Maternal anxiety in late pregnancy: effect on fetal movements and fetal heart rate

Early Human Development 67 (2002) 87 – 100 www.elsevier.com/locate/earlhumdev

Maternal anxiety in late pregnancy: effect on fetal movements and fetal heart rate Karin Sjo¨stro¨m a,*, Lil Valentin a, Thomas Thelin b, Karel Marsˇa´l a a

Department of Obstetrics and Gynecology, University Hospital, Malmo¨, S-205 02 Malmo¨, Sweden b Department of Psychiatry, University Hospital, Malmo¨, University of Lund, Lund, Sweden

Received 28 June 2001; received in revised form 7 December 2001; accepted 13 December 2001

Abstract Aim: To determine whether maternal state and trait anxiety levels affect fetal movements or fetal heart rate (FHR) in the third trimester. Subjects: Forty-one healthy pregnant nulliparous women not on medication and with a singleton pregnancy. Study design: Maternal anxiety was assessed using the Spielberger State- Trait Anxiety Inventory (Form Y) at 36 gestational weeks. The fetuses of the women were examined at 37 – 40 gestational weeks with ultrasound observation of fetal movements and cardiotocography (CTG). The results of the fetal examinations were compared between women with low and high anxiety scores (low scores being defined as scores below the median and high scores as scores equal to or above the median of the study population), and correlation analyses between anxiety scores and the outcome variables were performed. Outcome measures: The presence and duration (expressed as a percentage of the total examination time) of FHR patterns A, B, C, and D, the percentage duration of fetal movements in each FHR pattern, baseline FHR and FHR variability in each FHR pattern. Results: The presence of FHR patterns A, B, C, and D, the duration of FHR patterns A, B, and C, FHR variability in FHR patterns A, B, and C, baseline FHR and the percentage duration of fetal movements in each FHR pattern did not differ between women with low and high state and trait anxiety scores. In fetuses with FHR pattern D, the duration of FHR pattern D increased with increasing maternal trait anxiety scores, (rho = 0.88; p = 0.008), and FHR variability in FHR pattern D increased with maternal state and trait anxiety scores (r = 0.86, p = 0.01; r = 0.96, p = 0.001). Conclusion: Maternal anxiety does not seem to affect fetal movements or baseline FHR in late pregnancy, but there is a possible association between maternal anxiety and the duration of FHR pattern D and FHR variability in FHR pattern D. D 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Fetal heart rate; Fetal movements; Fetal heart rate variability; Maternal anxiety; State – Trait Anxiety Inventory (STAI)

*

Corresponding author. Tel.: +46-40-33-21-55; fax: +46-40-96-26-00. E-mail address: [email protected] (K. Sjo¨stro¨m).

0378-3782/02/$ - see front matter D 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 3 7 8 - 3 7 8 2 ( 0 1 ) 0 0 2 5 6 - 0

88

K. Sjo¨stro¨m et al. / Early Human Development 67 (2002) 87–100

1. Introduction In recent years, a growing interest in the association between maternal psychological well being during pregnancy and fetal welfare has developed. Anxiety is known to induce bodily changes in the mother, and the effect of maternal anxiety during pregnancy on pregnancy outcome has been studied [1]. The introduction of ultrasound technique into obstetrics has given us an opportunity to examine the human fetus non-invasively and has facilitated the study of effects of various maternal psychological conditions on the fetus. Thus, we have found an association between maternal trait anxiety and fetal cerebral and umbilical artery hemodynamics [2]. Other researchers have later shown a relationship between maternal state and trait anxiety and maternal uterine artery pulsatility index in late pregnancy [3]. A clinically interesting question is whether maternal anxiety might affect fetal behaviour recorded as fetal motor activity or fetal heart rate (FHR). Physiologically, periods with many fetal movements and high FHR variability (active states) alternate with periods with few fetal movements and low FHR variability (quiet states) [4]. Fetal behavioural states (state 1F, 2F, 3F, and 4F)— i.e., organised states of defined FHR patterns, body movements and eye movements are usually fully developed at 36 –38 gestational weeks [5]. Effects of the mother’s somatic or psychological condition on these states can be studied. The aim of this work was to examine whether maternal anxiety might affect fetal active and quiet states, baseline FHR, FHR variability, or fetal movements.

2. Subjects One hundred consecutive pregnant Swedish speaking nulliparous women attending the municipal antenatal clinic in Malmo¨, Sweden, were asked at their first visit to the antenatal clinic to participate in a prospective study on ’’the effect of psycho-social factors on pregnant women’s health’’. The inclusion criteria were: (1) healthy woman, (2) speaking Swedish fluently, (3) no regular medication, and (4) singleton pregnancy. The women were asked to participate in three interviews at 12, 25, and 36 gestational weeks and in one interview postpartum, and to undergo an examination of the fetus after the interview at 36 gestational weeks. All women were fully informed of the study, and 93 women fulfilled the inclusion criteria and gave their consent to participate. The fetal examination included cardiotocography (CTG) and ultrasound examination. Fourteen women were delivered before the planned fetal examination, four subsequently declined to participate, and 16 women were not able to attend the examination for personal reasons. Of the 59 women who underwent the fetal examination, eight were excluded because the fetal examination was interrupted before 90 min had elapsed because of maternal discomfort, and 10 women did not have complete information with regard to stress variables. Thus, 41 women were included. The biological and socio-demographic background variables of the 52 women excluded and the 41 women included are shown in Table 1. The women included were older and heavier than those excluded. Moreover, they belonged to higher social classes, a greater proportion was married or cohabiting, and a greater proportion worked full-time. The women’s self-reported alcohol intake was analysed retrospectively by a psychiatrist specialised in toxicomania. No woman was considered to abuse or to be at risk for an over-consumption of alcohol.

K. Sjo¨stro¨m et al. / Early Human Development 67 (2002) 87–100

89

Table 1 Maternal biological and socio-demographic variables at 12 gestational weeks Included (n = 41)

Excluded (n = 52)

p-Value

low middle high

26 67 14 17 15 9

24 63 14 34 14 4

0.002a 0.01a 0.37a 0.04c

Family social classb, n (%)

low middle high

18 (43.9) 11 (26.8) 12 (29.3)

28 (53.8) 14 (27.0) 10 (19.2)

0.53c

Marital status, n (%)

married cohabiting single

26 (63.4) 14 (34.1) 1 (2.4)

33 (63.5) 10 (19.2) 9 (17.3)

0.04c

Employment situation, n (%)

full-time part-time unemployed

36 (87.8) 2 (4.9) 3 (7.3)

35 (67.3) 13 (25.0) 4 (7.7)

0.03c

Smoking at beginning of pregnancy, n (%)

non-smokers smokers

29 (70.7) 12 (29.3)

31 (59.6) 21 (40.4)

0.29c

Smoking frequency (cigarettes /day) at 36 gestational weeks, n (%)

0 V 10 >10

31 (75.6) 6 (14.6) 4 (9.7)

34 (65.4) 10 (19.2) 8 (15.4)

0.55c

Age (years), mean (S.D.) Weight (kg), mean (S.D.) Weight increase (kg), mean (S.D.) Maternal social classb, n (%)

a b c

(3.9) (9.1) (4.2) (41.5) (36.6) (22.0)

(3.6) (9.4) (4.6) (65.4) (26.9) (7.7)

Student’s t-test. The Swedish Socioeconomic Classification system. v2 test.

The anxiety levels at 12 and 36 gestational weeks of the women included were similar to those of the women excluded (state anxiety score at 12 weeks 34 F 8.5 vs. 34 F 9.6, trait anxiety score at 12 weeks 33 F 9.6 vs. 33 F 9.8, values available for 41 and 46 women, respectively; state anxiety score at 36 weeks 32 F 7.5 vs. 31 F 6.9, trait anxiety score at 36 weeks 31 F 8.2 vs. 30 F 6.0, values available for 41 and 16 women, respectively). The pregnancies of all women included were uncomplicated, and all newborns had 5-min Apgar scores of 9 or 10. Twenty-three (56%) infants were boys. One baby was small-forgestational age (birth weight below the mean 2 S.D. of the reference population), and one was large for gestational age (birth weight above the mean + 2 S.D. of the reference population) [6]. No baby was malformed.

3. Methods The participating women were interviewed, and they also completed the Spielberger State – Trait Anxiety Inventory (STAI, Form-Y) [7] on the four interviews. The interviews were semi-structured and were carried out by the same researcher (KS). They included questions on obstetric history, current somatic and psychological health, and current

90

K. Sjo¨stro¨m et al. / Early Human Development 67 (2002) 87–100

psychosocial situation. The interviews took place at the women’s regular visits to the antenatal clinic. After the last interview during pregnancy, the women underwent a fetal examination including Doppler ultrasound examination of fetal blood flow, ultrasound examination of fetal movements, and CTG. The mean gestational age at the examination was 38 gestational weeks (range 37 –40). Thirty-four women were examined at z 38 gestational weeks. The women were instructed to have a cooked meal at noon but to abstain from smoking and from having coffee or tea after 1 p.m. on the examination day. No alcohol intake was allowed after 3 p.m. on the day before. A Doppler examination of fetal blood flow was performed starting at 3 p.m. Immediately thereafter, the maternal pulse was palpated and counted during 60 s, and blood pressure was measured in the right arm at heart level using a sphygmomanometer. A fruit was then offered to the woman, and she had the opportunity to visit the bathroom. All examinations of FHR and fetal movements started between 3:30 and 4 p.m. The same technician carried out all examinations. During the examination the women rested in a semi-recumbent position in a dimly lit and quiet room. FHR, fetal movements perceived by the pregnant woman and fetal movements seen by the technician were recorded simultaneously on paper using a three-channelled cardiotocograph (FM7/7L Sonicaid, Oxford Instruments Medical, Surrey, UK). The paper speed was 30 mm/min. FHR was recorded using Doppler ultrasound. Observations of fetal movements were made with the aid of a 3.5-MHz real-time linear array ultrasound scanner (Hitachi EUB 410, Hitachi Medical, Tokyo, Japan). The ultrasound transducer was positioned to include a cross-section of the fetal trunk and at least one fetal limb. It was held in place by the technician, who recorded all fetal movements she saw on the ultrasound screen using an event-marker, the push-button of the event-marker being pressed for as long as a movement was seen on the ultrasound screen. The signals from this event-marker were fed into the tocography channel of the cardiotocograph. The ultrasound screen was placed out of sight of the pregnant woman, who was asked to press the fetal movement eventmarker of the cardiotocograph as soon as she felt the fetus move and to press it for as long as she felt the fetal movement. The fetal movements recorded by the mother were not used for analysis in this study. Examination time was 120 min. The paper recordings were analysed manually off-line by one of the authors (L.V.), who had no knowledge of the maternal anxiety scores. The FHR pattern was classified as compatible with pattern A, B, C, or D [5] (Fig. 1), or as unclassifiable (e.g., because of technical problems). For each FHR pattern, the most common baseline FHR and the most common FHR variability as estimated from visual inspection of the paper recording were noted. The duration of each FHR pattern was expressed as a percentage of the total examination time. The duration of the fetal movements observed by ultrasound was calculated separately for each FHR pattern and was expressed as a percentage of the duration of that FHR pattern. Durations were measured in mm using a ruler. Fetal movements occurring V 6 s. apart (i.e., V 3 mm apart) were considered to belong to the same fetal movement. The STAI scale is a commonly used scale in all kinds of psychosomatic research [7]. The state part of the scale is regarded to be a very good instrument to reveal acute anxiety [7] and the trait scale has been found to reveal depression, negative affects, and anxiety [8]. The STAI scale consists of 40 statements describing various emotional states [7].

K. Sjo¨stro¨m et al. / Early Human Development 67 (2002) 87–100

91

Fig. 1. Examples of the four distinct fetal heart rate (FHR) patterns. FHR pattern A: FHR stable with a narrow oscillation bandwidth. Isolated accelerations occur strictly related to body movements. FHR pattern B: FHR with a wider oscillation bandwidth than pattern A, and frequent accelerations associated with body movements. Accelerations are indicated by dotted area. FHR pattern C: stable FHR, but with a wider oscillation bandwidth than pattern A and a more regular oscillation frequency than pattern B. No accelerations. FHR pattern D: FHR unstable, with large and prolonged accelerations often fused into sustained tachycardia. bpm = beats per minute. (Reprinted from Early Human Development, vol. 15, Mulder et al., 1987, Emergence of behavioural states in fetuses of type-1 diabetic women, p. 237, with permission from Elsevier.)

Twenty of these statements require the subjects to describe their emotional reactions in anxiety terms at a particular moment or period in time, i.e., state anxiety. Another 20 items require the subject to describe how they generally feel and their general tendency to respond to situations perceived as threatening, i.e., trait anxiety. In this study, the women scored their experienced state anxiety during the 3 months preceding each interview. We used the state anxiety reported to have been experienced between 25 and 36 weeks of gestation and the trait anxiety reported at 36 gestational weeks for analysis. On the basis of the state and trait anxiety scores, two groups were created: one with low anxiety scores (low anxiety group) and another with high anxiety scores (high anxiety group). Women in the low anxiety groups had scores below the median of the total study population, and women in the high anxiety groups had scores equal to or above the median. The low state anxiety group had scores from 20 to 31, the high state anxiety group

92

K. Sjo¨stro¨m et al. / Early Human Development 67 (2002) 87–100

had scores from 32 to 57, the low trait anxiety group had scores from 20 to 29, and the high trait anxiety group had scores from 30 to 54. The results of the fetal examinations were compared between the groups with low and high anxiety scores, the outcome measures being the presence and duration (expressed as a percentage of the total examination time) of FHR patterns A, B, C, and D, the percentage duration of fetal movements in each FHR pattern, and the baseline FHR and FHR variability in each FHR pattern. Possible relationships between maternal anxiety scores and the outcome variables were also assessed by visual inspection of scattergrams with anxiety scores on the x-axis and the outcome variable on the y-axis. Only fetuses manifesting the FHR pattern to be analysed were included in the analysis of the duration of that particular FHR pattern. Medical record data on the women’s pregnancy and delivery were retrieved retrospectively. All pregnancies were dated using ultrasound fetometry in the early second trimester [9], and all fetuses were screened for intra-uterine growth retardation by ultrasound fetometry at 32– 34 gestational week [10].

4. Statistical analysis Statistical analyses were carried out using the Statistical Package for the Social Sciences (SPSS) software (SPSS, Chicago, IL, USA, 1999) or StatXact-3 statistical program (Cytel Software, Cambridge, MA, USA, 1995), the latter software being used for calculations of the exact p-values for the Mann – Whitney test, when the sample size Table 2 The presence and duration (expressed as percentage of total examination time) of different fetal heart rate patterns in different anxiety groups State anxiety

p-Value Trait-anxiety High (n = 22)

68 (13/19) 15 (4 – 31)

68 (15/22) 13.5 (4 – 46)

1.0a 0.87b

FHRP B Present, % 100 (19/19) 95 (21/22) Duration, % Median (range) 86 (52 – 100) 80 (46 – 100)

1.0a 0.28b

FHRP C Present, % Duration, % Median (range)

16 (3/19) 9 (4 – 15)

23 (5/22) 20 (10 – 28)

0.70a 0.07c

20 (4/20) 12 (4 – 20)

19 (4/21) 23 (10 – 28)

1.0a 0.20c

FHRP D Present, % Duration, % Median (range)

16 (3/19) 7 (7 – 9)

18 (4/22) 20 (17 – 67)

1.0a 0.06c

15 (3/20) 7 (7 – 9)

19 (4/21) 20 (17 – 67)

1.0a 0.06c

FHRP A Present, % Duration, % Median (range)

FHRP = fetal heart rate pattern. a Fisher’s exact test. b Mann – Whitney U-test. c Mann – Whitney U-test, exact significance.

Low (n = 20)

p-Value

Low (n = 19)

80 (16/20) 16 (4 – 37)

High (n = 21) 57 (12/21) 13 (4 – 46)

100 (20/20) 95 (20/21) 80 (52 – 100) 83 (46 – 100)

0.18a 0.55b

1.0a 0.36b

K. Sjo¨stro¨m et al. / Early Human Development 67 (2002) 87–100

93

was small. Anxiety scores, FHR, and FHR variability were normally distributed, whereas the distributions of the duration of FHR patterns and fetal movements were skewed. Pearson’s product moment correlation coefficient (r) and Spearman’s rank correlation coefficient (rho) were used to assess correlation between normally distributed and skewed data, respectively. Student’s t-test and Mann – Whitney’s U-test were used to test the statistical significance of differences in normally distributed and skewed continuous data,

Table 3 Fetal heart rate and fetal heart rate variability in fetal heart rate patterns A, B, C, and D in different anxiety groups p-Valuea

State anxiety

FHRP A n FHR (bpm) Mean (S.D.) Range Variability (bpm) Mean (S.D.) Range FHRP B n FHR (bpm) Mean (S.D.) Range Variability (bpm) Mean (S.D.) Range FHRP C n FHR (bpm) Mean (S.D.) Range Variability (bpm) Mean (S.D.) Range FHRP D n FHR (bpm) Mean (S.D.) Range Variability (bpm) Mean (S.D.) Range

p-Valuea

Trait-anxiety

Low (n = 19)

High (n = 22)

Low (n = 20)

High (n = 21)

13

15

16

12

130 (8.7) 120 – 145

131 (7.5) 120 – 140

0.75

130 (8.0) 120 – 145

131 (8.1) 120 – 140

0.69

9 (2.0) 6 – 12

7 (2.6) 2 – 10

0.09

8 (2.3) 4 – 12

8 (2.8) 2 – 10

0.97

19

21

20

20

134 (7.0) 120 – 145

135 (7.7) 120 – 150

0.65

133 (6.6) 125 – 145

136 (7.9) 120 – 150

0.28

21 (3.0) 12 – 25

18 (4.0) 10 – 24

0.05

19 (3.2) 12 – 24

19 (4.2) 10 – 25

0.97

3

5

4

4

135 (5.0) 130 – 140

126 (5.5) 120 – 135

0.06

131 (8.5) 120 – 140

128 (5.0) 125 – 135

0.48

15 (3.1) 12 – 28

12 (4.3) 10 – 20

0.46

14 (2.8) 12 – 18

13 (5.0) 10 – 20

0.62

3

4

3

4

143 (5.8) 140 – 150

140 (4.1) 135 – 145

0.41

143 (5.8) 140 – 150

140 (4.1) 135 – 145

0.41

7 (2.9) 5 – 10

16 (4.4) 10 – 20

0.03

7 (2.9) 5 – 10

16 (4.4) 10 – 20

0.03

FHRP = fetal heart rate pattern. FHR = fetal heart rate. bpm = beats per minute. a Student’s t-test.

94

K. Sjo¨stro¨m et al. / Early Human Development 67 (2002) 87–100

respectively. The v2 test and Fisher’s exact test were used as appropriate to test the statistical significance of differences in categorical data. A p-value < 0.05 (two-tailed) was considered statistically significant. The Ethics committee of the Medical Faculty, University of Lund, Sweden, approved this study.

5. Results The mean duration of the fetal examinations was 115 min F 6.8 (S.D.), range 96– 123. There was a statistically significant negative correlation between trait anxiety score and the duration of the experiment (r = 0.32, p = 0.04). The mean state and trait anxiety scores were 32 F 7.5 (range 20– 57) and 31 F 8.2 (range 20 – 54), respectively. State and trait anxiety scores were strongly positively correlated (r = 0.70, p = 0.0001). Mean maternal systolic and diastolic blood pressures at the examination were 113 F 10.1 and 68 F 19.0 mm Hg, respectively, and mean maternal pulse rate was 83 beats per min F 10.8. Twenty-eight (70%) of 40 fetuses changed between active and quiet states during the examination (in one fetus a possible change between active and quiet states could not be evaluated because of long periods with technical problems). The proportion of fetuses changing between active and quiet states did not differ significantly between the low state anxiety group and the high state anxiety group (63% vs. 76%, i.e., 12/19 vs. 16/21; p = 0.5, v2 test) nor between the low trait anxiety group and the high trait anxiety group (75% vs. 65%, i.e., 15/20 vs. 13/20; p = 0.7, v2 test). FHR pattern A was found in 28 (68%) fetuses, pattern B in 40 (98%), pattern C in eight (19%), and pattern D in seven (17%) fetuses. Four (10%) fetuses had periods of unclassifiable patterns corresponding to 6% – 39% of the recording time. The proportion of fetuses with FHR patterns A, B, C, and D did not differ significantly between low and high state and trait anxiety groups (Table 2). The duration of FHR patterns A and B was similar in the low and high anxiety groups, but the duration of FHR pattern D tended to be longer in women with high state and trait anxiety scores (Table 2), and increased with increasing maternal trait anxiety scores (rho = 0.88; p = 0.008) and state anxiety scores (rho = 0.72; p = 0.07). The duration of FHR pattern C also tended to be longer in the high anxiety groups, but the differences were not statistically significant, and inspection of scattergrams showed no clear relationship between anxiety scores and the duration of FHR pattern C (Table 2). Baseline FHR and FHR variability in the different anxiety groups are shown in Table 3. There was no significant correlation between anxiety scores and baseline FHR in any FHR pattern, (visual inspection of scattergrams, calculation of correlation coefficients), and the baseline FHR in the different FHR patterns did not differ significantly between women

Fig. 2. (a) The relation between maternal state anxiety experienced between 25 and 36 gestational weeks and FHR variability in FHR pattern D at term (n = 7). The regression line is shown. (b) The relation between maternal trait anxiety assessed at 36 gestational weeks and FHR variability in FHR pattern D at term (n = 7). The regression line is shown. bpm = beats per minute.

K. Sjo¨stro¨m et al. / Early Human Development 67 (2002) 87–100

95

96

K. Sjo¨stro¨m et al. / Early Human Development 67 (2002) 87–100

Table 4 The percentage time the fetus spent moving in fetal heart rate patterns A, B, C, and D in different anxiety groups State anxiety

p-Value

Low (n = 19)

High (n = 22)

FHRP A n Movements, percentage time Median (range)

13 0.7 (0 – 6)

14* 1.2 (0 – 4)

FHRP B n Movements, percentage time Median (range)

19 12 (9 – 25)

20* 13 (5 – 32)

FHRP C n Movements, percentage time Median (range)

3 0 (0 – 3)

FHRP D n Movements, percentage time Median (range)

3 50 (18 – 73)

5 0.2 (0 – 1.1)

4 48 (30 – 65)

Trait-anxiety

p-Value

Low (n = 20)

High (n = 21)

0.39a

16 0.9 (0 – 6)

11* 1.2 (0 – 3)

0.76a

0.98a

20 12 (9 – 32)

19* 13 (4 – 27)

0.72a

0.93b

4 0 (0 – 3)

1.0b

3 50 (18 – 73)

4 0.4 (0 – 1.1)

4 48 (30 – 65)

0.49b

1.0b

* In one woman ultrasound examination of fetal movements was not performed. a Mann – Whitney U-test. b Mann – Whitney U-test, exact significance.

with low and high anxiety scores. There was also no significant correlation between state and trait anxiety scores and FHR variability in FHR patterns A, B, or C (visual inspection of scattergrams, calculation of correlation coefficients), but there was a statistically significant positive correlation between state and trait anxiety scores and FHR variability in FHR pattern D (r = 0.86, p = 0.01; and r = 0.96, p = 0.001); see Fig. 2. The variability in FHR pattern D was significantly higher in the high anxiety groups than in the low anxiety groups (Table 3). The percentage time that the fetus spent moving in each FHR pattern did not differ between the low and high anxiety groups (Table 4), and there was no correlation between maternal anxiety scores and fetal movements (visual inspection of scattergrams, calculation of correlation coefficients).

6. Discussion Three previous studies—two by Van den Bergh and co-workers and one by Groome and colleagues—have examined the relationship between maternal anxiety in the third trimester as assessed by the STAI and fetal behavioural state and fetal movements in different fetal behavioural states [11 – 13]. These studies differ methodologically from ours in several aspects: there are differences in definitions of state anxiety, dating of pregnancy,

K. Sjo¨stro¨m et al. / Early Human Development 67 (2002) 87–100

97

and gestational age at the examination, and there may be differences in the anxiety levels of the women. We did not determine the true fetal behavioural states because we did not register eye movements [5]. However, in late pregnancy, the FHR pattern probably reflects the current fetal behavioural state; at least FHR pattern A and B can probably distinguish quite accurately between state 1F and 2F [14]. Because our examinations were carried out at a mean gestational age of 38 weeks (only seven women were examined at 37 gestational weeks), when fetal behavioural states are established in almost all fetuses [5,14,15], the FHR pattern probably did reflect the current behavioural state in most cases. Thus, in our study, FHR pattern A can probably be taken to indicate state 1F, FHR pattern B state 2F, FHR pattern C state 3F, and FHR pattern D state 4F, see Fig. 1. The definition of state anxiety differed between our study and those of Van Den Bergh and colleagues [11,13]. In our study, the women estimated their state anxiety from 25 to 36 weeks of pregnancy, whereas in the studies of Van den Bergh and colleagues the women rated the anxiety they felt immediately before and after the fetal examination [11,13]. Moreover, the fetuses in one of the studies by Van Den Bergh and co-workers [11] were younger (36 – 37 gestational weeks) than those in our study. We determined gestational age on the basis of ultrasound fetometry in every case. The method of estimation of gestational age was not mentioned in the studies of Van den Bergh et al. [11,13]. In the study by Groome et al. [12], dating was based on the last menstrual period in some cases. The differences in results between our study and those cited [11 –13] may be — at least partly — explained by differences in methodology. Groome et al. [12] found that fetuses of women with high trait anxiety spent more time in behavioural state 1F than fetuses of women with low trait anxiety, and they reported the fetuses to move less during behavioural state 2F. In two different studies, Van den Bergh and colleagues found a significant positive correlation between maternal state anxiety in late pregnancy and fetal movements in states 2F, 3F, and 4F [13], and between maternal trait anxiety and fetal movements in behavioural state 4F [11]. We found no relation between maternal anxiety levels and fetal movements. In one of the studies by Van den Bergh et al. [11], a significant positive correlation between maternal trait anxiety and the duration of the time the fetus spent in state 4F was found. The latter finding supports our finding of a positive correlation between maternal trait anxiety scores and the duration of FHR pattern D (assuming FHR pattern D to represent state 4F). Our findings suggest that maternal anxiety might have an effect on the time the fetus is ’’awake’’ in late pregnancy (tendency of longer duration of FHR pattern C and D; Table 2). The association may not necessarily be causal but might reflect a genetic similarity between mother and fetus. The women in our study had state and trait anxiety scores (32 F 7.5 and 31 F 8.2, respectively) that were somewhat lower than the reference values reported for healthy nonpregnant women 19 – 39 years old (36 F 11.0 and 36 F 9.5, respectively) [7]. Thus, anxiety levels were probably within the normal range in our study. Nonetheless, it cannot be excluded that the women in our study represented a selected group with particularly low or high anxiety levels in the third trimester, because third trimester anxiety scores were unavailable for two thirds of the women excluded. However, this seems unlikely because women included and excluded had similar trait anxiety scores at 12 gestational weeks, and trait anxiety is known to be stable over time in non-pregnant women [7]. State anxiety

98

K. Sjo¨stro¨m et al. / Early Human Development 67 (2002) 87–100

scores, however, may well change over time [7]. In the studies by Groome et al. [12] and Van den Bergh and et al. [11,13] anxiety scores were not presented, with one exception: mean state anxiety score was 32 in one of the studies by Van den Bergh et al. [13]. The maternal anxiety levels are likely to have been lower in our study and in those cited above [11 –13] than in a report from Ianniruberto and Tajani [16]. They found that the fetuses of panic-stricken women manifested dramatic hyperkinesia followed by immobility of varying duration [16]. We wanted the testing procedure to be as comfortable as possible for the women, so that the testing procedure itself would not cause maternal anxiety. Van den Bergh et al. [13] and Groome et al. [12] took similar precautions. Still, we found a statistically significant positive correlation between maternal trait anxiety scores and the duration of the examination, indicating that women with the highest trait anxiety scores tended to break the experiment earlier than those with lower scores. However, the women who were excluded because of examination duration < 90 min did not have higher trait anxiety scores than those included (31 F 7.2 vs. 31 F 8.2), but only five of these eight women had completed the STAI at 36 gestational weeks. We did not find any association between maternal state and trait anxiety and baseline FHR. To the best of our knowledge, there is only one publication where a possible association between maternal anxiety and baseline FHR in late pregnancy was studied [17]. Monk et al. [17] found that stress induced high maternal state anxiety was associated with an increase in FHR. However, unlike us, Monk and co-workers examined state anxiety during the very experiment, and they did not take different FHR patterns or fetal behavioural states into account. Monk and co-workers measured anxiety not by STAI but by the State – Trait Personality Inventory (STPI), which is an anxiety scale similar to the STAI [18]. We found an association between maternal state and trait anxiety and FHR variability in FHR pattern D, high anxiety levels being associated with higher FHR variability. To the best of our knowledge, others have not studied possible associations between FHR variability and maternal anxiety during gestation. However, in a preliminary analysis of their data, DiPietro et al. [19] found that greater perceived stress by the mothers (measured as the combination of ‘‘hassles and uplifts’’) was associated with reduced FHR variability. This is interesting in view of our finding that FHR variability in FHR patterns A and B tended to be lower in the high state anxiety group (Table 3) and in view of a reported association between panic disorders and reduced heart rate variability in adults [20]. DiPietro and colleagues did not take different FHR patterns into account, but because FHR patterns A and B occur much more frequently than FHR patterns C and D, their finding is likely to be based mainly on FHR patterns A and B. The regulation of FHR variability is very complex [21], and it remains unclear by which mechanisms maternal anxiety could affect the FHR variability. It is an open question whether maternal anxiety during pregnancy and labour may affect FHR variability to such an extent that it could influence clinical judgements. To sum up, we have found no effect of maternal state and trait anxiety within the normal range on baseline FHR or fetal movements in late pregnancy. This does not exclude an effect of extreme maternal anxiety. There was a possible association between maternal anxiety and the duration of FHR pattern D and FHR variability in FHR pattern D,

K. Sjo¨stro¨m et al. / Early Human Development 67 (2002) 87–100

99

higher maternal anxiety scores being associated with longer duration of FHR pattern D and higher FHR variability in FHR pattern D.

Acknowledgements This work was supported by a grant from the Swedish Society of Medicine (the So¨derstro¨m-Ko¨nigska foundation).

References [1] Paarlberg KM, Vingerhoets AJM, Passchier J, Dekker GA, Van Geijn P. Psychosocial factors and pregnancy outcome: a review with emphasis on methodological issues. Psychosom Res 1995;39:563 – 95. [2] Sjo¨stro¨m K, Valentin L, Thelin T, Marsˇa´l K. Maternal anxiety in late pregnancy and fetal hemodynamics. Eur J Obstet Gynecol Reprod Biol 1997;74:149 – 55. [3] Teixeira JM, Fisk NM, Glover V. Association between maternal anxiety in pregnancy and increased uterine artery resistance index: a cohort based study. Br Med J 1999;16:153 – 7. [4] Timor-Tritsch IE, Dierker LJ, Hertz RH, Deagan NC, Rosen MG. Studies of antepartum behavioural state in the human fetus at term. Am J Obstet Gynecol 1978;132:524 – 8. [5] Nijhuis JG, Prechtl HFR, Martin Jr CB, Bots RSGM. Are there behavioural states in the human fetus? Early Hum Dev 1982;6:177 – 95. [6] Marsˇa´l K, Persson PH, Larsen T, Lilja H, Selbing A, Sultan B. Intrauterine growth curves based on ultrasonically estimated fetal weights. Acta Pediatr 1996;85:843 – 8. [7] Spielberger CD, Gorsuch RL, Lushene RE, Vagg PR, Jacobs GA. The State – Trait Anxiety Inventory: STAI. Palo Alto CA: Consulting Psychologist Press; 1970. [8] Bieling PJ, Antony MM, Swinson RP. The State – Trait Anxiety Inventory, Trait version: structure and content re-examined. Behav Res Ther 1997;36:777 – 88. [9] Persson PH, Weldner BM. Reliability of ultrasound fetometry in estimating gestational age in the second trimester. Acta Obstet Gynecol Scand 1986;65:481 – 3. [10] Eik-Nes SH, Persson PH, Gro¨ttum P, Marsˇa´l K. Prediction of fetal growth deviation by ultrasonic biometry. Acta Obstet Gynecol Scand 1983;62:117 – 23. [11] Van den Bergh B, Vandenberghe K, Danie¨ls H, Casaer P, Marcoen A. The relationship between maternal emotionality during pregnancy and the behavioural development of fetus and neonate. In: Gennser G, Marsˇa´l K, Svenningsen N, Lindstro¨m K, editors. International Conference on Fetal and Neonatal Physiological Measurements III. Proceedings of the 3rd International Conference of Fetal and Neonatal Physiological Measurements; Malmo¨/Ronneby, University Hospital Malmo¨, Sweden; 1989. p. 359 – 364. [12] Groome LJ, Swiber MJ, Bentz LS, Holland SB, Atterbury JL. Maternal anxiety during pregnancy: effect on fetal behaviour at 38 to 40 weeks of gestation. J Dev Behav Pediatr 1995;16:391 – 6. [13] Van den Bergh B, Mulder E, Visser G, Poelmann-Weesjes G, Bekedam D, Prechtl H. The effect of (induced) maternal emotions on fetal behaviour: a controlled study. Early Hum Dev 1989;19:9 – 19. [14] Pillai M, James D. The development of fetal heart rate patterns during normal pregnancy. Obstet Gynecol 1990;76:812 – 6. [15] Romanini C, Rizzo G. Fetal behaviour in normal and compromised fetuses. An overview. Early Hum Dev 1995;43:117 – 31. [16] Ianniruberto A, Tajani E. Ultrasonographic study of fetal movements. Semin Perinatol 1981;5:175 – 81. [17] Monk C, Fifer WP, Myers MM, Sloan RP, Trien L, Hurtado A. Maternal stress responses and anxiety during pregnancy: effects on fetal heart rate. Dev Psychobiol 2000;36:67 – 77. [18] Spielberger CD, Johnson E, Jacobs G, Ritterband R, Barker I, Crane R, et al. State – Trait Personality Inventory (Form Y). Palo Alto, CA: Mind Garden; 1995.

100

K. Sjo¨stro¨m et al. / Early Human Development 67 (2002) 87–100

[19] DiPietro J, Hodgson DM, Costigan A, Hilton SC, Johnson TRB. Fetal neurobehavioral development. Child Dev 1996;67:2553 – 67. [20] Friedman B, Thayer J. Autonomic balance revisited: panic anxiety and heart rate variability. J Psychosom Res 1998;44:133 – 51. [21] Dalton KJ, Phil D, Dawes GS, Patrick JE. The autonomic nervous system and fetal heart rate variability. Am J Obstet Gynecol 1983;146:456 – 62.