Association between latent tuberculosis infection and assisted reproductive outcomes: a systematic review and meta-analysis.

Tuberculosis (TB), a chronic infectious disease caused by Mycobacterium tuberculosis (Mtb), is a leading cause of death worldwide. In 2023, over 10.8 million new TB cases were reported globally, resulting in 1.25 million deaths, with the number increasing annually. 1 Five high-burden countries, including India, Indonesia and China, account for 56% of the global TB burden. 1 Studies have demonstrated that approximately 10%–20% of patients with pulmonary TB (PTB) also have genital TB (GTB). 2 5 Moreover, GTB primarily affects females of reproductive age, with 80%–90% of cases occurring in those aged 20–40. It is a common cause of infertility, as it can cause tubal obstruction (affecting 90% of cases), endometrial impairment (50%–60%) and ovarian dysfunction (10%–30%), thereby resulting in infertility. 6 8 With the recent increase in the use of assisted reproductive technology (ART) for women with infertility, such as artificial insemination and in vitro fertilisation (IVF) and embryo transfer (ET) (IVF-ET), the incidence of miliary TB during pregnancy has increased. The successful establishment of pregnancy through ART leads to elevated levels of oestrogen, progesterone and human chorionic gonadotropin, which can suppress T-cell function and increase vascular permeability. These changes collectively facilitate the reactivation of latent TB infection (LTBI) and the haematogenous dissemination of Mtb, ultimately giving rise to miliary TB. 9 10 This highlights the urgent need to strengthen TB screening and prevention before assisted reproduction in patients with infertility. LTBI is a state in which the body exhibits a persistent immune response to Mtb antigens without overt active TB. 11 A large-scale epidemiological study in 2015 confirmed that the prevalence of LTBI in the general population of China was 18.8%. 12 Approximately 5%–13% of women of reproductive age are estimated to have LTBI. 13 With the widespread use of ART, an increasing number of patients with infertility are becoming pregnant through IVF-ET. In areas with high TB prevalence, including China and India, clinical screening of infertility patients has revealed that approximately 9%–27.1% are affected by LTBI. 2 14 In China, the prevalence of infertility is estimated to be between 7% and 10%, fuelling more than one million cycles of ART treatment annually. 15 Therefore, the potential risks should not be underestimated. Research has identified that latent GTB (LGTB) adversely affects the ovarian reserve in infertile women, primarily manifesting as a significant decline in anti-Müllerian hormone (AMH) levels and antral follicle counts. 16 However, other studies reported that LTBI does not affect pregnancy outcomes in women. 2 14 Given these conflicting findings, there is an urgent need to conduct policy-oriented research on preventive treatment for infertile women with LTBI. Therefore, a meta-analysis is needed to aggregate studies on pregnancy outcomes in women with infertility affected by LTBI undergoing treatment with ART to clarify whether LTBI affects pregnancy outcomes and to provide guidance for clinical practice.

We performed a systematic review and meta-analysis. The meta-analysis of observational studies was conducted in accordance with the Meta-analysis Of Observational Studies in Epidemiology (MOOSE) guidelines. The findings are reported following the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines. 17 18 The protocol for this review is registered on PROSPERO (CRD42024605623). We used the recommended Grading of Recommendation, Assessment, Development and Evaluation (GRADE) process to summarise findings. 19 Two researchers independently searched the Embase, Web of Science and PubMed databases for studies published between inception and 1 September 2025, without language restrictions. The search terms used included Infertility , Latent Tuberculosis and Pregnancy Outcome , employing both MeSH terms and free-text terms. These keywords were identified by consulting related articles and reviews on similar topics. The search queries are listed in online supplemental table 1 . Two researchers independently screened all the included studies based on their titles and abstracts. Any disagreements were resolved by consultation with a third party. Articles were excluded if they did not pertain to the effect of LTBI on infertility or if they lacked necessary data for quantitative analysis (eg, outcome measures or adjusted effect estimates). The second screening step involved a full-text review and selection based on the inclusion and exclusion criteria. The inclusion criteria were as follows: (1) case–control or cohort studies; (2) comparison of at least two groups (LTBI and non-LTBI groups); (3) LTBI diagnosed based on tuberculin skin test (TST) or interferon-gamma release assay (IGRA) results 20 ; (4) availability of OR or relative risk to assess the relationship between LTBI and pregnancy outcomes after artificial insemination or IVF-ET in patients with infertility, with corresponding 95% CIs or sufficient data to calculate these ratios. The exclusion criteria were as follows: (1) studies not meeting the LTBI diagnostic criteria, such as TB infection diagnosed through reproductive system pathology and (2) studies with overlapping datasets from previously published studies. Two researchers independently evaluated the quality of all included studies, extracted and entered the data. The authors independently verified the validity of the quality assessment and data extraction. Discrepancies were resolved through discussion with a third investigator, if necessary. The extracted data included the first author, year of publication, location, population, study design (cross-sectional, case–control or cohort study), characteristics of the patients (sample size, age and LTBI diagnostic criteria), primary and secondary outcomes, OR and 95% CI for the primary outcomes. 21 Basic patient data are displayed in table 1 . ART, assisted reproductive technology; ICSI, intracytoplasmic sperm injection; IGRA, interferon-gamma release assay; IUI, intrauterine insemination; IVF/ET, in vitro fertilisation and embryo transfer; LTBI, latent tuberculosis infection; TST, tuberculin skin test. Two researchers independently assessed the quality of the included studies using the Newcastle-Ottawa Scale (NOS) scoring criteria. The NOS is a quality assessment tool for cohort studies and randomised controlled trials that include the representativeness of exposed and non-exposed groups, ascertainment of exposure, confirmation that outcomes of interest were not present at the start of the study, comparability between groups and outcome measurements. Each item was judged as either a ‘yes’ or ‘no’. After quality assessment, Stata V.18 software was used to perform a meta-analysis of the outcome indicators. 22 The primary outcome indicators of this meta-analysis were clinical pregnancy rate, live birth rate and miscarriage rate between patients with and without LTBI. The I² statistic was used to assess heterogeneity among the combined studies (I²≥50% indicates significant heterogeneity). Depending on the degree of heterogeneity, a fixed-effects or random-effects model was applied (I²≤50% for a fixed-effects model, 50%<I²≤75% for a random-effects model). 23 The model selection was prespecified in the study protocol based on the anticipated clinical and methodological heterogeneity among studies, with statistical heterogeneity (I² statistic) serving as a complementary criterion. To evaluate the robustness of our meta-analysis results, we performed sensitivity analyses using the leave-one-out method for all primary outcomes. This method involved systematically excluding each study one at a time and recalculating the pooled effect estimates to determine whether any single study disproportionately influenced the overall results. Publication bias was assessed using funnel plots and Egger’s test.

After reviewing the titles, abstracts and full texts, four studies met the criteria for inclusion in the meta-analysis. Figure 1 illustrates a flow chart of literature screening process, including the keywords used during the search, the number of articles identified in the databases (PubMed, Embase and Web of Science), the number of studies excluded and the reasons for exclusion. The characteristics of the included studies are summarised in table 1 . 2 13 14 24 Three of these articles were published in 2023, while our centre’s ambispective cohort study, while our centre’s cohort study, registered on ClinicalTrials.gov in 2019 (registration number NCT04443283 ) with its protocol published (Gai et al 6 ), was recently completed and has been published in 2025. 24 Four studies provided data on the comparison between LTBI and non-LTBI among pregnant women (see online supplemental table 2 ). All of them reported composite pregnancy outcomes including clinical pregnancy rate, miscarriage rate and live birth rate. This study assessed the quality of evidence using the GRADE approach. The evidence quality was rated as low for clinical pregnancy rate and live birth rate, downgraded mainly for imprecision, and as moderate for miscarriage rate, downgraded for risk of bias inherent in observational studies. None of the outcomes were downgraded for inconsistency or indirectness. Overall, the certainty of the evidence ranged from low to moderate (see online supplemental table 3 ). Among the four included studies, three were retrospective cohort studies, 2 13 14 and one was an ambispective cohort study. 24 All the studies used clear grouping methods, although whether they used allocation concealment strategies remained unclear. All studies provided detailed explanations regarding which patients were excluded from the study and who were lost to follow-up. Data from all studies were complete and selectively reported. One study had a relatively small sample size, which could have led to a selection bias. The quality of included studies was assessed using the NOS. Based on predefined thresholds (scores ≥7: low risk; 5–6: moderate risk; ≤4: high risk), the NOS scores for the four included studies were 6, 7, 6 and 7, corresponding to moderate, low, moderate and low risk of bias, respectively. Detailed scoring criteria are provided in online supplemental table 4 . All four studies reported clinical pregnancy rate. The studies exhibited no heterogeneity (p=0.136, I²=45.9%), and a fixed-effects model was used for the meta-analysis. No significant difference was observed in the clinical pregnancy rate between the LTBI and non-LTBI groups (OR 0.98, 95% CI 0.91 to 1.06, p=0.692). The forest plot is illustrated in figure 2A . Four studies reported the miscarriage rate. No evidence of heterogeneity was identified among the studies (p=0.756, I²=0), and a fixed-effects model was used for the meta-analysis. The results demonstrated a significantly higher miscarriage rate in the LTBI group than in the non-LTBI group (OR 1.14, 95% CI 1.00 to 1.31, p=0.049). The forest plot is illustrated in figure 2B . Four studies reported on live birth rate. The studies exhibited no significant heterogeneity (p=0.349, I²=8.9%), and a fixed-effects model was employed for the meta-analysis. The results indicated no significant differences in live birth rate between the LTBI and non-LTBI groups (OR 0.96, 95% CI 0.88 to 1.04, p=0.305). The forest plot is displayed in figure 2C . Sensitivity analyses were conducted using the clinical pregnancy, miscarriage and live birth rates as indicators. After sequentially excluding each study, miscarriage rate did not differ significantly from those before exclusion, indicating that the findings of this study were robust (see online supplemental figure 1 ). Egger’s tests were performed using clinical pregnancy, miscarriage and live birth rates as indicators. The results indicate a low likelihood of publication bias for clinical pregnancy rate and miscarriage rate (p=0.148 and p=0.332, respectively). However, the analysis suggests potential publication bias for the live birth rate (p=0.029) (see online supplemental table 5 ). To address potential publication bias, a trim-and-fill sensitivity analysis was performed using Duval and Tweedie’s non-parametric method. The adjusted pooled effect size remained consistent with the original analysis (95% CI 0.90 to 1.06), suggesting that the overall conclusion of the meta-analysis is robust. The scatter plots for each study were within the inverted funnel range and were generally symmetrical (see online supplemental figure 2 ).

Our meta-analysis revealed that LTBI is associated with a significantly higher miscarriage rate in infertile patients undergoing assisted reproduction compared with those without LTBI. However, no statistically significant differences were found in clinical pregnancy and live birth rates between LTBI and non-LTBI groups. These findings suggest that LTBI could adversely affect pregnancy maintenance, leading to an increased risk of miscarriage. This result underscores the importance of considering LTBI as a potential risk factor in reproductive outcomes for women undergoing assisted reproductive treatments, especially in TB-endemic regions. Our study highlights the need for further investigation into preventive strategies for LTBI to improve reproductive outcomes in this population. LTBI, defined as a persistent immune response to Mtb antigens, currently lacks definitive diagnostic standards. 11 25 The studies included in our meta-analysis used common diagnostic methods, namely the TST and IGRA. A retrospective cohort study conducted at our centre analysed infertile patients with untreated old/inactive PTB on chest X-ray and discovered that those patients had significantly lower clinical pregnancy and live birth rates after IVF-ET than those with normal chest X-ray. 26 This meta-analysis focused on the widely used TST and/or IGRA methods to determine LTBI status. LTBI significantly increases the risk of progression to active TB, particularly in high-risk populations (eg, those with HIV or immune suppression). 27 It is crucial to maintain immune balance, whether in the context of latent or active TB. 28 Successful implantation and pregnancy rely on factors such as uterine receptivity, endometrial regeneration and cytokine regulation. Immune factors associated with LGTB, such as vascular endothelial growth factor, various cytokines, leukaemia inhibitory factor, as well as cell adhesion molecules like E-cadherin, mucin-1, MECA-79 and alphavbeta3 integrin, may disrupt the endometrial microenvironment, leading to impaired endometrial receptivity, interference with embryo implantation and early development, and an increased risk of miscarriage. 29 Furthermore, the persistent immune activation state in patients with LTBI may affect systemic immune regulation and disrupt immune tolerance mechanisms during early pregnancy; the chronic inflammatory state may also activate the coagulation system, increasing the risk of microthrombus formation and compromising placental blood perfusion, ultimately contributing to early miscarriage. 10 In LGTB, a highly activated immune disorder characterised by T receptor cell and CD4 + lymphocyte activation has been observed, potentially contributing to implantation failure. 13 Additionally, the immunosuppressive effects of progesterone during IVF may amplify Th1 responses and reduce CD4 + T-cell proliferation, potentially exacerbating clinical manifestations in LTBI-positive patients. 30 Currently, international guidelines recommend preventive anti-TB treatment for high-risk groups within the general population to reduce progression to active TB and improve overall health outcomes. However, there are no specific guidelines addressing preventive anti-TB treatment in patients with infertility and LTBI, leaving this unique population without targeted recommendations. With the increasing use of IVF-ET among women with infertility, a growing number of reports have been observed on miliary PTB during pregnancy, mostly occurring in China, which has a high TB burden. This finding highlights the importance of enhanced TB screening and prevention. Some researchers advocate routine screening for LTBI in women with infertility before assisted reproductive treatment. 9 29 31 However, some studies showed no significant differences in pregnancy outcomes between women with and without LTBI undergoing IVF-ET. The findings of our meta-analysis highlight a critical issue: although clinical pregnancy and live birth rates were not significantly affected by LTBI, the miscarriage rate was notably elevated in patients with LTBI. This suggests that LTBI could negatively affect pregnancy maintenance, particularly by increasing the risk of miscarriage. Diagnosing LTBI requires a thorough exclusion of active TB, especially in infertile patients, where screening for GTB is just as crucial as screening for PTB. GTB is a common cause of female infertility, particularly in low-income and middle-income countries, with most cases secondary to PTB. 8 32 Globally, the incidence of infertility due to GTB ranges from 44% to 74%, with studies in India exhibiting a rate of 58%, underscoring the importance of early diagnosis. 2 33 Despite the established benefits of anti-TB treatment in high-risk populations, such as those with HIV, no guidelines support its use in LTBI-positive patients with infertility, primarily due to concerns about potential side effects, including liver and kidney toxicity, allergic reactions and delays in reproductive treatment. 34 This meta-analysis may enable countries to implement proactive and effective public health measures for TB control. 35 36 Therefore, further randomised controlled trials are urgently needed to evaluate the efficacy and safety of preventive anti-TB treatment in women with LTBI-positive infertility. Some studies suggest that anti-TB treatment, particularly based on endometrial TB PCR positivity, may improve assisted reproductive outcomes in infertility. 37 Research on LGTB or subclinical GTB has demonstrated that women with LGTB have significantly lower AMH levels and antral follicle counts than those in infertile women without LGTB, indicating diminished ovarian reserve. 16 However, anti-TB treatment in LGTB-positive women has been associated with improved embryo quality, implantation rate and pregnancy rate compared with those in infertile controls without LGTB, suggesting that anti-TB treatment improves the ovarian microenvironment in patients with LGTB. 16 Such studies could provide crucial evidence regarding whether treating LTBI before assisted reproduction can improve pregnancy outcomes by reducing the risk of miscarriage. Several potential confounding factors should be considered when interpreting our findings. Maternal age, infertility aetiology (eg, tubal factor, ovulatory disorders, endometriosis), lifestyle factors and differences in assisted reproductive techniques (such as IVF vs ICSI and fresh vs frozen ET) may all influence pregnancy outcomes. Among the included studies, some adjusted their analyses for age and infertility duration, while others did not provide sufficient details regarding adjustment for confounding variables. 2 13 14 24 Therefore, residual confounding cannot be ruled out. Importantly, our recently published prospective cohort study using multivariate analysis demonstrated that infertile patients with chest radiographic evidence of old TB lesions and positive IGRA results had a significantly higher risk of miscarriage and cumulative miscarriage rate compared with IGRA-negative patients. 24 These findings are consistent with the present meta-analysis and further highlight that LTBI deserves careful attention in the evaluation and management of infertility. Nevertheless, given the observational nature of both our cohort and the included studies, the current evidence should be interpreted as an association rather than a causal effect. This study had several limitations. First, the limited number of studies included in the meta-analysis may reduce the generalisability of our findings, the statistical power of the publication bias assessment is limited, and the results should be interpreted with caution. Second, all four studies were observational, highlighting the need for future randomised controlled trials on anti-TB treatment for LTBI and its impact on pregnancy outcomes. Additionally, since all four studies were conducted in China, a high TB-burden region, the findings may not be directly applicable to low-burden settings. Therefore, further research in diverse populations is warranted to enhance the generalisability of these results.

Based on low to moderate certainty evidence from observational studies, LTBI may be associated with an increased risk of miscarriage in infertile patients undergoing assisted reproduction, while it appears to have no significant effect on clinical pregnancy or live birth rates. Given the inherent limitations of the available evidence, further high-quality, well-controlled studies, particularly randomised controlled trials evaluating preventive anti-TB therapy, are needed to conclusively determine the effects of LTBI and to guide clinical management strategies for this patient population.