3.4. Differences in Food Intakes for Women ‘at risk’ and ‘not at risk’ of Developing post Natal Depression
There were no differences in the LC n-3 PUFA intakes for women ‘at risk’ and ‘not at risk’ of developing PND. There were no differences in grams per day for intakes for eggs, fish/seafood, meat and enriched food consumption between women ‘at risk’ and women ‘not at risk’ of PND (results not shown). There were no significant differences in intakes for EPA, DPA, DHA or total LC n-3 PUFA between the ‘at risk’ and the ‘not at risk’ groups. Even though this table suggests ‘at risk’ women had 20% higher DHA intakes than ‘not at risk’ women, this is due to one outlier in each of the two groups. These outliers consumed 588mg DHA (from ‘not at risk’ group) and 632mg DHA (from ‘at risk’ group). When these two outliers were deleted from the analysis, the mean DHA intakes in the ‘not at risk’ group and the ‘at risk’ group became 119mg and 118mg, respectively. The differences in LC n-3 PUFA intakes for women ‘at risk’ and ‘not at risk’ of developing postnatal depression is shown in Table 4
Comparison of LC n-3 PUFA intakes between pregnant women ‘not at risk’ and ‘at risk’ of developing PND as assessed by the Edinburgh Postnatal Depression Scale 1.
Comparison of LC n-3 PUFA intakes between pregnant women ‘not at risk’ and ‘at risk’ of developing PND as assessed by the Edinburgh Postnatal Depression Scale 1.
|Fatty acid||Women ‘not at risk’ of developing PND (n 76)||Women ‘at risk’ of developing PND (n 18)||P value*|
| ||Mean||SEM||Median||Mean||SEM||Median|| |
|LC n-3 PUFA mg/day||294||21||270||344||60||253||0.77|
There were no differences in EPA, DHA and LC n-3 PUFA intakes as assessed by the LC n-3 PUFA questionnaire between pregnant and non-pregnant women, however DPA intakes were significantly higher in pregnant women (Table 2
). Oily fleshed fish and seafood generally, particularly when harvested from cold-water regions, are a rich source of EPA and DHA [35
], whereas meat products, such as beef and lamb, contain a much higher DPA content [1
]. As the pregnant women consumed more meat than the non-pregnant women this would explain the differences in DPA intake between the two groups. In addition 5 of the 33 non-pregnant women in the study were vegetarians, which would contribute to the lower DPA intakes, whereas there were no vegetarians in the pregnant women group. However, the majority of the vegetarians (4 out of 5) still ate fish and eggs and a further analysis excluding the vegetarians still resulted in lower DPA intakes in the non-pregnant group (data not shown). Nevertheless, the overall intake of total LC n-3 PUFA was not significantly different between the pregnant and non-pregnant women groups.
The non-pregnant women group included eight women up to 14 years older than the pregnant women, therefore an analysis excluding all women aged > 41, was conducted. These results were the same as those found including the older women (data not shown). Furthermore, estimates of the LC n-3 PUFA dietary intakes in both the pregnant and non-pregnant women compared favourably with the results of the NNS95 (Table 2
). However, given that LC n-3 PUFA consumption is skewed to the right, comparisons of median intakes from both the pregnant and non-pregnant groups to the mean intakes in the NNS95 were not significantly different. Hence our results indicate that pregnant women in this study do not necessarily consume a higher intake of LC n-3 PUFA compared to non-pregnant women.
To our knowledge there are no other studies that have compared individual LC n-3 PUFA intake between pregnant and non-pregnant women. However, two previous longitudinal Dutch studies that reported the total n-3 PUFA (including ALA) of women during pregnancy found no significant changes in their dietary intakes pre and post-pregnancy [36
]. The Otto et al.
] study measured the amount of n-3 PUFA from a group of women pre-conception and then later at 10 weeks of pregnancy. While the Al et al.
] study (1996) measured intakes in women within the first 13 weeks, at 22 weeks and again at 32 weeks of pregnancy and found no differences in LC n-3 PUFA intakes.
Our results also show that there were no differences between the amount of daily LC n-3 PUFA consumed between women ‘at risk’ and ‘not at risk’ of developing PND groups (Table 4
). To date there are no other studies that have used a validated FFQ to compare the LC n-3 PUFA intakes from foods such as meat, seafood, eggs, n-3 PUFA enriched foods and fish oil supplements in women ‘at risk’ and ‘not at risk’ of developing PND. Even so these findings are consistent with a previous study between non-pregnant women aged 23–97 years categorised as depressed or non-depressed [38
] where they found no difference in LC n-3 PUFA intakes. However, one limitation of this study is the limited number of women at risk of developing PND available in the study. It is likely that women with depression may have been less likely to participate in a study or that some of the women screened as ‘not at risk of developing depression’ at the time of the study recruitment, later developed depression [33
]. Unfortunately, data on the women who later developed postnatal depression was not collected, as a second EPDS is not routinely administered, at the hospital, to women after they have given birth. Therefore it is important for hospital staff to screen all pregnant women during an antenatal visit to identify possible cases that maybe ‘at risk’ of developing PND. A lower EPDS cut-off score (≥10) was used by the hospital to identify all possible cases of depression. Using a cut-off score ≥10 as opposed to ≥13 or ≥15 will increase the sensitivity of the EPDS to reduce the number of false negatives but may increase the number of false positives [31
]. A validation study of the EPDS during pregnancy reported that a 14/15 cut-off score correctly identified all women with major depression (sensitivity 100%). Whereas using a lower 12/13 cut-off score identified the women with major depression and minor depression but also included some false positives [32
]. The hospital, in this study, intentionally used a lower EPDS cut-off score (≥10) as a safety net to ensure that no women with antenatal depression would be missed. Therefore, women who scores 10 or greater on the EPDS were regarded as ‘at risk’ of developing PND and were referred to a local community centre for counselling. Changing the cut-off scores to ≥13 and ≥15 to minimize false positives still did not produce any differences in LC n-3 PUFA intakes between the groups ‘at risk’ or ‘not at risk’ of developing PND (results not shown).
In this sample, the % rate of women ‘at risk’ of developing PND, using an EPDS cut-off score of ≥10, ≥13 and ≥15 were 19%, 13% and 12%, respectively, with a mean EPDS score 6.2 (SD 5.2). In comparison the Australian beyondblue
Postnatal Depression program [39
] where 40,333 pregnant women were screened for depression using the same EPDS cut-off scores was 20%, 9% and 5%, respectively, with a mean EPDS score 6.6 (SD 4.4). However, the pregnant women from our study deemed ‘at risk’ was screened at an average of 20 weeks gestation compared to the beyondblue
PND program where women were screened at an average of 26 weeks gestation [39
]. However, 52% of women in the beyondblue
PND program were recruited from the state of New South Wales. They were screened at an average of 16 weeks and % rates of antenatal depression, using a EPDS cut-off score of ≥10, ≥13 and ≥15, were 17%, 6% and 4%, respectively, mean score 7.5 (SD 3.9) [39
]. Our results show that our pregnant women screened for antenatal depression, had a similar prevalence rate to the beyondblue
PND program when using the EPDS cut-off score of ≥10 [39
]. Therefore our small study sample is a representative sample as the EPDS cut-off score of >
10 was used.
For this study, a difference of 200mg/day of LC n-3 PUFA was used in a power calculation to detect a significant difference between the women ‘at risk’ and ‘not at risk’ of developing PND. Therefore a sample size of 75 women was required for each group. While this was achieved for the ‘not at risk’ of developing PND group (n = 76) the ‘at risk’ of developing PND group contained only 18 women. A further power calculation for DHA alone (with a mean of 200mg/day and a standard deviation of 95 as found in our data results) showed that if the ‘at risk’ of developing PND group did contain 75 women the intakes would still not be significant. To detect a significant difference the sample size required would require at least 710 women total. Furthermore, Table 4
suggests that intakes of DHA in women ‘at risk’ of developing PND is higher than women ‘not at risk’ of developing PND. Therefore if this study had the statistical power, the results would not support our hypothesis.
Whether or not a difference of 200mg/day of DHA would have any clinical affect between women ‘at risk’ for depression or ‘not at risk’ for depression is still unclear. For this study, although not significantly different, the ‘not at risk’ women consumed 10 mg less DHA (96mg DHA) compared with the ‘at risk’ women who consumed 106mg DHA (Table 4
). Whereas, Koletzko et al.
] and The International Society for the Study of Fatty Acids and Lipids (ISSFAL) [40
] recommend during pregnancy at least 200mg/day of DHA alone, the median DHA intake found in our study and the NNS95 (83 mg/day) are much lower than this.
If DHA intakes are no different in women ‘at risk’ versus ‘not at risk’ of developing PND, then other factors may be considered. This includes metabolic abnormality in lipid metabolism or oxidative damage caused by smoking [42
]. Furthermore, social and economic factors may also contribute to the risk of developing depression [43
There are various recommendations for consumption of LC n-3 PUFA and they range from 135 to 1600mg [20
]. In Australia the National Health and Medical Research Council (NHMRC) Nutrient Reference Values, and recommend an adequate intake (AI) for LC n-3 PUFA at 160mg/day for adult men and 90mg/day for adult women [44
] but these recommendations are based on median/mean intakes of the Australian population [1
]. In pregnancy these AI recommendations are higher at 110mg/day for 14-18 year olds and 115mg/day for 19-50 year olds [44
]. The NHMRC also have Suggested Dietary Target (SDT) intakes for optimal health and the SDT for LC n-3 PUFA is 610mg/day for men and 430mg/day for women [44
] and these are based on 90th
percentile of Australian population intakes [1
]. Whilst it is difficult to establish an exact amount of DHA needed to supply the needs of the growing and developing foetus and replenish the mother’s maternal stores [10
], there are recommendations specifically for DHA. The ISSFAL proposed an adequate intake for pregnant women to consume > 200mg/day [40
] and the European consensus statement proposed at least 200mg/day of DHA alone [24
In this study the median intake shows that 50% of the women consumed an LC n-3 PUFA and DHA intake less than 268mg/day and 96mg/day, respectively. Furthermore, only 9% of the women consumed the proposed ideal recommended intake of > 200mg/day of DHA during pregnancy and lactation [40
]. If the foetus accumulates on average » 67 mg DHA/day from the maternal supply during the last trimester of pregnancy [45
], as predicted from post-mortem studies [46
], then the mother will have very little or no DHA left over to provide for her own physiological needs [10
]. Maternal plasma concentrations of a range of fatty acids are increased during pregnancy [12
] among these are the LC n-3 PUFA. There is no evidence of increased maternal LC n-3 PUFA intake during pregnancy [36
] so it is likely that the majority of these LC n-3 PUFA are mobilised from maternal stores with some contribution from oestrogen stimulated conversion of ALA to DHA [18
]. Interestingly the growing foetus preferentially takes up DHA and the omega-6 PUFA called arachidonic acid (AA, 20:4n-6) and these two fatty acids become important building blocks for the foetal brain [12
]. There is some evidence that women during pregnancy become deficient in DHA and as pregnancy continues their DHA deficiency index (as defined by the 22:5n-6/22:4n-6 ratio) increases during pregnancy [12
]. This DHA deficiency returns to normal 6 months post-partum [47
]. However this is not observed in all populations [47
]. It should be noted that it is difficult to deduce net transfer of fatty acids from mother to foetus from random measurement of maternal steady state plasma concentrations. Nonetheless, the high foetal demand and absolute requirement for DHA would suggest that pregnant women are under stress in terms of providing enough DHA to her growing foetus. A recent study reported that pregnant women lack the knowledge of the importance of LC n-3 PUFA consumption during pregnancy [48
]. Therefore, there is obviously a need to educate women of childbearing age about the benefits of increasing their LC n-3 PUFA intakes [48
] and the best alternatives are by consuming more fish products [20
There are various reports on actual intakes for women of reproductive age as outlined by Sioen et al.
]. The actual intakes of EPA range from 46 to 278mg/day, DPA intakes range from 10 to 75mg/day, DHA intakes range from 67 to 497mg/day and LC n-3 PUFA intakes range from 156–775mg/day [49
]. For our study, both pregnant and non-pregnant women, the average total LC n-3 PUFA intake of 263mg/day and 241mg/day, respectively, meets the AI recommendations of the NHMRC but falls short of the NHMRC SDT recommendations and the recently released National Heart Foundation of Australia’s recommended EPA and DHA intake of 500mg/day for healthy adults including pregnant women [50
] as well as the recommendations by ISSFAL [40
] and the European consensus [24
The results of our study showed no difference between the average amount of seafood, meat, eggs and enriched foods (grams per day) consumed between pregnant women ‘at risk’ of developing PND and those ‘not at risk’ of developing PND. In contrast to our study, a cross-sectional national New Zealand study of 4644 adult men and women (15+ years) reported that non-fish eaters scored significantly lower on a mental health test (poorer mental health) compared to the fish eaters’ on the same mental health test (better mental health) [51
]. Unlike our study, fish consumption was categorised into those who consumed no fish of any kind and those who consumed some kind of fish (ranging from ≥1 fish serving per month to ≥ 2 per day) [51
]. However, the National New Zealand study did not include and compare other foods that contain LC n-3 PUFA such as meat or enriched foods.
From the current study, 51% of the pregnant women were supplementing their diet with LC n-3 PUFA enriched foods such as milk, bread, eggs and fat spreads and/or fish oil supplements (Table 3
). This contributed to a total of 12% of their daily LC n-3 PUFA intakes; 5% EPA, 1% DPA, 22% DHA. Enriched foods can be very useful to top-up daily levels of LC n-3 PUFA when other enriched foods like fish and meat have already been incorporated into the diet [52
]. A recent study [53
] found that a group of overweight subjects with a habitual low fish intake could significantly increase their daily LC n-3 PUFA intake from 258mg to 1195 mg/day by incorporating up to eight various LC n-3 PUFA enriched foods into their diet. This same study showed that consumption of these enriched foods resulted in increased erythrocyte LC n-3 PUFA levels to a similar extent than if the subjects consumed fish/seafood [54
A limitation of this study was the lack of blood biomarkers from the women, such as erythrocyte LC n-3 PUFA levels which are indicative of tissue LC n-3 PUFA levels. However, a previous study found using the same FFQ that 53 healthy adults reported LC n-3 PUFA intakes parallel levels analysed in erythrocytes and plasma and these intakes [27
] were also comparable to those reported in the NNS95. Another limitation of our current study is the comparability of the pregnant and non-pregnant women. Pregnant women were recruited from a hospital, whilst the non-pregnant women were recruited from the University of Wollongong and the wider Wollongong community. Whilst it is recognised that women with a higher level of education or income were more aware of issues related to LC n-3 PUFA [48
], however, the results of our current study compare favourably to the intakes of women assessed in the NNS95, which covers a wide range of education and income levels. It must also be noted there was a difference in methodologies used between this study (FFQ including enriched foods) and the NNS95 (24 hr food recall and FFQ) [30
]. As well a difference of 9 years between this study (data collected 2004) and the NNS95 and within that time frame the availability of LC n-3 PUFA foods and the dietary habits of individuals have changed, especially with the introduction of LC n-3 PUFA enriched foods. For that reason we compared the LC n-3 PUFA intakes of pregnant and non-pregnant women, excluding the enriched foods, and their intakes compared favourably to those intakes of women reported in the NNS95. Furthermore due to limited time for subject recruitment 75 women for each of the groups was not achieved as only 18 women ‘at risk’ of developing PND and 33 non-pregnant women, compared to 76 women ‘not at risk’ of developing PND were recruited for this study. However, the data were also compared to that of the NNS95 which contained 5770 women.
In summary the LC n-3 PUFA intakes did not differ between pregnant women and non-pregnant women and were comparable to the intakes as recorded by the Australian NNS95. In our study, the average total LC n-3 PUFA intake for both pregnant and non-pregnant women meet the recommendations of some but not all recommended ranges for healthy adults. The LC n-3 PUFA intakes did not differ between the pregnant women ‘at risk’ of developing PND and ‘not at risk’ of developing PND. Although there is an impression that inadequate LC n-3 PUFA intakes in pregnant women may predispose them to developing postnatal depression due to growing foetal organ systems (particularly the nervous systems) requiring increased amounts of LC n-3 PUFA [23
], the results of this study do not currently support this deficiency concept.