Prevalence of and risk factors for suicidal ideation among perinatal women with HIV infection: a systematic review, meta-analysis, and meta-regression

Abstract

Background

The prevalence of suicidal ideation (SI) among HIV-positive pregnant women is a complex issue influenced by multiple risk factors. By addressing these risk factors and focusing on vulnerable regions, healthcare providers and policymakers can strive to alleviate the burden of SI in this population. The objective of this systematic review, meta-analysis, and meta-regression was to estimate the prevalence and identify risk factors for SI among HIV-positive pregnant women.

Methods

The review systematically searched PubMed, Scopus and Web of Science to identify relevant studies published until December 2024. A meta-analysis was conducted to summarize the prevalence and risk factors for SI among pregnant women with HIV infection. Sensitivity and meta-regression analysis were performed to explore the potential sources of heterogeneity in the distribution and determinants of suicidal behaviors within this at-risk group.

Results

The review identified 18 studies involving 5,242 participants. The overall prevalence of SI was 20.5%; 95% CI: 14.6 – 28.0% in pregnant women living with HIV, 19.8%; 95% CI: 12.6 – 29.6% in perinatal, and 14.9%; 95% CI: 7.8 – 26.5% in prenatal. The prevalence of SI was 17.1%, 20.3%, and 34.5% for the periods 2020–2024, 2015–2019, and 2000–2014, respectively. By gross domestic product (GDP), the prevalence was 16.9% in low, 23.0% in moderate, and 24.1% in high GDP countries, with the highest prevalence in the USA (24.2%). Among various risk factors, partner violence (OR = 1.44; 95% CI: 1.05–1.98), and higher education (OR = 0.80; 95% CI: 0.64–0.99) were identified as significant risk factors and protective factors, respectively. Meta-regression analysis indicated that GDP, partner violence, year, age, and depression were potential sources of heterogeneity, respectively.

Conclusion

The results indicated a high prevalence of SI among perinatal women living with HIV. The educational level, GDP, partner violence, and year were significant risk factors and potential sources of heterogeneity. It is crucial to incorporate specific questions about suicidal ideation into routine prenatal care for this population, even when depressive symptoms are not apparent. These findings underscore the need for a multifaceted approach to addressing suicidal ideation in HIV-positive prenatal women.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12888-025-07210-7.

Introduction

Women living with HIV, particularly in low and middle-income countries, face considerable psychological difficulties, including depression, stress, and anxiety, as a result of their HIV status. Research indicates that HIV positive women are at a heightened risk for experiencing more severe manifestations of mental health disorders [1, 2]. Women who are in their reproductive years are experiencing the most new HIV infections [3]. Pregnancy and childbirth are common experiences for many women living with HIV. Prenatal periods represent critical periods during which women may experience heightened vulnerability to depressive symptoms [4–7].

Research has shown that HIV-positive women experiencing psychological complications are more probable to face challenges in adhering to HIV treatments, experience accelerated illness development, have higher mortality rates, engage in amplified substance misuse, and encounter additional socioeconomic difficulties [1]. Despite significant advancements in treatment options that have improved overall well-being and abridged AIDS-related deceases among HIV-positive individuals worldwide [8], the risk of suicidality-encompassing suicidal thoughts, suicides and suicide attempts, remains higher for those with HIV compared to their uninfected peers in both developed and developing nations [9]. The periods surrounding childbirth, including the perinatal and postnatal phases, are associated with an increased the likelihood of suicidal behaviors (SBs). Compared to the general female population globally, perinatal women have lower rates of suicide attempts and completions, but pregnant women are more likely to experience suicidal ideation (SI) than the general population [10]. SI is common among perinatal women living with HIV, with its prevalence ranging from 8.2% [11] to 58% [12].

Several review studies have been conducted on the occurrence and risk factors of SBs among pregnant women or during the perinatal period. However, meta-analyses concerning suicidal behavior in pregnant women with HIV are not as well understood, particularly regarding the identification of factors influencing this prevalence and the exploration of key sources of heterogeneity [13–16]. Given the heightened vulnerability of pregnant women living with HIV, it is crucial to identify factors that contribute to their risk of SI. The prevalence of SI among HIV-positive pregnant women is a complex issue influenced by multiple risk factors. By addressing these risk factors and focusing on vulnerable regions, healthcare providers and policymakers can work towards mitigating the burden of SI in this population. The study conducted a systematic review and meta-analysis to address this gap and determine the prevalence and risk factors linked to SI among HIV-positive women during their pregnancy period and to explore differences in the prevalence and risk factors.

Materials and methods

PRISMA guidelines (8) were followed by this meta-analysis (8). The International Prospective Register of Systematic Reviews logged the protocol for this work (PROSPERO) (identifier: CRD42023450823).

Eligibility criteria

The review encompassed all records that evaluated SI and related risk factors among “perinatal” or “prenatal” women living with HIV. SI was measured using validated instruments noted in the eligible records. The term perinatal refers to the period surrounding birth, which includes both the prenatal (before birth) and postnatal (after birth) stages, usually extending up to a year postpartum. In contrast, prenatal specifically denotes the time during pregnancy, prior to birth. Published studies must meet the criteria of valid and well-defined methodology in order to be included, (1) records of enrolled HIV-positive perinatal or prenatal women, (2) studies identified HIV infection based on HIV standard test to confirm the infection, and (3) studies reported any information on suicidal ideation, thought or attempt. The exclusion criteria were: (1) Research that is presented in abstracts, case reports, case series, reviews, or practice guidelines, and (2) records enrolled only postpartum or postnatal cases.

To find related records published until December 2024, a comprehensive search was conducted in the PubMed, Scopus, and Web of Science (WOS) databases. Additionally, a manual search was performed on the reference lists of the identified records.

Search strategy

Scopus followed the following search strategy: TITLE-ABS-KEY (((“HIV” OR “human immunodeficiency virus” OR “acquired immunodeficiency syndrome” OR “AIDS”) AND (pregnant OR pregnancy OR perinatal OR antenatal OR prenatal)) AND (“Suicide” OR “suicide attempt*” OR “self-mutilation” OR “self-harm” OR “suicidal ideation*” OR “suicidal thought*” OR “suicidal plan” OR “completed suicide” OR “death by suicide” OR “mental health problem*” OR “mental health disorder*” OR “psychiatric disease*” OR “psychiatric disorder*” OR depression). PubMed and WOS had a search strategy similar to Scopus, and its table is included as a supplement (Appendix 1). Additionally, three reviewers performed an independent review of the reference lists and selected studies to guarantee that all pertinent articles were included in the evaluation.

Study selection

The complete text was examined to ascertain whether the articles fulfilled the inclusion criteria and were evaluated by three independent reviewers for each title and abstract. The three reviewers utilized their full texts to determine the eligibility of the chosen articles. A fourth reviewer was consulted to address any discrepancies.

Data extraction

Microsoft Excel spreadsheets were used to gather the information after extracting the data. The following dataset encompasses: baseline characteristics (authors, country, year, design), participants (study groups, age sample size by the study groups, pregnancy status and trimester of pregnancy, marital status, number of children, or number of pregnancies), HIV-related data (time of diagnosis, partner and children HIV status, disclosure of HIV), and psychological and physical issues (Depression, Suicide attempt, ideation, partner violence and etc.).

Quality assessment of included studies

Quality and risk of bias were assessed using ROBINS-I for non-randomized controlled trials. This instrument evaluates seven potential sources of bias, including confounding, selection bias, and information bias, as well as issues arising from deviations from intended interventions, missing data, bias in outcome measurement, and reporting [17]. The calculation of bias risk did not lead to the exclusion of any studies.

Synthesis methods

The absolute numbers and proportions were used to estimate pooled prevalence with 95% confidence intervals (CIs). The odds ratios (ORs) were pooled for assessing risk factors associated with suicidal ideation with 95% CI. The statistical heterogeneity between studies was analyzed using the I2 statistic. A fixed-effects or random-effects model was employed to estimate the pooled effect, depending on the heterogeneity, study design, and sample size. Low, moderate, and high levels of heterogeneity were represented by I2 values of 25%, 50%, and 75%. To identify the underlying heterogeneity, a subgroup meta-analysis was conducted. To investigate the effect of age, year, depression, and partner violence on suicidal ideation, a univariate meta-regression model was utilized [18, 19]. When decisions were made with arbitrary or unclear ranges, a sensitivity analysis was conducted. The evaluation of publication bias was done using Begg’s tests and Egger’s tests when there were at least 10 studies available for analysis. To determine the underlying cause of publication bias, a trim-and-fill analysis was applied when there was a noticeable publication bias. Comprehensive Meta-Analysis Version 3 was the tool used for conducting all analyses. All tests take into account the significance of P-values less than 0.05 [20].

Results

Study selection

Figure 1 displays the study flowchart. The review found 835 studies in PubMed, 1323 in Scopus, and 765 in WOS. 1502 studies were evaluated for their titles after removing duplicates. Of those, the abstract of 399 records were assessed and reviewed. After conducting an abstract survey, 31 studies were examined for full text. In conclusion, 18 studies met the inclusion criteria and were included in this systematic review and meta-analysis. The explanations for each exclusion of the records was provided in (Appendix 2).

Fig. 1.

PRISMA flow diagram

Characteristics of included studies

Table 1 outlines the baseline characteristics of the included studies [7, 11, 12, 21–35]. 18 studies were identified through the search process. The majority of the studies were conducted in Africa, with seven originating from Southern Africa [12, 26, 28–31, 35], four from Eastern Africa [7, 11, 23, 26], and two from Western Africa [21, 24]. Additionally, two articles were performed in the USA [32, 34], one in Thailand [33], and one in India [27]. Of the 18 studies reviewed, seven articles focused exclusively on prenatal patients, while the majority of the remaining studies included perinatal patients without distinguishing between pregnant and postpartum individuals. Regarding methodological approaches, seven studies lacked a control group, which precluded any comparative analysis. Six studies categorized perinatal HIV-positive patients based on the presence of suicidal ideation, allowing for a comparison of risk factors between two groups: those with suicidal ideation and those without. This facilitated an examination of the risk factors associated with each group. The remaining five studies categorized their participants based on various factors, including age, pregnancy status, mental health status, and the timing of HIV diagnosis. All studies presented the age of their participants as either a mean or a range, with ages spanning from a minimum of 18 years to a maximum of 35.8 years. Concerning the trimester of pregnancy at the time of enrollment, six studies did not provide information on the trimester of their participants. The majority of the studies provided data on the marital status and number of pregnancies of their participants; however, three studies did not report this information. Concerning information related to HIV, including the timing of diagnosis, disclosure to partners, partner HIV status, and receipt of antiretroviral therapy (ART), the majority of articles provided data. Intimate partner violence (IPV) was documented in ten studies, revealing a prevalence that ranged from a minimum of 15% to a maximum of 45%. Of these studies, five identified encompass various forms, including physical, psychological, and sexual violence.

Table 1.

Characteristics of the included studies

First Author, year	Country	Study Design	Participant Status and Population	Women status (Perinatal, prenatal)	Age (mean)	Trimester of pregnancy or weeks when enrolled	Marital Status, Number of Children, or Number of Pregnancies	HIV issues (Time of diagnosis, partner and children HIV status, disclosure of HIV)	Psychological and Physical issues (Depression, Suicide attempt, ideation, Partner violence and etc.)	Major Findings
Akinsolu et al. 2023 [21]	Nigeria	Cross-sectional	Cases number: 402 PHIV+ /No control group	Perinatal (prenatal and postpartum)	35.8 years	First trimester: 41.4% -Second trimester: 32.3% -Third trimester: 26.3%	Marital Status: Married: 93.5% Divorced: 2.3% Widowed: 2.7% Separated:1.5% Single: none	• Seropositive partner: 44.9% • Disclosed with partner: 58.2% • Years on ART: ≤1 year (19%) ≤ 5 Years (51.7%) >5 Years (46.5%) >10 Years (11.8)	• Depression prevalence: 63.9% • Mean EPDS score: 9.26 (antenatal)/13.3 (postpartum) • Suicidal ideation prevalence: 29.6% • Partner violence: 30.4%	• High prevalence of depressive symptoms among PHIV + during their perinatal period. • Low income, previous pregnancy complications, and gestational stage were associated with depression and perceived stress.
Gelaw et al. 2020 [11]	Ethiopia	Cross-sectional	Cases number: 414 PHIV+/No control group	Perinatal (prenatal and postpartum)	30.1 years (SD ± 4.65)	Not mentioned	Marital Status: Currently married: 72.5% Not currently married: 27.5%	• Seropositive partner: 59.4% • Disclosed HIV with partner: 74% • ART initiation before pregnancy: 63.3%	• Depression prevalence: 38.4% • Suicidal ideation prevalence: 8.2% • Suicide attempt prevalence: 1%	• High prevalence of perinatal depression among women living with HIV. • PHIV + need counseling to reduce HIV-related perceived stigma.
Harrington et al. 2018 [29]	Malawi	Cross-sectional	Cases number: 299 PHIV+/No control group	Perinatal (prenatal and postpartum)	26 years	22 (18–26) weeks	Marital Status: Currently married: 88% Not currently married: 12%	Not mentioned	•Depression prevalence: by EDPS: 10% by PHQ-9: 13% •Suicidal ideation prevalence: 2.4% • Intimate Partner Violence: 19%	• Higher prevalence of depression during the antenatal period than postpartum • Screening for depression during pregnancy should be integrated into antenatal HIV care
Knettel et al. 2020 [25]	Tanzania	Cross-sectional	Cases number: 200 PHIV+ PHIV + were compared as 2 groups based on having suicidal ideation (28 SI+/172 SI-)	Perinatal (prenatal and postpartum)	30 years (SD = ± 6)	Mean: 28 weeks	Marital Status: •SI + Currently married: 11.4% •SI- Currently married: 74% •SI + Not currently married: 32.1% •SI- Not currently married: 9.5%	• SI + Seropositive partner: 4.8% •SI- Seropositive partner: 20% •SI + Disclosed HIV: 14.2% •SI- Disclosed HIV: 66.5% •SI + newly diagnosed: 15.4% •SI- newly diagnosed: 5.5%	•SI + Suicidal ideation: 12.8% •SI + Partner violence: 20.3% •SI- Partner violence: 27.5%	• Significant decrease in suicidal ideation by 6 months post-partum • Significant correlation between anxiety and HIV stigma with suicidal ideation
Kulisewa et al. 2022 [12]	Malawi	Cross-sectional	Cases number: 73 PHIV+/No control group	Perinatal (prenatal and postpartum)	29 years	Not mentioned	Number of Pregnancies: P1: 12.5% P2: 37.5% P3: 25% P4: 16.7% P5+: 8.3%	• Time of Diagnosis: Newly diagnosed: 37.5%	• Mean EPDS score: 15 • Suicidal ideation prevalence: 58%	• Endorsement of the feasibility of integrated PND screening by most participants • Patient-centered counselling strategies were favored over medication by WLHIV as the acceptable treatment of choice
Kwalombota et al. 2002 [35]	Zambia	Cross-sectional	Cases number: 45 PHIV+ PHIV + were compared as 2 groups based on time of HIV diagnosis (Group A was 40 PHIV + diagnosed during pregnancy and group B was 5 PHIV + diagnosed before pregnancy)	Prenatal	between 20 and 35 years	Not mentioned	Not mentioned	• Time of Diagnosis: During Pregnancy: 89% Before Pregnancy: 11% • PHIV + in Group A did not know the HIV status of their partners. Women in Group B did know the HIV status of their partners.	• 95% from group A and 60% of group B had depression symptoms.	• Lesser depression, anxiety and suicidal ideation in women whose HIV are diagnosed before pregnancy
LeMasters et al. 2020 [26]	Malawi	Cross-sectional	Cases number: 73 PHIV+/No control group	Perinatal (prenatal and postpartum)	27 years	-Perinatal period: 58% -Postnatal period: 42%	Marital Status: Married: 83% Divorced: 8% Separated:8% Number of Pregnancies: P1: 13% P2: 38% P3: 33% P4: 17% P5+: 8%	• Time of Diagnosis: in last 6 months: 21% 6 months – 1 year: 4% 1–2 years: 4% 2 + years: 71%	• Depression prevalence: 33% • Suicidal ideation prevalence: 19%	• Receiving an HIV diagnosis unexpectedly during antenatal care was a key contributor to developing PND.
Levine et al. 2003 [34]	USA	Cross-sectional	Cases number: 56 PHIV+/No control group	Prenatal	27.9 years	Over 20 weeks	Not mentioned	• Time of Diagnosis: During Pregnancy: 23% Before Pregnancy: 77%	• Depression prevalence: 38% • Suicidal or Homicidal ideation: 14% • Domestic Violence history: 43% • Substance Abuse history: 54%	• High prevalence of self-reported depression and suicidal ideation in HIV-positive pregnant women
Mandell et al. 2022 [22]	South Africa	Cross-sectional	Cases number: 217 PHIV+ 82 SI+/135 SI-	Perinatal (prenatal and postpartum)	28.5 years (SD = ± 5.8)	- All: 17.5 weeks (SD = ± 5.8) - SI+: 16.5 weeks (SD = ± 5.5) - SI+: 18 weeks (SD = ± 5.1)	Marital Status: -SI+ & Currently married: 9% -SI- & Currently married: 13% -SI+ & Not currently married: 29% -SI- & Not currently married: 49%	Not mentioned	- Intimate Partner Violence (IPV): -SI + Psychological IPV:8.2% -SI- Psychological IPV: 5.6% -SI + Physical IPV: 4.3% -SI- Physical IPV: 2.2%	• Increased diastolic blood pressure was associated with both depression and suicidal ideation among pregnant women with HIV.
Ngocho et al. 2019 [7]	Tanzania	Cross-sectional	Cases number: 200 PHIV+/No control group PHIV + were compared as 2 groups based on probable depression	Prenatal	30 years	GA:28.4 weeks (SD = ± 5.6)	Marital Status: - Currently married: 86% - Not currently married: 14% Number of Pregnancies: - First Pregnancy: 22%	• Time of Diagnosis: Newly Diagnosed: 47% • Any HIV Status Disclosure: 79.5%	- Possible Depression: 25% - Probable Anxiety: 24.5% - Ever experienced Violence: 35%	• Depression was significantly associated with being single, food insecurity and HIV shame. • Anxiety was associated with being single, HIV shame and lifetime experience of violence.
Parcesepe et al. 2021 [24]	Cameroon	Cross-sectional	Cases number: 230 PHIV+/No control group PHIV + were compared as 2 groups based on probable common mental disorder SRQ-20 score	Prenatal	18–24 years: 21.3% 25 + years: 78.7%	Not mentioned	Marital Status: - Currently married: 30.9% - Not currently married: 69.1% Number of Pregnancies: 1: 9.1% 2–3: 43% 4+: 47.8%		- Probable Common Mental Disorder (CMD): SRQ-20, 9/10 cut-off: 26.7% SRQ-20, 7/18 cut-off: 42.2% Suicidal Ideation Prevalence: 14.8% - Intimate Partner Violence (IPV): Emotional IPV: 44.5% Physical IPV: 36.6% Sexual IPV: 31.3%	• All forms of IPV assessed were significantly associated with greater odds of probable CMD.
Rodriguez et al. 2017 [31]	South Africa	Cross-sectional	Cases number: 673 PHIV+ PHIV + were compared as 2 groups based on having suicidal ideation (261 SI+/412 SI-)	Prenatal	All: 28.39 years (SD = ± 5.73) SI+: 27.97 years (SD = ± 5.93) SI-: 28.66 years (SD = ± 5.59)	SI+: 17.21 weeks (SD = ± 5.74) SI-: 18.21 weeks (SD = ± 5.59)	Marital Status: -SI+ & Currently married: 16% -SI- & Currently married: 25% -SI+ & Not currently married: 23% -SI- & Not currently married: 36%	• Time of Diagnosis: - SI+ & During this Pregnancy: 22% - SI- & During this Pregnancy: 32% • HIV Serostatus of Partner: -SI+ & Seropositive partner: 9% -SI- & Seropositive partner: 16% • HIV Serostatus of Children: -SI+ & Seropositive children: 1% -SI- & Seropositive children: 3% • Any HIV Status Disclosure: - SI + Any HIV Status Disclosure: 18% - SI- Any HIV Status Disclosure: 46%	-Mean Psychological Intimate Partner Violence (IPV): 3.20% SI+ & Psychological IPV: 1.6% SI- & Psychological IPV: 1.4% -Mean Physical Intimate Partner Violence (IPV): 1.12% SI+ & Physical IPV: 0.7% SI- & Physical IPV: 0.7% Mean EPDS-9: SI+: 14.34 SI-: 9.39	• High rates of suicidal ideation among pregnant women with HIV • Depression had the largest association with suicidal ideation.
Rodriguez et al. 2018 [28]	South Africa	Cross-sectional	Cases number: 681 PHIV+ PHIV + were compared as 2 groups based on having suicidal ideation (266 SI+/415 SI-)	Perinatal (prenatal and postpartum)	All: 28.47 years (SD = ± 5.75) SI+: 28.02 years (SD = ± 5.93) SI-: 28.64 years (SD = ± 5.56)	8–24 weeks and 32 weeks	Marital Status: -SI+ & Currently married: 16% -SI- & Currently married: 25% -SI+ & Not currently married: 23% -SI- & Not currently married: 36%	• Time of Diagnosis: - SI+ & During this Pregnancy: 22% - SI- & During this Pregnancy: 32% • HIV Serostatus of Partner: - SI+ & Seropositive partner: 9% - SI- & Seropositive partner: 16% • HIV Serostatus of Children: -SI+ & Seropositive children: 1% -SI- & Seropositive children: 3% • HIV Status Disclosure: - SI + HIV Status Disclosure to Partner: 15% - SI- Any HIV Status Disclosure to Partner: 44%	- Depression prevalence: 10.82% SI+ & depressed: 5% SI- & depressed: 6%	• High rates of suicidal ideation among pregnant women with HIV
Sarna et al. 2019 [27]	India	Cross-sectional	Cases number: 200 PHIV+ PHIV + were compared as 2 groups based on women status (Prenatal and postpartum)	Perinatal	25.88 years (SD = ± 4.47)	26.4 weeks (SD = ± 12.1)	Marital Status: - Currently married: 98% - Not currently married: 2% Number of Pregnancies: 1: 32% 2: 30.5% 3: 23.5% 4+: 14%	• HIV Status Disclosure: - No Disclosure: 28.5% - Disclosure to family or friends: 71.5% • HIV Serostatus of Partner: - Seropositive partner: 61.5% - Seronegative/Untested partner: 38% • HIV Serostatus of Children: -at least One Seropositive child: 11% -No Seropositive child: 89%	Depression Prevalence: 52.5% Self-Harm Thoughts: 23% Mean EPDS Score: 12.69	• High rates of suicidal ideation and self-harming thoughts among pregnant women with HIV
Wong et al. 2017 [30]	South Africa	Cross-sectional	Cases number: 625 PHIV+ PHIV + were compared as 2 groups based on women age (18–24 years old and ≥ 25 years old)	Prenatal	28 years (SD = ± 5.3)	-First Trimester: 7% -Second Trimester: 57% -Third Trimester: 37%	Marital Status: - Currently married: 41% - Not currently married: 59% Number of Pregnancies: - First Pregnancy: 18%	• Time of Diagnosis: - During this Pregnancy: 55%	Depression Prevalence: 11% Self-Harm Thoughts: 6% Mean EPDS Score:5.3 - Intimate Partner Violence (IPV): Psychological IPV: 15% Physical IPV: 15% Sexual IPV: 2.2%	• More depressive symptoms and self-harming thoughts among younger HIV-infected pregnant women compared to older ones
Zewdu et al. 2021 [23]	Ethiopia	Cross-sectional	Cases number: 414 PHIV+ PHIV + were compared as 2 groups based on having suicidal ideation (34 SI+/380 SI-)	Perinatal (prenatal and postpartum)	30 years	Not mentioned	Number of Pregnancies: - First Pregnancy: 23.6%	• Time of Diagnosis: - During this Pregnancy: 63.3%	Depression Prevalence: 38.4% Suicidal Ideation Prevalence: 8.2% Suicidal Attempt Prevalence: 0.97%	• Low magnitude of suicidal ideation among HIV positive perinatal women • Perinatal depression, non-disclosed HIV status, and unplanned pregnancy are substantially associated with suicidal ideation
Aaron et al. 2015 [32]	USA	Case-Control	Cases number: 49 PHIV+/113 PHIV-	Perinatal (prenatal and postpartum)	• PHIV+: 28.1 (SD = ± 6.3) • PHIV-: 24.1(SD = ± 4.6)	Not mentioned	Marital Status: Currently married: 80.3% Not currently married: 19.8% Number of Children: none: 37.7% 1–3: 53.1% 3: 23.5% ≥ 4: 9.3%	• Time of Diagnosis: - During this Pregnancy: 22.4% • Pregnancy planned: 22.4% • Partner awareness of HIV status: 79.6%	• Suicidal attempt/ideation: (PHIV+:36.7 PHIV-: 14.2%) • Intimate Partner Violence (IPV): (PHIV+: 40.8% PHIV-: 15.9%) • Depression prevalence: (PHIV+: 57.1% PHIV-: 41.6%)	• CES-D scores did not differ prenatally or postpartum between women with and without HIV. • Prenatal state anxiety scores were higher in women with HIV but there were no differences postpartum.
Ross et al. 2009 [33]	Thailand	cross-sectional	Case number: 391 PHIV+/No control group	Prenatal	26.1 years (SD = ± 5.2)	-First trimester: 30% -Second trimester: 47% -Third trimester: 23%	Number of Pregnancies: - First Pregnancy: 55%	• Seropositive partner: 34.6%	• Depression prevalence: 78% - Mild Depression (CES-D = 16–22): 22.9% - Clinical Depression (CES-D ≥ 23): 55.1%	• High prevalence of depressive symptoms among PHIV + during their perinatal period • Positive association between physical symptoms and depressive symptoms • Negative association between self-esteem, emotional support and financial status with depressive symptoms

ART Antiretroviral Therapy, CES-D Center for Epidemiological Studies Depression Scale, EDPS Edinburgh Postnatal Depression Scale, HIV Human Immunodeficiency Virus, PHIV+ Pregnant women with HIV, PHQ-9 Patient Health Questionnaire-9, PND Perinatal Depression, SD Standard Deviation, SI Suicide Ideation

Quality assessment of included studies

The RoB-1 assessment is summarized in Fig. 2. It was found that there was no overall risk of bias. The second domain had only a low percentage (< unk > 15%) of serious bias risk, which is caused by participant selection bias. A moderate risk of bias (< 30%) was noted in confounding and selection of participants, missing data, and less than 10% in selection of reported results. There was no significant bias detected in third (classification of interventions), fourth (deviations from intended interventions) and sixth (measurement of outcomes) domains. Among the 17 studies, three [25, 34, 35] exhibited a moderate risk of confounding bias due to inadequate examination of depression status. Concerning bias from participant selection, six studies [11, 12, 21, 29, 33, 34] exhibited a moderate risk of bias due to the lack of a control group, which is crucial for achieving robust and reliable results. In the absence of a control group, it becomes difficult to ascertain whether the observed outcomes are truly a consequence of the intervention or are affected by other confounding factors. The study by Kwalombota et al. [35] was found to have a serious risk of bias in selecting participants; the researchers categorized participants into two groups: Group A consisted of 40 HIV-positive perinatal women diagnosed during pregnancy, while Group B included 5 HIV-positive perinatal women diagnosed prior to pregnancy. The study reported that 100% of Group A exhibited suicidal thoughts. Regarding bias arising from missing data, three studies [29, 34, 35] failed to provide any information on HIV-related issues such as partner HIV status, disclosure to partners, children’s HIV status, marital status, and time of diagnosis, despite indicating that they had collected this information. Only one study demonstrated bias related to the selection of reported results [35]. It exclusively presented the prevalence of anxiety and depression without providing additional details, such as the measurement tools used and the criteria for assessment.

Fig. 2.

Results of risk of bias assessment

Result of synthesis

Overall

A meta-analysis of 18 studies (n = 5,242) investigated the prevalence of SI among HIV-positive perinatal women, revealing an overall prevalence of 20.5% and an effect size of 0.205 (95% CI [0.146, 0.280], p < 0.001, I² = 96%). Notably, a study by Kwalombota et al. (2002) reported a particularly high SI prevalence of 91%. A sensitivity analysis excluding this study resulted in a revised SI prevalence of 17.9%, with an effect size of 0.179 (95% CI [0.128, 0.246], p < 0.001, I² = 96%). Additionally, six studies reported SI prevalence exceeding 25%, prompting a second sensitivity analysis that excluded these studies, which yielded a prevalence of 12.5% and an effect size of 0.125 (95% CI [0.097, 0.160], p < 0.001, I² = 83%). Table 2 represents the results of the meta-analysis. Forest plots are presented in Fig. 3 (overall prevalence of SI) and Fig. 4 (sensitivity analysis).

Table 2.

Result of synthesis (Meta-analysis and sub-group analysis)

Parameters	No. Studies	No. Events	No. Sample Size	Effect Model	Pooled Effect Size (CI: 95%)	P value*	Heterogeneity
							I2	P value
Prevalence of Suicidal Ideation Among Pregnant Women Living with HIV
Overall	18	1140	5242	Random	0.205 [0.146, 0.280]	0.00	96%	0.000
Sensitivity Analysis
Overalla	12	353	3175	Random	0.125 [0.097, 0.160]	0.00	83%	0.000
Overallb	17	1099	5197	Random	0.179 [0.128, 0.246]	0.00	96%	0.000
Grouped by Women Status
Perinatalb	11	640	2977	Random	0.198 [0.126, 0.296]	0.00	95%	0.000
Prenatal	6	459	2220	Random	0.149 [0.078, 0.265]	0.00	97%	0.000
Grouped by Year
2020–2024	8			Random	0.171 [0.099, 0.279]	0.000	94%	0.000
2015–2019	7			Random	0.203 [0.115, 0.331]	0.000	97%	0.000
2000–2014	3			Random	0.345 [0.152, 0.602]	0.247	96%	0.000
Grouped by Region
South Africa	7	699	2651	Random	0.209 [0.129, 0.321]	0.000	97%	0.000
East Africac	4	124	1228	Random	0.107 [0.054, 0.203]	0.00	69%	0.021
West Africa	2	153	632	Random	0.214 [0.088, 0.436]	0.015	94%	0.000
USA	2	26	95	Random	0.242 [0.094, 0.496]	0.05	85%	0.010
Thailand	1	51	391	-	0.130 [0.101, 0.168]	-	-	-
India	1	46	200	-	0.230 [0.177, 0.293]	-	-	-
Grouped by GDP
Low	14	1027	4892	Random	0.169 [0.114, 0.243]	0.000	97%	0.000
Moderate	1	46	200	-	0.230 [0.050, 0.630]	-	-	-
High	2	26	95	Random	0.241 [0.080, 0.538]	0.083	85%	0.010

^aKwalombota et al. 2002 [35], Akinsolu et al. 2023 [21], Aaron et al. 2015 [32], Rodriguez et al. 2017 [31], Rodriguez et al. 2018 [28], Mandell et al. 2022 [22]excluded because of reported prevalence over 25%

^bKwalombota et al. 2002 [35] excluded because of a reported prevalence of 91%

^cKwalombota et al. 2002 [35] excluded because of a reported prevalence of 91% (high risk of bias); *Bold P-values are significant

Fig. 3.

Prevalence of suicidal ideation among pregnant women living with HIV (Overall) This forest plot represents the overall prevalence of suicidal ideation among pregnant women living with HIV. The overall prevalence was 20.5% and the effect size was 0.205 (95% CI [0.146, 0.280], p < 0.001, I² = 96%)

Fig. 4.

Prevalence of Suicidal Ideation Among Pregnant Women Living with HIV (Sensitivity analysis-Overall¹)- Kwalombota et al. 2002, Akinsolu et al. 2023, Aaron et al. 2015, Rodriguez et al. 2017, Rodriguez et al. 2018, Mandell et al. 2022 excluded because of reported prevalence over 25% This forest plot represents the prevalence of suicidal ideation among pregnant women living with HIV (Sensitivity analysis-Overall1) after the exclusion of Kwalombota et al. 2002, Akinsolu et al. 2023, Aaron et al. 2015, Rodriguez et al. 2017, Rodriguez et al. 2018 and Mandell et al. 2022 because of a reported prevalence over 25%. This sensitivity analysis resulted in a revised SI prevalence of 12.5% and an effect size of 0.125 (95% CI [0.097, 0.160], p < 0.001, I² = 83%)

Subgroup analysis

Grouped by women status

A meta-analysis of 11 studies (n = 2,977) revealed that the overall prevalence of suicidal ideation (SI) among pregnant women living with HIV was 19.8%, with an effect size of 0.198 (95% CI [0.126, 0.296], p < 0.001, I² = 95%). Specifically, the prevalence of SI among HIV-positive pregnant women was 14.9%, with an effect size of 0.149 (95% CI [0.078, 0.265], p < 0.001, I² = 97%) (Fig. 5).

Fig. 5.

Subgroup meta-analysis by women pregnancy status This Forest Plot represents the prevalence of suicidal ideation among HIV-positive prenatal and perinatal women, grouped by women status. It reveals that the overall prevalence of SI among HIV-positive perinatal women was 19.8%, with an effect size of 0.198 (95% CI [0.126, 0.296], p < 0.001, I² = 95%) and the prevalence of SI among HIV-positive prenatal women was 14.9%, with an effect size of 0.149 (95% CI [0.078, 0.265], p < 0.001, I² = 97%)

Grouped by region

A meta-analysis of seven studies (n = 2,651) reported a prevalence of SI of 20.9% in South Africa, with an effect size of 0.209 (95% CI [0.129, 0.321], p < 0.001, I² = 97%). In East Africa, a meta-analysis of two studies found a SI prevalence of 11% among HIV-positive perinatal women, with an effect size of 0.107 (95% CI [0.054, 0.203], p < 0.001, I² = 69%). The prevalence of SI in West Africa (n = 632) was 21%, with an effect size of 0.214 (95% CI [0.088, 0.436], p = 0.015, I² = 94%). Lastly, a meta-analysis of two studies (n = 95) in the United States found a SI prevalence of 24%, which was not statistically significant, with an effect size of 0.242 (95% CI [0.094, 0.496], p = 0.05, I² = 85%).

Grouped by GDP

The meta-analysis of 14 studies conducted in low-GDP countries reported a prevalence of SI among HIV-positive perinatal women of approximately 17%, with an effect size of 0.169 (95% CI [0.114, 0.243], p < 0.001, I² = 97%). In contrast, the meta-analysis of two studies from high-GDP countries found a SI prevalence of 24%, but this result was not statistically significant, with an effect size of 0.241 (95% CI [0.080, 0.538], p = 0.083, I² = 85%).

Meta-regression

Correlation analysis examining the relationship between the prevalence of SI among HIV-positive perinatal women and potential moderators, including mean age of participants, year of study, depression prevalence, and GDP growth rate of the countries, found no significant associations (Table 3). Meta-regression plots are presented in (Figs. 6 and 7).

Table 3.

Meta-regression-Correlation between prevalence of suicidal ideation among pregnant women with HIV and associated moderators

Moderator	No. study		Coefficient	SE	Z value	P value	95% CI	R 2
Age		17	0.0149	0.086	0.17	0.863	[−0.155, 0.185]	0.00
Year		17	−0.066	0.036	−1.80	0.07	[−0.137, 0.005]	0.00
Depression Prevalence		13	−0.002	0.0121	−0.29	0.884	[−0.027, 0.020]	0.00
IPV Prevalence		10	−0.010	0.018	−0.58	0.561	[−0.04, 0.02]	0.03
GDP growth rate		18	−0.173	0.093	−1.85	0.06	[−0.356, 0.0103]	0.25

Fig. 6.

Meta-regression plot (prevalences of suicidal ideation and and depression) This Meta-regression represents the correlation between the prevalence of suicidal ideation among pregnant women living with HIV and depression prevalence. No significant correlation was revealed with a coefficient of −0.002 (95% CI [−0.027, 0.020], p = 0.884, R² = 0)

Fig. 7.

Meta-regression plot (SI prevalence and gross domestic product (GDP)) This Meta-regression represents the correlation between the prevalence of suicidal ideation among pregnant women living with HIV and GDP growth rate. No significant correlation was revealed with a coefficient of −0.173 (95% CI [−0.356, 0.0103], p = 0.6, R² = 0.25)

Risk factors associated with SI among HIV-positive perinatal women

The meta-analysis of four studies indicated that attainment of a higher education diploma or beyond was significantly associated with a reduced likelihood of SI among HIV-positive perinatal women (OR = 0.800, 95% CI = 0.642–0.995, p = 0.04). Additionally, experiencing partner violence was found to significantly increase the odds of SI in this population (OR = 1.447, 95% CI = 1.054–1.985, p = 0.022). No significant associations were observed between SI and the other factors examined (see Table 4).

Table 4.

Risk factors associated with suicidal ideation among pregnant women living with HIV

Risk factor	Category	No. study	Effect Model	Odds Ratio (95% CI)	P-value*	Heterogeneity
						I2	P-value
Marital status	Married/In relationship	5	Fixed	0.911 [0.751, 1.106]	0.348	48%	0.101
	Single/No commitment
Education	Diploma and higher	4	Fixed	0.800 [0.642, 0.995]	0.045	0%	0.734
	Below diploma
Disclosure of HIV	Disclosed with Partner	5	Random	0.625 [0.366, 1.067]	0.085	80.3%	0.000
	Not disclosed
Partner Violence	Experienced	4	Fixed	1.447 [1.054, 1.985]	0.022	0%	0.749
	Not experienced
Depression status	Depressed	4	Random	1.406 [0.421, 4.691]	0.580	86.2%	0.000
	Not depressed
Father’s HIV status	Positive	3	Fixed	0.809 [0.629, 1.040]	0.099	25.9%	0.259
	Negative
Planning for pregnancy	Planned	3	Random	0.595 [0.244, 1.448]	0.253	88%	0.000
	Not planned
HIV diagnosis	During pregnancy	3	Random	1.765 [0.888, 3.508]	0.105	87%	0.000
	Not during pregnancy

* Bold P-values are significant

Publication bias

The SI proportion among perinatal HIV-positive women was examined and it was found that there was a significant publication bias (Egger’s test: SE = 2.65, 95%CI [−11.3, −0.100], p = 0.04). Adding 2 studies on the right side of the scatter plot was necessary after implementing Duval and Tweedie’s trim and fill method. The effect size was calculated after making this adjustment (0.232, 95%CI: 0.168, 0.312).

Discussion

This review aimed to summarize the prevalence of SI among HIV-positive pregnant women. Adhering to rigorous inclusion and exclusion criteria, we identified 18 studies that fulfilled the eligibility requirements for meta-analysis. The analysis of these 18 qualifying articles indicates an overall SI prevalence of 20.5%. Importantly, a sensitivity analysis that excluded a study with a notably high SI prevalence adjusted the overall figure to 17.9% [35]. A subsequent sensitivity analysis, which omitted studies reporting SI prevalence greater than 25% [21, 22, 28, 31, 32, 35], resulted in a prevalence of 12.5%. Additionally, we performed a subgroup analysis to explore the factors that may influence SI prevalence. This analysis took into account the perinatal and prenatal status of the women, geographical region, GDP levels and year of publication. Moreover, we examined the relationship between SI and various moderators, as well as the impact of associated risk factors among HIV-positive pregnant women.

Prevalence of suicidal ideation based on women status

A meta-analysis of 6 studies revealed that the pooled proportion of SI among prenatal women living with HIV was 14.9%. Notably, the prevalence of SI among HIV-positive women during the perinatal period was found to be 19.8% suggesting a heightened occurrence of SI during this phase. There is evidence indicating a decline in SI from the prenatal period to the postpartum period [25], with reports suggesting that SI ceases following childbirth [28]. Further investigation is needed to determine the cause of high levels of SIs during pregnancy and how they reduce over time. The reason may be linked to worries about logistical and economic funding throughout pregnancy, as well as concern over the baby’s health and HIV status, HIV disclosures and subsequent stigma [7, 25]. It is well-documented that a mother may experience emotional instability in the period leading up to childbirth, largely attributable to the significant stress and uncertainty surrounding the delivery and the newborn. However, these feelings of pressure and uncertainty often diminish or completely resolve following the birth [36]. Furthermore, the act of giving birth can elicit profound joy in the mother, a phenomenon referred to as “joy at birth,” which has the potential to alleviate symptoms of depression [36].

Prevalence of suicidal ideation based on region and GDP

The findings from this meta-analysis revealed regional variations in the prevalence of SI among HIV-positive perinatal women, with West Africa showing the highest rate at 21%, closely followed by South Africa at 20.9%, and East Africa at 11%. These disparities underscore the complex interplay of factors influencing SI in different African regions. Accordingly, low-GDP countries reported a statistically significant prevalence of SI among HIV-positive perinatal women (approximately 17%). The high prevalence rates in West and South Africa and more precisely, low-GDP countries, are particularly concerning and warrant closer examination. These regions, like many parts of Africa, face numerous socioeconomic challenges that may contribute to the elevated SI rates among HIV-positive women. Factors such as poverty, limited access to healthcare, stigma associated with HIV, and gender inequality could all play significant roles in exacerbating mental health issues, including SI [21, 28, 35]. The lower prevalence in East Africa (11%) is noteworthy and may suggest the presence of protective factors or more effective support systems in this region. However, this rate is still alarmingly high and indicates a substantial mental health burden among HIV-positive women [25, 26, 37]. Psychosocial characteristics are also critical in understanding SI prevalence. The psychological impact of an HIV diagnosis, especially during pregnancy, can be profound. Women may experience fear, anxiety, depression, and concerns about their health and the health of their unborn child. Additionally, the stigma associated with HIV in many African communities can lead to social isolation, discrimination, and reduced self-esteem, further increasing the risk of SI [7]. The varying prevalence rates across regions suggest that cultural and societal factors specific to each area may influence SI risk. These findings highlight the urgent need for targeted interventions that address the specific challenges faced by HIV-positive women in different African regions. Comprehensive approaches that combine improved access to mental health services, HIV care, and social support are essential.

Prevalence of suicidal ideation and potential moderators

Our analysis showed no significant association between SI among HIV-positive pregnant women and factors including age, depression prevalence, GDP growth rate of countries, or the year in which studies were conducted. This lack of correlation may suggest that suicidal thoughts in this population may be influenced by more complex, multifaceted factors that are not easily predicted by demographic or economic indicators or even depressive disorder.

Wong et al. reported more depressive symptoms and self-harming thoughts among young HIV-infected pregnant women compared to older women [30]. On the contrary, Rodriguez et al. concluded that SI was not associated with age [28].

Even though the prevalence of SI is higher in low-GDP countries [25, 38], the lack of association with countries’ GDP growth rates suggests that economic development alone may not be sufficient to address the mental health challenges faced by this population. Even in countries experiencing economic growth, HIV-positive pregnant women may still struggle with suicidal thoughts, highlighting the need for targeted mental health interventions regardless of a country’s economic status.

Perhaps most concerning is the observation that there has been no significant improvement in preventing suicidal ideation among prenatal HIV-positive women over the years. This trend indicates a persistent gap in addressing the mental health needs of this vulnerable group. Despite advancements in HIV treatment and maternal care, the psychological burden of living with HIV during pregnancy continues to be a significant challenge [21]. This lack of progress over time underscores the urgent need for more effective, tailored interventions to address suicidal ideation in HIV-positive pregnant women.

While depression is often closely associated with suicidal thoughts, our analysis suggests that the relationship may not always be straightforward in this specific population. Gelaw et al., reported that depression was nearly four times as prevalent among women who had suicidal ideation as their counterparts [11]. Mandell et al. reported greater mean depressive symptoms in SI [22]. Moreover, Rodriguez et al. reported Depression had the largest association with SI [31].

In the study by knettel et al. the association between HIV, depression symptoms, and SI has been established, and these connections could be even more significant for pregnant and postpartum women [25]. Some studies have observed instances where SI was present in HIV-positive pregnant women who did not meet the full diagnostic criteria for depression. Akinsolu et al. demonstrated that the majority of HIV positive women without depressive symptoms experienced suicidal thoughts quite often [21]. Suicidal thoughts were not reported by most depressive symptoms respondents, suggesting that SI may not be as prevalent in this population as expected [21]. This highlights the changeability in the prevalence of SBs among people with and without depressive symptoms [23, 39, 40]. These finding underscores the importance of comprehensive mental health assessments that go beyond screening for depressive symptoms alone. Healthcare providers should be aware that suicidal thoughts may occur independently of clinically significant depression in HIV-positive pregnant women [23].

Risk factors associated with SI among HIV-positive perinatal women

The meta-analysis of four studies indicated that attainment of a higher education diploma or beyond was significantly associated with a reduced likelihood of SI among HIV-positive perinatal women. Education has been identified as a statistically significant risk factor for suicidal ideation in this population [24, 28, 32]. This correlation may be attributed to several factors. Lower educational attainment is frequently associated with reduced access to mental health resources, limited comprehension of HIV and its management, and fewer coping mechanisms to address the psychological burden of the diagnosis [12]. Furthermore, women with lower levels of education may encounter greater socioeconomic challenges, potentially exacerbating stress and contributing to suicidal thoughts [32].

Partner violence emerges as another critical risk factor in this study. The intersection of HIV status and intimate partner violence creates a particularly vulnerable situation for prenatal women [25]. Moreover, the apprehension surrounding HIV status disclosure to a violent partner may increase isolation and psychological distress, further elevating the risk of suicidal thoughts [24, 30].

Limitations

It is essential to recognize specific limitations when evaluating the reports presented. A subgroup meta-analysis was performed to address the variability among studies and the associated risk factors. Nonetheless, it is crucial to understand that the inherent limitations of the included studies rendered some level of heterogeneity unavoidable. The review is constrained by its reliance on narrative rather than statistical synthesis of the findings. Many of the risk factors were examined in only a limited number of studies that utilized diverse methodologies, making a meta-analysis inappropriate for most of these factors. While it is vital to explore suicidal ideation across all female demographics, this approach may have led to the inclusion of certain risk factors that do not universally apply to all women. Collectively, our findings highlight the intricate nature of suicidal ideation within this demographic and underscore the necessity for a more detailed comprehension of the contributing factors.

Implications for research, practice, and policy

Future research should focus on identifying the specific psychosocial, cultural, and clinical factors that contribute to suicidal thoughts in HIV-positive pregnant women, independent of age, economic conditions, or temporal trends. Developing and validating screening tools should be specifically developed to assess SI in HIV-positive pregnant women. Additionally, longitudinal studies examining the trajectory of mental health throughout pregnancy and postpartum periods could provide valuable insights into the timing and nature of interventions needed.

Conclusion

The results indicated an overall SI prevalence of 20.5%, with sensitivity analyses revealing a lower prevalence of 12.5%. Notably, the prevalence of SI was found to be more pronounced in countries with low GDP and those classified as developing. The analysis did not identify any significant correlations between SI and factors such as depression, intimate partner violence (IPV), age, or GDP growth rate. Conversely, higher educational attainment was associated with a reduced prevalence of SI, while an increase in partner violence correlated with a higher prevalence of SI.

In conclusion, the persistent nature of suicidal ideation in HIV-positive pregnant women, regardless of age, economic growth, or time, highlights a critical area for improvement in maternal and mental health care. More importantly, it is crucial to incorporate specific questions about suicidal ideation into routine prenatal care for this population, even when depressive symptoms are not apparent [41–43].

Authors’ contributions

SS: Conducting an evaluation, registering the protocol, conducting an independent evaluation, assessing the risk of bias, gathering data, analyzing it statistically, interpreting it, and writing the initial draft or revision of the final manuscript.ST-M: Independent reviewer, risk of bias assessment, statistical data analysis, revision of final manuscript.LF, EDE, LP, and LH: Independent reviewer, revision/editing of final manuscript.HA: Original idea, protocol development and edition, independent reviewer, revision/editing of final manuscript.

Funding

The review was funded by Tabriz University of Medical Sciences.

Data availability

The datasets used or analyzed during the present study are available from the corresponding author on reasonable request and as the supplementary material appendix.

Declarations

Ethics approval and consent to participate

The study is a review study. No human or animal sample were used. The study protocol was submitted and approved in the “PROSPERO” system (identifier: CRD42023450823).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.