Extracorporeal Life Support in Pregnancy: A Systematic Review

Abstract

Background

The use of extracorporeal life support (ECLS) has expanded to include unique populations such as peripartum women. This systematic review aims to (1) quantify the number of cases and indications for ECLS in women during the peripartum period reported in the literature and (2) report maternal and fetal complications and outcomes associated with peripartum ECLS.

Methods and Results

This review was registered in PROSPERO (CRD42018108142). MEDLINE, Embase, and CINAHL were searched for case reports, case series, and studies reporting cases of ECLS during the peripartum period that reported one or more of the following outcomes: maternal survival, maternal complications, fetal survival, and/or fetal complications. Qualitative assessment of 221 publications evaluated the number of cases, clinical details, and maternal and fetal outcomes of ECLS during the peripartum period. There were 358 women included and 68 reported fetal outcomes in cases where the mother was pregnant at the time of cannulation. The aggregate maternal survival at 30 days was 270 (75.4%) and at 1 year was 266 (74.3%); fetal survival was 44 (64.7%). The most common indications for ECLS overall in pregnancy included acute respiratory distress syndrome 177 (49.4%), cardiac failure 67 (18.7%), and cardiac arrest 57 (15.9%). The most common maternal complications included mild to moderate bleeding 66 (18.4%), severe bleeding requiring surgical intervention 48 (13.4%), and intracranial neurologic morbidity 19 (5.3%). The most commonly reported fetal complications included preterm delivery 33 (48.5%) and neonatal intensive care unit admission 19 (27.9%).

Conclusions

Reported rates of survival in ECLS in pregnant and postpartum women are high and major complications relatively low.

Nonstandard Abbreviations and Acronyms

ARDS

acute respiratory distress syndrome

ECLS

extracorporeal life support

ECMO

extracorporeal membrane oxygenation

Clinical Perspective

What Is New?

• With the increasing medical complexity of women of childbearing age, the potential for catastrophic complications of pregnancy may call for advanced therapies including extracorporeal life support.

What Are the Clinical Implications?

• Venovenous and venoarterial extracorporeal support may be implemented for a variety of severe cardiopulmonary conditions in pregnant patients with reasonable success and safety for mother and fetus.

Cardiovascular disease is currently the leading cause of pregnancy‐related mortality in the United States, contributing to more than a quarter (26%) of maternal deaths and as well as severe maternal morbidity.1, 2 Pregnant women are becoming more medically complex with chronic health conditions that may predispose them to cardiopulmonary complications during pregnancy.3, 4, 5, 6 Given the rise in conditions for which extracorporeal life support (ECLS) including extracorporeal membrane oxygenation (ECMO) may be indicated, providers must understand the uses and limitations of advanced medical therapies such as ECLS. The relative rarity of catastrophic cardiopulmonary disease in the peripartum period requiring ECLS makes it difficult to study. Existing systematic reviews are narrow in scope (for example, evaluating only certain causative conditions such as acute respiratory distress syndrome [ARDS]) and small in size (the largest studies report <70 patients).7, 8, 9, 10 Combining past studies for all conditions in a systematic review will enhance the understanding of how to identify patients who will benefit from ECLS, develop appropriate use, and provide data to counsel patients and their families. The aim of this systematic review is to perform a comprehensive search of ECLS in the pregnant and postpartum periods, to define the reported indications as well as maternal and fetal survival, and to identify associated complications.

Methods

This systematic review was registered in PROSPERO (CRD42018108142) and the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) guidelines were reviewed and maintained when performing this systematic review.11 This study was exempt from institutional review board approval as a systematic review of published literature. The authors declare that all supporting data are available within the article and its online supplementary files.

Literature Search

A comprehensive literature search was performed on September 11, 2018 for this systematic review using OVID MEDLINE (1946‐), Elsevier Embase (1947‐), and EBSCOhost CINAHL (1937‐) databases from time of inception to capture studies regarding ECLS in the pregnant and postpartum periods with no limit placed on language or date of publication, but studies that describe the use of ECMO in children were excluded. A selection of sentinel articles was used to generate search terms and test retrieval in all of the databases. The complete searches can be found in Data S1.

Study Selection

After the search was completed, 2 authors (E.E.N., A.N.C.) reviewed each abstract independently for consideration of full text review. The same 2 authors independently reviewed the full text articles for inclusion in the systematic review. A search using the Web of Science from February 28 through March 1, 2019 identified the articles in the bibliographies as well as citations of the selected articles and were also screened for inclusion. An additional study was found through preexisting subject knowledge and included. Once final articles were selected, data were extracted by one author (E.E.N.) and validated by another (A.N.C.). Google Translate was used for non‐English texts and information was collected from adequately translated articles for those that met inclusion criteria. Any discrepancies were resolved by discussion. Figure 1 depicts a flowchart of study selection.

Figure 1. Flowsheet for study inclusion/exclusion.

Flow diagram of search results and study selection.

Case reports, case series, meeting abstracts, correspondences, and any other type of study that reported one or more of the following were included: maternal survival, fetal survival, maternal complications, and/or fetal complications with the use of ECLS during pregnancy or within the postpartum period (the first 42 days after delivery). Studies in which the patients did not clearly undergo ECLS, underwent cardiopulmonary bypass for cardiac surgery only, or underwent immediate cardiac mechanical support (eg, left ventricular assist device, biventricular assist device) were excluded. Cases in which initiation of extracorporeal support was unclear were reviewed by 2 authors (E.E.N., M.E.B.) for inclusion and those where initiation was more than 42 days postpartum were excluded.

The following criteria were used for exclusion of studies: patients who were not pregnant or more than 42 days postpartum, studies not reporting either maternal or fetal outcomes, nonhuman studies, non‐English studies without adequate translation, editorials, narrative review articles, and systematic review articles. Articles that were excluded after full text review were recorded with reasons for exclusion. If more than one reason for exclusion was identified, articles were excluded based on the following hierarchical order: nonhuman experiment or unrelated to ECLS; unable to obtain full text or full English translation; review article or systematic review; commentary/editorials without original results; ECLS not performed; patient not pregnant; no ECLS outcomes reported; unable to differentiate outcomes from nonpregnant patients; patient >42 days postpartum; duplicate studies or overlapping cohorts.

Interpretation of Data

Data were collected for indication; maternal demographics; gestational age; timing, duration, and type of ECLS; and maternal and fetal morbidity and mortality. Maternal information collected included age, gravidity/parity, disease process, ECLS duration, complications, and mortality. Obstetrical information was collected for gestational age and delivery type. Fetal information was collected for gestational age at delivery, preterm delivery, neonatal intensive care unit admission, complications, and mortality. ECLS information was collected for support type (venoarterial, venovenous, or both [venoarterial+venovenous]), duration, and indication. For cohort studies, the mean age, gravidity and parity, gestational age, and duration on ECLS were recorded. One author collected the data (E.E.N.) and the second author (A.N.C.) reviewed each study and confirmed the data. Discrepancies in the data were resolved by discussion, and if still unresolved, a third author (M.E.B.) reviewed for resolution. Authors of published articles with incomplete information were contacted by e‐mail and/or telephone for further information.

Gravidity and parity was defined as described in reports on hospital admission. Vaginal deliveries included spontaneous vaginal deliveries and operative vaginal deliveries. Deliveries prior to 37 weeks were considered preterm deliveries. Any delivery prior to 32 weeks was considered to have been admitted to the neonatal intensive care unit whether or not it was explicitly stated, in accordance with the American Academy of Pediatrics guidelines on levels of neonatal care.12

Fetal outcomes were reported for patients who underwent ECLS while pregnant; fetal outcomes from patients who underwent ECLS immediately postpartum, postpartum, or unknown timing relative to delivery were not included. Ectopic pregnancies were not counted toward fetal mortality. Maternal cases <24 weeks with a fetal loss were reported as a spontaneous abortion and were not included in the preterm delivery count. Spontaneous abortions that occurred prior to ECLS were not included in fetal mortality. Therapeutic abortions were included in fetal mortality if the termination occurred on ELCS for maternal indications or if abortion was performed because of catastrophic fetal injury incurred due to ECLS. Fetal survival was reported within the outcome only if it was explicitly stated in the study.

Patients were classified as being antepartum, immediately postpartum, or postpartum. Antepartum patients were those cannulated and initiated on ECLS prior to delivery. Immediately postpartum cases were identified as those patients who were cannulated and initiated on ECLS within 24 hours of delivery. Postpartum cases included patients who were cannulated and initiated on ECLS >24 hours after delivery but within 42 days of parturition. Cases that were identified as postpartum but did not have a clear time frame of <24 hours of delivery were included in the postpartum category. Indications for ECLS were noted based on case reports; some patients had multiple indications for ECLS. Patients with underlying ischemic, structural, or valvular heart disease were defined as having heart disease as well as cardiac failure as the indication for ECLS.

Mechanical support was identified as venoarterial, venovenous, or both as described in the study. Extracorporeal carbon dioxide removal, extracorporeal arteriovenous carbon dioxide removal, interventional lung assist membrane, and percutaneous extracorporeal lung assist supports were included with venovenous therapies. Cases in which the patient underwent cardiopulmonary bypass and had been previously initiated on ECLS or underwent post‐cardiopulmonary bypass ECLS were included. Duration of ECLS was reported for all patients regardless of survival. Survival was reported within 30 days of discontinuation of ECLS therapy as the aim of this study is to evaluate the rescue survivability of ECLS; additional deaths within 1 year were also reported separately. Complications included mild to moderate bleeding (qualitatively described as mild to moderate and/or the need for transfusion without invasive intervention), severe bleeding (requiring surgical or endoscopic interventions), vascular complications (limb ischemia, pseudoaneurysm, wound infection), intracranial morbidity (hypoxic brain injury, intracranial hemorrhage, hemiplegia, cerebral infarct), deep vein thrombosis, and other neurologic morbidity (peripheral neuropathy, cognitive dysfunction, need for rehabilitation after discharge, myoclonic epilepsy, amnestic disorder). Some patients had multiple complications.

Quality Assessment

Because of the low level of evidence and high risk of bias, case reports and case series were not assessed for the level of quality. Given the limited information presented, meeting abstracts and correspondence were also not assessed for level of quality.

Results

Overall, 2116 studies were identified and reviewed and 221 studies met inclusion criteria. Table S1 contains the complete list of articles included and description of the studies and patients. These papers were published between 1974 and 2019. Figure 2 depicts the increasing rate of reporting of cases. Table 1 reports the patient characteristics. There were a total of 358 cases of ECLS during the peripartum period reported. The maternal survival at 30 days was 270 (75.4%) patients; maternal survival at 1 year was 266 (74.3%). Of the 358 cases, there were 210 (58.7%) cases that described the fetal outcome. There were 81 cases where the mother was on ECLS during pregnancy; of these cases, 68 (84.0%) had fetal outcomes reported. Fetal survival reported for patients who underwent ECLS during the antepartum period was 44 (64.7%) patients.

Figure 2. Case reports by year of publication.

Results of the case reports by year of publication. MSR indicates the maternal survival rate for each time period described.

Table 1.

Patient Characteristics

	Total Patients, n (%)
Maternal count	358 (100)
Antepartum cases with fetal outcome	68 (84.0)
Average maternal age, y (SD)	29.5 (6.1)
Median gravity/parity (range)	2/1 (G1P0–G9P8)
Median gestational age (SD)	24 wk (6.2)
Interquartile range	6.14 wk
Stage of pregnancy, n (%)
Antepartum	81 (22.6)
Immediate postpartum	69 (19.3)
Postpartum	126 (35.2)
Uncategorized	82 (22.9)
Delivery type, n (%)
Vaginal	67 (18.7)
Cesarean	158 (44.1)
Dilation and extraction	5 (1.4)
Not reported	128 (35.8)
Deliveries on extracorporeal membrane oxygenation, n (%)	35 (9.8)
Extracorporeal life support indications, n (%)a
Acute respiratory distress syndrome	177 (49.4)
Cardiac failure	67 (18.7)
Cardiac arrest	57 (15.9)
Peripartum cardiomyopathy	45 (12.6)
Pulmonary arterial hypertension	28 (7.8)
Amniotic fluid embolism	27 (7.5)
Pulmonary embolism	17 (4.7)
Heart disease	14 (3.9)
Septic shock	11 (3.1)
Asthma	7 (2.0)
Malignancy	7 (2.0)
Takotsubo cardiomyopathy	6 (1.7)
Pheochromocytoma	5 (1.4)
Spontaneous coronary artery dissection	3 (0.8)
Aspiration pneumonitis	2 (0.6)
Trauma	2 (0.6)
Distributive shock	1 (0.3)
Propofol infusion syndrome	1 (0.3)
Hemorrhagic shock	1 (0.3)
Diffuse alveolar hemorrhage	1 (0.3)
Pulmonary alveolar proteinosis	1 (0.3)
Interstitial lung disease	1 (0.3)
Sickle cell crisis	1 (0.3)
Aortic dissection	1 (0.3)
Transfusion related acute lung injury	1 (0.3)
Cerebral venous thrombosis	1 (0.3)
Pulmonary hemorrhage	1 (0.3)
Cystic fibrosis	1 (0.3)
Dengue fever	1 (0.3)
Extracorporeal membrane oxygenation cannulation mode, n (%)
Venoarterial	145 (40.5)
Venovenous	96 (26.8)
Both venoarterial and venovenous	19 (5.3)
Not reported	98 (27.4)
Maternal overall survival at 30 d, n (%)	270 (75.4)
Antepartum	65 (81.2)
Immediate postpartum	58 (84.1)
Postpartum	85 (67.5)
Timeframe unspecified	62 (75.6)
Maternal overall survival at 1 y, n (%)	266 (74.3)
Maternal complications, n (%)b
Mild to moderate bleeding	66 (18.4)
Severe bleeding requiring surgical intervention (laparotomy, hysterectomy)	48 (13.4)
Intracranial neurologic morbidity	19 (5.3)
Vascular complications (extremity ischemia, infection)	14 (3.9)
Deep vein thrombosis	10 (2.8)
Other (peripheral neuropathy, need for rehabilitation on discharge, hyperbilirubinemia, reperfusion injury, bradycardia with cannulation)	22 (6.1)
Fetal survival, n (%)	44 (64.7)
Fetal complications, n (%)b
Preterm delivery	33 (48.5)
Neonatal intensive care unit admission	19 (27.9)

^aSome cases had multiple indications for ECLS.

^bSome cases had multiple complications.

The most common indications for ECLS overall in pregnancy included ARDS 177 (49.4%), cardiac failure 67 (18.7%), and cardiac arrest 57 (15.9%). Table 2 reports the survival of patients based on indication for ECLS. The most common maternal complications included mild to moderate bleeding 66 (18.4%), severe bleeding requiring surgical intervention 48 (13.4%), and intracranial neurologic morbidity 19 (5.3%). There were 18 (5.0%) patients who had peripheral neurologic deficits or required rehabilitation resulting in an intact neurologic maternal survival of 245 (68.4%).

Table 2.

Extracorporeal Life Support Indications and Outcomes

	Total, n (%)	Survival, n (%)
Indications
Acute respiratory distress syndrome	177 (49.3)	141 (79.7)
Cardiac failure	67 (18.7)	52 (77.6)
Cardiac arrest	57 (15.9)	50 (87.7)
Peripartum cardiomyopathy	45 (12.5)	36 (78.3)
Pulmonary arterial hypertension	28 (7.8)	14 (50)
Amniotic fluid embolism	27 (7.5)	14 (51.9)
Pulmonary embolism	17 (4.7)	11 (64.7)
Heart disease	14 (3.9)	11 (78.6)
Septic shock	11 (3.1)	10 (90.9)

Outcomes separated by indication for extracorporeal support.

There were 81 (22.6%) cases in the antepartum period, 69 (19.3%) cases in the immediately postpartum period (<1 day), and 126 (35.2%) cases in the postpartum period (1–42 days). There were 82 (22.9%) cases that could not be classified into a defined time period. The most common antepartum indications were ARDS 53 (65.4%), cardiac failure 8 (9.9%), pulmonary arterial hypertension 7 (8.6%), and cardiac arrest 7 (8.6%). Complications included mild to moderate bleeding 27 (33.3%) and severe bleeding 7 (8.6%). Cannulation was venovenous in 46 (56.8%), venoarterial in 21 (25.9%), venoarterial+venovenous in 5 (6.2%), and unknown in 9 (11.1%). The average duration of ECLS was 10.8 days. Maternal survival was 65 (80.2%), fetal survival was 44 (64.7%). The most commonly reported fetal complications include preterm delivery 33 (48.5%) and neonatal intensive care unit admission 19 (27.9%). Fetal neurologic complications included intraventricular hemorrhage 2 (2.9%), ventriculomegaly 1 (1.5%), ventriculomegaly/cerebral ischemia 1 (1.5%), and asphyxia with therapeutic hypothermia 1 (1.5%).

The most common immediately postpartum indications included cardiac arrest 39 (56.6%), cardiac failure 16 (23.2%), and amniotic fluid embolism 15 (21.7%). Complications included mild to moderate bleeding 17 (24.6%) and severe bleeding 18 (26.1%). Cannulation was venoarterial in 53 (76.8%), venovenous in 6 (8.7%), venoarterial+venovenous in 6 (8.7%), and unknown in 4 (5.8%). The average duration of ECLS was 5.5 days and maternal survival was 58 (84.1%).

The most common postpartum indications were ARDS 50 (39.7%), peripartum cardiomyopathy 32 (25.4%), and cardiac failure 24 (19.0%). Complications included mild to moderate bleeding 17 (13.5%) and severe bleeding 18 (14.3%). Cannulation was venoarterial in 63 (50%), venovenous in 37 (29.4%), venovenous+venoarterial in 7 cases (5.6%), and unknown in 19 (15.1%). The average duration of ECLS was 17.9 days and maternal survival was 85 (67.5%).

There were 35 deliveries on ECLS at an average gestational age of 26.3 weeks. Women who had deliveries on ECLS had a survival of 27 (79.4%) and of the reported fetal outcomes, the fetal survival was 18 (56.3%). There were 14 (41.2%) vaginal deliveries, 1 (2.9%) vacuum‐assisted vaginal delivery, and 20 (57.1%) cesarean deliveries. Of the deliveries on ECLS, 16 (45.7%) were performed with the patient on anticoagulation including 10 (63%) cesarean deliveries and 6 vaginal deliveries (37%). Anticoagulation was held for the delivery in 3 (8.6%) cases and 16 (45.7%) did not explicitly outline anticoagulation management. Additional details on these cases can be found in Table S2. There were 11 cases including cardiopulmonary bypass pre‐ or post‐ECLS. Additional details on these cases can be found in Table S3.

We also analyzed cases based on the cannulation mode as seen in Table 3. Extracorporeal carbon dioxide removal removal 4 (1.1%), extracorporeal arteriovenous carbon dioxide removal removal 1 (0.28%), interventional lung assist membrane 2 (0.56%), and percutaneous extracorporeal lung assist 1 (0.28%) were included within the venovenous cases. Venoarterial cases were more likely to be of shorter duration (average 6.9 days) and for cases of cardiac arrest (28.3%), cardiac failure (26.2%) or peripartum cardiomyopathy (25.5%). Venovenous cases were on average longer in duration (average 18.8 days) and the vast majority were indicated for ARDS 74 (77.1%) followed by asthma 6 (6.25%) and cardiac failure 4 (4.2%). A smaller percentage of patients (5.3%) had both venoarterial or venovenous at different times during the peripartum period and this was most frequently used for cardiac arrest 6 (31.6%), ARDS 6 (31.6%), or cardiac failure 4 (21.1%). The survival between these groups as well as reports where the cannulation mode was unknown was comparable.

Table 3.

Cannulation Outcomes

Venoarterial	Total, n (%)	Venoarterial+Venovenous	Total, n (%)
	145 (40.5)		19 (5.3)
Indications		Indications
Cardiac arrest	41 (28.3)	Cardiac arrest	6 (31.6)
Cardiac failure	39 (26.9)	ARDS	6 (31.6)
Peripartum cardiomyopathy	37 (25.5)	Cardiac failure	4 (21.1)
Duration	6.9 d	Duration	13.9 d
Maternal survival	105 (72.4)	Maternal survival	14 (73.7)
Maternal complications		Maternal complications
Mild to moderate bleeding	29 (20.0)	Mild to moderate bleeding	6 (31.6)
Severe bleeding	23 (15.9)	Severe bleeding	5 (26.3)
Intracranial complications	10 (6.9)	Vascular complications	3 (15.8)

Venovenous	96 (26.8)	Unknown Cannulation	98 (27.4)
Indications		Indications
ARDS	74 (77.1)	ARDS	83 (84.7)
Asthma	7 (7.3)	Cardiac failure	20 (20.4)
Cardiac failure	4 (4.2)	Cardiac arrest	7 (7.1)
Duration	18.8 d	Duration	12.7 d
Maternal survival	77 (80.2)	Maternal survival	74 (76.3)
Maternal complications		Maternal complications
Mild to moderate bleeding	29 (30.2)	Mild to moderate bleeding	4 (4.1)
Severe bleeding	12 (12.5)	Severe bleeding	8 (8.2)
Intracranial complications	6 (6.3)	Deep vein thrombosis	5 (5.1)

Indications and outcomes separated by cannulation type. ARDS indicates acute respiratory distress syndrome.

Discussion

In this study, we found an overall 30‐day survival rate of 75.4% (n=270) for mothers requiring ECLS and a survival rate of 64.7% (n=44) for fetuses exposed to ECLS. However, our fetal survival rate is limited by missing fetal outcomes and lack of long‐term follow‐up. Complications associated with ECLS in pregnant patients were consistent with the general population including bleeding, deep vein thrombosis, and vascular complications.13, 14, 15 The indications for ECLS and cannulation methods varied based on the timing relative to delivery. From the few reviews published to date, the survival rate for pregnant patients undergoing ECLS has been significantly higher than the overall survival with adult ECLS for pulmonary (59%) or cardiac (43%) causes16 with reported maternal survival rates ranging from 70% to 80%7, 8, 9, 10 and reported fetal survival rates 65% to 72%.7, 8, 9, 10

Despite concerns that pregnant or immediately postpartum women may be at risk of bleeding complications and/or at risk for thromboembolic phenomenon in the immediate postpartum period, our findings do not reflect this concern and suggest that the conditions leading these women to require ECLS may be reversible and in fact, more amenable to mechanical support than the standard adult ECLS. Mild to moderate bleeding complications were noted in 18.4% (n=66) of cases and severe bleeding requiring operation was present in 13.4% (n=48) of cases; this was comparable to the reported range in other adults studies ranging from 28% to 32%.13, 14, 15 Deep vein thromboses and vascular complications were uncommon at 2.8% (n=10) and 3.9% (n=14) respectively and comparable to the ranges reported for limb ischemia in other adult populations, 2% to 14%.13, 14, 15, 17 Collectively, our findings demonstrate that pregnant patients had more favorable survival than prior reported rates for the general population with similar rates of complications.13, 14, 15, 17 These patients represent an overall younger, healthier group who are more likely to have ECLS for acute, reversible indications than the general population of ECLS cases and our findings reflect this underlying favorability for better outcomes.

Survival varied depending on the indication for ECLS. Women who were cannulated for cardiac arrest had a survival rate of 87.7%; comparing this with a population‐based study on maternal cardiac arrest that reported a 58.9% survival suggests that ECLS may have a role in improving clinical outcomes in this context.18 Neurologically intact survival for patients who undergo extracorporeal cardiopulmonary resuscitation has been reported at 28.5%, in this study it was found to be 78.9% with neurologic deficits including hemiparesis 1 (1.8%), limb motor deficits 2 (3.5%), hand weakness 1 (1.8%), and need for rehabilitation on discharge 2 (3.5%).19 The current literature regarding amniotic fluid embolism reports a wide range of survival from 39% to 89%.20, 21, 22 This review found an overall survival of 51.9% (n=14), consistent with the current literature; however, the rate of neurologically intact survival was 74% (n=20) compared with 15% in previously reported literature.20 Pulmonary arterial hypertension has a very poor prognosis in pregnant women and prior case series have reported very poor survival rates with ECLS (16.7%); however, our study reported a survival of 50% (n=14).23 Several reports of antepartum placement of venous and arterial sheaths for potential rapid initiation of ECLS in patients with severe pulmonary arterial hypertension suggest that these anticipatory interventions may improve outcomes in these high‐risk patients.24, 25

We separated peripartum patients who underwent ECLS into 3 time periods and observed that there were different indications and survival rates for advanced mechanical support at different stages of pregnancy. The majority of cases during the antepartum period were for cases of ARDS and roughly one‐third of those patients delivered while on ECLS. The decision to deliver the neonate versus continue ECLS in these patients is one that requires multidisciplinary discussion and is made on an individual basis depending on the disease process, evolution of illness, institutional experience, and expert opinion. Notably, survival in the immediately postpartum group was highest at 84.1% (n=58). The higher risk of bleeding in the immediately postpartum group (mild to moderate bleeding 17 [24.6%] and severe bleeding 18 [26.1%]) may not be surprising given the number of cases associated with amniotic fluid embolism and postpartum hemorrhage, 2 pathologies known to induce coagulopathy and hyperfibrinolysis. Of the 3 groups, postpartum ECLS had the longest average duration of support and lowest survival rates.

Cardiac arrest was the most common indication in the immediately postpartum period and many of these reports are cases of extracorporeal cardiopulmonary resuscitation, defined as application of venoarterial ECMO during cardiopulmonary resuscitation. The 57 cases of cardiac arrest in this review had very favorable survival (87.7%) compared with the general adult population survival with extracorporeal cardiopulmonary resuscitation (29%).16 These findings support consideration and use of extracorporeal cardiopulmonary resuscitation in pregnant and immediately postpartum patients as an advantageous group.

Our study has several limitations. These results may be confounded by publication bias as authors may be more likely to report favorable outcomes than poor ones. It may be that the current literature is unduly optimistic with successful cases being overrepresented but could also be underreporting of uneventful but successful cases. Another limitation is that the current data come from case reports and case series. We considered estimating our outcomes using 95% CIs; however, we did not have an unexposed population without ECLS in order to compare our outcomes. Additionally, when calculating CIs to estimate a risk/proportion, there is also the underlying assumption that the observations come from the same population with the same “true” risk and that these are randomly drawn from the population. We rejected this assumption as there is likely to be publication bias when synthesizing results from case reports and case series. Sample size limitations were also a major concern in the consideration of fitting a random effects model. Therefore, we included only the proportion of what is reported in the literature in our sample. There were 4 (1.1%) cases in which the authors noted maternal deaths more than 30 days postpartum that were included to estimate a 1‐year maternal survival, however, outcomes at 1 year were not uniformly followed nor reported and this may be an underestimation. We are unable to draw any conclusions about maternal survival over time because of the limited reports prior to 2009. However, the contemporary reports from 2009 to 2019 have a maternal survival rate of 75%. Prospective and detailed reporting with multicenter collaboration may help to better evaluate the use of ECLS in pregnancy including indications, complications, outcomes, and best management strategies for this unique population.

The findings of this study are encouraging but one also has to consider the long‐term functional outcomes in these patients. Although studies suggest that patients who have undergone ECMO are able to achieve reasonable physical and neuropsychologic recovery, functional deficits may persist.26, 27, 28, 29, 30 This is an important consideration in this young population as they may benefit from targeted medical or psychosocial rehabilitation.31 Intracranial neurologic morbidity occurred in 5.3% (n=19) of patients and overall neurologically intact survival in this study was 68.4% (n=245). These findings suggest a lower incidence of poor neurologic outcome and mortality compared with other ECMO studies that cite neurologic morbidity in up to 16.0% of patients.17, 19, 32, 33 Another limitation to this study is the lack of information on fetal outcomes as they were only reported in 84.0% (n=68) of antepartum cases. With these data alone, it is challenging to accurately predict the fetal risk portended with maternal ECLS. Prospective studies may better elucidate the association of maternal and fetal outcomes with ECLS use during pregnancy.

These limitations notwithstanding, our findings prompt providers to consider the use of ECLS during the peripartum period as a potentially lifesaving intervention for patients with cardiac and/or respiratory failure. Centers that manage high‐risk pregnancies, particularly those that include women with congenital heart disease, cardiovascular disease, and/or severe respiratory illness, should be prepared to initiate and manage patients with ECLS in the event of a devastating cardiopulmonary event during the peripartum period.

Conclusions

ECLS in the peripartum period has been successfully utilized with a maternal survival rate of 75.4% and should be considered in catastrophic cardiopulmonary conditions. Pregnancy brings a set of challenges and unique considerations for ECLS. At present, there are no formal guidelines to lead physicians to best manage these patients and future directions of research include optimal anticoagulation strategies, choice of cannulation sites, fetal monitoring, and/or method and timing of delivery. Future studies may want to assess long‐term outcomes of neonates born to women who underwent ECLS during the antepartum period. Moving forward, to more clearly assess outcomes, an emphasis should be made on reporting all cases of ECLS in pregnancy in a prospective manner with more granular case details to establish an inclusive assessment of outcomes and complications in this unique population.

Sources of Funding

This work was supported by the University of Michigan Department of Anesthesiology.

Disclosures

None.

Supporting information

Data S1 Tables S1–S3 References 7, 8, 9, 23, and 34–250

(J Am Heart Assoc. 2020;9:e016072 DOI: 10.1161/JAHA.119.016072.)