficulties at 36 months, as compared with those who had not been exposed to maternal anxiety (Table 3). These associations remained, but were somewhat reduced in size after controlling for a number of potential confounders, including maternal anxiety measured at 6 months. Examination of different cut-points did not indicate fundamentally different findings, e.g. 90th percentile suggested a positive change in OR of 0.32 at 6 months and 0.58 at 36 months. Table 3 The Effect of Prenatal Anxiety on Infant Difficulties 6 months and Emotional Difficulties 36 months in Full cohort   Infant Difficulties 6 months  Emotional Difficulties 36 months  Logistic regression  Crude  Adjusted*  Crude  Adjusted**    OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  Sample size(=n)  (n=80,238)  (n=55,065)  (n=52,608)  (n=37,982)    No Anxiety (reference)  1  1  1  1  17th week only  1.63 (1.51, 1.76)  1.37 (1.24, 1.51)  1.91 (1.75, 2.09)  1.47 (1.32, 1.65)  30th week only  1.51 (1.4, 1.63)  1.21 (1.1, 1.34)  1.96 (1.8, 2.14)  1.5 (1.34, 1.67)  Both 17th and 30th week  2.09 (1.94, 2.27)  1.38 (1.23, 1.55)  2.72 (2.47, 2.99)  1.84 (1.62, 2.1)    Multiple linear regression  Crude  Adjusted*  Crude  Adjusted**    Beta (95% CI)  Beta (95% CI)  Beta (95% CI)  Beta (95% CI)    General Anxiety          17th week only  0.28 (0.24, 0.31)  0.16 (0.12, 0.2)  0.12 (0.1, 0.13)  0.06 (0.05, 0.08)  30th week only  0.34 (0.3, 0.37)  0.16 (0.12, 0.21)  0.16 (0.14, 0.17)  0.09 (0.07, 0.1)  Both 17th and 30th week  0.5 (0.46, 0.54)  0.28 (0.19, 0.36)  0.23 (0.21, 0.24)  0.16 (0.12, 0.19)    Infant Difficulties 6 months  Emotional Difficulties 36 months  Logistic regression  Crude  Adjusted*  Crude  Adjusted**    OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  Sample size(=n)  (n=80,238)  (n=55,065)  (n=52,608)  (n=37,982)    No Anxiety (reference)  1  1  1  1  17th week only  1.63 (1.51, 1.76)  1.37 (1.24, 1.51)  1.91 (1.75, 2.09)  1.47 (1.32, 1.65)  30th week only  1.51 (1.4, 1.63)  1.21 (1.1, 1.34)  1.96 (1.8, 2.14)  1.5 (1.34, 1.67)  Both 17th and 30th week  2.09 (1.94, 2.27)  1.38 (1.23, 1.55)  2.72 (2.47, 2.99)  1.84 (1.62, 2.1)    Multiple linear regression  Crude  Adjusted*  Crude  Adjusted**    Beta (95% CI)  Beta (95% CI)  Beta (95% CI)  Beta (95% CI)    General Anxiety          17th week only  0.28 (0.24, 0.31)  0.16 (0.12, 0.2)  0.12 (0.1, 0.13)  0.06 (0.05, 0.08)  30th week only  0.34 (0.3, 0.37)  0.16 (0.12, 0.21)  0.16 (0.14, 0.17)  0.09 (0.07, 0.1)  Both 17th and 30th week  0.5 (0.46, 0.54)  0.28 (0.19, 0.36)  0.23 (0.21, 0.24)  0.16 (0.12, 0.19)  *Adjusted for maternal anxiety 6 months, mother’s age, education, marital status, parity, partner relation, alcohol intake, smoking status, birth complication, child’s sex, birth weight, somatic disease, gestational age. **Also adjusted for infant difficulties at 6 months. Bold are significant at p<0.001. We also examined these associations using multiple regression analyses with continuous measures for the full sample. The results showed moderate associations with infant difficulties at 6 months, as well as emotional difficulties at 36 months. The follow-up analyses at 36 months showed that there was no difference in effects, when we examined the associations between prenatal maternal anxiety on infant difficulties at 6 months and emotional difficulties at 36 months, suggesting that some overall liability is involved. Sibling comparison analyses We examined the same associations in a sibling comparison design, to control for potential social and genetic confounders. In the crude and adjusted sibling analyses for the conditional logistic regression, no associations were found between prenatal maternal anxieties, measured at either the 17th or 30th week of gestation or both, on infant difficulties at 6 months or emotional difficulties at 36 months of age (Table 4). Table 4 The Effect of Prenatal Anxiety on Infant Difficulties 6 months and Emotional Difficulties 36 months in sibling sub-sample (full siblings)   Infant Difficulties at 6 months  Emotional Difficulties at 36 months  Conditional logistic regression  Crude  Adjusted*  Crude  Adjusted**  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  Sample size(=n)  (n=21,571)  (n=15,664)  (n=13,011)  (n=9,844)    No Anxiety (reference)  1  1  1  1  17th week only  1.12 (0.84, 1.49)  0.92 (0.61, 1.4)  1.33 (0.9, 1.96)  0.78 (0.46, 1.33)  30th week only  1.15 (0.86, 1.53)  0.91 (0.59, 1.4)  0.99 (0.69, 1.44)  1.04 (0.6, 1.77)  Both 17th and 30th week***  1.3 (0.9, 1.87)  1.01 (0.57, 1.79)  1.07 (0.64, 1.78)  1.13 (0.54, 2.36)    Multiple linear regression  Crude Beta (95% CI)  Adjusted*  Beta (95% CI)  Crude Beta (95% CI)  Adjusted**  Beta (95% CI)  Sample size(=n)  (n=10,594)  (n=5,935)  (n=6,393)  (n=3,877)    General Anxiety          17th week only  0.047 (−0.001, 0.095)  0.033 (−0.031, 0.096)  0.005 (−0.013, 0.023)  0.001 (−0.022, 0.025)  30th week only  0.04 (−0.022, 0.102)  0.02 (−0.061, 0.101)  −0.011 (−0.034, 0.012)  −0.024 (−0.055, 0.006)  Both 17th and 30th week***  0.065 (−0.046, 0.175)  −0.03 (−0.174, 0.115)  0.015 (−0.027, 0.058)  −0.002 (−0.054, 0.051)    Infant Difficulties at 6 months  Emotional Difficulties at 36 months  Conditional logistic regression  Crude  Adjusted*  Crude  Adjusted**  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  Sample size(=n)  (n=21,571)  (n=15,664)  (n=13,011)  (n=9,844)    No Anxiety (reference)  1  1  1  1  17th week only  1.12 (0.84, 1.49)  0.92 (0.61, 1.4)  1.33 (0.9, 1.96)  0.78 (0.46, 1.33)  30th week only  1.15 (0.86, 1.53)  0.91 (0.59, 1.4)  0.99 (0.69, 1.44)  1.04 (0.6, 1.77)  Both 17th and 30th week***  1.3 (0.9, 1.87)  1.01 (0.57, 1.79)  1.07 (0.64, 1.78)  1.13 (0.54, 2.36)    Multiple linear regression  Crude Beta (95% CI)  Adjusted*  Beta (95% CI)  Crude Beta (95% CI)  Adjusted**  Beta (95% CI)  Sample size(=n)  (n=10,594)  (n=5,935)  (n=6,393)  (n=3,877)    General Anxiety          17th week only  0.047 (−0.001, 0.095)  0.033 (−0.031, 0.096)  0.005 (−0.013, 0.023)  0.001 (−0.022, 0.025)  30th week only  0.04 (−0.022, 0.102)  0.02 (−0.061, 0.101)  −0.011 (−0.034, 0.012)  −0.024 (−0.055, 0.006)  Both 17th and 30th week***  0.065 (−0.046, 0.175)  −0.03 (−0.174, 0.115)  0.015 (−0.027, 0.058)  −0.002 (−0.054, 0.051)  *Adjusted for maternal anxiety 6 months, mother’s age, education, marital status, parity, partner relation, alcohol intake, smoking status, birth complication, child’s sex, birth weight, somatic disease, gestational age. **Also adjusted for Infant Difficulties at 6 months. *** β^17th week+β^30th week+β^17th weekβ^30th week. We also examined these associations using multiple regression analyses (Table 4). In contrast to what was found in the full cohort, there were no associations between prenatal maternal anxiety and either infant difficulties at 6 months or emotional difficulties at 36 months. However, a moderate association with maternal anxiety reported at 6 months postnatally remained. Discussion Using a large population-based longitudinal cohort study we assessed the link between prenatal maternal anxiety and child emotional difficulties using a sibling design. There are two key findings that stand out. First, the conventional full-cohort analyses replicated common findings,9 but the same analyses, controlled for pleiotropic genes or environmental family factors, did not. Second, the findings were essentially the same whether we examined infant difficulties at 6 months or emotional difficulties at 36 months. In the first analyses using the full cohort, children prenatally exposed to maternal anxiety had twice the risk of infant difficulties at 6 months, and almost three times the risk of emotional difficulties at 36 months. This association did not change substantially after controlling for multiple covariates. This is consistent with a number of other studies3,6,9 and is almost an exact replicate of the recent findings by O’Donnell et al.3 They found that a 2-fold increase in risk of a probable child mental disorder was associated with exposure to prenatal maternal anxiety.3 This is equivalent to our findings from the full-cohort analyses (both logistic and multiple regression). However, once the same analyses were conducted using a sibling design, these associations were no longer found either for infant difficulties at 6 months or emotional difficulties measured at 36 months. This suggests that the substantial findings on the full cohort are likely to be confounded by pleiotropic genetic or constant environmental family factors. Although the use of sibling design of discordant siblings reflects a reduction in sample size, a major strength of our study is the consistency in the findings across 6 and 36 months outcomes, as well as across both logistic regression and multiple regression analyses. Our findings further suggest little support for there being an independent prenatal effect on child difficulties and suggest that statistical control for confounders does not deal adequately with the issues, but that the sibling design comparing exposed and non-exposed siblings does. However, although the within-pair estimates will not be confounded by factors that are shared by the siblings, bias could still occur due to non-shared factors.31 Therefore, several covariates (i.e. birth order, maternal age etc.) were adjusted for, without significant changes to the results. However, although our findings lend little support for there being an independent prenatal effect on child emotional difficulties, there is still the possibility that there may be a prenatal effect on other child outcomes. To our knowledge this is the first study to examine the link between prenatal maternal anxiety and child difficulties at both 6 months and 36 months of age in siblings that differ in exposure to prenatal maternal anxiety. Because we cannot randomize pregnant mothers to anxiety exposure and because twins will always be concordant in prenatal risk exposure, the sibling-comparison design is optimal when examining prenatal risk effects and offers a powerful quasi-experimental approach to study prenatal risks. There are, however, limitations that need to be mentioned. The sample size in MoBa limits the possibility of providing clinical interviews in assessing anxiety in the mothers, or clinical diagnosis of difficulties, such as anxiety, in children. However, this study used validated questionnaires that, although they are,short-scales, have been found to correlate highly with the original large-scale questionnaires.25,26,30 Bias could still occur due to attrition or because of halo effects. Maternal reports of infant difficulties and emotional difficulties are subjective and therefore their levels of stress during the pre- and postnatal periods could affect their perceptions of child difficulties. That is, anxious mothers may report their child´s difficulties as more negative than non-anxious mothers. However, using the same informant–the mother–is unavoidable when investigating behaviour in very young children within large-scale population cohort studies, and is a common feature for similar studies (for example, the Avon Longitudinal Study of Parents and Children; the Quebec Longitudinal Study of Child Development). Bias due to selective recruitment is another possible limitation regarding prevalence, but has minimal influence on associations,22,32,33 which was the focus in this study. In sum, our finding suggests that statistical control for confounders does not deal adequately with issues of confounding, but that a sibling design comparing exposed and non-exposed siblings does.15 Funding The Norwegian Mother and Child Cohort Study is supported by the Norwegian Ministry of Health and Care Services and the Ministry of Education and Research, NIH/NIEHS (contract no N01-ES-75558), NIH/NINDS (grant no.1 UO1 NS 047537–01 and grant no.2 UO1 NS 047537–06A1). Acknowledgement We are grateful to all the participating families in Norway who take part in this ongoing cohort study. Conflict of interest: None declared. References 1 Glover V, O’Connor TG, O’Donnell KJ. Prenatal stress and the programming of the HPA axis. Neurosci Biobehav Rev  2010; 35: 17– 22. Google Scholar CrossRef Search ADS PubMed  2 Davis EP, Snidman N, Wadhwa PD, Glynn LM, Schetter CD, Sandman CA. Prenatal maternal anxiety and depression predict negative behavioral reactivity in infancy. Infancy  2004; 6: 319– 31. Google Scholar CrossRef Search ADS   3 O’Donnell KJ, Glover V, Barker ED, O’Connor TG. The persisting effect of maternal mood in pregnancy on childhood psychopathology. Dev Psychopath  2014; 26: 393– 403. Google Scholar CrossRef Search ADS   4 DiPietro JA, Novak MFSX, Costigan KA, Atella LD, Reusing SP. Maternal psychological distress during pregnancy in relation to child development at age two. Child Dev  2006; 77: 573– 87. Google Scholar CrossRef Search ADS PubMed  5 McGrath JM, Records K, Rice M. Maternal depression and infant characteristics. Infant Behav Dev  2008; 31: 71– 80. Google Scholar CrossRef Search ADS PubMed  6 Kingsbury M, Weeks M, MacKinnon N, et al.   Stressful life events during pregnancy and offspring depression: evidence from a prospective cohort study. J Am Acad Child Adolesc Psychiatry  2016; 55: 709– 16. Google Scholar CrossRef Search ADS PubMed  7 Talge NM, Neal C, Glover V. Antenatal maternal stress and long-term effects on child neurodevelopment: How and Why? J Child Psychol Psychiatry  2007; 48: 245– 62. Google Scholar CrossRef Search ADS PubMed  8 Austin M-P, Leader LR, Reilly N. Prenatal stress, the hypothalamic-pituitary-adrenal axis, and fetal and infant neurobehavior. Early Hum Dev  2005; 81: 917– 26. Google Scholar CrossRef Search ADS PubMed  9 Glover V, O’Connor TG. Effects of antenatal stress and anxiety. Implications for development and psychiatry. Br J Psychiatry  2002; 180: 389– 91. Google Scholar CrossRef Search ADS PubMed  10 Welberg LA, Seckl JR. Prenatal stress, glucocorticoids and the programming of the brain. J Neuroendocrinol  2001; 13: 113– 28. Google Scholar CrossRef Search ADS PubMed  11 Weinstock M. Alterations induced by gestational stress in brain morphology and behaviour of the offspring. Brain Behav Immun  2001; 19: 296– 308. Google Scholar CrossRef Search ADS   12 Thapar A, Rutter M. Do prenatal risk factors cause psychiatric disorder? Be wary of causal claims. Br J Psychiatry  2009; 195: 100– 01. Google Scholar CrossRef Search ADS PubMed  13 Rutter M, Pickles A. Annual research review: threats to the validity of child psychiatry and psychology. J Child Psychol Psychiatry  2016; 57: 398– 416. Google Scholar CrossRef Search ADS PubMed  14 Rutter M. Annual research review: resilience: clinical implications. J Child Psychol Psychiatry  2012; 4: 474– 87. 15 D’Onofrio BM, Lahey BB, Turkheimer E, Lichtenstein P. Critical need for family-based, quasi-experimental designs in integrating genetic and social science research. Am J Public Health  2013; 103: 46– 55. Google Scholar CrossRef Search ADS   16 Rice F, Harold GT, Boivin J, van den Bree M, Hay DF, Thapar A. The links between prenatal stress and offspring development and psychopathology: disentangling environmental and inherited influences. Psych Med  2010; 40: 335– 45. Google Scholar CrossRef Search ADS   17 D’Onofrio BM, van Hulle CA, Waldman ID, et al.   Smoking during pregnancy and offspring externalizing problems: An exploration of genetic and environmental confounds. Dev Psychopath  2008; 20: 139– 64. 18 Obel C, Linnet KM, Henriksen TB, et al.   Smoking during pregnancy and hyperactivity-innattention in the offspring – comparing results from three Nordic cohorts. Int J Epidemiol  2009; 38: 698– 705. Google Scholar CrossRef Search ADS PubMed  19 Obel C, Zhu JL, Olsen J, et al.   The risk of attention deficit hyperactivity disorder in children exposed to maternal smoking during pregnancy – a re-examination using a sibling design. J Child Psychol Psychiatry  2016; 57: 532– 37. Google Scholar CrossRef Search ADS PubMed  20 Rice F, Harold GT, Boivin J, Hay DF, van den Bree M, Thapar A. Disentangling prenatal and inherited influences in humans with an experimental design. Proc Natl Acad Sci U S A  2009; 106: 2464– 67. Google Scholar CrossRef Search ADS PubMed  21 Rutter M. Proceeding from observed correlation to causal inference: The use of natural experiments. Perspect Psychol Sc  2007; 2: 377– 95. Google Scholar CrossRef Search ADS   22 Magnus P, Birke C, Vejrup K, et al.   Cohort Profile Update: The Norwegian Mother and Child Cohort Study (MoBa). Int J Epidemiol  2016; 45: 382– 88. Google Scholar CrossRef Search ADS PubMed  23 Irgens LM. The Medical Birth Registry of Norway. Epidemiological research and surveillance throughout 30 years. Acta Obstet Gynaecol Scand  2000; 79: 435– 39. Google Scholar CrossRef Search ADS   24 Schreuder P, Alsaker E. The Norwegian Mother and Child Cohort Study (MoBa) – MoBa recruitment and logistics. Norsk Epidemiol  2014; 24: 23– 27. Google Scholar CrossRef Search ADS   25 Tambs K, Moum T. How well can a few questionnaire items indicate anxiety and depression? Acta Psychiatr Scand  1993; 87: 364– 67. Google Scholar CrossRef Search ADS PubMed  26 Tambs K, Roysamb E. Selection of questions to short-form versions of original psychometric instruments in MoBa. Norsk Epidemiol  2014; 24: 195– 201. Google Scholar CrossRef Search ADS   27 Brandlistuen RE, Ystrom E, Nulman I, Nordeng H. Behavioural effects of fetal antidepressant exposure in a Norwegian cohort of discordant siblings. Int J Epidemiol  2015; 44: 1397– 407. Google Scholar CrossRef Search ADS PubMed  28 Bates JE, Freeland CAB, Lounsbury ML. Measurement of infant difficultness. Child Dev  1979; 50: 794– 803. Google Scholar CrossRef Search ADS PubMed  29 Achenbach TM. Assessment and Taxonomy of Child and Adolescent Psychopathology . London: Sage, 1985. 30 Zachrisson HD, Dearing E, Lekhal R, Toppelberg CO. Little evidence that time in child care causes externalizing problems during early childhood in Norway. Child Dev  2013; 84: 1152– 70. Google Scholar CrossRef Search ADS PubMed  31 Frisell T, Öberg S, Kuja-Halkola R, Sjölander A. Sibling comparison designs: bias from non-shared confounders and measurement error. Epidemiology  2012; 7: 713– 20. Google Scholar CrossRef Search ADS   32 Nilsen MR, Vollset SE, Gjessing HK, et al.   Self-selection and bias in a large prospective pregnancy cohort in Norway. Paediatr Perinat Epidemiol  2009; 23: 597– 608. Google Scholar CrossRef Search ADS PubMed  33 Wolke D, Waylen A, Samara M, et al.   Selective drop-out in longitudinal studies and non-biased prediction of behaviour disorders. Br J Psychiatry  2009; 195: 249– 56. Google Scholar CrossRef Search ADS PubMed  © The Author 2017; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png International Journal of Epidemiology Oxford University Press

Re-examining the link between prenatal maternal anxiety and child emotional difficulties, using a sibling design

Loading next page...
 
/lp/ou_press/re-examining-the-link-between-prenatal-maternal-anxiety-and-child-hqjksb0inb
Publisher
Oxford University Press
Copyright
© The Author 2017; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association
ISSN
0300-5771
eISSN
1464-3685
D.O.I.
10.1093/ije/dyx186
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Prenatal exposure to maternal anxiety has been associated with child emotional difficulties in a number of epidemiological studies. One key concern, however, is that this link is vulnerable to confounding by pleiotropic genes or environmental family factors. Methods Data on 82 383 mothers and children from the population-based Mother and Child Cohort Study and data on 21 980 siblings were used in this study. Mothers filled out questionnaires for each unique pregnancy, for infant difficulties at 6 months and for emotional difficulties at 36 months. The link between prenatal maternal anxiety and child difficulties were examined using logistic regression analyses and multiple linear regression analyses for the full study sample and the sibling sample. Results In the conventional full-cohort analyses, prenatal exposure to maternal anxiety was associated with child difficulties at both 6 months [odds ratio (OR) = 2.1 (1.94–2.27)] and 36 months [OR = 2.72 (2.47–2.99)]. The findings were essentially the same whether we examined difficulties at 6 months or at 36 months. However, these associations were no longer present once we controlled for potential social and genetic confounders in the sibling comparison analyses, either at 6 months [OR = 1.32 (0.91–1.90)] or at 36 months [OR = 1.28 (0.63–2.60)]. Findings from multiple regression analyses with continuous measures were essentially the same. Conclusions Our finding lends little support for there being an independent prenatal effect on child emotional difficulties; rather, our findings suggest that the link between prenatal maternal anxiety and child difficulties could be confounded by pleiotropic genes or environmental family factors. Prenatal, anxiety, sibling design, emotional difficulties Key Messages In the full cohort (n = 82 383), we found that children exposed to maternal anxiety in pregnancy had about twice the risk for short- and long-term difficulties in the child. When the same association was examined within a sibling design (n = 21 980) of this cohort, this association disappeared. The findings from our sibling analyses suggest that the association between prenatal maternal anxiety and emotional difficulties is confounded by genetic or other familial factors. Introduction During the past half-century, animal research and parallel human findings suggest that prenatal exposure to maternal stress and anxiety influence offspring psychopathology.1 Prenatal maternal anxiety has been associated with both short- and long-term mental health problems in the child (i.e. difficulties such as emotional reactivity, symptoms of anxiety, somatic complaints and sleep problems).1–7 This link may be mediated through a programming effect, perhaps through increased production of cortisol, which can influence fetal brain development.8–10 Although findings in animal models have been robust in showing the link between elevated glucocorticoid levels following stress and adverse outcome in the offspring,1,10,11 scepticism has been raised as to the validity of the concepts and methods of several of the findings in human studies. Some interpretations are based on inconsistent results, low sample sizes, poor research designs and lack of appropriate genetic and postnatal controls.12,13 One key concern is that observational studies are vulnerable to confounding by pleiotropic genes.14,15 Rice and colleagues16 used a ‘prenatal cross-fostering’ design to examine the intra-uterine environment where pregnant mothers were either related or unrelated to their child, as a result of in vitro fertilization (IVF). The results suggest that the associations with prenatal stress were caused by either postnatal risks or common genes. However, this finding was limited by the use of retrospective recall of prenatal stress, and small sample size. Another powerful way of examining prenatal risk effects within epidemiological studies is through the use of sibling designs. For example, D’Onofrio and colleagues17 examined the well-established link between smoking during pregnancy and antisocial behaviour in a sibling design where they compared siblings that differed in their exposure to maternal smoking during pregnancy. Findings showed that siblings discordant in exposure to prenatal smoking did not differ in their risk for antisocial behaviour. This finding has been confirmed by other quasi-experimental studies.18–20 The implication from these studies is that relying on measured covariates to account for confounding could result in false-positive conclusions, because of insufficient control for postnatal exposure or pleiotropic genes. A sibling comparison design, where one sibling has been exposed to prenatal maternal anxiety and the other has not, would control for both genetic and environmental factors that are shared by the siblings. Because each child receives a random set of its mother´s genes through the process of meiosis,21 the sibling design provides a good control for pleiotropic genes that influence maternal anxiety and behaviour as well as emotional difficulties in the child.15 Using data from pregnancy questionnaires as well as responses to questions on child behaviour at 6 and 36 months, we aim to examine the risk associated with maternal anxiety in pregnancy on short- (6 months) and long-term (36 months) outcomes in the child, and to control potential social and genetic confounding using a sibling comparison design. Methods Study population and data collection This study is a subproject of the Norwegian Mother and Child Cohort Study (MoBa) conducted by the Norwegian Institute of Public Health.22 MoBa is a cohort comprising 113 000 pregnancies recruited from 1999 to 2008 with a participation rate of 40.6%.22 All but two of a total of 52 hospitals around Norway agreed to participate in the recruitment to the study. Women were invited to participate when they attended the routine ultrasound examination offered to all pregnant women at 17–18 weeks of gestation [www.fhi.no/moba-en]. The mothers filled out questionnaires about characteristics including their age, education, smoking and drinking habits, partner harmony and anxiety at the 17th and 30th gestational week. In addition, the current study used data from the questionnaires about child somatic health, infant difficulties and emotional difficulties at 6 months and 36 months after birth. The cohort was also linked to the Medical Birth Registry of Norway (MBRN).23 The MBRN contains detailed medical information about the infant, including birthweight and birth complications. The maternal questionnaire response rates at the 17th and 30th week of gestation, and at 6 and 36 months after birth, were approximately 95.1%, 91.4%, 87.0% and 61.4%, respectively.24 Written informed consent was obtained from all participating women. The Regional Committees for Medical and Health Research Ethics (REK) and the Norwegian Data Inspectorate have approved the study. Within MoBa, a total of 15 256 mothers participated with more than one pregnancy. Mothers filled out questionnaires for each unique pregnancy and for infant difficulties at 6 months, as well as for emotional difficulties at 36 months. For women participating with three or more pregnancies, two siblings were randomly selected. We used version 7 of the quality-assured data files for participants recruited in the period 1999 to 2009, see flow chart (Figure 1). Figure 1 View largeDownload slide Flow chart of sibling sample. Figure 1 View largeDownload slide Flow chart of sibling sample. Measure of maternal general anxiety Mothers reported on symptoms of anxiety using a validated short version of the Hopkins Symptom Checklist, the SCL-525 and SCL-8.26 The Symptom Checklist scale (SCL) is scored on a Likert scale ranging from 1 to 4. The short scale has been found to be valid with a correlation of 0.92 with the SCL-25.26 Assessments of anxiety were made twice during pregnancy (17th and 30th gestational weeks), and again when the child was 6 months old. Two out of five questions from the SCL-5 and four out of eight questions from the SCL-8 measured anxiety at 17th and 30th gestational weeks and at 6 months, respectively. The mean score for the 17th and 30th gestational weeks and 6 months measure ranged 1–2 and 1–4, respectively. Both continuous scores and a quasi-clinical cut-off (in the absence of an established cut-off, the top 15% of the sample was identified, which is equivalent to other studies, e.g. O’Donnell and colleagues,3 and Brandlistuen and colleagues27), are used in the analyses. Exploring alternative cut-points suggested essentially the same results. Based on these dichotomized variables, a four-level factor variable was constructed to represent ‘no exposure to maternal anxiety’, ‘high anxiety in week 17 only’, ‘high anxiety in week 30 only’ or ‘both’. Main outcome variables: Infant difficulties at 6 months were measured at 6 months postpartum by nine items from the Infant Characteristic Questionnaire (ICQ).28 Mothers responded to seven ordered response categories. The responses were scored 1–7, and the mean score calculated (range 1–7). The categorical variable was based on one standard deviation (SD) above the mean, and coded as low infant difficulties (‘0’), and high infant difficulties (‘1’). That is, the top 15%, which is equivalent to one standard deviation away from the centre, is regarded as high infant difficulties. Emotional difficulties were measured by 10 items from the Child Behaviour Checklist (CBCL/TRF),29 which represented four subscales: emotionally reactive, anxious/depressed, somatic complaints and sleep problems. Mothers reported the extent to which they agreed with the difficulties statements on a three-point Likert scale, from not true (‘0’), sometimes true (‘1’) and often true (‘2’). Mean scores were calculated and ranged 0–2, and the Cronbach`s alpha was 0.56. The categorical variable was based on one SD above the mean, and coded as low emotional difficulties (‘0’), and high emotional difficulties (‘1’). That is, approximately the top 15% is regarded as high emotional difficulties. A team of clinical and developmental psychologists selected the items from the CBCL used in MoBa. The selected items were based on theoretical and empirical representativeness and have been found to be representative with a correlation of 0.92 to the full scale.30 Assessment of potential confounders Potential confounding factors were considered if the variable could be associated with both prenatal maternal anxiety and child outcomes (i.e. causing the two to correlate), and were included in the adjusted model if associated with the exposure (prenatal maternal anxiety) or one of the two outcome measures. The following variables were considered as potential controls: alcohol consumption during pregnancy (coded as never ‘0’ and more than once a month ‘1’); and smoking in pregnancy (coded as never ‘0’, sometimes ‘1’ or daily ‘2’). In addition we controlled for: maternal anxiety reported at 6 months, low = ‘0’ and high = ‘1’; partner (dis)harmony (good partner relationship = ‘0’, and poor partner relationship = ‘1’); somatic disease (not present at 6 months ‘0’ and present ‘1’); marital status (married/living together ‘0’ and single ‘1’); maternal education (coded as higher university degree of +4 years ‘0’, college/university of 3 years ‘1’, college 1–2 years ‘2’, secondary school ‘3’); and maternal age (coded as <25 ‘0’, 25–29 ‘1’, 30–34 ‘2’ and >= 35 ‘3’). We also controlled for the following variables extracted from the MBRN: parity was coded 0 ‘0’, and >=1 ‘1’; gestational age (coded as <37 ‘0’, and >=37 ‘1’; birth complications, coded as yes ‘1’ or no ‘0’; child sex (girl ‘0’, boy ‘1’); and birthweight (coded as >=2500 g = ‘0’, <2500 g = ‘1’). Statistical analyses We present descriptive data on the full sample and the sibling sample, including the distribution on confounders and covariates for infant difficulties at 6 months and emotional difficulties at 36 months. After reporting descriptive data, we present the analyses for the full sample (n = 82 383) and the sibling sample (n = 21 980). We present the results from the logistic regression analyses and multiple linear regression analyses for the full study sample and the sibling sample. We also adjusted for the following control variables: maternal age, educational level, marital status, parity, partner relationship, prenatal alcohol consumption, prenatal smoking, birth complications, child’s gender, birthweight, child’s somatic diseases and child’s gestational age. We also examined the potential mediating effect of the control variables: birth complication, birthweight, gestational age and child somatic disease, in a proportion; however, the mediation effect, was not clearly present. The mediating effect of the four variables was 0.11%, 0.09%, 6.27% and 1.42% respectively. For the sibling data with the binary outcomes, we used a conditional logistic regression model containing a family-specific term in the linear predictor that captures environmental and genetic factors common to the family. When conditioning on the number of siblings in each family with the reported outcome, these nuisance terms vanish, but the remaining terms are unchanged. This gives a conditional logistic regression model and we used the ‘clogit’ function in R to fit the model. Also for the continuous scores, we can include family-specific terms in the linear predictors that will cancel out when conditioning on the total reported scores in each family. Equivalently, as we considered only two siblings in each family, we analysed the differences in the siblings’ scores using differences in sibling covariates as explanatory variables. To handle missing values, we used pairwise deletion of missing. For the full cohort (n = 82 383), the crude model deleted 2145 observations due to missing. In the sibling design (n = 21 980), only 413 were taken out when we have the crude model fitted. The total number of discordant siblings was 1476 at 6 months and 786 at 36 months. Results The characteristics of the full cohort sample are shown in Table 1 and the sibling sub-sample in Table 2. In the sibling sub-sample, the proportion of children with difficulties at 6 months after birth was 20.4% among mothers reporting anxiety at week 17 and 20% among mothers with anxiety at week 30. This is comparable to 22%, which was found in the full cohort. In the sibling sub-sample, approximately 26% of the children exposed to high prenatal maternal anxiety at both 17th and 30th weeks of gestation had emotional difficulties at 36 months. This is somewhat lower than for the full cohort, which was approximately 30% at both 17th and 30th weeks of gestation. Of 10 990 pairs of siblings (n = 21 980), 86.3% and 86.9% turned out to be concordant for not being exposed to maternal anxiety at the 17th and 30th weeks of gestation, respectively. Also, 1.8% (17th week) and 2.5% (30th week) of the sibling pairs were both exposed to anxiety at these two points in time. However, 9.9% and 9.7% differed in exposure for maternal anxiety at the 17th and 30th week, respectively. The proportion of children with high score on child difficulties at both 6 and 36 months was 3.4%. Table 1 Characteristics of the Full Cohort   Infant Difficulties at 6 months   Emotional Difficulties 36 months   N, total  n  %  N, total  n  %  Total  82,383  12,613  15.3  53,917  9,750  18.1  General anxiety week 17   Low  73,028  10,488  14.36  48,431  8,151  16.83   High  8,074  1,852  22.94  4,802  1,446  30.11   Missing  1,281  273  21.31  684  153  22.37  General anxiety week 30   Low  73,230  10,569  14.43  48,536  8,145  16.78   High  8,547  1,893  22.15  5,087  1,528  30.04   Missing  606  151  24.92  294  77  26.19  General anxiety 6 months   Low  75,207  10,801  14.36  49,732  8,404  16.90   High  6,828  1,721  25.21  4,041  1,306  32.32   Missing  348  91  26.15  144  40  27.78  Mother's age   <25  8,405  1,566  18.63  4,833  1,077  22.28   25–29  27,478  4,409  16.05  17,957  3,266  18.19   30–34  32,081  4,643  14.47  21,412  3,673  17.15   >35  14,314  1,980  13.83  9,646  1,721  17.84   Missing  105  15  14.29  69  13  18.84  Education   University 4+  18,828  2,803  14.89  13,191  2,232  16.92   College/University 3y  32,948  4,756  14.43  22,525  3,701  16.43   College 1–2y  11,316  1,738  15.36  6,938  1,386  19.98   Secondary school  15,253  2,647  17.35  8,790  1,900  21.62   Missing  4,038  669  16.57  2,473  531  21.47  Marital status   Married/partners  79,325  12,050  15.19  52,069  9,311  17.88   Single  2,953  548  18.56  1,779  426  23.95   Missing  105  15  14.29  69  13  18.84  Partner harmony   High  69,814  10,115  14.49  46,386  7,964  17.17   Low  7,990  1,647  20.61  4,910  1,239  25.23   Missing  4,579  851  18.58  2,621  547  20.87  Alcohol in pregnancy   Never  67,959  10,167  14.96  45,772  8,140  17.78   >1 a month  480  82  17.08  341  93  27.27   Missing  13,944  2,364  16.95  7,804  1,517  19.44  Smoking status   Never  74,456  11,341  15.23  49,406  8,744  17.70   Sometimes  3,502  557  15.91  2,022  439  21.71   Daily  3,146  483  15.35  1,716  408  23.78   Missing  1,279  232  18.14  773  159  20.57  Birth complication   No  59,757  9,327  15.61  39,589  7,376  18.63   Yes  22,521  3,271  14.52  14,259  2,361  16.56   Missing  105  15  14.29  69  13  18.84  Parity   0(=0)  37,734  6,399  16.96  25,788  5,347  20.73   >=1(=1)  44,544  6,199  13.92  28,060  4,390  15.65   Missing  105  15  14.29  69  13  18.84  Child sex   Girls  40,170  5,720  14.24  26,317  4,902  18.63   Boys  42,108  6878  16.33  27,531  4835  17.56   Missing  105  15  14.29  69  13  18.84  Birth weight   >2500  80,245  12,168  15.16  52,468  9,412  17.94   <2500  1,989  422  21.22  1,358  321  23.64   Missing  149  23  15.44  91  17  18.68  Somatic disease 6m   No  56,759  7,100  12.51  39,115  6,610  16.90  vYes  17,590  4,420  25.13  11,271  2,451  21.75   Missing  8,034  1,093  13.60  3,531  689  19.51  Gestational age  v<37  3,533  681  19.28  2,412  515  21.35  v>=37  78,420  11,858  15.12  51,219  9,184  17.93  vMissing  430  74  17.21  286  51  17.83    Infant Difficulties at 6 months   Emotional Difficulties 36 months   N, total  n  %  N, total  n  %  Total  82,383  12,613  15.3  53,917  9,750  18.1  General anxiety week 17   Low  73,028  10,488  14.36  48,431  8,151  16.83   High  8,074  1,852  22.94  4,802  1,446  30.11   Missing  1,281  273  21.31  684  153  22.37  General anxiety week 30   Low  73,230  10,569  14.43  48,536  8,145  16.78   High  8,547  1,893  22.15  5,087  1,528  30.04   Missing  606  151  24.92  294  77  26.19  General anxiety 6 months   Low  75,207  10,801  14.36  49,732  8,404  16.90   High  6,828  1,721  25.21  4,041  1,306  32.32   Missing  348  91  26.15  144  40  27.78  Mother's age   <25  8,405  1,566  18.63  4,833  1,077  22.28   25–29  27,478  4,409  16.05  17,957  3,266  18.19   30–34  32,081  4,643  14.47  21,412  3,673  17.15   >35  14,314  1,980  13.83  9,646  1,721  17.84   Missing  105  15  14.29  69  13  18.84  Education   University 4+  18,828  2,803  14.89  13,191  2,232  16.92   College/University 3y  32,948  4,756  14.43  22,525  3,701  16.43   College 1–2y  11,316  1,738  15.36  6,938  1,386  19.98   Secondary school  15,253  2,647  17.35  8,790  1,900  21.62   Missing  4,038  669  16.57  2,473  531  21.47  Marital status   Married/partners  79,325  12,050  15.19  52,069  9,311  17.88   Single  2,953  548  18.56  1,779  426  23.95   Missing  105  15  14.29  69  13  18.84  Partner harmony   High  69,814  10,115  14.49  46,386  7,964  17.17   Low  7,990  1,647  20.61  4,910  1,239  25.23   Missing  4,579  851  18.58  2,621  547  20.87  Alcohol in pregnancy   Never  67,959  10,167  14.96  45,772  8,140  17.78   >1 a month  480  82  17.08  341  93  27.27   Missing  13,944  2,364  16.95  7,804  1,517  19.44  Smoking status   Never  74,456  11,341  15.23  49,406  8,744  17.70   Sometimes  3,502  557  15.91  2,022  439  21.71   Daily  3,146  483  15.35  1,716  408  23.78   Missing  1,279  232  18.14  773  159  20.57  Birth complication   No  59,757  9,327  15.61  39,589  7,376  18.63   Yes  22,521  3,271  14.52  14,259  2,361  16.56   Missing  105  15  14.29  69  13  18.84  Parity   0(=0)  37,734  6,399  16.96  25,788  5,347  20.73   >=1(=1)  44,544  6,199  13.92  28,060  4,390  15.65   Missing  105  15  14.29  69  13  18.84  Child sex   Girls  40,170  5,720  14.24  26,317  4,902  18.63   Boys  42,108  6878  16.33  27,531  4835  17.56   Missing  105  15  14.29  69  13  18.84  Birth weight   >2500  80,245  12,168  15.16  52,468  9,412  17.94   <2500  1,989  422  21.22  1,358  321  23.64   Missing  149  23  15.44  91  17  18.68  Somatic disease 6m   No  56,759  7,100  12.51  39,115  6,610  16.90  vYes  17,590  4,420  25.13  11,271  2,451  21.75   Missing  8,034  1,093  13.60  3,531  689  19.51  Gestational age  v<37  3,533  681  19.28  2,412  515  21.35  v>=37  78,420  11,858  15.12  51,219  9,184  17.93  vMissing  430  74  17.21  286  51  17.83  Note: n and % corresponds to infant difficulties and emotional difficulties. Table 2 Characteristics of the Siblings Sub-sample   Infant Difficulties at 6 months   Emotional Difficulties 36 months   N, total  n  %  N, total  n  %  Total  21,980  3,067  14.0  13,252  1,926  14.5  General anxiety week 17   Low  20,241  2,721  13.44  12,385  1,705  13.77   High  1,505  307  20.40  756  198  26.19   Missing  234  39  16.67  111  23  20.72  General anxiety week 30   Low  20,248  2,719  13.43  12,372  1,698  13.72   High  1,628  326  20.02  830  218  26.27   Missing  104  22  21.15  50  10  20.00  General anxiety 6 months   Low  20,546  2,743  13.35  12,497  1,706  13.65   High  1,374  309  22.49  734  210  28.61   Missing  60  15  25.00  21  10  47.62  Mother's age   <25  1,889  310  16.41  944  168  17.80   25–29  7,828  1,120  14.31  4,683  680  14.52   30–34  9,122  1,221  13.39  5,673  811  14.30   >35  3,111  409  13.15  1,933  262  13.55   Missing  30  7  23.33  19  5  26.32  Education   University 4+  5,781  790  13.67  3,711  530  14.28   College/University 3y  9,837  1,297  13.18  6,087  827  13.59   College 1–2y  2,504  370  14.78  1,411  237  16.80   Secondary school  2,967  476  16.04  1,551  242  15.60   Missing  891  134  15.04  492  90  18.29  Marital status   Married/partners  21,543  2,997  13.91  13,031  1,896  14.55   Single  407  63  15.48  202  25  12.38   Missing  30  7  23.33  19  5  26.32  Partner harmony   High  19,213  2,559  13.32  11,713  1,637  13.98   Low  1,905  377  19.79  1,083  223  20.59   Missing  862  131  15.20  456  66  14.47  Alcohol in pregnancy   Never  18,580  2,553  13.74  11,447  1,650  14.41   >1 a month  89  9  10.11  55  11  20.00   Missing  3,311  505  15.25  1,750  265  15.14  Smoking status   Never  20,571  2,855  13.88  12,535  1,815  14.48   Sometimes  601  84  13.98  306  50  16.34   Daily  483  82  16.98  225  38  16.89   Missing  325  46  14.15  186  23  12.37  Birth complication   No  15,369  2,209  14.37  9,350  1,410  15.08   Yes  6,581  851  12.93  3,883  511  13.16   Missing  30  7  23.33  19  5  26.32  Parity   0(=0)  8,295  1,289  15.54  5,049  889  17.61   >=1(=1)  13,655  1,771  12.97  8,184  1,032  12.61   Missing  30  7  23.33  19  5  26.32  Child sex   Girls  10,707  1,384  12.93  6,407  989  15.44   Boys  11,243  1676  14.91  6,826  932  13.65   Missing  30  7  23.33  19  5  26.32  Birth weight   >2500  21,505  2,968  13.80  12,964  1,873  14.45   <2500  433  90  20.79  263  48  18.25   Missing  42  9  21.43  25  5  20.00  Somatic disease 6M   No  15,393  1,758  11.42  9,732  1,323  13.59   Yes  4,751  1,090  22.94  2,813  489  17.38   Missing  1,836  219  11.93  707  114  16.12  Gestational age   <37  816  148  18.14  499  81  16.23   >=37  21,065  2,900  13.77  12,698  1,837  14.47   Missing  99  19  19.19  55  8  14.55    Infant Difficulties at 6 months   Emotional Difficulties 36 months   N, total  n  %  N, total  n  %  Total  21,980  3,067  14.0  13,252  1,926  14.5  General anxiety week 17   Low  20,241  2,721  13.44  12,385  1,705  13.77   High  1,505  307  20.40  756  198  26.19   Missing  234  39  16.67  111  23  20.72  General anxiety week 30   Low  20,248  2,719  13.43  12,372  1,698  13.72   High  1,628  326  20.02  830  218  26.27   Missing  104  22  21.15  50  10  20.00  General anxiety 6 months   Low  20,546  2,743  13.35  12,497  1,706  13.65   High  1,374  309  22.49  734  210  28.61   Missing  60  15  25.00  21  10  47.62  Mother's age   <25  1,889  310  16.41  944  168  17.80   25–29  7,828  1,120  14.31  4,683  680  14.52   30–34  9,122  1,221  13.39  5,673  811  14.30   >35  3,111  409  13.15  1,933  262  13.55   Missing  30  7  23.33  19  5  26.32  Education   University 4+  5,781  790  13.67  3,711  530  14.28   College/University 3y  9,837  1,297  13.18  6,087  827  13.59   College 1–2y  2,504  370  14.78  1,411  237  16.80   Secondary school  2,967  476  16.04  1,551  242  15.60   Missing  891  134  15.04  492  90  18.29  Marital status   Married/partners  21,543  2,997  13.91  13,031  1,896  14.55   Single  407  63  15.48  202  25  12.38   Missing  30  7  23.33  19  5  26.32  Partner harmony   High  19,213  2,559  13.32  11,713  1,637  13.98   Low  1,905  377  19.79  1,083  223  20.59   Missing  862  131  15.20  456  66  14.47  Alcohol in pregnancy   Never  18,580  2,553  13.74  11,447  1,650  14.41   >1 a month  89  9  10.11  55  11  20.00   Missing  3,311  505  15.25  1,750  265  15.14  Smoking status   Never  20,571  2,855  13.88  12,535  1,815  14.48   Sometimes  601  84  13.98  306  50  16.34   Daily  483  82  16.98  225  38  16.89   Missing  325  46  14.15  186  23  12.37  Birth complication   No  15,369  2,209  14.37  9,350  1,410  15.08   Yes  6,581  851  12.93  3,883  511  13.16   Missing  30  7  23.33  19  5  26.32  Parity   0(=0)  8,295  1,289  15.54  5,049  889  17.61   >=1(=1)  13,655  1,771  12.97  8,184  1,032  12.61   Missing  30  7  23.33  19  5  26.32  Child sex   Girls  10,707  1,384  12.93  6,407  989  15.44   Boys  11,243  1676  14.91  6,826  932  13.65   Missing  30  7  23.33  19  5  26.32  Birth weight   >2500  21,505  2,968  13.80  12,964  1,873  14.45   <2500  433  90  20.79  263  48  18.25   Missing  42  9  21.43  25  5  20.00  Somatic disease 6M   No  15,393  1,758  11.42  9,732  1,323  13.59   Yes  4,751  1,090  22.94  2,813  489  17.38   Missing  1,836  219  11.93  707  114  16.12  Gestational age   <37  816  148  18.14  499  81  16.23   >=37  21,065  2,900  13.77  12,698  1,837  14.47   Missing  99  19  19.19  55  8  14.55  Note: n and % corresponds to infant difficulties and emotional difficulties. Full cohort analyses In the logistic regression analyses on the full cohort, we found that children exposed to maternal anxiety at both 17th and 30th week of gestation had a higher risk of infant difficulties at 6 months and emotional difficulties at 36 months, as compared with those who had not been exposed to maternal anxiety (Table 3). These associations remained, but were somewhat reduced in size after controlling for a number of potential confounders, including maternal anxiety measured at 6 months. Examination of different cut-points did not indicate fundamentally different findings, e.g. 90th percentile suggested a positive change in OR of 0.32 at 6 months and 0.58 at 36 months. Table 3 The Effect of Prenatal Anxiety on Infant Difficulties 6 months and Emotional Difficulties 36 months in Full cohort   Infant Difficulties 6 months  Emotional Difficulties 36 months  Logistic regression  Crude  Adjusted*  Crude  Adjusted**    OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  Sample size(=n)  (n=80,238)  (n=55,065)  (n=52,608)  (n=37,982)    No Anxiety (reference)  1  1  1  1  17th week only  1.63 (1.51, 1.76)  1.37 (1.24, 1.51)  1.91 (1.75, 2.09)  1.47 (1.32, 1.65)  30th week only  1.51 (1.4, 1.63)  1.21 (1.1, 1.34)  1.96 (1.8, 2.14)  1.5 (1.34, 1.67)  Both 17th and 30th week  2.09 (1.94, 2.27)  1.38 (1.23, 1.55)  2.72 (2.47, 2.99)  1.84 (1.62, 2.1)    Multiple linear regression  Crude  Adjusted*  Crude  Adjusted**    Beta (95% CI)  Beta (95% CI)  Beta (95% CI)  Beta (95% CI)    General Anxiety          17th week only  0.28 (0.24, 0.31)  0.16 (0.12, 0.2)  0.12 (0.1, 0.13)  0.06 (0.05, 0.08)  30th week only  0.34 (0.3, 0.37)  0.16 (0.12, 0.21)  0.16 (0.14, 0.17)  0.09 (0.07, 0.1)  Both 17th and 30th week  0.5 (0.46, 0.54)  0.28 (0.19, 0.36)  0.23 (0.21, 0.24)  0.16 (0.12, 0.19)    Infant Difficulties 6 months  Emotional Difficulties 36 months  Logistic regression  Crude  Adjusted*  Crude  Adjusted**    OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  Sample size(=n)  (n=80,238)  (n=55,065)  (n=52,608)  (n=37,982)    No Anxiety (reference)  1  1  1  1  17th week only  1.63 (1.51, 1.76)  1.37 (1.24, 1.51)  1.91 (1.75, 2.09)  1.47 (1.32, 1.65)  30th week only  1.51 (1.4, 1.63)  1.21 (1.1, 1.34)  1.96 (1.8, 2.14)  1.5 (1.34, 1.67)  Both 17th and 30th week  2.09 (1.94, 2.27)  1.38 (1.23, 1.55)  2.72 (2.47, 2.99)  1.84 (1.62, 2.1)    Multiple linear regression  Crude  Adjusted*  Crude  Adjusted**    Beta (95% CI)  Beta (95% CI)  Beta (95% CI)  Beta (95% CI)    General Anxiety          17th week only  0.28 (0.24, 0.31)  0.16 (0.12, 0.2)  0.12 (0.1, 0.13)  0.06 (0.05, 0.08)  30th week only  0.34 (0.3, 0.37)  0.16 (0.12, 0.21)  0.16 (0.14, 0.17)  0.09 (0.07, 0.1)  Both 17th and 30th week  0.5 (0.46, 0.54)  0.28 (0.19, 0.36)  0.23 (0.21, 0.24)  0.16 (0.12, 0.19)  *Adjusted for maternal anxiety 6 months, mother’s age, education, marital status, parity, partner relation, alcohol intake, smoking status, birth complication, child’s sex, birth weight, somatic disease, gestational age. **Also adjusted for infant difficulties at 6 months. Bold are significant at p<0.001. We also examined these associations using multiple regression analyses with continuous measures for the full sample. The results showed moderate associations with infant difficulties at 6 months, as well as emotional difficulties at 36 months. The follow-up analyses at 36 months showed that there was no difference in effects, when we examined the associations between prenatal maternal anxiety on infant difficulties at 6 months and emotional difficulties at 36 months, suggesting that some overall liability is involved. Sibling comparison analyses We examined the same associations in a sibling comparison design, to control for potential social and genetic confounders. In the crude and adjusted sibling analyses for the conditional logistic regression, no associations were found between prenatal maternal anxieties, measured at either the 17th or 30th week of gestation or both, on infant difficulties at 6 months or emotional difficulties at 36 months of age (Table 4). Table 4 The Effect of Prenatal Anxiety on Infant Difficulties 6 months and Emotional Difficulties 36 months in sibling sub-sample (full siblings)   Infant Difficulties at 6 months  Emotional Difficulties at 36 months  Conditional logistic regression  Crude  Adjusted*  Crude  Adjusted**  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  Sample size(=n)  (n=21,571)  (n=15,664)  (n=13,011)  (n=9,844)    No Anxiety (reference)  1  1  1  1  17th week only  1.12 (0.84, 1.49)  0.92 (0.61, 1.4)  1.33 (0.9, 1.96)  0.78 (0.46, 1.33)  30th week only  1.15 (0.86, 1.53)  0.91 (0.59, 1.4)  0.99 (0.69, 1.44)  1.04 (0.6, 1.77)  Both 17th and 30th week***  1.3 (0.9, 1.87)  1.01 (0.57, 1.79)  1.07 (0.64, 1.78)  1.13 (0.54, 2.36)    Multiple linear regression  Crude Beta (95% CI)  Adjusted*  Beta (95% CI)  Crude Beta (95% CI)  Adjusted**  Beta (95% CI)  Sample size(=n)  (n=10,594)  (n=5,935)  (n=6,393)  (n=3,877)    General Anxiety          17th week only  0.047 (−0.001, 0.095)  0.033 (−0.031, 0.096)  0.005 (−0.013, 0.023)  0.001 (−0.022, 0.025)  30th week only  0.04 (−0.022, 0.102)  0.02 (−0.061, 0.101)  −0.011 (−0.034, 0.012)  −0.024 (−0.055, 0.006)  Both 17th and 30th week***  0.065 (−0.046, 0.175)  −0.03 (−0.174, 0.115)  0.015 (−0.027, 0.058)  −0.002 (−0.054, 0.051)    Infant Difficulties at 6 months  Emotional Difficulties at 36 months  Conditional logistic regression  Crude  Adjusted*  Crude  Adjusted**  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  Sample size(=n)  (n=21,571)  (n=15,664)  (n=13,011)  (n=9,844)    No Anxiety (reference)  1  1  1  1  17th week only  1.12 (0.84, 1.49)  0.92 (0.61, 1.4)  1.33 (0.9, 1.96)  0.78 (0.46, 1.33)  30th week only  1.15 (0.86, 1.53)  0.91 (0.59, 1.4)  0.99 (0.69, 1.44)  1.04 (0.6, 1.77)  Both 17th and 30th week***  1.3 (0.9, 1.87)  1.01 (0.57, 1.79)  1.07 (0.64, 1.78)  1.13 (0.54, 2.36)    Multiple linear regression  Crude Beta (95% CI)  Adjusted*  Beta (95% CI)  Crude Beta (95% CI)  Adjusted**  Beta (95% CI)  Sample size(=n)  (n=10,594)  (n=5,935)  (n=6,393)  (n=3,877)    General Anxiety          17th week only  0.047 (−0.001, 0.095)  0.033 (−0.031, 0.096)  0.005 (−0.013, 0.023)  0.001 (−0.022, 0.025)  30th week only  0.04 (−0.022, 0.102)  0.02 (−0.061, 0.101)  −0.011 (−0.034, 0.012)  −0.024 (−0.055, 0.006)  Both 17th and 30th week***  0.065 (−0.046, 0.175)  −0.03 (−0.174, 0.115)  0.015 (−0.027, 0.058)  −0.002 (−0.054, 0.051)  *Adjusted for maternal anxiety 6 months, mother’s age, education, marital status, parity, partner relation, alcohol intake, smoking status, birth complication, child’s sex, birth weight, somatic disease, gestational age. **Also adjusted for Infant Difficulties at 6 months. *** β^17th week+β^30th week+β^17th weekβ^30th week. We also examined these associations using multiple regression analyses (Table 4). In contrast to what was found in the full cohort, there were no associations between prenatal maternal anxiety and either infant difficulties at 6 months or emotional difficulties at 36 months. However, a moderate association with maternal anxiety reported at 6 months postnatally remained. Discussion Using a large population-based longitudinal cohort study we assessed the link between prenatal maternal anxiety and child emotional difficulties using a sibling design. There are two key findings that stand out. First, the conventional full-cohort analyses replicated common findings,9 but the same analyses, controlled for pleiotropic genes or environmental family factors, did not. Second, the findings were essentially the same whether we examined infant difficulties at 6 months or emotional difficulties at 36 months. In the first analyses using the full cohort, children prenatally exposed to maternal anxiety had twice the risk of infant difficulties at 6 months, and almost three times the risk of emotional difficulties at 36 months. This association did not change substantially after controlling for multiple covariates. This is consistent with a number of other studies3,6,9 and is almost an exact replicate of the recent findings by O’Donnell et al.3 They found that a 2-fold increase in risk of a probable child mental disorder was associated with exposure to prenatal maternal anxiety.3 This is equivalent to our findings from the full-cohort analyses (both logistic and multiple regression). However, once the same analyses were conducted using a sibling design, these associations were no longer found either for infant difficulties at 6 months or emotional difficulties measured at 36 months. This suggests that the substantial findings on the full cohort are likely to be confounded by pleiotropic genetic or constant environmental family factors. Although the use of sibling design of discordant siblings reflects a reduction in sample size, a major strength of our study is the consistency in the findings across 6 and 36 months outcomes, as well as across both logistic regression and multiple regression analyses. Our findings further suggest little support for there being an independent prenatal effect on child difficulties and suggest that statistical control for confounders does not deal adequately with the issues, but that the sibling design comparing exposed and non-exposed siblings does. However, although the within-pair estimates will not be confounded by factors that are shared by the siblings, bias could still occur due to non-shared factors.31 Therefore, several covariates (i.e. birth order, maternal age etc.) were adjusted for, without significant changes to the results. However, although our findings lend little support for there being an independent prenatal effect on child emotional difficulties, there is still the possibility that there may be a prenatal effect on other child outcomes. To our knowledge this is the first study to examine the link between prenatal maternal anxiety and child difficulties at both 6 months and 36 months of age in siblings that differ in exposure to prenatal maternal anxiety. Because we cannot randomize pregnant mothers to anxiety exposure and because twins will always be concordant in prenatal risk exposure, the sibling-comparison design is optimal when examining prenatal risk effects and offers a powerful quasi-experimental approach to study prenatal risks. There are, however, limitations that need to be mentioned. The sample size in MoBa limits the possibility of providing clinical interviews in assessing anxiety in the mothers, or clinical diagnosis of difficulties, such as anxiety, in children. However, this study used validated questionnaires that, although they are,short-scales, have been found to correlate highly with the original large-scale questionnaires.25,26,30 Bias could still occur due to attrition or because of halo effects. Maternal reports of infant difficulties and emotional difficulties are subjective and therefore their levels of stress during the pre- and postnatal periods could affect their perceptions of child difficulties. That is, anxious mothers may report their child´s difficulties as more negative than non-anxious mothers. However, using the same informant–the mother–is unavoidable when investigating behaviour in very young children within large-scale population cohort studies, and is a common feature for similar studies (for example, the Avon Longitudinal Study of Parents and Children; the Quebec Longitudinal Study of Child Development). Bias due to selective recruitment is another possible limitation regarding prevalence, but has minimal influence on associations,22,32,33 which was the focus in this study. In sum, our finding suggests that statistical control for confounders does not deal adequately with issues of confounding, but that a sibling design comparing exposed and non-exposed siblings does.15 Funding The Norwegian Mother and Child Cohort Study is supported by the Norwegian Ministry of Health and Care Services and the Ministry of Education and Research, NIH/NIEHS (contract no N01-ES-75558), NIH/NINDS (grant no.1 UO1 NS 047537–01 and grant no.2 UO1 NS 047537–06A1). Acknowledgement We are grateful to all the participating families in Norway who take part in this ongoing cohort study. Conflict of interest: None declared. References 1 Glover V, O’Connor TG, O’Donnell KJ. Prenatal stress and the programming of the HPA axis. Neurosci Biobehav Rev  2010; 35: 17– 22. Google Scholar CrossRef Search ADS PubMed  2 Davis EP, Snidman N, Wadhwa PD, Glynn LM, Schetter CD, Sandman CA. Prenatal maternal anxiety and depression predict negative behavioral reactivity in infancy. Infancy  2004; 6: 319– 31. Google Scholar CrossRef Search ADS   3 O’Donnell KJ, Glover V, Barker ED, O’Connor TG. The persisting effect of maternal mood in pregnancy on childhood psychopathology. Dev Psychopath  2014; 26: 393– 403. Google Scholar CrossRef Search ADS   4 DiPietro JA, Novak MFSX, Costigan KA, Atella LD, Reusing SP. Maternal psychological distress during pregnancy in relation to child development at age two. Child Dev  2006; 77: 573– 87. Google Scholar CrossRef Search ADS PubMed  5 McGrath JM, Records K, Rice M. Maternal depression and infant characteristics. Infant Behav Dev  2008; 31: 71– 80. Google Scholar CrossRef Search ADS PubMed  6 Kingsbury M, Weeks M, MacKinnon N, et al.   Stressful life events during pregnancy and offspring depression: evidence from a prospective cohort study. J Am Acad Child Adolesc Psychiatry  2016; 55: 709– 16. Google Scholar CrossRef Search ADS PubMed  7 Talge NM, Neal C, Glover V. Antenatal maternal stress and long-term effects on child neurodevelopment: How and Why? J Child Psychol Psychiatry  2007; 48: 245– 62. Google Scholar CrossRef Search ADS PubMed  8 Austin M-P, Leader LR, Reilly N. Prenatal stress, the hypothalamic-pituitary-adrenal axis, and fetal and infant neurobehavior. Early Hum Dev  2005; 81: 917– 26. Google Scholar CrossRef Search ADS PubMed  9 Glover V, O’Connor TG. Effects of antenatal stress and anxiety. Implications for development and psychiatry. Br J Psychiatry  2002; 180: 389– 91. Google Scholar CrossRef Search ADS PubMed  10 Welberg LA, Seckl JR. Prenatal stress, glucocorticoids and the programming of the brain. J Neuroendocrinol  2001; 13: 113– 28. Google Scholar CrossRef Search ADS PubMed  11 Weinstock M. Alterations induced by gestational stress in brain morphology and behaviour of the offspring. Brain Behav Immun  2001; 19: 296– 308. Google Scholar CrossRef Search ADS   12 Thapar A, Rutter M. Do prenatal risk factors cause psychiatric disorder? Be wary of causal claims. Br J Psychiatry  2009; 195: 100– 01. Google Scholar CrossRef Search ADS PubMed  13 Rutter M, Pickles A. Annual research review: threats to the validity of child psychiatry and psychology. J Child Psychol Psychiatry  2016; 57: 398– 416. Google Scholar CrossRef Search ADS PubMed  14 Rutter M. Annual research review: resilience: clinical implications. J Child Psychol Psychiatry  2012; 4: 474– 87. 15 D’Onofrio BM, Lahey BB, Turkheimer E, Lichtenstein P. Critical need for family-based, quasi-experimental designs in integrating genetic and social science research. Am J Public Health  2013; 103: 46– 55. Google Scholar CrossRef Search ADS   16 Rice F, Harold GT, Boivin J, van den Bree M, Hay DF, Thapar A. The links between prenatal stress and offspring development and psychopathology: disentangling environmental and inherited influences. Psych Med  2010; 40: 335– 45. Google Scholar CrossRef Search ADS   17 D’Onofrio BM, van Hulle CA, Waldman ID, et al.   Smoking during pregnancy and offspring externalizing problems: An exploration of genetic and environmental confounds. Dev Psychopath  2008; 20: 139– 64. 18 Obel C, Linnet KM, Henriksen TB, et al.   Smoking during pregnancy and hyperactivity-innattention in the offspring – comparing results from three Nordic cohorts. Int J Epidemiol  2009; 38: 698– 705. Google Scholar CrossRef Search ADS PubMed  19 Obel C, Zhu JL, Olsen J, et al.   The risk of attention deficit hyperactivity disorder in children exposed to maternal smoking during pregnancy – a re-examination using a sibling design. J Child Psychol Psychiatry  2016; 57: 532– 37. Google Scholar CrossRef Search ADS PubMed  20 Rice F, Harold GT, Boivin J, Hay DF, van den Bree M, Thapar A. Disentangling prenatal and inherited influences in humans with an experimental design. Proc Natl Acad Sci U S A  2009; 106: 2464– 67. Google Scholar CrossRef Search ADS PubMed  21 Rutter M. Proceeding from observed correlation to causal inference: The use of natural experiments. Perspect Psychol Sc  2007; 2: 377– 95. Google Scholar CrossRef Search ADS   22 Magnus P, Birke C, Vejrup K, et al.   Cohort Profile Update: The Norwegian Mother and Child Cohort Study (MoBa). Int J Epidemiol  2016; 45: 382– 88. Google Scholar CrossRef Search ADS PubMed  23 Irgens LM. The Medical Birth Registry of Norway. Epidemiological research and surveillance throughout 30 years. Acta Obstet Gynaecol Scand  2000; 79: 435– 39. Google Scholar CrossRef Search ADS   24 Schreuder P, Alsaker E. The Norwegian Mother and Child Cohort Study (MoBa) – MoBa recruitment and logistics. Norsk Epidemiol  2014; 24: 23– 27. Google Scholar CrossRef Search ADS   25 Tambs K, Moum T. How well can a few questionnaire items indicate anxiety and depression? Acta Psychiatr Scand  1993; 87: 364– 67. Google Scholar CrossRef Search ADS PubMed  26 Tambs K, Roysamb E. Selection of questions to short-form versions of original psychometric instruments in MoBa. Norsk Epidemiol  2014; 24: 195– 201. Google Scholar CrossRef Search ADS   27 Brandlistuen RE, Ystrom E, Nulman I, Nordeng H. Behavioural effects of fetal antidepressant exposure in a Norwegian cohort of discordant siblings. Int J Epidemiol  2015; 44: 1397– 407. Google Scholar CrossRef Search ADS PubMed  28 Bates JE, Freeland CAB, Lounsbury ML. Measurement of infant difficultness. Child Dev  1979; 50: 794– 803. Google Scholar CrossRef Search ADS PubMed  29 Achenbach TM. Assessment and Taxonomy of Child and Adolescent Psychopathology . London: Sage, 1985. 30 Zachrisson HD, Dearing E, Lekhal R, Toppelberg CO. Little evidence that time in child care causes externalizing problems during early childhood in Norway. Child Dev  2013; 84: 1152– 70. Google Scholar CrossRef Search ADS PubMed  31 Frisell T, Öberg S, Kuja-Halkola R, Sjölander A. Sibling comparison designs: bias from non-shared confounders and measurement error. Epidemiology  2012; 7: 713– 20. Google Scholar CrossRef Search ADS   32 Nilsen MR, Vollset SE, Gjessing HK, et al.   Self-selection and bias in a large prospective pregnancy cohort in Norway. Paediatr Perinat Epidemiol  2009; 23: 597– 608. Google Scholar CrossRef Search ADS PubMed  33 Wolke D, Waylen A, Samara M, et al.   Selective drop-out in longitudinal studies and non-biased prediction of behaviour disorders. Br J Psychiatry  2009; 195: 249– 56. Google Scholar CrossRef Search ADS PubMed  © The Author 2017; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association

Journal

International Journal of EpidemiologyOxford University Press

Published: Feb 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Unlimited reading

Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.

Stay up to date

Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.

Organize your research

It’s easy to organize your research with our built-in tools.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve Freelancer

DeepDyve Pro

Price
FREE
$49/month

$360/year
Save searches from
Google Scholar,
PubMed
Create lists to
organize your research
Export lists, citations
Read DeepDyve articles
Abstract access only
Unlimited access to over
18 million full-text articles
Print
20 pages/month
PDF Discount
20% off