Association of Ethnicity With Ovarian Reserve: A Systematic Review and Meta‐Analysis

ABSTRACT

Background and Aims

Ovarian reserve markers are reported to differ across ethnic groups, though the sources of this variation are unclear. We aimed to document this variation and examine the environmental, nutritional, and sociocultural factors that may explain it.

Methods

Observational studies were included if they enrolled women aged 18–45 and reported ovarian reserve markers (AMH, AFC, or FSH) stratified by an explicitly defined ethnicity classification. Four databases were searched. Risk of bias was assessed with the Joanna Briggs Institute checklist. A random‐effects meta‐analysis estimated standardized mean differences (SMD) with 95% confidence intervals (CI) and prediction intervals; heterogeneity was assessed with τ² and I². The two‐study FSH comparison was treated as exploratory.

Results

Ten studies (10,349 women, nine countries) were included in the narrative synthesis; six contributed to the AMH meta‐analysis and two to the exploratory FSH analysis. Most studies (7/10) had a low risk of bias. Individually, most studies reported higher ovarian reserve markers in White European women than in women of Middle Eastern, South Asian, or some Latin American or African descent. The pooled estimates for Asian versus European women were not statistically significant and were accompanied by extreme heterogeneity: AMH (6 studies: SMD –0.68, 95% CI –1.84 to 0.48; p = 0.19; I² = 99.1%) and FSH (2 studies: SMD –1.36, 95% CI –15.25 to 12.54; p = 0.43; I² = 89.8%). The wide FSH interval reflects the fragility of pooling only two studies.

Conclusion

The consistent, clinically relevant finding here is qualitative: individual studies repeatedly report lower ovarian reserve markers outside White European populations, most plausibly reflecting environmental, nutritional, and socioeconomic exposures rather than fixed biology. The pooled estimates are exploratory given the small number of studies and near‐total heterogeneity, and should not be over‐interpreted. Population‐specific AMH reference ranges, informed by studies that rigorously adjust for these confounders, are needed for equitable clinical decision‐making.

Trial Registration

PROSPERO Registration: CRD420251026342.

Abbreviations

AFC

antral follicle count

AMH

Anti‐Mullerian Hormone

FSH

Follicle‐Stimulating Hormone

Obs

Observational

UAE

United Arab Emirates

1. Introduction

Ovarian reserve refers to the quantity and quality of a woman's remaining oocytes, which determine her reproductive potential [1]. Ovarian reserve is positively associated with fertility outcomes and the success of in‐vitro fertilization [2]. Diminished ovarian reserve leads to a reduced number of follicles and, consequently, lower chances of conception [3]. Ovarian reserve can be quantified through hormonal biomarkers such as anti‐Müllerian hormone (AMH), antral follicle count (AFC), and follicle‐stimulating hormone (FSH), with AMH being the most reliable due to its stability across the menstrual cycle [4].

The decline in ovarian reserve is influenced by a range of factors. While aging remains the most significant determinant, genetic predisposition, autoimmune conditions, environmental exposures, lifestyle choices—such as smoking, poor diet, and chronic stress—and ethnic background, which likely reflects the cumulative effect of these environmental and sociocultural exposures, also contribute [5, 6]. Additionally, medical interventions such as ovarian surgery and chemotherapy may accelerate ovarian depletion [7]. Given the increasing prevalence of infertility, early identification of individuals at risk for diminished ovarian reserve is essential for guiding reproductive planning and optimizing fertility treatment [8].

Ethnicity, as used in clinical research, is best understood as a proxy variable that captures a complex cluster of environmental exposures, dietary patterns, socioeconomic conditions, and cultural practices—rather than a fixed biological trait. With this in mind, reported differences in ovarian reserve markers across ethnic groups likely reflect variation in lived conditions more than inherent biological differences. Begum et al., one of the studies included in this review, demonstrated this directly: Bangladeshi women who had migrated to the UK as children had substantially higher ovarian reserve than those who migrated as adults or remained in Bangladesh—showing that environmental change, not ethnicity per se, drives much of the observed variation. Nonetheless, documenting how ovarian reserve markers vary by ethnic group in clinical settings carries practical importance. Most reference ranges for AMH were established in predominantly European populations; if these thresholds are applied universally, women from other ethnic backgrounds may be incorrectly classified as having diminished ovarian reserve or may be inappropriately reassured. Studies have reported differences in AMH, AFC, and FSH across ethnic groups [9, 10, 11]. Some evidence suggests that women from certain backgrounds may show differences in the rate of ovarian reserve decline, though it remains unclear how much of this reflects environmental exposures rather than ethnicity itself [12]. In some populations, consanguineous marriages are relatively common and may increase the likelihood of autosomal recessive variants affecting ovarian function, representing one potential environmental and cultural contributor to observed variation [13, 14]. While some studies suggest an increased prevalence of primary ovarian insufficiency, others have found no association with ethnicity [15].

This systematic review and meta‐analysis documents reported variation in ovarian reserve markers across ethnic groups and examines the environmental, nutritional, and sociocultural factors that may explain the observed differences. The goal is not to identify ethnic groups with inherently superior or inferior ovarian reserve, but to determine whether clinically measured ovarian reserve markers differ systematically across populations—and whether current reference ranges, largely derived from European women, can be applied equitably across all groups. The findings are intended to inform more culturally sensitive fertility counseling and to highlight where population‐specific reference data are needed.

2. Methodology

2.1. Protocol and Registration

This study was registered with PROSPERO(CRD420251026342) and was conducted in accordance with the Cochrane Handbook [16] and reported following the PRISMA guideline [17].

2.2. Eligibility criteria

Eligibility criteria were based on population, exposure, comparison, outcome, and study design (PECOS) principles; P: Included studies enrolled women of reproductive age, typically between 18 and 45 years. Studies involving participants with known reproductive disorders (e.g., polycystic ovary syndrome, endometriosis) were eligible if ovarian reserve markers were reported separately or if statistical adjustments for these conditions were made; E: Ethnicity or racial background was considered the primary exposure. Studies were included if they reported ovarian reserve markers stratified by ethnicity or assessed the association between ethnicity and ovarian reserve; C: Studies that included comparisons between two or more ethnic or racial groups were eligible; O: Studies were required to report at least one ovarian reserve marker, including AMH, AFC, or FSH; and S: Eligible studies included observational designs such as cross‐sectional, cohort, and case‐control studies. Clinical trials were included only if they reported relevant baseline data stratified by ethnicity. Animal studies, case reports, narrative reviews, editorials, and commentaries were excluded. Regarding ethnicity classification: consistent with the PROSPERO protocol, studies were required to use an explicitly defined ethnicity or racial classification—whether participant self‐report, country of birth, a named governmental or census‐based framework, or a clearly described researcher‐assigned system. Studies that referred to ethnicity only in vague terms without stating the basis for group assignment were excluded. Given the heterogeneity in classification systems across studies, we documented how each included study defined and operationalized ethnicity, and this is reported in Table 1.

Table 1.

Characteristics of the included studies.

Author	Setting; Design	Sample size; age	Groups	Outcome measured	Main findings
Begum 2015 [9]	Bangladesh & UK; Cross‐sectional Obs	179; 35–59 yr	G1: Sedentees (Bangladesh reident, n = 36) G2: Adult Bangladesh migrants (> 16, UK, n = 53) G3: Child Bangladesh migrants (< menarche, UK, n = 40) G4: Europeans (London, n = 50)	AMH, FSH, Inhibin B, AFC	Child migrants (p < 0.01) and women of Europeandescent had a highly significantlycompared to Bangladeshi sedentees and adult migrants
Bhide 2015 [10]	UK; Cross‐sectional Obs	865: 25–45 yr	G1: Afro‐Caribbean (n = 150) G2:White European (n = 384) G3: Asian (n = 239)	Serum AMH (pmol/l)	Asian have high ovarian reserve than European
Carlos Iglesias 2014 [6]	Spain; Cross‐sectional Obs	465; ~35 yrs	G1: Spanish (n = 229) G2: Indian (n = 236)	AMH, AFC, FSH	Spanish women had lower ovarian reserve than Indian women
Bleil 2014 [11]	California; Cross‐sectional Obs	947; 25–45 yr	G1: White American (n = 277) G2: African American (n = 237) G3: Latina (n = 220) G4: Chinese (n = 213)	AMH level	The AMH levels were higher among African American compared with Latina and Chinese women
Laura Melado 2021 [18]	UAE; Cross‐sectional	2495; 19–50	G1: Arabian peninsula(Yemen, Oman, Qatar, Bahrain, Kuwait, Saudi Arabia and the United Arab Emirates)	AMH, AFC	There is reduced ovarian reserve identified in the Middle Easternethnical group
Kyweluk 2018 [19]	Mexico; Cross‐sectional Obs	97; 40–60 yrs	G1: Maya American (n = 44) G2: Non‐Maya (n = 39) G3: Unclassified (n = 14)	AMH	Maya women were over five times more likely to have undetectable AMH levels than non‐Maya women.
Schuh‐Huerta 2012 [15]	Mexico; Cross‐sectional Obs	432; 25–45	G1: White European (n = 232) G2: African American (n = 200)	AMH, FSH, AFC	AMH level was significantly higher in White European women than African American but no difference for FSH
Tabbalat 2018 [20]	USA; Cross sectional Obs	763; ~31.6 yr	G1: Arabian Peninsula (n = 217) G2: White European (n = 546)	FSH, AFC	Arabian Peninsula women had higher FSH, lower AFC
Gromski 2022 [21]	Europe & India; Cross‐sectional Obs	2758; 25–45 yrs	G1: Healthy European (n = 758) G2: Healthy Indian (n = 400) G3: Infertile Indian (n = 1600)	AMH	AMH is significantly lower in Indian women compared to Europeans
Nelson 2020 [22]	Europe & China; Cross‐sectional Obs	1348; 25–45 yrs	G1: European (n = 887) G2: Chinese (n = 461)	AMH	Chinese women had lower AMH than European women

2.3. Research Question

Do clinically measured ovarian reserve markers—including AMH, AFC, and FSH—differ across ethnic groups in women of reproductive age, and what environmental, nutritional, or sociocultural factors may account for the observed variation?

2.4. Information Sources and Search

Two authors (MR, SA) in collaboration developed the search strategy in PubMed, then independently performed an unrestricted search across four databases (PubMed, Web of Science, Scopus, and Cochrane CENTRAL) from inception to January 25, 2025 (Table 1). Reference lists of eligible papers were manually screened to identify additional studies.

2.5. Selection Process

The study selection was manually managed using Microsoft Excel 2010. Duplicates were detected based on their titles, marked in red, and sorted for easy removal. The first step involved screening titles and abstracts of all studies against the eligibility criteria. Full texts of the studies that passed this initial screening were then retrieved and carefully reviewed. Two authors (MR, SA) independently carried out the selection, with any disagreements resolved through consultation with a third author (WN), who made the final decision.

2.6. Data Collection Process and Items

Data were independently extracted by two authors (MR, SAF) using a customized form, which included key study details such as the author, study setting, and design. The total sample size and age range of participants for each group within the study were recorded. Information about the groups, including ethnic or racial background and sample sizes, was extracted. Outcome measures such as AMH, FSH, AFC, and Inhibin B were documented for each study. The main findings, particularly differences in ovarian reserve markers between ethnic or racial groups, were summarized.

2.7. Risk of Bias

The risk of bias for the included observational studies was assessed using the JBI Critical Appraisal Checklist for Analytical Cross‐Sectional Studies. Two authors (WN, SA) independently evaluated each study's risk of bias, considering factors such as study design, participant selection, control of confounder, follow up, measurement of outcomes, and statistical analysis. Any discrepancies between the authors were resolved through discussion with a third author (MR).

2.8. Effect Measures and Synthesis Measures

Data synthesis was performed using R software version 4.3.3 for studies that included data on the relevant outcome variables. Due to clinical differences such as study design, ethnicity, and patient age, a random‐effects model was used to estimate the average distribution of true effects [23]. Following recent guidelines [24], the restricted maximum likelihood method was applied to estimate heterogeneity. Heterogeneity between studies was assessed through forest plots, as well as by calculating tau² and the I² statistic [25]. To accurately interpret the results from random‐effects meta‐analyses, 95% prediction intervals were calculated for analyses involving three or more studies. This method considered the existing heterogeneity in treatment effects and provided a range of potential effects for similar future clinical trials [26]. Publication bias was assessed through a contour‐enhanced funnel plot and Egger's test.

All statistical tests were two‐sided, with a significance threshold set a priori at α = 0.05. Meta‐analyses were conducted in R (version 4.3.3) using the meta package. Standardized mean difference (SMD) with 95% confidence interval (CI) was used as the effect measure to accommodate differing assay platforms and units across studies; τ² denotes the estimated between‐study variance, and I² the percentage of total variability attributable to heterogeneity rather than chance, interpreted using conventional thresholds of low (< 25%), moderate (25%–50%), substantial (50%–75%), and high (> 75%) heterogeneity. The AMH meta‐analysis (six studies) was pre‐specified in the PROSPERO protocol. The FSH meta‐analysis was not part of the original pre‐specified analysis plan; because only two studies contributed extractable data, it is reported here as an exploratory, hypothesis‐generating analysis rather than a definitive estimate, and its results should be interpreted with substantially greater caution than the AMH comparison. No sensitivity or subgroup analyses (e.g., by ethnicity classification method or assay platform) were pre‐specified or performed, given the small number of studies available for pooling within each outcome; this is discussed further as a limitation below. As this review analyzed only previously published, de‐identified, aggregate data, ethical approval and individual informed consent were not required for its conduct.

3. Results

3.1. Study Selection

The electronic database search yielded 2495 records, with no additional records identified from registers. After removing 1069 duplicate records, 1427 records remained for screening. Of these, 1296 records were excluded based on titles and abstracts, leaving 131 full‐text reports for assessment. Upon full‐text review, 121 articles were excluded for the following reasons: ineligible outcome (n = 103), ineligible population (n = 7), irrelevant (n = 8), commentary (n = 1), review (n = 1), and unclear (n = 1). One additional record was identified via citation searching. A total of 10 studies met the eligibility criteria and were included in the final review (Figure 1).

Figure 1.

PRISMA flowchart of the study selection process.

3.2. Study Characteristics

This review included 10 cross‐sectional observational studies conducted in nine different countries (Bangladesh, China, Europe, India, Mexico, Spain, United Arab Emirates, United Kingdom, and United States), comprising a total of 10,349 women with age ranges from 19 to 60 years. Most studies used AMH (n = 9, 90%), FSH (n = 4, 40%), AFC (n = 5, 50%), and, in some cases, inhibin B as markers of ovarian reserve. Most studies identified significant ethnic disparities. In a bicultural study by Begum et al., child Bangladeshi migrants and European women in the UK had significantly higher ovarian reserve markers compared with adult Bangladeshi migrants and Bangladeshi residents (p < 0.01). Bhide et al. found that Asian women had higher AMH levels than White Europeans, suggesting better ovarian reserve in Asian populations. In Carlos Iglesias et al., Indian women in Spain showed higher AMH and AFC compared to Spanish women. Bleil et al. reported that African American women had higher AMH levels than Latina and Chinese women, while Schuh‐Huerta et al. observed higher AMH in White European women than African Americans, although no significant difference in FSH was found. Studies focusing on Middle Eastern populations indicated lower ovarian reserve. Melado et al. reported reduced AMH and AFC in women from the Arabian Peninsula, a finding corroborated by Tabbalat et al., who found higher FSH and lower AFC in this group compared to White European women in the US. In a population‐based study in Mexico, Kyweluk et al. found that Maya women were over five times more likely to have undetectable AMH compared with non‐Maya women. A large multinational comparison by Gromski et al. reported significantly lower AMH levels in both healthy and infertile Indian women compared to healthy Europeans. Similarly, Nelson et al. found that Chinese women had lower AMH levels than European women (Table 1).

3.3. Risk of Bias in Studies

The risk of bias assessment for the included studies is summarized in Table 2. Most studies demonstrated a low risk of bias across the majority of domains, particularly in the clear definition of exposure and outcomes, valid and reliable measurement techniques, and the use of appropriate statistical analyses. Out of the 10 studies, seven (Bleil 2014, Melado 2021, Kyweluk 2018, Schuh‐Huerta 2012, Tabbalat 2018, Gromski 2022, and Nelson 2020) were rated as having a low overall risk of bias, primarily due to consistent performance across all criteria, despite the follow‐up‐related domains being marked as “not applicable” (N/A), which is acceptable for cross‐sectional designs. In contrast, three studies (Begum 2016, Bhide 2015, and Carlos Iglesias 2014) were judged to have a moderate risk of bias. These studies generally failed to adjust for confounders, and one study (Bhide 2015) also did not apply appropriate statistical analyses, increasing concerns about the reliability of their findings. The domains related to follow‐up were marked as not applicable (N/A) across all studies due to cross‐sectional designs.

Table 2.

Risk of Bias for cross sectional observational studies through Joanna Briggs Institute (JBI).

Study	1. Were exposure factors clearly defined?	2. Was the cohort recruited in an acceptable way?	3. Were confounders identified and adjusted for?	4. Were groups comparable at baseline?	5. Was exposure measured validly and reliably?	6. Were outcomes clearly defined and measured reliably?	7. Was the follow‐up period adequate?	8. Was follow‐up complete, and were losses to follow‐up accounted for?	9. Were appropriate statistical analyses used?	10. Were results clearly reported?	Overall Risk of Bias
Begum 2015	Yes	Yes	No	Yes	Yes	Yes	N/A	N/A	Yes	Yes	Moderate
Bhide 2015	Yes	Yes	Yes	Yes	Yes	Yes	N/A	N/A	No	Yes	Moderate
Carlos Iglesias 2014	Yes	Yes	No	Yes	Yes	Yes	N/A	N/A	Yes	Yes	Moderate
Bleil 2014	Yes	Yes	Yes	Yes	Yes	Yes	N/A	N/A	Yes	Yes	Low
Laura Melado 2021	Yes	Yes	Unclear	Yes	Yes	Yes	N/A	N/A	Yes	Yes	Low
Kyweluk 2018	Yes	Yes	Yes	Yes	Yes	Yes	N/A	N/A	Yes	Yes	Low
Schuh‐Huerta 2012	Yes	Yes	Yes	Yes	Yes	Yes	N/A	N/A	Yes	Yes	Low
Tabbalat 2018	Yes	Yes	Yes	Yes	Yes	Yes	N/A	N/A	Yes	Yes	Low
Gromski 2022	Yes	Yes	Yes	Yes	Yes	Yes	N/A	N/A	Yes	Yes	Low
Nelson 2020	Yes	Yes	Yes	Yes	Yes	Yes	N/A	N/A	Yes	Yes	Low

3.4. Comparing AMH and FSH Between Asian and European Ethnicities

The meta‐analysis included six studies comparing AMH levels between Asian and European populations. The pooled SMD was negative (SMD = −0.68; 95% CI: −1.84 to 0.48), nominally suggesting lower AMH in Asian than European women, but this difference was not statistically significant (p = 0.19). Heterogeneity was extremely high (I² = 99.1%), and the 95% prediction interval (−3.95 to 2.61) spanned from a substantial apparent Asian deficit to a substantial apparent European deficit, indicating that the pooled estimate is not a reliable summary of the true effect and should be interpreted with caution. A separate, exploratory meta‐analysis of FSH was performed using only two studies; this comparison was not part of the pre‐specified analysis plan and is reported for completeness rather than as a substantive finding. The pooled SMD was negative (SMD = −1.355; 95% CI: −15.25 to 12.54) and not statistically significant (p = 0.43); the extremely wide confidence interval, spanning from a very large apparent Asian deficit to a very large apparent European deficit, reflects the inherent statistical instability of pooling only two studies and means this estimate carries little independent evidential weight. Heterogeneity was high (I² = 89.8%), though τ² could not be reliably estimated from only two studies. Overall, neither the AMH nor the FSH meta‐analysis provided statistically significant or reliable evidence of an ethnic difference; given the extreme heterogeneity for both outcomes and the very small number of studies contributing to the FSH estimate, these pooled results should be regarded as exploratory rather than conclusive, and the qualitative pattern seen across the individual studies (Table 1) should be weighted more heavily than these meta‐analytic point estimates (Table 3; Figures 2, 3, and 4).

Table 3.

Results of meta‐analyses (≥ 2 studies) comparing AMH and FSH between Asian and European ethnicities.

Outcome	n	SMD [95% CI]	p	τ2 [95% CI]	I2 [95% CI]	95% prediction
AMH	6	−0.68 [−1.84; 0.48]	0.19	1.19 [0.45; 7.48]	99.1% [98.8%; 99.4%]	−3.95; 2.61
FSH	2	−1.355 [−15.25; 12.54]	0.43	2.169 [—]	89.8% [62.4%; 97.2%]	—

Abbreviations: CI, confidence interval; n, number of studies; P, p‐value; SMD, standardized mean difference; ‐, not calculable.

Figure 2.

Forest plot for comparing AMH between Asian and European ethnicities.

Figure 4.

Forest plot for comparing FSH between Asian and European ethnicities.

3.5. Publication Bias

Visual inspection of the funnel plot for AMH levels comparing Asian and European populations suggested slight asymmetry (Figure 5). Egger's test indicated no statistically significant evidence of small‐study effects (p = 0.06), although the result approached significance. Most studies were distributed symmetrically around the mean effect size, but a few studies with larger standard errors deviated to the left, suggesting a potential for mild publication bias. With only six studies contributing to this comparison, Egger's test is substantially underpowered to detect funnel‐plot asymmetry, and this result should be regarded as inconclusive rather than reassuring (Figure 5).

Figure 5.

Funnel plot for AMH for Asian versus European ethnicity.

4. Discussion

This systematic review and meta‐analysis examined reported variation in ovarian reserve markers across ethnic groups, drawing on 10 cross‐sectional observational studies covering 10,349 women across nine countries. The majority of studies measured AMH (n = 9; 90%), with AFC (n = 5; 50%) and FSH (n = 4; 40%) also assessed. Before interpreting the findings, it is important to be explicit about what this review can and cannot tell us. Ethnicity, as used across the included studies, is not a biological category—it is a social and demographic variable that tracks a cluster of environmental conditions, dietary patterns, geographic histories, and socioeconomic circumstances. Observed differences in ovarian reserve markers between ethnic groups, therefore most plausibly reflect variation in these upstream determinants rather than fixed genetic traits. The clearest evidence for this comes from within our own dataset: Begum et al. [9], in a study explicitly titled “Ethnicity or environment,” found that Bangladeshi women who had migrated to the UK as children had significantly higher ovarian reserve than adult Bangladeshi migrants and Bangladeshi women who had never migrated. The migrant women had not changed their ethnicity; what changed was their environment. This finding is consistent with an environmental explanation for the broader pattern observed across this review, though it derives from a single study and does not itself establish that migration or environmental change explains the differences reported by the other included studies. The environmental hypothesis remains, at this stage, an interpretation that is consistent with the available evidence rather than one this review directly tests; confirming it will require future studies that explicitly measure and adjust for the relevant environmental and socioeconomic exposures.

Most of the included studies reported that White European women were observed to have higher ovarian reserve markers than women of Middle Eastern, South Asian, and African descent in the populations studied. However, the pooled estimates were not significant for AMH (6 studies: SMD –0.68, 95% CI –1.84 to 0.48; 95% PI –3.95 to 2.61; p = 0.19) or FSH (2 studies: SMD –1.36, 95% CI –15.25 to 12.54; p = 0.43) between Asian and European women. The meta‐analysis was restricted to Asian versus European comparisons because only these groups had sufficient studies with extractable data to permit pooling; Middle Eastern, Latin American, and African populations were represented by single studies each, precluding formal meta‐analysis for those comparisons. Both pooled estimates showed very high heterogeneity (I² = 99.1% for AMH; I² = 89.8% for FSH). While high heterogeneity is expected given the multicentric and multi‐ethnic nature of the included studies, a meaningful proportion of this variability can be attributed to identifiable methodological differences. Studies differed substantially in the commercial AMH assay platforms used, which are known to produce systematically different absolute values and introduce meaningful inter‐assay variation.24 Additionally, age distributions varied across studies, which is particularly important given the steep age‐related decline in AMH; studies enrolling older participants inherently report lower values irrespective of ethnicity. Differences in participant selection—some studies recruiting from fertility clinics, others from the general population — introduced a further source of structural heterogeneity that cannot be resolved through statistical adjustment alone. The wide 95% prediction interval for AMH (–3.95 to 2.61) reflects this uncertainty and indicates that in future studies, the true effect could plausibly range from a meaningful Asian deficit to a meaningful European deficit. Given these limitations, the pooled SMD should be interpreted as an approximation of the average effect across a heterogeneous body of literature rather than a precise estimate of the true ethnic difference. This also raises a design question worth stating directly: the AMH meta‐analysis was restricted to an aggregate Asian‐versus‐European comparison because these were the only two groupings with enough studies to pool, yet the review's own thesis is that ethnicity labels of this breadth obscure more than they reveal. The pooled estimate is therefore best read as a starting point that motivates the subgroup analysis in Figure 3, not as the review's primary finding.

Figure 3.

Forest plot for subgroup comparison of AMH between Asian and European ethnicities.

Several important factors likely contributed to this high degree of heterogeneity. Firstly, within the broad “Asian” category, there is immense genetic, environmental, and sociocultural diversity [3, 27]. South Asians, East Asians, and Middle Eastern populations, although grouped together in some studies, exhibit markedly different reproductive patterns, nutritional statuses, health behaviors, and genetic predispositions [28]. Additionally, assay variability played a significant role, as different studies used different AMH assay kits and had variations in sample handling, leading to potential measurement biases [29]. Another critical source of heterogeneity was the age distribution of participants. Given the natural age‐related decline in ovarian reserve, studies with older average participant ages might have reported lower AMH and higher FSH levels, thus influencing pooled estimates [13, 30].

Furthermore, social and environmental determinants deserve consideration as contributing factors to the observed ethnic variability. Two of the included studies—Melado et al. and Tabbalat et al.—specifically reported reduced ovarian reserve in women from the Arabian Peninsula, a region where consanguineous marriage is relatively prevalent. Consanguinity increases the risk of autosomal recessive variants affecting ovarian function, and while none of the included studies formally adjusted for consanguinity as a covariate, its potential confounding role cannot be dismissed in populations where it is common [13, 31]. Recent studies have identified mutations related to primary ovarian insufficiency that are more frequent in consanguineous populations, leading to reduced ovarian follicular pools and earlier decline in ovarian reserve [13, 14]. Malnutrition, particularly during fetal development, infancy, and adolescence, has lasting effects on reproductive capacity and may partially account for lower ovarian reserve markers observed in South Asian and some Latin American populations [32]. Nutritional deficiencies common in parts of Asia and among migrant populations—including vitamin D deficiency and protein‐energy malnutrition—can impair folliculogenesis and are associated with lower AMH levels [33, 34]. These genetic and environmental factors were inadequately controlled for in most included studies, and their interaction likely accounts for a meaningful portion of the observed heterogeneity.

Several limitations of this review must be acknowledged. The most fundamental is the heterogeneity of ethnicity classification across included studies. Studies used different and often incompatible approaches—self‐report, country of birth, researcher assignment—and the same label (e.g., “Asian”) referred to entirely different population groups depending on the country of study. This makes pooling across studies conceptually problematic, regardless of statistical approach, and is a primary reason the I² exceeded 99%. Future studies must adopt standardized, explicitly defined classification frameworks—such as those used by the UK Office for National Statistics or the US Office of Management and Budget—and report the basis for group assignment transparently. Second, the cross‐sectional design of all included studies precludes any causal inference regarding the observed variation [35] in ovarian reserve markers between groups. Most studies also failed to adjust for key confounders—BMI, smoking, alcohol intake, physical activity, and socioeconomic status—all of which independently influence ovarian reserve and many of which differ systematically between the ethnic groups being compared. The inability to separate ethnic group membership from these environmental and socioeconomic factors is a core interpretive limitation: it means this review cannot conclude that particular ethnic groups inherently have higher or lower ovarian reserve, only that ovarian reserve markers were observed to differ across unadjusted, heterogeneous study populations that also differed systematically in exposures unrelated to ethnicity per se. Relatedly, we did not perform sensitivity or subgroup analyses—for example, restricting the AMH meta‐analysis to studies sharing a single ethnicity classification method or AMH assay platform—because the small number of studies contributing to each outcome left insufficient studies to do so meaningfully; this limits our ability to identify which specific sources of heterogeneity drove the pooled estimates and is a further reason the point estimates reported here should be treated as hypothesis‐generating rather than definitive. Additionally, the extreme statistical heterogeneity substantially reduces confidence in the pooled estimates, and publication bias cannot be fully excluded despite a non‐significant Egger's test (p = 0.06). Certain ethnic subgroups—notably Southeast Asians, indigenous populations, and sub‐Saharan African women—were underrepresented, and findings cannot be extrapolated to these groups. Finally, no included study linked ovarian reserve markers to actual reproductive outcomes, which limits the clinical relevance of the observed variation.

Future studies should use longitudinal designs, following women from early reproductive age to midlife across clearly defined and consistently classified ethnic groups. Critically, they must measure and adjust for the environmental and socioeconomic factors—nutrition, migration history, vitamin D status, socioeconomic deprivation, and healthcare access—that most likely explain the variation currently attributed to ethnicity. Standardized AMH assay platforms and reporting protocols should be adopted across sites to eliminate inter‐laboratory variability. Subgroup analyses should treat “Asian” and “African” not as monolithic categories but as diverse groupings requiring separate treatment. The ultimate aim should be developing population‐appropriate AMH reference ranges that allow clinicians to assess ovarian reserve in context, rather than applying thresholds derived from one population to all.

5. Conclusion

Clinically measured ovarian reserve markers vary across ethnic groups, and this variation is more plausibly explained by differences in environmental conditions, nutrition, and socioeconomic circumstances than by fixed biological traits, although this review's design does not allow that explanation to be directly tested. The Begum et al. migration study in our dataset offers proof of concept for an environmental explanation, showing that ovarian reserve markers differed by migration history within a single ethnic group; whether the same mechanism accounts for the differences reported across the other included studies remains a hypothesis for future research to test directly, not a conclusion this review establishes. The meta‐analysis found no statistically significant difference in AMH or FSH between Asian and European women, though this result is constrained by a very small number of studies and near‐total between‐study heterogeneity, limiting how much can be concluded either way. The more consistent and clinically relevant observation across the qualitative synthesis is that standard AMH reference ranges —developed predominantly in European populations — may not apply universally, particularly for women of Middle Eastern, South Asian, and Latin American descent. Developing population‐appropriate reference ranges, informed by research that rigorously adjusts for environmental confounders, should be the next priority for this field.

Author Contributions

Mahnaz Raees: methodology, validation, conceptualization, visualization, writing – original draft, writing – review and editing, funding acquisition, investigation, software, formal analysis, project administration, data curation, supervision, resources. Shahzadi Saima Hussain: conceptualization, investigation, funding acquisition, writing – original draft, methodology, validation, visualization, writing – review and editing, resources, supervision, data curation, project administration, formal analysis, software. Sumayya Asif: funding acquisition, investigation, conceptualization, writing – original draft, writing – review and editing, visualization, validation, methodology, software, formal analysis, project administration, data curation, supervision, resources. Syed Suleman Faisal: funding acquisition, writing – original draft, validation, methodology, visualization, writing – review and editing. Syed Abdullah Faisal: funding acquisition, writing – original draft, writing – review and editing, visualization, validation, methodology. Shehroz Khan: methodology, validation, visualization, writing – review and editing, writing – original draft, funding acquisition. Waqas Naseem: conceptualization, investigation, funding acquisition, writing – original draft, methodology, validation, visualization, writing – review and editing, software, formal analysis, project administration, data curation, supervision, resources.

Use of Artificial Intelligence

No generative artificial intelligence tools were used to write, analyse, or generate any content, data, or results reported in this manuscript. AI‐assisted grammar/language‐polishing software was not used beyond standard word‐processing spell‐check.

Funding

The authors have nothing to report.

Ethics Statement

The authors have nothing to report.

Consent

The authors have nothing to report.

Conflicts of Interest

The authors whose names are listed in title page certify that they have NO affiliations with or involvement in any organization or entity with any financial interest (such as honoraria; educational grants; participation in speakers’ bureaus; membership, employment, consultancies, stock ownership, or other equity interest; and expert testimony or patent‐licensing arrangements), or non‐financial interest (such as personal or professional relationships, affiliations, knowledge or beliefs) in the subject matter or materials discussed in this manuscript.

Data Access and Authorship Approval

All authors have read and approved the final version of the manuscript. Waqas Naseem, the corresponding author and manuscript guarantor, had full access to all of the data in this study and takes complete responsibility for the integrity of the data and the accuracy of the data analysis.

Permission for Reuse of Images

Not applicable. Figure 1 (PRISMA flowchart) and all other figures in this manuscript are original and were created by the authors for this study; no image or figure from a previously published source has been reused or adapted.

Transparency Statement

Waqas Naseem, the manuscript guarantor, affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained.

Raees M., Hussain S. S., Asif S., et al., “Association of Ethnicity With Ovarian Reserve: A Systematic Review and Meta‐Analysis,” Health Science Reports 9 (2026): e73211, 10.1002/hsr2.73211.

Data Availability Statement

All data analyzed in this systematic review and meta‐analysis are derived from previously published sources, each of which is cited within the article. The data extraction spreadsheet and analysis code compiled by the authors for this review are available from the corresponding author upon reasonable request.