To explore the risk of adverse maternal and perinatal outcomes in women with antepartum bleeding of unknown origin (ABUO).
To explore the risk of adverse maternal and perinatal outcomes in women with antepartum bleeding of unknown origin (ABUO).
Cohort study based on data extracted from the Aberdeen Maternity and Neonatal Databank. Exposure was antepartum haemorrhage occurring after the first trimester not attributable to placenta praevia or placental abruption.
Aberdeen Maternity Hospital, Aberdeen, Scotland, UK.
All primigravidae delivering between 1976 and 2010.
Data were analysed using univariate and multivariate statistical methods.
Pre-eclampsia, induced labour, mode of delivery, preterm delivery, postpartum haemorrhage, admission to neonatal unit, perinatal death.
Between 1976 and 2010, there were 7517 women with ABUO and 68 423 women without ABUO in the cohort. Women with ABUO were more likely to be non-smokers, belong to a lower social class, and have a higher body mass index. ABUO was a significant risk factor for induced labour (adjusted odds ratio, aOR, 1.23; 95% CI 1.16–1.31), preterm delivery at <37 weeks of gestation (aOR 2.30; 95% CI 2.11–2.50), and postpartum haemorrhage (aOR 1.15; 95% CI 1.06–1.25). There was no significant association detected with pre-eclampsia (aOR 0.93; 95% CI 0.83–1.05). Whereas there was an increased risk of low birthweight (aOR 0.90; 95% CI 0.79–1.03) and stillbirth (aOR 0.92; 95% CI 0.66–1.30) with ABUO on univariate analysis, once adjusted for confounding factors this risk was non-significant.
Pregnancies complicated by ABUO are at a greater risk of preterm delivery and induced labour. There was no increase seen in perinatal mortality after adjusting for preterm birth.
Several publications have previously established adverse pregnancy outcomes associated with placenta praevia or placental abruption. Where antepartum haemorrhage (APH) is not attributable to either of these serious conditions, the literature is sparse and sometimes conflicting regarding the associated risk of adverse perinatal outcomes and management of these cases. Yet, the condition of antepartum bleeding of unknown origin (ABUO) is by far the most common form of APH, contributing to 50% of all APH, and may complicate upto 6% of pregnancies. Diagnosis is by the exclusion of placental causes using ultrasound examination.[3, 4] In the absence of clear guidelines, ABUO is often managed as cases of mild acute placental separation.
In their review, Magann et al. found consistent associations between ABUO and preterm delivery, stillbirth, and congenital anomalies. There was considerable heterogeneity in the definitions of exposure and outcomes in the ten studies included in the review. Many of the studies did not report the magnitude of association for crude or adjusted analyses. An updated search in Medline identified five additional articles assessing other adverse perinatal outcomes associated with ABUO, but considerable uncertainty remained with regard to associations with pre-eclampsia and eclampsia, induction of labour, operative delivery, preterm prelabour rupture of membranes (PPROM), and malpresentation and malposition.[7-11] Magann called for larger prospective studies with the ability to correct for confounding factors such as maternal age, parity, gestational age, and comorbidities.
Given the conflicting evidence surrounding the risks associated with ABUO and the uncertainity regarding its management, we aimed to assess the risks of adverse pregnancy outcomes associated with ABUO using our population-based database of routinely collected obstetric data. Although essentially retrospective in nature, we felt the analysis of this large database would add to the available evidence, and perhaps provide directions for future research.
The Aberdeen Maternity and Neonatal Databank was established in 1950, and since then has been continuously recording pregnancy-related events occurring in women resident in Aberdeen city and district, in the north-east of Scotland. Complications related to pregnancy and delivery, as well as related comorbidities, are coded using criteria taken from the ninth and tenth revisions of the International Classification of Diseases (ICD-9 and -10). Details of the background and structure of the databank are available from the website www.abdn.ac.uk/amnd. Data were extracted from this database using predetermined criteria.
The study population was limited to primigravid women in order to minimise the confounding effect of parity. We included all women who delivered singleton babies after 24 completed weeks of gestation between 1976 and 2010 in Aberdeen Maternity Hospital.
A cohort study design was used where the exposed cohort comprised women with ABUO, and the unexposed cohort comprised women without APH. The outcomes assessed were classified into two groups: maternal and perinatal. Maternal outcomes included gestational hypertension, pre-eclampsia and eclampsia, PPROM, induction of labour, mode of delivery, and postpartum haemorrhage (PPH). Perinatal outcomes included preterm birth (delivery before 37 weeks of gestation), low birthweight (of less than 2500 g), stillbirth, neonatal death, malposition or presentation, and admission to neonatal intensive care.
Univariate analysis was conducted to identify one-to-one relationships between each maternal or perinatal outcome and ABUO without adjusting for confounding factors. Here, the crude odds ratio was identified as a measure of association. Binary logistic regression was used to identify the relationship between the explanatory variable ABUO and each dichotomous outcome variable. Similarly, multinomial logistic regression was performed when the outcome variable had more than two categories.
Multivariate logistic regression analysis was carried out to identify the relationship between each maternal or perinatal outcome and ABUO by adjusting for confounding variables. The sociodemographic variables that showed significant association at the 5% level of significance with ABUO on univariate analysis were included in the model to obtain adjusted odds ratios with 95% confidence intervals.
Permission to carry out the study was granted by the steering committee of the Aberdeen Maternity and Neonatal Databank. As only anonymised routinely collected data were used in the analysis, formal ethical approval was not considered necessary by the North of Scotland Research Ethics Service.
At first, the data set with a total sample size of 76 886 primigravid women was obtained from the AMND. Data cleaning began by filtering out 650 cases of placenta praevia and abruptio placentae. Women with missing or out-of-range values in the outcome variables were excluded from the analysis. Statistical analysis was carried out on 75 940 primigravid women. Figure 1 shows the flow diagram of study population selection.
As a complication after the first trimester of pregnancy, ABUO occurred in 9.89% of pregnancies (n = 7517). Table 1 presents the sociodemographic characteristics of women with and without ABUO. The mean age (SD) was 25.70 (5.66) years for women with ABUO and 25.78 (5.29) years for women without APH. This difference was not statistically significant. Of women with ABUO, 39% belonged to a manual social class (registrar general's occupation-based social class), as did 38.1% of women with no APH. Women with ABUO were more likely to be non-smokers or ex-smokers, compared with women without APH. A significantly higher proportion of women were single (20.7 versus 17.6%; P < 0.001) among women with ABUO. There was a statistically significant difference between the median (interquartile range, IQR) body mass index (BMI) of women in the ABUO group versus women who did not experience APH: 23.44 (21.38–26.45) versus 23.75 (21.61–26.67). Social class, smoking, marital status, and BMI were found to be significantly associated with ABUO.
|Characteristics||ABUO (n = 7517)||No APH (n = 68 423)||P value|
|Age at delivery||25.70 (5.66)a||25.78 (5.29)a||0.02|
|Never smoked||3791||50.4||32 673||47.8||<0.001|
|Current smoker||2061||27.4||18 539||27.1|
|Not single||5960||79.3||56 383||82.4||<0.001|
|Body mass index||23.44 (21.38–26.45)b||23.75 (21.61–26.67)b||<0.001|
Table 2 represents the univariate and multivariate comparison of maternal outcomes between the exposed and unexposed cohorts. The proportions of women with gestational hypertension (19.8 versus 21.9%), pre-eclampsia and eclampsia (5.9 versus 6.9%), and other hypertension (0.7 versus 0.9%) were significantly less for women with ABUO, compared with those without APH. However, multivariate analysis showed that the association with gestational hypertension (adjusted OR, aOR 0.98; 95% CI 0.91–1.05), pre-eclampsia and eclampsia (adj. OR 0.93; 95% CI 0.83, 1.05), and other hypertension (adj. OR 0.92; 95% CI 0.69, 1.25) was no longer statistically significant after adjusting for confounding factors. The proportion of women with PPROM was equal in both groups (0.8%).
|Characteristics||ABUO (n = 7517)||No APH (n = 68 423)||Unadjusted||P value||Adjustedb||P value|
|No||%||No||%||OR (95% CI)||OR (95% CI)|
|Pre-eclampsia and eclampsia|
|No||5529||73.6||48 056||70.2||1.00||1.00 0|
|Gestational hypertension||1492||19.8||15 011||21.9||0.86 (0.81–0.92)||<0.001||0.98 (0.91–1.05)||0.53|
|Pre-eclampsia and eclampsia||441||5.9||4746||6.9||0.81 (0.73–0.89)||<0.001||0.93 (0.83–1.05)||0.26|
|Other||55||0.7||610||0.9||0.78 (0.59–1.03)||0.08||0.92 (0.69–1.25)||0.61|
|Yes||62||0.8||524||0.8||1.08 (0.83–1.40)||1.43 (1.00–2.03)|
|Yes||2872||38.2||23 960||35.0||1.15 (1.09–1.20)||1.23 (1.16–1.31) c|
|Type of delivery|
|Instrumental||2060||27.4||19 547||28.6||0.97 (0.92–1.02)||0.27||1.03 (0.97–1.10)e||0.31|
|Assisted breech and breech extraction||86||1.1||781||1.1||1.01 (0.81–1.27)||0.91||0.85 (0.64–1.13)e||0.27|
|Caesarean delivery||1515||20.2||12 644||18.5||1.10 (1.03–1.17)||0.003||1.01 (0.94–1.08)e||0.78|
|Non-caesarean delivery||6002||79.8||55 779||81.5||1.00||1.00|
|Elective||273||3.6||2449||3.6||1.04 (0.91–1.18)||0.59||0.77 (0.66–0.90)f||0.001|
|Emergency||1242||16.5||10 195||14.9||1.13 (1.06–1.21)||<0.001||1.06 (0.98–1.14)f||0.13|
|Yes||922||12.3||6974||10.2||1.24 (1.15–1.33)||1.15 (1.06–1.25) d|
The proportion of women having induced labour in the ABUO group was 38.2%, and that among women with no APH was 35% (P < 0.001). After adjusting for confounding factors, the risk of induced labour remained statistically significant (aOR 1.23; 95% CI 1.16–1.31).
Out of all of the women with ABUO, 12.3% had PPH, which was slightly higher than among women who did not have APH (10.2%). The odds of PPH were higher (1.15; 95% CI 1.06–1.25) in women with ABUO compared with women without APH, after adjusting for confounding factors.
Caesarean delivery was slightly higher (20.2%) in the ABUO group than in women without APH (18.5%; P < 0.001); however, after adjusting for confounding factors, this difference was no longer significant.
Table 3 shows the results regarding the relationship between perinatal outcomes and ABUO. The proportion of preterm delivery was 14.1% among women with non-specific APH and 6.4% among women without APH. Even after adjusting for confounding factors, an elevated risk of preterm delivery remained associated with ABUO (aOR 2.30; 95% CI 2.11–2.50).
|Characteristics||ABUO (n = 7517)||No APH (n = 68 423)||Unadjusted||P value||Adjustedb||P value|
|No||%||No||%||OR (95% CI)||OR (95% CI)|
|Preterm delivery (<37 weeks of gestation)|
|Yes||1060||14.1||4397||6.4||2.39 (2.22–2.57)||2.30 (2.11–2.50) c|
|Yes||99||1.3||418||0.6||2.17 (1.74–2.71)||0.92 (0.66–1.30)d|
|Yes||49||0.7||280||0.4||1.60 (1.18–2.16)||0.92 (0.61–1.38)e|
|Low birthweight (<2500 g)|
|Yes||935||12.4||4537||6.6||2.00 (1.86–2.16)||0.90 (0.79–1.03)f|
|Baby presentation at delivery|
|Abnormal||1333||17.7||12 369||18.1||0.98 (0.92–1.04)||0.96 (0.90–1.03)g|
|Admission to Neonatal Intensive Care Unit|
|Yes||1054||14.0||6585||9.6||1.53 (1.43–1.64)||1.03 (0.94–1.12)h|
The proportion of stillbirths was 1.3% among women with ABUO and 0.6% among women without APH. There was 2.17 (95% CI 1.74–2.71) times increased odds of stillbirth among women with ABUO compared with women without APH on univariate analysis. However, there was no independent association seen between ABUO and stillbirth, after adjusting for other confounding factors, notably preterm birth in the multivariate analysis (aOR 0.92; 95% CI 0.66–1.30). Similarly, the proportion of neonatal deaths was 0.7% in the ABUO group and 0.4% in the group without APH. There was a 1.60 (95% CI 1.18–2.16) times increased odds of neonatal death in women with ABUO compared with women without APH (P = 0.04) after unadjusted analysis, but this association was no longer statistically significant after adjusting for confounding factors in the multivariate model (aOR 0.92; 95% CI 0.61–1.38).
The proportion of low birthweight (LBW) among women with ABUO (12.4%) was almost twice that for women without APH (6.6%). LBW was significantly associated with ABUO, with a two-fold elevated risk in the unadjusted analysis (OR 2.00, 95% CI 1.86–2.16); however, the adjusted odds ratio (aOR 0.90; 95% CI 0.79–1.03) for LBW was statistically non-significant.
The proportion of admissions to the neonatal intensive care unit (NICU) was significantly higher among the group of women with ABUO than in women with no APH, but no significant association between ABUO and admission to NICU was found after multivariate analysis (aOR 1.03; 95% CI 0.94–1.12).
The results of this study indicate that non-specific bleeding after the first trimester complicates 9.89% of pregnancies. ABUO was linked to some adverse outcomes. ABUO was significantly associated with induced labour and PPH, whereas no association with hypertensive disorders of pregnancy, PPROM, and elective or emergency caesarean was seen. ABUO also appears to be strongly associated with preterm delivery; however, other adverse perinatal outcomes, such as stillbirth, neonatal death, low birthweight, and admission to the NICU appear to have no association with ABUO, after adjusting for confounding factors.
To our knowledge, this is the largest epidemiological study exploring the association between ABUO and maternal and perinatal complications. The AMND offers a unique opportunity to study the effects of ABUO using routinely collected population-based data at minimal cost. Contemporaneous data entry, using consistent coding criteria and various quality-assurance checks make the data valid and reliable, and also minimises recall bias. The study was carried out on data that date back to 1976. Year of delivery was adjusted in the analysis to control for any changes in the definition of exposure and other outcome variables, as well as in clinical practice. The effect of parity on maternal and perinatal outcomes was prevented by restricting the study to primigravid women because the risk of adverse pregnancy outcomes varies according to parity. The ability to account for a large number of covariates is also a strength of this analysis.
As the study consisted of a very large sample it is possible that the differences detected between exposed and non-exposed groups may result from chance alone. Women experiencing light bleeding during the antepartum period may not seek hospital consultation, which may lead to those cases being missed, thereby resulting in selection bias. There is no differentiation made in the AMND between single or recurrent episodes of ABUO, the amount of bleeding, or for the gestation at which the ABUO occurred, as long as it occurred after the first trimester. We also acknowledge that some minor cases of placental separation may have been misclassified as ABUO. As the study is based on secondary data, the presence of missing values and the retrospective nature of data collection may have biased the results to some extent. As the study was restricted to primigravid women, the results may not be generalisable to multigravid women or women with restricted healthcare access.
According to the results of this study the prevalence of ABUO was 9.89%, which is higher than has been reported elsewhere: 1–2% and 5%.[6-8, 11] This may be because the AMND collects data from Aberdeen Maternity Hospital, which is a tertiary maternity care hospital where complicated pregnancies are over-represented. Moreover, the current study was restricted to nulliparous women, where the prevalence of APH is known to be higher.
Although the absolute risk of pre-eclampsia and eclampsia was found to be lower in women with ABUO, the association was not statistically significant on multivariate analysis. Pre-eclampsia is a condition that typically develops in late pregnancy. It is possible that the increased rates of preterm delivery associated with ABUO may have contributed to the slight reduction in risk of pre-eclampsia seen in this study. The increased rates of preterm delivery may also explain the non-significant association of ABUO with stillbirth found on multivariate analysis. Incidentally, Harlev et al. and McCormack et al.[9, 11] also found a significantly lower risk of pre-eclampsia and eclampsia in women with ABUO. On the other hand, Koifman et al. reported an increased risk of pre-eclampsia in these women.
Induction of labour appeared to be significantly increased in both univariate and multivariate analyses. This was supported by the results of McCormack et al., who showed an elevated risk of induction at term among women with ABUO. Although it is difficult to be certain that the reason for induction of labour was mainly ABUO, adjusting for confounding factors should have addressed this uncertainty. Episodes of ABUO, especially if significant or recurrent, may result in anxiety in both the woman and the healthcare professionals, and may explain the increased induction rate. However, a significantly lower risk of induction of labour was found by Koifman et al.
We also found that women with ABUO had an increased risk of PPH in both univariate and multivariate analyses, showing a highly significant association between them. This was supported by Harlev et al., who showed a significant increase in the risk of PPH among women with idiopathic vaginal bleeding on crude analysis. None of the published studies assessed the adjusted risk of PPH.
A strong and consistent association has been reported between ABUO and preterm delivery.[6-11] Similarly, the results of studies conducted by Berkowitz et al., Signore et al., and Sipila et al.[14-16] were also in agreement with the results of this study, which showed that women with ABUO were more than twice as likely to have preterm delivery. Applying Hill's criteria of causality, we can therefore say that at least in the case of preterm delivery, the association is likely to be causal. We cannot definitely infer causality from the other associations detected in this study, as the literature is conflicting and the likelihood of type-1 error is possible.
Whereas the strong association with preterm delivery may be at least partially related to the lowered threshold for intervention in cases of ABUO, the increased risk of preterm delivery remained even after adjusting for induction of labour, indicating a susceptibility to spontaneous preterm delivery in women with ABUO. ABUO might therefore be a risk marker for preterm delivery, indicating a need for vigilance on the part of clinicians, and for the preparation for appropriate care in advance, such as the administration of corticosteroids.17 The risk of stillbirth and hypertensive disorders did not differ in women with and without ABUO, but it is interesting to speculate how fetal wellbeing would be affected had labour not been induced early or occurred spontaneously, as there is an inverse relationship between preterm delivery and late stillbirths. It is impossible to make any recommendations regarding early delivery of ABUO pregnancies based on retrospective data, but this analysis perhaps paves the way for a randomised controlled trial comparing expectant management with appropriate monitoring versus early delivery for cases of ABUO.
Pregnancies complicated by ABUO are at greater risk of preterm delivery; the risks of stillbirth and neonatal mortality appear not to be increased with ABUO once preterm births are taken into account.
The authors confirm that they have no conflicts of interest.
S.B. conducted the literature search and main analyses, and wrote the first draft of the article. E.A.R. supervised the statistical analysis. A.S. provided the clinical perspective and commented on the article. S.B. was responsible for overall supervision, the design of the study, and for facilitating data extraction. All authors contributed to writing the final draft.
Approval was obtained from the steering committee of the Aberdeen Maternity and Neonatal Databank (ref. no. sb/amnd/1–12; 7 March 2012).
This project was funded by the University of Aberdeen as part of the MSc in Public Health and Health Services Research.
The authors wish to thank Ms Linda Murdoch and Ms Katie Wilde for extracting the data from AMND.
Second- and early third-trimester vaginal bleeding is abnormal, and can be associated with placenta praevia or placental abruption. In the absence of these complications, vaginal bleeding at this stage of gestation can be called antepartum bleeding of unknown origin (ABUO), and can complicate 1–10% of pregnancies. The general impression of ABUO is that most of it is likely to result from small placental separation, and there must be some overlap between ABUO and abruptio placentae: both diagnoses of exclusion.
This article is the largest epidemiological study exploring the association between ABUO and maternal and perinatal complications. When compared with most other literature this study controlled for confounding factors of risks, and therefore reaches more reliable conclusions. The biggest shortcomings are the lack of specific data on volume or recurrence of bleeding, gestational ages at diagnosis of ABUO (e.g. was ABUO at term included?), a significant number of ‘missing’ data, and data on management. As clinicians, we would like to frame this study in the context of direct patient management.
In all cases, there is no substitute for a complete history and physical examination, as this will help exclude other causes of antepartum bleeding. Primary efforts should be aimed at confirming that the source of bleeding is vaginal, and subsequently ruling out any vaginal or cervical causes for bleeding (lesions, polyps, malignancy, etc.). The physical examination should include a speculum examination to exclude the above diagnoses and quantify the volume of bleeding, as women that experience heavy bleeding are at highest risk for poor outcomes (Towers et al. Am J Obstet Gynecol 2008;198:684.e1–684.e5). Ultrasound should exclude other causes of antepartum bleeding such as placenta praevia, vasa praevia, and potentially large abruptions, although ultrasound is not particularly sensitive to detect this complication (Glantz et al. J Ultrasound Med 2002;21:837–840). Transvaginal cervical length measurement can help determine the magnitude of risk of preterm birth (PTB) (Ramaeker et al. Am J Obstet Gynecol 2012;206:224.e1–224.e4), and external fetal and tocomonitoring may be helpful to assess fetal status and risk of PTB.
The risk of PTB is increased more than twice (aOR 2.3), and this is consistent with most studies. The risk of induction increased significantly by 23%, but this increase is not consistent with other studies. Postpartum haemorrhage also increased significantly, but also modestly, by 15%, which is consistent with a prior study. From this study and a summary of the literature, there seems to be no association between ABUO and risk of pre-eclampsia, low birthweight, stillbirth, or neonatal mortality.
The primary objective in women with ABUO should be to prevent or mitigate the adverse outcomes of PTB. Some women with ABUO at between 24 and 33 weeks of gestation may warrant antenatal steroid treatment, especially if short cervix is diagnosed in conjunction with bleeding. When maternal and fetal status is stable, outpatient treatment appears to be a viable option, but whether antenatal surveillance would improve fetal outcomes remains unclear (Ajayi et al. BJOG 1992;99:122–125). Induction of labour at term specifically for ABUO does not appear to prevent perinatal complications, and obstetrical indications should guide labour management of these patients. Future prospective studies to elicit which women are at greatest risk for PTB will be an important factor in determining management.
V.B. is a scientific editor for BJOG.
C Schoen & V Berghella
Division of Maternal Fetal Medicine, Department of Obstetrics and Gynecology, Jefferson University Hospital PA, USA
A patient, G3 P1+1, presented with an episode of antepartum haemorrhage at 26 week of her otherwise uncomplicated pregnancy. On examination, her genital tract was normal. Ultrasound showed a normal intrauterine pregnancy with no placenta praevia or abruption. She asked, “should I worry?”.
|Participants||All primigravidae delivering between 1976 and 2010 in Aberdeen Maternity Hospital (UK)|
|Events||Antepartum haemorrhage (APH) of unknown origin|
|Comparison||No antepartum haemorrhage|
|Outcomes||Preeclampsia, induced labour, mode of delivery, preterm delivery, postpartum haemorrhage, admission to neonatal unit, perinatal death|
|Study design||Cohort study (retrospective data extraction using a prospectively collected database)|
Women's Health Research Unit, Queen Mary,
University of London, London, UK
Join us at #BlueJC: You can collaborate with us to host a #BlueJC session. Find out more on our Journal Club page at www.BJOG.org. Follow @BJOGTweets to stay updated on scheduled #BlueJC sessions. For background of #BlueJC, please refer to BJOG 2013;120:657–660. http://bit.ly/10VaiRZ
Complementary slide set suitable for your local journal club is also available with this paper. See Supporting Information for details.