The prevalence rates, 16.2% (30/185) and 13.0% (21/162), of maternal postnatal depression correspond to earlier studies and depression in women, in general (Pritchard and Harris 1996, Miller 2002). The purpose was to detect postnatal depression and to find out the possible association between labour analgesia and postnatal depression and to detect possible seasonal variations, not to establish a clinical diagnosis.
Paternal self-reported postnatal depression was lower than that in mothers, which is a finding similar founding to prior studies (Ballard 1994, Matthey et al. 2000). A high rate of zero symptoms reporting by men was also found. Matthey and others have suggested that gender specific symptoms may be possible. Thus, more studies are needed to assess stress and anxiety in men. Paternal depression was associated with maternal depression, but no conclusions can be drawn of the causal relationship, other than there seemed to be strong couple co-morbidity. Fathers had no depressive symptoms unless their partner had been depressed immediately after delivery or during the first four months postpartum.
The taboo of this illness must be aborted. More discussion and more attention on both the physical and emotional well-being of the mother in the maternal ward and well baby-clinics and more frequent usage of detection questionnaires as helping device is needed to prevent postnatal depression. Mental help expertise can be provided by a well-baby clinic nurse, GP or psychiatrist (Stotland 1999). Different professionals, nurses, doctors and social workers may work together, and mental health professionals can be consulted in stead of referral. In Finland ”Baby Blues”-groups have been organised for postnatally depressed mothers in different cities by the Mother and Child Home and Shelters (Ensi- ja turvakodit). Education and support groups can be found in self-help networks, such as Postpartum Support International (http://www.chss.iup.edu/postpartum/) or Depression After Delivery (http://www.depressionafterdelivery.com/).
We found postnatal depression to be associated with maternal age and the analgesia used during delivery. The prevalence of maternal postnatal depression also varied when comparing different seasons and seasonality. In 1997 the average age of parturients in Finland was 29.8 years. Mothers, aged 30 years or less, had in general more depression than the mothers aged 31 years or more. However, maternal age did not affect the persistence of depressive symptoms.
Studies of the association of modern labour analgesia and postnatal depression are rare. In fact, only one Finnish study was found, which evaluated how the perceived labour pain was associated with postnatal depression. Inadequate pain relief and emergency CS highly predicted disappointment with delivery, but not increased the risk of postnatal depression (Saisto et al. 2001). The mode of delivery and postpartum mood has been studied a lot, and there is still debate as to whether CS, especially emergency CS, is a risk for postpartum mental health (Culp 1989, Boyce 1992, Reynolds 1997).
In this study CS was not associated with immediate or later postpartum maternal depressive symptoms, although the emergency CS mothers were younger than other mothers and their children were more often treated at the NICU compared to other children. On the other hand, the analgesia used during VD seemed to protect the mother from immediate depressive symptoms. And the epidural/PCB analgesia group had less persistant depressive symptoms compared to the group of mothers with no analgesia during VD. These findings suggest that adequate pain relief gives the mother a better basis for coping with labour pain and might therefore improve the immediate circumstances for the mother to recover and be better able to bond with the newborn.
The length of labour has been found to correlate with reported pain (Niven and Gisberg 1984, Scott et al. 1999). In our study, mothers who received epidural analgesia or PCB had the longest duration of delivery, which may have influenced the need of pain relief. According to a systematic review by Leighton and Halpern (2002), epidural analgesia does not increase the duration of stage I of delivery. In our study, the mothers without analgesia had the shortest mean length of labour (406 minutes) and the shortest mean length of time spent in the delivery room (255 minutes), which may both have influenced to the fact that no pain relief was administered. As pain relief had favourable consequences to the postpartum maternal well-being and might diminish the risk of postnatal depressive symptoms, our results suggest that adequate pain relief is recommendable for all parturients regardless of the length of labour.
Parity was not associated with postnatal depression within these different pain relief groups, although there were more primiparas in the epidural/PCB analgesia group than in the other groups with VD. However, more emergency CS’s were performed on younger mothers. This underlines the importance of the availability of satisfactory labour analgesia not only to primiparas but to all parturients.
Even if both Seasonal affective disorder (SAD) (Rosenthal et al. 1984, Okawa et al. 1996, Maskall et al. 1997, Saeed and Bruce 1998) and postnatal depression (e.g. Warner 1996, Beck 1996, Fontaine 1997) are widely studied topics, their possible association has rarely been discussed and hardly ever been studied. During this study no previous publications on seasonal variation in postnatal depression were found.
We found more mild depression (the EPDS 10) in the autumn immediately after delivery, and less depression (the EPDS 13) in the spring measured at four months postpartum. When using classification by the amount of light there was more depression during the dark time (1.58; 1.05–2.11) immediately postpartum. This result supports the view that biological factors may have a role in the aetiology of the immediate postnatal depression.
A few hormones have been claimed to cause postnatal depression, e.g. extremely low levels of prolactin or excessive levels of progesterone. SAD, in turn is thought to be strongly related to lack of serotonin (Thorell et al. 1999, Jepson et al. 1999). But, it is also associated with changes in melatonin synthesis and secretion (Lewy et al. 1980), and circadian rhythms (Wirz-Justice et al. 1993). The symptoms of SAD are partly those of general depression, but the core symptoms include sadness, anxiety, inability to interact with other people and tiredness. The prognosis is usual good and symptoms mild (Partonen 1998). Additionally, SAD with minor depressive episodes is more frequent in women, whereas SAD with major depressive episodes in commonly experienced by men (Blazer et al. 1998). It could be that a subgroup of women is more vulnerable to biologic changes and therefore more sensitive to seasonal variation of postnatal mood in the immediate postpartum time. Indeed, in the Oulu area there is only 4,5 hours of daylight in the darkest time of the year.
Interestingly, it has been noticed that children who are born at summer have a statistically bigger birth weight and –height. The longer height can be a beneficial factor at later life, but the larger weight can predispose to overweight at adulthood (Partonen 2002). This shows that seasonal variability has been associated with and assessed in relate to birth and infant variables.
Some limitations must be considered: We did not measure depressive symptoms during pregnancy, and we have only the medical records of the mothers on their prior mental health. Knowing the fact that several mothers hide or deny their depressive symptoms during pregnancy or postpartum, the prior depression can not be fully ruled out.
Finnish mothers reported more surprise and less distress than American mothers. Finnish mothers also reported more anger than American mothers, but the difference was minimal. One explanation for Finns reporting more surprise than Americans could be methodological. There were pictures of the bright blue-eyed girl who were interpreted by Finns as surprised compared to American interpretations of interest. This could be partly due to the close Finnish translations of “surprise” and “interest”. It might also be cultural: American people are usually more expressive than Finns.
The interpretations of distress can not be explained by translation difficulties or other issues related to the method. It should be remembered that the Finnish group included mothers with depression, compared to American mothers, of whom the depression status was not assessed.
Although, the Finnish and the American vocabulary and lexicon may be considered almost 100% equal, some cultural variability may confound the available data. Thus, dimensional analysis must be performed in a bigger sample in order to be sure of the cultural stability using a method originally based on a foreign language. It seems probable that differences between Finnish and Croatian mothers in 7 out of 13 affect categories in the pilot study are probably due to poor validation work of the Croatian lexicon and small sample size of the Finnish pilot mothers. However, the replicated result of the differences between American and Finnish mothers supports the hypothesis that the more blended and more ambiguous the facial expression, the more cultural circumstances and customs direct what we see and interpret.
Maternal affect interpretation was associated with maternal postnatal depressive symptoms. Immediately postpartum, depressed mothers reported more anger than mothers without depressive symptoms. Mothers with high depressive scores once reported less joy immediately postpartum than other mothers. The mothers of the ERA subgroup who were repeatedly rated as depressed reported significantly more sadness compared to other ERA subgroup women. This is a similar finding to the study of American depressed mothers. Those depressed women interpreted infants to show more fear and anxiety than non-depressed mothers, and within the depressed group, the interpretations of fear correlated with the severity of depression (Zahn-Waxler and Wagner 1993).
The results indicate that depression characterises the interpretation of emotions from pure expression, as depressed mothers see less joy and more sadness from still infant facial features. Immediately after delivery, also anger was interpreted less by depressed mothers. We could assume that depressed mothers were interpreting and seeing less intensive emotions on infants’ faces than not depressed mothers. However, the method must be thoroughly validated (dimensional analysis) before setting permanent assumptions. Additionally, the IFP was completed only once, which limits the drawing of any far-reaching conclusions. But it is possible that IFP could predict the maternal postnatal mood better than a self-report questionnaire, if depressed mothers do not recognise their depressive symptoms or are not willing to admit having them.
The quality of the mother-infant interaction of occasionally depressed women did not differ from not depressed mothers at ten months postpartum. Statistical differences were found when comparing mothers who had depressive symptoms once to the mothers who had persistent depressive symptoms. The main idea was to evaluate how the persistent postnatal depression affects the early mother-infant relationship.
Mothers with persistent depressive symptoms showed less negative affect in general, not only toward their infants, which were ten months old at the time of assessment. Their dyad contained less anger and were less anxious than those of other mothers. The result is not surprising when thinking about the nature of postnatal depression: silent, secret and even smiling. Only few maternal negative expressions of emotions could associate with excessive fatigue of the mothers. Mothers could have protected their infants from negative affects as well. Depressed mothers often try very hard of being good mothers and they “try to take right steps” with the child without feelings (Stern 1994). Mothers can also deny their depressive feelings or express positive feelings instead of negative ones (reaction formation) (Small et al. 1994). Also, the fact that during videotapings there was a stranger standing by next to the mother and the infant could have affected to mothers behaviour.
The children of persistently depressed mothers were slightly less impulsive. They showed less social initiatives and had fewer visual contacts with their mother. They were less expressive and receptive to their mother than children of not depressed or occasionally depressed women. In general, they had fewer social contacts with their mother. This is an opposite finding from Stern’s (1994) which has shown that children of depressed mothers attempt to invite their mothers to interact; e.g. try to wake mothers alive. However, it corresponds to the finding that mother’s behaviour is imitated by the infant: the mother is withdrawn and the infant builds her own world and is harder to drawn in social contact with the “avoidant” mother (Field 1992). The result may also be considered similar to Beck’s finding (1996) that also infant variables have a significant contribution to the early mother-infant relation. However, we do not know whether the infants were less responsive to their mother because of maternal depression, or the maternal depression was partly a consequence of infant behaviour.
The dyad was characterised by lack of mutual interaction; there were fleeting periods of unconnectedness and less reciprocity. Some of the statistically significant differences were minimal and thus thought to be without clinical relevance. Contrary to the study of chronically depressed mothers and their 36-month-old infants (NICHD 1999), maternal depressive scores in our study did not affect maternal sensitivity in interaction with their infants at videotaping.
Part of the focus of this study was in linguistics, and the linguistic analysis was made by a qualified logopaedist. Language comprehension skills of three-and half-year-old children of depressed mothers were equal to children of not depressed mothers. Differences were found in language expression; the more persistent the maternal depressive symptoms, the poorer the child performed on the Expressive Language Scale. However, considerable individual differences were found within the group of depressive mothers. Thus, it is likely that the possible effects of maternal depression on children’s linguistic development are not straightforward.
During the interaction assessment depressed mothers did not speak or use gestures or other facilitation less frequently than other mothers. The differences were found in the infant variables and dyadic attunement. Infants of persistently depressed mothers had less visual contact and were harder to draw to social interaction with their mother than the infants of mothers who did not report depressive symptoms. This could be due to mother’s incapability (due to fatigue) to create social interaction. Infants could have also rejected maternal play or other social initiations. Also, the methodological aspect must be considered: researcher used her own toys, which could have held infant’s interest. Dyads of depressed mothers and their infants had less reciprocal moments than not depressed mothers and their infants, but their overall dyad was coloured by less anxiety than the dyads of not depressed mothers and their infants. All these issues may influence how a child learns and uses vocal expressions, but more studies are needed, preferably with a larger group.
In conclusion, persistent maternal depression characterised some issues in the mother-infant interaction at 10 months postpartum. Depressed mothers’ behaviour in interaction with their children seemed to be more empty and blank, instead of aggressive or angry. The mothers were physically present, but partly psychologically absent (Green 1986). The continuance of postnatal depression may pose a threat to the genial and active early relationship between the infant and the mother. Despite the association between maternal depressive scores and the reduced quality of some of the items of early interaction, we would like to underline the variability of the quality of early interaction between the depressed mothers and their children. Differences in some aspects of interaction were slight and may be interpreted as indicative. Nevertheless, the level of maternal depression contributes to the quality of mutual interaction.