Vaginal bleeding in early pregnancy: patterns, predictors, and association with miscarriage Public Deposited

Downloadable Content

Download PDF
Last Modified
  • March 22, 2019
  • Hasan, Reem
    • Affiliation: Gillings School of Global Public Health, Department of Epidemiology
  • First-trimester vaginal bleeding is common in pregnancy; however, few data have described the distribution, characteristics, and predictors of early bleeding episodes. The relationship between bleeding and miscarriage is not well understood. Data from Right From the Start (RFTS), a prospective, community-based pregnancy cohort were used for all analyses. We used descriptive statistics to characterize first-trimester bleeding episodes and logistic regression to identify predictors of bleeding. Bleeding characteristics (such as heaviness, duration, timing, and color) predictive of miscarriage were identified using classification and regression trees. The relationship between bleeding and miscarriage was modeled using discrete-time hazard models. We compared retrospectively collected bleeding reports from the first trimester interview with prospective data from a daily diary to obtain sensitivity, specificity, and kappa statistics. Log-linear models were used to identify predictors of agreement. In all analyses, we removed bleeding episodes that ended within four days of miscarriage. Approximately one-fourth of participants reported bleeding in early pregnancy, mostly spotting or light bleeding episodes. Most episodes lasted fewer than 3 days, and most occurred between gestational weeks 5 and 8. Heavy episodes, reported by about 2% of women, were more likely to be painful, of longer duration, and red in color. Predictors of bleeding were age (particularly between 28 and 34), increasing education, nulliparity, and menstrual cycle length less than 27 days or greater than 33 days. Maternal conditions (diabetes, fibroids), prior pregnancy outcomes (miscarriage, induced abortion), reproductive tract infections, smoking, and alcohol intake were also predictive of bleeding. Women who reported heavy bleeding had nearly three times the risk of miscarriage compared to women without bleeding (OR 3.0, 95% CI 1.9, 4.6). Spotting and light bleeding episodes were not related to miscarriage. Bleeding episodes and characteristics were reported with high levels of agreement in the diary and interview. No predictors of agreement were identified in this analysis. To summarize, we found that vaginal bleeding was a common first-trimester symptom. However, the majority of episodes were spotting or light bleeding episodes, which did not confer an increased risk for miscarriage. Although few women reported heavy bleeding, heavy bleeding was more strongly related to pregnancy loss.
Date of publication
Resource type
Rights statement
  • In Copyright
  • Baird, Donna
  • Open access

This work has no parents.