Discussion

Among patients with pre-existing CVD in pregnancy, the overall prevalence of probable clinical symptoms was 10.4–32.2% for post-pregnancy depression, 22.0–28.5% for post-pregnancy anxiety, and 26.4% for mild or 18.2% for moderate post-pregnancy PTSD symptoms. Post-pregnancy depression was more commonly reported (seven studies), and post-pregnancy anxiety and PTSD were rarely reported (two studies and one study, respectively). Regarding post-pregnancy depression, the EPDS was the most used scale for determining probable clinical depression. The HADS and DASS-21 were both used for post-pregnancy anxiety symptom assessment and the IES-R was only used for post-pregnancy PTSD symptom assessment. Of the two studies investigating the association between pre-existing CVD in pregnancy and PPD symptoms or diagnosis, neither reported a significant association either initially ^{Reference Katon, Russo and Gavin56} or after adjustment for covariates. ^{Reference Zipursky, Thiruchelvam and Redelmeier60} Additionally, one study reported no association between previous heart failure, heart disease or atrial fibrillation and hospital admission for PPD after adjustment for covariates. ^{Reference Abu-Zaid, Gari, Alsharif, Alshahrani, Khadawardi and Ahmed57} No studies investigated the association between pre-existing CVD in pregnancy and post-pregnancy anxiety or PTSD symptoms. Of the three studies reporting associations between WHO risk classification scores and PPD symptoms, none reported significant associations. ^{Reference Panelli, Sherwin, Lee, Leonard, Miller and Miller54,Reference Andrade, Metz, Martin, Cox, Ou and Hammad55,Reference Freiberger, Beckmann, Freilinger, Kaemmerer, Huber and Nagdyman58} One study investigated the correlation between WHO risk classification scores and post-pregnancy anxiety and PTSD symptoms, finding no association for anxiety and a significant association for PTSD symptoms. ^{Reference Freiberger, Beckmann, Freilinger, Kaemmerer, Huber and Nagdyman58}

Our findings suggest that the prevalence of post-pregnancy depressive symptoms among individuals with pre-existing CVD may be comparable to the general postpartum population, although interpretation is limited by the small number of studies included in the review. Most studies (n = 3) reported PPD symptom prevalence to be 14–19%, which is similar to global estimates of PPD affecting 17% of postpartum individuals. ^{Reference Wang, Liu, Shuai, Cai, Fu and Liu23,Reference Panelli, Sherwin, Lee, Leonard, Miller and Miller54,Reference Andrade, Metz, Martin, Cox, Ou and Hammad55,Reference Hutchens, Frawley and Sullivan59} In contrast, studies of high-risk pregnancies characterised by obstetric complications such as preeclampsia have reported the prevalence of PPD symptoms to be 27–44%. ^{Reference Brandon, Trivedi, Hynan, Miltenberger, Labat and Rifkin20–Reference Hoedjes, Berks, Vogel, Franx, Bangma and Darlington22} Two studies in this review, however, reported PPD symptom prevalences not consistent with estimates in the general population, with one reporting a lower prevalence (10.4%) and one reporting a higher prevalence (32.2%). ^{Reference Katon, Russo and Gavin56,Reference Freiberger, Beckmann, Freilinger, Kaemmerer, Huber and Nagdyman58} The wide range in PPD symptom prevalence in this review is likely driven by substantial heterogeneity in study design, including differences in timing of assessment, screening tool utilised and classification of CVD.

The timing of measurement in relation to delivery may explain some of the variation in PPD symptom prevalence. Prevalence lowest (10.4%) in the only study that measured PPD symptoms prospectively within 6 weeks postpartum. ^{Reference Katon, Russo and Gavin56} In contrast, prevalence was highest (32.2%) in the study that assessed depressive symptoms retrospectively to the postpartum period an average of 11 years after delivery among individuals with CHD. ^{Reference Freiberger, Beckmann, Freilinger, Kaemmerer, Huber and Nagdyman58} Notably, the prevalence of depressive symptoms at the time of measurement was much lower (15%) in the same study. ^{Reference Freiberger, Beckmann, Freilinger, Kaemmerer, Huber and Nagdyman58} Evidence suggests that recalling depressive episodes is often inaccurate; only 40% of respondents with an episode of depression 10 years prior were able to accurately recall the episode in a study of non-pregnant individuals. ^{Reference Patten, Williams, Lavorato, Bulloch, D’Arcy and Streiner61} Thus, this discrepancy in depressive symptom prevalence may be driven by recall bias. Although Hutchens et al’s study also measured depressive symptoms in the years following delivery, participants only reported symptoms at the time of measurement, with 14% of those with CHD screening positive for depression. ^{Reference Hutchens, Frawley and Sullivan59} Overall, the prevalence of PPD symptoms was highest when assessed retrospectively over a decade postpartum, followed by studies that retrospectively assessed PPD symptoms within the immediate postpartum period. Future studies should ideally implore a prospective cohort study design to accurately assess PPD symptom prevalence within 6 weeks postpartum, which is the recommended time frame to screen for PPD. ¹⁴

Measures used to screen for depressive symptoms varied across the studies, which also may have contributed to the variation in probable depression prevalence. Depressive symptoms were highest (18–32%) among the two studies utilising the EPDS, ^{Reference Panelli, Sherwin, Lee, Leonard, Miller and Miller54,Reference Freiberger, Beckmann, Freilinger, Kaemmerer, Huber and Nagdyman58} followed by the study utilising the DASS-21 to assess current depressive symptoms several years postpartum (14%). ^{Reference Hutchens, Frawley and Sullivan59} The study utilising the PHQ-9 reported the lowest prevalence of PPD symptoms (10.4%). ^{Reference Katon, Russo and Gavin56} Conceptual differences in these screening tools may explain some of these differences. A network analysis study of over 2000 pregnant people found that the EPDS emphasises psychological symptoms (i.e. feeling sad or miserable) whereas the PHQ-9 prioritises the physical symptoms of depression (i.e. fatigue). ^{Reference Liu, Li, Wang, Yin, Xiao and Hu62} Those with CVD may attribute their physical symptoms, such as fatigue, to their CVD rather than to their depression, which could potentially influence symptom reporting. However, no literature to date has investigated how measurement of depression differs across screening tools in individuals with CVD, making it difficult to understand whether this influenced our findings. In the perinatal population, however, evidence suggests that these instruments operate similarly in screening for PPD. A recent systematic review reported nearly identical operating characteristics between the PHQ-9 and EPDS in detecting perinatal depression. ^{Reference Wang, Liu, Shuai, Cai, Fu and Liu23} Similarly, the area under the curve (AUC) for the DASS-21 scale in identifying current depressive episode compared to DSM-IV criteria has been found to be similar to the AUC for the EPDS, at 0.76 v. 0.77, respectively. ^{Reference Moya, Larson, Stewart, Fisher, Mwangi and Phiri63} Therefore, the existing literature suggests that the PHQ-9, EPDS and DASS-21 similarly detect postpartum depressive symptoms.

Differences in screening thresholds between the two studies utilising the EPDS may also explain the variation in PPD symptom prevalence. In Freiberger et al’s study assessing PPD symptoms retrospectively a mean of 11 years postpartum, a cut-off of ≥9 was used as suggested by validation studies conducted among non-postpartum samples. ^{Reference Freiberger, Beckmann, Freilinger, Kaemmerer, Huber and Nagdyman58} In contrast, Panelli et al’s study assessed depressive symptoms within 6 weeks postpartum and thus utilised the validated cut-off of ≥10 to screen for PPD. ^{Reference Panelli, Sherwin, Lee, Leonard, Miller and Miller54} As the prevalence of probable PPD was much higher in Freiberger et al’s study (32.2%) compared with Panelli et al’s study (18%), the lower cut-off value may explain why more cases of probable depression were detected. Current guidelines suggest utilising either the EPDS or the PHQ-9 to screen for PPD. ¹⁴ These screening instruments have been found to have nearly identical operating characteristics, with similar sensitivities and specificities tools. ^{Reference Wang, Liu, Shuai, Cai, Fu and Liu23,Reference Levis, Negeri, Sun, Benedetti and Thombs64} Therefore, future studies investigating PPD after pregnancies complicated by pre-existing CVD should employ either the EPDS or the PHQ-9, using the cut-off values recommended for the study population.

Although all studies examined cases where CVD was pre-existing before pregnancy, the timing and nature of the diagnosis differed. In some studies, CVD was congenital and identified at birth or during childhood, whereas in others, it was acquired and diagnosed either years before pregnancy or just before conception. Additionally, some studies reported PPD symptom prevalences separately for congenital and acquired CVD, whereas others combined these conditions in their analyses. No clear pattern in post-pregnancy depressive symptom prevalence by CVD subtype emerged. Only one study compared the prevalence of depressive symptoms between those with congenital and pre-existing acquired diseases, reporting that 20% of those with congenital diseases screened positive for PPD symptoms compared with 15% of those with acquired conditions. ^{Reference Panelli, Sherwin, Lee, Leonard, Miller and Miller54} There are distinct psychosocial stressors that may be experienced by those with CHD or pre-existing acquired CVD during pregnancy. For instance, individuals with congenital CVD may experience anxiety about transmitting their condition to their offspring, whereas those with acquired CVD may not face the same concern. ^{Reference Hutchens, Frawley and Sullivan65} Conversely, individuals with acquired CVD may struggle with adapting to a new diagnosis, a challenge that has been observed in studies of chronic illness among non-pregnant adults. ^{Reference Stanton, Revenson and Tennen66} Since pregnancy is often a significant life stressor, navigating it alongside a recent CVD diagnosis may further heighten vulnerability to postpartum mental illness. Given the potential implications of CVD classification and timing of diagnosis on postpartum mental health, future research on PPD should differentiate between congenital and pre-existing acquired CVD.

Of the two studies investigating the association between pre-existing CVD and PPD symptoms, neither found a significant association in adjusted models. ^{Reference Katon, Russo and Gavin56,Reference Zipursky, Thiruchelvam and Redelmeier60} It is important to note, however, that one study adjusted for length of stay at delivery and number of clinic visits during pregnancy, both of which remained significantly associated with PPD symptoms in adjusted models. ^{Reference Zipursky, Thiruchelvam and Redelmeier60} These variables may reflect morbidity from CVD and thus may be mediators of the relationship between pre-existing CVD and PPD. For example, a study published using data from the National Inpatient Sample reported that pregnant people with CVD had significantly prolonged hospital stays compared with those without (3.9 v. 2.6 days, p < 0.001). ^{Reference Williamson, Altendahl, Martinez, Ng, Lin and Benharash67} Interestingly, another study in this review reported that participants with a positive PPD screen were significantly more likely to have had antepartum anticoagulation, blood transfusion and maternal–infant postpartum separation. ^{Reference Panelli, Sherwin, Lee, Leonard, Miller and Miller54} Therefore, future studies should consider indicators of CVD morbidity as potential mediators when assessing the association between pre-existing CVD and PPD. Although the second study assessing the association between pre-existing CVD and PPD symptoms reported null findings, the association approached significance (adjusted odds ratio 1.92, 95% CI 0.98–3.79, p = 0.06) and thus may be detected in studies with greater power (n = 67 with CVD in this study). ^{Reference Katon, Russo and Gavin56} Consistent with these findings, an additional study examining heart disease in pregnancy and hospital admission for PPD also reported no association. ^{Reference Abu-Zaid, Gari, Alsharif, Alshahrani, Khadawardi and Ahmed57} However, interpretation is limited because hospital admission for PPD is rare, occurring in fewer than 1% of cases, and therefore does not accurately reflect PPD morbidity. ^{Reference Hippman, Kaur, Carrion, Baetz, Buffam and Dyck68} Overall, more research is needed to determine if there is an association between pre-existing CVD during pregnancy and PPD. Future studies should ensure adequate statistical power with larger sample sizes and consider how markers of CVD morbidity may affect the relationship between pre-existing CVD and PPD. Additionally, no studies investigated the association between pre-existing CVD and post-pregnancy anxiety or PTSD, which warrants future research.

None of the three studies investigating WHO risk classification scores and PPD symptoms reported a significant association. ^{Reference Panelli, Sherwin, Lee, Leonard, Miller and Miller54,Reference Andrade, Metz, Martin, Cox, Ou and Hammad55,Reference Freiberger, Beckmann, Freilinger, Kaemmerer, Huber and Nagdyman58} Additionally, one study noted a trend toward higher WHO risk classification scores among those who screened positive for PPD, ^{Reference Panelli, Sherwin, Lee, Leonard, Miller and Miller54} which is in contrast to another study that reported a trend toward lower WHO risk classification scores among those who screened positive for PPD. ^{Reference Andrade, Metz, Martin, Cox, Ou and Hammad55} These conflicting results may be reflective of the limited number of participants with a WHO risk classification score of 4 (n = 21 total in the three studies). ^{1,Reference Panelli, Sherwin, Lee, Leonard, Miller and Miller54,Reference Andrade, Metz, Martin, Cox, Ou and Hammad55,Reference Freiberger, Beckmann, Freilinger, Kaemmerer, Huber and Nagdyman58} This category includes conditions contraindicated in pregnancy, such as pulmonary arterial hypertension, because of the extremely high risk of maternal mortality or severe morbidity during pregnancy, and current guidelines recommend pregnancy termination counselling for these patients. ¹ Future studies that include a greater number of participants with WHO risk classification scores 3 and 4 are needed to better understand the association between WHO risk classification scores and PPD.

Our post-pregnancy anxiety symptom prevalence are higher (24.8–28.5%) than the estimated global prevalence of postpartum anxiety of 12.3%, although estimates are limited to two studies. ^{Reference Freiberger, Beckmann, Freilinger, Kaemmerer, Huber and Nagdyman58,Reference Hutchens, Frawley and Sullivan59,Reference Feldman, Hibara, Ye, Macaranas, Larkin and Hendrix69} However, both studies assessed current anxiety symptoms an average of 5–11 years postpartum, indicating a gap in the literature on postpartum anxiety in this patient population. It is possible that our anxiety symptom prevalence results reflect higher baseline levels of anxiety found in individuals with both acquired and congenital CVD, which is often attributable to illness-related stressors. ^{Reference Karami, Kazeminia, Karami, Salimi, Ziapour and Janjani70} In both studies, the prevalence of current anxiety symptoms exceeded the prevalence of depressive symptoms, indicating that anxiety may be especially salient among those with pre-existing CVD. ^{Reference Freiberger, Beckmann, Freilinger, Kaemmerer, Huber and Nagdyman58,Reference Hutchens, Frawley and Sullivan59} Qualitative research among pregnant individuals with pre-existing CVD has reported that many experience fears of cardiac deterioration and anxiety about the potential impact of their CVD on fetal health. ^{Reference Hutchens, Frawley and Sullivan65} Only one study investigated the association between WHO risk classification score and anxiety, reporting no significant association. ^{Reference Freiberger, Beckmann, Freilinger, Kaemmerer, Huber and Nagdyman58} As no studies assessed anxiety in the immediate postpartum period, there is a gap in the literature regarding the prevalence of anxiety following a pregnancy complicated by pre-existing CVD and the association between pre-existing CVD in pregnancy and anxiety. Future studies should ideally employ prospective cohort study designs, using validated perinatal anxiety measures such as the Generalised Anxiety Disorder-7 screener or the State-Trait Anxiety Inventory Short Form to fill this gap. ¹

Similarly, our postpartum PTSD symptom prevalence (18.2% for mild PTSD and 26.4% for moderate PTSD) is somewhat higher than what has been reported in the literature, although this result is limited to one study. ^{Reference Heyne, Kazmierczak, Souday, Horesh, Lambregtse-van den Berg and Weigl71} For example, a recent meta-analysis reported that the prevalence of postpartum PTSD was estimated to be 3.1% in the general population and 15.7% in at-risk populations, defined by psychiatric history, trauma history and perinatal risk. ^{Reference Grekin and O’Hara72} It is important to note that the study included in our review measured PTSD symptoms retrospectively to the postpartum period an average of 11 years after delivery, which introduces the possibility of recall bias and thus may inflate symptom prevalence. However, this study also reported that higher WHO risk classification scores were associated with greater postpartum PTSD symptoms, suggesting that disease severity may affect traumatic experiences. ^{Reference Freiberger, Beckmann, Freilinger, Kaemmerer, Huber and Nagdyman58} Evidence from non-pregnant samples has found that CVD is associated with increased rates of PTSD, likely because of exposure to traumatic acute cardiac events. ^{Reference Lunkenheimer, Garatva, Steubl and Baumeister73–Reference O’Donnell, Schwartz Longacre, Cohen, Fayad, Gillespie and Liberzon75} Qualitative research has revealed that those with pre-existing CVD during pregnancy fear the potentially life-threatening impact of their disease on their health and their child’s health. ^{Reference Hutchens, Frawley and Sullivan65} Given that the DSM-5 classifies PTSD as ‘exposure to actual or threatened death, serious injury or sexual violence to oneself, or others and subsequent symptoms’, it is plausible that experiencing a pregnancy complicated by pre-existing CVD may be a traumatic experience for some individuals and increase risk of developing PTSD. ⁷⁶ As only one study reported on this outcome, prospective studies are needed to assess the prevalence of postpartum PTSD in this population and to investigate the association between pre-existing CVD and postpartum PTSD. Future studies should use validated tools for assessing PTSD such as the Primary Care PTSD Screen, which is recommended for use in perinatal populations. ¹⁴

This scoping review highlights substantial gaps in the literature regarding pre-existing CVD during pregnancy and post-pregnancy mental health outcomes, necessitating future research. Our recommendations for future studies are listed in Table 1. We recommend that future studies employ a prospective cohort study design to minimise recall bias and to accurately capture post-pregnancy mental health outcomes within the immediate postpartum period (6 weeks) and up to 1 year postpartum. Samples should include those with CHD and those with acquired CVD, to compare mental health outcomes between these groups, as they may experience unique psychosocial stressors. ^{Reference Stanton, Revenson and Tennen66} We also recommend investigating the association between pre-existing CVD diagnosis and mental health outcomes, exploring measures of disease morbidity as potential mediators of these relationships. Literature regarding PTSD and anxiety is especially limited in this patient population, and current evidence suggests that these mental health conditions may be especially salient. Future studies should employ either validated screening tools to capture symptom prevalence or utilise clinician diagnosis.

Table 1

Recommendations for future studies

CVD, cardiovascular disease; PTSD, post-traumatic stress disorder; EPDS, Edinburgh Postnatal Depression Scale; PHQ-9, Patient Health Questionnaire-9; GAD-7, Generalised Anxiety Disorder-7; STAI-6, 6-Item State–Trait Anxiety Inventory; PC-PTSD, Primary Care Posttraumatic Stress Disorder Screen; WHO, World Health Organization.

Few clinical recommendations can be made from this scoping review because of the limited evidence and high study heterogeneity. However, this review does highlight the importance of screening patients with CVD for mental illness in the postpartum period. Although current guidelines recommend universal screening for depression and anxiety in the postpartum period, real-world data suggests screening rates are as low as 3.0% for depression and 2.0% for anxiety. ^{14,Reference Solomonov, Kerchner, Dai, Kwon, Callaghan and Schier77,Reference Sharma, Gaffey, Hameed, Kasparian, Mauricio and Marsh78} Screening for postpartum mental health conditions among those with CVD may be especially important, as postpartum mental illness have been found to increase CVD risk later in life. For example, PPD has been found to be associated with a 30–50% increased risk of hypertension, ischaemic heart disease and heart failure up to 20 years postpartum. ^{Reference Lu, Valdimarsdóttir, Wei, Chen, Andreassen and Fang13} For individuals already experiencing CVD morbidity, this increased risk may have compounding effects on future disease morbidity. To reduce this risk, current guidelines recommend integrating mental health specialists as needed into the prenatal care team for individuals with CVD during pregnancy and utilising community and national resources for support. ^{1,Reference Sharma, Gaffey, Hameed, Kasparian, Mauricio and Marsh78} Future research is needed to clarify whether preventive mental health support is necessary in this population and to identify the most effective strategies for integrating this care into clinical practice.

Our scoping review has several strengths. First, two independent reviewers screened all titles and abstracts and full texts for inclusion in the scoping review and two independent reviewers extracted all data included in the scoping review. A third reviewer was used as a tie breaker in the case of disagreements. Second, our broad inclusion criteria of multiple mental health outcomes (depression, anxiety and PTSD) and multiple time frames beyond the immediate postpartum period provides a thorough overview of the current literature on pre-existing CVD and postpartum mental health, allowing for identification of significant gaps in the literature. Third, our review was conducted in accordance with established scoping review methodology, ^{Reference Levac, Colquhoun and O’Brien46,Reference Tricco, Lillie, Zarin, O’Brien, Colquhoun and Levac79} with a pre-registered protocol on PROSPERO and additional scoping review registration on Open Science Framework. ^{Reference Becene, Jasani, Grimshaw and Vartany48} Finally, our literature search was conducted by a medical research librarian who ensured we captured studies not retrieved by database searches through citation chasing.

There are several limitations to this scoping review. First, we assessed a small number of studies, as only seven studies met criteria for inclusion in the scoping review. Second, there was high heterogeneity in study design among the included studies; studies relied on different screening tools (EPDS, PHQ-9, DASS, HADS) at different time points (6 weeks postpartum, 1 year postpartum, retrospectively 5 years postpartum, retrospectively 11 years postpartum), to assess depression, anxiety and PTSD symptoms. Third, two studies assessed postpartum mental health symptoms retrospectively an average of 5 and 11 years postpartum, introducing the possibility of significant recall bias. Fourth, only one study reported the prevalence of probable clinical depression for both those with congenital and pre-existing acquired CVD, limiting our ability to assess differences in depression by type of CVD. Fifth, only two studies reported results for anxiety and only one study reported results for PTSD, demonstrating a significant gap in literature. Sixth, four of the seven included studies had small sample sizes (n < 150), limiting their power to adequately detect statistically significant differences in mental health outcomes or provide representative estimates of probable clinical levels of depression, anxiety and PTSD.

Seventh, four of the seven studies included utilised data from the USA, which restricts generalisability of the findings to other countries. Finally, we did not evaluate study quality or risk of bias because of the high heterogeneity of study findings.

The literature examining pre-existing CVD during pregnancy and postpartum mental health disorders is limited. Future research examining depression, anxiety and PTSD in the postpartum period is needed among this patient population. Future studies should ideally employ a prospective cohort study design to evaluate the differing effects of acquired versus CHD and CVD severity. Postpartum mental health disorders should be assessed via validated screening tools or clinician diagnosis. As those with pre-existing CVD and mental illness are especially vulnerable to adverse cardiac events, it is vital to understand these risk factors in this population.