Study design and participants

This is a multicenter two-phase observational prospective cohort study conducted from June 2020 (one month after the first lockdown in Greece) to August 2020, involving 5 hospitals, throughout Greece. The study was approved by the ethics committees of all hospitals, and it was registered on clinicaltrials.gov. All procedures were in accordance with the World Medical Association Helsinki Declaration and signed informed consent was obtained from all participants.

In the first phase of the study, all parturients presenting for labor during June 2020, were asked to participate. Exclusion criteria were inability to read and write in Greek, history of an acute psychotic episode, intoxication or confusion. Each local investigator explained the aims of the study and a written informed consent was signed. The set of Phase 1 questionnaires (Generalized Anxiety Disorder-GAD-7 and a self-reported questionnaire about COVID-19 related concerns) was administered to all parturients. Demographic and medical variables were also recorded from the medical records. The second phase was conducted at 6–8 weeks postpartum by telephone and all women were asked to complete the Edinburgh Postnatal Depression Scale (EPDS). This is because PPD usually manifests at 6–8 weeks postpartum and in order to avoid the confounding impact of maternity blues (30–80% during the first 15 days postpartum). If a woman recorded suicidal ideation or a high score in the study tools, she was referred to a mental health service. Also, they were informed about the healthcare services and help lines operated during the pandemic.

Measures and study instruments

Sociodemographic and pregnancy variables were collected through a questionnaire and medical records. COVID-19 related concerns were collected using a 38-items structured questionnaire grouped in 5 domains (pregnancy, family, adequacy of information, obstetrics/gynecology, and anesthetic concerns). Fourteen items were dichotomous (Y/N) or with multiple responses (e.g. ‘where did you get information regarding viral transmission to the baby’ with possible answers such as physician, media, social media, other) and 24 items were scored on a 10-point Likert scale (0–9, in general 0 refers to ‘not at all’ and 9 to ‘extremely’, e.g. how much are you concerned about your pregnancy due to pandemic?). The questionnaire was developed specifically for this study after careful examination of the available regarding common concerns and anxieties among pregnant women. Subsequently all items were reviewed by all authors based on their expertise, specialty, and empirical knowledge at the first wave of pandemic in Greece, and consensus was established^17-25. Two parturients that were not included in the study sample were involved in this research by means of assisting in developing these research questions and in their prioritization in this structured questionnaire. Reliability analysis revealed a good strength of associations (Cronbach alpha=0.78).

PPD was assessed with the EPDS²⁶, which is a 10-item self-report tool. Mothers were asked to check the response that is closest to how they have been feeling in the last 7 days in a 4-point Likert scale (0–3; score range: 0–30). PPD was defined as an EPDS score of ≥13. The reliability of the scale was high (Cronbach’s alpha=0.87).

Anxiety symptoms were assessed by the Greek version of GAD-7 which rates the frequency of anxiety symptoms in the last 2 weeks on a Likert scale ranging from 0–3 (‘not at all’ to ‘nearly every day’) with a standard cut-off point of ≥10 indicating moderate to severe anxiety symptoms²⁷.

Statistical analysis

All statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS) version 26 for Mac. Summary statistics for all variables were calculated using means and standard deviation for continuous variables and frequency and proportions for categorical variables. The primary outcome variable was the EPDS score at 6–8 weeks postpartum. To assess the association of COVID-19 related concerns with anxiety and depressive symptoms, Pearson’s correlations tests were carried out, followed by partial correlations adjusted for age, demographic and obstetric variables. To assess the independent associations of COVID-19 related concerns with anxiety and depression, two separate binary logistic regression analyses were performed with dependent variable the GAD-7 score (<10 and ≥10) and EPDS score (<13 and ≥13), respectively. Independent variables in each analysis were the most significant associations of COVID-19-related concerns derived from the correlation analysis.

Demographic characteristics of the study sample

In total, 339 parturients enrolled in the study. There was a 2.65% dropout rate between the first and second phase of the study (9 participants were not located at 6 weeks). Finally, the sample consisted of 330 women with a mean age of 32.3 years (SD: 5.6) [maternal age <35, (n=207; 62.7%) and maternal age ≥35 (n=123; 37.3%)], at a gestational age of 38 weeks (SD: 1.9). The majority were married (n=300; 90.6%), 63.1% were employed (n=209) and 85.5% (n=282) had completed secondary education (high school). Labor was conducted via caesarean (n=211; 64%) or vaginal (n=119; 36%) delivery. Preterm deliveries accounted for 10% of the study sample (n=33) and term deliveries for 90% (n=297).

Prevalence of postnatal depression and generalized anxiety disorder

In total, 43 mothers scored 13 or above in EPDS indicating that the prevalence of MDD (Major Depressive Disorder) was 13.2%. As far as anxiety symptoms are concerned, the mean score of GAD-7 before delivery was 6.3±5.0 indicating mild anxiety for parturients, while one-fourth (24.8%; n=82) could be identified as having probable, generalized anxiety disorder, when a cut-off point of ≥10 was used.

Concerns about COVID-19

Figure 1 shows the mean scores for the items related to COVID-19 concerns. Concerns about pregnancy and pandemic and transmission to fetus/newborn were moderate. They also seemed quite worried about the health of older children (5.4±3.2). Moreover, antenatal care seemed mildly affected due to the pandemic; however, a proportion of 36% (n=119) reported a decrease in willingness to attend antenatal education courses and a similar proportion of 34% (n=112) reported fewer antenatal visits to an obstetrician/gynecologist than planned. Also, mild concerns about COVID-19 transmission were reported because of their contact with the anesthesiologist as a first line doctor.

Figure 1

Questions (sorted in relation to pregnancy, newborn, family, information on COVID-19, obstetrics and anesthesiologist) related to concerns about COVID-19 and mean values

Correlations between COVID-19 concerns and anxiety symptoms

In univariate analysis (Table 1), a number of variables were associated with anxiety symptoms. A belief about transmission of the virus to the baby was positively correlated with anxiety score. Anxiety was significantly associated with the impact of COVID-19 outbreak on family relationships while the covariates did not affect this relationship. Parturients were anxious when their daily contact with their husband was affected by his work during the lockdown. As far as information related to transmission, our results revealed a negative relationship with anxiety, indicating that the better the information the less the anxiety. Finally, antenatal care as reflected in willingness to attend childbirth classes and prenatal visits was related to the GAD-7 total score. Partial correlation showed that certain associations were affected by the covariates like parturients’ age and gestational age (Table 1).

Separate univariate analysis (Table 2) showed that parturients aged ≥35 years were more anxious when they believed that the virus could be transmitted to the baby and less anxious when they were well informed about fetal transmission. The latter was also observed in parturients aged <35 years. Even so, younger women were more anxious if their daily contact with their husband was affected during the lockdown. In respect to gestational age, the anxiety symptoms were positively associated with their isolation from their partner due to his working during the lockdown only in preterm pregnancies. The same association was observed in term pregnancies along with the belief about viral transmission to the baby. Finally, in term pregnancies, a negative correlation was found between anxiety symptoms and the amount of information regarding the fetal transmission.

Subsequent regression analysis (Table 3) showed that the odds of being assessed with generalized anxiety disorder were higher in parturients who reported strong impact of pandemic on family relations (p=0.003, 95% CI: 1.060–1.316).

Correlations between COVID-19 concerns and depressive symptoms at 6–8 weeks postpartum

Univariable analysis in Table 4 showed that different COVID-19 related variables correlated with EPDS depressive symptoms severity. When various variables were considered, only three remained significant. Depressive symptoms were positively associated with the degree of the emotional impact to the parturient due to the change in her daily contact with her husband because of his work in lockdown (p=0.018). Moreover, there were positive correlations with the degree of the anxiety related to her contact with the anesthesiologist as a possible source of COVID-19 transmission (p=0.043) as well as the belief that the anesthesiologist, as a first line doctor, could transmit the infection to her and the baby (p=0.023). Subsequent regression analysis (Table 5) showed that the odds of having a postnatal major depressive episode were higher only in parturients that reported higher levels of anxiety (p=0.000; 95% CI: 1.063–1.221).

Main findings

In our study, the prevalence of PPD at 6 weeks postpartum was 13.2%. Data from Greece regarding the prevalence of PPD prior to the pandemic have revealed similar rates^28-30. Gonidakis et al.²⁹ found that the prevalence of PPD one month after delivery was 12.5% using the cut-off point of 12. Similarly, Koutra et al.³⁰, using the same cut-off with our study (≥13), found that the prevalence of PPD was 13.6% at 8 weeks postpartum. Moreover, Leonardou et al.²⁸ used the cut-off point of 11/12 and found that the prevalence of PPD was 12.4%. Vivilaki et al.³¹ in 2007–2008, in a sample of postpartum women, reported that 6.7% of mothers diagnosed with major depression, 11.7% with moderate and 31.7% with mild depression according to BDI-II scores. Therefore, the prevalence of PPD during the first phase of COVID-19 pandemic was not increased, most probably because Greece was not hit hard by the first pandemic wave. Confirmed cases during the study period (June 2020–August 2020) ranged from 2.937 on 1 June to 10.757 on 1 September; the latter represents a rate near to 1% of the total population. Moreover, the number of deaths were low compared to other countries (n=271 on 1 September 2020). Moreover, our study was conducted over a period where no strict restrictive measures were implemented. Although a number of studies report an increase in PPD, a recent meta-analysis showed that even though the EPDS score was higher in the pandemic compared to the nonpandemic period, it did not reach the level of statistical significance^12-14,32. In contrast, parturients reported high antenatal levels of anxiety, since 24.8% of them scored ≥10 in GAD-7. A recent meta-analysis, before the pandemic showed that self-reported anxiety symptoms prevalence was 24.6% in the third trimester and the overall prevalence of any anxiety disorder was 15.2%¹⁹. During the pandemic, two studies using GAD-7 reported lower rates compared to our study, of moderate anxiety (14% and 21.6%) in countries hit hard by COVID-19^13,33. Similarly, in a multinational study of 3097 pregnant women, 11% scored ≥10 in GAD-7¹³. Exception to these rates is the study of Preis et al.³³ which revealed substantial levels of moderate to severe anxiety up to 43.3%. Evidence suggests that prenatal anxiety increases with the severity of the measures imposed. Interestingly, our recruitment was undertaken in a period where strict restrictive measures were not applied since Greece was not hit so hard in the first pandemic wave. One explanation is that family and social support during pregnancy in Greek families is very high and this was disrupted during lockdown, right before recruitment. Additionally, parturients were not allowed to have companions in the antenatal visits and in the hospital before delivery. By tradition, in Greece parturients have frequent visits to the hospital, and social support is associated with increased mental health and emotional well-being³⁴. Still, we cannot ignore that anxiety measurement (even if GAD refers to the last 2 weeks) was undertaken right before delivery and this per se represents a stressful moment.

Our findings suggest that parturients were more anxious when they believed that SARS-CoV-2 virus could be transmitted by them to their baby. Anxiety could be due to lack of solid scientific evidence about transplacental transmission at the time the study was conducted³⁵. However, even if the evidence was against transmission, parturients could be influenced by the so-called infodemic phenomenon. Moreover, antenatal care was correlated with GAD-7 scores probably because of the fear of transmission. Finally, anxiety symptoms were associated with the decreased contact with the partner, probably due to lack of emotional support.

In our study, anxiety of advanced maternal age mothers was correlated with the belief that the virus could be transmitted to the baby, something that was not observed in the younger group. Advanced maternal age and the fear of related complications in pregnancy could be the explanation. In contrast, younger mothers were anxious when they have lost their daily contact with their husband. Anxiety of preterm pregnancy mothers (in contrast with term pregnancies) is not associated with beliefs about transmission to the baby probably because their primary source of anxiety is the survival or the complications to the newborn due to the preterm pregnancy itself^36,37. In general, little is known about predictive risk factors for anxiety disorders in pregnancy and postnatal period, being a neglected area of research^38-42. In our study during the pandemic, parturients being at higher risk for generalized anxiety disorder were those whose family relations were affected the most. This is indicative of the importance of marital and social support during pregnancy that a woman calls on during stressful events⁴³.

Depressive symptoms were correlated with various COVID-19 concerns. The PPD was positively associated with the reduced daily contact with the husband and with the fear that anesthesiologists could be a possible source of transmission. Previous studies related to SARS outbreak showed that the relationship with the partner was compromised due to interruption of intimacy and fear of transmission^34,44. Additionally, women who experience higher levels of support from their partners during the second trimester of pregnancy present lower levels of psychological implications postnatally⁴⁴. Therefore, it comes as no surprise that COVID-19 outbreak preventive measures related to partners’ distancing implemented in our study population could predispose to PPD. Moreover, anesthesiologists who come in close contact with severely ill patients are exposed to airborne viral transmission and are among the medical professionals of highest risk of infection⁴⁵. Furthermore, their contribution to the current pandemic management is vastly praised by the social media and the public⁴⁶. Considering that the public is overexposed to social media and TV news as a source of information during this pandemic⁴⁷, parturients were directly exposed to the infodemic phenomenon. However, concerning predictive factors, in our study only high levels of anxiety were associated with postnatal depression, a finding consistent with the literature^48-50.

Limitations

Our study has some limitations that need to be addressed. One lies in its cross-sectional nature. Yet it involves 5 centers throughout Greece (three located in Athens where half of the total population lives). Moreover, it assesses depressive symptoms at 6 weeks postpartum in order to avoid the ‘baby blues’ period. Although we used the most common PPD tool, we did not confirm the diagnosis using a structured interview. Another limitation is the fact that the 6 weeks’ postpartum assessment was performed by telephone contact where women underreport their symptoms. However, there are reports that EPDS administration over the telephone has high reliability⁵¹. Finally, the study was conducted during the first pandemic wave where COVID-19 confirmed cases and deaths were not high in Greece. However, to the best of our knowledge, this is the first study in Greece assessing the prevalence of PPD during the COVID-19 pandemic.

Future research

Unfortunately, the pandemic is still on the rise and new mutations appear that may be a trigger for psychological distress for the population and of course for pregnant women. Future longitudinal research is needed to better identify high risk parturients and to provide the best maternal and child care both prenatal and postnatal. Future research should focus on psychosocial features such as the general support to pregnant women, e.g. marital and social support during pregnancy as well as medical staff support just after the labor, with well validated tools. Finally, since countries do not have the same infection rate throughout their region (and in different times this was the case in Greece), it is important to conduct further prevalence and longitudinal studies (throughout pregnancy) with larger samples in different regions to identify the effect of covid-infection rate in maternal psychological distress.