Massive Hemorrhage: A Report from the Anesthesia Closed Claims Project

MASSIVE hemorrhage related to surgical procedures is a rare, but potentially lethal, event that may manifest in the operating room (OR), obstetric suite, intensive care unit (ICU), or postanesthesia care unit (PACU). Anticipating the potential for hemorrhage, and the corresponding requirement for fluid resuscitation and transfusion, is a core requirement of a successful anesthetic.¹  Hemorrhage is expected in certain cases, for example, major peripheral vascular procedures, cardiac and liver surgery, and radical cancer resections. Hemorrhage is less routine, but anticipated, in trauma cases, emergency general surgery procedures, major orthopedics, and obstetrics. Failure by anesthesiologists and surgeons to recognize and appropriately treat hemorrhage in a timely manner may result in significant morbidity and mortality.

Advances in resuscitation have improved patient survival during the last several decades, as techniques developed in combat casualty care have been adopted in civilian hospitals and elective surgery.²  These techniques include algorithms and scoring systems to anticipate the onset of hemorrhagic shock,^3,4  the surgical concept of “damage control,”^3,5,6  the use of deliberate hypotensive resuscitation,^4–7  early replacement of coagulation factors,^2,4–9  and improved laboratory testing and hemodynamic monitoring.¹⁰  Technological advances include better and more rapid means of obtaining large-bore intravenous access,⁴  improved fluid warmers and rapid infusion systems,⁶  and the ongoing evolution of surgical instruments such as staplers and cautery devices. Despite these advances, however, massive hemorrhage remains a life-threatening complication of surgery and obstetrics.

To obtain a perspective on factors associated with patient injury from massive hemorrhage during anesthesia care in the United States, we examined closed malpractice claims from the past 2 decades, comparing claims for massive hemorrhage to other surgical and obstetric malpractice claims. Our hypothesis was that a detailed review of these most serious events could reveal common factors associated with malpractice claims for massive hemorrhage. We also examined the actions and reactions of the providers involved in massive hemorrhage claims in an attempt to identify recurring patterns of patient harm that could inform anesthesia education and system redesign.

The Anesthesia Closed Claims Project database is a structured collection of closed anesthesia malpractice claims described in detail elsewhere.¹¹  In brief, on-site anesthesiologist-reviewers abstracted data from closed anesthesia malpractice claims onto detailed data collection instruments at participating professional liability companies across the United States. The panel of 22 companies (at the time of this study) insured over one third of practicing anesthesiologists in the United States. Information was collected from medical records, consultant evaluations, expert witness reports, claims manager summaries, and legal summaries. Data collected included patient demographics, type of surgery, details regarding anesthesia care, patient outcomes, and legal outcomes. The on-site reviewer evaluated the standard of care, outcome, severity of injury, and cause of injury (i.e., damaging event). The severity of injury score used the National Association of Insurance Commissioners’ 10-point scale, which ranges from 0 (no apparent injury) to 9 (death).¹²  This scale was collapsed into three categories for this analysis: death (score = 9), permanent disabling injuries (score = 6 to 8), and temporary or minor injuries (score = 0 to 5).

Appropriateness of anesthesia care was assessed by the on-site reviewer as appropriate (based on reasonable or prudent practice at the time of the event), substandard, or impossible to judge. The reliability of these evaluations has been judged acceptable.¹³  The on-site reviewer also summarized the claim in a brief narrative, including the sequence of events and causes of injury. The Closed Claims Project Investigator Committee reviewed the claims, and any disagreements in assessments were resolved by Committee members.

For this study, we used the Anesthesia Closed Claims Project database of 9,799 claims. Inclusion criteria were claims for injuries occurring with surgical or obstetric anesthesia care between 1995 and 2011. Of the 3,211 surgical or obstetric claims from this time period, 141 had a primary damaging event of hemorrhage (hemorrhage claims) and 3,070 had some other primary damaging event (other claims). Background information on the demographics of anesthesia care in the United States was obtained from the National Anesthesia Clinical Outcomes Registry (NACOR) Participant User File of January 2014, by direct query.¹⁴  This information, drawn from approximately 20% of all anesthesia practices in the nation from 2010 to 2013, was used to provide context for case numbers and patient demographics in the Closed Claims Project database.

Risk factors for hemorrhage and coagulopathy and the location(s) within the treatment facility where hemorrhage was apparent were abstracted from the claim narratives by two of the authors (R.P.D. and L.A.L.). Obstetric risk factors for hemorrhage were placenta accreta, increta, or percreta; retained placenta; placenta previa; placenta abruption; uterine atony; and uterine rupture. Obstetric risk factors for coagulopathy were amniotic fluid embolus, intrauterine fetal demise of a week or more, and placental abruption. Surgical risk factors for hemorrhage were spine surgery, major vascular surgery, cardiac surgery, liver surgery, large tumor surgery, robotic/laparoscopic/minimally invasive surgery, and surgery after a major trauma event. Surgical risk factors for coagulopathy were preexisting use of anticoagulants or platelet inhibitors, liver disease, and preoperative increased partial thromboplastin time. In addition to these predefined risk factors, other obstetric and surgical risk factors for hemorrhage and coagulopathy were abstracted when present in the claim narratives to assure completeness. The locations within the treatment facility where the hemorrhage was apparent included the OR, the PACU, the ICU, the ward or floor, or other locations.

Two authors (R.P.D. and L.A.L.) judged the following factors based on claim narratives with a third author (K.B.D.) serving as tie breaker for disagreements: whether the hemorrhage was the result of an unexpected organ or vessel injury, whether or not the hemorrhage caused an immediate change in vital signs, and the timeliness of the diagnosis of hemorrhage, transfusion of blood products, or return to the OR if bleeding became apparent postoperatively. These were global assessments of the team (surgeon, anesthesiologist, ICU, etc.), not restricted to judgments of anesthesia care. The contribution of anesthesia and surgery to the patient’s injury was judged as no contribution, some contribution, or totally responsible for injury. Communication issues contributing to the adverse outcome were identified and grouped into thematic categories.

Interrater reliability for judgments of factors in hemorrhage claims was measured using kappa (κ) scores calculated on the initial two author judgments before tie-breaking by the third author. All payments made to the plaintiff were extracted from the database and adjusted to 2012 dollar amounts with the Consumer Price Index.^* Median and interquartile range were reported for payments because they were not normally distributed. Claims with no payment were excluded from calculation of median and interquartile range. Demographics and legal outcomes of hemorrhage claims were compared with those in other surgical/obstetric anesthesia claims using Fisher exact test, Pearson chi-square, t test for equality of means, or Mann–Whitney U test with P value less than 0.05 as the criterion for statistical significance and two-tailed tests. All statistical analysis used SPSS 19 for Windows (IBM Corporation, Armonk, NY). Comparison of Closed Claims cases and NACOR cases did not incorporate a formal statistical analysis because comparison involving 13 million records risks generating statistical significance when the actual clinical differences are trivial. The comparison is therefore presented as qualitative rather than formally quantitative.

Hemorrhage occurred in 141 (4%) of 3,211 surgical and obstetric anesthesia claims. Hemorrhage was the primary damaging event in 9.6% of claims filed on behalf of obstetric patients versus 3.5% of claims filed on behalf of patients undergoing other surgical procedures (P < 0.001). The 43 obstetric patients in hemorrhage claims underwent cesarean delivery (n = 23), vaginal delivery (n = 16), dilation and curettage for fetal demise (n = 3), and intrautero laparoscopic photocoagulation (n = 1). Thoracic or lumbar spine procedures (24% of hemorrhage claims) and cesarean delivery (16% of hemorrhage claims) occurred more frequently in hemorrhage claims than in other claims (P < 0.001 and P = 0.002, respectively; table 1). Obstetric cases account for 8% of all anesthetics in NACOR, and spine cases account for 3%; both of these proportions are lower than in the Anesthesia Closed Claims Project registry overall and much lower than in hemorrhage claims. Although hemorrhage claims were more likely to involve emergency procedures (29%) compared with other claims (18%, P = 0.003; table 1), they were not more likely to occur in association with a traumatic injury (4%; table 1). Patient demographics (age, sex, American Society of Anesthesiologists physical status) were similar between other (nonhemorrhage) claims from the Anesthesia Closed Claims Project and overall national case demographics drawn from NACOR (table 1).

Claims associated with hemorrhage had a greater severity of injury than other claims, with over three fourths resulting in death (P < 0.001; table 2). The anesthesia care was more often assessed as less than appropriate and a payment made more often to the plaintiff in hemorrhage claims compared with other claims (P < 0.001; table 2). Payments were greater for hemorrhage claims (median $607,750) than for other obstetric and surgical claims ($276,000, P < 0.001; table 2).

Most (78%) of the hemorrhage cases had risk factors for hemorrhage and 20% had independent risk factors for coagulopathy. The risk factors varied between obstetric and surgical procedures (table 3). Among the 43 obstetric claims, 74% presented with at least one risk factor for hemorrhage, with the most common risk factors including placenta accreta, increta, or percreta (n = 13); retained placenta (n = 10); and uterine atony (n = 7; table 3). More than one fourth of obstetric patients (28%) had risk factors for coagulopathy, most commonly amniotic fluid embolus (n = 5), placental abruption (n = 3), and intrauterine fetal demise (n = 3; table 3).

Almost 80% of the 98 surgical claims had at least one risk factor for hemorrhage, including thoracic or lumbar spine surgery (n = 35), robotic/laparoscopic/minimally invasive surgery (n = 18), or major vascular or cardiac surgery (n = 16; table 3). Most (77%, n = 27) of the spine procedures were fusions and half the procedures (n = 17) were described as “multi-level.” Sixteen percent (n = 16) of the surgical claims had risk factors for coagulopathy, primarily preexisting use of anticoagulants or platelet inhibitors (n = 9; table 3).

In most claims, hemorrhage was apparent in the OR (72%) or in the PACU (21%; table 4). In about one third of all of the claims, the hemorrhage was the direct result of an unexpected organ or vessel injury (κ = 0.903). In about one fourth of all the claims, the hemorrhagic event caused an immediate change in vital signs (κ = 0.763).

A timely diagnosis of hemorrhage was judged to have occurred in only 31% of the claims (κ = 0.658; tables 4 and 5). Transfusion did not occur in a timely manner in most of the claims (86%, κ = 0.690; tables 4 and 5). A timely return to the OR occurred in only 11% of the 52 claims where the hemorrhage was first recognized outside of the OR (κ = 0.945; table 4). In 88% of claims, the anesthesiologist contributed to some degree to the injury experienced by the patient (κ = 0.515). In 99% of claims, the surgeon contributed to the patient’s injury (κ = 0.472).

For 12 of the 18 robotic, laparoscopic, or minimally invasive procedures, hemorrhage was the result of unexpected organ (e.g., liver) or vessel (e.g., cystic artery, superior mesenteric artery, portal vein, iliac artery vein, aorta) injury. Bleeding was apparent in the OR in two thirds (n = 12 of 18), in the PACU (n = 4), in the ICU (n = 3), or on the ward (n = 4, some had multiple locations). A timely diagnosis of hemorrhage occurred in only 5 of the 18 cases.

Sixty percent of the hemorrhage claims (n = 81) had at least one communication breakdown occur (table 6). Over half of the communication problems occurred between the anesthesiologist and the surgeon or obstetrician. The other communication problems were more systematic issues, for example, arising during supervision of cases or with follow-up communication issues with PACU, ICU, or floor personnel.

This review of the Closed Claims Project database identified 141 claims resulting from perioperative hemorrhage. Although claims associated with massive hemorrhage were a small proportion of all surgical and obstetric claims (4%), the high percentage resulting in death or permanent injury raises concern. Analysis of hemorrhage claims reveals several lessons about patient safety (table 7).

The most obvious lesson is the type of cases that resulted in malpractice claims. Although hemorrhage is common in certain major surgeries (e.g., liver transplantations, open-heart procedures) and trauma care,¹⁵  these cases were underrepresented in the closed hemorrhage malpractice claims. Hemorrhage claims were most common in obstetrics (30%), thoracic, or lumbar spine surgery (24%). Hemorrhage also occurred in low-risk laparoscopic, robotic, or minimally invasive procedures. Data from the Manufacturer and User Facility Device Experience database corroborates these findings with deaths related to major vascular lacerations during robotic surgical cases. A query limited to the brand name “Da Vinci” or manufacturer name “Intuitive” for 2006 to 2013 identified 135 deaths with 30 (22%) of these associated with intraoperative hemorrhage.^†

Failure to immediately recognize ongoing hemorrhage was a common problem in hemorrhage claims. The careful provider must recognize this risk and consider the possibility of hemorrhage whenever a patient exhibits nonroutine clinical signs or symptoms. Warning signs of hemorrhage that might have been recognized earlier were identified in many of the hemorrhage claims. Although hindsight is 20:20 and not all injuries are salvageable, failure to recognize and respond to life-threatening hemorrhage was one of the more common findings in this collection of closed claims.

Delays in recognition and communication of a developing emergency contributed to many of adverse outcomes in the hemorrhage claims. The importance of effective and timely communication between the anesthesia team and the surgeon or obstetrician cannot be overemphasized. Anesthesiologists can be instrumental in calling for reexploring a patient for bleeding. As in disasters in aviation, nuclear power, and other high-risk disciplines, protocols that emphasize interpersonal communications during crisis situations can have a positive impact. Training programs such as Team Strategies and Tools to Enhance Performance and Patient Safety^‡ provide multidisciplinary training to improve teamwork skills in healthcare settings with a focus on effective communication.¹⁶ 

Timely transfusion did not occur in a majority of these claims. Once life-threatening hemorrhage is recognized, it is important to have an organized response. Both military and civilian trauma education emphasizes team-based care supported by standard practices and well-rehearsed institutional protocols that begin with the ability to call for help, such as a system to notify additional anesthesia personnel and experienced gynecologists, trauma, or vascular surgeons.^4–6  Ensuring an adequate supply, easy ordering, and rapid delivery of blood products are critical components of these protocols. Availability of such resources should be factored into the risk–benefit equation for any surgical plan.

Obstetric claims represented a third of the hemorrhage cases in this study. Throughout the world, postpartum hemorrhage is the most common cause of maternal mortality. In the United States, hemorrhage caused 9.7% of pregnancy-related deaths between 1998 and 2005.¹⁷  Although some obstetric hemorrhage events are unanticipated, identifying patients at risk, such as those with abnormal placentation or with risk factors for uterine atony, can improve earlier recognition and treatment of hemorrhage.^18,19  Preparation for an anticipated postpartum hemorrhage includes development of a multidisciplinary care plan. Consultation between the specialties (including hematology and interventional radiology) regarding patient management can improve communication and reduce maternal mortality.²⁰  Several initiatives are underway in the United States to improve maternal safety.²¹ 

In some of the claims reviewed, there were no warning signs of impending hemorrhage in a surgery or delivery that should have been routine. It behooves practitioners in resource-limited settings to have protocols in place for dealing with massive hemorrhage. The most important step might be the drill for rapid consultation outside the facility, with rapid patient transfer to a referral hospital or trauma center. This is one of the core lessons of the Advanced Trauma Life Support curriculum of the American College of Surgeons²²  and might be beneficial in nontraumatic hemorrhage as well. Referral of obstetrics patients with high hemorrhage risk to a tertiary care center for delivery should also be considered for disorders diagnosed on routine ultrasound (e.g., placenta previa; placenta accrete, increta, percreta). These cases represented about a third of the obstetric claims for hemorrhage, emphasizing that hemorrhage cannot always be predicted before labor and delivery.

Training of OR and obstetric teams in crew resource management with ongoing practice for hemorrhage emergencies has the potential to improve team coordination and mitigate poor outcomes.^23,24  Training can take various forms. One approach is to set up the simulation in the actual OR with no prior notice to the staff.²⁵  The Mobile Obstetrics Emergencies Simulator system, which has been aligned with Team Strategies and Tools to Enhance Performance and Patient Safety, demonstrates how simulation training and clinical drills performed in a facility’s obstetric unit can improve team performance. Issues specific to each unit can be addressed in the setting in which they occur.²⁶ 

Another key component of communications is the massive transfusion protocol (MTP), now common at many university hospitals, obstetric units, and level 1 trauma centers.^3,27–29  This prearranged order set for the blood bank can be initiated at the point of care with a single phone call. The MTP provides for rapid delivery of blood products to the bedside, beginning with uncrossmatched and emergency-release products if necessary, and plans for continued delivery as the resuscitation progresses. Activation of the MTP is one component of the “crisis checklist” that has been recommended for dealing with intraoperative emergencies.³⁰  A third of the obstetric hemorrhage cases followed vaginal delivery. Although MTPs specific for obstetric units have been described,²⁸  there are no data to suggest that a specific obstetric protocol is better than a standardized institutional MTP. Systems factors may need to be optimized to ensure that the MTP functions well in the obstetric environment.

Massive transfusion protocols are designed to facilitate the early replacement of clotting factors if coagulopathy is present or a high risk. A typical MTP calls for the blood bank to send 6 units of erythrocytes, 4 units of plasma, and 1 apheresis pack of platelets to the OR as rapidly as possible, followed with similar “transfusion packs” at regular intervals until the crisis is resolved.^3,27  The optimal ratio of erythrocytes, plasma, and platelets is controversial,^31,32  but current recommendations begin with empiric replacement of coagulation factors until hemorrhage has slowed sufficiently to allow for precise assessment of clotting function.⁶  Trauma patients, presenting with tissue injury and shock, can become coagulopathic very early after injury.⁸  In pregnancy, the placenta is an important source of immune activation that can trigger early, massive coagulopathy.³³  Elective surgical cases with massive hemorrhage may also deteriorate toward the final common pathway of death from hemorrhage: uncorrectable coagulopathy, persistent acidosis, circulatory exhaustion, and eventual cardiac arrest. Frequent laboratory testing and rapid return of results are essential during the dynamic course of hemorrhagic shock.⁶ 

The limitations of closed claims analysis have been previously described, including selection bias, nonrandom retrospective data collection, outcome bias, and possible geographic imbalance in data collection.³⁴  The data are limited to information gathered by insurance companies for claims resolution, and the database lacks a denominator of anesthetics for estimating risk.³⁴  The NACOR cases used to provide a general context come from a more recently compiled registry (2010 to 2013) and may themselves be subject to selection biases. Comparisons should be interpreted with caution.

Important data on factors in the hemorrhage claims, such as timeliness of diagnosis and treatment, was based on secondary analysis of claim narratives provided by the project on-site reviewer rather than systematic primary data abstraction from the insurance company claim files. Reliability of reviewer judgments in these hemorrhage claims was good to excellent on most items. Inclusion of a third reviewer in such assessments provides improved reliability. The lowest reliability was observed on the assessment of the anesthesiologist and surgeon contribution to the patient’s injury. The κ value on these assessments reflects the high prevalence of reviewer judgments that these providers contributed to the injury, leaving little room for agreement beyond chance. Our measure of interrater reliability (kappa) calculates agreement beyond that which would be expected by chance, and results in low scores in situations of skewed data distributions such as observed on these judgments.³⁵  Therefore, the seemingly low reliability on these items, which was still in the acceptable range, reflects the inherent traits of the statistical methodology.

In summary, the most common types of procedures involved in claims associated with hemorrhage were obstetrics, thoracic or lumbar spine surgery, and robotic/laparoscopic surgery. Lack of communication and absence of organized responses to massive hemorrhage were common and associated with inadequate preparation, delays in diagnosis, and/or effective treatment. Evolving, evidence-based treatment for uncontrolled hemorrhage may lead to a reduction in the frequency and severity of these events.

The authors thank the closed claims reviewers from the American Society of Anesthesiologists (Schaumburg, Illinois) and appreciate the participation of the following liability insurance companies who have given permission to be acknowledged: Anesthesia Service Medical Group, Inc., San Diego, California; Armed Forces Institute of Pathology, Silver Spring, Maryland; COPIC Insurance Company, Denver, Colorado; Department of Veterans Affairs, Washington, D.C.; ISMIE Mutual Insurance Company, Chicago, Illinois; MAG Mutual Insurance Company, Atlanta, Georgia; Medical Liability Mutual Insurance Company, New York, New York; Midwest Medical Insurance Company, Minneapolis, Minnesota; Mutual Insurance Company of Arizona, Phoenix, Arizona; NORCAL Mutual Insurance Company, San Francisco, California; Pennsylvania Medical Society Liability Insurance Company, Mechanicsburg, Pennsylvania; Physicians Insurance A Mutual Company, Seattle, Washington; Preferred Physicians Medical Risk Retention Group, Shawnee Mission, Kansas; Medical Professional Mutual Insurance Company, Boston, Massachusetts; Risk Management Foundation, Cambridge, Massachusetts; State Volunteer Mutual Insurance Company, Brentwood, Tennessee; The Doctors’ Company, Napa, California; The University of Texas System, Austin, Texas; and Utah Medical Insurance Association, Salt Lake City, Utah.

Supported in part by the American Society of Anesthesiologists (ASA) and the Anesthesia Quality Institute (AQI), Schaumburg, Illinois. All opinions expressed are those of the authors and do not reflect the policy of the ASA or AQI.

The authors declare no competing interests.