Bleeding in the second trimester is a common and distressing presentation in pregnancy, yet its clinical implications remain poorly defined and there are no universally accepted guidelines. This systematic review and meta-analysis aims to evaluate the association of unprovoked vaginal bleeding (UVB) in the second trimester and preterm birth (PTB), and to synthesise the available evidence to inform clinical practice. Across included studies, second trimester UVB was found to be significantly associated with PTB (OR: 6.72, 95%CI 3.36-17.47), highlighting its role as an important marker of adverse pregnancy outcomes. We showcase our findings as proof of a higher risk pathology that can lead to significant pregnancy outcomes. Greater awareness of this association between UVB and PTB is needed to support risk stratification, closer surveillance, as well as timely investigations and management.

Antepartum bleeding is a common complication of pregnancy which, although not always associated with adverse outcomes, remains a well-recognised risk factor for maternal and perinatal morbidity and mortality. Beyond its clinical implications, it is frequently a source of emotional distress for many patients, making timely assessment, clear communication and emotional support a central component of care. Existing literature has predominantly focused on the effects of bleeding in the first and third trimesters, where clearer diagnostic pathways and management strategies have evolved. However, bleeding in the second trimester remains comparatively underexplored, despite its clinical uncertainty and significant psychological burden for pregnant people.

Second trimester bleeding (13-24+6 weeks) affects up to 1.1% of all pregnancies [1]. It is typically medically assessed to distinguish between provoked or unprovoked causes, with no clear aetiology identified in over half of cases. Provoked bleeding would be attributed to identifiable factors such as vaginal infections, cervical ectropion or inflammation, presence of fibroids or polyps, placental abnormalities including low lying placentas or subchorionic haematomas. In these cases, reassurance can often be provided. In contrast, unprovoked vaginal bleeding (UVB) refers to bleeding without an identifiable cause following clinical assessment, where both reassurance and accurate risk stratification for the future remain challenging.

Moreover, the management of UVB in the second trimester is limited and inconsistent. Patients commonly attend maternity triage units for review where assessment may include history taking, laboratory investigations, confirmation of fetal viability and speculum examination. However, there are no specific national guidelines directing management. Care is largely expectant, with decisions regarding admission or surveillance varying between clinicians and institutions. Ultrasound follow up for UVB is not justified under the NHS during the second trimester. Fetal surveillance is not feasible until the third trimester, and fetal movements do not have an established pattern until 28 weeks. Obstetric assessment alone may not always provide adequate explanation. Thus, it can be frustrating to patients and challenging to healthcare professionals to provide support. 

Despite these limitations, second trimester UVB has been associated with adverse pregnancy outcomes including miscarriage, preterm pre-labour rupture of membranes (PPROM) and preterm birth (PTB) [1-3]. Among these, PTB represents the most clinically significant outcome, given its strong association with neonatal morbidity, risks associated with prematurity and the increased rates of neonatal intensive care unit (NICU) admission and lower birth weight [2, 4, 5]. Furthermore, recurrent bleeding episodes appear to confer an even greater risk of adverse outcomes, particularly PTB and PPROM [6].

However, the existing literature is limited by heterogeneity in definitions regarding the timing of bleeding, and the inconsistent characterisation between provoked and unprovoked bleeding. As a result, there is a lack of robust evidence addressing unprovoked second trimester bleeding and its clinical implications. Given emerging evidence suggesting that UVB is most strongly associated in a third of patients with PTB compared to other adverse outcomes(7); this study aims to characterize the risk of PTB following 2nd trimester UVB and to better define its clinical significance for patient counselling and management.

2.1 Search strategy

The present systematic review and meta-analysis aimed to comprehensively synthesize the available evidence regarding the association between second trimester per vaginal bleeding (PVB) and the risk of preterm birth (PTB), in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [8]. Adherence to PRISMA standards was intended to promote transparency, methodological rigor, and reproducibility, thereby strengthening the validity of the findings. A systematic literature search was conducted across multiple electronic databases, including PubMed and MEDLINE including articles published up to December 2025, in English. The search strategy combined the following terms: (“per vaginal bleeding” OR “vaginal bleeding” OR “antepartum hemorrhage” OR “antepartum haemorrhage” OR PVB OR “unprovoked vaginal bleeding” OR “idiopathic bleeding in pregnancy”) AND (“second trimester” OR “mid-pregnancy”) AND (“preterm birth” OR “preterm delivery” OR “PTB” OR “prematurity”) AND (“outcomes” OR “risk factors” OR “association”). No geographical restrictions were applied. Additionally, the reference lists of all included studies were manually screened to identify any further relevant publications not captured in the initial search. Two independent reviewers screened titles and abstracts, followed by full-text evaluation of potentially eligible studies. Any disagreements were resolved through discussion and consensus. The study protocol was registered in the PROSPERO international prospective register of systematic reviews (protocol number: CRD420261367031). As this meta-analysis was based exclusively on previously published data, ethical approval and patient consent were not required.

2.2 Eligibility criteria

Studies were deemed eligible if they were original research articles (prospective or retrospective cohort or case-control studies) involving pregnant women with documented unprovoked second-trimester per vaginal bleeding and reported outcomes related to preterm birth, or provided sufficient quantitative data (e.g., odds ratios, relative risks, or extractable raw data) to assess this association. Unprovoked bleeding was defined as no known cause being associated with it, being a placenta pathology such as placenta previa, vasa previa, placenta accreta, or a cervical or vaginal bleeding point. Studies were excluded if they were non-primary research articles (including reviews, meta-analyses, editorials, letters, or case reports), did not specifically address second-trimester bleeding, lacked adequate quantitative data for analysis, involved non-human subjects or were not published in English, or represented duplicate publications or overlapping study populations.

2.3 Data Extraction & Quality Assessment

Data was independently extracted by two reviewers using a standardized form. Extracted variables included study characteristics, baseline population data, and details of exposure and outcomes. Effect estimates were recorded where available; otherwise, raw data was extracted for calculations. Discrepancies were resolved by consensus. Study quality was parallelly assessed by two reviewers using the Newcastle–Ottawa Scale (NOS) as per standard procedure [9]. Discrepancies were also resolved through consensus. As per sensitivity analysis and risk of bias evaluation, studies scoring below 7 were considered of lower methodological quality and were excluded. Detailed assessments are included in Table 1.

2.4 Statistical Analysis

All analyses were conducted using IBM SPSS Statistics version 31.0. The genetic inverse variance method was used with odds ratios (ORs) converted to log scale (lnOR) and standard errors were calculated. Where not reported, ORs and confidence intervals (CI) were derived from raw data. Heterogeneity was assessed by Cochran’s Q-test and the I² statistic (p < 0>

In addition, a retrospective analysis was completed with aim to include our results and contribute to the assessment of the significance of bleeding in the second trimester. Patient electronic records of all women attending the Rosie Hospital in Cambridge, UK with antepartum vaginal bleeding between January 2019 and December 2021 were accessed. This period was chosen because of accessibility linking patients to preterm birth outcomes following the above inclusion and exclusion criteria. Records were reviewed and only women with UVB in the second trimester, from 13+0 weeks until 24+6 weeks presenting, were included. These results have not been published before. 

Table 1: The Newcastle-Ottawa scale for assessing the risk of bias of the included studies.

The search strategy yielded a total of 1858 articles, out of those 126 passed the initial article selection and under more detailed assessment only 13 satisfied the inclusion criteria and were included in the final analysis (Figure. 1). The characteristics of the included studies are summarized, including study design, sample size, and key findings (table 2). The results of the meta-analysis, including the overall effect size and confidence intervals, are presented. Subgroup analyses and sensitivity analyses are also reported to explore potential sources of heterogeneity.

Figure 1: Flowchart of the study selection procedure

Table 2: Characteristics of included studies.

3.1 Main findings 

All included studies provided data regarding the exact number and mean age of patients. In total, 420,307 patients were included in the present study and the mean age ranged from 27 to 31 years. A total of 14 studies, including our results, were included in the quantitative synthesis, encompassing 5,601 cases of second trimester per vaginal bleeding associated with preterm birth and 414,706 controls across diverse ethnic populations.

Due to substantial heterogeneity observed among the studies included (I2 = 99%, Cochrane Q test = 559.04, p-value < 0.001), the random effects model was employed. The overall odds ratio (OR) for PTB due to 2nd trimester PVB compared to PTB non-relevant to PVB was 6.72 (95% CI 3.35-13.47, p-value<0.001). All the included studies exhibited statistical significance, despite one that showed marginal significance (10) (Figure. 2).

Figure 2: Forest plot demonstrating the summary odds ratio for PTB due to 2nd trimester PVB compared to non-related to bleeding PTB. Studies are listed in chronological order.

.2 Subgroup analysis

3.2.1 By age: Subgroup analysis based on maternal age demonstrated that the association between second-trimester PVB and PTB remained statistically significant within each examined age category. The overall pooled OR estimate of 11.56 (95% CI 2.28–58.55, p < 0.001), suggesting that maternal age may influence the strength of the association between PVB and PTB. (Figure.3)

Figure 3: Forest plot of the subgroup analysis stratified by maternal age, examining the association between second-trimester per vaginal bleeding (PVB) and preterm birth (PTB).

3.2.2 By smoking status, hypertension during pregnancy and parity: The meta-regression analysis based on data only from three studies, was performed to assess the effect of smoking prevalence, hypertension and parity on the association between second-trimester PVB and PTB. Only three studies provided patient characteristics for the above information. Smoking and hypertension showed a positive correlation but this was not statistically significant. Multiparity demonstrated a significant positive correlation. These findings, contradicting current knowledge, can only be explained because of the small size of studies and participants, as many papers did not provide this information. 

3.3 Sensitivity analysis

Through a sensitivity analysis that considered the weight of each study, three studies conducted by by Ӧzdemirci et al., Karim et al., and Lipitz et al. were removed as they carried the lowest weight in the meta-analysis results Following this exclusion, the pooled odds ratio (OR) was 4.64 (95% CI 2.78–7.75, p<0.001), indicating a significant association between second-trimester PVB and PTB. High heterogeneity persisted (I² = 98%, Cochrane Q = 535.59, p < 0.001), suggesting that these studies did not substantially influence the overall effect estimate or heterogeneity. (Figure. 7)

Notably, a sensitivity analysis restricted to high-quality studies (NOS score ≥ 7) resulted in the exclusion of the same studies, yielding consistent findings and further confirming the stability and robustness of the observed association. (Figure. 4).

Figure 4: Forest plot demonstrating the summary odds ratio for PTB due to 2nd trimester UVB compared to non-related to bleeding PTB, after excluding the studies with the lowest weight.  Studies are listed in chronological order.

3.4 Publication bias

Upon inspection, the funnel plot of studies included in the primary analysis showed evidence of asymmetry suggestive of publication bias, which was confirmed statistically (Egger P=0.008).  (Fig. 5). After 3 studies that appeared to be the cause of the asymmetry [Karim, Ozdemirci, LIpitz] were removed, little to no evidence of underlying bias could be found. Visual inspection of the updated funnel plot demonstrated improved symmetry, with a more balanced distribution of studies around the pooled effect estimate, suggesting a reduced likelihood of small-study effects. This observation was further supported by Egger’s regression test, which no longer indicated statistically significant asymmetry (p = 0.032). (Figure 6) These findings suggest that, after removal of the influential studies, the risk of publication bias is attenuated, thereby increasing confidence in the stability and reliability of the meta-analysis results.

Conversely, when visually assessing the funnel plots of studies encompassed within the majority of subgroup analyses, no significant asymmetry indicative of publication bias was evident.

Figure 7: Funnel plot of the 8 studies included in the meta-analysis

Second trimester unprovoked vaginal bleeding represents a clinically significant, yet incompletely characterised obstetric event. In this systematic review and meta-analysis, we demonstrate a strong and consistent association between second trimester bleeding and preterm birth (PTB), with a pooled odds ratio of approximately 6.72. This magnitude of effect indicates that women experiencing second trimester bleeding are at substantially increased risk of PTB compared to those without bleeding.

The physiology of this association strengthens the clinical relevance of our findings. Vaginal bleeding during pregnancy may reflect underlying decidual instability, subclinical inflammation or early cervical remodelling. It is well recognised clinically that blood is a uterine irritant [11]. Mechanistically, thrombin generation following intrauterine bleeding has been shown to promote uterine contractility and weaken fetal membranes, contributing to both preterm labour and preterm pre-labour rupture of membranes (PPROM) [12, 13].

Unprovoked second-trimester vaginal bleeding is increasingly recognised as a clinical manifestation of underlying placental and decidual dysfunction, different to placenta previa. Specifically, it can represent signs of occult placental abruption, defective trophoblast invasion, or decidual vascular fragility. Occult placental abruption is a recognised mechanism of UVB in late pregnancy. Chronic abruption can present with periodic small amounts of bleeding, but it is often concealed, with a suspected prevalence up to 18%; thus the timing and severity cannot always be determined by visible loss [14]. Defective trophoblast invasion, particularly in early pregnancy, is implicated in early bleeding presentations and later placental complications. Pregnancies with threatened miscarriage presentations have increased risk of placenta previa (OR 1.62), UVB in the third trimester (OR 2.47) and placental abruption (OR 5.6)[15]. And these are the considered high-risk causes of UVB that require extra attention. 

In addition, second trimester bleeding may represent an early manifestation of pathological processes such as uterine or intra-amniotic infection or inflammation, or cervical pathology. Subclinical intra-amniotic inflammation, which may not be readily detectable, has been associated with both vaginal bleeding and adverse pregnancy outcomes. Supporting this, Musilova et al. recently demonstrated that the attenuation of intra-amniotic fluid inflammation following antibiotic therapy is achievable following antibiotics in the second trimester in patients experiencing UVB. This could support a possible mechanistic pathway linking UVB to PTB as patients with inflammation had a mean delivery of 28 vs 37 weeks in the no inflammation group [16]. 

Cervical factors may also play a key role. Conditions such as cervical ectropion, polyps, local infection, and early dilatation can present with bleeding. Studies have shown that local infections can induce pro-inflammatory cytokines IL-6, IL-8, TNF-a and prostaglandins that promote cervical ripening and membrane remodelling[17]. Furthermore, pathogens such as Gardnerella vaginalis have been shown to stimulate cervical metalloproteinase activity, that contributes to collage degradation and disruption of the epithelial integrity, which overall can result in premature cervical changes and ripening [18]. And in some cases as well, cervical remodelling, can occur in conjunction with placental pathologies and intrauterine inflammation. 

Therefore, second trimester UVB is not a benign symptom but rather a clinical marker of underlying biological processes that can predispose patients to preterm birth. Thus, bleeding should be considered as a clinical sign to prompt for targeted investigations to evaluate for infection- inflammation, placental problems, such as small detachments and haematomas, as well as other cervical pathologies. We believe that this systematic approach will improve the management and outcomes of patients with second trimester bleeding to differentiate high and low risk placental, cervical or inflammatory causes and when these transition to higher complication risk. 

Our findings align with previous studies demonstrating an association between vaginal bleeding and adverse pregnancy outcomes, including PTB and PPROM. Prior literature has often grouped bleeding across all trimesters or failed to distinguish between provoked and unprovoked bleeding. By focusing specifically on unprovoked second trimester bleeding, our study provides a more clinically relevant estimate of risk during a gestational window where prognostic uncertainty is high and management strategies are limited. However, our study does not show the expected association with smoking, hypertension and parity. Vascular factors pertaining to smoking and hypertension have been well established due to the risk of abruption in the third trimester, where they appear to almost double and quadruple that risk respectively [19]. 

Notably, our local cohort further supports this relationship, demonstrating that second trimester bleeding was more strongly associated with PTB than with other adverse outcomes. This reinforces the significance of PTB as the primary clinical endpoint in unprovoked second trimester bleeding and supports the rationale for focusing risk stratification efforts and preventative interventions on preterm delivery.

Despite the strength of the observed association, several limitations should be considered. First, heterogeneity across studies was substantial, reflecting variation in study deign, population characteristics, timing and definition of the bleeding presentation. Moreover, a few of the included studies examined patients with bleeding in the 1st or 3 trimester, along with 2nd trimester findings. We ensured that these studies had clear descriptive information regarding the bleeding presentation; however certain gestational outcomes were often combined, which may have introduced some misclassification bias. Lastly, evidence of potential publication bias was observed, with an Egger’s test p=0.039, suggesting that smaller studies with null findings may be underrepresented, potentially inflating the observed effect size.

From a clinical standpoint, these findings have important implications despite the above limitations. Second trimester bleeding should not be considered a benign or self-limiting symptom but rather a important clinical sign of increased risk for preterm birth. It is a clinical challenge to distinguish high-risk bleeding causes vs low-risk, and in cases of low-risk causes with recurrent presentations, be able to adapt management strategies when low-risk causes have resulted in chronic, more severe pathologies and transform to higher risk presentations. Recognition of this association should prompt careful evaluation, enhanced surveillance and patient counselling. However, the absence of standardized management pathways highlights another important gap in current clinical practices.

In conclusion, our study is clearly demonstrating that unprovoked vaginal bleeding during the second trimester is strongly associated with an increased risk of preterm birth and should be recognised as a clinically significant risk marker rather than a benign finding. Our analysis demonstrates a consistent and substantial association across study designs, supporting both the robustness and clinical relevance of this relationship.

These findings emphasise the importance of prompt evaluation and targeted investigation in women presenting with second-trimester bleeding, including assessment for infection, cervical pathology, and placental abnormalities. Improved recognition of this association may facilitate earlier risk stratification, enhanced surveillance, and the development of preventative strategies aimed at reducing preterm birth. Further prospective research is required to better characterise the underlying mechanisms, refine risk prediction, and establish evidence-based management pathways for this high-risk population.