Amir K. Jaffer, MD, MBA

Of the 36 million US hospitalizations each year, 22% are surgical.¹ Although less frequent than medical hospitalizations, surgical hospitalizations are more than twice as costly.² Additionally, surgical hospitalizations are on average longer than medical hospitalizations.² Given the increased scrutiny on cost and efficiency of care, attention has turned to optimizing perioperative care. Hospitalists are well positioned to provide specific expertise in the complex interdisciplinary medical management of surgical patients.

In recent decades, multiple models of hospitalist involvement in perioperative care have evolved across the United States.^3-19 To consolidate knowledge and experience and to develop a framework for providing the best care for surgical patients, the Society of Hospital Medicine organized the Perioperative Care Work Group in 2015. This framework was designed for interdisciplinary collaboration in building and strengthening perioperative care programs.

The Society of Hospital Medicine recognized hospital medicine programs’ need for guidance in developing collaborative care in perioperative medicine and appointed the Perioperative Care Work Group in May 2015. Work group members are perioperative medicine experts from US medical centers. They have extensive knowledge of the literature as well as administrative and clinical experience in a variety of perioperative care models.

Topic Development. Initial work was focused on reviewing and discussing multiple models of perioperative care and exploring the roles that hospital medicine physicians have within these models. Useful information was summarized to guide hospitals and physicians in designing, implementing, and expanding patient-centric perioperative medicine services with a focus on preoperative and postoperative care. A final document was created; it outlines system-level issues in perioperative care, organized by perioperative phases.

Initial Framework. Group members submitted written descriptions of key issues in each of 4 phases: (1) preoperative, (2) day of surgery, (3) postoperative inpatient, and (4) postdischarge. These descriptions were merged and reviewed by the content experts. Editing and discussion from the entire group were incorporated into the final matrix, which highlighted (1) perioperative phase definitions, (2) requirements for patients to move to next phase, (3) elements of care coordination typically provided by surgery, anesthesiology, and medicine disciplines, (4) concerns and risks particular to each phase, (5) unique considerations for each phase, (6) suggested metrics of success, and (7) key questions for determining the effectiveness of perioperative care in an institution. All members provided final evaluation and editing.

Final Approval. The Perioperative Care Matrix for Inpatient Surgeries (PCMIS) was presented to the board of the Society of Hospital Medicine in fall 2015 and was approved for use in centering and directing discussions regarding perioperative care.

Models of Care. The Perioperative Care Work Group surveyed examples of hospitalist engagement in perioperative care and synthesized these into synopses of existing models of care for the preoperative, day-of-surgery, postoperative-inpatient, and postdischarge phases.

Defining Key Concepts and Issues

Hospitalists have participated in a variety of perioperative roles for more than a decade. Roles include performing in-depth preoperative assessments, providing oversight to presurgical advanced practice provider assessments, providing inpatient comanagement and consultation both before and after surgery, and providing postdischarge follow-up within the surgical period for medical comorbidities.

Although a comprehensive look at the entire perioperative period is important, 4 specific phases were defined to guide this work (Figure). The phases identified were based on time relative to surgery, with unique considerations as to the overall perioperative period. Concerns and potential risks specific to each phase were considered (Table 1).

Models of Care

Multiple examples of hospitalist involvement were collected to inform the program development guidelines. The specifics noted among the reviewed practice models are described here.

Preoperative. In some centers, all patients scheduled for surgery are required to undergo evaluation at the institution’s preoperative clinic. At most others, referral to the preoperative clinic is at the discretion of the surgical specialists, who have been informed of the clinic’s available resources. Factors determining whether a patient has an in-person clinic visit, undergoes a telephone-based medical evaluation, or has a referral deferred to the primary care physician (PCP) include patient complexity and surgery-specific risk. Patients who have major medical comorbidities (eg, chronic lung or heart disease) or are undergoing higher risk procedures (eg, those lasting >1 hour, laparotomy) most often undergo a formal clinic evaluation. Often, even for a patient whose preoperative evaluation is completed by a PCP, the preoperative nursing staff will call before surgery to provide instructions and to confirm that preoperative planning is complete. Confirmation includes ensuring that the surgery consent and preoperative history and physical examination documents are in the medical record, and that all recommended tests have been performed. If deficiencies are found, surgical and preoperative clinic staff are notified.

The United States is focused on sensible, high-value care. Perioperative care is burgeoning with opportunities for improvement, including reducing avoidable complications, developing systems for early recognition and treatment of complications, and streamlining processes to shorten length of stay and improve patient experience. The PCMIS provides the needed platform to catalyze detailed collaborative work between disciplines engaged in perioperative care.

As average age and level of medical comorbidity increase among surgical patients, hospitalists will increasingly be called on to assist in perioperative care. Hospitalists have long been involved in caring for medically complex surgical patients, through comanagement, consultation, and preoperative evaluations. As a provider group, hospitalists have comprehensive skills in quality and systems improvement, and in program development across hospital systems nationwide. Hospitalists have demonstrated their value by focusing on improving patient outcomes and enhancing patient engagement and experiences. Additionally, the perioperative period is fraught with multiple and complicated handoffs, a problem area for which hospital medicine has pioneered solutions and developed unique expertise. Hospital medicine is well prepared to provide skilled and proven leadership in the timely development, improvement, and expansion of perioperative care for this increasingly older and chronically ill population.

Hospitalists are established in multiple perioperative roles for high-risk surgical patients and have the opportunity to expand optimal patient-centric perioperative care systems working in close concert with surgeons and anesthesiologists. The basics of developing these systems include (1) assessing risk for medical complications, (2) planning for perioperative care, (3) developing programs aimed at risk reduction for preventable complications and early identification and intervention for unavoidable complications, and (4) guiding quality improvement efforts, including planning for frequent handoffs and transitions.

As a key partner in developing comprehensive programs in perioperative care, hospital medicine will continue to shape the future of hospital care for all patients. The PCMIS, as developed with support from the Society of Hospital Medicine, will aid efforts to achieve the best perioperative care models for our surgical patients.

Disclosures

Financial activities outside the submitted work: Drs. Pfeifer and Jaffer report payment for development of educational presentations; Dr. Grant reports payment for expert testimony pertaining to hospital medicine; Drs. Grant and Jaffer report royalties from publishing; Drs. Thompson, Pfiefer, Grant, Slawski, and Jaffer report travel expenses for speaking and serving on national committees; and Drs. Slawski and Jaffer serve on the board of the Society of Perioperative Assessment and Quality Improvement. The other authors have nothing to report.

Of the 36 million US hospitalizations each year, 22% are surgical.¹ Although less frequent than medical hospitalizations, surgical hospitalizations are more than twice as costly.² Additionally, surgical hospitalizations are on average longer than medical hospitalizations.² Given the increased scrutiny on cost and efficiency of care, attention has turned to optimizing perioperative care. Hospitalists are well positioned to provide specific expertise in the complex interdisciplinary medical management of surgical patients.

In recent decades, multiple models of hospitalist involvement in perioperative care have evolved across the United States.^3-19 To consolidate knowledge and experience and to develop a framework for providing the best care for surgical patients, the Society of Hospital Medicine organized the Perioperative Care Work Group in 2015. This framework was designed for interdisciplinary collaboration in building and strengthening perioperative care programs.

The Society of Hospital Medicine recognized hospital medicine programs’ need for guidance in developing collaborative care in perioperative medicine and appointed the Perioperative Care Work Group in May 2015. Work group members are perioperative medicine experts from US medical centers. They have extensive knowledge of the literature as well as administrative and clinical experience in a variety of perioperative care models.

Topic Development. Initial work was focused on reviewing and discussing multiple models of perioperative care and exploring the roles that hospital medicine physicians have within these models. Useful information was summarized to guide hospitals and physicians in designing, implementing, and expanding patient-centric perioperative medicine services with a focus on preoperative and postoperative care. A final document was created; it outlines system-level issues in perioperative care, organized by perioperative phases.

Initial Framework. Group members submitted written descriptions of key issues in each of 4 phases: (1) preoperative, (2) day of surgery, (3) postoperative inpatient, and (4) postdischarge. These descriptions were merged and reviewed by the content experts. Editing and discussion from the entire group were incorporated into the final matrix, which highlighted (1) perioperative phase definitions, (2) requirements for patients to move to next phase, (3) elements of care coordination typically provided by surgery, anesthesiology, and medicine disciplines, (4) concerns and risks particular to each phase, (5) unique considerations for each phase, (6) suggested metrics of success, and (7) key questions for determining the effectiveness of perioperative care in an institution. All members provided final evaluation and editing.

Final Approval. The Perioperative Care Matrix for Inpatient Surgeries (PCMIS) was presented to the board of the Society of Hospital Medicine in fall 2015 and was approved for use in centering and directing discussions regarding perioperative care.

Models of Care. The Perioperative Care Work Group surveyed examples of hospitalist engagement in perioperative care and synthesized these into synopses of existing models of care for the preoperative, day-of-surgery, postoperative-inpatient, and postdischarge phases.

Defining Key Concepts and Issues

Hospitalists have participated in a variety of perioperative roles for more than a decade. Roles include performing in-depth preoperative assessments, providing oversight to presurgical advanced practice provider assessments, providing inpatient comanagement and consultation both before and after surgery, and providing postdischarge follow-up within the surgical period for medical comorbidities.

Although a comprehensive look at the entire perioperative period is important, 4 specific phases were defined to guide this work (Figure). The phases identified were based on time relative to surgery, with unique considerations as to the overall perioperative period. Concerns and potential risks specific to each phase were considered (Table 1).

Models of Care

Multiple examples of hospitalist involvement were collected to inform the program development guidelines. The specifics noted among the reviewed practice models are described here.

Preoperative. In some centers, all patients scheduled for surgery are required to undergo evaluation at the institution’s preoperative clinic. At most others, referral to the preoperative clinic is at the discretion of the surgical specialists, who have been informed of the clinic’s available resources. Factors determining whether a patient has an in-person clinic visit, undergoes a telephone-based medical evaluation, or has a referral deferred to the primary care physician (PCP) include patient complexity and surgery-specific risk. Patients who have major medical comorbidities (eg, chronic lung or heart disease) or are undergoing higher risk procedures (eg, those lasting >1 hour, laparotomy) most often undergo a formal clinic evaluation. Often, even for a patient whose preoperative evaluation is completed by a PCP, the preoperative nursing staff will call before surgery to provide instructions and to confirm that preoperative planning is complete. Confirmation includes ensuring that the surgery consent and preoperative history and physical examination documents are in the medical record, and that all recommended tests have been performed. If deficiencies are found, surgical and preoperative clinic staff are notified.

The United States is focused on sensible, high-value care. Perioperative care is burgeoning with opportunities for improvement, including reducing avoidable complications, developing systems for early recognition and treatment of complications, and streamlining processes to shorten length of stay and improve patient experience. The PCMIS provides the needed platform to catalyze detailed collaborative work between disciplines engaged in perioperative care.

As average age and level of medical comorbidity increase among surgical patients, hospitalists will increasingly be called on to assist in perioperative care. Hospitalists have long been involved in caring for medically complex surgical patients, through comanagement, consultation, and preoperative evaluations. As a provider group, hospitalists have comprehensive skills in quality and systems improvement, and in program development across hospital systems nationwide. Hospitalists have demonstrated their value by focusing on improving patient outcomes and enhancing patient engagement and experiences. Additionally, the perioperative period is fraught with multiple and complicated handoffs, a problem area for which hospital medicine has pioneered solutions and developed unique expertise. Hospital medicine is well prepared to provide skilled and proven leadership in the timely development, improvement, and expansion of perioperative care for this increasingly older and chronically ill population.

Hospitalists are established in multiple perioperative roles for high-risk surgical patients and have the opportunity to expand optimal patient-centric perioperative care systems working in close concert with surgeons and anesthesiologists. The basics of developing these systems include (1) assessing risk for medical complications, (2) planning for perioperative care, (3) developing programs aimed at risk reduction for preventable complications and early identification and intervention for unavoidable complications, and (4) guiding quality improvement efforts, including planning for frequent handoffs and transitions.

As a key partner in developing comprehensive programs in perioperative care, hospital medicine will continue to shape the future of hospital care for all patients. The PCMIS, as developed with support from the Society of Hospital Medicine, will aid efforts to achieve the best perioperative care models for our surgical patients.

Disclosures

Financial activities outside the submitted work: Drs. Pfeifer and Jaffer report payment for development of educational presentations; Dr. Grant reports payment for expert testimony pertaining to hospital medicine; Drs. Grant and Jaffer report royalties from publishing; Drs. Thompson, Pfiefer, Grant, Slawski, and Jaffer report travel expenses for speaking and serving on national committees; and Drs. Slawski and Jaffer serve on the board of the Society of Perioperative Assessment and Quality Improvement. The other authors have nothing to report.

Of the 36 million US hospitalizations each year, 22% are surgical.¹ Although less frequent than medical hospitalizations, surgical hospitalizations are more than twice as costly.² Additionally, surgical hospitalizations are on average longer than medical hospitalizations.² Given the increased scrutiny on cost and efficiency of care, attention has turned to optimizing perioperative care. Hospitalists are well positioned to provide specific expertise in the complex interdisciplinary medical management of surgical patients.

In recent decades, multiple models of hospitalist involvement in perioperative care have evolved across the United States.^3-19 To consolidate knowledge and experience and to develop a framework for providing the best care for surgical patients, the Society of Hospital Medicine organized the Perioperative Care Work Group in 2015. This framework was designed for interdisciplinary collaboration in building and strengthening perioperative care programs.

The Society of Hospital Medicine recognized hospital medicine programs’ need for guidance in developing collaborative care in perioperative medicine and appointed the Perioperative Care Work Group in May 2015. Work group members are perioperative medicine experts from US medical centers. They have extensive knowledge of the literature as well as administrative and clinical experience in a variety of perioperative care models.

Topic Development. Initial work was focused on reviewing and discussing multiple models of perioperative care and exploring the roles that hospital medicine physicians have within these models. Useful information was summarized to guide hospitals and physicians in designing, implementing, and expanding patient-centric perioperative medicine services with a focus on preoperative and postoperative care. A final document was created; it outlines system-level issues in perioperative care, organized by perioperative phases.

Initial Framework. Group members submitted written descriptions of key issues in each of 4 phases: (1) preoperative, (2) day of surgery, (3) postoperative inpatient, and (4) postdischarge. These descriptions were merged and reviewed by the content experts. Editing and discussion from the entire group were incorporated into the final matrix, which highlighted (1) perioperative phase definitions, (2) requirements for patients to move to next phase, (3) elements of care coordination typically provided by surgery, anesthesiology, and medicine disciplines, (4) concerns and risks particular to each phase, (5) unique considerations for each phase, (6) suggested metrics of success, and (7) key questions for determining the effectiveness of perioperative care in an institution. All members provided final evaluation and editing.

Final Approval. The Perioperative Care Matrix for Inpatient Surgeries (PCMIS) was presented to the board of the Society of Hospital Medicine in fall 2015 and was approved for use in centering and directing discussions regarding perioperative care.

Models of Care. The Perioperative Care Work Group surveyed examples of hospitalist engagement in perioperative care and synthesized these into synopses of existing models of care for the preoperative, day-of-surgery, postoperative-inpatient, and postdischarge phases.

Defining Key Concepts and Issues

Hospitalists have participated in a variety of perioperative roles for more than a decade. Roles include performing in-depth preoperative assessments, providing oversight to presurgical advanced practice provider assessments, providing inpatient comanagement and consultation both before and after surgery, and providing postdischarge follow-up within the surgical period for medical comorbidities.

Although a comprehensive look at the entire perioperative period is important, 4 specific phases were defined to guide this work (Figure). The phases identified were based on time relative to surgery, with unique considerations as to the overall perioperative period. Concerns and potential risks specific to each phase were considered (Table 1).

Models of Care

Multiple examples of hospitalist involvement were collected to inform the program development guidelines. The specifics noted among the reviewed practice models are described here.

Preoperative. In some centers, all patients scheduled for surgery are required to undergo evaluation at the institution’s preoperative clinic. At most others, referral to the preoperative clinic is at the discretion of the surgical specialists, who have been informed of the clinic’s available resources. Factors determining whether a patient has an in-person clinic visit, undergoes a telephone-based medical evaluation, or has a referral deferred to the primary care physician (PCP) include patient complexity and surgery-specific risk. Patients who have major medical comorbidities (eg, chronic lung or heart disease) or are undergoing higher risk procedures (eg, those lasting >1 hour, laparotomy) most often undergo a formal clinic evaluation. Often, even for a patient whose preoperative evaluation is completed by a PCP, the preoperative nursing staff will call before surgery to provide instructions and to confirm that preoperative planning is complete. Confirmation includes ensuring that the surgery consent and preoperative history and physical examination documents are in the medical record, and that all recommended tests have been performed. If deficiencies are found, surgical and preoperative clinic staff are notified.

The United States is focused on sensible, high-value care. Perioperative care is burgeoning with opportunities for improvement, including reducing avoidable complications, developing systems for early recognition and treatment of complications, and streamlining processes to shorten length of stay and improve patient experience. The PCMIS provides the needed platform to catalyze detailed collaborative work between disciplines engaged in perioperative care.

As average age and level of medical comorbidity increase among surgical patients, hospitalists will increasingly be called on to assist in perioperative care. Hospitalists have long been involved in caring for medically complex surgical patients, through comanagement, consultation, and preoperative evaluations. As a provider group, hospitalists have comprehensive skills in quality and systems improvement, and in program development across hospital systems nationwide. Hospitalists have demonstrated their value by focusing on improving patient outcomes and enhancing patient engagement and experiences. Additionally, the perioperative period is fraught with multiple and complicated handoffs, a problem area for which hospital medicine has pioneered solutions and developed unique expertise. Hospital medicine is well prepared to provide skilled and proven leadership in the timely development, improvement, and expansion of perioperative care for this increasingly older and chronically ill population.

Hospitalists are established in multiple perioperative roles for high-risk surgical patients and have the opportunity to expand optimal patient-centric perioperative care systems working in close concert with surgeons and anesthesiologists. The basics of developing these systems include (1) assessing risk for medical complications, (2) planning for perioperative care, (3) developing programs aimed at risk reduction for preventable complications and early identification and intervention for unavoidable complications, and (4) guiding quality improvement efforts, including planning for frequent handoffs and transitions.

As a key partner in developing comprehensive programs in perioperative care, hospital medicine will continue to shape the future of hospital care for all patients. The PCMIS, as developed with support from the Society of Hospital Medicine, will aid efforts to achieve the best perioperative care models for our surgical patients.

Disclosures

Financial activities outside the submitted work: Drs. Pfeifer and Jaffer report payment for development of educational presentations; Dr. Grant reports payment for expert testimony pertaining to hospital medicine; Drs. Grant and Jaffer report royalties from publishing; Drs. Thompson, Pfiefer, Grant, Slawski, and Jaffer report travel expenses for speaking and serving on national committees; and Drs. Slawski and Jaffer serve on the board of the Society of Perioperative Assessment and Quality Improvement. The other authors have nothing to report.

In Reply: We thank Dr. Jandali for his thoughtful comments on our article. We acknowledge that there may be a small subset of patients in whom low-intensity warfarin may be worth trying—such as patients with a history of idiopathic or recurrent venous thromboembolism in whom problematic (but not life-threatening) bleeding recurs—but only when the international normalized ratio (INR) is at the high end of the therapeutic range or slightly above it. However, when attempting to apply the results from PREVENT¹ and ELATE² to clinical practice and the management of anticoagulation after hemorrhage, it is important to note that in ELATE there was a higher incidence of recurrent thromboembolism in patients on lower-intensity anticoagulation than in those on conventional treatment, and no significant difference in major bleeding was noted between the high- and low-intensity groups.

We acknowledge, though, that the rates of major bleeding were surprisingly low in the high-intensity group in this study relative to historical controls and so may not apply to all patients.

It is also important to recognize that several studies have evaluated low-intensity dosing for stroke prophylaxis in atrial fibrillation with generally disappointing results, and at present, expert opinion continues to support a therapeutic INR goal of 2.0 to 3.0.³

Therefore, we believe that low-intensity warfarin treatment is only appropriate to try in a very small subset of carefully selected patients with a history of venous thromboembolism who have proven that they cannot tolerate full-dose warfarin and in whom a trial of low-dose warfarin treatment carries acceptable risk.

In Reply: We thank Dr. Jandali for his thoughtful comments on our article. We acknowledge that there may be a small subset of patients in whom low-intensity warfarin may be worth trying—such as patients with a history of idiopathic or recurrent venous thromboembolism in whom problematic (but not life-threatening) bleeding recurs—but only when the international normalized ratio (INR) is at the high end of the therapeutic range or slightly above it. However, when attempting to apply the results from PREVENT¹ and ELATE² to clinical practice and the management of anticoagulation after hemorrhage, it is important to note that in ELATE there was a higher incidence of recurrent thromboembolism in patients on lower-intensity anticoagulation than in those on conventional treatment, and no significant difference in major bleeding was noted between the high- and low-intensity groups.

We acknowledge, though, that the rates of major bleeding were surprisingly low in the high-intensity group in this study relative to historical controls and so may not apply to all patients.

It is also important to recognize that several studies have evaluated low-intensity dosing for stroke prophylaxis in atrial fibrillation with generally disappointing results, and at present, expert opinion continues to support a therapeutic INR goal of 2.0 to 3.0.³

Therefore, we believe that low-intensity warfarin treatment is only appropriate to try in a very small subset of carefully selected patients with a history of venous thromboembolism who have proven that they cannot tolerate full-dose warfarin and in whom a trial of low-dose warfarin treatment carries acceptable risk.

In Reply: We thank Dr. Jandali for his thoughtful comments on our article. We acknowledge that there may be a small subset of patients in whom low-intensity warfarin may be worth trying—such as patients with a history of idiopathic or recurrent venous thromboembolism in whom problematic (but not life-threatening) bleeding recurs—but only when the international normalized ratio (INR) is at the high end of the therapeutic range or slightly above it. However, when attempting to apply the results from PREVENT¹ and ELATE² to clinical practice and the management of anticoagulation after hemorrhage, it is important to note that in ELATE there was a higher incidence of recurrent thromboembolism in patients on lower-intensity anticoagulation than in those on conventional treatment, and no significant difference in major bleeding was noted between the high- and low-intensity groups.

We acknowledge, though, that the rates of major bleeding were surprisingly low in the high-intensity group in this study relative to historical controls and so may not apply to all patients.

It is also important to recognize that several studies have evaluated low-intensity dosing for stroke prophylaxis in atrial fibrillation with generally disappointing results, and at present, expert opinion continues to support a therapeutic INR goal of 2.0 to 3.0.³

Therefore, we believe that low-intensity warfarin treatment is only appropriate to try in a very small subset of carefully selected patients with a history of venous thromboembolism who have proven that they cannot tolerate full-dose warfarin and in whom a trial of low-dose warfarin treatment carries acceptable risk.

If a patient receiving anticoagulant therapy suffers a bleeding event, the patient and physician must decide whether and how soon to restart the therapy, and with what agent.

Foremost on our minds tends to be the risk of another hemorrhage. Subtler to appreciate immediately after an event is the continued risk of thrombosis, often from the same medical condition that prompted anticoagulation therapy in the first place (Table 1).

Complicating the decision, there may be a rebound effect: some thrombotic events such as pulmonary embolism and atrial fibrillation-related stroke may be more likely to occur in the first weeks after stopping warfarin than during similar intervals in patients who have not been taking it.^1–3 The same thing may happen with the newer, target-specific oral anticoagulants.^4–6

Although we have evidence-based guidelines for initiating and managing anticoagulant therapy, ample data on adverse events, and protocols for reversing anticoagulation if bleeding occurs, we do not have clear guidelines on restarting anticoagulation after a hemorrhagic event.

In this article, we outline a practical framework for approaching this clinical dilemma. Used in conjunction with consideration of a patient’s values and preferences as well as input from experts, this framework can help clinicians guide their patients through this challenging clinical decision. It consists of five questions:

• Why is the patient on anticoagulation, and what is the risk of thromboembolism without it?
• What was the clinical impact of the hemorrhage, and what is the risk of rebleeding if anticoagulation is resumed?
• What additional patient factors should be taken into consideration?
• How long should we wait before restarting anticoagulation?
• Would a newer drug be a better choice?

BLEEDING OCCURS IN 2% TO 3% OF PATIENTS PER YEAR

Most of our information on anticoagulation is about vitamin K antagonists—principally warfarin, in use since the 1950s. Among patients taking warfarin outside of clinical trials, the risk of major bleeding is estimated at 2% to 3% per year.⁷

However, the target-specific oral anticoagulants rivaroxaban (Xarelto), apixaban (Eliquis),  dabigatran (Pradaxa) and edoxaban (Savaysa) are being used more and more, and we include them in our discussion insofar as we have information on them. The rates of bleeding with these new drugs in clinical trials have been comparable to or lower than those with warfarin.⁸ Postmarketing surveillance is under way.

WHY IS THE PATIENT ON ANTICOAGULATION? WHAT IS THE RISK WITHOUT IT?

Common, evidence-based indications for anticoagulation are to prevent complications in patients with venous thromboembolism and to prevent stroke in patients with atrial fibrillation or a mechanical heart valve. Other uses, such as in heart failure and its sequelae, pulmonary hypertension, and splanchnic or hepatic vein thrombosis, have less robust evidence to support them.

When anticoagulation-related bleeding occurs, it is essential to review why the patient is taking the drug and the risk of thromboembolism without it. Some indications pose a higher risk of thromboembolism than others and so argue more strongly for continuing the treatment.

Douketis et al⁹ developed a risk-stratification scheme for perioperative thromboembolism. We have modified it by adding the CHA₂DS₂-VASc score (Table 2),^9–11 and believe it can be used more widely.

High-risk indications

Conditions that pose a high risk of thrombosis almost always require restarting anticoagulation. Here, the most appropriate question nearly always is not if anticoagulation should be restarted, but when. Examples:

• A mechanical mitral valve
• Antiphospholipid antibody syndrome with recurrent thromboembolic events.

Lower-risk indications

Lower-risk indications allow more leeway in determining if anticoagulation should be resumed. The most straightforward cases fall well within established guidelines. Examples:

• Atrial fibrillation and a CHA₂DS₂-VASc score of 1. The 2014 guidelines from the American College of Cardiology, American Heart Association, and Heart Rhythm Society¹⁰ suggest that patients with nonvalvular atrial fibrillation and a CHA₂DS₂-VASc score of 1 have three options: an oral anticoagulant, aspirin, and no antithrombotic therapy. If such a patient on anticoagulant therapy subsequently experiences a major gastrointestinal hemorrhage requiring transfusion and intensive care and no definitively treatable source of bleeding is found on endoscopy, one can argue that the risks of continued anticoagulation (recurrent bleeding) now exceed the benefits and that the patient would be better served by aspirin or even no antithrombotic therapy.

• After 6 months of anticoagulation for unprovoked deep vein thrombosis. Several studies showed that aspirin reduced the risk of recurrent venous thromboembolism in patients who completed an initial 6-month course of anticoagulation.^12–15 Though these studies did not specifically compare aspirin with warfarin or target-specific oral anticoagulants in preventing recurrent venous thromboembolism after a hemorrhage, it is reasonable to extrapolate their results to this situation.

If the risk of recurrent hemorrhage on anticoagulation is considered to be too great, then aspirin is an alternative to no anticoagulation, as it reduces the risk of recurrent venous thromboembolism.¹⁶ However, we advise caution if the bleeding lesion may be specifically exacerbated by aspirin, particularly upper gastrointestinal ulcers.

Moderate-risk indications

• After a partial course of anticoagulation for provoked venous thromboembolism. Suppose a patient in the 10th week of a planned 12-week course of anticoagulation for a surgically provoked, first deep vein thrombosis presents with abdominal pain and is found to have a retroperitoneal hematoma. In light of the risk of recurrent bleeding vs the benefit of resuming anticoagulation for the limited remaining period, her 12-week treatment course can reasonably be shortened to 10 weeks.

The risk of recurrent venous thromboembolism when a patient is off anticoagulation decreases with time from the initial event. The highest risk, estimated at 0.3% to 1.3% per day, is in the first 4 weeks, falling to 0.03% to 0.2% per day in weeks 5 through 12, and 0.05% per day thereafter.^17–20

The risk of recurrent venous thromboembolism is greatest immediately after the event and decreases over time

Additionally, a pooled analysis of seven randomized trials suggests that patients with isolated, distal deep vein thrombosis provoked by a temporary risk factor did not have a high risk of recurrence after being treated for 4 to 6 weeks.²¹ These analyses are based on vitamin K antagonists, though it seems reasonable to extrapolate this information to the target-specific oral anticoagulants.

More challenging are situations in which the evidence supporting the initial or continued need for anticoagulation is less robust, such as in heart failure, pulmonary hypertension, or splanchnic and hepatic vein thrombosis. In these cases, the lack of strong evidence supporting the use of anticoagulation should make us hesitate to resume it after bleeding.

WHAT WAS THE CLINICAL IMPACT? WHAT IS THE RISK OF REBLEEDING?

Different groups have defined major and minor bleeding in different ways.^22,23 Several have proposed criteria to standardize how bleeding events (on warfarin and otherwise) are classified,^23–25 but the definitions differ.

Specifically, all agree that a “major” bleeding event is one that is fatal, involves bleeding into a major organ, or leads to a substantial decline in hemoglobin level. However, the Thrombolysis in Myocardial Infarction trials use a decline of more than 5 g/dL in their definition,^23,25 while the International Society on Thrombosis and Haemostasis uses 2 g/dL.²⁴

Here, we review the clinical impact of the most common sources of anticoagulation-related hemorrhage—gastrointestinal, soft tissue, and urinary tract²⁶—as well as intracerebral hemorrhage, a less common but more uniformly devastating event.²⁷

Clinical impact of gastrointestinal hemorrhage

Each year, about 4.5% of patients taking warfarin have a gastrointestinal hemorrhage, though not all of these events are major.²⁸ Evolving data suggest that the newer agents (particularly dabigatran, rivaroxaban, and edoxaban) pose a higher risk of gastrointestinal bleeding than warfarin.²⁹ Patients may need plasma and blood transfusions and intravenous phytonadione, all of which carry risks, albeit small.

Frequently, endoscopy is needed to find the source of bleeding and to control it. If this does not work, angiographic intervention to infuse vasoconstrictors or embolic coils into the culprit artery may be required, and some patients need surgery. Each intervention carries its own risk.

Clinical impact of soft-tissue hemorrhage

Soft-tissue hemorrhage accounts for more than 20% of warfarin-related bleeding events²⁶; as yet, we know of no data on the rate with the new drugs. Soft-tissue hemorrhage is often localized to the large muscles of the retroperitoneum and legs. Though retroperitoneal hemorrhage accounts for a relatively small portion of soft-tissue hemorrhages, it is associated with high rates of morbidity and death and will therefore be our focus.²⁶

Some indications for anticoagulation pose a higher risk of thromboembolism than others

Much of the clinical impact of retroperitoneal hemorrhage is from a mass effect that causes abdominal compartment syndrome, hydroureter, ileus, abscess formation, and acute and chronic pain. At least 20% of cases are associated with femoral neuropathy. It can also lead to deep vein thrombosis from venous compression, coupled with hypercoagulability in response to bleeding. Brisk bleeding can lead to shock and death, and the mortality rate in retroperitoneal hemorrhage is estimated at 20% or higher.³⁰

In many cases, the retroperitoneal hemorrhage will self-tamponade and the blood will be reabsorbed once the bleeding has stopped, but uncontrolled bleeding may require surgical or angiographic intervention.³⁰

Clinical impact of urinary tract hemorrhage

Gross or microscopic hematuria can be found in an estimated 2% to 24% of patients taking warfarin^31–33; data are lacking for the target-specific oral anticoagulants. Interventions required to manage urinary tract bleeding include bladder irrigation and, less often, transfusion.³¹ Since a significant number of cases of hematuria are due to neoplastic disease,³² a diagnostic workup with radiographic imaging of the upper tract and cystoscopy of the lower tract is usually required.³¹ While life-threatening hemorrhage is uncommon, complications such as transient urinary obstruction from clots may occur.

Clinical impact of intracranial hemorrhage

Intracranial hemorrhage is the most feared and deadly of the bleeding complications of anticoagulation. The incidence in patients on warfarin is estimated at 2% to 3% per year, which is markedly higher than the estimated incidence of 25 per 100,000 person-years in the general population.³⁴ Emerging data indicate that the newer drugs are also associated with a risk of intracranial hemorrhage, though the risk is about half that with vitamin K antagonists.³⁵ Intracranial hemorrhage leads to death or disability in 76% of cases, compared with 3% of cases of bleeding from the gastrointestinal or urinary tract.²⁷

Regardless of the source of bleeding, hospitalization is likely to be required and may be  prolonged, with attendant risks of nosocomial harms such as infection.

Risk of rebleeding

Given the scope and severity of anticoagulation-related bleeding, there is strong interest in predicting and preventing it. By some estimates, the incidence of recurrent bleeding after resuming vitamin K antagonists is 8% to 13%.²² Although there are several indices for predicting the risk of major bleeding when starting anticoagulation, there are currently no validated tools to estimate a patient’s risk of rebleeding.³⁶

The patient factor that most consistently predicts major bleeding is a history of bleeding, particularly from the gastrointestinal tract. Finding and controlling the source of bleeding is important.^26,37 For example, a patient with gross hematuria who is found on cystoscopy to have a urothelial papilloma is unlikely to have rebleeding if the tumor is successfully resected and serial follow-up shows no regrowth. In contrast, consider a patient with a major gastrointestinal hemorrhage, the source of which remains elusive after upper, lower, and capsule endoscopy or, alternatively, is suspected to be from one of multiple angiodysplastic lesions. Without definitive source management, this patient faces a high risk of rebleeding.

With or without anticoagulation, after a first intracranial hemorrhage the risk of another one is estimated at 2% to 4% per year.³⁴ An observational study found a recurrence rate of 7.5% when vitamin K antagonist therapy was started after an intracranial hemorrhage (though not all patients were on a vitamin K antagonist at the time of the first hemorrhage).³⁸

Evolving data suggest the newer oral agents pose a higher risk of GI bleeding

Patients with lobar hemorrhage and those with suspected cerebral amyloid angiopathy may be at particularly high risk if anticoagulation is resumed. Conversely, initial events attributed to uncontrolled hypertension that subsequently can be well controlled may portend a lower risk of rebleeding.³⁴ For other types of intracranial hemorrhage, recurrence rates can be even higher. Irrespective of anticoagulation, one prospective study estimated the crude annual rebleeding rate with untreated arteriovenous malformations to be 7%.³⁹ In chronic subdural hematoma, the recurrence rate after initial drainage has been estimated at 9.2% to 26.5%, with use of anticoagulants (in this case, vitamin K antagonists) being an independent predictor of recurrence.⁴⁰

WHAT OTHER PATIENT FACTORS NEED CONSIDERATION?

Target INR on warfarin

An important factor influencing the risk of bleeding with warfarin is the intensity of this therapy.³⁷ A meta-analysis⁴¹ found that the risks of major hemorrhage and thromboembolism are minimized if the goal international normalized ratio (INR) is 2.0 to 3.0. When considering resuming anticoagulation after bleeding, make sure the therapeutic target is appropriate.³⁷

Table 3 summarizes recommended therapeutic ranges for frequently encountered indications for warfarin.^36,42,43

INR at time of the event and challenges in controlling it

The decision to resume anticoagulation in patients who bled while using warfarin must take into account the actual INR at the time of the event.

For example, consider a patient whose INR values are consistently in the therapeutic range.  While on vacation, he receives ciprofloxacin for acute prostatitis from an urgent care team, and no adjustment to INR monitoring or warfarin dose is made. Several days later, he presents with lower gastrointestinal bleeding. His INR is 8, and colonoscopy reveals diverticulosis with a bleeding vessel, responsive to endoscopic therapy. After controlling the source of bleeding and reinforcing the need to always review new medications for potential interactions with anticoagulation, it is reasonable to expect that he once again will be able to keep his INR in the therapeutic range.

A patient on anticoagulation for the same indication but who has a history of repeated supratherapeutic levels, poor adherence, or poor access to INR monitoring poses very different concerns about resuming anticoagulation (as well as which agent to use, as we discuss below).

Of note, a high INR alone does not explain bleeding. It is estimated that a workup for gastrointestinal bleeding and gross hematuria uncovers previously undetected lesions in approximately one-third of cases involving warfarin.²⁶ A similar malignancy-unmasking effect is now recognized in patients using the target-specific oral agents who experience gastrointestinal bleeding.⁴⁴ Accordingly, we recommend a comprehensive source evaluation for any anticoagulation-related hemorrhage.

Comorbid conditions

Comorbid conditions associated with bleeding include cancer, end-stage renal disease, liver disease, arterial hypertension, prior stroke, and alcohol abuse.^37,45 Gait instability, regardless of cause, may also increase the risk of trauma-related hemorrhage, but some have estimated that a patient would need to fall multiple times per week to contraindicate anticoagulation on the basis of falls alone.⁴⁶

Concurrent medications

Concomitant therapies, including antiplatelet drugs and nonsteroidal anti-inflammatory drugs, increase bleeding risk.^47,48 Aspirin and the nonsteroidals, in addition to having antiplatelet effects, also can cause gastric erosion.³⁷ In evaluating whether and when to restart anticoagulation, it is advisable to review the role that concomitant therapies may have had in the index bleeding event and to evaluate the risks and benefits of these other agents.

The factor that most consistently predicts major bleeding is a history of bleeding, particularly gastrointestinal bleeding

Additionally, warfarin has many interactions. Although the newer drugs are lauded for having fewer interactions, they are not completely free of them, and the potential for interactions must always be reviewed.⁴⁹ Further, unlike warfarin therapy, therapy with the newer agents is not routinely monitored with laboratory tests, so toxicity (or underdosing) may not be recognized until an adverse clinical event occurs. Ultimately, it may be safer to resume anticoagulation after a contributing drug can be safely discontinued.

Advanced age

The influence that the patient’s age should have on the decision to restart anticoagulation is unclear. Although the risk of intracranial hemorrhage increases with age, particularly after age 80, limited data exist in this population, particularly with regard to rebleeding. Further, age is a major risk factor for most thrombotic events, including venous thromboembolism and stroke from atrial fibrillation, so although the risks of anticoagulation may be higher, the benefits may also be higher than in younger patients.^37,46 We discourage using age alone as a reason to withhold anticoagulation after a hemorrhage.

HOW LONG SHOULD WE WAIT TO RESTART ANTICOAGULATION?

We lack conclusive data on how long to wait to restart anticoagulation after an anticoagulation-associated hemorrhage.

The decision is complicated by evidence suggesting a rebound effect, with an increased risk of pulmonary embolism and atrial fibrillation-related stroke during the first 90 days of interruption of therapy with warfarin as well as with target-specific oral anticoagulants.^3–8 In anticoagulation-associated retroperitoneal bleeding, there is increased risk of deep vein thrombosis from compression, even if venous thromboembolism was not the initial indication for anticoagulation.³⁰

In patients with intracranial hemorrhage, evidence suggests that the intracranial hemorrhage itself increases the risk of arterial and venous thromboembolic events. Irrespective of whether a patient was previously on anticoagulation, the risk of arterial and venous thromboembolic events approaches 7% during the initial intracranial hemorrhage-related hospitalization and 9% during the first 90 days.^34,50,51

To date, the only information we have about when to resume anticoagulation comes from patients taking vitamin K antagonists.

Timing after gastrointestinal bleeding

Small case series suggest that in the first 2 months after warfarin-associated gastrointestinal bleeding, there is substantial risk of rebleeding when anticoagulation is resumed—and of thrombosis when it is not.^52,53 Two retrospective cohort studies may provide some guidance in this dilemma.^28,54

A workup for GI bleeding and gross hematuria uncovers previously undetected lesions in about one-third of cases involving warfarin

Witt et al²⁸ followed 442 patients who presented with gastrointestinal bleeding from any site during warfarin therapy for varied indications for up to 90 days after the index bleeding event. The risk of death was three times lower in patients who restarted warfarin than in those who did not, and their rate of thrombotic events was 10 times lower. The risk of recurrent gastrointestinal bleeding was statistically insignificant, and there were no fatal bleeding events. Anticoagulant therapy was generally resumed within 1 week of the bleeding event, at a median of 4 days.^28,55

Qureshi et al⁵⁴ performed a retrospective cohort study of 1,329 patients with  nonvalvular atrial fibrillation who had experienced a gastrointestinal hemorrhage while taking warfarin. They found that resuming warfarin after 7 days was not associated with a higher risk of recurrent gastrointestinal bleeding and that the rates of death and thromboembolism were lower than in patients who resumed warfarin after 30 days. On the other hand, the risk of recurrent gastrointestinal bleeding was significantly greater if therapy was resumed within the first week.

In view of these studies, we believe that most patients should resume anticoagulation after 4 to 7 days of interruption after gastrointestinal bleeding.⁵⁵

Timing after soft-tissue hemorrhage

The literature on resuming anticoagulation after soft-tissue hemorrhage is sparse. A retrospective study⁵² looked at this question in patients with spontaneous rectal sheath hematoma who had been receiving  antiplatelet drugs, intravenous heparin, vitamin K antagonists, or a combination of these, but not target-specific agents. More than half of the patients were on vitamin K antagonists at the time of hemorrhage. Analysis suggested that when benefits of resuming anticoagulation are believed to outweigh risks, it is reasonable to resume anticoagulation 4 days after the index event.⁵⁶

Timing after intracranial hemorrhage

Anticoagulation should not be considered within the first 24 hours after intracranial hemorrhage, as over 70% of patients develop some amount of hematoma expansion during this time.^34,57 The period thereafter poses a challenge, as the risk of hematoma expansion decreases while the risk of arterial and venous thromboembolism is ongoing and cumulative.⁵⁰

Perhaps surprisingly, national guidelines suggest starting prophylactic-dosed anticoagulation early in all intracranial hemorrhage patients, including those not previously on warfarin.^58,59 In a randomized trial, Boeer et al⁶⁰ concluded that starting low-dose subcutaneous heparin the day after an intracranial hemorrhage decreased the risk of thromboembolism without increasing the risk of rebleeding.⁶⁰ Dickmann et al⁶¹ similarly concluded that there was no increased risk of rebleeding with early prophylactic-dosed subcutaneous heparin.⁶¹ Optimal mechanical thromboprophylaxis, including graduated compression stockings and intermittent pneumatic compression stockings, is also encouraged.³⁴

We discourage using age alone as a reason to withhold anticoagulation after a hemorrhage

Expert opinion remains divided on when and if anticoagulants should be resumed.^34,62 The American Heart Association suggests that in nonvalvular atrial fibrillation, long-term anticoagulation should be avoided after spontaneous lobar hemorrhage; antiplatelet agents can be considered instead.⁵⁸ In nonlobar hemorrhage, the American Heart Association suggests that anticoagulation be considered, depending on strength of indication, 7 to 10 days after the onset.⁵⁸ The European Stroke Initiative suggests patients with strong indications for anticoagulation be restarted on warfarin 10 to 14 days after the event, depending on the risk of thromboembolism and recurrent intracranial hemorrhage.⁵⁹ Others suggest delaying resumption to 10 to 30 weeks after an index intracranial hemorrhage.⁶³

Overall, in the immediate acute period of intracranial hemorrhage, most patients will likely benefit from acute reversal of anticoagulation, followed by institution of prophylactic-dose anticoagulation after the first 24 hours. Going forward, patients who remain at higher risk of a recurrence of anticoagulant-related intracranial hemorrhage (such as those with lobar hemorrhage, suspected cerebral amyloid angiopathy, and other high-risk factors) than of thromboembolic events may be best managed without anticoagulants. Alternatively, patients with deep hemispheric intracranial hemorrhage, hypertension that can be well controlled, and a high risk of serious thromboembolism may experience net benefit from restarting anticoagulation.³⁴

We recommend considering restarting anticoagulation 7 days after the onset of intracranial hemorrhage in patients at high risk of thromboembolism and after at least 14 days for patients at lower risk (Table 2). Discussions with neurologic and neurosurgical consultants should also inform this timing decision.

WOULD A NEWER DRUG BE A BETTER CHOICE?

The emergence of target-specific oral anticoagulants, including factor Xa inhibitors such as rivaroxaban, apixaban, and edoxaban and the direct thrombin inhibitor dabigatran etexilate, presents further challenges in managing anticoagulation after hemorrhage. Table 4 summarizes the current FDA-approved indications.^64–67

These newer agents are attractive because, compared with warfarin, they have wider therapeutic windows, faster onset and offset of action, and fewer drug and food interactions.⁶⁸ A meta-analysis of data available to date suggests that the new drugs, compared with warfarin, show a favorable risk-benefit profile with reductions in stroke, intracranial hemorrhage, and mortality with similar overall major bleeding rates, except for a possible increase in gastrointestinal bleeding.⁶⁸

However, when managing anticoagulation after a bleeding event, the newer agents are challenging for two reasons: they may be associated with a higher incidence of gastrointestinal bleeding than warfarin, and they lack the typical reversal agents that can be used to manage an acute bleeding event.^68,69

In individual studies comparing warfarin with dabigatran,⁷⁰ rivaroxaban,⁷¹ apixaban,⁷² or edoxaban⁷³ for stroke prevention in patients with atrial fibrillation, there was no significant difference in the rate of major bleeding between dabigatran in its higher dose (150 mg twice a day) or rivaroxaban compared with warfarin.^70,71 The risk of major bleeding was actually lower with apixaban⁷² and edoxaban.⁷³

In regard to specific types of major bleeding, the rate of intracranial hemorrhage was significantly lower with dabigatran, rivaroxaban, apixaban, and edoxaban than with warfarin.^35,68–73 Some have proposed that since the brain is high in tissue factor, inhibition of tissue factor-factor VIIa complexes by vitamin K antagonists leaves the brain vulnerable to hemorrhage. Others suggest that the targeted mechanism of target-specific agents, as opposed to the multiple pathways in both the intrinsic and extrinsic coagulation cascade that vitamin K antagonists affect, may explain this difference.^35,74,75

However, some studies suggest that rivaroxaban and the higher doses of dabigatran and edoxaban are associated with higher rates of major gastrointestinal bleeding compared with warfarin.^69–71,76 But apixaban demonstrated no significant difference in gastrointestinal bleeding, and instead demonstrated rates of gastrointestinal bleeding comparable to that with aspirin for stroke prevention in atrial fibrillation.⁷²

The new oral anticoagulants lack antidotes or reversal agents such as phytonadione and fresh-frozen plasma that are available to manage warfarin-associated bleeding events. Other proposed reversal options for the new agents include activated charcoal (if the drugs were taken recently enough to remain in the gastrointestinal tract) and concentrated clotting factor product, though research is ongoing in regards to the most appropriate use in clinical practice.^37,69 Unlike rivaroxaban and apixaban, dabigatran has low plasma protein binding and is dialyzable, which provides another strategy in managing dabigatran-related bleeding.⁶⁹

We believe most patients should resume anticoagulation after 4 to 7 days of interruption after GI bleeding

Of note, the above bleeding risk calculations relate to the first anticoagulant-related bleeding event, though presumably the same risk comparison across agents may be applicable to rebleeding events. Given the data above, when anticoagulation is to be resumed after an intracranial hemorrhage, the risk of rebleeding, particularly in the form of recurrent intracranial hemorrhage, may be lower if a target-specific oral anticoagulant is used.⁷⁵ Similarly, when anticoagulation is to be resumed after a gastrointestinal bleeding event, reinitiation with warfarin or apixaban therapy may present the lowest risk of recurrent gastrointestinal rebleeding. In other sources of bleeding, such as retroperitoneal bleeding, we suggest consideration of transitioning to warfarin, given the availability of reversal agents in the event of recurrent bleeding.

Other important drug-specific factors that must be noted when selecting an agent with which to resume anticoagulation after a hemorrhage include the following:

• In patients with significant renal impairment, the choice of agent will be limited to a vitamin K antagonist.⁷⁷
• A meta-analysis of randomized clinical trials suggests that in the elderly (age 75 and older) target-specific oral anticoagulants did not cause excess bleeding and were associated with at least equal efficacy compared with vitamin K antagonists.⁷⁸
• Target-specific oral anticoagulants may be beneficial in patients who have challenges in achieving INR targets, as evidence suggests that switching to them is associated with a reduction in bleeding for patients who struggle to maintain an appropriately therapeutic INR.⁶⁸ On the other hand, if there is concern that a patient may occasionally miss doses of an anticoagulant, given the rapid onset and offset of action of target-specific agents compared with warfarin, a missed dose of a target-specific agent may result in faster dissolution of anticoagulant effect and increased risk of thrombotic events, and lapses in anticoagulation will not be identified by routine drug monitoring.^6–8,75 As such, it is vital to have a frank discussion with any patient who has difficulty maintaining therapeutic INRs on warfarin treatment to make sure that he or she is not missing doses.
• If there is no clear and compelling reason to select a particular agent, cost considerations should be taken into account. We have included estimated 30-day pricing for the various agents in Table 4. 

If a patient receiving anticoagulant therapy suffers a bleeding event, the patient and physician must decide whether and how soon to restart the therapy, and with what agent.

Foremost on our minds tends to be the risk of another hemorrhage. Subtler to appreciate immediately after an event is the continued risk of thrombosis, often from the same medical condition that prompted anticoagulation therapy in the first place (Table 1).

Complicating the decision, there may be a rebound effect: some thrombotic events such as pulmonary embolism and atrial fibrillation-related stroke may be more likely to occur in the first weeks after stopping warfarin than during similar intervals in patients who have not been taking it.^1–3 The same thing may happen with the newer, target-specific oral anticoagulants.^4–6

Although we have evidence-based guidelines for initiating and managing anticoagulant therapy, ample data on adverse events, and protocols for reversing anticoagulation if bleeding occurs, we do not have clear guidelines on restarting anticoagulation after a hemorrhagic event.

In this article, we outline a practical framework for approaching this clinical dilemma. Used in conjunction with consideration of a patient’s values and preferences as well as input from experts, this framework can help clinicians guide their patients through this challenging clinical decision. It consists of five questions:

• Why is the patient on anticoagulation, and what is the risk of thromboembolism without it?
• What was the clinical impact of the hemorrhage, and what is the risk of rebleeding if anticoagulation is resumed?
• What additional patient factors should be taken into consideration?
• How long should we wait before restarting anticoagulation?
• Would a newer drug be a better choice?

BLEEDING OCCURS IN 2% TO 3% OF PATIENTS PER YEAR

Most of our information on anticoagulation is about vitamin K antagonists—principally warfarin, in use since the 1950s. Among patients taking warfarin outside of clinical trials, the risk of major bleeding is estimated at 2% to 3% per year.⁷

However, the target-specific oral anticoagulants rivaroxaban (Xarelto), apixaban (Eliquis),  dabigatran (Pradaxa) and edoxaban (Savaysa) are being used more and more, and we include them in our discussion insofar as we have information on them. The rates of bleeding with these new drugs in clinical trials have been comparable to or lower than those with warfarin.⁸ Postmarketing surveillance is under way.

WHY IS THE PATIENT ON ANTICOAGULATION? WHAT IS THE RISK WITHOUT IT?

Common, evidence-based indications for anticoagulation are to prevent complications in patients with venous thromboembolism and to prevent stroke in patients with atrial fibrillation or a mechanical heart valve. Other uses, such as in heart failure and its sequelae, pulmonary hypertension, and splanchnic or hepatic vein thrombosis, have less robust evidence to support them.

When anticoagulation-related bleeding occurs, it is essential to review why the patient is taking the drug and the risk of thromboembolism without it. Some indications pose a higher risk of thromboembolism than others and so argue more strongly for continuing the treatment.

Douketis et al⁹ developed a risk-stratification scheme for perioperative thromboembolism. We have modified it by adding the CHA₂DS₂-VASc score (Table 2),^9–11 and believe it can be used more widely.

High-risk indications

Conditions that pose a high risk of thrombosis almost always require restarting anticoagulation. Here, the most appropriate question nearly always is not if anticoagulation should be restarted, but when. Examples:

• A mechanical mitral valve
• Antiphospholipid antibody syndrome with recurrent thromboembolic events.

Lower-risk indications

Lower-risk indications allow more leeway in determining if anticoagulation should be resumed. The most straightforward cases fall well within established guidelines. Examples:

• Atrial fibrillation and a CHA₂DS₂-VASc score of 1. The 2014 guidelines from the American College of Cardiology, American Heart Association, and Heart Rhythm Society¹⁰ suggest that patients with nonvalvular atrial fibrillation and a CHA₂DS₂-VASc score of 1 have three options: an oral anticoagulant, aspirin, and no antithrombotic therapy. If such a patient on anticoagulant therapy subsequently experiences a major gastrointestinal hemorrhage requiring transfusion and intensive care and no definitively treatable source of bleeding is found on endoscopy, one can argue that the risks of continued anticoagulation (recurrent bleeding) now exceed the benefits and that the patient would be better served by aspirin or even no antithrombotic therapy.

• After 6 months of anticoagulation for unprovoked deep vein thrombosis. Several studies showed that aspirin reduced the risk of recurrent venous thromboembolism in patients who completed an initial 6-month course of anticoagulation.^12–15 Though these studies did not specifically compare aspirin with warfarin or target-specific oral anticoagulants in preventing recurrent venous thromboembolism after a hemorrhage, it is reasonable to extrapolate their results to this situation.

If the risk of recurrent hemorrhage on anticoagulation is considered to be too great, then aspirin is an alternative to no anticoagulation, as it reduces the risk of recurrent venous thromboembolism.¹⁶ However, we advise caution if the bleeding lesion may be specifically exacerbated by aspirin, particularly upper gastrointestinal ulcers.

Moderate-risk indications

• After a partial course of anticoagulation for provoked venous thromboembolism. Suppose a patient in the 10th week of a planned 12-week course of anticoagulation for a surgically provoked, first deep vein thrombosis presents with abdominal pain and is found to have a retroperitoneal hematoma. In light of the risk of recurrent bleeding vs the benefit of resuming anticoagulation for the limited remaining period, her 12-week treatment course can reasonably be shortened to 10 weeks.

The risk of recurrent venous thromboembolism when a patient is off anticoagulation decreases with time from the initial event. The highest risk, estimated at 0.3% to 1.3% per day, is in the first 4 weeks, falling to 0.03% to 0.2% per day in weeks 5 through 12, and 0.05% per day thereafter.^17–20

The risk of recurrent venous thromboembolism is greatest immediately after the event and decreases over time

Additionally, a pooled analysis of seven randomized trials suggests that patients with isolated, distal deep vein thrombosis provoked by a temporary risk factor did not have a high risk of recurrence after being treated for 4 to 6 weeks.²¹ These analyses are based on vitamin K antagonists, though it seems reasonable to extrapolate this information to the target-specific oral anticoagulants.

More challenging are situations in which the evidence supporting the initial or continued need for anticoagulation is less robust, such as in heart failure, pulmonary hypertension, or splanchnic and hepatic vein thrombosis. In these cases, the lack of strong evidence supporting the use of anticoagulation should make us hesitate to resume it after bleeding.

WHAT WAS THE CLINICAL IMPACT? WHAT IS THE RISK OF REBLEEDING?

Different groups have defined major and minor bleeding in different ways.^22,23 Several have proposed criteria to standardize how bleeding events (on warfarin and otherwise) are classified,^23–25 but the definitions differ.

Specifically, all agree that a “major” bleeding event is one that is fatal, involves bleeding into a major organ, or leads to a substantial decline in hemoglobin level. However, the Thrombolysis in Myocardial Infarction trials use a decline of more than 5 g/dL in their definition,^23,25 while the International Society on Thrombosis and Haemostasis uses 2 g/dL.²⁴

Here, we review the clinical impact of the most common sources of anticoagulation-related hemorrhage—gastrointestinal, soft tissue, and urinary tract²⁶—as well as intracerebral hemorrhage, a less common but more uniformly devastating event.²⁷

Clinical impact of gastrointestinal hemorrhage

Each year, about 4.5% of patients taking warfarin have a gastrointestinal hemorrhage, though not all of these events are major.²⁸ Evolving data suggest that the newer agents (particularly dabigatran, rivaroxaban, and edoxaban) pose a higher risk of gastrointestinal bleeding than warfarin.²⁹ Patients may need plasma and blood transfusions and intravenous phytonadione, all of which carry risks, albeit small.

Frequently, endoscopy is needed to find the source of bleeding and to control it. If this does not work, angiographic intervention to infuse vasoconstrictors or embolic coils into the culprit artery may be required, and some patients need surgery. Each intervention carries its own risk.

Clinical impact of soft-tissue hemorrhage

Soft-tissue hemorrhage accounts for more than 20% of warfarin-related bleeding events²⁶; as yet, we know of no data on the rate with the new drugs. Soft-tissue hemorrhage is often localized to the large muscles of the retroperitoneum and legs. Though retroperitoneal hemorrhage accounts for a relatively small portion of soft-tissue hemorrhages, it is associated with high rates of morbidity and death and will therefore be our focus.²⁶

Some indications for anticoagulation pose a higher risk of thromboembolism than others

Much of the clinical impact of retroperitoneal hemorrhage is from a mass effect that causes abdominal compartment syndrome, hydroureter, ileus, abscess formation, and acute and chronic pain. At least 20% of cases are associated with femoral neuropathy. It can also lead to deep vein thrombosis from venous compression, coupled with hypercoagulability in response to bleeding. Brisk bleeding can lead to shock and death, and the mortality rate in retroperitoneal hemorrhage is estimated at 20% or higher.³⁰

In many cases, the retroperitoneal hemorrhage will self-tamponade and the blood will be reabsorbed once the bleeding has stopped, but uncontrolled bleeding may require surgical or angiographic intervention.³⁰

Clinical impact of urinary tract hemorrhage

Gross or microscopic hematuria can be found in an estimated 2% to 24% of patients taking warfarin^31–33; data are lacking for the target-specific oral anticoagulants. Interventions required to manage urinary tract bleeding include bladder irrigation and, less often, transfusion.³¹ Since a significant number of cases of hematuria are due to neoplastic disease,³² a diagnostic workup with radiographic imaging of the upper tract and cystoscopy of the lower tract is usually required.³¹ While life-threatening hemorrhage is uncommon, complications such as transient urinary obstruction from clots may occur.

Clinical impact of intracranial hemorrhage

Intracranial hemorrhage is the most feared and deadly of the bleeding complications of anticoagulation. The incidence in patients on warfarin is estimated at 2% to 3% per year, which is markedly higher than the estimated incidence of 25 per 100,000 person-years in the general population.³⁴ Emerging data indicate that the newer drugs are also associated with a risk of intracranial hemorrhage, though the risk is about half that with vitamin K antagonists.³⁵ Intracranial hemorrhage leads to death or disability in 76% of cases, compared with 3% of cases of bleeding from the gastrointestinal or urinary tract.²⁷

Regardless of the source of bleeding, hospitalization is likely to be required and may be  prolonged, with attendant risks of nosocomial harms such as infection.

Risk of rebleeding

Given the scope and severity of anticoagulation-related bleeding, there is strong interest in predicting and preventing it. By some estimates, the incidence of recurrent bleeding after resuming vitamin K antagonists is 8% to 13%.²² Although there are several indices for predicting the risk of major bleeding when starting anticoagulation, there are currently no validated tools to estimate a patient’s risk of rebleeding.³⁶

The patient factor that most consistently predicts major bleeding is a history of bleeding, particularly from the gastrointestinal tract. Finding and controlling the source of bleeding is important.^26,37 For example, a patient with gross hematuria who is found on cystoscopy to have a urothelial papilloma is unlikely to have rebleeding if the tumor is successfully resected and serial follow-up shows no regrowth. In contrast, consider a patient with a major gastrointestinal hemorrhage, the source of which remains elusive after upper, lower, and capsule endoscopy or, alternatively, is suspected to be from one of multiple angiodysplastic lesions. Without definitive source management, this patient faces a high risk of rebleeding.

With or without anticoagulation, after a first intracranial hemorrhage the risk of another one is estimated at 2% to 4% per year.³⁴ An observational study found a recurrence rate of 7.5% when vitamin K antagonist therapy was started after an intracranial hemorrhage (though not all patients were on a vitamin K antagonist at the time of the first hemorrhage).³⁸

Evolving data suggest the newer oral agents pose a higher risk of GI bleeding

Patients with lobar hemorrhage and those with suspected cerebral amyloid angiopathy may be at particularly high risk if anticoagulation is resumed. Conversely, initial events attributed to uncontrolled hypertension that subsequently can be well controlled may portend a lower risk of rebleeding.³⁴ For other types of intracranial hemorrhage, recurrence rates can be even higher. Irrespective of anticoagulation, one prospective study estimated the crude annual rebleeding rate with untreated arteriovenous malformations to be 7%.³⁹ In chronic subdural hematoma, the recurrence rate after initial drainage has been estimated at 9.2% to 26.5%, with use of anticoagulants (in this case, vitamin K antagonists) being an independent predictor of recurrence.⁴⁰

WHAT OTHER PATIENT FACTORS NEED CONSIDERATION?

Target INR on warfarin

An important factor influencing the risk of bleeding with warfarin is the intensity of this therapy.³⁷ A meta-analysis⁴¹ found that the risks of major hemorrhage and thromboembolism are minimized if the goal international normalized ratio (INR) is 2.0 to 3.0. When considering resuming anticoagulation after bleeding, make sure the therapeutic target is appropriate.³⁷

Table 3 summarizes recommended therapeutic ranges for frequently encountered indications for warfarin.^36,42,43

INR at time of the event and challenges in controlling it

The decision to resume anticoagulation in patients who bled while using warfarin must take into account the actual INR at the time of the event.

For example, consider a patient whose INR values are consistently in the therapeutic range.  While on vacation, he receives ciprofloxacin for acute prostatitis from an urgent care team, and no adjustment to INR monitoring or warfarin dose is made. Several days later, he presents with lower gastrointestinal bleeding. His INR is 8, and colonoscopy reveals diverticulosis with a bleeding vessel, responsive to endoscopic therapy. After controlling the source of bleeding and reinforcing the need to always review new medications for potential interactions with anticoagulation, it is reasonable to expect that he once again will be able to keep his INR in the therapeutic range.

A patient on anticoagulation for the same indication but who has a history of repeated supratherapeutic levels, poor adherence, or poor access to INR monitoring poses very different concerns about resuming anticoagulation (as well as which agent to use, as we discuss below).

Of note, a high INR alone does not explain bleeding. It is estimated that a workup for gastrointestinal bleeding and gross hematuria uncovers previously undetected lesions in approximately one-third of cases involving warfarin.²⁶ A similar malignancy-unmasking effect is now recognized in patients using the target-specific oral agents who experience gastrointestinal bleeding.⁴⁴ Accordingly, we recommend a comprehensive source evaluation for any anticoagulation-related hemorrhage.

Comorbid conditions

Comorbid conditions associated with bleeding include cancer, end-stage renal disease, liver disease, arterial hypertension, prior stroke, and alcohol abuse.^37,45 Gait instability, regardless of cause, may also increase the risk of trauma-related hemorrhage, but some have estimated that a patient would need to fall multiple times per week to contraindicate anticoagulation on the basis of falls alone.⁴⁶

Concurrent medications

Concomitant therapies, including antiplatelet drugs and nonsteroidal anti-inflammatory drugs, increase bleeding risk.^47,48 Aspirin and the nonsteroidals, in addition to having antiplatelet effects, also can cause gastric erosion.³⁷ In evaluating whether and when to restart anticoagulation, it is advisable to review the role that concomitant therapies may have had in the index bleeding event and to evaluate the risks and benefits of these other agents.

The factor that most consistently predicts major bleeding is a history of bleeding, particularly gastrointestinal bleeding

Additionally, warfarin has many interactions. Although the newer drugs are lauded for having fewer interactions, they are not completely free of them, and the potential for interactions must always be reviewed.⁴⁹ Further, unlike warfarin therapy, therapy with the newer agents is not routinely monitored with laboratory tests, so toxicity (or underdosing) may not be recognized until an adverse clinical event occurs. Ultimately, it may be safer to resume anticoagulation after a contributing drug can be safely discontinued.

Advanced age

The influence that the patient’s age should have on the decision to restart anticoagulation is unclear. Although the risk of intracranial hemorrhage increases with age, particularly after age 80, limited data exist in this population, particularly with regard to rebleeding. Further, age is a major risk factor for most thrombotic events, including venous thromboembolism and stroke from atrial fibrillation, so although the risks of anticoagulation may be higher, the benefits may also be higher than in younger patients.^37,46 We discourage using age alone as a reason to withhold anticoagulation after a hemorrhage.

HOW LONG SHOULD WE WAIT TO RESTART ANTICOAGULATION?

We lack conclusive data on how long to wait to restart anticoagulation after an anticoagulation-associated hemorrhage.

The decision is complicated by evidence suggesting a rebound effect, with an increased risk of pulmonary embolism and atrial fibrillation-related stroke during the first 90 days of interruption of therapy with warfarin as well as with target-specific oral anticoagulants.^3–8 In anticoagulation-associated retroperitoneal bleeding, there is increased risk of deep vein thrombosis from compression, even if venous thromboembolism was not the initial indication for anticoagulation.³⁰

In patients with intracranial hemorrhage, evidence suggests that the intracranial hemorrhage itself increases the risk of arterial and venous thromboembolic events. Irrespective of whether a patient was previously on anticoagulation, the risk of arterial and venous thromboembolic events approaches 7% during the initial intracranial hemorrhage-related hospitalization and 9% during the first 90 days.^34,50,51

To date, the only information we have about when to resume anticoagulation comes from patients taking vitamin K antagonists.

Timing after gastrointestinal bleeding

Small case series suggest that in the first 2 months after warfarin-associated gastrointestinal bleeding, there is substantial risk of rebleeding when anticoagulation is resumed—and of thrombosis when it is not.^52,53 Two retrospective cohort studies may provide some guidance in this dilemma.^28,54

A workup for GI bleeding and gross hematuria uncovers previously undetected lesions in about one-third of cases involving warfarin

Witt et al²⁸ followed 442 patients who presented with gastrointestinal bleeding from any site during warfarin therapy for varied indications for up to 90 days after the index bleeding event. The risk of death was three times lower in patients who restarted warfarin than in those who did not, and their rate of thrombotic events was 10 times lower. The risk of recurrent gastrointestinal bleeding was statistically insignificant, and there were no fatal bleeding events. Anticoagulant therapy was generally resumed within 1 week of the bleeding event, at a median of 4 days.^28,55

Qureshi et al⁵⁴ performed a retrospective cohort study of 1,329 patients with  nonvalvular atrial fibrillation who had experienced a gastrointestinal hemorrhage while taking warfarin. They found that resuming warfarin after 7 days was not associated with a higher risk of recurrent gastrointestinal bleeding and that the rates of death and thromboembolism were lower than in patients who resumed warfarin after 30 days. On the other hand, the risk of recurrent gastrointestinal bleeding was significantly greater if therapy was resumed within the first week.

In view of these studies, we believe that most patients should resume anticoagulation after 4 to 7 days of interruption after gastrointestinal bleeding.⁵⁵

Timing after soft-tissue hemorrhage

The literature on resuming anticoagulation after soft-tissue hemorrhage is sparse. A retrospective study⁵² looked at this question in patients with spontaneous rectal sheath hematoma who had been receiving  antiplatelet drugs, intravenous heparin, vitamin K antagonists, or a combination of these, but not target-specific agents. More than half of the patients were on vitamin K antagonists at the time of hemorrhage. Analysis suggested that when benefits of resuming anticoagulation are believed to outweigh risks, it is reasonable to resume anticoagulation 4 days after the index event.⁵⁶

Timing after intracranial hemorrhage

Anticoagulation should not be considered within the first 24 hours after intracranial hemorrhage, as over 70% of patients develop some amount of hematoma expansion during this time.^34,57 The period thereafter poses a challenge, as the risk of hematoma expansion decreases while the risk of arterial and venous thromboembolism is ongoing and cumulative.⁵⁰

Perhaps surprisingly, national guidelines suggest starting prophylactic-dosed anticoagulation early in all intracranial hemorrhage patients, including those not previously on warfarin.^58,59 In a randomized trial, Boeer et al⁶⁰ concluded that starting low-dose subcutaneous heparin the day after an intracranial hemorrhage decreased the risk of thromboembolism without increasing the risk of rebleeding.⁶⁰ Dickmann et al⁶¹ similarly concluded that there was no increased risk of rebleeding with early prophylactic-dosed subcutaneous heparin.⁶¹ Optimal mechanical thromboprophylaxis, including graduated compression stockings and intermittent pneumatic compression stockings, is also encouraged.³⁴

We discourage using age alone as a reason to withhold anticoagulation after a hemorrhage

Expert opinion remains divided on when and if anticoagulants should be resumed.^34,62 The American Heart Association suggests that in nonvalvular atrial fibrillation, long-term anticoagulation should be avoided after spontaneous lobar hemorrhage; antiplatelet agents can be considered instead.⁵⁸ In nonlobar hemorrhage, the American Heart Association suggests that anticoagulation be considered, depending on strength of indication, 7 to 10 days after the onset.⁵⁸ The European Stroke Initiative suggests patients with strong indications for anticoagulation be restarted on warfarin 10 to 14 days after the event, depending on the risk of thromboembolism and recurrent intracranial hemorrhage.⁵⁹ Others suggest delaying resumption to 10 to 30 weeks after an index intracranial hemorrhage.⁶³

Overall, in the immediate acute period of intracranial hemorrhage, most patients will likely benefit from acute reversal of anticoagulation, followed by institution of prophylactic-dose anticoagulation after the first 24 hours. Going forward, patients who remain at higher risk of a recurrence of anticoagulant-related intracranial hemorrhage (such as those with lobar hemorrhage, suspected cerebral amyloid angiopathy, and other high-risk factors) than of thromboembolic events may be best managed without anticoagulants. Alternatively, patients with deep hemispheric intracranial hemorrhage, hypertension that can be well controlled, and a high risk of serious thromboembolism may experience net benefit from restarting anticoagulation.³⁴

We recommend considering restarting anticoagulation 7 days after the onset of intracranial hemorrhage in patients at high risk of thromboembolism and after at least 14 days for patients at lower risk (Table 2). Discussions with neurologic and neurosurgical consultants should also inform this timing decision.

WOULD A NEWER DRUG BE A BETTER CHOICE?

The emergence of target-specific oral anticoagulants, including factor Xa inhibitors such as rivaroxaban, apixaban, and edoxaban and the direct thrombin inhibitor dabigatran etexilate, presents further challenges in managing anticoagulation after hemorrhage. Table 4 summarizes the current FDA-approved indications.^64–67

These newer agents are attractive because, compared with warfarin, they have wider therapeutic windows, faster onset and offset of action, and fewer drug and food interactions.⁶⁸ A meta-analysis of data available to date suggests that the new drugs, compared with warfarin, show a favorable risk-benefit profile with reductions in stroke, intracranial hemorrhage, and mortality with similar overall major bleeding rates, except for a possible increase in gastrointestinal bleeding.⁶⁸

However, when managing anticoagulation after a bleeding event, the newer agents are challenging for two reasons: they may be associated with a higher incidence of gastrointestinal bleeding than warfarin, and they lack the typical reversal agents that can be used to manage an acute bleeding event.^68,69

In individual studies comparing warfarin with dabigatran,⁷⁰ rivaroxaban,⁷¹ apixaban,⁷² or edoxaban⁷³ for stroke prevention in patients with atrial fibrillation, there was no significant difference in the rate of major bleeding between dabigatran in its higher dose (150 mg twice a day) or rivaroxaban compared with warfarin.^70,71 The risk of major bleeding was actually lower with apixaban⁷² and edoxaban.⁷³

In regard to specific types of major bleeding, the rate of intracranial hemorrhage was significantly lower with dabigatran, rivaroxaban, apixaban, and edoxaban than with warfarin.^35,68–73 Some have proposed that since the brain is high in tissue factor, inhibition of tissue factor-factor VIIa complexes by vitamin K antagonists leaves the brain vulnerable to hemorrhage. Others suggest that the targeted mechanism of target-specific agents, as opposed to the multiple pathways in both the intrinsic and extrinsic coagulation cascade that vitamin K antagonists affect, may explain this difference.^35,74,75

However, some studies suggest that rivaroxaban and the higher doses of dabigatran and edoxaban are associated with higher rates of major gastrointestinal bleeding compared with warfarin.^69–71,76 But apixaban demonstrated no significant difference in gastrointestinal bleeding, and instead demonstrated rates of gastrointestinal bleeding comparable to that with aspirin for stroke prevention in atrial fibrillation.⁷²

The new oral anticoagulants lack antidotes or reversal agents such as phytonadione and fresh-frozen plasma that are available to manage warfarin-associated bleeding events. Other proposed reversal options for the new agents include activated charcoal (if the drugs were taken recently enough to remain in the gastrointestinal tract) and concentrated clotting factor product, though research is ongoing in regards to the most appropriate use in clinical practice.^37,69 Unlike rivaroxaban and apixaban, dabigatran has low plasma protein binding and is dialyzable, which provides another strategy in managing dabigatran-related bleeding.⁶⁹

We believe most patients should resume anticoagulation after 4 to 7 days of interruption after GI bleeding

Of note, the above bleeding risk calculations relate to the first anticoagulant-related bleeding event, though presumably the same risk comparison across agents may be applicable to rebleeding events. Given the data above, when anticoagulation is to be resumed after an intracranial hemorrhage, the risk of rebleeding, particularly in the form of recurrent intracranial hemorrhage, may be lower if a target-specific oral anticoagulant is used.⁷⁵ Similarly, when anticoagulation is to be resumed after a gastrointestinal bleeding event, reinitiation with warfarin or apixaban therapy may present the lowest risk of recurrent gastrointestinal rebleeding. In other sources of bleeding, such as retroperitoneal bleeding, we suggest consideration of transitioning to warfarin, given the availability of reversal agents in the event of recurrent bleeding.

Other important drug-specific factors that must be noted when selecting an agent with which to resume anticoagulation after a hemorrhage include the following:

• In patients with significant renal impairment, the choice of agent will be limited to a vitamin K antagonist.⁷⁷
• A meta-analysis of randomized clinical trials suggests that in the elderly (age 75 and older) target-specific oral anticoagulants did not cause excess bleeding and were associated with at least equal efficacy compared with vitamin K antagonists.⁷⁸
• Target-specific oral anticoagulants may be beneficial in patients who have challenges in achieving INR targets, as evidence suggests that switching to them is associated with a reduction in bleeding for patients who struggle to maintain an appropriately therapeutic INR.⁶⁸ On the other hand, if there is concern that a patient may occasionally miss doses of an anticoagulant, given the rapid onset and offset of action of target-specific agents compared with warfarin, a missed dose of a target-specific agent may result in faster dissolution of anticoagulant effect and increased risk of thrombotic events, and lapses in anticoagulation will not be identified by routine drug monitoring.^6–8,75 As such, it is vital to have a frank discussion with any patient who has difficulty maintaining therapeutic INRs on warfarin treatment to make sure that he or she is not missing doses.
• If there is no clear and compelling reason to select a particular agent, cost considerations should be taken into account. We have included estimated 30-day pricing for the various agents in Table 4. 

If a patient receiving anticoagulant therapy suffers a bleeding event, the patient and physician must decide whether and how soon to restart the therapy, and with what agent.

Foremost on our minds tends to be the risk of another hemorrhage. Subtler to appreciate immediately after an event is the continued risk of thrombosis, often from the same medical condition that prompted anticoagulation therapy in the first place (Table 1).

Complicating the decision, there may be a rebound effect: some thrombotic events such as pulmonary embolism and atrial fibrillation-related stroke may be more likely to occur in the first weeks after stopping warfarin than during similar intervals in patients who have not been taking it.^1–3 The same thing may happen with the newer, target-specific oral anticoagulants.^4–6

Although we have evidence-based guidelines for initiating and managing anticoagulant therapy, ample data on adverse events, and protocols for reversing anticoagulation if bleeding occurs, we do not have clear guidelines on restarting anticoagulation after a hemorrhagic event.

In this article, we outline a practical framework for approaching this clinical dilemma. Used in conjunction with consideration of a patient’s values and preferences as well as input from experts, this framework can help clinicians guide their patients through this challenging clinical decision. It consists of five questions:

• Why is the patient on anticoagulation, and what is the risk of thromboembolism without it?
• What was the clinical impact of the hemorrhage, and what is the risk of rebleeding if anticoagulation is resumed?
• What additional patient factors should be taken into consideration?
• How long should we wait before restarting anticoagulation?
• Would a newer drug be a better choice?

BLEEDING OCCURS IN 2% TO 3% OF PATIENTS PER YEAR

Most of our information on anticoagulation is about vitamin K antagonists—principally warfarin, in use since the 1950s. Among patients taking warfarin outside of clinical trials, the risk of major bleeding is estimated at 2% to 3% per year.⁷

However, the target-specific oral anticoagulants rivaroxaban (Xarelto), apixaban (Eliquis),  dabigatran (Pradaxa) and edoxaban (Savaysa) are being used more and more, and we include them in our discussion insofar as we have information on them. The rates of bleeding with these new drugs in clinical trials have been comparable to or lower than those with warfarin.⁸ Postmarketing surveillance is under way.

WHY IS THE PATIENT ON ANTICOAGULATION? WHAT IS THE RISK WITHOUT IT?

Common, evidence-based indications for anticoagulation are to prevent complications in patients with venous thromboembolism and to prevent stroke in patients with atrial fibrillation or a mechanical heart valve. Other uses, such as in heart failure and its sequelae, pulmonary hypertension, and splanchnic or hepatic vein thrombosis, have less robust evidence to support them.

When anticoagulation-related bleeding occurs, it is essential to review why the patient is taking the drug and the risk of thromboembolism without it. Some indications pose a higher risk of thromboembolism than others and so argue more strongly for continuing the treatment.

Douketis et al⁹ developed a risk-stratification scheme for perioperative thromboembolism. We have modified it by adding the CHA₂DS₂-VASc score (Table 2),^9–11 and believe it can be used more widely.

High-risk indications

Conditions that pose a high risk of thrombosis almost always require restarting anticoagulation. Here, the most appropriate question nearly always is not if anticoagulation should be restarted, but when. Examples:

• A mechanical mitral valve
• Antiphospholipid antibody syndrome with recurrent thromboembolic events.

Lower-risk indications

Lower-risk indications allow more leeway in determining if anticoagulation should be resumed. The most straightforward cases fall well within established guidelines. Examples:

• Atrial fibrillation and a CHA₂DS₂-VASc score of 1. The 2014 guidelines from the American College of Cardiology, American Heart Association, and Heart Rhythm Society¹⁰ suggest that patients with nonvalvular atrial fibrillation and a CHA₂DS₂-VASc score of 1 have three options: an oral anticoagulant, aspirin, and no antithrombotic therapy. If such a patient on anticoagulant therapy subsequently experiences a major gastrointestinal hemorrhage requiring transfusion and intensive care and no definitively treatable source of bleeding is found on endoscopy, one can argue that the risks of continued anticoagulation (recurrent bleeding) now exceed the benefits and that the patient would be better served by aspirin or even no antithrombotic therapy.

• After 6 months of anticoagulation for unprovoked deep vein thrombosis. Several studies showed that aspirin reduced the risk of recurrent venous thromboembolism in patients who completed an initial 6-month course of anticoagulation.^12–15 Though these studies did not specifically compare aspirin with warfarin or target-specific oral anticoagulants in preventing recurrent venous thromboembolism after a hemorrhage, it is reasonable to extrapolate their results to this situation.

If the risk of recurrent hemorrhage on anticoagulation is considered to be too great, then aspirin is an alternative to no anticoagulation, as it reduces the risk of recurrent venous thromboembolism.¹⁶ However, we advise caution if the bleeding lesion may be specifically exacerbated by aspirin, particularly upper gastrointestinal ulcers.

Moderate-risk indications

• After a partial course of anticoagulation for provoked venous thromboembolism. Suppose a patient in the 10th week of a planned 12-week course of anticoagulation for a surgically provoked, first deep vein thrombosis presents with abdominal pain and is found to have a retroperitoneal hematoma. In light of the risk of recurrent bleeding vs the benefit of resuming anticoagulation for the limited remaining period, her 12-week treatment course can reasonably be shortened to 10 weeks.

The risk of recurrent venous thromboembolism when a patient is off anticoagulation decreases with time from the initial event. The highest risk, estimated at 0.3% to 1.3% per day, is in the first 4 weeks, falling to 0.03% to 0.2% per day in weeks 5 through 12, and 0.05% per day thereafter.^17–20

The risk of recurrent venous thromboembolism is greatest immediately after the event and decreases over time

Additionally, a pooled analysis of seven randomized trials suggests that patients with isolated, distal deep vein thrombosis provoked by a temporary risk factor did not have a high risk of recurrence after being treated for 4 to 6 weeks.²¹ These analyses are based on vitamin K antagonists, though it seems reasonable to extrapolate this information to the target-specific oral anticoagulants.

More challenging are situations in which the evidence supporting the initial or continued need for anticoagulation is less robust, such as in heart failure, pulmonary hypertension, or splanchnic and hepatic vein thrombosis. In these cases, the lack of strong evidence supporting the use of anticoagulation should make us hesitate to resume it after bleeding.

WHAT WAS THE CLINICAL IMPACT? WHAT IS THE RISK OF REBLEEDING?

Different groups have defined major and minor bleeding in different ways.^22,23 Several have proposed criteria to standardize how bleeding events (on warfarin and otherwise) are classified,^23–25 but the definitions differ.

Specifically, all agree that a “major” bleeding event is one that is fatal, involves bleeding into a major organ, or leads to a substantial decline in hemoglobin level. However, the Thrombolysis in Myocardial Infarction trials use a decline of more than 5 g/dL in their definition,^23,25 while the International Society on Thrombosis and Haemostasis uses 2 g/dL.²⁴

Here, we review the clinical impact of the most common sources of anticoagulation-related hemorrhage—gastrointestinal, soft tissue, and urinary tract²⁶—as well as intracerebral hemorrhage, a less common but more uniformly devastating event.²⁷

Clinical impact of gastrointestinal hemorrhage

Each year, about 4.5% of patients taking warfarin have a gastrointestinal hemorrhage, though not all of these events are major.²⁸ Evolving data suggest that the newer agents (particularly dabigatran, rivaroxaban, and edoxaban) pose a higher risk of gastrointestinal bleeding than warfarin.²⁹ Patients may need plasma and blood transfusions and intravenous phytonadione, all of which carry risks, albeit small.

Frequently, endoscopy is needed to find the source of bleeding and to control it. If this does not work, angiographic intervention to infuse vasoconstrictors or embolic coils into the culprit artery may be required, and some patients need surgery. Each intervention carries its own risk.

Clinical impact of soft-tissue hemorrhage

Soft-tissue hemorrhage accounts for more than 20% of warfarin-related bleeding events²⁶; as yet, we know of no data on the rate with the new drugs. Soft-tissue hemorrhage is often localized to the large muscles of the retroperitoneum and legs. Though retroperitoneal hemorrhage accounts for a relatively small portion of soft-tissue hemorrhages, it is associated with high rates of morbidity and death and will therefore be our focus.²⁶

Some indications for anticoagulation pose a higher risk of thromboembolism than others

Much of the clinical impact of retroperitoneal hemorrhage is from a mass effect that causes abdominal compartment syndrome, hydroureter, ileus, abscess formation, and acute and chronic pain. At least 20% of cases are associated with femoral neuropathy. It can also lead to deep vein thrombosis from venous compression, coupled with hypercoagulability in response to bleeding. Brisk bleeding can lead to shock and death, and the mortality rate in retroperitoneal hemorrhage is estimated at 20% or higher.³⁰

In many cases, the retroperitoneal hemorrhage will self-tamponade and the blood will be reabsorbed once the bleeding has stopped, but uncontrolled bleeding may require surgical or angiographic intervention.³⁰

Clinical impact of urinary tract hemorrhage

Gross or microscopic hematuria can be found in an estimated 2% to 24% of patients taking warfarin^31–33; data are lacking for the target-specific oral anticoagulants. Interventions required to manage urinary tract bleeding include bladder irrigation and, less often, transfusion.³¹ Since a significant number of cases of hematuria are due to neoplastic disease,³² a diagnostic workup with radiographic imaging of the upper tract and cystoscopy of the lower tract is usually required.³¹ While life-threatening hemorrhage is uncommon, complications such as transient urinary obstruction from clots may occur.

Clinical impact of intracranial hemorrhage

Intracranial hemorrhage is the most feared and deadly of the bleeding complications of anticoagulation. The incidence in patients on warfarin is estimated at 2% to 3% per year, which is markedly higher than the estimated incidence of 25 per 100,000 person-years in the general population.³⁴ Emerging data indicate that the newer drugs are also associated with a risk of intracranial hemorrhage, though the risk is about half that with vitamin K antagonists.³⁵ Intracranial hemorrhage leads to death or disability in 76% of cases, compared with 3% of cases of bleeding from the gastrointestinal or urinary tract.²⁷

Regardless of the source of bleeding, hospitalization is likely to be required and may be  prolonged, with attendant risks of nosocomial harms such as infection.

Risk of rebleeding

Given the scope and severity of anticoagulation-related bleeding, there is strong interest in predicting and preventing it. By some estimates, the incidence of recurrent bleeding after resuming vitamin K antagonists is 8% to 13%.²² Although there are several indices for predicting the risk of major bleeding when starting anticoagulation, there are currently no validated tools to estimate a patient’s risk of rebleeding.³⁶

The patient factor that most consistently predicts major bleeding is a history of bleeding, particularly from the gastrointestinal tract. Finding and controlling the source of bleeding is important.^26,37 For example, a patient with gross hematuria who is found on cystoscopy to have a urothelial papilloma is unlikely to have rebleeding if the tumor is successfully resected and serial follow-up shows no regrowth. In contrast, consider a patient with a major gastrointestinal hemorrhage, the source of which remains elusive after upper, lower, and capsule endoscopy or, alternatively, is suspected to be from one of multiple angiodysplastic lesions. Without definitive source management, this patient faces a high risk of rebleeding.

With or without anticoagulation, after a first intracranial hemorrhage the risk of another one is estimated at 2% to 4% per year.³⁴ An observational study found a recurrence rate of 7.5% when vitamin K antagonist therapy was started after an intracranial hemorrhage (though not all patients were on a vitamin K antagonist at the time of the first hemorrhage).³⁸

Evolving data suggest the newer oral agents pose a higher risk of GI bleeding

Patients with lobar hemorrhage and those with suspected cerebral amyloid angiopathy may be at particularly high risk if anticoagulation is resumed. Conversely, initial events attributed to uncontrolled hypertension that subsequently can be well controlled may portend a lower risk of rebleeding.³⁴ For other types of intracranial hemorrhage, recurrence rates can be even higher. Irrespective of anticoagulation, one prospective study estimated the crude annual rebleeding rate with untreated arteriovenous malformations to be 7%.³⁹ In chronic subdural hematoma, the recurrence rate after initial drainage has been estimated at 9.2% to 26.5%, with use of anticoagulants (in this case, vitamin K antagonists) being an independent predictor of recurrence.⁴⁰

WHAT OTHER PATIENT FACTORS NEED CONSIDERATION?

Target INR on warfarin

An important factor influencing the risk of bleeding with warfarin is the intensity of this therapy.³⁷ A meta-analysis⁴¹ found that the risks of major hemorrhage and thromboembolism are minimized if the goal international normalized ratio (INR) is 2.0 to 3.0. When considering resuming anticoagulation after bleeding, make sure the therapeutic target is appropriate.³⁷

Table 3 summarizes recommended therapeutic ranges for frequently encountered indications for warfarin.^36,42,43

INR at time of the event and challenges in controlling it

The decision to resume anticoagulation in patients who bled while using warfarin must take into account the actual INR at the time of the event.

For example, consider a patient whose INR values are consistently in the therapeutic range.  While on vacation, he receives ciprofloxacin for acute prostatitis from an urgent care team, and no adjustment to INR monitoring or warfarin dose is made. Several days later, he presents with lower gastrointestinal bleeding. His INR is 8, and colonoscopy reveals diverticulosis with a bleeding vessel, responsive to endoscopic therapy. After controlling the source of bleeding and reinforcing the need to always review new medications for potential interactions with anticoagulation, it is reasonable to expect that he once again will be able to keep his INR in the therapeutic range.

A patient on anticoagulation for the same indication but who has a history of repeated supratherapeutic levels, poor adherence, or poor access to INR monitoring poses very different concerns about resuming anticoagulation (as well as which agent to use, as we discuss below).

Of note, a high INR alone does not explain bleeding. It is estimated that a workup for gastrointestinal bleeding and gross hematuria uncovers previously undetected lesions in approximately one-third of cases involving warfarin.²⁶ A similar malignancy-unmasking effect is now recognized in patients using the target-specific oral agents who experience gastrointestinal bleeding.⁴⁴ Accordingly, we recommend a comprehensive source evaluation for any anticoagulation-related hemorrhage.

Comorbid conditions

Comorbid conditions associated with bleeding include cancer, end-stage renal disease, liver disease, arterial hypertension, prior stroke, and alcohol abuse.^37,45 Gait instability, regardless of cause, may also increase the risk of trauma-related hemorrhage, but some have estimated that a patient would need to fall multiple times per week to contraindicate anticoagulation on the basis of falls alone.⁴⁶

Concurrent medications

Concomitant therapies, including antiplatelet drugs and nonsteroidal anti-inflammatory drugs, increase bleeding risk.^47,48 Aspirin and the nonsteroidals, in addition to having antiplatelet effects, also can cause gastric erosion.³⁷ In evaluating whether and when to restart anticoagulation, it is advisable to review the role that concomitant therapies may have had in the index bleeding event and to evaluate the risks and benefits of these other agents.

The factor that most consistently predicts major bleeding is a history of bleeding, particularly gastrointestinal bleeding

Additionally, warfarin has many interactions. Although the newer drugs are lauded for having fewer interactions, they are not completely free of them, and the potential for interactions must always be reviewed.⁴⁹ Further, unlike warfarin therapy, therapy with the newer agents is not routinely monitored with laboratory tests, so toxicity (or underdosing) may not be recognized until an adverse clinical event occurs. Ultimately, it may be safer to resume anticoagulation after a contributing drug can be safely discontinued.

Advanced age

The influence that the patient’s age should have on the decision to restart anticoagulation is unclear. Although the risk of intracranial hemorrhage increases with age, particularly after age 80, limited data exist in this population, particularly with regard to rebleeding. Further, age is a major risk factor for most thrombotic events, including venous thromboembolism and stroke from atrial fibrillation, so although the risks of anticoagulation may be higher, the benefits may also be higher than in younger patients.^37,46 We discourage using age alone as a reason to withhold anticoagulation after a hemorrhage.

HOW LONG SHOULD WE WAIT TO RESTART ANTICOAGULATION?

We lack conclusive data on how long to wait to restart anticoagulation after an anticoagulation-associated hemorrhage.

The decision is complicated by evidence suggesting a rebound effect, with an increased risk of pulmonary embolism and atrial fibrillation-related stroke during the first 90 days of interruption of therapy with warfarin as well as with target-specific oral anticoagulants.^3–8 In anticoagulation-associated retroperitoneal bleeding, there is increased risk of deep vein thrombosis from compression, even if venous thromboembolism was not the initial indication for anticoagulation.³⁰

In patients with intracranial hemorrhage, evidence suggests that the intracranial hemorrhage itself increases the risk of arterial and venous thromboembolic events. Irrespective of whether a patient was previously on anticoagulation, the risk of arterial and venous thromboembolic events approaches 7% during the initial intracranial hemorrhage-related hospitalization and 9% during the first 90 days.^34,50,51

To date, the only information we have about when to resume anticoagulation comes from patients taking vitamin K antagonists.

Timing after gastrointestinal bleeding

Small case series suggest that in the first 2 months after warfarin-associated gastrointestinal bleeding, there is substantial risk of rebleeding when anticoagulation is resumed—and of thrombosis when it is not.^52,53 Two retrospective cohort studies may provide some guidance in this dilemma.^28,54

A workup for GI bleeding and gross hematuria uncovers previously undetected lesions in about one-third of cases involving warfarin

Witt et al²⁸ followed 442 patients who presented with gastrointestinal bleeding from any site during warfarin therapy for varied indications for up to 90 days after the index bleeding event. The risk of death was three times lower in patients who restarted warfarin than in those who did not, and their rate of thrombotic events was 10 times lower. The risk of recurrent gastrointestinal bleeding was statistically insignificant, and there were no fatal bleeding events. Anticoagulant therapy was generally resumed within 1 week of the bleeding event, at a median of 4 days.^28,55

Qureshi et al⁵⁴ performed a retrospective cohort study of 1,329 patients with  nonvalvular atrial fibrillation who had experienced a gastrointestinal hemorrhage while taking warfarin. They found that resuming warfarin after 7 days was not associated with a higher risk of recurrent gastrointestinal bleeding and that the rates of death and thromboembolism were lower than in patients who resumed warfarin after 30 days. On the other hand, the risk of recurrent gastrointestinal bleeding was significantly greater if therapy was resumed within the first week.

In view of these studies, we believe that most patients should resume anticoagulation after 4 to 7 days of interruption after gastrointestinal bleeding.⁵⁵

Timing after soft-tissue hemorrhage

The literature on resuming anticoagulation after soft-tissue hemorrhage is sparse. A retrospective study⁵² looked at this question in patients with spontaneous rectal sheath hematoma who had been receiving  antiplatelet drugs, intravenous heparin, vitamin K antagonists, or a combination of these, but not target-specific agents. More than half of the patients were on vitamin K antagonists at the time of hemorrhage. Analysis suggested that when benefits of resuming anticoagulation are believed to outweigh risks, it is reasonable to resume anticoagulation 4 days after the index event.⁵⁶

Timing after intracranial hemorrhage

Anticoagulation should not be considered within the first 24 hours after intracranial hemorrhage, as over 70% of patients develop some amount of hematoma expansion during this time.^34,57 The period thereafter poses a challenge, as the risk of hematoma expansion decreases while the risk of arterial and venous thromboembolism is ongoing and cumulative.⁵⁰

Perhaps surprisingly, national guidelines suggest starting prophylactic-dosed anticoagulation early in all intracranial hemorrhage patients, including those not previously on warfarin.^58,59 In a randomized trial, Boeer et al⁶⁰ concluded that starting low-dose subcutaneous heparin the day after an intracranial hemorrhage decreased the risk of thromboembolism without increasing the risk of rebleeding.⁶⁰ Dickmann et al⁶¹ similarly concluded that there was no increased risk of rebleeding with early prophylactic-dosed subcutaneous heparin.⁶¹ Optimal mechanical thromboprophylaxis, including graduated compression stockings and intermittent pneumatic compression stockings, is also encouraged.³⁴

We discourage using age alone as a reason to withhold anticoagulation after a hemorrhage

Expert opinion remains divided on when and if anticoagulants should be resumed.^34,62 The American Heart Association suggests that in nonvalvular atrial fibrillation, long-term anticoagulation should be avoided after spontaneous lobar hemorrhage; antiplatelet agents can be considered instead.⁵⁸ In nonlobar hemorrhage, the American Heart Association suggests that anticoagulation be considered, depending on strength of indication, 7 to 10 days after the onset.⁵⁸ The European Stroke Initiative suggests patients with strong indications for anticoagulation be restarted on warfarin 10 to 14 days after the event, depending on the risk of thromboembolism and recurrent intracranial hemorrhage.⁵⁹ Others suggest delaying resumption to 10 to 30 weeks after an index intracranial hemorrhage.⁶³

Overall, in the immediate acute period of intracranial hemorrhage, most patients will likely benefit from acute reversal of anticoagulation, followed by institution of prophylactic-dose anticoagulation after the first 24 hours. Going forward, patients who remain at higher risk of a recurrence of anticoagulant-related intracranial hemorrhage (such as those with lobar hemorrhage, suspected cerebral amyloid angiopathy, and other high-risk factors) than of thromboembolic events may be best managed without anticoagulants. Alternatively, patients with deep hemispheric intracranial hemorrhage, hypertension that can be well controlled, and a high risk of serious thromboembolism may experience net benefit from restarting anticoagulation.³⁴

We recommend considering restarting anticoagulation 7 days after the onset of intracranial hemorrhage in patients at high risk of thromboembolism and after at least 14 days for patients at lower risk (Table 2). Discussions with neurologic and neurosurgical consultants should also inform this timing decision.

WOULD A NEWER DRUG BE A BETTER CHOICE?

The emergence of target-specific oral anticoagulants, including factor Xa inhibitors such as rivaroxaban, apixaban, and edoxaban and the direct thrombin inhibitor dabigatran etexilate, presents further challenges in managing anticoagulation after hemorrhage. Table 4 summarizes the current FDA-approved indications.^64–67

These newer agents are attractive because, compared with warfarin, they have wider therapeutic windows, faster onset and offset of action, and fewer drug and food interactions.⁶⁸ A meta-analysis of data available to date suggests that the new drugs, compared with warfarin, show a favorable risk-benefit profile with reductions in stroke, intracranial hemorrhage, and mortality with similar overall major bleeding rates, except for a possible increase in gastrointestinal bleeding.⁶⁸

However, when managing anticoagulation after a bleeding event, the newer agents are challenging for two reasons: they may be associated with a higher incidence of gastrointestinal bleeding than warfarin, and they lack the typical reversal agents that can be used to manage an acute bleeding event.^68,69

In individual studies comparing warfarin with dabigatran,⁷⁰ rivaroxaban,⁷¹ apixaban,⁷² or edoxaban⁷³ for stroke prevention in patients with atrial fibrillation, there was no significant difference in the rate of major bleeding between dabigatran in its higher dose (150 mg twice a day) or rivaroxaban compared with warfarin.^70,71 The risk of major bleeding was actually lower with apixaban⁷² and edoxaban.⁷³

In regard to specific types of major bleeding, the rate of intracranial hemorrhage was significantly lower with dabigatran, rivaroxaban, apixaban, and edoxaban than with warfarin.^35,68–73 Some have proposed that since the brain is high in tissue factor, inhibition of tissue factor-factor VIIa complexes by vitamin K antagonists leaves the brain vulnerable to hemorrhage. Others suggest that the targeted mechanism of target-specific agents, as opposed to the multiple pathways in both the intrinsic and extrinsic coagulation cascade that vitamin K antagonists affect, may explain this difference.^35,74,75

However, some studies suggest that rivaroxaban and the higher doses of dabigatran and edoxaban are associated with higher rates of major gastrointestinal bleeding compared with warfarin.^69–71,76 But apixaban demonstrated no significant difference in gastrointestinal bleeding, and instead demonstrated rates of gastrointestinal bleeding comparable to that with aspirin for stroke prevention in atrial fibrillation.⁷²

The new oral anticoagulants lack antidotes or reversal agents such as phytonadione and fresh-frozen plasma that are available to manage warfarin-associated bleeding events. Other proposed reversal options for the new agents include activated charcoal (if the drugs were taken recently enough to remain in the gastrointestinal tract) and concentrated clotting factor product, though research is ongoing in regards to the most appropriate use in clinical practice.^37,69 Unlike rivaroxaban and apixaban, dabigatran has low plasma protein binding and is dialyzable, which provides another strategy in managing dabigatran-related bleeding.⁶⁹

We believe most patients should resume anticoagulation after 4 to 7 days of interruption after GI bleeding

Of note, the above bleeding risk calculations relate to the first anticoagulant-related bleeding event, though presumably the same risk comparison across agents may be applicable to rebleeding events. Given the data above, when anticoagulation is to be resumed after an intracranial hemorrhage, the risk of rebleeding, particularly in the form of recurrent intracranial hemorrhage, may be lower if a target-specific oral anticoagulant is used.⁷⁵ Similarly, when anticoagulation is to be resumed after a gastrointestinal bleeding event, reinitiation with warfarin or apixaban therapy may present the lowest risk of recurrent gastrointestinal rebleeding. In other sources of bleeding, such as retroperitoneal bleeding, we suggest consideration of transitioning to warfarin, given the availability of reversal agents in the event of recurrent bleeding.

Other important drug-specific factors that must be noted when selecting an agent with which to resume anticoagulation after a hemorrhage include the following:

• In patients with significant renal impairment, the choice of agent will be limited to a vitamin K antagonist.⁷⁷
• A meta-analysis of randomized clinical trials suggests that in the elderly (age 75 and older) target-specific oral anticoagulants did not cause excess bleeding and were associated with at least equal efficacy compared with vitamin K antagonists.⁷⁸
• Target-specific oral anticoagulants may be beneficial in patients who have challenges in achieving INR targets, as evidence suggests that switching to them is associated with a reduction in bleeding for patients who struggle to maintain an appropriately therapeutic INR.⁶⁸ On the other hand, if there is concern that a patient may occasionally miss doses of an anticoagulant, given the rapid onset and offset of action of target-specific agents compared with warfarin, a missed dose of a target-specific agent may result in faster dissolution of anticoagulant effect and increased risk of thrombotic events, and lapses in anticoagulation will not be identified by routine drug monitoring.^6–8,75 As such, it is vital to have a frank discussion with any patient who has difficulty maintaining therapeutic INRs on warfarin treatment to make sure that he or she is not missing doses.
• If there is no clear and compelling reason to select a particular agent, cost considerations should be taken into account. We have included estimated 30-day pricing for the various agents in Table 4. 

A number of studies published in the last few years will likely affect the way we practice medicine in the hospital. Here, we will use a hypothetical case scenario to focus on the issues of anticoagulants, patient safety, quality improvement, critical care, transitions of care, and perioperative medicine.

AN ELDERLY MAN WITH NEW-ONSET ATRIAL FIBRILLATION

P.G. is an 80-year-old man with a history of hypertension and type 2 diabetes mellitus who is admitted with new-onset atrial fibrillation. In the hospital, his heart rate is brought under control with intravenous metoprolol (Lopressor). On discharge, he will be followed by his primary care physician (PCP). He does not have access to an anticoagulation clinic.

1. What are this patient’s options for stroke prevention?

• Aspirin 81 mg daily and clopidogrel (Plavix) 75 mg daily
• Warfarin (Coumadin) with a target international normalized ratio (INR) of 2.0 to 3.0
• Aspirin mg daily by itself
• Dabigatran (Pradaxa) 150 mg daily

A new oral anticoagulant agent

In deciding what type of anticoagulation to give to a patient with atrial fibrillation, it is useful to look at the CHADS₂ score (1 point each for congestive heart failure, hypertension, age 75 or older, and diabetes mellitus; 2 points for prior stroke or transient ischemic attack. This patient has a CHADS₂ score of 3, indicating that he should receive warfarin. An alternative is dabigatran, the first new anticoagulant agent in more than 50 years.

In a multicenter, international trial, Connolly et al¹ randomized 18,113 patients (mean age 71, 64% men) to receive dabigatran 110 mg twice daily, dabigatran 150 mg twice daily, or warfarin with a target INR of 2.0 to 3.0. In this noninferiority trial, dabigatran was given in a blinded manner, but the use of warfarin was open-label. Patients were eligible if they had atrial fibrillation at screening or within the previous 6 months and were at risk of stroke—ie, if they had at least one of the following: a history of stroke or transient ischemic attack, a left ventricular ejection fraction of less than 40%, symptoms of congestive heart failure (New York Heart Association class II or higher), and an age of 75 or older or an age of 65 to 74 with diabetes mellitus, hypertension, or coronary artery disease.

At a mean follow-up of 2 years, the rate of stroke or systolic embolism was 1.69% per year in the warfarin group compared with 1.1% in the higher-dose dabigatran group (relative risk 0.66, 95% confidence interval [CI] 0.53–0.82, P < .001). The rates of major hemorrhage were similar between these two groups. Comparing lower-dose dabigatran and warfarin, the rates of stroke or systolic embolism were not significantly different, but the rate of major bleeding was significantly lower with lower-dose dabigatran.

In a trial in patients with acute venous thromboembolism, Schulman et al² found that dabigatran was not inferior to warfarin in preventing venous thromboembolism.

Guidelines from the American College of Cardiology Foundation and the American Heart Association now endorse dabigatran as an alternative to warfarin for patients with atrial fibrillation.³ However, the guidelines state that it should be reserved for those patients who:

• Do not have a prosthetic heart valve or hemodynamically significant valve disease
• Have good kidney function (dabigatran is cleared by the kidney; the creatinine clearance rate should be greater than 30 mL/min for patients to receive dabigatran 150 mg twice a day, and at least 15 mL/min to receive 75 mg twice a day)
• Do not have severe hepatic dysfunction (which would impair baseline clotting function).

They note that other factors to consider are whether the patient:

• Can comply with the twice-daily dosing required
• Can afford the drug
• Has access to an anticoagulation management program (which would argue in favor of using warfarin).

Dabigatran is not yet approved to prevent venous thromboembolism.

CASE CONTINUED: HE GETS AN INFECTION

P.G. is started on dabigatran 150 mg by mouth twice a day.

While in the hospital he develops shortness of breath and needs intravenous furosemide (Lasix). Because he has bad veins, a percutaneous intravenous central catheter (PICC) line is placed. However, 2 days later, his temperature is 101.5°F, and his systolic blood pressure is 70 mm Hg. He is transferred to the medical intensive care unit (ICU) for treatment of sepsis. The anticoagulant is held, the PICC line is removed, and a new central catheter is inserted.

2. Which of the following directions is incorrect?

• Wash your hands before inserting the catheter. The accompanying nurse is required to directly observe this procedure or, if this step is not observed, to confirm that the physician did it.
• Before inserting the catheter, clean the patient’s skin with chlorhexidine antiseptic.
• Place sterile drapes over the entire patient.
• Wear any mask, hat, gown, and gloves available.
• Put a sterile dressing over the catheter.

A checklist can prevent infections when inserting central catheters

A checklist developed at Johns Hopkins Hospital consists of the five statements above, except for the second to last one—you should wear a sterile mask, hat, gown and gloves. This is important to ensure that sterility is not broken at any point during the procedure.

Pronovost et al⁴ launched a multicenter initiative at 90 ICUs, predominantly in the state of Michigan, to implement interventions to improve staff culture and teamwork and to translate research into practice by increasing the extent to which these five evidence-based recommendations were applied. The mean rate of catheter-related blood stream infections at baseline was 7.7%; this dropped to 2.8% during the implementation period, 2.3% in the first 3 months after implementation, 1.3% in months 16 through 18, and 1.1% in months 34 through 36, demonstrating that the gains from this quality-improvement project were sustainable.

If this intervention and collaborative model were implemented in all ICUs across the United States and if similar success rates were achieved, substantial and sustained reductions could be made in the 82,000 infections, 28,000 deaths, and $2.3 billion in costs attributed to these infections annually.

CASE CONTINUED: HE IS RESUSCITATED

P.G. is started on a 1-L fluid bolus but he remains hypotensive, necessitating a norepinephrine drip. He does well for about 6 hours, but in the middle of the night he develops ventricular tachycardia and ventricular fibrillation, and a code is called. He is successfully resuscitated, but the family is looking for prognostic information.

3. What are P.G.’s chances of surviving and leaving the hospital?

• 5%
• 8%
• 15%
• 23%

A registry of cardiopulmonary resuscitation

Tian et al⁵ evaluated outcomes in the largest registry of cardiopulmonary resuscitation to date. In this analysis, 49,656 adult patients with a first cardiopulmonary arrest occurring in an ICU between January 1, 2000, and August 26, 2008, were evaluated for their outcomes on pressors vs those not on pressors.

Other independent predictors of a lower survival rate were nonwhite race, mechanical ventilation, having three or more immediate causes of cardiopulmonary arrest, age 65 years or older, and cardiopulmonary arrest occurring at night or over the weekend.

Fortunately, for our patient, survival rates were higher for patients with ventricular tachycardia or fibrillation than with other causes of cardiopulmonary arrest: 22.6% for those on pressors (like our patient) and 40.7% for those on no pressors.

CASE CONTINUED: HE RECOVERS AND GOES HOME

P.G. makes a remarkable recovery and is now ready to go home. It is the weekend, and you are unable to schedule a follow-up appointment before his discharge, so you ask him to make an appointment with his PCP.

4. What is the likelihood that P.G. will be readmitted within 1 month?

• 5%
• 12%
• 20%
• 25%
• 30%

The importance of follow-up with a primary care physician

Misky et al,⁶ in a small study, attempted to identify the characteristics and outcomes of discharged patients who lack timely follow-up with a PCP. They prospectively enrolled 65 patients admitted to University of Colorado Hospital, an urban 425-bed tertiary care center, collecting information about patient demographics, diagnosis, payer source, and PCPs. After discharge, they called the patients to determine their PCP follow-up and readmission status. Thirty-day readmission rates and hospital length of stay were compared in patients with and without timely PCP follow-up (ie, within 4 weeks).

Patients lacking timely PCP follow-up were 10 times more likely to be readmitted (odds ratio [OR] = 9.9, P = .04): the rate was 21% in patients lacking timely PCP follow-up vs 3% in patients with timely PCP follow-up, P = .03. Lack of insurance was associated with lower rates of timely PCP follow-up: 29% vs 56% (P = .06), but did not independently increase the readmission rate or length of stay (OR = 1.0, P = .96). Index hospital length of stay was longer in patients lacking timely PCP follow-up: 4.4 days vs 6.3 days, P = 0.11.

Comment. Nearly half of the patients in this study, who were discharged from a large urban academic center, lacked timely follow-up with a PCP, resulting in higher rates of readmission and a nonsignificant trend toward longer length of stay. Timely follow-up is necessary for vulnerable patients.

Since the lack of timely PCP follow-up results in higher readmission rates and possibly a longer length of stay, a PCP appointment at discharge should perhaps be considered a core quality measure. This would be problematic in our American health care system, in which many patients lack health insurance and do not have a PCP.

A MAN UNDERGOING GASTRIC BYPASS SURGERY

A 55-year-old morbidly obese man (body mass index 45 kg/m²) with a history of type 2 diabetes mellitus, chronic renal insufficiency (serum creatinine level 2.1 mg/dL), hypercholesterolemia, and previous stroke is scheduled for gastric bypass surgery. His functional capacity is low, but he is able to do his activities of daily living. He reports having dyspnea on exertion and intermittently at rest, but no chest pain. His medications include insulin, atorvastatin (Lipitor), aspirin, and atenolol (Tenormin). He is afebrile; his blood pressure is 130/80 mm Hg, pulse 75, and oxygen saturation 97% on room air. His baseline electrocardiogram shows no Q waves.

5. Which of the following is an appropriate next step before proceeding to surgery?

• Echocardiography
• Cardiac catheterization
• Dobutamine stress echocardiography or adenosine thallium scanning
• No cardiac testing is necessary before surgery

Wijeysundera et al⁷ performed a retrospective cohort study of patients who underwent elective surgery at acute care hospitals in Ontario, Canada, in the years 1994 through 2004. The aim was to determine the association of noninvasive cardiac stress testing before surgery with survival rates and length of hospital stay. Included were 271,082 patients, of whom 23,991 (8.9%) underwent stress testing less than 6 months before surgery. These patients were matched with 46,120 who did not undergo testing.

One year after surgery, fewer patients who underwent stress testing had died: 1,622 (7.0%) vs 1,738 (7.5%); hazard ratio 0.92, 95% CI 0.86–0.99, P = .03. The number needed to treat (ie, to be tested) to prevent one death was 221. The tested patients also had a shorter mean hospital stay: 8.72 vs 8.96 days, a difference of 0.24 days (95% CI −0.07 to −0.43; P < .001).

However, the elderly patients (ie, older than 66 years) who underwent testing were more likely to be on beta-blockers and statins than those who did not undergo testing, which may be a confounding factor.

Furthermore, the benefit was all in the patients at intermediate or high risk. The authors performed a subgroup analysis, dividing the patients on the basis of their Revised Cardiac Risk Index (RCRI; 1 point each for ischemic heart disease, congestive heart failure, cerebrovascular disease, diabetes, renal insufficiency, and high-risk surgery).⁸ Patients with an RCRI of 0 points (indicating low risk) actually had a higher risk of death with testing than without testing: hazard ratio 1.35 (95% CI 1.03–1.74), number needed to harm 179—ie, for every 179 low-risk patients tested, one excess death occurred. Those with an RCRI of 1 or 2 points (indicating intermediate risk) had a hazard ratio of 0.92 with testing (95% CI 085–0.99), and those with an RCRI of 3 to 6 points (indicating high risk) had a hazard ratio of 0.80 with testing (95% CI 0.67- 0.97; number needed to treat = 38).

Comment. These findings indicate that cardiac stress testing should be done selectively before noncardiac surgery, and primarily for patients at high risk (with an RCRI of 3 or higher) and in some patients at intermediate risk, but not in patients at low risk, in whom it may be harmful. Stress testing may change patient management because a positive stress test allows one to start a beta-blocker or a statin, use more aggressive intraoperative and postoperative care, and identify patients who have indications for revascularization.

A number of studies published in the last few years will likely affect the way we practice medicine in the hospital. Here, we will use a hypothetical case scenario to focus on the issues of anticoagulants, patient safety, quality improvement, critical care, transitions of care, and perioperative medicine.

AN ELDERLY MAN WITH NEW-ONSET ATRIAL FIBRILLATION

P.G. is an 80-year-old man with a history of hypertension and type 2 diabetes mellitus who is admitted with new-onset atrial fibrillation. In the hospital, his heart rate is brought under control with intravenous metoprolol (Lopressor). On discharge, he will be followed by his primary care physician (PCP). He does not have access to an anticoagulation clinic.

1. What are this patient’s options for stroke prevention?

• Aspirin 81 mg daily and clopidogrel (Plavix) 75 mg daily
• Warfarin (Coumadin) with a target international normalized ratio (INR) of 2.0 to 3.0
• Aspirin mg daily by itself
• Dabigatran (Pradaxa) 150 mg daily

A new oral anticoagulant agent

In deciding what type of anticoagulation to give to a patient with atrial fibrillation, it is useful to look at the CHADS₂ score (1 point each for congestive heart failure, hypertension, age 75 or older, and diabetes mellitus; 2 points for prior stroke or transient ischemic attack. This patient has a CHADS₂ score of 3, indicating that he should receive warfarin. An alternative is dabigatran, the first new anticoagulant agent in more than 50 years.

In a multicenter, international trial, Connolly et al¹ randomized 18,113 patients (mean age 71, 64% men) to receive dabigatran 110 mg twice daily, dabigatran 150 mg twice daily, or warfarin with a target INR of 2.0 to 3.0. In this noninferiority trial, dabigatran was given in a blinded manner, but the use of warfarin was open-label. Patients were eligible if they had atrial fibrillation at screening or within the previous 6 months and were at risk of stroke—ie, if they had at least one of the following: a history of stroke or transient ischemic attack, a left ventricular ejection fraction of less than 40%, symptoms of congestive heart failure (New York Heart Association class II or higher), and an age of 75 or older or an age of 65 to 74 with diabetes mellitus, hypertension, or coronary artery disease.

At a mean follow-up of 2 years, the rate of stroke or systolic embolism was 1.69% per year in the warfarin group compared with 1.1% in the higher-dose dabigatran group (relative risk 0.66, 95% confidence interval [CI] 0.53–0.82, P < .001). The rates of major hemorrhage were similar between these two groups. Comparing lower-dose dabigatran and warfarin, the rates of stroke or systolic embolism were not significantly different, but the rate of major bleeding was significantly lower with lower-dose dabigatran.

In a trial in patients with acute venous thromboembolism, Schulman et al² found that dabigatran was not inferior to warfarin in preventing venous thromboembolism.

Guidelines from the American College of Cardiology Foundation and the American Heart Association now endorse dabigatran as an alternative to warfarin for patients with atrial fibrillation.³ However, the guidelines state that it should be reserved for those patients who:

• Do not have a prosthetic heart valve or hemodynamically significant valve disease
• Have good kidney function (dabigatran is cleared by the kidney; the creatinine clearance rate should be greater than 30 mL/min for patients to receive dabigatran 150 mg twice a day, and at least 15 mL/min to receive 75 mg twice a day)
• Do not have severe hepatic dysfunction (which would impair baseline clotting function).

They note that other factors to consider are whether the patient:

• Can comply with the twice-daily dosing required
• Can afford the drug
• Has access to an anticoagulation management program (which would argue in favor of using warfarin).

Dabigatran is not yet approved to prevent venous thromboembolism.

CASE CONTINUED: HE GETS AN INFECTION

P.G. is started on dabigatran 150 mg by mouth twice a day.

While in the hospital he develops shortness of breath and needs intravenous furosemide (Lasix). Because he has bad veins, a percutaneous intravenous central catheter (PICC) line is placed. However, 2 days later, his temperature is 101.5°F, and his systolic blood pressure is 70 mm Hg. He is transferred to the medical intensive care unit (ICU) for treatment of sepsis. The anticoagulant is held, the PICC line is removed, and a new central catheter is inserted.

2. Which of the following directions is incorrect?

• Wash your hands before inserting the catheter. The accompanying nurse is required to directly observe this procedure or, if this step is not observed, to confirm that the physician did it.
• Before inserting the catheter, clean the patient’s skin with chlorhexidine antiseptic.
• Place sterile drapes over the entire patient.
• Wear any mask, hat, gown, and gloves available.
• Put a sterile dressing over the catheter.

A checklist can prevent infections when inserting central catheters

A checklist developed at Johns Hopkins Hospital consists of the five statements above, except for the second to last one—you should wear a sterile mask, hat, gown and gloves. This is important to ensure that sterility is not broken at any point during the procedure.

Pronovost et al⁴ launched a multicenter initiative at 90 ICUs, predominantly in the state of Michigan, to implement interventions to improve staff culture and teamwork and to translate research into practice by increasing the extent to which these five evidence-based recommendations were applied. The mean rate of catheter-related blood stream infections at baseline was 7.7%; this dropped to 2.8% during the implementation period, 2.3% in the first 3 months after implementation, 1.3% in months 16 through 18, and 1.1% in months 34 through 36, demonstrating that the gains from this quality-improvement project were sustainable.

If this intervention and collaborative model were implemented in all ICUs across the United States and if similar success rates were achieved, substantial and sustained reductions could be made in the 82,000 infections, 28,000 deaths, and $2.3 billion in costs attributed to these infections annually.

CASE CONTINUED: HE IS RESUSCITATED

P.G. is started on a 1-L fluid bolus but he remains hypotensive, necessitating a norepinephrine drip. He does well for about 6 hours, but in the middle of the night he develops ventricular tachycardia and ventricular fibrillation, and a code is called. He is successfully resuscitated, but the family is looking for prognostic information.

3. What are P.G.’s chances of surviving and leaving the hospital?

• 5%
• 8%
• 15%
• 23%

A registry of cardiopulmonary resuscitation

Tian et al⁵ evaluated outcomes in the largest registry of cardiopulmonary resuscitation to date. In this analysis, 49,656 adult patients with a first cardiopulmonary arrest occurring in an ICU between January 1, 2000, and August 26, 2008, were evaluated for their outcomes on pressors vs those not on pressors.

Other independent predictors of a lower survival rate were nonwhite race, mechanical ventilation, having three or more immediate causes of cardiopulmonary arrest, age 65 years or older, and cardiopulmonary arrest occurring at night or over the weekend.

Fortunately, for our patient, survival rates were higher for patients with ventricular tachycardia or fibrillation than with other causes of cardiopulmonary arrest: 22.6% for those on pressors (like our patient) and 40.7% for those on no pressors.

CASE CONTINUED: HE RECOVERS AND GOES HOME

P.G. makes a remarkable recovery and is now ready to go home. It is the weekend, and you are unable to schedule a follow-up appointment before his discharge, so you ask him to make an appointment with his PCP.

4. What is the likelihood that P.G. will be readmitted within 1 month?

• 5%
• 12%
• 20%
• 25%
• 30%

The importance of follow-up with a primary care physician

Misky et al,⁶ in a small study, attempted to identify the characteristics and outcomes of discharged patients who lack timely follow-up with a PCP. They prospectively enrolled 65 patients admitted to University of Colorado Hospital, an urban 425-bed tertiary care center, collecting information about patient demographics, diagnosis, payer source, and PCPs. After discharge, they called the patients to determine their PCP follow-up and readmission status. Thirty-day readmission rates and hospital length of stay were compared in patients with and without timely PCP follow-up (ie, within 4 weeks).

Patients lacking timely PCP follow-up were 10 times more likely to be readmitted (odds ratio [OR] = 9.9, P = .04): the rate was 21% in patients lacking timely PCP follow-up vs 3% in patients with timely PCP follow-up, P = .03. Lack of insurance was associated with lower rates of timely PCP follow-up: 29% vs 56% (P = .06), but did not independently increase the readmission rate or length of stay (OR = 1.0, P = .96). Index hospital length of stay was longer in patients lacking timely PCP follow-up: 4.4 days vs 6.3 days, P = 0.11.

Comment. Nearly half of the patients in this study, who were discharged from a large urban academic center, lacked timely follow-up with a PCP, resulting in higher rates of readmission and a nonsignificant trend toward longer length of stay. Timely follow-up is necessary for vulnerable patients.

Since the lack of timely PCP follow-up results in higher readmission rates and possibly a longer length of stay, a PCP appointment at discharge should perhaps be considered a core quality measure. This would be problematic in our American health care system, in which many patients lack health insurance and do not have a PCP.

A MAN UNDERGOING GASTRIC BYPASS SURGERY

A 55-year-old morbidly obese man (body mass index 45 kg/m²) with a history of type 2 diabetes mellitus, chronic renal insufficiency (serum creatinine level 2.1 mg/dL), hypercholesterolemia, and previous stroke is scheduled for gastric bypass surgery. His functional capacity is low, but he is able to do his activities of daily living. He reports having dyspnea on exertion and intermittently at rest, but no chest pain. His medications include insulin, atorvastatin (Lipitor), aspirin, and atenolol (Tenormin). He is afebrile; his blood pressure is 130/80 mm Hg, pulse 75, and oxygen saturation 97% on room air. His baseline electrocardiogram shows no Q waves.

5. Which of the following is an appropriate next step before proceeding to surgery?

• Echocardiography
• Cardiac catheterization
• Dobutamine stress echocardiography or adenosine thallium scanning
• No cardiac testing is necessary before surgery

Wijeysundera et al⁷ performed a retrospective cohort study of patients who underwent elective surgery at acute care hospitals in Ontario, Canada, in the years 1994 through 2004. The aim was to determine the association of noninvasive cardiac stress testing before surgery with survival rates and length of hospital stay. Included were 271,082 patients, of whom 23,991 (8.9%) underwent stress testing less than 6 months before surgery. These patients were matched with 46,120 who did not undergo testing.

One year after surgery, fewer patients who underwent stress testing had died: 1,622 (7.0%) vs 1,738 (7.5%); hazard ratio 0.92, 95% CI 0.86–0.99, P = .03. The number needed to treat (ie, to be tested) to prevent one death was 221. The tested patients also had a shorter mean hospital stay: 8.72 vs 8.96 days, a difference of 0.24 days (95% CI −0.07 to −0.43; P < .001).

However, the elderly patients (ie, older than 66 years) who underwent testing were more likely to be on beta-blockers and statins than those who did not undergo testing, which may be a confounding factor.

Furthermore, the benefit was all in the patients at intermediate or high risk. The authors performed a subgroup analysis, dividing the patients on the basis of their Revised Cardiac Risk Index (RCRI; 1 point each for ischemic heart disease, congestive heart failure, cerebrovascular disease, diabetes, renal insufficiency, and high-risk surgery).⁸ Patients with an RCRI of 0 points (indicating low risk) actually had a higher risk of death with testing than without testing: hazard ratio 1.35 (95% CI 1.03–1.74), number needed to harm 179—ie, for every 179 low-risk patients tested, one excess death occurred. Those with an RCRI of 1 or 2 points (indicating intermediate risk) had a hazard ratio of 0.92 with testing (95% CI 085–0.99), and those with an RCRI of 3 to 6 points (indicating high risk) had a hazard ratio of 0.80 with testing (95% CI 0.67- 0.97; number needed to treat = 38).

Comment. These findings indicate that cardiac stress testing should be done selectively before noncardiac surgery, and primarily for patients at high risk (with an RCRI of 3 or higher) and in some patients at intermediate risk, but not in patients at low risk, in whom it may be harmful. Stress testing may change patient management because a positive stress test allows one to start a beta-blocker or a statin, use more aggressive intraoperative and postoperative care, and identify patients who have indications for revascularization.

A number of studies published in the last few years will likely affect the way we practice medicine in the hospital. Here, we will use a hypothetical case scenario to focus on the issues of anticoagulants, patient safety, quality improvement, critical care, transitions of care, and perioperative medicine.

AN ELDERLY MAN WITH NEW-ONSET ATRIAL FIBRILLATION

P.G. is an 80-year-old man with a history of hypertension and type 2 diabetes mellitus who is admitted with new-onset atrial fibrillation. In the hospital, his heart rate is brought under control with intravenous metoprolol (Lopressor). On discharge, he will be followed by his primary care physician (PCP). He does not have access to an anticoagulation clinic.

1. What are this patient’s options for stroke prevention?

• Aspirin 81 mg daily and clopidogrel (Plavix) 75 mg daily
• Warfarin (Coumadin) with a target international normalized ratio (INR) of 2.0 to 3.0
• Aspirin mg daily by itself
• Dabigatran (Pradaxa) 150 mg daily

A new oral anticoagulant agent

In deciding what type of anticoagulation to give to a patient with atrial fibrillation, it is useful to look at the CHADS₂ score (1 point each for congestive heart failure, hypertension, age 75 or older, and diabetes mellitus; 2 points for prior stroke or transient ischemic attack. This patient has a CHADS₂ score of 3, indicating that he should receive warfarin. An alternative is dabigatran, the first new anticoagulant agent in more than 50 years.

In a multicenter, international trial, Connolly et al¹ randomized 18,113 patients (mean age 71, 64% men) to receive dabigatran 110 mg twice daily, dabigatran 150 mg twice daily, or warfarin with a target INR of 2.0 to 3.0. In this noninferiority trial, dabigatran was given in a blinded manner, but the use of warfarin was open-label. Patients were eligible if they had atrial fibrillation at screening or within the previous 6 months and were at risk of stroke—ie, if they had at least one of the following: a history of stroke or transient ischemic attack, a left ventricular ejection fraction of less than 40%, symptoms of congestive heart failure (New York Heart Association class II or higher), and an age of 75 or older or an age of 65 to 74 with diabetes mellitus, hypertension, or coronary artery disease.

At a mean follow-up of 2 years, the rate of stroke or systolic embolism was 1.69% per year in the warfarin group compared with 1.1% in the higher-dose dabigatran group (relative risk 0.66, 95% confidence interval [CI] 0.53–0.82, P < .001). The rates of major hemorrhage were similar between these two groups. Comparing lower-dose dabigatran and warfarin, the rates of stroke or systolic embolism were not significantly different, but the rate of major bleeding was significantly lower with lower-dose dabigatran.

In a trial in patients with acute venous thromboembolism, Schulman et al² found that dabigatran was not inferior to warfarin in preventing venous thromboembolism.

Guidelines from the American College of Cardiology Foundation and the American Heart Association now endorse dabigatran as an alternative to warfarin for patients with atrial fibrillation.³ However, the guidelines state that it should be reserved for those patients who:

• Do not have a prosthetic heart valve or hemodynamically significant valve disease
• Have good kidney function (dabigatran is cleared by the kidney; the creatinine clearance rate should be greater than 30 mL/min for patients to receive dabigatran 150 mg twice a day, and at least 15 mL/min to receive 75 mg twice a day)
• Do not have severe hepatic dysfunction (which would impair baseline clotting function).

They note that other factors to consider are whether the patient:

• Can comply with the twice-daily dosing required
• Can afford the drug
• Has access to an anticoagulation management program (which would argue in favor of using warfarin).

Dabigatran is not yet approved to prevent venous thromboembolism.

CASE CONTINUED: HE GETS AN INFECTION

P.G. is started on dabigatran 150 mg by mouth twice a day.

While in the hospital he develops shortness of breath and needs intravenous furosemide (Lasix). Because he has bad veins, a percutaneous intravenous central catheter (PICC) line is placed. However, 2 days later, his temperature is 101.5°F, and his systolic blood pressure is 70 mm Hg. He is transferred to the medical intensive care unit (ICU) for treatment of sepsis. The anticoagulant is held, the PICC line is removed, and a new central catheter is inserted.

2. Which of the following directions is incorrect?

• Wash your hands before inserting the catheter. The accompanying nurse is required to directly observe this procedure or, if this step is not observed, to confirm that the physician did it.
• Before inserting the catheter, clean the patient’s skin with chlorhexidine antiseptic.
• Place sterile drapes over the entire patient.
• Wear any mask, hat, gown, and gloves available.
• Put a sterile dressing over the catheter.

A checklist can prevent infections when inserting central catheters

A checklist developed at Johns Hopkins Hospital consists of the five statements above, except for the second to last one—you should wear a sterile mask, hat, gown and gloves. This is important to ensure that sterility is not broken at any point during the procedure.

Pronovost et al⁴ launched a multicenter initiative at 90 ICUs, predominantly in the state of Michigan, to implement interventions to improve staff culture and teamwork and to translate research into practice by increasing the extent to which these five evidence-based recommendations were applied. The mean rate of catheter-related blood stream infections at baseline was 7.7%; this dropped to 2.8% during the implementation period, 2.3% in the first 3 months after implementation, 1.3% in months 16 through 18, and 1.1% in months 34 through 36, demonstrating that the gains from this quality-improvement project were sustainable.

If this intervention and collaborative model were implemented in all ICUs across the United States and if similar success rates were achieved, substantial and sustained reductions could be made in the 82,000 infections, 28,000 deaths, and $2.3 billion in costs attributed to these infections annually.

CASE CONTINUED: HE IS RESUSCITATED

P.G. is started on a 1-L fluid bolus but he remains hypotensive, necessitating a norepinephrine drip. He does well for about 6 hours, but in the middle of the night he develops ventricular tachycardia and ventricular fibrillation, and a code is called. He is successfully resuscitated, but the family is looking for prognostic information.

3. What are P.G.’s chances of surviving and leaving the hospital?

• 5%
• 8%
• 15%
• 23%

A registry of cardiopulmonary resuscitation

Tian et al⁵ evaluated outcomes in the largest registry of cardiopulmonary resuscitation to date. In this analysis, 49,656 adult patients with a first cardiopulmonary arrest occurring in an ICU between January 1, 2000, and August 26, 2008, were evaluated for their outcomes on pressors vs those not on pressors.

Other independent predictors of a lower survival rate were nonwhite race, mechanical ventilation, having three or more immediate causes of cardiopulmonary arrest, age 65 years or older, and cardiopulmonary arrest occurring at night or over the weekend.

Fortunately, for our patient, survival rates were higher for patients with ventricular tachycardia or fibrillation than with other causes of cardiopulmonary arrest: 22.6% for those on pressors (like our patient) and 40.7% for those on no pressors.

CASE CONTINUED: HE RECOVERS AND GOES HOME

P.G. makes a remarkable recovery and is now ready to go home. It is the weekend, and you are unable to schedule a follow-up appointment before his discharge, so you ask him to make an appointment with his PCP.

4. What is the likelihood that P.G. will be readmitted within 1 month?

• 5%
• 12%
• 20%
• 25%
• 30%

The importance of follow-up with a primary care physician

Misky et al,⁶ in a small study, attempted to identify the characteristics and outcomes of discharged patients who lack timely follow-up with a PCP. They prospectively enrolled 65 patients admitted to University of Colorado Hospital, an urban 425-bed tertiary care center, collecting information about patient demographics, diagnosis, payer source, and PCPs. After discharge, they called the patients to determine their PCP follow-up and readmission status. Thirty-day readmission rates and hospital length of stay were compared in patients with and without timely PCP follow-up (ie, within 4 weeks).

Patients lacking timely PCP follow-up were 10 times more likely to be readmitted (odds ratio [OR] = 9.9, P = .04): the rate was 21% in patients lacking timely PCP follow-up vs 3% in patients with timely PCP follow-up, P = .03. Lack of insurance was associated with lower rates of timely PCP follow-up: 29% vs 56% (P = .06), but did not independently increase the readmission rate or length of stay (OR = 1.0, P = .96). Index hospital length of stay was longer in patients lacking timely PCP follow-up: 4.4 days vs 6.3 days, P = 0.11.

Comment. Nearly half of the patients in this study, who were discharged from a large urban academic center, lacked timely follow-up with a PCP, resulting in higher rates of readmission and a nonsignificant trend toward longer length of stay. Timely follow-up is necessary for vulnerable patients.

Since the lack of timely PCP follow-up results in higher readmission rates and possibly a longer length of stay, a PCP appointment at discharge should perhaps be considered a core quality measure. This would be problematic in our American health care system, in which many patients lack health insurance and do not have a PCP.

A MAN UNDERGOING GASTRIC BYPASS SURGERY

A 55-year-old morbidly obese man (body mass index 45 kg/m²) with a history of type 2 diabetes mellitus, chronic renal insufficiency (serum creatinine level 2.1 mg/dL), hypercholesterolemia, and previous stroke is scheduled for gastric bypass surgery. His functional capacity is low, but he is able to do his activities of daily living. He reports having dyspnea on exertion and intermittently at rest, but no chest pain. His medications include insulin, atorvastatin (Lipitor), aspirin, and atenolol (Tenormin). He is afebrile; his blood pressure is 130/80 mm Hg, pulse 75, and oxygen saturation 97% on room air. His baseline electrocardiogram shows no Q waves.

5. Which of the following is an appropriate next step before proceeding to surgery?

• Echocardiography
• Cardiac catheterization
• Dobutamine stress echocardiography or adenosine thallium scanning
• No cardiac testing is necessary before surgery

Wijeysundera et al⁷ performed a retrospective cohort study of patients who underwent elective surgery at acute care hospitals in Ontario, Canada, in the years 1994 through 2004. The aim was to determine the association of noninvasive cardiac stress testing before surgery with survival rates and length of hospital stay. Included were 271,082 patients, of whom 23,991 (8.9%) underwent stress testing less than 6 months before surgery. These patients were matched with 46,120 who did not undergo testing.

One year after surgery, fewer patients who underwent stress testing had died: 1,622 (7.0%) vs 1,738 (7.5%); hazard ratio 0.92, 95% CI 0.86–0.99, P = .03. The number needed to treat (ie, to be tested) to prevent one death was 221. The tested patients also had a shorter mean hospital stay: 8.72 vs 8.96 days, a difference of 0.24 days (95% CI −0.07 to −0.43; P < .001).

However, the elderly patients (ie, older than 66 years) who underwent testing were more likely to be on beta-blockers and statins than those who did not undergo testing, which may be a confounding factor.

Furthermore, the benefit was all in the patients at intermediate or high risk. The authors performed a subgroup analysis, dividing the patients on the basis of their Revised Cardiac Risk Index (RCRI; 1 point each for ischemic heart disease, congestive heart failure, cerebrovascular disease, diabetes, renal insufficiency, and high-risk surgery).⁸ Patients with an RCRI of 0 points (indicating low risk) actually had a higher risk of death with testing than without testing: hazard ratio 1.35 (95% CI 1.03–1.74), number needed to harm 179—ie, for every 179 low-risk patients tested, one excess death occurred. Those with an RCRI of 1 or 2 points (indicating intermediate risk) had a hazard ratio of 0.92 with testing (95% CI 085–0.99), and those with an RCRI of 3 to 6 points (indicating high risk) had a hazard ratio of 0.80 with testing (95% CI 0.67- 0.97; number needed to treat = 38).

Comment. These findings indicate that cardiac stress testing should be done selectively before noncardiac surgery, and primarily for patients at high risk (with an RCRI of 3 or higher) and in some patients at intermediate risk, but not in patients at low risk, in whom it may be harmful. Stress testing may change patient management because a positive stress test allows one to start a beta-blocker or a statin, use more aggressive intraoperative and postoperative care, and identify patients who have indications for revascularization.