Systematic review and meta-analysis of complementary and alternative medicine use during pregnancy in Iran

Abstract

Objective

This systematic review and meta-analysis aimed to determine the prevalence, patterns and associated factors of complementary and alternative medicine (CAM) use among pregnant women in Iran.

Design

A systematic review and meta-analysis of observational studies.

Data sources

A comprehensive search was conducted in PubMed/MEDLINE, Web of Science, Scopus, ScienceDirect and major Iranian databases from inception to 30 November 2024.

Eligibility criteria for selecting studies

We included observational studies published in peer-reviewed journals that assessed CAM use among pregnant women in Iran and reported prevalence estimates or relevant associated factors.

Data extraction and synthesis

Data extraction and quality assessment were performed independently by two reviewers using a standardised form and the Quality Assessment Tool. Meta-analyses of proportions were performed using the DerSimonian and Laird random effects model. The between-study heterogeneity was assessed using the I-squared (I²) statistic. Subgroup analysis, 95% prediction intervals (95% PrIs) and sensitivity analysis were conducted to explore the sources of heterogeneity and to evaluate the robustness of the overall effects, respectively. This study followed Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and was registered with PROSPERO.

Results

20 studies with a total of 8106 participants were included. The pooled prevalence of CAM use was 49% (95% CI 39% to 59%). The between-study heterogeneity was very high (I²=94.69%) with a wide 95% PrIs range of 3%–94%. The very high between-study heterogeneity and the wide range of PrI for the pooled prevalence were not explained by the quality of the studies, geographical regions of Iran or the methods of data collection. Mentha longifolia, Zataria multiflora and Boswellia thurifera were the most commonly used herbs. Across the included studies, the first trimester of pregnancy was most frequently reported as the period of CAM use, with socioeconomic factors identified as significant predictors. Family and friends were the primary sources of information regarding CAM. The most common reasons for CAM use included gastrointestinal issues, respiratory problems such as colds and coughs, and the desire to improve general health. Only 31% of participants disclosed their use of CAM to their healthcare provider.

Conclusions

Nearly half of pregnant women in Iran use CAM, yet disclosure to healthcare providers is low. An actionable implication is the critical need to integrate routine enquiry about CAM use into standard antenatal care to ensure safe practice. Further research is recommended to evaluate the effectiveness, safety and outcomes of CAM use among pregnant women.

PROSPERO registration

CRD42024618490

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• A comprehensive and systematic literature search was performed across multiple international and regional databases to identify all relevant studies.

• To ensure accurate statistical synthesis of prevalence data, we employed the Freeman-Tukey double arcsine transformation for variance stabilisation and the Metaprop command to keep CIs within a logical range.

• Between-study heterogeneity was addressed by using the DerSimonian and Laird random effects model. We further investigated sources of heterogeneity through subgroup and sensitivity analyses, and reported 95% prediction intervals.

• A key methodological limitation is the high statistical heterogeneity, which is inherent in prevalence meta-analyses and was exacerbated by inconsistencies in measurement tools (eg, varied questionnaires and reporting items) across the included primary studies.

Introduction

Complementary and alternative medicine (CAM) encompasses a wide range of diagnostic and therapeutic interventions that are not typically included in conventional medical care. This includes herbal remedies, homeopathic treatments, manipulative and body-based therapies, and energy-oriented modalities such as acupuncture.^{1 2} Recent studies indicate a rising prevalence of CAM use, particularly medicinal plants, in developing countries, with usage rates reaching as high as 90%.^{3 4} The holistic approach and cost-effectiveness of CAM have greatly contributed to its widespread adoption in modern times.^{5 6}

Overall, CAM usage is reported more frequently by females than by males,⁷ and it is especially prevalent among pregnant women worldwide.¹ Studies indicate that 69% of pregnant women in the USA and 51% in Germany use CAM.⁸ The rising CAM use among pregnant women is primarily aimed at alleviating pregnancy-related issues such as nausea, vomiting, back pain, common colds and respiratory diseases.⁹ The most popular modalities used include massage, vitamin and mineral supplements, herbal remedies, relaxation therapies and aromatherapy.¹⁰ Nevertheless, the use of CAM products during pregnancy, especially when not recommended by healthcare providers, is not consistently supported by definitive scientific research.¹¹ There is potential for complications, such as teratogenic effects, embryonic toxicity or harm to newborns, which highlights the need for further research in this area to establish clearer conclusions.¹²

In Iran, several studies have assessed the use of CAM among pregnant women. However, no research has specifically reviewed or explored the prevalence of CAM use in this population. The present study aims to systematically review descriptive studies on various types of CAM used by pregnant women in Iran to evaluate multiple items, such as the frequency of CAM use, identifying the most commonly used types of CAM, determining the primary reasons for its use and understanding the indications for CAM among Iranian pregnant women.

Methods

The study protocol has been reviewed and approved by the Medical Ethics Committee of Shiraz University of Medical Sciences with approval ID: IR.SUMS.REC.1403.262. This systematic review was conducted using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and checklist.¹³ The protocol for this review was registered in PROSPERO (identifier: CRD42024618490).

Search strategy

The researcher identified keywords and developed a search strategy. Relevant keywords were extracted using tools such as Medical Subject Headings. An appropriate search strategy was tailored to each database. The search strategy was Peer Review of Electronic Search Strategies—checked before implementation.

A comprehensive systematic search was performed across multiple international and regional databases, including PubMed/Medline, Scopus, Web of Science, Google Scholar, Magiran and the Scientific Information Database. The final search was conducted on 30 November 2024, with no time limits or language restrictions applied. The complete and detailed search strategies, including all keywords and syntax used for each database, are provided in online supplemental file (search strategy) to ensure full reproducibility.

Eligibility criteria

The studies were selected based on these inclusion criteria: descriptive studies conducted in Iran that examined the prevalence and pattern of CAM use in pregnant women. Studies were excluded if they met any of the following criteria: (1) ineligible study design: clinical trials, review articles, commentaries, letters or any non-observational study design. (2) Ineligible population: studies not conducted on a population of pregnant women in Iran. (3) Ineligible outcome: studies that did not investigate or report quantitative data on the prevalence or use of CAM or herbal medicines

Literature screening and evaluation process

The screening and selection of articles were conducted independently by two authors (SMN and MH) in three stages. Initially, 815 articles were identified during the search. The subsequent screening process involved a complete review of titles, abstracts and full texts to identify studies that met the inclusion criteria.

Study selection and screening process

The search results from all databases were imported into EndNote X7 reference management software for deduplication. The remaining unique citations were then uploaded to Microsoft Excel to facilitate the screening process. The study selection was performed independently by two reviewers (SMN and MH) in two stages: title and abstract screening, based on the predefined eligibility criteria. Full-text assessment: the full text of potentially relevant studies was retrieved and screened independently by the same two reviewers. Any disagreements at either stage were resolved through discussion. If a consensus could not be reached, the senior reviewer (MHH) adjudicated the decision.

Data extraction and management

Data from the included studies were extracted independently by two reviewers (SMN and MH) using a standardised data extraction form in Microsoft Excel. The form was designed to capture the following data: study characteristics, including first author, publication year, study location, design, sample size, data collection method and response rate. The primary outcome was the prevalence of CAM use among pregnant women. Secondary data: types of CAM used, sources of information about CAM, disclosure of CAM use to a healthcare provider, reasons for CAM use, pregnancy trimester for CAM use and factors associated with a higher probability of CAM use. The completed extractions were compared, and any discrepancies were identified and resolved through consensus or, when necessary, by consultation with the third reviewer (MHH).

Quality assessment

We used the Quality Assessment Tool (QAT) to evaluate the quality of the articles included in our study. This tool was developed by Bishop et al,¹⁴ and is based on the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement, which outlines essential criteria for reporting observational research, including surveys. The tool assesses four domains: study methods (four points), sampling (two points), reporting of individuals’ characteristics (three points) and CAM usage (six points). The original tool has a maximum score of 17 points. However, one item, ‘recall bias—low risk’ (two points), is only applicable to prospective cohort studies. Given that all studies included in our review were retrospective or cross-sectional in design, this item was deemed not applicable. Therefore, the maximum achievable score was adjusted to 15 points to accurately reflect the relevant quality domains for our study context. QAT was categorised as follows: a total score of 0–5 was defined as high risk of bias, 6–10 as moderate risk and 11–15 as low risk of bias.

Evidence synthesis and statistical analyses

The primary outcomes were the pooled prevalence estimates of CAM use, the most prevalent medicinal herbs and CAM use disclosure, along with their 95% CIs. To avoid double-counting, if a study reported multiple results on the prevalence of CAM use, the prevalence encompassing the greatest number of CAM items was included in the CAM use meta-analysis. The prevalence of the three most frequently used herbs was analysed separately. Data analysis was performed using the Metaprop command in Stata V.17 (StataCorp, College Station, TX), a tool specifically designed for conducting meta-analyses of proportions.¹⁵ The Metaprop command enables the inclusion of studies reporting proportions of 0% or 100% in the meta-analysis while ensuring that CIs remain within the 0–1 range.¹⁵ Moreover, the Freeman-Tukey double arcsine transformation was applied to stabilise variances.¹⁶ The Wilson score interval was used to calculate the corresponding 95% CIs.¹⁷ Meta-analyses of proportions were conducted using the DerSimonian and Laird random effects model¹⁸ for the prevalence of CAM use and CAM use disclosure.

Additionally, heterogeneity among the studies was evaluated using the Cochrane Q test and the I-squared statistic (I²). A Cochrane Q test yielding a p value below 0.05, alongside an I² value greater than 30%, was interpreted as evidence of statistically significant heterogeneity.¹⁹ I² values were categorised as follows: 30%–60% indicating moderate heterogeneity, 50%–90% representing substantial heterogeneity and values of 75% or higher reflecting considerable heterogeneity between studies. Subgroup analyses, 95% prediction intervals (95% PrIs) and sensitivity analyses were conducted to explore sources of heterogeneity and evaluate the robustness of the overall effects, respectively. Subgroup analyses were based on the data collection method (questionnaire/interview), geographical region of Iran (north/centre/south) and risk of bias assessment results (high, moderate or low risk). The 95% PrIs were computed to represent the expected range of true prevalence values in future comparable settings. Since these PrIs were derived on a transformed scale, the lower bounds occasionally assumed negative values; for clarity, such values were truncated at 0%. Furthermore, a leave-one-out sensitivity analysis was performed, in which each study was sequentially excluded, and the meta-analysis was repeated using the same random-effects model and transformation as in the primary analysis. The pooled prevalence estimates from these iterations were then compared with the overall pooled estimates to assess the stability of the findings and identify any study exerting disproportionate influence.

Publication bias was assessed using funnel plot analysis, along with Egger’s test²⁰ and Begg’s adjusted rank correlation test²¹ to detect statistical asymmetry. The trim and fill method was planned to be used to detect and adjust for potential publication bias identified in the meta-analysis results. Statistical analyses were performed using Stata, V.17 (StataCorp, College Station, TX). A p value of ≤0.05 was considered statistically significant.

Patient and public involvement

Patients and/or the public were not involved in the design, conduct, reporting or dissemination plans of this research.

Results

We selected 20 eligible studies out of 815 articles identified through our search strategy. The selection process is shown in figure 1. A list of all studies excluded after full-text assessment, with specific reasons for their exclusion, is provided in online supplemental table 1.

Figure 1. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart of the study screening and selection. WOS, Web of Science.

Characteristics of the included studies

Research has been conducted on the prevalence of CAM use among pregnant women in various regions of Iran. These regions include the North (Mazandaran,22,24 Guilan²⁵ and Sari²⁶), the South (Ahvaz,²⁷ Kazeroon,²⁸ Shiraz,²⁹ Khuzestan,³⁰ Kerman^{31 32} and Shahrekord³³), the Centre (Tehran,^{34 35} Isfahan³⁶ and Kashan³⁷), the East (Mashhad³⁸ and Bojnord³⁹) and the West (Tabriz^{40 41}). 14 studies focused primarily on herbal medicine, while six examined various traditional medicine (TM)/CAM modalities. Data collection methods in this research area primarily rely on face-to-face interviews and questionnaire-based approaches, including both self-report and semistructured formats. Sample sizes in the studies varied from 150³¹ to 919³⁸ participants. Response rates ranged from 80.20% to 100%, with three studies achieving full participation^{23 32 37} (table 1).

Table 1. Summary table: characteristics of the included studies in the systematic review.

No	Author	Study aims	Location	Sample size	Response rate (%)	Participants’ mean age (SD)	Mode of data collection	CAM studied
1	Abdolahi et al22	Investigating the level of recognition and use of complementary and traditional medicine in pregnant women	Mazandaran	400	87.5	27.9±5.2	Questionnaire	TM/CAM
2	Abdollahi and Yazdani Cherati24	Investigate the prevalence of and characteristics related to the use of HM among pregnant women	Mazandaran	320	80.20	27.97±5.24	Self-report questionnaire	Herbal medicine
3	Azari et al41	Determine the frequency and type of herbal medicine used by pregnant women	Tabriz	318	–	–	Questionnaire	Herbal medicine
4	Biazar et al25	Explore the pattern of herbal use during pregnancy	Rasht	836	95	≤20; 219 (25.3) 20–30; 467 (54.1) ≥30; 178 (20.6)	Face-to-face interview	Herbal medicine
5	Hashem et al34	Determining the frequency of the use of herbal medicines in a group of pregnant women	Tehran	618	97	27.03±4.8	Face-to-face interview	Herbal medicines
6	Ekrasarian et al26	Assessing the knowledge, attitude and practice of pregnant women on using medicinal plants during pregnancy	Sari	307	–	28.19±4.8	Questionnaire	Herbal medicine
7	Ghasemi et al23	Determining the frequency of consumption of medicinal plants and its relationship with health literacy and the attitude of pregnant women towards medicinal plants	Mazandaran	322	100	20–40 (83.2)	Questionnaire	Herbal medicine
8	Goli et al36	Assess the status of using CAM by women in the pregnancy period	Isfahan	600	99.5	28.5	Face-to-face interview	CAM
9	Hosseini et al39	Evaluate the prevalence of medicinal herbs’ consumption in pregnant women	Bojnord	350	–	25.8±5.6	Questionnaire	Herbal medicine
10	Karimian et al37	Categorise the predictors of self-medication with herbal remedies during pregnancy based on the theory of planned behaviour	Kashan	300	100	28.7±5.4	Face-to-face interview	Herbal medicine
11	Khadivzadeh et al38	Determining the use of TM/CAM during pregnancy and the reasons behind it from the mother’s perspective	Mashhad	919	–	25.7±5.1	Self-report questionnaire	TM/CAM
12	Khandan et al32	Investigate the application of CAM in pregnant women	Kerman	226	100	≤18; 3 (1.33) 18–35; 200 (88.5) ≥35; 23 (10.17)	Face-to-face interview	CAM
13	Molavi Vardanjani et al29	Evaluate the use of CAM products and their related factors among pregnant women	Shiraz	365	85	29.30±5.40	Face-to-face interview	CAM
14	Saber et al31	Identify all types of medicinal plants and synthetic medicines used by pregnant women	Kerman	150	–	26.2±0.4	Face-to-face interview	Herbal medicine
15	Sattari et al40	Evaluate the use and attitudes of pregnant women towards herbal medicine use in pregnancy	Tabriz	400	–	26.4±5.2	Questionnaire	Herbal medicine
16	Sereshty et al33	Determining the consumption of medicinal plants in pregnant women	Shahrekord	447	–	26.8±5.95	Face-to-face interview	Herbal medicine
17	Shabani et al35	Evaluate the frequency of medicinal plant usage and its influencing factors among pregnant mothers living in Tehran	Tehran	326	–	28±6.31	Questionnaire	Herbal medicine
18	Soleymani and Makvandi27	Investigate the rate of use of herbal medications during pregnancy and some of the related factors	Ahvaz	219	–	25.8±5.4	Face-to-face interview	Herbal medicine
19	Tabatabaee28	Evaluate the herbal drug utilisation pattern during pregnancy	Kazeroon	513	96.8	25.5±4.7	Face-to-face interview	Herbal medicine
20	Yazdi et al30	Understand the use of TM/CAM modalities in pregnant women of different ethnic groups in southern Iran	Khuzestan	170	–	26.6±5.67	Questionnaire	TM/CAM

n (%), data presented as age categories.

CAM, complementary and alternative medicine; HM, herbal medicine; TM, traditional medicine.

Patients' demographic

Overall, 8106 pregnant women participated in these studies. The mean age ranges from 25.5 to 29.3 years. Two studies provide a more detailed breakdown of age groups, showing that most participants are within the reproductive age group of 18–35 years.^{25 32}

Prevalence of CAM use

The lowest prevalence was 19% reported by Soleymani and Makvandi et al in Ahvaz²⁷; and the highest prevalence was 84% reported by Khadivzadeh and Ghabel in Mashhad³⁸ (table 2).

Table 2. Pattern of CAM use among pregnant women in the included studies.

Author	Three most frequently used CAM/herbs (%)	Prevalence of CAM usage (%)	The most common time of use of CAM	Effect of CAM use (%)	Most common CAM use reason	Most common indication for CAM use	Adverse effects of CAM use	Main information sources	Disclosing CAM use to their physician (%)	The main cause of not disclosing CAM use to physicians	Variables that increase the probability of CAM use
Abdolahi et al22	CAM Medicinal plants (9.5%) Prayer therapy (5.6%) Hydrotherapy (5%)	39.2%	The entire pregnancy (39.2%)	43.8%	Previous beneficial experience of oneself or others (49%) Being safe or having no side effects (39%)	–	–	–	–	–	–
Abdollahi and Yazdani Cherati24	Herb Citrus aurantium (30.97%) Mentha piperita (19.81%) Borago officinalis (19.46%)	48.4%	Third trimester of pregnancy (25.96)	–	–	Promote fetal health and intelligence (35.40%) Promote the women’s health status (32.5%) Relieve discomfort during pregnancy (25.1%)	–	Friends and relatives (30.5%) Their mothers (26.3%) Mass media (11%)	50%	Perceiving the safe use (31%) Perception of disclosure as unimportant (19%)	Prior use of herbs (OR 36.1 (95% CI 13.4 to 97.1)) Levels of education (upper secondary) (OR 1.78 (95% CI 1.02 to 3.12)) Lived in their own house (OR 2.15 (95% CI 1.37 to 3.38))
Azari et al41		31.4%	Second trimester of pregnancy (36%)	–	–	Common cold (36%) Pleasant taste (10%)		Family (70%)	25%	–	–
Biazar et al25	Herb Cinnamomum verum (n=34; 15.4%) Mentha longifolia (n=33; 14.9%) Descurainia Sophia (n=29; 13.1%)	19.6%	–	–	–	Reducing blood sugar (15.4%; n=34) GI problems/UTI (14.9%; n=33) Common cold (13.1%; n=33)	–	–	–	–	Higher education (p=0.001) Urban residence (p=0.008)
Dabaghian et al34	Herb M. piperita (n=197; 32.8%) Boswellia thurifera (n=158; 26.3%) D. Sophia (n=148; 24.7%)	67%	First trimester of pregnancy	–	Less side effects than conventional medicine (59%) Prefer herbal medicine to conventional medicine in pregnancy (28%) Used as routine in family (26.4%)	Bloating and stomach ache (30.2%) Respiratory infections (18.7%) Nausea and vomiting (11.5)	–	Family and friends (60.2%) Book, newspaper, magazine and internet (49.7%)	37.8%	Lack of inquiry from medical providers (43%) Perception of disclosure as unimportant (17.2%)	Gestational age (p=0.02) Family size (p=0.001) Children number (p <0.001) Medical insurance (p=0.04) Attitude (OR 0.89 (95% CI 0.86 to 0.92))
Ekrasarian et al26	Herb M. longifolia (31%) C. aurantium (28.5%) D. sophia (22%)	48%	The entire pregnancy (61.2%)	–	Be safe and have no complications (80.1%)	Maintaining the health of the mother and the baby	–	By their mothers (37.8%)	–	–	–
Ghasemi et al23	Herb M. longifolia (n=65; 48.5%) B. officinalis (n=53; 39.5%) Zingiber officinale (n=43; 32%)	41.6% (n=134)	First and third trimester of pregnancy (65.5%)	–	–	Digestive problems (nausea, flatulence and constipation) Respiratory problems (cold, cough)	–	Physician Internet	22.4%	Lack of inquiry from medical providers (50.6) Perception of disclosure as unimportant (25.9)	Attitude towards medicinal plants (p <0.001) Disease history (p <0.05)
Goli et al36	CAM Prayer (n=370; 61.7%) Medicinal plants (n=111; 18.5%) Massage (n=39; 6.4%)	66.8%	–	–	–	–	–	–	–	–	Being employed (p=0.02) Higher education (p <0.001)
Hosseini et al39	Herb M. longifolia (47.7%) Zataria multiflora (46.8%) Camellia sinensis (28.8%)	63.4%	–	–	–	Common cold (33.9%) Sedation (27.1%) Stomach upset (25.2%)	–	Family (53.4%) Radio and TV (22.3%)	57.3%	–	Husband’s occupation (p= 0.047) Husband’s education (p=0.035)
Karimian et al37	–	1%	–	–	–	Heartburn (n=82; 28.6%) Constipation (n=80; 27.9%) Nausea (n=75; 26.1%) Anaemia (n=67; 23.3%)	–	–	–	–	Subjective norms among pregnant women (p <0.05)
Khadivzadeh and Ghabel38	CAM Herbal preparations (49.2%) Diet/nutrition change (42.3%) Vitamin and mineral (30.1%)	83.7%	First trimester of pregnancy (64.3)	42.9%	Natural and safe to use, especially for the fetus (53%)	GI disorders (39.3%)	3.3% report mild or moderate form of nausea, gastric pain, low blood pressure and urticarial	Obstetricians and midwives (41.5%) Friend, colleague and family members (33.5%)	–	–	–
Khandan et al32	CAM Herbal medicine (n=188; 82.81%) Praying (n=18; 7.93%) Cupping (n=7; 3.10%) Herb Honey (n=139; 21.89%) M. longifolia (n=71; 11.18%) Matricaria chamomilla (n=64; 10.08%)	83.62% (n=189)	–	41.15% (n=93)	Safety due to fewer side effects, 26.99% (n=61)	Morning sickness (n=61; 24.02%) Cold (n=46; 18.11%)	–	Family, friends and acquaintances (65.93%) Personal study (14.61%) Media (13.27%) Healthcare providers (6.19%)	21.24%	–	Gestational age (p=0.003) OR 0.056
Molavi Vardanjani et al29	Herb Cichorium intybus (33%) Phoenix dactylifera (18%) B. thurifera (17%)	56.92%	Third trimester of pregnancy	51.83% of all participants believed using CAM had been effective for them	–	Improve their general condition (52%) Manage digestive problems (44%)	–	Relatives, friends and acquaintances (44.43%)	27%	–	Participant’s opinions on CAM side effects (p=0.009) Opinions on side effects of chemical drugs (p <0.001) Having a history of CAM use, either in previous pregnancies (OR 9.8; p <0.001)
Saber et al31	Herb M. longifolia (30%) B. officinalis (19.3%) B. thurifera (18.7%)	71.3%	First trimester of pregnancy (42.5)	–	–	Improving GI problems and reducing nausea (66.6%) Strengthening the nerves and sedation (26%) Increasing the fetus’ IQ (20%)	–	Their own diagnosis (54.7%) Recommendations from others (36.3%) Physicians’ prescription (0.9%)	–	–	Housewives (OR 4.4 (95% CI 1.1 to 17.5)) Person’s perspective on the safety of herbal medicine (OR 4.5 (95% CI 2.0 to 10.2))
Sattari et al40		22.3%	–	–	Be safe use herbal remedies during pregnancy (39.8%)	–	–	Physician (46.1%) Self-medication (44.9%)	–	–	–
Sereshti et al33	Herb Echium amoenum D. sophia (8%)	51.9 %	Third trimester of pregnancy (40.6%)	57/4	–	Common cold (29.1%) Abdominal pain (17.1%) Induction of labour (11.7%)	Allergic reactions (7.1%)	Relative (74.2) Family members (66.8) Physician (2.4)	44.1%	Being natural and harmless herbs (39.1) Lack of inquiry from medical providers (34.8) Fear of the doctor’s negative attitude (10.6)	Level of education (p=0.004)
Shabani et al35	Herb Z. officinale (n=25; 7.66%) M. piperita (n=18; 5.52%) Z. multiflora (n=16; 4.90%)	32.2%	–	–	Less adverse effects in comparison with conventional drugs (n=23; 21.9%) Good experience reported by family and friends (n=19; 18%) More effective in comparison with conventional drugs (n=6; 5.7%)	Gastric symptoms (nausea/morning sickness, bloating, stomachache and constipation) Cold symptoms The tendency to increase the neonate’s IQ		Family or friends (35%) Own data (14.3%) Books, magazines and herbal sellers (7.6%) Medical doctors and midwives (11.3%) Googling (6.6%) Others (4.7%).	–	–	–
Soleymani and Makvandi 27	Herb Z. multiflora (n=17; 40.4%) Olea europaea (n=11; 26.1%) Peganum harmala (n=8; 19%)	19.2% (n=42)	–	–	–	Cold Cough	–	Family and friends (38.1%) Internet (28.6) Healthcare providers (21.4%) Scientific books and journals (11.9%)	17.7%	–	Level of education (p=0.001)
Tabatabaee28	Herb Trachyspermum ammi (n=50; 22.6%) Z. multiflora (n=28; 12.6%) Ocimum basilicum (n=28; 12.6%)	30.8%	First trimester of pregnancy (36.7)	–	–	GI problem (32.1%) Nausea and vomiting (20.2%)	–	Family (87.3%) Primary maternity care providers (7.6%)	–	–	–
Yazdi et al30	CAM Herbal medicine (93%, n=74) Dry herbal water vapour (6.3%, n=5) Animal-derived drugs (1.3%, n=1)	46.5%	First trimester of pregnancy and third trimester		Effectiveness (n= 61; 77%) Availability (n=8; 10%) Affordability (n=6; 7%) Safety (n=4; 6%)	–	98.7% (78) of them reported that they had no complications from TM/CAM use	Recommendation of friends and family (89.9%)	11.1%	–	Ethnicity (being Arab) (p=0.014)) History of complications in previous pregnancy (OR 4.31 (95% CI 2.13 to 8.74)) Higher gestational age (OR 1.065 (95% CI 1.02 to 1.10))

CAM, complementary and alternative medicine; GI, Gastrointestinal; TM, traditional medicine; UTI, urinary tract infection.

Type of CAM used

13 studies have identified medicinal plants that are common among pregnant women. Based on the results, Mentha longifolia has been used more for gastrointestinal (GI) problems, Zataria multiflora for cold and cough, and Boswellia thurifera for increasing children’s intelligence. Additionally, five studies reported on CAM use during pregnancy. The most popular CAM modalities were herbal medicine (37.27%), prayer therapy (17.71%) and dietary supplements (16.7%).

Timing of CAM use during pregnancy

12 of the included studies reported information on the timing of CAM use during pregnancy. Due to heterogeneity in reporting formats and the lack of consistent trimester-specific prevalence data at the individual participant level, a patient-level meta-analysis of trimester-specific CAM use was not feasible. Therefore, the following findings represent a study-level descriptive summary, based on the number and proportion of studies identifying each trimester as the most common period of CAM use.

Specifically, the first trimester was reported as the most common period of CAM use in 50% of studies (6/12).^{22 26 28 31 34 38} The third trimester was identified as the peak period of CAM use in 25% of studies (3/12).^{24 29 33} Two studies (17%) reported that CAM use was common in both the first and third trimesters,^{23 30} whereas one study (8%) reported the second trimester as the predominant period of use.⁴¹ Overall, this study-level summary suggests that the first trimester is most frequently reported as the primary period of CAM use among pregnant women in Iran; however, these findings should not be interpreted as pooled trimester-specific prevalence estimates at the patient level.

Motivations behind CAM use

Eight studies explored the reasons for using CAM. Among users, 44.5% reported choosing CAM due to its perceived safety or fewer side effects. Moreover, 11.2% of pregnant women mentioned that previous positive experiences influenced the decision to use CAM. Other common reasons for preferring CAM include its effectiveness,³⁰ preference for herbal medicine,³⁴ availability³⁰ and affordability.³⁰

Predictors of CAM use

13 studies identified various factors that increase the probability of CAM use. These predictors were categorised into three main domains, as detailed below.

Socioeconomic factors

Education level was highlighted in four studies.^{24 25 27 33} One study indicated that ethnicity also influenced CAM use.³⁰ Additionally, employment status (being employed,³⁶ housewives³¹), husband’s occupation and education,³⁹ medical insurance,³⁴ living in their own house,²⁴ and urban residence²⁵ were other predictors of CAM use.

Attitude and belief towards CAM

This encompasses attitudes towards medicinal plants,²³ participants’ opinions on the side effects of CAM,²⁹ individual views on the safety of herbal medicine³¹ and subjective norms among pregnant women.³⁷

Gestational age, family characteristics and medical history

Gestational age,^{30 32 34} family size and number of children³⁴ expressed as factors that increase the probability of CAM use.

Also, the history of CAM use in previous pregnancies,^{24 29} disease history,²³ side effects associated with conventional medications²⁹ and complications from previous pregnancies³⁰ were identified as significant predictors of CAM use.

A synthesis of the quantitatively reported adjusted effect sizes (table 3) reveals a clear hierarchy among these predictors. A history of CAM use emerged as the strongest predictor, with an exceptionally high OR (36.1). Socioeconomic factors such as home ownership (OR 2.15) and education level (OR 1.78) demonstrated modest but significant associations. Notably, a positive perception of herbal medicines’ safety was a strong attitudinal predictor (OR 4.5), comparable in strength to a history of obstetric complications (OR 4.31). The association with gestational age was inconsistent, with one study reporting a slight increase in odds per week (OR 1.07) and another reporting a counter-intuitive, strong negative association (OR 0.06) that warrants cautious interpretation.

Table 3. Summary of factors associated with CAM use during pregnancy in Iran.

Predictor category	Specific factor	Effect size (95% CI)	Supporting study
Socioeconomic and demographic	Education	OR 1.78 (95% CI 1.02 to 3.12)	24
	Lived in their own house	OR 2.15 (95% CI 1.37 to 3.38)	24
	Employment (housewives)	OR 4.4 (95% CI 1.1 to 17.5)	31
Attitudes and beliefs	Attitude	OR 0.89 (95% CI 0.86 to 0.92)	34
	Perception of herbal medicine safety	OR 4.5 (95% CI 2.0 to 10.2)	31
Medical and obstetric history	History of CAM use	OR 36.1 (95% CI 13.4 to 97.1)	24
	History of CAM use in previous pregnancies	OR 9.8	29
	Complications in the previous pregnancy	OR 4.31 (95% CI 2.13 to 8.74)	30
Pregnancy-related	Higher gestational age	OR 1.065 (95% CI 1.02 to 1.10)	30
	Gestational age	OR 0.056	32

CAM, complementary and alternative medicine.

Reasons for CAM use

12 studies have specifically identified the most common complications that lead to the use of CAM during pregnancy. Among these, 30.6% of pregnant women used CAM during pregnancy for GI problems, including symptoms such as nausea, bloating, stomach aches and constipation. Respiratory problems, such as colds and coughs, were the second most common reason at 10.32%, followed by 6.9% of participants using CAM to improve their general health.

CAM information sources

16 studies indicated the most common sources of information for using CAM. Among users, 48.98% reported obtaining information from family and friends, 12.65% from healthcare providers and 10.36% from books and media.

Disclosure of CAM use

Four studies identified reasons why patients did not disclose their CAM use. A significant portion of pregnant women (34.09%) cited a lack of inquiry from medical providers as the primary reason for not disclosing their CAM usage. Other reasons included the belief CAM was safe (16.05%) and the perception of disclosure as unimportant (14.67%).

Self-reported efficacy and adverse effects of CAM use

Five studies explored the opinions of pregnant women regarding the efficacy of CAM. Remarkably, 49.46% of participants believed that using CAM had been beneficial for them.^{24 29 32 33 38} Three studies specifically addressed the side effects of CAM use. One study reported allergic reactions occurring in 7.1% of users.³³ Another study found that mild or moderate side effects, such as nausea, gastric pain, hypotension and urticaria, affected 3.3% of participants.³⁸ In contrast, one study showed that 98.7% of users reported no complications associated with TM or CAM use.³⁰ The data on CAM use are presented in table 2.

Quality assessment of the included studies

The methodological quality of the included studies is shown in table 4. The majority of studies were judged to be of high methodological quality. Additionally, the inter-rater reliability of the initial independent assessments was evaluated by categorising studies as ‘low risk’ or ‘moderate risk’ of bias. The agreement between the two reviewers was substantial (Cohen’s κ=0.73). All initial assessments, including discrepancies that were resolved through discussion, are detailed in online supplemental table 2.

Table 4. Details of the scoring of included articles using the Quality Assessment Tool.

Study	Study methods (four points)	Sampling (two points)	Reporting of individuals’ characteristics (three points)	CAM usage (six points)	Total score (out of 15)
Abdolahi et al22	1	2	3	6	12
Abdollahi and Yazdani Cherati24	3	2	3	6	14
Azari et al41	1	1	3	6	11
Biazar et al25	2	2	3	5	12
Dabaghian et al34	2	2	3	6	13
Ekrasarian et al26	1	1	3	4	9
Ghasemi et al23	2	2	3	6	13
Goli et al36	1	2	3	3	9
Hosseini et al39	2	1	3	5	11
Karimian et al37	2	2	3	5	12
Khadivzadeh and Ghabel38	1	1	3	6	11
Khandan et al32	1	1	3	6	11
Molavi Vardanjani et al29	3	2	3	6	14
Saber et al31	1	1	3	6	11
Sattari et al40	1	1	3	3	8
Sereshti et al33	1	1	3	6	11
Shabani et al35	2	1	3	5	11
Soleymani and Makvand27	1	1	3	5	10
Tabatabaee et al28	3	2	3	6	14
Yazdi et al30	2	1	3	6	12

Findings from the meta-analysis

Prevalence of CAM use

A total of 20 publications with a total number of 8106 pregnant women were included in the prevalence meta-analysis. The pooled prevalence of CAM use in pregnant women was 49% (pooled prevalence 0.49 (95% CI 0.39 to 0.59)). Between-study heterogeneity was considerable (I²=94.69%) with a wide 95% PrI range of 0.03%–0.94% (figure 2). About subgroup analyses, the considerable between-study heterogeneity of the pooled prevalence was not explained by the quality of studies, geographical region of Iran or the methods of data gathering (table 5).

Figure 2. Forest plot presenting the pooled prevalence of complementary and alternative medicine use in pregnant women (n=20, conducted using the DerSimonian and Laird random effects model).

Table 5. Meta-analysis of overall prevalence of CAM use and CAM disclosure among pregnant women, with subgroup analyses (all analyses were performed using the DerSimonian and Laird random effects model).

Study group	Meta-analysis			Heterogeneity
	Number of participants	Pooled prevalence (95% CI)	Q statistic	P within group	I2 (%)	P between group
CAM use
Overall	8106	0.49 (0.39 to 0.59)	357.86	0.000	94.69	–
Risk of bias assessment						0.75
Moderate	2152	0.50 (0.39 to 0.59)	258.70	0.000	94.97
Low risk	5954	0.46 (0.28 to 0.65)	97.69	0.000	94.88
Geographical region of Iran						0.58
North	4172	0.44 (0.28 to 0.61)	231.59	0.000	96.55
Centre	1844	0.56 (0.41 to 0.71)	31.18	0.000	90.38
South	2090	0.51 (0.36 to 0.67)	31.17	0.000	91.88
Type of data gathering						0.52
Self-administer	3832	0.45 (0.32 to 0.60)	169.74	0.000	94.70
Interview	4274	0.52 (0.38 to 0.66)	187.08	0.000	95.19
CAM disclosure
Overall	3355	0.32 (0.23 to 0.41)	59.85	0.000	–
Risk of bias assessment						0.82
Moderate	445	0.33 (0.22 to 0.43)	41.99	0.000
Low risk	2910	0.30 (0.09 to 0.51)	17.49	0.000
Geographical region of Iran						0.18
North	1310	0.39 (0.21 to 0.56)	30.74	0.000
Centre	618	0.38 (0.30 to 0.46)	0.00	–
South	1427	0.25 (0.13 to 0.37)	19.87	0.001
Type of data gathering						0.75
Self-administer	1480	0.33 (0.17 to 0.50)	42.57	0.000	42.57
Interview	1875	0.30 (0.21 to 0.40)	16.52	0.000	16.52

CAM, complementary and alternative medicine.

The most frequently mentioned herbs include M. longifolia, Z. multiflora and B. thurifera. A total of seven publications (n=2410) were included in the prevalence M. longifolia use. The pooled prevalence of M. longifolia use was 21% (k=7, pooled prevalence 0.21 (95% CI 0.10 to 0.32)), and a considerable between-study heterogeneity was observed (I²=84.64, p<0.000) (table 6). The 95% PrI was approximately 0%–57%, indicating that future studies may observe prevalences anywhere within this wide range (figure 3). Subgroup analyses did not explain the considerable between-study heterogeneity (table 6).

Table 6. Meta-analysis of prevalence of common herbal medicine use among pregnant women, with subgroup analyses (all analyses were performed using the DerSimonian and Laird random effects model).

Herbs	Study group	Meta-analysis			Heterogeneity
		Number of participants	Pooled prevalence (95% CI)	Q statistic	P within group	I2 (%)	P between group
Boswellia thurifera	Overall	3349	0.09 (0.01 to 0.16)	32.73	0.000	78.61
	Risk of bias assessment						0.27
	Moderate	219	0.009 (−0.12 to 0.65)	0.00	_	_
	Low risk	3130	0.10 (0.01 to 0.18)	31.36	0.000	80.87
	Geographical region of Iran						0.23
	North	1158	0.01 (−0.04 to 0.07)	0.10	0.751	0.00
	Centre	944	0.15 (−0.06 to 0.36)	9.95	0.002	89.95
	South	1247	0.09 (−0.005 to 0.18)	7.91	0.048	62.05
	Type of data gathering						0.26
	Self-administer	648	0.03 (−0.04 to 0.11)	0.02	0.000	0.00
	Interview	2701	0.11 (0.008 to 0.20)	30.57	0.01	83.61
Mentha longifolia	Overall	2410	0.21 (0.10 to 0.31)	39.06	0.000	84.64	0.92
	Risk of bias assessment
	Moderate	752	0.22 (0.03 to 0.40)	13.40	0.001	85.08
	Low risk	1658	0.20 (0.06 to 0.35)	22.66	0.000	86.76
	Geographical region of Iran						0.99
	North	1815	0.21 (0.06 to 0.35)	26.75	0.000	88.78
	South	120	0.21 (0.01 to 0.40)	11.70	0.003	82.91
	Type of data gathering						0.20
	Self-administer	979	0.27 (0.20 to 0.34)	169.74	2.36	14.40
	Interview	1431	0.16 (0.006 to 0.31)	187.08	20.87	85.62
Zataria multiflora	Overall	3775	0.10 (0.04 to 0.16)	26.33	0.004	69.61	0.70
	Risk of bias assessment
	Moderate	445	0.12 (0.03 to 0.22)	0.91	0.339	0.00
	Low risk	3330	0.10 (0.03 to 0.17)	24.81	0.000	75.82
	Geographical region of Iran						0.29
	North	1508	0.14 (−0.03 to 0.31)	20.43	0.000	90.21
	Centre	944	0.04 (−0.03 to 0.12)	0.01	0.937	0.00
	South	1323	0.11 (0.05 to 0.17)	3.47	0.324	13.34
	Type of data gathering						0.29
	Self-administer	998	0.16 (0.01 to 0.30)	11.10	0.004	81.98
	Interview	2777	0.07 (0.02 to 0.12)	8.54	0.13	41.43

Figure 3. Forest plots indicating the pooled prevalence of commonly used herbs ((A) Boswellia thurifera, (B) Mentha longifolia and (C) Zataria multiflora) among pregnant women (k=8, k=7 and k=9 for the meta-analyses of B. thurifera, M. longifolia and Z. multiflora, respectively). The analyses were conducted using the DerSimonian and Laird random effects model. Since these prediction intervals were derived on a transformed scale, the lower bounds occasionally assumed negative values; for interpretation, such values should be truncated at 0%.

Nine publications, encompassing a total of 3775 pregnant women, were included in the meta-analysis of the prevalence of Z. multiflora use. The pooled prevalence of Z. multiflora use was 10% (k=9; pooled prevalence 0.10 (95% CI 0.04 to 0.16)), with substantial between-study heterogeneity (I² = 69.61%, p=0.004) (table 6). The 95% PrI ranged from approximately 0% to 29%, indicating that future studies may observe prevalence rates anywhere within this wide range (figure 3). Subgroup analyses did not account for the substantial heterogeneity observed between studies (table 6).

A total of eight publications (n=3349) were included in the meta-analysis of the prevalence of B. thurifera use. The pooled prevalence of B. thurifera use was 9% (k=8; pooled prevalence 0.09 (95% CI 0.01 to 0.16)), with moderate-to-substantial between-study heterogeneity (I² = 78.61%, p<0.001) (table 6). The 95% PrI ranged from approximately 0% to 34%, suggesting that prevalence estimates in future research may fall anywhere within this broad range (figure 3). Subgroup analyses did not explain the moderate-to-substantial heterogeneity observed between studies (table 6).

Disclosure of CAM use

Out of 10 studies (n=3355) that reported on the disclosure of CAM use, 32% of users informed their physicians about it (pooled prevalence 0.32 (95% CI 0.23 to 0.41)). Between-study heterogeneity was considerable (I²=84.96%) with a wide 95% PrI range of 0.00%–0.64% (figure 4). According to table 5, the considerable between-study heterogeneity of the pooled prevalence was not explained by the quality of studies or the methods of data gathering.

Figure 4. Prevalence of pregnant women disclosing CAM use to their physicians (n=10, conducted using the Der Simonian and Laird random effects model).

Sensitivity analysis and publication bias

Sensitivity analysis demonstrated that the omission of any single study did not significantly alter the pooled prevalence estimates of CAM use and CAM disclosure. When each study was excluded in turn, the pooled prevalence varied only slightly, ranging from 47% to 51% for CAM use compared with the overall estimate of 49%, and from 29% to 34% for CAM disclosure compared with the overall estimate of 32%. All recalculated estimates fell within the original CI, indicating that no individual study exerted undue influence on the meta-analytic results. Additionally, across herbal remedies, the leave-one-out sensitivity analyses confirmed the robustness of the meta-analytic findings. The pooled prevalence estimates remained stable within narrow ranges (7%–12%, 5%–10% and 19%–24% for B. thurifera, Z. multiflora and M. longifolia, respectively) compared with the overall estimates of 10%, 9% and 21% for B. thurifera, Z. multiflora and M. longifolia, respectively. No individual study meaningfully affected the overall results (online supplemental figure 1).

According to the visual inspection of the funnel plots, there is no evidence of publication bias concerning the prevalence estimates of CAM use (online supplemental figure 2a). Additionally, the results of Egger’s test (p=0.91) and Begg’s test (p=0.94) support these findings. Regarding CAM use disclosure, visual inspection of the funnel plots indicates some degree of publication bias, which is confirmed by Egger’s test (p=0.02) and Begg’s test (p=0.03) (online supplemental figure 2b). However, the nonparametric trim-and-fill analysis did not impute any missing studies, indicating no evidence of publication bias. The adjusted pooled effect was identical to the observed estimate (0.342 (95% CI 0.308 to 0.376)), suggesting that the results are robust and unlikely to be influenced by small-study effects. For the most frequently used herbs, visual inspection of the funnel plots revealed no evidence of publication bias regarding the prevalence of B. thurifera, M. longifolia and Z. multiflora (online supplemental figure 2). Additionally, the results of Egger’s test (p=0.84, p=0.27 and p=0.11 for B. thurifera, M. longifolia and Z. multiflora, respectively) and Begg’s test (p=0.90, p=1.00 and p=0.12 for B. thurifera, M. longifolia and Z. multiflora, respectively) support these findings.

Discussion

This study explains various CAM modalities and medicinal herbs used by Iranian pregnant women, aiming to highlight the most prevalent CAM practices, reasons for CAM use and identify any patterns in CAM preferences. The findings regarding the prevalence of CAM use among pregnant women in various regions of Iran offer important insights into local practices and preferences for traditional and alternative medical approaches. In the 20 articles analysed, the average prevalence of CAM use by pregnant women was 49%, which ranged from 39% to 59%. The findings of studies conducted in other parts of the world indicated a high prevalence of CAM use among pregnant women. In Turkey, 71.5%,⁴² Northwest Ethiopia, 89.3%⁴³ and in Kenya, 50.7%⁹ of pregnant women used CAM during their pregnancy. However, unregulated or inappropriate CAM use can have sequelae for maternal and fetal health. Therefore, establishing guidelines for the safe use of CAM treatments is crucial.44,46

In this study, the most popular herbs were M. longifolia, Z. multiflora and B. thurifera. Similarly, other systematic reviews have identified these herbs as commonly used during pregnancy.^{47 48} Also, the use of B. thurifera during pregnancy has cultural significance, particularly in some Middle Eastern societies, where it is believed to enhance the intelligence of children.⁴⁹ This review revealed that medicinal herbs were the most prevalent CAM modality used by pregnant women, followed by prayer therapy and dietary supplements. Similar to the commonly used types of CAM reported in other studies.^{50 51} In this regard, a study in the Netherlands evaluated 83 prayer-related healing reports from 2015 to 2020. It found that various forms of prayer (personal, group and liturgical) often led to significant healing experiences.⁵²

The most common period for CAM use was the first trimester of pregnancy. Also, one UK study reported that two-thirds of women use CAM during early pregnancy.⁵³ CAM therapies can interact with conventional medications, potentially reducing their effectiveness or causing adverse reactions. This raises concerns, especially in the critical first-trimester pregnancy, as it may result in malformations or developmental issues in the fetus.^{42 54} Therefore, healthcare providers must openly discuss these practices with patients to mitigate the risks associated with their use during this sensitive period.

According to the included studies, a larger portion of users reported choosing CAM due to its perceived safety and fewer side effects, and prior experiences. Prasad et al⁵⁵ also showed that many pregnant women believe CAM poses fewer risks than conventional medicine and feel it does not adversely affect the fetus. Past positive experiences with CAM significantly influence current usage.⁵⁵ Nonetheless, there is scepticism among healthcare providers about safety.⁵⁶ This scepticism may hinder its integration into standard care practices and lead to a lack of communication about CAM’s potential risks and benefits, emphasising the need for improved awareness and education among healthcare professionals and patients.

The present review revealed substantial heterogeneity (I²=94.69%) with a wide predictive interval (3%–96%) for CAM use, which remained unexplained despite conducting subgroup analyses based on study quality, geographical regions of Iran and data collection methods, as well as reporting PrIs and assessing sensitivity analyses. Similar challenges were observed in the analyses of the most prevalent herb use and CAM disclosure. Consistent with our findings, other systematic reviews and meta-analyses57,61 on the prevalence of CAM use in various populations have reported similarly high heterogeneity. The observed substantial heterogeneity in the reporting of CAM use across various studies highlights significant variations in the definitions and perceptions of these practices. This inconsistency can be attributed to multiple factors, including inherent differences in research methods and study designs.⁶² In addition, Asanad,⁵⁷ in the assessment of the prevalence of CAM use in Saudi Arabia, revealed that significant inconsistencies in the definitions and measurement methods of CAM practices across the included studies might be the sources of this heterogeneity. Additionally, another meta-analysis⁶³ on the prevalence of CAM use reported that upper-middle-income countries such as Malaysia exhibited the highest prevalence of CAM usage. This may be attributed to the multicultural and multiracial diversity of Malaysian citizens, which encourages the exploration of various types of CAM available in the country.

The considerable heterogeneity observed in the pooled prevalence estimate may be substantially explained by variations in socioeconomic status across the included studies. Our findings qualitatively confirmed that the likelihood of CAM use might be influenced by socioeconomic determinants such as education, income and occupation, which subsequently shape individuals’ attitudes and access to these therapies. For instance, higher education levels are associated with increased CAM utilisation, potentially because more educated individuals tend to have greater health literacy and more positive attitudes towards alternative treatments.⁶⁴ This socioeconomic gradient is not unique to Iran; international evidence, such as from Japan, also links factors like hope, which often correlates with educational attainment, to higher CAM usage.⁶⁵ Therefore, the marked socioeconomic diversity among the study populations across different regions of Iran likely served as a key moderator, contributing to the wide variation in prevalence rates and ultimately to the high statistical heterogeneity observed in our meta-analysis. This underscores the complex interplay between socioeconomic and psychological factors in determining CAM use.⁶⁶

This systematic review shows that GI problems were a common reason for CAM use among pregnant women. This could be due to hormonal changes that affect GI motility. These symptoms can significantly impact pregnant women’s quality of life, leading many to seek relief through various treatment options, including CAM.⁶⁷ A systematic review addressed that CAM for nausea and vomiting was superior to placebo or conventional medicine, with fewer side effects.⁶⁷ However, some herbs may affect liver function or alter hormone levels.⁶⁸

Almost half of users reported obtaining information from family and friends. This highlights the importance of personal networks in shaping health-related decisions. A narrative review indicated that many users rely on these interpersonal connections for guidance.⁶⁹ The trust placed in family and friends as sources of information is consistent with broader consumer behaviour trends, where personal recommendations are often viewed as more credible than advertising or other formal sources.⁷⁰ The reliance on non-professional sources for CAM information raises concerns about the accuracy and safety of the practices being adopted.⁷¹ This reliance may lead to the spread of misinformation, underscoring the need for improved communication between healthcare providers and patients regarding CAM practices.

Only 31% of users informed their physician about CAM use, and the reasons for non-disclosure were a lack of inquiry from medical providers, the belief that CAM was safe, and the perception of disclosure as unimportant. Similar to our study, a meta-analysis study revealed a 33% disclosure rate with the same reasons for non-disclosure.⁷² However, standardised assessments have shown that discussing CAM can significantly enhance patient–provider communication, fostering a supportive environment.⁷³ Despite these benefits, many patients remain unaware of the potential risks associated with interactions between CAM and conventional treatments. One study found that 14 patients were at risk of CAM–cancer interactions.⁷⁴ Therefore, there is a critical need for improved communication and education between patients and healthcare providers. To bridge this communication gap and mitigate potential risks, the development of clear clinical guidelines is imperative. These should empower healthcare providers to address CAM use within routine prenatal care proactively.

The strengths of the present study are as follows: first, we applied the Metaprop command, which ensures that CIs remain within the 0–1 range.¹⁵ Second, the Freeman-Tukey double arcsine transformation was used to stabilise variances.¹⁶ Third, meta-analyses of proportions were conducted using the DerSimonian and Laird random effects model¹⁸ to address high between-study heterogeneity. Additionally, we employed various methods to identify the sources of heterogeneity, including subgroup analyses, 95% PrIs and sensitivity analyses.

Limitations and future perspectives

This study on the prevalence of CAM use among pregnant women in Iran, despite its scientific value, may face certain limitations at both the review and study levels. At the review level, first, although we employed a random-effects model to deal with the between-study heterogeneity, the analysis is challenged by the very high heterogeneity, typical for this type of study. This heterogeneity may stem from variations in the questionnaires used across the included studies, leading some studies to omit certain items that others included. Second, the wide 95% PrI for the assessed outcomes—prevalence of CAM use, the most popular herbs used and CAM disclosure—indicates a broad expected range of true prevalence values in future comparable settings, reflecting low generalisability of these findings. Third, a formal GRADE (Grading of Recommendations Assessment, Development and Evaluation) assessment was not performed. GRADE was originally developed to assess the certainty of evidence for comparative effect estimates, and its direct application to prevalence meta-analyses is methodologically challenging.⁷⁵ Prevalence syntheses are inherently characterised by substantial between-study heterogeneity and wide CI and PrI,⁷⁵ reflecting genuine variability across populations, settings and measurement approaches rather than methodological flaws. Applying standard GRADE domains in this context may therefore result in misleadingly low certainty ratings that do not accurately represent the descriptive nature or methodological quality of the underlying evidence.

At the study level, most studies did not report reasons for non-disclosure of CAM use to healthcare providers. Moreover, many studies lacked information on the effectiveness, safety or outcomes of CAM use among pregnant women. Given the importance of these issues and the promotion of maternal and infant safety, future research should address these gaps.

Conclusion

This systematic review and meta-analysis indicate that nearly half of pregnant women in Iran report using CAM, most commonly herbal therapies, to address a range of pregnancy-related conditions. However, the pooled prevalence estimate should be interpreted with caution, as the very high between-study heterogeneity and the wide 95% PrIs reflect substantial variability across populations, settings and study methodologies. Consequently, the true prevalence of CAM use may differ considerably between individual contexts and regions.

Despite this variability, the consistently high prevalence reported across studies underscores the widespread use of CAM among pregnant women and highlights a critical gap in communication between patients and healthcare providers, as disclosure rates remain low. These findings emphasise the need for routine enquiry about CAM use within antenatal care and for improved education of both healthcare professionals and pregnant women regarding the potential risks and benefits of CAM use during pregnancy. Further high-quality, standardised research is required to better characterise patterns of use, safety and pregnancy-related outcomes associated with CAM use in this population. In addition, based on our findings, we propose the following actionable recommendations for clinical practice and policy.

Documentation of the product and dose

When CAM use is reported, the specific product name, dosage, frequency and route of administration should be meticulously documented in the patient’s medical record to facilitate accurate risk assessment and ongoing monitoring.

Interaction counselling

Healthcare professionals need appropriate training and resources to effectively analyse possible interactions between commonly used herbs (eg, M. longifolia, Z. multiflora, B. thurifera) and conventional medications, as well as potential effects on pregnancy outcomes.

Pharmacovigilance and labelling guidance

National health authorities should establish pharmacovigilance systems specifically for monitoring adverse events related to the use of CAM during pregnancy. Furthermore, clear labelling guidelines for herbal products warning about use during pregnancy are urgently needed to protect this vulnerable population.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.