Care of the Postpartum Patient in the Emergency Department: A Systematic Review with Implications for Maternal Mortality

Abstract

Objective

Approximately one-third of maternal deaths occur postpartum. Little is known about the intersection between the postpartum period, emergency department (ED) use, and opportunities to reduce maternal mortality. The primary objectives of this systematic review are to explore the incidence of postpartum ED use, identify postpartum disease states that are evaluated in the ED, and summarize postpartum ED use by race/ethnicity and payor source.

Study Design

We searched PubMed, Embase, Cumulative Index of Nursing and Allied Health Literature (CINAHL), ClinicalTrials.gov, Cochrane CENTRAL, Social Services Abstracts, and Scopus from inception to September 19, 2019. Each identified abstract was screened by two authors; the full-text manuscripts of all studies deemed to be potential candidates were then reviewed by the same two authors and included if they were full-text, peer-reviewed articles in the English language with primary patient data reporting care of a female in the ED in the postpartum period, defined as up to 1 year after the end of pregnancy.

Results

A total of 620 were screened, 354 records were excluded and 266 full-text articles were reviewed. Of the 266 full-text articles, 178 were included in the systematic review; of these, 108 were case reports. Incidence of ED use by postpartum females varied from 4.8 to 12.2% in the general population. Infection was the most common reason for postpartum ED evaluation. Young females of minority race and those with public insurance were more likely than whites and those with private insurance to use the ED.

Conclusion

As many as 12% of postpartum women seek care in the ED. Young minority women of lower socioeconomic status are more likely to use the ED. Since approximately one-third of maternal deaths occur in the postpartum period, successful efforts to reduce maternal mortality must include ED stakeholders. This study is registered with the Systematic Review Registration (identifier: CRD42020151126).

Since the Centers for Disease Control and Prevention (CDC) began measuring pregnancy-related deaths in 1986, the United States has experienced a steady increase, from 7.2 to 17.2 pregnancy-related deaths per 100,000 live births in 2015. Approximately three-fifths of these maternal deaths are likely preventable.^1-3 Cardiovascular complications are responsible for one-third of pregnancy-related deaths, and the CDC postulates that the rise in maternal deaths in the United States may be in part due to an increase in the number of pregnant women with chronic diseases, such as hypertension, diabetes, and heart disease.⁴ The CDC reports that one-third of pregnancy-related deaths occur 1 week to 1 year after delivery.⁵ Leading causes of maternal mortality up to 1 year after delivery include hypertensive disorders of pregnancy, pulmonary embolism, and peripartum cardiomyopathy. Little is known about the location of health care delivery for these postpartum cases that result in death. In the United States, many women rely on Medicaid throughout pregnancy for their health insurance. The CDC reported that in 2016, nearly half of all deliveries were paid for by either Medicaid or self-pay.⁶ Under the Affordable Care Act, all states extend full Medicaid eligibility 60 days after delivery. However, in most states that did not expand Medicaid, coverage is terminated after 60 days unless the woman requalifies for Medicaid services at the income threshold for nonpregnant adults. One study reported that 55% of women had a gap in their insurance coverage in the first 6 months postpartum.⁷ It is possible that lack of insurance coverage may result in decreased routine postpartum care and increased emergency department (ED) visits.

Little is known about the magnitude of postpartum ED use, what factors contribute to postpartum ED use, and what postpartum disease states are commonly managed in the ED. Therefore, we sought to summarize available studies by conducting a systematic review to identify the incidence of postpartum ED use, outline postpartum disease states that commonly present to the ED, and summarize postpartum ED use by race/ethnicity and payor source.

Materials and Methods

Sources

An a priori protocol was drafted and registered with Prospero according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (protocol: CRD 151126).⁸ The University of Alabama at Birmingham’s Institutional Review Board (IRB) declared this review IRB exempt, as it did not qualify as human subjects research. The search strategy was developed in collaboration with a skilled medical librarian (C.H.S). PubMed, Scopus, Embase, Cumulative Index of Nursing and Allied Health Literature (CINAHL), ClinicalTrials.gov, Cochrane CENTRAL, and Social Services Abstracts were searched from inception to September 19, 2019. Searches included all languages and excluded no study types. We used the following search terms: “emergency service,” or “emergency medical services,” or “emergency services,” or “emergency ward,” or “emergency unit,” or “emergency room,” or “emergency department,” and “postpartum period” or “peripartum,” or “postpartum,” or “postnatal period,” or “postnatal,” or “puerperium,” or “studies with female participants.” The full search strategy is listed in ►Supplementary Appendix A (available in the online version).

Study Selection

Full-length, published articles written in English were included if they reported primary data on postpartum females obtaining medical care in the ED. As per the CDC definition, we defined postpartum as the first 365 days after the end of a pregnancy,⁴ which was the completion of any pregnancy, including live births, stillbirths, miscarriages, or terminations at any gestational age. While broader than other definitions of maternal mortality, this definition was selected in accordance with information published by the CDC which states that one-third of pregnancy-related deaths occur between 1 week and 1 year postpartum. All study types from all countries were included. Studies were included if some, but not all, patients met the above criteria; results are reported only for those patients that met criteria.

Studies were excluded for the following reasons: if they documented postpartum patients not seen in the ED, if the encounter originated as a transfer from another hospital and the receiving hospital used the ED as part of patient workflow to facilitate inpatient admission, if patients were transferred to the ED on the day of delivery or during the delivery-associated hospitalization, if birth occurred in the ED, if the patient underwent a home birth and subsequently presented to the ED for immediate postpartum care, if the ED visit was for management of ectopic pregnancy, if the study reported on patients currently pregnant or outside the 1-year postpartum period, if the infant was the patient during a pediatric ED visit, or if the full-text manuscript was published solely in a non-English language. Conflicts were resolved through discussion with the senior author. Data collection was performed independently by K.A.M. into piloted forms.

Studies included in this review were summarized into the following categories: manuscripts reporting the incidence of ED use in postpartum women in the general population, manuscripts reporting the incidence of ED use in postpartum women pertaining to a specific subgroup (e.g., women with intellectual disability), studies reporting postpartum ED use classified by diagnosis, case reports documenting postpartum ED use classified by diagnosis, studies that addressed social determinants of health (defined as “conditions in the environments in which people are born, live, learn, work, play, worship, and age that affect … health”⁹) and postpartum ED use, and studies that primarily reported public health interventions and the effect on postpartum ED use. The Newcastle–Ottawa Quality Assessment Scale was then used to assess the study quality of all included studies that were not case reports.

Results

A total of 772 records were identified via the exhaustive search performed by the medical librarian (C.H.S.). Search results from each database were as follows: (1) PubMed: 308; (2) Embase: 91; (3) CINAHL: 83; (4) ClinicalTrials.gov: 2; (5) Cochrane CENTRAL: 137; (6) Social Services Abstracts: 59; and (7) Scopus: 92. The citations were uploaded to Covidence, an online systematic review citation manager. After removing 152 duplicates, 620 citations were reviewed by two authors (K.A.M. and R.S.). Of these, 354 abstracts were excluded, and two authors (K.A.M and R.S) reviewed 266 full-text articles to determine study eligibility. In all, 178 studies were included in the qualitative synthesis (►Fig. 1). Of the 178 included studies, 108 were case reports or case series including ≤3 patients. Because databases were searched from inception to present, included articles date from 1984 to 2019. The vast majority of the studies were performed in the United States (49%) and Canada (20%), but studies from Europe (13%), Australia (6%), the Middle East (6%), Africa (1%), and Asia (5%) also met criteria (►Supplementary Fig. S1, available in the online version). Case reports were similarly distributed.

Fig. 1.

Flow diagram of included studies.

Incidence of ED use ranged from 4.8 to 8.3% in U.S. studies to as high as 12.2% in Sweden (►Table 1).^10-12 Of the studies reporting incidence of ED use in subpopulations (n=13), four dealt with maternal morbidity from disorders related to the pregnancy, three with intellectual disability, two with intimate partner violence (IPV), two with immigrant or refugee status, one with postpartum depression, one with HIV, and one on patients with neonates requiring neonatal intensive care unit (NICU) stay (►Table 2). Notably, Mitra et al reported that between 91 and 365 days of postpartum, 43% of women with intellectual disability utilized the ED.¹³

Table 1.

Incidence of ED use in postpartum women in the general population

Study	Country (year)	PP period (d)	Incidence of PP ED use	Other findings
Batra et al11	The United States (2017)	90	8.3%; 5.8% used the ED 3+ times	ED use is more common in women <20 years old; with Medicaid, severe maternal morbidity at delivery, or antepartum complications; or who underwent cesarean delivery
Clark et al10	The United States (2010)	42–200	4.8%	58% of conditions were related to pregnancy; half of women were evaluated within 10 days of discharge from the delivery-associated hospitalization
Vikstrom et al12	Sweden (2018)	30	12.2%	ED use is more common in women without routine postpartum care. Women who had a cesarean birth, assisted vaginal birth, or sphincter injury were more likely to use the ED

Abbreviations: ED, emergency department; d, days; PP, postpartum.

Table 2.

Incidence of ED use in postpartum women in specific subgroups

Study	Region (year)	PP period	Specific population addressed	Incidence of PP ED use	Other key findings
Brown et al75	Canada (2017)	6 weeks	Women with ID	With: 16.6% Without: 7.9%	PP women with ID were more likely to use the ED and used the ED more often than PP women without ID for both medical and psychiatric reasons and were more likely to be admitted
Fabiyi et al77	The United States (2019)	1 year	Gestational diabetes, hypertensive disorders of pregnancy	15%	Pregnancy complications were not independently associated with ED use, but lower education and less insurance continuity was associated with increased PP ED use
França and McManus36	The United States (2018)	40 weeks	Postpartum depression	36.7 visits for depression/10,000 deliveries	Hospital visits and ED use for postpartum depression are increasing
Harris et al78	The United States (2015)	6 months	Women with pregnancy complications	With: 27.7% Without: 23.6%	Pregnancy complications, preconception disease, and younger age increase PP ED use
Kothari et al30	The United States (2009)	4 weeks pregnant to 12 weeks postpartum	IPV survivors	13% among perinatal women. Sample too small for statistical comparison	Perinatal women who experience IPV are no less likely to go to an ED than women outside the peripartum period. However, they are more likely to use the ED in the 6 months prior to the assault. Perinatal women assaulted were more likely to be young, unmarried, and Black. Perinatal women may seek treatment for less severe injuries and are unlikely to volunteer information about IPV without specifically being asked
Lewkowitz et al79	The United States (2019)	1 year	Women with severe maternal morbidity	With: 1.6% Without: 0.6%	Study specifically looked at psychiatric ED visits only
Luo et al80	Canada (2018)	6 weeks	Women with hypertensive disorders of pregnancy	9.4%	Women with multiple antihypertensives and abnormal labs had more visits. Postpartum community care programs may decrease visits
Mitra et al13	The United States (2018)	1 year	Women with ID	1 –42 days With: 13.8% Without: 3.7% 43–90 days With: 21.6% Without: 6.4% 91–365 days With: 43.0% Without: 18.2%	Women with ID were more likely to go to the ED >2 times, less likely to have primary care or nonurgent visits, and more likely to visit for mental health
Mitra et al76	The United States (2019)	1 year	Women with ID	1–42 days With: 11.7% Without: 3.5% 43–90 days With: 7.2% Without: 2.5% 91–365 days With: 19.5% Without: 11.0%	Women with ID had twice as many ED visits and 4 times as many repeat visits as women without ID. Women with ID had more admissions
Nannini et al26	The United States (2011)	1 year	Survivors of assault	0.9%	10.4% of these women visited at least two times. PP patients were less likely to visit for assault than first and second trimester women, but more likely than third trimester women
Spicer et al81	Spain (2008)	1 year	Mothers of babies requiring NICU care	9% within the first 3 months, 5% in months 4–12	Visits or calls to health care providers, including ED visits, were most commonly self-reported to occur in the first 3 months postpartum. Physicians were the most heavily utilized group throughout the postpartum period
Sparrow-Downes et al82	Canada (2019)	1 year	Women with HIV	With: 3.3% Without: 1.1%	These visits were for psychiatric reasons only, although women with HIV also had more non-psychiatric visits. Women with HIV 3 times as likely as those without HIV to use the ED
Vigod	Canada (2016)	1 year	Immigrant women	Immigrants: 0.6% Nonimmigrants: 1.3%	These visits were for psychiatric reasons only. Immigrants were less likely than nonimmigrants to use the ED in the postpartum period
Vigod et al83	Canada (2017)	1 year	Refugee women	Refugees: 1% Other immigrants: 0.5%	These visits were for psychiatric reasons only. Refugees were more likely than other immigrants to use the ED

Abbreviations: ED, emergency department; ID, intellectual disability; IPV, intimate partner violence; PP, postpartum.

Studies that dealt primarily with specific diagnoses of postpartum women presenting to the ED (n=147), rather than incidence of ED use, were separated into two groups as follows: (1) clinical research studies (including randomized controlled trials, case-control, cross-sectional, and cohort studies), and (2) case reports, defined as studies reporting outcomes of ≤3 women. Clinical research studies (n=39) were then grouped by the broad diagnosis addressed (►Table 3).

Table 3.

Included studies^a reporting postpartum ED use classified by diagnosis

Pathology (number of studies)	Study	Country (year)	Pathology	Findings
Cardiovascular (2)	Kamel et al14	The United States (2016)	Aortic dissection and rupture	Significantly increased risk of aortic dissection and rupture during pregnancy and postpartum period compared with non-peripartum women
	Wang and Wang15	China (2018)	PPCM	Four percent of postpartum women presenting to the ED had PPCM; among PPCM patients presenting to the ED, 9% had hypertension and 56% had physiologic dyspnea of pregnancy
Hypertensive disorders of pregnancy (3)	Brousseau et al17	The United States (2017)	Postpartum hypertension	Women with postpartum hypertension typically present within 1 week after delivery and are likely to be readmitted. Many knew they had hypertension; surveillance may not be sufficient to curb readmission rates
	Sharara	Qatar (2012)	Eclampsia prevalence in Qatar	Eclampsia occurred in 0.31/1,000 deliveries, and 31.4% of women who developed eclampsia were postpartum. Of the postpartum patients, most had their first seizure within 2–72 hours after delivery
	Yancey et al11	The United States (2011)	Preeclampsia	Patients most commonly presented 3–10 days postpartum, with headache as the most common chief complaint, followed by vision changes, and gastrointestinal upset; 95% had proteinuria, elevated uric acid, or abnormal liver function tests. Almost all had elevated blood pressure in the ED
Infectious disease (9)	Agarwal	India (2019)	Sepsis	The Sepsis in Obstetrics Score had a positive predictive value of 83% for severe sepsis and good correlation with organ failure in peripartum women
	Albright	The United States (2014)	Sepsis	The Sepsis in Obstetrics Score identified peripartum patients who were at high risk for ICU admissions
	Albright	The United States (2017)	Sepsis	The Sepsis in Obstetrics Score is valid to identify risk of sepsis-related ICU admissions, with a negative predictive value of 98.6%
	Amir	Australia (2006)	Mastitis (Staphylococcus aureus)	No association was found between maternal nasal colonization of S. aureus and S. aureus mastitis; however, more infants of mothers with S. aureus mastitis were nasal carriers. Nipple damage positively correlated with mastitis risk
	Amir	Australia (2006)	Mastitis	Negative emotions associated with the physical symptoms of mastitis may lead some women to stop breastfeeding.
	Bosma	The United States (2016)	Breast abscess	(12% of patients were postpartum in this study.) Nearly half of women presenting to the ED who believed they had breast abscesses were diagnosed with the disease; the remainder had mastitis. Appropriate follow-up and/or imaging did not occur in approximately one-fourth of cases
	Hu	Australia (2017)	Breast abscess/mastitis	Of women presenting to the ED for infectious breast disease, 16.8% were referred to the ED by general practitioners, 2.5% by midwives, and 80.7% self-presented; 42% were admitted for only 1 day or were discharged and the rest were admitted. In all, 1.7% of women in the study re-presented to the ED at least once
	Parriot	The United States (2016)	Postpartum infection	ED visits and readmissions were grouped together. Clinical and hospital volumes did not affect postdischarge infection rates in postpartum women. UTIs and surgical complications were most common; 7% of infections after initial hospital discharge involved sepsis or blood stream infection
	Yokoe et al20	The United States (2001)	Postpartum infection	94% of infections in the postpartum period were detected after discharge. In the cohort, 64% of ED visits for women with vaginal delivery were infection related, compared with 80% for women with cesarean delivery
Multiple pathologies (2)	Brousseau et al16	The United States (2018)	Multiple	The median time to ED visit was 7.5 days postpartum. Common chief complaints included: wound complication (17.5%), fever (17.1%), abdominal pain (15.9%), headache/dizziness (12.3%), breast problem (10.7%), and hypertension (10.3%)
	Cano Alonso	Spain (2009)	Multiple	Use of MDCT in EDs to diagnose acute female pelvic disease was assessed. While ultrasound is the modality of choice for acute pelvic pain in the ED, MDCT can be used when ultrasound is equivocal or gynecologic pathology is not at the top of the differential diagnosis. The study included women not in the postpartum period as well as those in the postpartum period
Neurologic (4)	Basurto Ona	Spain (2015)	PDPH	13 RCTs were reviewed. Not all patients were postpartum. Compared with placebo, caffeine was effective at treating PDPH. Gabapentin, hydrocortisone, and theophylline lowered pain severity. Evidence for sumatriptan, adrenocorticotropic hormone, pregabalin, and cosyntropin was lacking
	Behcet et al21	Tukey (2007)	Stroke	Strokes were most likely to occur in the third trimester and postpartum period, particularly the first week postpartum. Most commonly these strokes were caused by preeclampsia and eclampsia. The mortality rate for all patients with stroke was 34.2%, but the rate was higher when hemorrhage was present
	Houtchens	The United States (2018)	MS	MS relapse rates dropped during pregnancy, rose during the first 6 postpartum months, and dropped again at months 6–12 postpartum. Women did not commonly use disease-modifying drugs before or during pregnancy, but use increased during the postpartum period
	Karadas et al22	Turkey (2014)	Cerebral venous sinus thrombosis	Of patients presenting to the ED, headache was the most common symptom, and the most common predisposing condition was the postpartum state
OB/GYN (1)	La Rosa	The United States (2019)	Outcomes after cesarean	Patients were randomized to azithromycin or placebo after cesarean. Women prescribed azithromycin had lower rates of unexpected visits
Psychiatric (11)	Barker	Canada (2016)	Psychiatric disease	For 60.4% of women, their first mental health physician contact since delivery was in the ED. Most of the presentations were for mood or anxiety disorders; 13.6% of presenting women were admitted. Women having first psychiatric physician contact in the ED were more likely to have low income, residential instability, material deprivation and to live in rural areas
	Bryan	The United States (1999)	Postpartum depression	3.7% of women had postpartum depression during the first postpartum year. 0.5% had preexisting depression that continued through the postpartum period. This incidence was lower than prospective studies had previously found
	Glasser	Israel (2018)	Postpartum suicidality	Compared to same-age women, postpartum women were less likely to attempt suicide. Among postpartum deaths, 3.6% were secondary to suicides, 1% to injuries of undetermined intent, 1.5% to accidents, and 4.7% to assaults
	Kramlinger	The United States (1984)	Psychiatric disease and thyroid hormone level	One patient in the population was postpartum. Postpartum patients with known psychotic disease may have high thyroxine secondary to the postpartum state and be clinically euthyroid. Thyroxine normalized in this patient within days
	Lai et al23	Taiwan (2004)	Postpartum psychiatric disease with catatonia	Postpartum patients presenting to the ED most often had depression (53.3%), but one-fifth had mania, one-fifth had schizophrenia, and 1 person had psychosis due to medical condition. Of 4 patients with catatonic features, 3 improved with intramuscular lorazepam
	Lewkowitz et al79	The United States (2019)	Postpartum psychiatric illness in stillbirth vs. live birth	Stillbirth after 23 weeks’ gestation increased readmission risk for psychiatric illness by a factor of 2.5. This most commonly presented as depression or anxiety, but increased risk of substance abuse-related ED visits in this population is also noted
	Polachek et al24	Canada (2016)	Psychiatric disease	Of women admitted to psychiatric services in the postpartum period, 4.9% represented to the ED for psychiatric reasons within 1 week. Women with prior psychiatric admissions before the postpartum period used the ED more often and had less physician follow-up
	Rochon-Terry	Canada (2016)	Psychiatric hospitalizations in postpartum women with schizophrenia	Among women with schizophrenia, 10% had premature deliveries. Their postpartum risk of visiting the ED for psychiatric reasons, while lower in the postpartum period than it was prior to conception, was higher in the first 9 days of postpartum. Self-harm visits were less likely during the peripartum period
	Polachek	Canada (2018)	Postpartum psychiatric admissions	Non-admission ED visits were observed in 11.4% within 1 week postpartum, 15.7% within 1 month, and 33.9% within 1 year. Postpartum women without recent mental illness presented around 4.5 months with affective and adjustment disorders. Those with a history of severe mental illness presented within 9 days for affective and psychotic disorders. Those with non-psychotic mental health history came at 6 months for adjustment and affective disorders. Those with social and substance abuse history came at 7 months and were most likely to be involuntary, pose threats to self, and return to the ED within 1 year
	Polachek	Canada (2017)	Psychiatric admissions postpartum and follow-up ED visits; length of postpartum psych stay	Postpartum women previously admitted for psychiatric disease had higher rates of re-presenting to the ED within 7 days, but lower rates of re-presentation 30 days and 1 year after discharge
	Xu	Australia (2017)	Postpartum depression	Women who had more ED presentations during pregnancy had more PP ED visits for depression.
Trauma/IPV (2)	Nannini et al38	The United States (2008)	Patterns of physical injury during ED visits for assault	Of the 0.9% of women visiting the ED for assault postpartum, the most frequent injuries were superficial contusions, open wounds, and other head and neck injuries. Postpartum women were less likely to have trauma to the abdomen or torso, but visits for assault with fracture were twice as common in the postpartum period than during pregnancy. More than 5% of visits were multisystem injuries and could not be classified to one region of the body
	Nannini et al38	The United States (2008)	Injury etiology and ED visits in pregnant and postpartum women	22 deaths occurred in the cohort, and 7 of these were due to injury. 13.6% of women had an injury-related visit, with 55% occurring postpartum, and 91% were ED visits without admission. A vast majority (89%) were unintentional injury: most were motor vehicle accidents, followed by falls and then overexertion. Assault accounted for 5.4% of injuries. Self-inflicted injuries accounted for <1%. 22% of women visiting for injury visited more than once
Venous thromboembolic events (5)	Goodacre et al33	The United Kingdom (2018)	PE	In pregnant and postpartum women suspected of having PE, D-dimer and clinical decision rules were not cost effective, accurate, or effective at selecting women for diagnostic imaging
	Goodacre et al32	The United Kingdom (2019)	PE	In pregnant and postpartum women suspected of having PE, D-dimer and clinical decision rules were not useful in a clinical setting
	Kabrhel et al34	The United States (2010)	PE	1% of the women included in the study were early postpartum (<4 weeks). Postpartum women often had a positive D-dimer test regardless of whether or not they had a PE
	Kamel et al14	The United States (2014)	Thrombotic events	Until at least 12 weeks postpartum, women were at greater risk of thrombosis, but the risk after 6 weeks, while increased, had a low absolute value
	Robison	The United States (2017)	Heparin protocol in PP venous thromboembolism	The protocol implemented increased enoxaparin use from <1% to >30% without differences in ED visits between groups. However, the study was underpowered to assess a difference in ED visits

Abbreviations: CT, computed tomography; ED, emergency department; ICU, intensive care unit; IM, intramuscular; IPV, intimate partner violence; MDCT, multidetector CT; MS, multiple sclerosis; OB/GYN, obstetrics and gynecology; PDPH, postdural puncture headache; PE, pulmonary embolism; PP, postpartum; PPCM, peripartum cardiomyopathy; PPD, postpartum depression; RCT, randomized controlled trial; UTI, urinary tract infection.

^aExcluding case reports.

Studies reporting postpartum ED use focused on diagnosis were broadly grouped into the following categories: cardiovascular, hypertensive disorders of pregnancy, infection, neurology, obstetrics, psychiatric disease, trauma and IPV, and venous thromboembolic events (►Table 3). The majority of these 147 studies dealt with psychiatric disease (28%) or infectious disease (23%), followed by pulmonary embolism (PE), thrombotic disease (13%), and neurologic diseases (10%).

Cardiovascular studies reported on the increased risk of aortic dissection¹⁴ and peripartum cardiomyopathy¹⁵ in postpartum women. Hypertensive disorders of pregnancy were also a common reason for postpartum women to undergo ED evaluation.¹⁶ In the United States, women with postpartum hypertensive diseases of pregnancy presented very early after delivery, around the first week to 10 days.^17,18 While rare, postpartum preeclampsia and eclampsia can occur 6 weeks after the pregnancy ends or later, and one study found that all of the mothers with late-onset preeclampsia were readmitted within 3 weeks of delivery. These women were more likely to be over 40 years of age, Black or Latinx, have higher BMI and gestational diabetes.¹⁹

Nine studies reporting postpartum ED evaluation of infection met inclusion criteria. Among these studies, one by Yokoe et al concluded that depending on mode of delivery, 64 to 80% of postpartum visits to the ED were related to infection.²⁰

Four studies addressed neurologic diseases, with one finding stroke mortality to be more likely in late pregnancy and the postpartum period, especially the first week postpartum, with a mortality rate of 34%.²¹ Another found that for ED patients with cerebral venous sinus thrombosis, headache was the most common chief complaint and the most common patient demographic was postpartum women.²²

Postpartum depression was a common reason for ED evaluation.²³ Some studies reported that postpartum women may use the ED for psychiatric reasons and return to the ED for ongoing psychiatric care.^24,25 IPV in postpartum ED patients was also commonly reported. Approximately 1% of postpartum women presented specifically for assault or IPV. Of postpartum women visiting the ED for IPV or assault, 10% visited multiple times.²⁶ One study reported a prevalence of any type of IPV at 33.8% at 6 or 12 months after childbirth, with 8.5% of women experiencing physical, and 4% experiencing sexual IPV,²⁷ while other studies reported rates of physical IPV at 16.1 to 17.7% in adolescent mothers²⁸ and 12.7% among low-income mothers.²⁹ It is important to note that in the United States, women who are victims of violence are more likely to be young, unmarried, or Black and are unlikely to volunteer information about IPV unless asked.³⁰

Five studies addressed pulmonary embolism and thrombosis. One study found an increased thrombotic risk lasting at least 12 weeks after delivery.³¹ Other studies found that D-dimers and clinical decision rules to address the likelihood of deep vein thrombosis or pulmonary thromboembolism were not valid or clinically useful in pregnant and postpartum women.^32-34

Seven studies describing social determinants of health and postpartum ED use were included (►Table 4). Five studies were from the United States, one from Canada, and one from Bangladesh and Burkina Faso. Consistently, the U.S. studies concluded that the following factors increase the risk of ED use in the postpartum period: younger age, public insurance coverage (Medicaid), unmarried status, lower socioeconomic status, minority race, lower education level, prenatal morbidity, and delivery complications.^11,35-38 Lack of transportation access was a barrier in Bangladesh and Burkina Faso,³⁹ whereas no difference was found in postpartum ED use based on socioeconomic status in the Canadian population.⁴⁰

Table 4.

Included studies describing social determinants of health and postpartum ED use

Study	Location (year)	PP period (subgroup if applicable)	Risk factors for ED use
Alam et al39	Bangladesh and Burkina Faso (2016)	1 year	Lack of reliable transportation access
Batra et al11	The United States(2017)	90 days	Less than 20 years old, Medicaid, complicated pregnancy/delivery or cesarean, low income
Bryant et al35	The United States (2016)	2 months-2 years	Younger age, unmarried, publicly insured, obese, ED visits earlier in the postpartum period
França and McManus36	The United States (2018)	40 weeks (postpartum depression)	Increased hospital encounters during pregnancy, public insurance
Kurtz Landy et al40	Canada (2008)	Perinatal–4 weeks	No difference in ED use found between socioeconomically advantaged and disadvantaged women
Nannini et al38	The United States (2008)	Perinatal-1 year	Less educated, minority race, public insurance, young age, unmarried
Sheen et al37	The United States (2019)	42 days	Primiparity, public insurance, younger age, prenatal morbidity, weekend delivery, delivery complications

Abbreviations: ED, emergency department; PP, postpartum.

Eleven studies pertaining to public health interventions associated with postpartum ED use were included (►Table 5). One study that evaluated Minnesotan mothers identified that mothers with longer hospital stays, mothers without early follow-up (before 6 weeks), and mothers who delivered via cesarean were more likely to use ED services in the postpartum period.⁴¹ Other studies within this category reported on the length of hospital stay, implementing education measures, reasons for ED referral from clinic, follow-up care, type of delivering practitioner, and patient satisfaction. Many of these studies similarly determined that longer hospital stays correlate with more ED visits.^42,43 Implementation of postpartum clinics or home visits did not tend to have an effect,^44-46 although women at risk for depression who self-excluded from home visits were more likely to visit an ED.⁴⁷

Table 5.

Included studies reporting public health interventions associated with PP ED use

Study	Location	Concept	Findings
Barimani	Sweden	PP support satisfaction and ED visit frequency	Women with delivery complications were less satisfied and used EDs more often
Ferriols Perez	Spain	Referrals to the ED and the adequacy of that referral in pregnant and PP women	In PP patients, 33.8% visited the ED for adequate motives, 41.6% for moderately adequate motives, and 24.7% for inadequate motives
Johnson et al46	The United States	Psychosocial education during pregnancy hospital stay and follow-up care	Despite less time in the hospital, postdischarge maternal ED visits and/or readmits did not increase. Differences in newborn ED visits and/or readmits were not significant
Laliberté et al45	Canada	Breastfeeding clinic and ED visits	No difference was found in the rate of ED visits by mothers who received care at the PP breastfeeding clinic
Lieu et al44	The United States	Establishment of a PP care center and the implications for ED visits for both mothers and infants	No significant differences in maternal (or newborn) ED visits were found after establishing a PP care center
Madlon-Kay and DeFor41	The United States	Implementation of Minnesota length-of-stay laws requiring home visit for shorter postnatal hospital stay, frequency of ED visits and admittance to hospital	Mothers with longer hospital stays, mothers without early follow-up (before 6 weeks), and mothers with cesarean delivery were more likely to use urgent care or ED services PP
Mandl	The United States	Mother’s length of hospital stay and follow-up ED visits	Longer maternal hospital stays correlated with increased maternal ED use
Murray et al47	The United Kingdom	Self-exclusion from health visiting support in women at risk for PPD	Women who self-excluded from health visiting support services were more likely to use EDs instead of attending regular appointments. These women were also of lower education and younger
Paul	The United States	Planned follow-up visits vs. home visits in breastfeeding mothers and PP health care utilization.	Planned follow-up vs. home health had no effect on ED visits or readmissions
Pérez-Martinez	Spain	Midwife vs. obstetrician delivery and complications	No significant difference was found in the number of women attending the ED after discharge between women who were delivered by midwives vs. obstetricians
Watt et al42	Canada	PP hospitalization length of stay and ED visits	With longer stays, women experienced more hospital readmissions, but fewer had contacts with community physicians and EDs

Abbreviation: ED, emergency department; PP, postpartum.

Of the 108 case reports (►Supplementary Table S1, available in the online version), 37% dealt with cardiovascular disease, including aortic dissection,⁴⁸ peripartum cardiomyopathy,^15,49-56 spontaneous coronary artery dissection,^48,57-68 and ST-elevation myocardial infarction.^62,63,69-71 The cardiovascular category also had the highest number of cases that were associated with maternal mortality (n=3).^48,54,72 The other five cases of maternal mortality were due to pheochromocytoma, malignant fibrous histiocytoma, concomitant central venous thrombosis and pulmonary thromboembolism, missed abortion, and status asthmaticus. Patients most commonly presented within the first 3 months of postpartum. Notably, many of the 14 neurology case studies dealt with cerebral venous thrombosis, and one (with concomitant pulmonary thromboembolism) resulted in death due to cerebral tonsillar herniation in the intensive care unit.⁷³

Risk of bias among the 70 clinical research studies was determined by two authors (K.A.M. and R.S) using the Newcastle–Ottawa Quality Assessment Scale (►Table 6).⁷⁴ Six studies (9%) received <5 points, while 64 (91%) studies received ≥5 points, indicating the vast majority of studies had a low risk of bias.

Table 6.

Newcastle–Ottawa quality assessment scale for included studies^a

Study	Country	Year	Article/study type	Subjects (n)	Journal	Newcastle–Ottawa scale risk of bias
Agarwal	India	2019	Cross-sectional	100	Obstet Med	6
Alam et al39	Bangladesh, Burkina Faso	2016	Mixed methods	640	Int J Gynaecol Obstet	5
Albright	The United States	2014	Retrospective cohort	850	Am J Obstet Gynecol	6
Albright	The United States	2017	Retrospective cohort	1,250	Obstet Gynecol	6
Amir	Australia	2006	Case control	199	BMC Fam Pract	7
Amir	Australia	2006	Case control	94	Aust Fam Physician	8
Barimani	Sweden	2014	Cross-sectional	546	Scand J Caring Sci	3
Barker	Canada	2016	Nested case control	8,728	Archives of Women’s Mental Health	6
Basurto Ona	Spain	2015	Systematic review	479	Cochrane Database Syst Rev	8
Batra et al11	The United States	2017	Retrospective cross-sectional	1,071,232	Obstet Gynecol	6
Behcet et al21	Turkey	2007	Retrospective cross-sectional	38	Neurosciences (Riyadh)	6
Bosma	The United States	2016	Retrospective cross-sectional	185	Emergency Radiol	7
Brousseau et al17	The United States	2017	Retrospective cohort	252	Hypertens Pregnancy	6
Brousseau et al16	The United States	2018	Retrospective cohort	5,708	J Womens Health	6
Brown et al75	Canada	2017	Cross-sectional	382,116	J Womens Health	8
Bryan	The United States	1999	Retrospective cross-sectional	403	J Reprod Med	4
Bryant et al35	The United States	2016	Retrospective cohort	14,220	Matern Child Health	6
Cano Alonso	Spain	2009	Retrospective observational	NR	Emerg Radiol	6
Clark et al10	The United States	2010	Retrospective cross-sectional	222,084	Am J Obstet Gynecol	6
Fabiyi et al77	The United States	2019	Retrospective cohort	304	Journal of Women’s Health	6
Ferriols Perez	Spain	2017	Observational retrospective	668	J Obstet Gynecol	6
França and McManus36	The United States	2018	Observational retrospective	3,213,111	PLOS One	6
Glasser	Israel	2018	Observational retrospective	20,259	Isr J Health Policy Res	5
Goodacre et al32	The United Kingdom	2019	Retrospective cross-section	522	BJOG: An International Journal of Obstetrics and Gynaecology	6
Goodacre et al33	The United Kingdom	2018	Case control; cross-sectional	440; 265	Health Technol Assess	8
Harris et al78	The United States	2015	Retrospective cohort	26,074	Journal of Women’s Health	8
Houtchens	The United States	2018	Retrospective administrative claims database	2,158	Neurology	6
Hu	Australia	2017	Retrospective observational	119	Pract Midwife	6
Johnson et al46	The United States	2000	Case control and cohort survey	1,427	Womens Health Issues	7
Kabrhel et al34	The United States	2010	Prospective observational	4,356	Academic Emergency Medicine	6
Kamel et al31	The United States	2014	Retrospective crossover cohort	1,687,930	The New England Journal of Medicine	6
Kamel et al14	The United States	2016	Cohort cross-over	4,933,697	Circulation	6
Karadas et al22	Turkey	2014	Retrospective cross-sectional	51	J Pak Med Assoc	6
Kothari et al30	The United States	2009	Retrospective observational	964	J Women’s Health	8
Kramlinger	The United States	1984	Retrospective cross-sectional	106	Am J Med	4
Kurtz Landy et al40	Canada	2008	Cross-sectional	726	BMC Health Serv Res	6
La Rosa	The United States	2019	Randomized controlled trial	1,019	Amer L Perinatol	7
Lai et al23	Taiwan	2004	Prospective observational	15	Psychiatry and Clinical Neurosciences	6
Lalibert et al45	Canada	2016	Randomized controlled trial	572	PLOS One	7
Lewkowitz	The United States	2019	Retrospective cohort	1,203,050	Am J Obstet Gynecol	5
Lewkowitz et al79	The United States	2019	Retrospective cohort	1,193,968	Obstet Gynecol	8
Lieu et al44	The United States	1998	Double prospective cohort	800	Pediatrics	5
Luo et al80	Canada	2018	Retrospective cohort	319	J Maternal Fetal Neonatal Med	7
Madlon-Kay and DeFor41	The United States	2005	Retrospective cross-sectional	22,944	J Am Board Fam Pract	6
Mandl	The United States	1999	Prospective cohort	1,200	Arch Pediatr Adolesc Med	4
Mitra et al76	The United States	2019	Retrospective cohort	596,464	J Epidemiol Community Health	8
Mitra et al13	The United States	2018	Retrospective cohort	779,513	Maternal and Child Health Journal	8
Murray et al47	The United Kingdom	2003	Nested cohort case control	158	J Public Health Med	6
Nannini	The United States	2011	Retrospective cohort	273,345	MIDIRS Midwifery Digest	6
Nannini	The United States	2008	Retrospective cohort	1,468	Nursing Research	6
Nannini et al38	The United States	2008	Retrospective cohort	100,051	J Midwifery Womens Health	6
Parriott	The United States	2016	Retrospective cohort	217,954	Am J Infect Control	6
Paul	The United States	2012	Randomized controlled trial	1,154	Arch Pediatr Adolesc Med	7
Pérez-Martinez	Spain	2019	Quasi experimental retrospective cohort	2,621	Midwifery	5
Polachek et al24	Canada	2016	Retrospective population-based cohort	1,071	General Hospital Psychiatry	8
Rochon-Terry	Canada	2016	Population-based retrospective cohort	1,433	J Clin Psychiatry	6
Robison	The United States	2017	Retrospective cohort	9,766	Obstet Gynecol	5
Sharara	Qatar	2012	Retrospective case review	224,809	Qatar Medical Journal	6
Shlomi Polachek	Canada	2018	Population-based retrospective cohort	1,702	Psychiatry Research	6
Shlomi Polachek et al25	Canada	2017	Population-based retrospective cohort	1,702	Archives of Women’s Mental Health	7
Sparrow-Downes et al82	Canada	2019	Retrospective population-based cohort	861,365	AIDS Care	8
Spicer et al81	Canada	2008	Observational longitudinal	152	Advances in Neonatal Care	3
Vigod	Canada	2016	Population-based retrospective cohort	450,622	The Canadian Journal of Psychiatry	8
Vigod et al83	Canada	2017	Population-based retrospective cohort	123,231	Archives of Women’s Mental Health	6
Vikström et al12	Sweden	2018	Retrospective cross-sectional	28,963	Journal of Clinical Nursing	6
Wang and Wang15	China	2018	Retrospective observational	262	Medicine (Baltimore)	6
Watt et al42	Canada	2005	Retrospective cohort	6250	BMC Pregnancy and Childbirth	4
Xu	Australia	2017	Retrospective cohort	154,328	BMC Emerg Med	6
Yancey et al11	The United States	2011	Retrospective case series	22	J Emerg Med	5
Yokoe et al20	The United States	2001	Retrospective cross-sectional	2,746	Emerging Infectious Diseases	6

Abbreviation: NR, not reported.

^aExcluding case reports with ≤3 patients.

Discussion

This systematic review demonstrates that as many as 12.2% of women worldwide use the ED in the postpartum period. This is of critical importance to the maternal mortality discussion as Clark and colleagues found that 58% of the conditions for which women seek ED care in the postpartum period were pregnancy related.¹⁰ More vulnerable populations, such as those with intellectual disability,^13,75,76 prenatal complications and maternal morbidity,^77-80 mothers with sick neonates,⁸¹ and those with HIV,⁸² as well as women of color and younger, less educated, refugee,⁸³ unmarried, and/or low-income women, were more likely to use the ED in the postpartum period.^11,35-38 It is important to highlight that Black women are at least three times more likely to die during the postpartum period than non-Hispanic White counterparts.⁴ Racial disparities in maternal mortality continue after adjusting for education, location (within the United States), and age.⁸⁴ More frequent ED use by a group of women who are at higher risk for maternal mortality may offer an opportunity for an ED-based intervention that could improve outcomes and decrease morbidity and mortality.

To be clear, ED use is often dictated by what sources of care are available to various segments of the population. Increased use by people with lower socioeconomic status or government insurance may indicate that they have no other source of care. However, it may also signal that they have more acute care needs. As eloquently stated by a scholar in this field, “adjustment for sociodemographic and reproductive factors has not explained the racial gap in pregnancy-related mortality in most studies.”⁸⁵

The strength of this review comes largely from the exhaustive search by an experienced senior medical research librarian (C.H.S), as well as independent review of hundreds of articles by two individuals and the assignment of Newcastle–Ottawa risk by two independent reviewers. By including all study types and an international scope, as well as the entire postpartum period, this review covers a wide population and assesses multiple aspects of why postpartum women present to the ED and the implications with respect to maternal mortality.

Limitations

This review does have limitations. First, we only included articles published in the English language and thus likely missed important information presented in other languages, skewing the results toward implications in English-speaking countries. Second, the reviewers were not able to distinguish in many cases that the gestational age at which the pregnancy ended and acknowledge that postpartum complications significantly vary between first trimester losses and term deliveries. Notably, we excluded home births, even though an estimated 1 out of 62 births occurred outside the hospital in 2017.⁸⁶ Because we defined the postpartum period as the day after birth through 1 year after birth, deliveries that took place in the ED were excluded which leaves out intrapartum and immediate postpartum pathology that may lead to maternal mortality in the ED. Despite reviewing 178 articles that reported care of the postpartum patient in the ED, very little information exists on maternal morbidity and mortality in the ED. Only two studies of postpartum women presenting to the ED addressed mortality as a primary outcome: one U.S. study dealt with IPV³⁸ and the other addressed strokes in Turkey.²¹ This systematic review revealed opportunities for future study. Specifically, despite this exhaustive review, the global incidence of postpartum ED use is still uncertain, and incidences of conditions cared for in the ED cannot be ascertained without a common denominator. Future studies using the Nationwide Emergency Department Sample (NEDS) may provide contemporary data on rates of U.S. postpartum ED utilization.

Conclusion

Maternal mortality is an international health crisis. Postpartum women frequently use the ED and, in the ED, are frequently diagnosed with illnesses that are leading causes of maternal mortality. This provides an opportunity for EDs to incorporate postpartum status as a “vital sign” for all females of child-bearing age seeking care in the ED to identify this high-risk group. Knowledge that the patient is postpartum and in the “fourth trimester” allows the provider to prioritize the differential diagnosis accordingly. In addition, allowing ED providers to extend the period in which they classify a patient as postpartum to a full year after delivery may better serve women who experience diseases late into the postpartum period, such as peripartum cardiomyopathy, the leading cause of postpartum death.³ Further research is needed to fully understand the incidence of postpartum ED use and inform the development of potential interventions to reduce unnecessary postpartum ED use. This research may identify opportunities for maternal mortality review committees and perinatal quality collaboration efforts to reduce postpartum morbidity and mortality through ED-based interventions.

Key Points.

• Up to 12% of postpartum women seek care in the ED.

• One-third of maternal deaths occur postpartum.

• Maternal mortality reduction efforts should include ED stakeholders.