Introduction and epidemiology
The obstetric population remains difficult to assess when critically ill. Vital sign changes generally considered foreboding may be normal in pregnancy, often leading to delayed diagnosis. Furthermore, obstetric hemorrhage can be complicated by difficult identification of cause, whereas quantification of blood loss is often grossly inaccurate or impossible . Resuscitation and transfusion endpoints may be less distinct and resuscitation must account for ongoing blood loss.
Despite significant medical advances over the last century, bleeding remains the leading cause of death among parturients worldwide and a major cause of mortality in the United States . According to a 2014 World Health Organization analysis, the risks of maternal deaths from postpartum hemorrhage (PPH) in developed and developing regions are approximately 16% and 27%, respectively . In the United States, there are approximately 650 maternal deaths per year from hemorrhage , accounting for about 11% of all peripartum mortality between 2006 and 2011 .
The definition of PPH varies worldwide; in the United States, greater than 500 mL blood loss (EBL) after vaginal delivery (VD) and greater than 1000 mL after cesarean delivery (CD) are accepted . More pertinent than a definition, specialists agreed that women with greater than 1000 mL EBL with bleeding refractory to usual measures in the first 24 hours postpartum are at increased risk of hemorrhage-related morbidity . Several obstetric-specific abnormalities contribute to PPH in addition to general population risk factors, such as surgical intervention or trauma . Significant baseline systemic diseases can be major contributing factors in the incidence of hemorrhage. That being said, uterine atony accounts for up to 80% for PPH, and a large portion of patients with clinically significant bleeding is devoid of risk factors .
Bleeding risk assessment in pregnancy and specific causes
As noted, it may be difficult to predict if women will have significant PPH. Thankfully, most parturients can tolerate acute EBL of less than 1000 mL with little or no clinical significance . As such, it may be more important to identify patients at risk for morbidity and mortality to improve resource utilization, readiness, and even survival.
Uterine rupture, placenta previa, and placental abruption can all result in antepartum hemorrhage with implications for both fetal and maternal well-being. Postpartum bleeding is generally related to uterine atony, retained or invasive placenta, genital tract trauma, and surgical complications. However, antepartum and postpartum bleeding causes often overlap. For example, uterine rupture often causes antepartum hemorrhage but will also make postpartum hemostasis more complicated. The next section discusses PPH abnormalities, risk factors, and obstetric, surgical, and anesthesiologist management considerations.
The gravid uterus blood supply typically comes from 2 uterine arteries and 2 ovarian arteries in addition to vast collateralization . Total arterial blood flow term uterus is between 500 and 900 mL/min, around 12% of total cardiac output. Thus, unabated uterine bleeding could cause rapid maternal exsanguination. The key steps in hemostasis are uterine contraction and vasoconstriction, which occur postpartum in response to high levels of endogenous oxytocin and prostaglandin F2-alpha. Should the uterus remain atonic however, control of blood loss can be challenging.
For unknown reasons, preeclampsia, white and Hispanic race, and obesity are independent patient factors for atony . Chorioamnionitis also results in atony from local inflammation and acidosis of uterine tissue impairing contraction. Polyhydramnios, fetal macrosomia, multiple-gestation pregnancies, and high parity cause uterine atony due to impaired actin-myosin interaction. Prolonged or induced/augmented labor results in functional downregulation of oxytocin receptors and decreased uterine contraction . Similarly, arrested or precipitous labor may result in exhausted muscle . The presence of tocolytic agents, such as magnesium, calcium channel blockers, beta-agonists, or volatile anesthetics, can also impair uterine contraction . Interestingly, amnioinfusion may be protective against uterine atony because of a washout effect on proinflammatory mediators or bacteria . Because no specific constellation of risk factors can reliably predict atony, global preparedness for and recognition of PPH is crucial.
The third stage of labor should be aggressively managed with the overarching goal to increase uterine tone or create tamponade to slow or stop bleeding. Management starts with bimanual uterine compression but may require uterine packing, intrauterine saline balloon tamponade (Bakri Ballon, Cook Medical, Bloomington, IN), or a brace (“B-lynch”) uterine suture. These measures may either provide satisfactory tone or temporize bleeding to allow more diagnostic time or access to a more definitive measure. If bleeding remains significant, more aggressive steps including emergent hysterectomy should be considered life-saving measures . Ongoing bleeding with adequate tone should prompt evaluation for other contributing sources.
Uterine inversion presents as a protruding, briskly bleeding vaginal mass. Incidence is ∼1:2500 deliveries. It has decreased significantly since the advent of active third-stage management . Inversion is typically related to aggressive cord traction or fundal pressure. Other risk factors include nulliparity, fundal placentation, magnesium sulfate treatment, macrosomia, uterine anomalies, precipitous labor, and invasive placentation . Uterotonics may result in conversion from incomplete to complete inversion; thus, all uterotonics should be held for suspected inversion. Bleeding can be significant and requires transfusion in almost 50% of cases. The uterus must be replaced as soon as possible, often requiring extreme tocolysis with terbutaline, nitroglycerin, and even high concentrations of volatile anesthetics.
Retained products of conception
The presence of retained placenta or products of conception (POC) should prompt immediate manual or instrumental evacuation of the uterus, potentially in the operating room. A recent Cochrane Review showed no benefit to adding prostaglandin uterotonics in the setting of retained placenta. If evacuation is unsuccessful with grossly adherent placenta, the diagnosis is placenta accreta until proven otherwise (see later discussion).
Genital tract trauma
Lacerations of the perineum, vagina, or cervix are common following VD . Risk factors include nulliparity, macrosomia, precipitous delivery, operative VD, and episiotomy . Female genital mutilation also increases risk . Some degree of trauma may complicate greater than 60% of VD, but 80% of cases are minor . Lacerations are graded 1 to 4 corresponding to superficial to deep injury. More severe lacerations (grade 3–4) can bleed profusely and may involve surrounding structures, including the urethra, anal sphincter, and vaginal arteries. The latter may cause occult vaginal hematoma causing delayed, unexplained perineal pain . Small hematomas can be managed expectantly; however, large or expanding hematomas should be explored. Patients may need transfusion upon evacuation when relief of tamponade may result in profuse arterial bleeding. Repair of severe lacerations in the operating room may be preferable because up to half of women requiring surgical intervention require transfusion .
Presentation of uterine rupture is markedly varied, but is often associated with unremitting abdominal pain. There should be a high index of suspicion in at-risk patients if typical epidural “top-up” strategies provide inadequate relief or if pain timing is unrelated to contractions. Shoulder pain from referred diaphragmatic irritation is a concerning complaint . Tocographic changes are fairly common; however, no specific pattern is highly correlative . A very specific sign is the clear impression of a fetal part (such as a hand or foot) on the abdominal wall, but this is very rare. Loss of fetal station or change in presenting part may be more common. In addition, fetal bradycardia is common, indicating significant fetal distress , and may even be predictive of uterine rupture .
The overall incidence of uterine rupture is slightly less than 0.1%. The major risk factor is a history of CD, a 40-fold increase compared with women with only a previous VD . Prostaglandins should be avoided in trial of labor after cesarean (TOLAC) because of an incidence of uterine rupture of greater than 2% . If low-dose oxytocin increases rupture risk, it is minimal; however, rupture may be more likely with higher doses . Other major risk factors include previous uterine rupture and history of non–low-transverse hysterotomy or other full-thickness uterine incision . Multiple previous low-transverse CDs, advanced maternal age, macrosomia greater than 4000 g, and short interpregnancy interval may contribute to risk, but data are inconsistent . Primary uterine rupture is extraordinarily rare; however, those patients are more likely to undergo hysterectomy (35% vs 2.4%) compared with the TOLAC population .
Uterine rupture has a 60-fold increase in the risk of fetal demise , but maternal death after rupture is rare (0%–0.2% in the United States and Canada) . Still, greater than 20% of peripartum hysterectomies are required for rupture-related PPH control . Early diagnosis and management are key factors to minimizing maternal morbidity.
Suspicion of rupture should result in rapid assessment and likely delivery. VD is possible; however, CD is often required and has the added benefit of immediate surgical access. Timeliness of delivery may necessitate general anesthesia, but dosing of an indwelling epidural catheter can be attempted. A midline incision may be required for optimal access , and obstetricians must decide whether repair is possible or if hysterectomy is indicated . Hemodynamic instability should be treated aggressively with early blood transfusion, and communication is paramount to help guide clinical decision making.
Placenta previa is diagnosed when the placenta obstructs part or all of the cervical os. Placenta previa is distinct from vasa previa, the latter defined as fetal vessels, rather than placenta, that run near or over cervical os. Both may result in significant fetal morbidity, but placenta previa also carries significant concurrent maternal risk. Maternal blood loss can be marked and is classically painless; however, abdominal pain and contractions may be seen
The frequency is approximately 5/1000 deliveries worldwide . Risk factors include multiple gestation, uterine scarring, including previous CD and prior placenta previa. History of smoking, cocaine use, or advanced maternal age may also increase risk .
Diagnosis is typically made during routine antenatal ultrasonography (US). Transvaginal US is more sensitive at ruling out previa but is not necessary if low-lying placenta is not suspected. Unless bleeding, pain, or preterm contractions occur, outpatient management is advised. If symptoms do occur, patients are often admitted to the antepartum unit for observation and fetal protective strategies (eg, steroids) .
Scheduled CD is usually performed around 37 weeks’ gestation. The uterine incision should be adjusted, potentially guided by intraoperative placental mapping, to avoid cutting into an anterior placenta, because this increases the likelihood of maternal hemorrhage . After delivery, significant PPH may occur because of diffuse bleeding of the lower uterine segment placental bed. Aside from typical methods to control bleeding, suture tamponade (“oversewing”) of the area to minimize blood loss may be warranted . Preoperative large-bore venous access should be obtained, and blood products should be available before incision.
During implantation, the trophoblast does not normally invade beyond the endometrial decidua. Placenta accreta implies invasion beyond the decidua but not into the myometrium and may also refer to grossly adherent retained POC. Increta refers to an invasion into the myometrial wall, and percreta describes placental invasion beyond uterine serosa into the abdomen and potentially into surrounding structures (see Fig. 3 ).
The frequency of invasive placentation is increasing dramatically and may now exceed 1/500 pregnancies. Probability of accreta with repeat CD is impressively correlated from 0.24% for primary cesarean to 6.74% for the sixth CD. When previa is present, this risk becomes exponential; 3% of patients with previa undergoing primary cesarean had placenta accreta, but this increased to 11%, 40%, 61%, and 67% with the second, third, fourth, and fifth cesarean. In fact, any history of endometrial trespass will markedly increase the risk of invasion with subsequent pregnancy, including uterine curettage, myomectomy, pelvic radiation, or endometrial ablation, the last being especially high risk. Other risk factors include smoking, in vitro fertilization, advanced maternal age, multiparity, and short interpregnancy interval.
Diagnosis is often made on routine US examination, and there should be a high suspicion in patients with a history of uterine scarring and a previa. Sensitivity and specificity for US diagnosis have been reported as high as 90% to 100% ; however, these decrease significantly when radiologists are both inexperienced and blinded to clinical history. MRI can at times more accurately predict depth and topography of invasion; however, routine use of MRI is limited to major referral centers because of the need for a specialized radiologist. Despite unreliable imaging, antepartum rather than intrapartum diagnosis is unsurprisingly associated with improved outcomes.
Invasive placentation is strongly associated with PPH, and maternal mortality has been reported in up to 7% of accreta cases , although this is likely lower with improved diagnosis and multidisciplinary delivery planning. Planned delivery at a center comfortable with abnormal placentation may improve outcomes. Delivery timing balances maximizing fetal age with risk of emergency delivery. In the absence of bleeding, most deliveries occur between 34 and 36 weeks’ gestation. Intraoperative placental mapping can help assure an incision remote from the placenta.
Delivery is usually followed by hysterectomy during which massive acute PPH is expected. Surgical technique and hemostasis can be complicated by pregnancy-induced vascularity and distorted anatomy. Current median delivery EBL is 2 to 3 L, although EBL may exceed 5 L in greater than 40% of cases. Median transfusion requirement is 5 units of packed red blood cells (PRBC) but commonly exceeds 10 units. Even patients who survive initial resuscitation often develop significant morbidity, including disseminated intravascular coagulation (DIC), renal failure, acute respiratory distress syndrome, cardiac failure, and sepsis.
Occasionally, a uterus-sparing technique is possible in which the placental resorption is induced or a focal placental excision is performed; however, both techniques place the patient at high risk for PPH and require close postoperative monitoring, and hysterectomy may still be required. Fetal outcomes are typically good; however, outcomes correlate with gestational age.
There is no consensus regarding the best anesthetic technique during CD for invasive placentation and no evidence strongly in favor of one technique over another; thus, either neuraxial or general anesthesia can be acceptable taking into account patient and surgical considerations. The major role of the anesthesiologist is clear: adequate venous and arterial access and goal-directed (at times massive) resuscitation with hemodynamic support.
Injury to surrounding organs and tissues during delivery is possible with any surgical intervention in the obstetric population. The large pelvic blood vessels, mesentery, bowel, and bladder all may be injured during CD or hysterectomy. Repair of injuries may require consult with subspecialty surgery teams and may require adjustment to anesthetic plan.
Premature separation of the placenta from endometrium, whether complete or incomplete, complicates ∼1 in 200 pregnancies. The classic symptom is painful vaginal bleeding with or without abdominal tenderness or irritable contraction pattern. Major risk factors are hypertensive disorders and abdominal trauma, but also include substance use (cocaine or tobacco), advanced maternal age, polyhydramnios, premature rupture of membranes, leiomyomas, and personal history of placental abruption. Maternal and fetal well-being are both at risk, and urgent or emergency delivery may be necessary.
Morbidity associated with placental abruption is marked, with resultant fetal demise in up to one-third of cases. In addition, maternal tissue thromboplastin release causes massive coagulation cascade activation leading to DIC. VD is generally preferred over CD, but depending on fetal status, CD may be indicated. Neuraxial anesthesia can be attempted in the absence of coagulopathy if time allows.
Both inherited and acquired coagulopathy increase peripartum bleeding. Evidence is lacking regarding whether prophylactic treatment is advantageous in patients with known coagulopathy; and therefore, a highly individualized case-by-case approach is best with early hematology consultation when possible. In addition, cardiology or neurology consultation may be helpful in patients who are taking chronic antiplatelet or anticoagulation therapy for concurrent medical problems. Such consultations may help the anesthesia provider optimize timing of neuraxial blockade.
Patients who refuse transfusion
Patients refusing blood product administration can be particularly difficult to treat during marked PPH. Jehovah’s Witnesses (JWs) are the largest group (approximately 1.2 million people in the United States) who consistently refuse blood products.
Most JW patients refuse whole blood, PRBC, fresh frozen plasma (FFP), and platelets (PLT). However, some will accept blood “fractions,” including albumin, cryoprecipitate, and clotting factors as well as therapies such as hemodilution, cell salvage (CS), and cardiopulmonary bypass if blood is maintained in continuity with the patient’s circulation. Therapies usually widely acceptable to JWs include nonprotein colloid (hetastarch), recombinant factor VII (rFVII), recombinant erythropoietin, antifibrinolytics, and intravenous (IV) iron.
JW parturients are within their rights to refuse blood transfusion, even in life-threatening maternal or fetal scenarios. Some JW patients are well versed in the blood products or fractions that they will accept; however, all require an explicit consent process to detail and clearly document what is and is not accepted. Unfortunately, transfusion refusal is associated with a significantly increased risk for maternal morbidity and mortality, which should be fully disclosed during the consent/refusal process. One of the more striking studies was from the Netherlands, where JW parturients had a 6 times overall increased mortality risk, but a 130 times increased risk for mortality when considering only cases due to PPH.
Preparation and anticipation for peripartum hemorrhage
More than 90% of deaths related to PPH may be preventable, and mitigating hemorrhage remains a key maternal safety initiative. Protocol-driven models improve care, decrease intensive care admissions, and even reduce transfusion requirements. However, as of 2014, survey data had demonstrated that greater than 30% of obstetric care units and greater than 20% of academic centers did not have a protocol for PPH treatment.
In 2015, the National Partnership for Maternal Safety (NPMS) released the Consensus Bundle on Maternal Hemorrhage and recommended that PPH protocols be implemented on every obstetric unit. Although resource-driven institutional variability is expected, intradepartmental standardization is encouraged. The safety bundle distills “Readiness,” “Recognition and Prevention,” “Response,” and “Reporting and System Learning” with a total of 13 key elements, summarized in Box 1
Hemorrhage cart with supplies, checklist, and instruction cards for intrauterine balloons and compressions stitches
Immediate access to hemorrhage medications (kit or equivalent)
Establish a response team – who to call when help is needed (blood bank, advanced gynecologic surgery, other support and tertiary services)
Establish massive and emergency release transfusion protocols (type-O negative/uncrossmatched)
Unit education on protocols, unit-based drills (with post-drill debriefs)
RECOGNITION & PREVENTION
Assessment of hemorrhage risk (prenatal, on admission, and at other appropriate times)
Measurement of cumulative blood loss (formal, as quantitative as possible)
Active management of the 3rd stage of labor (department-wide protocol)
Unit-standard, stage-based, obstetric hemorrhage emergency management plan with checklist
Support program for patients, families, and staff for all significant hemorrhages
Establish a culture of huddles for high risk patients and post-event debriefs to identify successes and opportunities
Multidisciplinary review of serious hemorrhages for systems issues
Monitor outcomes and process metrics in perinatal quality improvement (QI) committee
Multidisciplinary preparedness and case review are recommended and can improve outcomes in complex cases. Depending on the case, a multidisciplinary team might include obstetric anesthesiologists, obstetricians, maternal fetal medicine specialists, gynecologic oncologists, interventional radiology (IR), nursing staff, pathology, critical care medicine, pharmacy, and other surgical subspecialists. Early antepartum discussion can optimize patient outcomes and formalize care plans in case of unexpected admission or need for emergency delivery. Simulation training, including a postscenario debriefing for PPH, is also recommended to foster staff familiarity and understanding of protocols and treatment principles. A widespread notification system, such as a group page or electronic board alert, can alert the entire care team that a PPH is occurring.
Use of CS in obstetrics can decrease the need for allogenic blood transfusion. Potential indications during CD include severe maternal anemia, rare blood types/difficult crossmatching, JWs, and invasive placentation. Although CS is not cost-effective or feasible for routine CD, a recent economic analysis found that it is cost-effective for high-risk cases. Unfortunately, CS may not be available when needed. A survey of UK obstetric units revealed only 21% of units had 24-hour access to CS. Thus, cases for which the care team anticipates using CS should likely be planned far in advance.
Considerations unique to obstetrics include a separate suction to waste amniotic fluid and use of a leukocyte depletion filter. A recent focused review highlighted published studies that examined CS safety in obstetric patients, and no serious complications were reported. The historical concern regarding risk of inducing anaphylactoid syndrome of pregnancy (amniotic fluid embolism, AFE) due to retransfusion of fetal tissue or amniotic fluid/cells has been largely dispelled. The leukocyte depletion filter decreases cell counts to equivalent or less than those found in maternal blood after the normal delivery process , and to the authors’ knowledge, no definitive published reports of AFE due to CS exist.
When antepartum patients present with known PPH risk factors, clinicians can attempt to optimize hemoglobin and have blood products available. Unfortunately, previously low-risk women account for approximately 40% of PPH. For women identified early, oral and at times IV iron can increase hemoglobin levels. Erythropoietin (in addition to iron) has been used to both optimize antepartum hemoglobin and restore red blood cell (RBC) mass after PPH, most commonly in populations that refuse blood transfusions. Response requires at least 2 to 3 days, and more commonly will take at least a week for noticeable improvement.
All parturients should be stratified based on hemorrhage risk factors, and appropriate venous access and blood bank status should be assured. It is neither cost-effective nor feasible in most institutions to have blood available for every parturient. The American Society of Anesthesiology Practice Guidelines for Obstetric Anesthesia recommend that a routine crossmatch is not necessary for healthy women. The authors’ institutional practice is to send blood samples for hold to the blood bank for all patients with anticipated uncomplicated VD and obtain a type and screen for uncomplicated patients presenting for CD. Additional crossmatching is based on hemorrhage risk. A sample of risk stratification for blood availability is shown in Table 1 .
|Recommended blood bank status||Patient factors or relevant medical conditions|
|Draw and hold||
|Type and screen||
|Crossmatch ≥2 units PRBC||
Severe baseline anemia
Active bleeding on admission
Known coagulation disorder, including HELLP (hemolysis, elevated liver enzymes, low platelet count) syndrome
Intrauterine fetal demise
History of difficult crossmatch or positive antibodies on type and screen
|Crossmatch ≥4 units PRBC, FFP, ± PLT||
Suspected placenta accreta, increta, or percreta based on imaging
Planned cesarean hysterectomy
Most parturients will need venous access, and institutional policy may require all patients to have an IV cannula. Often this equates to a single 18-gauge IV, but depends on the culture and practice patterns of specific obstetric units. In patients with known risk of PPH (eg, placenta previa), additional IV access is likely warranted. Although central venous access is a rarity, it should be considered when massive hemorrhage is likely or ongoing. Similarly, invasive blood pressure monitoring can be helpful during PPH to ensure rapid access to blood samples for laboratory analysis as well as timely assessment of hemodynamic changes.
Response to peripartum hemorrhage
Blood loss estimation and measurement
Clinicians are notoriously poor at estimating blood loss during and after PPH. Educational programs are effective at improving estimation accuracy ; however, ongoing refresher tactics are crucial to maintain accuracy. Visual estimation can be improved with calibrated drapes and marking suction canisters before and after amniotic fluid is collected. Gravimetric measurement during PPH can accurately calculate blood loss. Gravimetric estimation involves weighing any absorbent material (ie, sheets, Chux pads, sponges) and subtracting dry weight from saturated weight, wherein the assumption is 1 g = 1 mL blood.
Recently, a digital blood loss estimation system has been investigated. The Triton System (Triton, Gauss Surgical, Los Altos, CA) is a tablet-based mobile platform that scans blood-containing sponges to estimate hemoglobin mass and track ongoing EBL. One small study found the Triton system significantly more accurate for estimating blood loss compared with a gravimetric technique. However, given the cost of implementation and potential need for a dedicated team member to take photos, this technology requires additional investigation, specifically in acute PPH, before widespread adoption can be recommended. Regardless of the technique or techniques used, blood loss should be constantly reassessed during and after acute PPH.
Maternal early warning systems
The Joint Commission and NPMS both strongly recommend that all obstetric units implement a maternal early warning system (MEWS) to promote early detection and assessment of parturients at risk for morbidity. MEWS are not specific to hemorrhage-related morbidity; however, the criteria in MEWS should identify patients suffering from ongoing hemorrhage. The MEWS criteria have evolved to reflect “normal” vital signs during pregnancy, and several publications detail proven systems. Shields and colleagues recently found a significant reduction in severe maternal morbidity, including severe hemorrhage when an MEWS was used.
Healthy parturients should have basically normal baseline laboratory values. A mild dilutional anemia, mild leukocytosis, and a mild dilution of PLT are common. Production of clotting factors increases with normal or slightly decreased international normalized ratio (INR; 0.9–1.0) and partial thromboplastin time (PTT), and elevated fibrinogen (400–600 mg/dL).
Timely transfusion during PPH often depends on rapid assessment of hematologic and coagulation parameters. The most common panel of tests done during a PPH include complete blood count, arterial blood gas (including electrolytes), prothrombin time (PT)/INR, PTT, and fibrinogen. Full point-of-care testing on labor and delivery (eg, iSTAT machine) is not widely used, although it can be helpful if available.
Thromboelastography (TEG) and rotational thromboelastometry (ROTEM) have become popular during PPH in many institutions. Use of this technology is long standing in cardiac and liver surgery. Both TEG and ROTEM can be used as a dynamic point-of-care device and usually provide at least initial results within 30 minutes. The main advantage is the complete coagulation picture provided rather than the segmented view from isolated coagulation tests.
Several investigations recently examined “normal” values for both TEG and ROTEM in parturients. It is difficult to use standard reference ranges developed in the nonpregnant population because of the aforementioned hematologic changes during pregnancy. de Lange and colleagues reported little change in values for ROTEM parameters from either active labor or precesarean to the postpartum period ( Table 2 ).
|Component||Healthy parturient (term, predelivery)||Healthy parturient (term, postpartum)||Nonpregnant|
|R time (min)||5.8–7.0||5.0–6.6||4–8|
|K time (min)||1.3–2.0||1.1–1.8||0–4|
|Ly30 (%)||0.2–1.6 a||0.6–0.7||0–8|
|CI||1.2||1.8||−3 to 3|
|CT (s)||45 (41–50)||45 (40–49)||31–66|
|CFT (s)||69 (62–81)||73 (63–86)||41–154|
|Alpha-angle (degrees)||77 (67–83)||76 (74–79)||63–83|
|MCF (mm)||71 (42–78)||71 (68–74)||42–78|
|ML (%)||7 (0–41)||8 (3–12)||0–44|
Macafee and colleagues ran TEG samples precesarean and postoperatively and reported reference ranges in healthy women. Compared with the midpoint of the nonpregnant reference range, parturients had larger maximum amplitude (MA), alpha-angle, and coagulation index and smaller percent lysis at 30 minutes (Ly30) TEG indices (see Table 2 ). Shreeve and colleagues reported reference ranges for TEG indices for women throughout pregnancy (first trimester, second trimester, third trimester, and women in labor) and found a significantly positive correlation between pregnancy and alpha-angle and MA, and a significantly negative correlation between pregnancy and K time and Ly30 (see Table 2 ).
Peripartum hemorrhage protocols
The need for all obstetric units to maintain a hemorrhage protocol was previously discussed. A staged approach to PPH management within a protocol is advised by NPMS starting with risk assessment and recognition of hemorrhage (stage 0) with increasing interventions as thresholds of blood loss are met. A sample stepwise protocol is described in later discussion and summarized in Table 3 .
|Stage of hemorrhage||Interventions and consideration|
Patients typically remain in the initial stage of a PPH protocol until they reach 1000 mL EBL. During this time, routine therapies are used. Stage 1 is reached once EBL reaches 1000 mL, at which point the patient requires increased assessment. A type and screen should be obtained if not already. Stage 2 of a PPH protocol is reached when the EBL is 1000 to 1500 mL with ongoing bleeding despite intervention performed during stage 1 but maintained hemodynamic stability. Activating a massive transfusion protocol (MTP), starting transfusion or other more aggressive pharmacologic therapy at this stage depends largely on the emerging clinical scenario.
Stage 3 of a PPH protocol typically indicates uncontrolled, ongoing hemorrhage. Clinicians should consider escalation of anesthetic and surgical management. The care team should also strongly consider activating the obstetric MTP and bring blood products to the bedside with a provision to “keep ahead” to replace what is transfused. An example of a first round of an MTP for PPH is 6 units PRBC, 6 units FFP, 6 units (pooled) PLT, and 10 units (pooled) cryoprecipitate. When activating an MTP, consideration should be made as to not waste resources, especially when blood bank supplies are limited.
Blood product transfusion
Transfusion practices vary widely, and many clinicians are hesitant to transfuse parturients until significantly anemic or symptomatic from blood loss because of concern for transfusion-related complications. The fact that parturients are often young, otherwise healthy patients whose vital signs remain stable until the onset of sudden hemodynamic instability also contributes to this reluctance. During CD, both obstetricians and anesthesiologists estimate blood loss and jointly initiate transfusion; however, in VDs, anesthesiologists are usually not present until blood loss is significant, so the obstetrician may unilaterally handle transfusion decision making. Similarly, in units without a robust multidisciplinary huddle or sign-out system, the anesthesia team may not know a patient required postpartum transfusion.
A recent survey to assess transfusion knowledge in members of the Society of Obstetricians and Gynaecologists of Canada revealed that although respondents performed well on questions related to RBC transfusion and anemia management, scores for non-RBC transfusion (PLT, cryoprecipitate, FFP, tranexamic acid [TXA]) were quite poor. These results emphasize the need for anesthesia team involvement in the care for parturients who may require transfusion.
RBC and FFP transfusion should be based on laboratory analysis or the clinical status of the patient with consideration of ongoing bleeding. In patients with massive acute bleeding and no available crossmatched blood, “trauma blood” or O-negative blood can be used. Transfusion of more than 1 to 2 units PRBC should follow an established protocol, either institutional or based on national guidelines. Many clinicians advocate mimicking MTP ratios used for trauma patients. In a survey of obstetric anesthesia unit directors regarding PPH protocols, 48% reported a 1:1 PRBC:FFP ratio and 35% indicated a 1:1:1 PRBC:FFP:PLT transfusion ratio. To the authors’ knowledge, there is no national recommendation regarding optimal transfusion ratio during PPH, and decisions should be made on a case-by-case basis.
PLT transfusion during PPH is somewhat controversial. Some investigators argue that PLT should be transfused early in a fixed ratio to RBC and FFP, similar to trauma patients. However, PPH-related coagulopathy patterns remain distinct from those of trauma patients, and the literature remains scarce on the best evidence-based strategy for PLT transfusion during PPH. In 2016, Jones and colleagues published a prospective, observational study of PLT counts and need for transfusion during moderate to severe PPH. Platelet transfusion occurred in 3.4% of all PPH patients and was more likely in patients with lower baseline PLT count, those with EBL greater than 2500 mL, and those with lower ROTEM Fibtem A5 values. Interestingly, 32 patients would have received PLT if a 4 RBC:4FFP:1PLT ratio was used; however, only 7 of those patients were actually transfused PLT and the ratio-based transfusion would have missed 5 patients who ultimately received a PLT transfusion, demonstrating PLT transfusion is rarely needed even in moderate to severe PPH without underlying thrombocytopenia, placental abruption, or PPH greater than 5000 mL.
Hypofibrinogenemia to less than 200 mg/dL is strongly predictive of severe PPH, and fibrinogen should be aggressively repleted during obstetric hemorrhage. FFP and cryoprecipitate can both provide repletion of fibrinogen, but require thawing and have volume and infection considerations. FFP, cryoprecipitate, and fibrinogen concentrate contain approximately 2.5, 15, and 20 g/L of fibrinogen, respectively. An average dose of fibrinogen concentrate (2 g) should increase the plasma fibrinogen level by approximately 100 mg/dL, depending on speed of ongoing blood loss and fibrinolysis. In a retrospective review of 99 patients who received a 3-g dose of fibrinogen concentrate during PPH treatment, fibrinogen level increased from an average of 70.5 mg/dL to 187 mg/dL (32.9 mg/dL/g fibrinogen concentrate). No adverse events were associated with fibrinogen concentrate use. A prospective study (FIB-PPH) using prophylactic fibrinogen concentrate (2-g dose) administration in patients with diagnosed PPH but normal fibrinogen did not find a difference in blood transfusion between the fibrinogen concentrate and placebo groups. However, parturients suffering rapid, massive hemorrhage could not be recruited, and because this population is likely most at risk for hypofibrinogenemia, there may have been a difference if those patients were included. Two prospective randomized controlled trials of fibrinogen concentrate are currently ongoing. The publication of these trials will provide important evidence either for or against the use of fibrinogen concentrate during PPH.
Transfusion thresholds in obstetric hemorrhage are often left to the discretion of the clinical care team because of the dynamic nature of the blood loss pattern. However, one group developed an algorithm for FFP, cryoprecipitate, and PLT transfusion based on TEG (with kaolin) results during PPH using pregnancy-specific reference ranges. A recent study in trauma patients similarly examined RapidTEG thresholds for transfusion in patients at risk for massive transfusion. Although the thresholds in parturients would likely vary slightly (eg, higher MA), the studies strengthen the argument for TEG-guided transfusion.
One institution has incorporated use of ROTEM in their obstetric MTP to guide transfusion of clotting factors, including fibrinogen concentrate.Initially, the protocol did not include fibrinogen concentrate; however, nearly 10% of the patients who underwent transfusion guided by the protocol developed transfusion-associated circulatory overload, compared with zero patients once fibrinogen concentrate was used.
As discussed earlier, RBC salvage can be used during CD, especially in cases of anticipated large volume blood loss. Two recent studies further explored CS in more routine CD. Milne and colleagues reported 884 cases during which intraoperative CS was used. Sufficient blood was collected for reinfusion in only 21% of the overall number of cases, and only 13% of the 748 routine CD patients received RBC reinfusion. Red cell reinfusion was far more likely for the patients undergoing cesarean hysterectomy and for patients with uncontrolled bleeding following delivery. The Cell SALVage in Obstetrics Trial, completed in 2016, is a prospective randomized controlled trial that examined roughly 1500 women assigned to CS during CD compared with women who received standard of care. Initial analysis indicates a slightly lower rate of allogenic blood transfusion but increased rate of hemorrhage in the CS group. Based on the current evidence, it seems CS in routine CDs is unwarranted.
CS in VDs has also been investigated. Teare and colleagues collected blood lost after VD from 50 women, washed the blood in a CS machine, and analyzed the postwash solution for cell counts and bacterial contamination. The investigators found that blood can be effectively collected by CS after VD, and the overall amount of hemolysis, washout of non-RBC components, and bacterial contamination was similar to the results seen in CD and CS quality control testing. An accompanying editorial points out that in some patients such as JWs, CS during VD could be considered a life-saving measure. At this time, this technology is likely far from being adopting during even moderate PPHs following VD.
Pharmacologic treatment options
Prophylactic administration of uterotonic agents immediately after delivery has decreased average postpartum bleeding worldwide. The most commonly administered drug in this setting, oxytocin, was the subject of a 2013Cochrane Review that included greater than 20 trials and 10,000 patients undergoing VD. Oxytocin at any dose resulted in an average decrease of greater than 500 mL EBL compared with placebo. Prophylactic oxytocin was also superior to ergot alkaloid administration, and the prophylactic combination conferred no added benefit. Oxytocin is the similarly preferred uterotonic during CD.
Administration practices for prophylactic oxytocin vary widely. IV bolus administration, continuous “wide-open” infusions, and controlled-rate infusions have all been described. Oxytocin has a remarkably low 90% effective dose (ED 90 ) of 0.35 IU and ED 100 of 0.5 IU bolus in nonlaboring patients undergoing elective CD. Study of controlled infusions has demonstrated similar potency, with one dose-finding study demonstrating an ED 90 of only 15 IU per hour (0.4 IU/min). Patients that require CD after exposure to oxytocin labor augmentation or induction demonstrate a relative oxytocin resistance, and the ED 90 is increased to approximately 3 IU bolus. When given as a controlled infusion, the oxytocin ED 90 for CD patients previously exposed to oxytocin was 44.2 IU/h compared with 16.2 IU/h in the nonlaboring group, again demonstrating that a higher oxytocin dose should be considered in this population.
An oxytocin protocol of 3-IU bolus dose was compared with 30-IU/500 cc bag infusion with no difference in side effects, maternal hemodynamics, uterine tone, or EBL. However, large bolus doses, especially greater than 5 to 10 IU, can cause marked tachycardia, hemodynamic instability, and ST depression, and even low doses can cause similar effects if given rapidly (<15–30 seconds).
Regardless of initial oxytocin strategy, after adequate tone is achieved, a maintenance infusion of oxytocin helps prevent reversion to an atonic state. Ideal dosing for maintenance of tone is unclear. A comparison of high- (15 IU/h) and low- (2.5 IU/h) dose maintenance oxytocin infusion revealed no difference in uterine tone or EBL .
Atony refractory to oxytocin typically requires treatment with other uterotonics. Ergot alkaloids (eg, methylergonovine), prostaglandin F2alpha (carboprost), and prostaglandin E (misoprostol) are all options to manage atony-related bleeding. Unfortunately, algorithms and dosing regimens vary widely across the world. The major deciding factor for which can and should be administered is typically the side-effect profile. A summary of secondary uterotonics is included in Table 4 .
|Uterotonic agent a||Dose/route||Frequency/maximum dose||Relative contraindications|
|Ergot alkaloids (methylergonovine, ergonovine)||0.2 mg IM or IU||Every 2–4 h||
|15-methyl-PGF2alpha (hemabate, carboprost)||0.25 mg IM||Every 15–90 min, 8 doses (2 mg)||
|Prostaglandin E1 (misoprostol, cytotec)||400–1000 μg rectal, IU, sublingual b||Once||None|
TXA stabilizes clot formation by preventing plasmin formation, thus stopping fibrin breakdown. It has been shown to decrease blood loss, need for transfusion, and mortality in several surgical patient populations. Several investigators have proposed TXA as a way to mitigate blood loss in ongoing PPH or as a prophylactic measure to prevent PPH. To follow initial studies that suggested TXA could reduce blood loss and the incidence of PPH when given prophylactically , Ducloy-Bouthors and colleagues randomized women with ongoing PPH following VD to receive TXA or placebo. The TXA group had lower blood loss from enrollment to the 6-hour point and less frequent RBC transfusion compared with the control group.
Since that time, interest in peripartum TXA has continued, and 2 meta-analyses were recently published. The investigators report the usual dose of TXA (1 g or 10 mg/kg) , decreased blood loss in the TXA group (approximately 150 mL for CD and 85 mL for VD), a reduced incidence of PPH in the TXA group , and no increased risk for thromboembolic side effects in the TXA group.
The World Maternal Antifibrinolytic (WOMAN) Trial, an international, multicenter, randomized, placebo-controlled trial of TXA during PPH was published in late April 2017. A total of 20,060 women with a diagnosis of PPH were randomly assigned to either 1 g TXA or placebo. Death due to PPH was significantly lower in the TXA group, especially if TXA was administered within 3 hours of birth; however, the composite endpoint of death from all causes or hysterectomy was not reduced in the treatment group. There was also no difference in the incidence of thromboembolic events between groups.
Given prior studies and the results of the WOMAN Trial supporting both benefit and safety of TXA in instances of PPH, it can be considered a likely safe treatment in hemorrhaging patients that could potentially decrease blood loss and death from PPH. However, the authors still cannot currently endorse widespread prophylactic use and suggest prophylactic TXA be reserved for patients at very high risk of PPH or those who refuse transfusion or present issues for blood availability.
Recombinant factor VII
rFVII has been reported in cases of severe PPH with coagulopathy. Evidence is largely based on case reports and series as well as several national registry reports. The success rate at controlling hemorrhage is greater than 80% with an average dose of 80 to 90 μg/kg. Thromboembolic complications have been reported in nearly all of the registry reports and in a recent randomized open-label trial of rFVII in refractory PPH. rFVII should be used with extreme caution in cases of life-threatening hemorrhage only after other factors contributing to uncontrollable hemorrhage have been optimized. Repeated doses of rFVII can be used, but if bleeding continues unabated after 2 doses, it is unlikely that additional doses of rFVII will be effective.
Prothrombin complex concentrate
Prothrombin complex concentrate (PCC) contains factors II (prothrombin), VII, IX, and X and is obtained from concentrated pooled plasma, thus creating a low-volume, high-activity treatment of factor deficiency or reversal of vitamin K antagonists. Although PCCs are used in massive hemorrhage from trauma, there is scant mention of use in treating PPH. One randomized trial is currently listed ( NCT01910675 , clinicaltrials.gov , Accessed March 12, 2017) that compares PCC and fibrinogen concentrate versus FFP (and fibrinogen if needed) in patients with PPH 2000 to 3000 mL. It is unclear if the study is still recruiting at this time. Currently, PCC is not recommended for use in obstetric hemorrhage.
Use of recombinant human soluble thrombomodulin (rhTM) in cases of DIC in obstetric patients was recently examined. rhTM is an anticoagulant agent designed to reduce excessive thrombin activation, thus modulating the clotting cascade. Compared with the control group (DIC diagnosis who did not receive rhTM), patients treated with rhTM had significantly higher PLT and significantly improved fibrinogen levels and PT/INR values. Patients who received rhTM also had lower PLT transfusion volume compared with the control group. rhTM is not currently available in the United States, and prospective studies are required to further evaluate the therapy.
There are no randomized trials of desmopressin (DDAVP) use as part of a PPH treatment protocol. Reports of DDAVP use in parturients center around prophylactic administration in patients with known hemophilia A or von Willibrand disease (VWD). A systematic review of 30 studies including 216 patients treated with DDAVP during pregnancy reported that only 5/172 patients who received DDAVP prophylaxis suffered blood loss adequate to diagnose hemorrhage. One case of water intoxication was reported, but otherwise, DDAVP was well tolerated. DDAVP should be given prophylactically if indicated to parturients diagnosed with applicable bleeding disorders (ie, VWD) and can also be considered during severe obstetric hemorrhage when clinicians think PLT dysfunction is contributing to ongoing bleeding. A recent Cochrane Review of DDAVP use in parturients reiterates the lack of high-quality evidence for or against routine use and that decisions about use in prevention or treatment of acute obstetric bleeding should be made on a case-by-case basis.
Maintenance of normothermia and normocalcemia and avoidance of marked acidosis (pH <7.2) are important for effective clot formation. Calcium repletion should be aggressive with ongoing PPH, especially in the setting of blood transfusion. An in vitro model indicates perturbations in calcium may lead to atony. Magnesium therapy impairs normal calcium-induced uterine contraction, and in life-threatening hemorrhage, IV calcium can be considered on a case-by-case basis.
Interventional radiology and arterial embolization
Refractory PPH requires invasive maneuvers. Previously, open uterine or internal iliac vessel ligation and hysterectomy were the only options, but because of collateralization, vessel ligation is prone to failure. Hysterectomy results in loss of reproductive ability and is associated with additional high-volume blood loss. IR offers a nonsurgical option for definitive or temporizing treatments in severe PPH cases, which may be particularly advantageous after VD when surgical access is not immediate. National guidelines vary worldwide regarding IR in the setting of PPH, but in general, transcatheter embolization is considered a viable option. The American College of, Obstetricians and Gynecologists recommends embolization only if bleeding is “persistent, stable, and nonexcessive”, implying that rapid ongoing hemorrhage with hemodynamic instability is not amenable to IR. Mode of delivery does not affect clinical outcome, need for repeat embolization, or surgical intervention. Pelvic artery embolization procedures are successful in greater than 90% of patients. Alternatively, selective supplementary embolization of the round ligament artery may also be successful.
The use of prophylactic transcatheter balloon occlusion of pelvic vasculature in high-risk CD is a topic of both interest and controversy. A recent review of 15 case reports and 5 small studies could not conclude whether the potential decrease in EBL outweighed associated maternal complication risks (ie, thromboembolic events, arterial dissection), which were reported between 6% and 16%. In 2012, a retrospective report of a single institution’s 21 years of experience with balloon occlusion concluded that blood loss benefit does indeed outweigh risks, and since that time, one small trial and 2 retrospective reviews have demonstrated similar benefit. However, a randomized controlled trial completed in 2015 failed to show difference in blood loss or packed red cell units transfused. It has been pointed out that both arms of this study demonstrated particularly high average blood loss, which may skew results ; however, at this time, use of occlusion balloons is recommended only for consideration on a case-by-case basis.
Obstetric hemorrhage remains a significant source of maternal morbidity and mortality. Although many conditions convey increased risk, uterine atony remains the major cause of PPH. Preparation for hemorrhage should be commonplace in all obstetric units and should include the components of the NPMS hemorrhage bundle. Individual patients should undergo risk assessment starting in the antenatal period and should be optimized if possible. Early and ongoing recognition of PPH is crucial with stepwise, protocol-based treatment, including blood transfusion and a multitude of pharmacologic options. Optimal care during obstetric hemorrhage remains an area of active research, and we look forward to further advances in preventing and treating PPH in the future.