Prevalence of depression and anxiety in gestational trophoblastic disease: A systematic review and meta-analysis

Literature search

Literature searches were performed on PubMed, ScienceDirect, Cochrane Library, EBSCOHost, ProQuest, and Google Scholar for articles published between 27 February 2000 and 27 February 2024. This review excluded articles in languages other than English. The keywords applied were (Depression) OR (Depressive Symptoms) OR (Anxiety) OR (Anxiety Symptoms)) AND (Gestational Trophoblastic Disease) OR (Hydatidiform Mole) OR (Gestational Trophoblastic Neoplasia)) OR (Trophoblastic Neoplasm) and their respective Medical Subject Headings (MeSH) terms, if applicable.

Eligibility criteria

The inclusion criteria for our study encompassed cases with a confirmed diagnosis of GTD that reported on the prevalence of anxiety and depression using validated questionnaires or diagnostic interviews. We excluded case report, commentaries, letters, non-full-text articles, and unpublished or non-peer-reviewed materials. Single case reports were excluded due to their tendency to document only severe or atypical cases, which provide the least robust evidence. Additionally, studies that focused exclusively on patients with severe features were not included in this review.

Data extraction

Four assessors evaluated the selected studies’ quality utilizing the Newcastle–Ottawa Scale. ⁸ The reporting of this review adheres to the guidelines provided by the PRISMA statement and MOOSE recommendations. ⁹ Both reviewers carefully examined all abstracts independently to ensure consistent study selection. Discussions involving a third reviewer (CNRS) resolved disagreements on study inclusion. For datasets with identical cohorts and outcomes, only the study with the most complete data set was included to prevent duplication of cases.

Statistical analyses

Data were synthesized using a proportion-based meta-analysis. Prior to analysis, logit transformations were applied to the data, and depression and anxiety frequency estimates were obtained using a random-effects model. Meta-analysis computations were conducted with MedCalc software, version 19.5.1. ¹⁰ The significance of the overall effect was determined with a threshold of a p-value less than 0.05 in a two-tailed test. Study variability was assessed using the I² index, where an I² value below 50% indicated acceptable homogeneity, warranting the use of the Mantel-Haenszel fixed-effects model. Conversely, studies with an I² value above 50% were analyzed using the DerSimonian and Laird random-effects model. To evaluate publication bias, Egger’s linear regression test was employed, complemented by Begg’s rank correlation test. A p-value greater than 0.05 on these tests suggested an absence of publication bias. This systematic review and meta-analysis is registered with PROSPERO under registration number CRD42024525746.

Characteristics of included studies

The initial electronic database search yielded 4167 studies, from which 11 articles were identified for eligibility assessment after screening titles and abstracts. One article was excluded due to the unavailability of full text, another was an unpublished report, and two studies were omitted due to a lack of reported prevalence data. Subsequently, seven studies were considered appropriate for inclusion in this meta-analysis.^5,11–16 The search process and selection methods employed in this study are outlined in detail in Figure 1. This meta-analysis synthesized findings from seven distinct studies, involving a total of 462 participants. Among these studies, four utilized a cross-sectional design, two were retrospective, and one was prospective. However, two studies omitted reporting on anxiety prevalence. The review spanned various geographical locations, including Australia, the Netherlands, the USA, Italy, and Brazil, ensuring a diverse representation of populations (Table 1). Most of the studies included were conducted using methodological rigor, which was upheld by subjecting all studies to assess the operation of the New Castle–Ottawa Scale (Table 2).

OPEN IN VIEWER

Figure 1.

PRISMA flow chart outlines the methodology used to conduct the literature search and select articles suitable for inclusion in this meta-analysis.

OPEN IN VIEWER

Table 1.

Characteristics of studies included.

No	Author, year	Country	Study design	Population	Time since diagnosis	Number of sample	Age (years)	Partner (%)	Having children (%)	Previous pregnancy loss or history of miscarriage	First pregnancy
1	Blok LJ et al, 2021 11	Netherlands	Prospective multicenter cohort study	GTD	39 (25–50) days	60	32 ± 6.9	59 (98)	34 (57)	21 (35)	-
2	Di Mattei VE et al, 2015 5	Italy	Cross-sectional	GTD	6.6 ± 8.3 months	37	37.1 ± 7.5	26 (70.3)	22 (59.5)	-	15 (40.5)
3	Franca ACG et al, 2022 12	Brazil	Retrospective cohort study	GTD	-	186	31.2 ± 6.9	-	94 (50.5)	-	71 (38.2)
4	Jewell EL et al, 2018 13	USA	Retrospective study	GTN	4.54 ± 3.14 years	51	35.6 ± 10.4	43 (84)	42 (82)	-	-
5	Petersen RW et al, 2005 14	Australia	Cross-sectional	Molar pregnancy	-	74	-	-	-	-	-
6	Stafford L et al, 2011 15	Australia	Cross-sectional	GTD	4.7 ± 3.3 years	176	32.1 ± 3.89	165 (94)	91 (52)	-	71 (40)
7	Tenorio PJ et al, 2023 16	Brazil	Cross-sectional	GTD	3 months after pregnancy loss	64	25.87 ± 8.48	57 (87.7)	-	16 (25)	-

GTD: Gestational trophoblastic disease.

OPEN IN VIEWER

Table 2.

Newcastle Ottawa Scale of studies included in this review.

No	Study	Selection	Comparability	Outcome		Judgment
1	Blok LJ et al, 2021 11	⭑⭑⭑	⭑⭑	⭑	⭑	Good
2	Di Mattei VE et al, 2015 5	⭑⭑	⭑⭑	⭑	⭑	Satisfactory
3	Franca ACG et al, 2022 12	⭑⭑⭑	⭑⭑	⭑	⭑	Good
4	Jewell EL et al, 2018 13	⭑⭑⭑	⭑⭑	⭑	⭑	Good
5	Petersen RW et al, 2005 14	⭑⭑⭑⭑	⭑⭑	⭑⭑	⭑	Very good
6	Stafford L et al, 2011 15	⭑⭑	⭑⭑	⭑⭑	⭑	Good
7	Tenorio PJ et al, 2023 16	⭑⭑⭑	⭑⭑	⭑⭑	⭑	Very good

The aggregated data yielded a mean maternal age of 32.3 ± 3.91 years. About four studies utilized the Hospital Anxiety and Depression Scale (HADS-A) to evaluate anxiety, while one study utilized The State-Trait Anxiety Inventory (STAI-Y). Around three studies utilized the Hospital Anxiety and Depression Scale (HADS-D) to assess depression, with two studies using the Center for Epidemiologic Studies–Depression Scale (CES-D), one study utilizing the Beck Depression Inventory-Short Form (BDI-SF), and one employing diagnostic interviews (Table 3).

OPEN IN VIEWER

Table 3.

Prevalence of anxiety and depression.

No	Author, year	Number of sample	Anxiety or Depression (Measurement scale)	Anxiety		Depression
				Score	Prevalence (%)	Score	Prevalence (%)
1	Blok LJ et al, 2021 11	60	HADS-A ≥8 HADS-D ≥8	7.7 ± 4.5	28 (47)	12.9 ± 8.5	16 (27)
2	Di Mattei VE et al, 2015 5	37	STAI-Y ≥60 BDI-SF 9/10	42.7 ± 10.1	2/37	5.4 ± 4.21	6 (16.2)
3	Franca ACG et al, 2022 12	186	Diagnostic interview	-	-	-	17 (9.1)
4	Jewell EL et al, 2018 13	51	CES-D ≥16	-	-	10.5 ± 8.96	10 (20)
5	Petersen RW et al, 2005 14	74	HADS-A ≥8 HADS-D ≥8	-	41 (55.4)	-	13 (17.6)
6	Stafford L et al, 2011 15	176	HADS-A ≥8 CES-D ≥16	10.0 ± 10.4	38 (22)	5.67 ± 3.89	46 (26)
7	Tenorio PJ et al, 2023 16	64	HADS-A ≥8 HADS-D ≥8	8 (5-12.5)	34 (53.1)	6 (4–10)	28 (43.8)

Prevalence of depression among patients with gestational trophoblastic disease

The depression pooled event rate was calculated to be 0.222 (95% CI; 0.141–0.315), indicating a prevalence of depression at 22.2% throughout the study population (Figure 2). Analysis revealed no indication of publication bias as demonstrated by statistical tests including Begg’s and Egger’s tests of intercept. However, there was notable heterogeneity between studies (Q = 41.00, p < .0001, I² = 85.37).

OPEN IN VIEWER

Figure 2.

Forest plot of 7 studies assessing prevalence of depression in gestational trophoblastic disease.

Prevalence of anxiety among patients with gestational trophoblastic disease

The meta-analysis determined an anxiety pooled event rate of 0.349 (95% CI: 0.177–0.545), indicating a 34.9% prevalence of depression across the studies examined (Figure 3). Statistical tests, including Begg’s test and Egger’s test of intercept, revealed no evidence of publication bias. However, significant heterogeneity was observed between studies (Q = 61.43, p < .0001, I² = 93.49).

OPEN IN VIEWER

Figure 3.

Forest plot of 5 studies assessing prevalence of anxiety in gestational trophoblastic disease.

Discussion

This systematic review revealed that the prevalence of anxiety and depression among individuals with GTD significantly exceeds that reported in other forms of pregnancy loss, such as miscarriage, where depression affects approximately 5.4% to 18.6% of individuals and anxiety is observed in about 17% of cases.^17–22 The investigation unveiled that approximately 25% of the affected cohort exhibited symptoms indicative of depression, while roughly one-third displayed signs suggestive of anxiety. Many patients grappling with pregnancy loss embark on a grieving journey to mourn the departure of a cherished one. This emotional process endures for an extended duration, spanning from months to even years. ²³

The emotional toll associated with pregnancy loss typically encompasses a grieving process reminiscent of mourning the loss of a loved one, persisting over varying periods, from months to years. ²⁴ Nevertheless, GTD introduces complexities extending beyond mere pregnancy loss. Despite GTD’s high responsiveness to treatment with favorable prognoses for complete recovery, women diagnosed with GTD must navigate a dual challenge encompassing not only the loss of pregnancy but also the necessity for surgical interventions, the potential requirement for chemotherapy, confronting potentially life-threatening diagnoses, and encountering hurdles in subsequent attempts at conception.^11,16

Lower educational attainment and lack of partner support are associated with the onset of psychiatric disorders following a diagnosis of GTD.^15,16 Difficulties in adopting effective coping strategies and accessing essential resources during crises, particularly among women with limited education, can exacerbate these issues.^16,25 Partner support is recognized as a protective factor for mental health during pregnancy and the transition to motherhood. In contrast, the absence of such support is a significant risk factor for the development of anxiety and depression, especially in the context of GTD.^5,26 Inadequate partner support may act as a barrier to achieving a higher quality of life and managing the illness. Low educational attainment increases the likelihood of experiencing depressive symptoms following pregnancy loss by a factor of 3.4. Additionally, inadequate partner support is linked to a 4.4-fold increase in the risk of developing depression and a fourfold increase in the risk of developing anxiety. ¹⁶

Anxiety symptoms in patients with GTD may be associated with routine β-hCG monitoring, as many experience heightened distress before weekly testing, often triggered by fears of disease recurrence, infertility, and the emotional burden of delaying future pregnancy.^5,13 Surveillance through such monitoring has been demonstrated to induce a notable manifestation of depressive symptoms, mainly when patients become preoccupied with the levels, particularly regarding concerns about recurrence and the potential progression to malignancies. ¹⁶

In this investigation, symptoms of depression were reported less frequently than those of anxiety, a trend consistently observed in the included studies.^11,14,16 While anxiety may be driven by ongoing uncertainty and surveillance, depression appears to be more closely tied to the deeper emotional toll of losing a pregnancy. Many individuals experience an initial emotional shift—from the excitement of anticipating a healthy baby to the devastating reality of a potentially life-threatening diagnosis. This abrupt transition often triggers feelings of shock, grief, and disbelief.^14,16

The presence of existing children appears to alleviate anxiety among women diagnosed with GTD, potentially due to the fulfillment of their reproductive desires.^14,15 Research indicates that women who have children, whether before or after the diagnosis, generally experience lower levels of anxiety and depression compared to those without children. However, delays in childbearing caused by routine β-hCG monitoring also contribute to these symptoms.^5,16