Hi. To everyone watching, my name is Emily Naum, and I am a dual-trained critical care physician and obstetric anesthesiologist at Massachusetts General Hospital. I'm here today to speak about perioperative intensive care of the high-risk obstetrics patient. To quickly go through goals for this session, I'll review current trends in maternal morbidity and mortality, ICU admission characteristics in this population. We'll go through pregnancy physiology and how that affects our ability to detect deterioration in pregnant women, and identify opportunities for nontraditional critical care administration models in these patients. So let's start with some epidemiology. Severe maternal morbidity are CDC-defined unexpected outcomes of labor and delivery that result in significant short- or long-term consequences to a woman's health. There are 21 indicators with conditions in various systems as noted here. The rate of severe maternal morbidity is rising over the last two decades by almost 200% overall since 1993. This increase has been mostly driven by blood transfusions, but even when you exclude those, the rate of severe maternal morbidity increased by about 20% over time from 1993 to 2014. Severe maternal morbidity is associated with ICU admission as well as mortality. Recognizing who's at risk for these complications when they appear on the labor floor is important to maintain vigilance and monitoring in those people who are most likely to experience morbidity. A study was done in 2013 of more than 850,000 women using Medicaid data to look at conditions associated with the CDC-defined diagnoses for severe maternal morbidity. The conditions were weighted, and a comorbidity index was created and subsequently validated in the clinical setting. I'll talk more about this later as a way to flag at-risk patients, so let's put a pin in this for now. Medical conditions associated with increased morbidity include chronic heart, liver, and renal disease, lupus, and HIV. Obstetric conditions include preeclampsia with severe features, multiple distation, and prior cesarean delivery. Social risk factors include public or absent health insurance as well as lower education levels and lower incomes. Non-Hispanic racial and ethnic disparities have also persisted in maternal morbidity and mortality. This gap exists in every racial and ethnic minority category compared with deliveries among non-Hispanic white women. Non-Hispanic black women are particularly affected with severe maternal morbidity occurring in 230 per 10,000 delivery hospitalizations compared to only 139 in non-Hispanic white women. Maternal mortality is a very real problem in both low-income and high-income countries. As you can see here, the U.S. maternal mortality rate is quoted at 17.3 per 100,000 live births based on a study of pregnancy-related mortality by the CDC. This is substantially increased over the last two decades from 10 in 100,000 in 1990, which is closer to the current mortality rate in the U.K. An important point to be made here is that several large-scale studies have shown that ICU admission and maternal mortalities have been deemed potentially avoidable in a significant number of cases. Two of the most common factors include lack of recognition of complexity or seriousness of the condition by the caregiver and a lack of knowledge and skills by staff. This figure comes from a retrospective study of maternal mortality in the United States. It demonstrates the leading cause of maternal death in the U.S. as cardiovascular conditions with slightly more than a quarter of deaths followed by pre-existing illness, sepsis, and then hemorrhage. Of note, this study did not include psychiatric conditions, which has been increasingly highlighted as a leading cause of maternal morbidity. A recent article in JAMA Network Open found that the number four leading cause of pregnancy-related maternal death is perinatal mood and anxiety disorders. Reviewing the trends, the rates of maternal deaths from hemorrhage, hypertensive disorders of pregnancy, and anesthesia complications have decreased over time. However, one condition that deserves highlighting is that of maternal sepsis, with a mortality rate that remains around 12% over the last three decades, despite the publication and implementation of the surviving sepsis guidelines. This figure comes from the UK Embrace, which is a collaboration that conducts surveillance and inquiry into maternal deaths. It shows similar trends to the U.S. reports, with cardiac disease leading the causes of death, followed by thromboembolism, neurologic morbidity, and here they include psychiatric disease. Following this, we see that pre-existing illness and sepsis that we see in higher numbers in the U.S. maternal mortality study. Importantly, they found that deaths related to pregnancy-related infections actually increased from 2012 to 2018, again highlighting the need for better detection and management in this at-risk population. Mortality in maternal health and outcomes remains an ongoing battle. This figure comes from the CDC, and it demonstrates clear racial disparities in the incidence of pregnancy-related mortality. Variability in the risk of death by race and ethnicity may be due to several factors, including access to care, quality of care, prevalence of chronic diseases, structural racism, and implicit biases. So how often do we actually see pregnant patients in the ICU? The estimates in ICU admission are challenging due to lack of consistency in definition, and it's likely underestimated due to confounders like excluding women who were critically ill but not ultimately admitted to an ICU, or women who didn't suffer significant organ failure. Overall, 1-3% of peripartum women require ICU admission or critical care services in the U.S. each year. This is comparable to rates in other developed nations. Most of these admissions are postpartum. The causes of ICU admission do not actually differ significantly in high-income versus low- and middle-income countries, although there is a significantly higher maternal mortality rate in low- and middle-income countries compared to high-income. Hemorrhage and hypertensive disorders of pregnancy account for the majority of maternal ICU admissions, although the exact case mix differs widely in each country and even within different regions. This is likely multifactorial and related to differences in preconception counseling, prenatal care, socioeconomic and environmental factors, surgical practices, and regional ICU admission guidelines. This figure shows the causes of maternal ICU admission in high-income countries across several studies ranging from 2006 to 2019. These were all retrospective analyses of admission diagnoses in the U.S., Canada, the Netherlands, Italy, France, and New Zealand. This table comes from a study of obstetric ICU admissions in New Jersey that highlights significant risk factors for ICU admission, including hypertensive disorders of pregnancy, cesarean delivery, multiple gestation, and hemorrhage. This study looked at over 18 million live births in the U.S. between 2012 and 2016. The authors developed a predictive model looking at factors associated with maternal ICU admission and identified cesarean delivery, hypertensive disorders, and induction of labor as the highest risk factors for critical care needs. Other notable risk factors include a BMI greater than 50, and again, to highlight the racial disparities with Black and Hispanic women having an increased risk of admission compared with White women. I mentioned before that we would circle back to severe maternal morbidity. We discussed the 2013 study that created the comorbidity index score where various maternal and obstetric conditions are assigned a weighted number and the total score is calculated. This score was primarily calculated to estimate morbidity, but this figure shows one of the secondary outcomes of that study was ICU admission, which was also found to be associated with higher comorbidity index scores, or CMI scores. Clinically, simplified tools have allowed us to determine these scores, and they can be used to make a universal baseline assessment of every patient on a labor and delivery floor. A higher initial CMI score draws attention to those patients who may require closer monitoring and help to influence providers to escalate care more promptly. If you know that they're at risk of being sick, you may be more likely to tune in to even subtle abnormalities when they develop and capture evolving complications earlier in their time course. One cannot give a maternal critical care lecture without a brief review of the physiologic changes of pregnancy. Maternal changes of pregnancy are related to the hyperdynamic state. Increased cardiac output comes from an increased stroke volume and heart rate and increased contractility. SVR and PVR decrease throughout pregnancy, resulting in lower systolic blood pressures, which nature in the second trimester. Pulmonary changes include an increased tidal volume related to increased anterior-posterior diameter of the chest, and this results in a respiratory alkalosis that's compensated by renal bicarbonate excretion. There's a physiologic anemia of pregnancy due to a disproportionate increase in plasma volume compared to red blood cell volume. Most clotting factors increase, resulting in a hypercoagulable state. White blood cell count increases in pregnancy, and this is generally a reflection of an increase in the number of polymorphonuclear cells with the appearance of immature granulocytic forms. Levels of IgA, IgG, and IgM are unchanged during gestation, but humoral antibody titers to certain viruses are decreased. Renal blood flow and creatinine clearance increase throughout the pregnancy, resulting in a lower serum creatinine. And progesterone effects on the GI system include biliary stasis. This contributes to the increased incidence of gallbladder disease during pregnancy. And finally, increased abdominal pressure and reduced lower esophageal sphincter tone increase the risk of gastric aspiration in pregnant patients. Keeping in mind these physiologic changes, I want to reflect upon the scoring systems that we have to detect sepsis in the general population and challenge you to apply them in the peripartum period. The SIRS criteria actually describes normal pregnancy physiology in three out of the four values that we look at. These have been found to be very sensitive but poorly specific in pregnancy. The QSOFA score is more specific, however, the sensitivity has been found to be only 50% in pregnant patients who had known sepsis due to the rarity of mental status changes and unreliable estimates of respiratory rates. The SOFA score has better predictive value for patients who are admitted to the ICU for in-hospital mortality than SIRS or QSOFA, but poor specificity for maternal sepsis again due to the physiologic changes of pregnancy. And finally, the APACHE-2 score consistently overestimates mortality risks for pregnant and recently pregnant women. Because of these issues detecting vital sign changes in pregnant patients, comprehensive early warning systems have been developed specifically for obstetric patients. These include the Maternal Early Warning Criteria or MEW criteria and the Modified Early Warning Score and variants of these have been studied and implemented across the world. The MEW criteria require one or more of several physiologic alterations to trigger a clinician evaluation and possible escalation of care if needed. A case control study in 2018 looking at patients with maternal sepsis showed that the MEW criteria had a sensitivity of 82% and a specificity of 87% and that this performed significantly better than SIRS or QSOFA. Another score, the Modified Early Obstetric Warning Score, uses similar physiologic variables to the MEW criteria but with additional components that are noted here. It includes different levels of trigger based on the degree of abnormality. A case control validation study looking at overall maternal ICU admission, so not just for sepsis, showed that this modified score had a high sensitivity at 96% and a specificity of 73% when using one or more red trigger values. Notably, the four variables that were significantly associated with ICU admission at greater than 24 hours were maximum temperature, heart rate, systolic blood pressure, and respiratory rate. Integrating these warning systems into clinical pathways has real appeal to standardize evaluation and early intervention in at-risk or deteriorating patients. This figure comes from a study published by a group in California where they reported nearly 120,000 deliveries after implementation of this tool. It includes early warning triggers and pathway specific recommendations to expedite treatment for the most common areas of morbidity, sepsis, cardiac dysfunction, hypertension, and hemorrhage. This group actually noted a statistically significant reduction in severe maternal morbidity when comparing baseline data to after implementation of the tool. Identifying patients who are developing critical illness is really challenging, and overcoming the cognitive bias that most of these patients are healthy young women who couldn't possibly be developing critical illness is another real challenge. These kinds of systems help to create a universal check-in for providers on labor and delivery floors to objectively evaluate the patients and proactively manage complications to prevent further morbidity. Our biggest challenge is when these patients exist in the gray part of the spectrum of critical illness. Sometimes though, there are clear indications for admission. Mechanical ventilation, vasopressor support or invasive monitoring, ongoing bleeding, acute liver, cardiac, or neurologic failure all generally require an ICU admission at most institutions. As our pregnant population evolves into more complex patients with more chronic illness, we should consider directions for the future. An important evolution in practice is determining the best location to deliver care to critically ill obstetric patients, particularly those who are not yet delivered. The advantages of delivering on the labor floor are increased physical space, familiarity with obstetric interventions and access to operating rooms, and a reduced risk of nosocomial infections. Conversely, the advantages of delivering in the ICU are immediate availability of ICU providers, advanced monitoring, and higher-level interventions. The choice really depends on the degree of patient instability and the interventions required. The emphasis should remain on maternal well-being as the primary goal. Providers who regularly work in labor and delivery have the greatest familiarity with obstetric interventions, obstetric medications, which are not routine in the ICU, and close access to operating room facilities for potential delivery. Although many intensivists consider the safest place for all patients in the hospital to be the ICU, this is a rare and highly specialized patient population. The majority of intensivists do not routinely care for pregnant patients, and especially not in the immediate peri-delivery period. So those providers who do have experience with obstetric, obstetric anesthesia, high-risk medical, and or surgical care to pregnant patients on a more regular basis should be available to provide direct or consultant care. Finally, recall that I mentioned psychiatric illness is a leading cause of death and depression and PTSD are very real complications from ICU admission. The importance of emotional attachment and maternal-fetal bonding must not be overlooked. An alluring prospect is the idea of graduated escalation of critical care delivery. A nice article published in 2019 outlines the cosmic paradigm. This is a four-step system of early multidisciplinary consultation, automated surveillance, monitoring, and intensive care aimed to identify high-risk and clinically deteriorating patients. The proposed concept is that at-risk patients are best managed on the labor and delivery floor with a remote ICU telemedicine service, except in the most extraordinary cases. This model provides a monitoring framework on labor and delivery using nurses who have both ICU and obstetric training, ICU-level patient monitoring, and telemedicine consultation and select high-risk OB patients to add an additional layer of vigilance for patients who may not otherwise be sick enough for ICU admission, but when you want to bring an ICU level of care to labor and delivery. Proponents of this model highlight the importance of immediate proximity to obstetricians, obstetric anesthesiologists, and the operating room. I want to close by taking a high-level view in maternal critical care with a big step back. Preconception counseling and medical optimization is crucial in high-risk obstetric patients. Upon admission, patients should be assessed for risk of morbidity, and they should be continuously monitored for development of complications using maternal-specific warning systems. If critical illness develops, there are bundles in place for common maternal conditions that are associated with morbidity, and these emphasize prompt assessment and treatment, multidisciplinary involvement of care teams, and escalation of care when appropriate. Thank you for taking the time to listen. My email is listed here on this slide. I welcome any and all feedback or conversation surrounding maternal critical care. Please feel free to contact me. Looking forward to meeting in real life whenever we're back to in-person meetings. Thank you so much again for listening.

perioperative intensive care

maternal morbidity and mortality

severe maternal morbidity

racial and ethnic disparities

Maternal Early Warning Criteria (MEW)