Amniotic Fluid Embolism:

Latest Concepts and Current Research

This site is intended to provide background on Amniotic Fluid Embolism and give readers information about on-going research about this leading cause of maternal mortality in developed countries.

Research History


            In 1993, Dr. Benson published his first paper regarding amniotic fluid embolism.[1] This was a case series of three patients with amniotic fluid embolism that he had personally cared.  The first patient, encountered during is obstetrics-gynecology residency at Prentice Women’s Hospital of Northwestern, experienced sudden shock and loss of consciousness shortly after birth.  She ultimately survived but received 108 units of blood products and required a hysterectomy to control her bleeding. 

            The second patient collapsed at home, but responded to CPR administered by her husband.   She underwent emergency cesarean section upon arrival at the hospital but developed disseminated intravascular coagulation within an hour of delivery and required 8 units of blood products.  Her child was born alive but developed cerebral palsy.  The patient developed Hepatitis C from the blood transfusion as this was shortly before the availability of Hepatitis C testing.

            Finally, the third patient experienced no significant adverse events.  She experienced a transient drop in her blood pressure with a concomitant increase in her pulse.  The hypotension responded well to intravenous hydration, although she had a persistent and substantial elevation in her pulse.  Laboratory evaluation revealed a substantial drop in her clotting factors that did not require treatment.

            Taken together, Dr. Benson attempted to create a new definition of amniotic fluid embolism dependent on pulse and blood pressure alterations.  The third patient in the series, who had only transient symptoms and laboratory abnormalities, suggested that AFE might more common than previously believed and less fatal than the oft-cited 90%.

            This proposed definition did not survive the test of time because of the substantial variation in which AFE can present to say nothing of the fact that some patients die within minutes and for that reason alone would not meet the blood pressure and pulse criteria.  However, two concepts emerged from the paper that did have future utility.  First, the notion that there are more nonfatal cases than previously recognized (as had been suggested by others, previously) has been born out with lower fatality rates reported throughout the world.  Second, it was proposed that this was not an embolic event as classically understood but rather a specific type of allergic reaction to fetal material leaking into the maternal circulation.  Although the notion that AFE might be some sort of immune response to the fetus had been proposed earlier, the contribution of this 1993 paper was to specifically suggest that AFE might actually be an anaphylactic reaction to the fetus.

Link 1



            Within a year or two of the 1993 case series, Dr. Benson realized that the anaphylaxis idea might be provable in humans.  To date, trying to reproduce the disease in animals has simply not been very helpful.  Mammals simply do not reliably become ill when amniotic fluid is artificially injected into their circulation.  In fact, this has been attempted within the same species using the animal’s own amniotic fluid, fluid from another animal, and even fluid with meconium (fetal bowel movement) in it.  The only approach that makes animals sick consistently (and even then, not in all cases) is to inject fluid from another species.  Obviously, this limits the utility of studying the disease in animals since it simply does not resemble the human experience.

            Yet the anaphylaxis hypothesis is testable—in humans.  In 1996, Dr. Benson, and a senior scientist in the diagnostic division of Abbott Laboratories, published a brief paper describing the possibility of testing for anaphylaxis in AFE patients.[2] For the past two decades, anaphylaxis has been confirmed in the laboratory by checking for Tryptase levels in the blood stream.  Anaphylaxis occurs when the immune system recognizes foreign material when a specific type of immune protein, Immunoglobulin Type E (IgE) binds to the invader and then attaches to Mast cells.  Mast cells, present in both tissue and the circulation, release large amounts of histamine when IgE attaches to their surface.  Histamine causes blood vessels to dilate which increases blood flow to the area in which the foreign material is present.  Of course, this can also cause blood pressure to drop.  Another effect of histamine is to cause airway constriction which accounts for the difficulty breathing in anaphylaxis.  Unfortunately for the lab, histamine disappears from the blood stream within minutes.  Yet mast cells also release another protein, Tryptase, which stays in the blood stream for hours.  Although the function of Tryptase is not known, it has served as a useful marker for anaphylaxis.  Benson and Lindburgh suggested tryptase could serve as a useful test for the anaphylaxis hyptothesis.  On this basis, Benson created a clinical protocol to test for tryptase and other immune markers in patients with AFE at Evanston Hospital.  (Evanston Hospital is tertiary care facility in the northern suburbs of Chicago that has merged with adjacent hospitals over the years.  At one point, this hospital system was known as Evanston Northwestern Healthcare and is now know as North Shore University Health System.)



            In 2001, Dr. Benson and colleagues, both at Evanston Hospital and in Japan published a study in which 9 AFE patients and 21 healthy controls were tested for both anaphylaxis and Complement activation.[3]  No evidence for anaphylaxis was found in any of the AFE patients while all of the eight patients in whom it could be tested, showed evidence of substantial Complement activation.  The control group of healthy, labor women also did not show any evidence of anaphylaxis.  Of interest is the fact that while their Complement levels all remained in the normal range, they too, showed evidence of mild Complement activation.  The biological significance of this novel observation is unknown although it is felt to be important.  The phenomenon of sub-clinical complement activation during normal parturition remains an area of active research interest.

Link 2



            In this year two investigations were published.  The first was a study of 12 healthy laboring patients in which Complement levels and a specific fetal antigen were measured.[4]  The women were compensated volunteers who had a second IV established solely for research purposes.  The study was intended to specifically test the hypothesis that the shaking chills that some women experience during labor were the specific result of fetal antigen leaking into the maternal circulation—and possibly resulting in Complement activation.  The hypothesis was proven wrong although the phenomenon of Complement activation during normal parturition was observed again (in a different group of heathy women).

Link 3


            The second paper published in 2006 reflected a collaboration between Dr. Benson and a group of Norwegian anesthesiologists.[5]  The anesthesiologists were studying a series of 12 patients who experience cardiovascular collapse shortly after receiving epidural anesthesia during labor.  The authors suggested that the presence of blood clotting abnormalities could be used to distinguish those patients with AFE from those who suffered a drug reaction.

Link 4



            In 2007, Dr. Benson published “A Hypothesis Regarding complement Activation and Amniotic Fluid Embolism” in the journal, Medical Hypotheses.[6]  This paper suggested that AFE might not be an embolic event resulting from the physical obstruction of maternal pulmonary capillaries by fetal material.  Rather, the disordered physiology might be caused be fetal antigens leaking in the maternal circulation and activating the immune system.  The clotting abnormalities sometimes seen with AFE may be a secondary result of immune activation.  The nature of the fetal antigen may be a common antigen presented abnormally (timing, frequency or amount) or by small amounts of an uncommon antigen.

Link 5


            In a letter to the editor, Dr. Benson, and a Japanese colleague, Hidezazu Oi, presented a possible case of mild amniotic fluid embolism in which a patient experienced transient shortness of breath and a blood pressure drop that resolved within ten minutes of onset.[7]  She was enrolled in the AFE protocol and consent was obtained although she ultimately did not require treatment and had a normal vaginal birth.  The unusual aspect of her hospital course is that she demonstrated several abnormalities of her clotting laboratories that did not fully resolve for a week.

Link 6



            In 2008, a description of shaking rigors encountered by roughly 450 patients in normal labor was published.[8]  Over half of first time mothers experienced these chills and the typical episode lasted for 20 minutes.  There was no relationship between rigor onset and rupture of membranes or delivery, which seemed to refute the informal hypothesis that the chills may be related to these events.  The most important association with shaking rigors was epidural anesthesia and maternal fever—although much of the variation in experience remained unaccounted for.

Link 7

             At the annual meeting of the Society for Gynecologic Investigation, held in San Diego, Dr. Benson presented the preliminary description of his research into disseminated intravascular coagulation—a clotting disorder frequently seen with AFE.[9] The study was specifically intended to test the hypothesis that DIC would be associated with abnormal amounts of fetal material in the maternal circulation while other causes of life-threatening maternal hemorrhage would not be.  59 women who underwent a hysterectomy at the time of childbirth to control hemorrhage were evaluated.  Portions of their uterine walls were taken out of storage and examined microscopically.  Specifically, the blood vessels in the uterine wall were examined for the presence or absence of fetal material (skin cells, bile, hair).

            There were seven different possible diagnoses for the hemorrhage that required emergency surgery.  Each diagnostic category had fetal material in the maternal blood vessels approximately one-third of the time.  DIC as a category did not stand out one way or the other.  Furthermore, presence or absence of fetal material was not related to total blood loss.  As a result, the study cast considerable doubt on the idea that fetal material per se leaking into the maternal circulation causes obstetrical DIC.  By extension, it raises some doubt about the link between abnormal amounts of fetal material and AFE as well.

Link 8


How is Research Done?


            Medical research in humans has come a long way from the dissection of corpses and animals in the day of William Harvey.  Two separate disciplines, ethics and statistics have informed modern practice and are necessary to understand for an appreciation of some of the research challenges posed by amniotic fluid embolism.

Human Research Ethics in Practice

            First and foremost in enrolling patients in research study is the necessity of obtaining informed consent.   The first published, generally accepted, guidelines concerning ethical human research came out of war crimes trials in Nuremburg, Germany at the end of World War II.   The Nuremberg Code emphasized the rights of the individual to choose to participate—or not—and to end study participation at any time.  Furthermore, individual risk and suffering had to be minimized and a specific scientific benefit has to be derived from the study.  The ten principals that emerged were further codified by the Declaration of Helsinki, first published by the World Medical Association in 1964, and revised periodically. 

            The potential ethical problems in human research (outside of the atrocities of World War II) are perhaps no where better exemplified than in the Tuskegee syphilis experiment conducted by the US Public Health Service in Tuskegee Alabama from 1932 to 1972.   Black men with syphilis were studied to learn about natural disease progression.  At the start of the study, the slightly effective treatment that was available was offered.  However, when penicillin was developed as a treatment for syphilis by 1947, study participants were not offered the cure.  In fact, diagnostic testing for neurosyphylis, in the form of a spinal tap, was presented to study participants as a “treatment.”  The study came to an abrupt end in 1972 when a Public Health Service employee, Peter Buxton, went to the press after complaining with PHS for six years about the study.  128 men of the original 400 had died of syphilis, 40 of their wives had been infected and 19 of their children had been born with congenital syphilis.

                        As a result of the Tuskegee study scandal, the US Department of Health, Education, and Welfare commissioned a study known as the Belmont Report to inform future legislation and oversight of medical research in the United States.  Issued in 1976, the report emphasized basic ethical principles which are now taught in medical schools.  “Respect for persons” reflects the believe that people are autonomous and that those with diminished autonomy should be protected.  “Beneficence” describes the ethical duty that goes beyond respecting autonomy of people to secure their well-being.  (For instance, a patient may request a procedure or treatment but if a physician does not believe it will be helpful, is ethically obligated to decline.)  Finally, “justice” refers to the relationship between a study’s benefits and its risks—not just to individuals but as these risks and benefits affect different populations differently.

            Although little recognized, the Tuskegee study reentered American consciousness in the 2008 presidential campaign between Barak Obama and John McCain.  Mr. Obama’s minister, Reverend Jeremiah Wright, had been widely televised as stating during one of his sermons that, “The United States Government had developed the AIDS virus to commit genocide on black people.”  While many commentators were outraged at this outlandish claim, it was apparent from their remarks that they were unfamiliar with the reality of the Tuskegee study in which the government did withhold treatment from black men for a sexually transmitted infection.  Of course, Reverend Wright failed to mention that the United States Government has established a huge oversight mechanism to be sure that another Tuskegee study never happens.

            As a result of the Belmont Report and decades of legislation, human research in the United States has to be approved at the outset and then monitored by Institutional Review Boards.  Comprised of both lay people and physicians, the IRB’s are responsible for assuring that all studies that receive federal funding provide adequate protection and supervision of study participants.  IRB approval is required before any patient can be enrolled in a study and study review is required at least annually.

            Obtaining Institutional Review Board approval for a study is a highly specialized, formal process.  For example, the principal investigator who submits the protocol has to provide documentation that he or she has reviewed the National Institutes of Health course on medical research—which includes a discussion of the Nuremberg Code, the Helsinki Declaration, the Belmont Report, and the.  In general, IRB submissions are created with the help of a clinical research nurse who has specialized training in the rules and regulations surrounding clinical research at that institution (i.e. hospital or university).  Once approved, the researcher has to submit annual reports concerning patient enrollment, outcomes, data, and adverse events as well as statements regarding potential financial conflicts of interest.   In general, even studies funded by foundations and not by federal grants try to adhere to federal guidelines—both for ethical reasons.  Obviously, research subjects in privately funded studies should not have less protection than those in federal studies.  

            These ethical considerations make research more difficult and more expensive.  This is not an argument that research regulations should be relaxed.  Rather, this discussion is intended to help the reader appreciate the complexity of medical research and give insight as to why it is so expensive (and of course as to why donations are needed.)

            Although current research on AFE does not involve treatment (with potentially dangerous side effects) obtaining ANY blood or tissue samples from hospitalized patients requires informed consent.  As AFE patients are often too ill to given consent, research consent can be obtained from family members.  Since AFE is entirely unpredictable, and can occur at any time, an effective consent program requires the availability of specifically designated, highly trained individual 7 days a week.  Typically, a group clinical research nurses at a tertiary care center who participate in a variety of research studies, the University or Hospital requires substantial funding to support these nurses.

            To give but one example of the practical considerations in AFE research, even the simple collection of placentas is a significant undertaking.  Placentas (the afterbirth) are normally discarded after delivery, without any specific discussion or consent by the patient.  Although federal guidelines do not specifically call for obtaining consent for the use of tissues that are normally discarded, the IRB at Dr. Benson’s hospital has decided that informed consent is required.  Dr. Benson thought this was wise since adherence to the highest ethical standard is only prudent, and since any medical journal in which the work is published will want to know about research participant consent.  The point is that even obtaining tissue that would be normally discarded is a highly regulated process (that requires effort, time and money).

            To given another example, even though much speculation about AFE is published, physicians at institutions which do not already have an AFE research protocol in place cannot simply obtain research specimens and send them to Dr. Benson for analysis.  Failure to get informed consent from the patient in the context of an already existing human research protocol approved by that institution’s IRB could potentially place the hospital or university in violation of federal law and jeopardize all of the facility’s other research.  In short, to efficiently study a disease which occurs every one in 8 to 20,000 pregnancies, research protocols are needed at institutions in which tens of thousands of deliveries are taking place.  Again, money is needed—“no bucks, no Buck Rogers.”



            Another aspect of modern medical research is the use of statistics to inform study design and data reporting.  Both statistics itself and its interface with medicine is evolving.  One frequently cited statistic in news reporting is “significance.”  This is a reference to a calculated statistic that the difference seen between two groups is due to chance.  If the probability (“p-value”) is less than 5% that the results are due to random sampling error, the difference is said to be statistically significant.  However over the past two decades, the power of a study to detect differences between two groups has become more important.  Given certain assumptions about likely differences between two study populations and their intended enrollment size, how likely is it that the study will actually find a difference if one exists?

            As can be seen, medical research encompasses far more than enrolling patients and obtaining lab results.  A specialized statistician is usually needed to be sure that the study results are reported in the proper perspective.   In many of the AFE studies reported here, Jennifer Beaumont, was the biostatistician and is cited as a coauthor.  Of course, research biostatisticians are not directly involved in patient care and need funding from outside the institution.  In turn researchers who use them have their research accounts billed.    Statistical analysis is an important improvement in medical research but adds further expense to study publication. 

She did NOT have AFE.  Her vaginal birth was normal.  WHY?  We do not know.

 The Need For More Research

            Amniotic Fluid Embolism kills or injures several hundred mother-baby pairs in the United States every year.   Research is needed to understand the cause of the disease so that it can be prevented or more effectively treated.  Yet the importance of AFE extends far beyond its devastated victims.  At the present, there are no broadly accepted theories or mechanisms explaining how women can tolerate the presence of their fetus—50% of which comes from an unrelated father.  IF AFE does represent a maternal immune response to the baby, why don’t all mothers develop an allergic response?  Mothers cannot necessarily receive blood safely from their children nor can they invariably accept tissue transplants from their offspring.  Yet they routinely tolerate the presences of several pounds of foreign tissue during pregnancy with no apparent ill affect?  How is this possible?  The answer is key to progress on cancer, other diseases of pregnancy, autoimmune disease and transplantation biology.  It remains one of the leading mysteries confronting 21rst century biology.  It is hoped that AFE research, in which the apparent abnormal activation of the human immune system occurs, can help both its victims as well as answer the question, how can healthy pregnancies occur in the first place?