Etiology of stillbirth in a tertiary care center: a retrospective cohort study assessing ultrasound, laboratory, and pathology investigations

Abstract Background Canadian stillbirth data are limited, and a significant proportion of pregnancies resulting in stillbirth have no attributable cause. The objective of this study was to characterize stillbirth case investigations and management at a tertiary care hospital in Ontario, Canada. Methods This was a retrospective chart review study of all cases of singleton stillbirth at The Ottawa Hospital between 1 January 2012 and 31 December 2017. Terminations and multiples stillbirths were excluded. Chart reviews were conducted to extract maternal sociodemographic, obstetrical, and fetal characteristics, including results from antenatal ultrasounds, autopsy, placenta pathology, and laboratory investigations. Results A total of 155 eligible cases of stillbirth were identified, resulting in a 6-year stillbirth rate of 4.2 per 1000 total births. The median maternal age was 31.0 years (IQR: 29.0, 35.0) and the median gestational age at delivery was 28 weeks (IQR: 24, 35). A total of 9 (5.8%) pregnant individuals had a history of previous stillbirth. Of the 155 stillbirths, 35% underwent the full suite of post-loss laboratory, placental, and fetal autopsy investigations. 63.2% of cases had post-loss laboratory investigations completed. 76% and 71% of cases had fetal autopsy and placenta pathology evaluations completed, respectively. Antenatal characteristics associated with stillbirth included fetal anomalies/genetic markers (27.1%), umbilical cord and placental anomalies (24.5%), fetal growth abnormalities (27.7%), cervical/uterine abnormalities (11.6%), and amniotic fluid abnormalities (25.1%). The most common autopsy findings included evidence of infection (22.7%), fetal anomalies (12.6%), and fetal hypoxia (10%). The most common placental pathology findings included features of placental insufficiency (21.8%), retroplacental abnormalities (16.3%), and umbilical cord accident/infarct (15.4%). Conclusions Our findings demonstrate that as many as two-thirds of singleton stillbirth cases at our center did not receive the post-perinatal loss investigations recommended by clinical practice guidelines. More thorough collection of post-stillbirth data at all levels (institutional, provincial, national) is warranted to improve our understanding of stillbirth epidemiology, etiology, and management in Canada.


Introduction
Stillbirth is a devastating obstetrical outcome defined as the death of a fetus at �20 gestational weeks or where the fetal weight is �500 g [1].In 2021, the stillbirth rate in Canada was approximately 8.6 per 1000 live births [2].The etiology of stillbirth varies widely from case to case, and may be multifactorial with maternal, fetal, and placental origins.Maternal risk factors include increased maternal body mass index (BMI), nulliparity, advanced maternal age, smoking in pregnancy, previous stillbirth, preexisting diabetes or hypertension, and hypertensive disorders of pregnancy [3,4].Population-based studies demonstrate that leading fetal risk factors for stillbirth include fetal growth restriction, fetal infection or sepsis, and congenital or genetic anomalies [5,6].Placental dysfunction and disease are also attributable causes of stillbirths [7].A systematic review of placental pathologies associated with stillbirth found that factors contributing to fetal demise included placental abruption, infarction, chorioamnionitis, and villous immaturity [8].Diffuse placental insufficiency can precipitate stillbirth by mechanisms of inadequate nutrient and oxygen delivery, as well as producing small-for-gestational age or under-developed fetuses [7].
The physiological mechanisms underlying stillbirth are not well understood, and many cases are attributed to unknown causes or unexplained conditions, with no clear maternal, fetal, or placental contributions [9].Despite this, it remains essential that care providers investigate possible contributing factors to stillbirth cases [10].The Society of Obstetricians and Gynaecologists of Canada (SOGC) provides guidelines to support the systematic investigation and management of pregnancies affected by stillbirth, provide support to the family, and guide future pregnancies [1].Routine institutional practices may include fetal autopsy, placental pathology examination, and genetic karyotyping; however, parental consent and limitations in local hospital resources may preclude some or all these investigations.Furthermore, documentation of stillbirths in Canadian is not completed by hospitals, but is reliant on parents to report fetal deaths with the Office of Vital Statistics, creating additional mechanisms for data inaccuracy and under-reporting [11].As a result, there remains a paucity of information on the causes and risk factors of stillbirth in Canada [12].A single-center retrospective chart review study from Regina, Alberta reported a 5-year stillbirth rate of 0.87% [13].In this sample of 180 stillbirth deliveries and 183 fetuses, over 40% of stillbirths were attributed to unexplained etiology.Although placental histopathology was completed for 100% of cases, the fetal autopsy rate was 42.5%, and karyotyping was done in 20.2% of cases [13].The most common risk factors for stillbirth in this sample were maternal obesity, late prenatal care, and smoking in pregnancy.Another single-center study from Montreal, Quebec described stillbirth case management and findings from over 20 years of data [14].Among 289 stillbirth cases, 70 had to be excluded because they did not undergo an autopsy examination, and one was excluded because placental pathology examination was not performed.The most common cause of stillbirth was unknown etiology (26.7%), followed by placental factors (19.8%) and placental abruption (12.9%) [14].
Comprehensive data on maternal sociodemographic information, medical histories, and obstetrical outcomes are critical to identifying population-specific risk factors for stillbirth, and for informing care management.The objective of this study was to characterize stillbirth cases, workup, and management at a tertiary care hospital in Ontario, Canada.

Study design, setting, and population
This was a retrospective chart review study of all cases of singleton stillbirth at The Ottawa Hospital (TOH) in Ottawa, Canada, between 1 January 2012 and 31 December 2017.Stillbirths were defined as the death of a fetus >20 weeks gestation or weighing more than 500 g [1].Terminations of pregnancy (TOP) and twin/multiple stillbirths were excluded.
Potentially eligible cases were identified via the Better Outcomes Registry & Network (BORN) Ontario, a prescribed registry, which captures maternal demographic, pregnancy and labor and delivery characteristics and newborn outcome details for all births in Ontario for administrative and statistical purposes [15].Hospitals submitting data to BORN Ontario, including TOH, have access to site-specific data within the BORN Information System to support quality-improvement initiatives and research.Following identification in BORN Ontario, case eligibility was confirmed by chart review, and pre-specified demographic and clinical data were extracted.
Research ethics approval was granted by the Ottawa Health Science Network Research Ethics Board (OHSN REB #20170526-01H).

Data collection
Charts were reviewed for common maternal and antenatal risk factors for stillbirth.Maternal demographic and obstetrical characteristics included maternal age at delivery, pre-pregnancy BMI, gestational age at delivery, parity, and numbers of prior term pregnancies, pre-term pregnancies, abortuses, and stillbirths.We also collected data on maternal substance use during pregnancy, tobacco smoking in pregnancy, medication use in pregnancy, and the type of antenatal care provider attending to the index pregnancy.Preexisting maternal conditions including diabetes, hypertension, lupus or autoimmune disease, asthma, mental health conditions, herpes simplex virus, venous thromboembolic disease, and obesity were also recorded.
Labor characteristics and complications captured during data extraction included indication for induction, type of labor analgesia, group B streptococcus (GBS) status, delivery type, fetal presentation, the occurrence or absence of intrapartum fever, induced cervical ripening, abnormal fetal heart rate (AFHR) as defined by SOGC guideline, equivalent to NICHD category II or III [16,17], cord prolapse, uterine rupture, shoulder dystocia, and nuchal cord.
Ultrasound findings included fetal anomalies, soft markers for aneuploidy; Doppler evaluation, umbilical cord, or placental anomalies; as well as growth, cervical/uterine, and amniotic fluid abnormalities.
Fetal autopsy observations included whether there was evidence of fetal hypoxia, infection, meconium, oligohydramnios, hydrops fetalis, trauma, and gross fetal anomalies.Extracted placenta pathology details included observations of placental structural abnormalities, umbilical cord accidents or infarcts, retroplacental abnormalities, growth abnormalities and features of placental insufficiency.Placentas from our institution are investigated using the 2016 Amsterdam Consensus Statement [18].
TOH provides institutional guidelines outlining appropriate investigative protocols for perinatal loss.These guidelines recommend assessment of maternal coagulation factors, autoimmune factors, and bacterial/viral infections.Laboratory studies are carried out using maternal serum samples and vaginal or rectal swabs where appropriate.Findings of these investigations were also extracted.Coagulation factors included: international normalized ratio (INR), prothrombin time (PTT), D-dimer, fibrinogen, protein C, protein S, lupus anticoagulant, factor V Leiden, and prothrombin gene.Autoimmune factors included anticardiolipin and anti-nuclear antibody (ANA) levels.Factors arising from infection work-up consisted of: toxoplasma, rubella, cytomegalovirus (CMV), herpes, parvovirus, mumps, varicella-zoster (VZV), HIV, hepatitis B surface antigen (HBsAg), and hepatitis C (HCV).Additional laboratory data included homocysteine, Kleihauer-Betke, glycated hemoglobin levels (HbA1c) and GBS.

Analysis
Descriptive statistics were used to report the demographic and clinical characteristics of our sample.The distribution of the variables described above was summarized using frequencies (n) and percentages (%) for categorical data, while median and interquartile ranges (IQRs) were used to describe continuous variables.The overall stillbirth rate at TOH during the study period was calculated by dividing the number of stillbirths by the total number of births per year, multiplied by 1000.The proportion of individuals for whom the recommended clinical workup, including fetal autopsy, placental pathology, and additional institutional post-perinatal loss investigations were conducted was calculated.For laboratory data subject to standardized reference ranges, values were categorized as "normal" or "abnormal", with careful consideration of gestational age and other clinical correlations as necessary [19].

Sample characteristics
From 1 January 2012 to 31 December 2017, there were 37,053 births at TOH, of which 243 were stillbirths.Of these, 88 cases (36.2%) of twin/multiple stillbirths or TOPs were excluded, resulting in a final sample of 155 singleton stillbirths.The singleton stillbirth rate at TOH over the 6-year study period was 4.2 per 1000 births.

Institution post-perinatal loss laboratory investigations
Post-perinatal laboratory loss investigations were conducted for 98 individuals in our sample (63.2%).Maternal INR and PTT tests results came back as atypical for 2% and 4% of our sample, respectively.Pertinent d-dimer values were reported for 29.6% of patients, whereas atypical results for fibrinogen were evident in only 3% of patients.Factor V Leiden levels, lupus anticoagulant and prothrombin gene levels were relevant in 2%, 1%, and 1% of patients, respectively.Infection serology is tabulated in Table 3.One individual in the cohort was GBS positive, but some may not have been tested yet due to gestational age at screening being 35-36 weeks.Positive Kleihauer test was reported in 4% of patients.Abnormal HbA1C levels were detected in 12% of patients.ANA was detected in 18% of patients, while anti-cardiolipin IgG and IgM were reported in 3% and 0%, respectively.Relevant homocysteine levels were not observed in our cohort.

Discussion
In this retrospective chart review study, we identified a total of 155 stillbirths that occurred at our tertiary care center between 2012 and 2017.The six-year singleton stillbirth rate was 4.2 per 1000 births.In total, 35% of this sample underwent the complete set of post-perinatal loss investigations recommended by national clinical practice guidelines, including a review of antenatal ultrasounds, maternal and fetal laboratory testing, and fetal autopsy and placenta pathologic examinations.Of the 155 cases, over 16% of cases exhibited no clear evidence as to the cause of stillbirth.The prevalence of singleton at our center is relatively low compared to published provincial (9.9/ 1000 total births), national (8.6/1000 total births), international (5.74/1000 live births), and global (13.9/1000 live births) estimates of stillbirth [2,11,20,21].TOH is a tertiary care center offering high-risk obstetrical care services as well as a good volume of routine obstetrical care patients and this may balance it out to explain the relatively low stillbirth rate at our center compared to provincial and national measures.The exclusion of stillbirths occurring in multi-fetal pregnancies may have contributed to the lower prevalence estimate as well.It was not surprising to find evidence of fetal and placental stigmata of stillbirth in our sample.Ascending infection, diverse congenital abnormalities, and peripartum hypoxia are well documented in the literature as potential causes and contributors to stillbirth [22,23], as is placental insufficiency caused by maternal or fetal vascular malperfusion, retroplacental insults compromising the maternal-fetal interface and cord accidents negatively affecting the fetal circulation [8,24].The association between ultrasound findings including fetal, Doppler, growth, uterine and amniotic fluid abnormalities and stillbirth is also supported by current literature.A large retrospective cohort study of singleton pregnancies resulting in stillbirth reported a 15-fold increase in the risk of stillbirth when isolated major congenital abnormalities were detected compared to stillbirths without fetal evidence of congenital anomalies (OR 15.2, 95%CI 11.0-20.9)[25].Findings from another study of singleton pregnancies complicated by fetal growth restriction demonstrated that fetal Doppler abnormalities including high resistance umbilical artery flow and abnormal ductus venosus indices before 34 weeks gestation were more frequent in stillborn compared to liveborn infants [26].The authors of this study also reported that over 90% of stillbirths occurred within one week of abnormal biophysical profile scanning, suggesting an association between fetal wellbeing monitoring by ultrasound and stillbirth [26].Finally, a single-center analysis from Texas found that stillbirth rates after 34 weeks gestation were significantly higher for pregnancies complicated by oligohydramnios compared to pregnancies with normal amniotic fluid indices (1.4% vs. 0.3%; p < .03)[27].
The completion rate of post-perinatal loss investigations recommended by practice guidelines in our sample is quite high (35%) compared to other reported experiences.A six-year, multicenter Australian study found that no stillbirth cases between 2013 and 2018 were investigated in compliance with Perinatal Mortality Clinical Practice Guidelines from the Perinatal Society of Australia and New Zealand (PSANZ) [22].Although all causes of stillbirth should be thoroughly investigated, limitations in data availability, practitioner preference and expertise, as well as emotional and psychological burden to the family typically hinder this [28].The most useful investigations for elucidating the cause of stillbirth include fetal autopsy and pathological examination of the placenta [28,29].Fetal autopsy rates (77%) at our institution were greater than those reported in studies from Australia (52%) and the United States (21%) [22,30].Oliver et al. found that the fetal autopsy rates were higher when losses occurred after 28 weeks gestation, suggesting that less invasive investigations may be insufficient to ascertain the possible causes of late pregnancy loss [30].Pathological examination of the placenta was completed in just 71% in our cohort compared to approximately 90% reported by other tertiary centers [22,31].The relatively low rate of placental investigations in our sample was unexpected considering the proximity of our center to specialized perinatal pathologists at a neighboring pediatric hospital.However, routine pathologic examinations may not occur in scenarios where an obvious cause has been identified, the parents deny consent for the examination or the family wishes to retain the placenta for cultural reasons.Decisions regarding these invasive methods of investigation are highly distressing to families.There remains no guidelines for interventions to support parents through decisions regarding invasive postmortem investigative methods after stillbirth [32].Noninvasive or minimally invasive autopsy techniques should be considered as an alternative to traditional autopsy in tertiary care centers.For example, a combination of postmortem magnetic resonance imaging (MRI) and endoscopic "keyhole" autopsy allows for noninvasive anatomic analysis and tissue biopsy as required, without the negative cosmetic effects of traditional autopsy [33].Indeed, evidence from the United Kingdom, where minimally invasive fetal autopsy techniques are more commonly used suggests that for cases of stillbirth, noninvasive biopsy and imaging techniques are favorable investigative options [34].
The overall utility of post-perinatal loss laboratory investigations in determining the cause of stillbirth remains unclear [29,35].At our institution, post-perinatal loss laboratory investigations for blood-based maternal markers were completed in 63% of stillbirth cases.Several hematology markers demonstrated relevant results, including INR, PTT, d-dimer, fibrinogen, factor V lupus anticoagulant, anti-cardiolipin IgG, and prothrombin gene.The clinical utility of these markers may be limited in pregnancy, however, as abnormal values may simply reflect normal physiologic changes associated with the hypercoagulable state of pregnancy [36,37].Post-perinatal loss, abnormal values may indicate undiagnosed obstetric conditions linked to poor obstetrical or fetal outcomes such as preeclampsia [38], but again findings should be interpreted with caution.The Stillbirth Collaborative Research Network (SCRN) has previously linked Factor V Leiden mutations with an increased odds of stillbirth (OR 87.44, 95%CI 7.88-970.92)[39].The SCRN did not find associations between adverse fetal outcome and prothrombin gene mutation in their cohort, which is supported by other large US studies [39,40].Antiphospholipid syndrome antibodies have been linked with stillbirth; however, as in our study, the SCRN reported a low prevalence of these antibodies in their cohort [41].Bacterial and viral infections are well documented in cases of stillbirth [42], and indeed, our sample exhibited atypical results from infection workup including those for rubella, CMV, HSV, and parvovirus IgM antibodies, as well as HBsAg and GBS.There was also evidence of immunological perturbations, based on Kleihauer-Betke and HbA1c test results, which may indicate fetal-maternal hemorrhage (FMH) or maternal diabetes, respectively.Although the SCRN reported FMH and gestational diabetes in 43.6% and 91% of stillbirth cases, respectively [6], these findings were not directly attributed to stillbirths in this cohort.Despite 63% of our cohort having completed bloodbased laboratory investigations, the results were not clinically informative to ascertain etiology of stillbirth.Given the potential distress, these investigations may cause for affected families, the value of these investigations for determining likely causes and recurrence risk of stillbirth warrants reconsideration.
Limitations of this study include the use of a singlecenter study design.However, by conducting detailed chart reviews and extractions, we were able to generate comprehensive profiles of all eligible stillbirth cases in this study, which would not have been possible with a population-based assessment.A limitation of our study was that fetal genetic evaluations and karyotyping were completed but were not available through our site database as they are processed at a different institution, and as a result we cannot comment on the relative contribution of genetic factors to stillbirths in our sample.
Our findings demonstrate that approximately twothirds of stillbirth cases at our center did not receive the complete set of post-perinatal loss investigations recommended by clinical practice guidelines.Despite the demonstrated utility of placenta pathology and fetal autopsy examinations for identifying causal factors, such investigations were not completed for 25% and 30% of cases, respectively.Where detailed data on stillbirths in Canada continue to rely on chart review studies to inform on the risk factors and current case management practices, hospital/departmental policies on post-perinatal loss investigations should be clarified.Ultimately, more thorough and systematic collection of post-stillbirth data at all levels (institutional, provincial, national) is warranted to improve our understanding of stillbirth epidemiology, etiology, and management in Canada.